By Robert Watt
Dennis Schulting’s Kant’s Radical Subjectivism: Perspectives on the Transcendental Deduction (KRS) is full of big ideas. Its central claim is that Kant is a “radical subjectivist about the possibility of knowledge” (p. 35), and that recognising this is crucial if we want to save the Transcendental Deduction (henceforth ‘the Deduction’) from the “standard charges of incoherence, inconsistency, or relativism/scepticism” (p. 22). In the course of defending this claim, Schulting addresses a number of important issues regarding the interpretation of the Deduction, including the alleged logical gap in Kant’s argument, the continuing debate between conceptualists and non-conceptualists, and his own suggestion in his previous book Kant’s Deduction and Apperception: Explaining the Categories (KDA) that the Deduction contains a “logical derivation” of the Categories from self-consciousness.
In the following remarks I am going to focus on three aspects of KRS. The first aspect is the central claim that Kant is a radical subjectivist. In Section I, I use Schulting’s distinction between ‘bad’ subjectivism and radical subjectivism to shed light on the latter, before explaining how the claim that Kant is a radical subjectivist helps Schulting to solve the problem of the alleged logical gap in the Deduction. I also raise two questions. The first is how exactly radical subjectivism is compatible with epistemic humility, and the second is whether the claim that Kant is a radical subjectivist is itself such a radical claim.
The second aspect of KRS that I am going to consider is Schulting’s discussion of the debate between conceptualists and non-conceptualists. In Section II, I endorse Schulting’s claim that while non-conceptualists are right to deny that Kant regards the application of the categories as a necessary condition of empirical intuition, conceptualists are right to insist that he regards the application of the categories as a necessary condition of “genuine perception” (p. 266). What renders these claims compatible is the fact that empirical intuition is not the same thing as genuine perception: genuine perception is objective in a way that empirical intuition is not. I suggest that Schulting’s attempt to explain what the objectivity of genuine perception consists in is the weak spot in his discussion.
The third and final aspect of KRS that I am going to discuss in these remarks is also the aspect of the book that I find least convincing. This is Schulting’s response to objections against his claim in KDA that the Deduction contains a logical derivation of the categories from self-consciousness. In Section III, I outline four reasons why I find this claim so difficult to accept.
What is radical subjectivism about the possibility of knowledge? Schulting contrasts radical subjectivism with bad subjectivism. According to bad subjectivism, “what we know is true relative merely to our own perspective, because that is just how we are psychologically (or culturally, or epistemologically, etc.) disposed” (pp. 16–17).
Bad subjectivism tells us that the only facts we can know are the facts according to our particular psychological/cultural/epistemological framework. We can know, for example, that objects are spatiotemporal according to our framework, but we cannot know that objects are spatiotemporal simpliciter. In other words, we cannot know whether our framework corresponds to reality.
By contrast, radical subjectivism tells us that “subjective agency is first constitutive of objectivity, so that there would not even be anything objective, any nature, without the subjective forms of the understanding” (p. 17). It tells us that “objectivity itself is dependent on our forms of cognition, contrary to the ‘bad’, psychological subjectivism […] which says that the categories (and the forms of intuition) are merely our ways of cognising, but do not reach the real objects themselves” (p. 17). What radical subjectivism denies is that we can conceive of some reality, independent of our framework, with which to compare our framework. The only reality with which we can compare our framework is reality according to our framework. And it is trivial that our framework corresponds to reality according to our framework.
Radical subjectivism agrees with bad subjectivism that the only facts we can know are the facts according to our framework. It agrees that our knowledge cannot outstrip the subjective facts. But whereas bad subjectivism contrasts the subjective facts that we can know with the objective facts that we cannot know, radical subjectivism argues that the objective facts are constituted by the facts according to our framework. And of course we can know that the facts according to our framework are the same as the facts according to our framework. Radical subjectivism argues that the objective facts cannot outstrip the subjective facts.
Here’s how Schulting thinks this applies to Kant. Kant argues that if we want to think of objects, we have to think of them as unities, pluralities and totalities, realities and negations, standing in relations of inherence and causality, etc. He thinks that the categories constitute our conceptual framework. Suppose Kant added to this the sceptical claim that we cannot know whether the objects that we think of as unities, pluralities and totalities, realities and negations, standing in relations of inherence and causality, etc., really are unities, pluralities and totalities, realities and negations, standing in relations of inherence and causality, etc. Suppose he doubted that our categorial framework corresponds to reality. In that case he would count as a bad subjectivist. In fact, however, Kant argues that if we want to think of a reality with which to compare our categorial framework, we have to think of it as a reality consisting of unities, pluralities and totalities, realities and negations, relations of inherence and causality, etc. And of course we can know that our conceptual framework corresponds to this reality. This makes Kant a radical subjectivist.
Why ascribe radical subjectivism to Kant? To my mind, the big payoff is the one that Schulting discusses in Chapter 4 of KRS. It helps us to explain why there isn’t a serious logical gap in the Deduction. Here’s the reason why there appears to be such a gap. Part of the aim of the Deduction is to prove that there are objects instantiating the categories: unities, pluralities and totalities, realities and negations, substances and causes, etc. If it fails to do this, it leaves open the possibility that the categories are “entirely empty” (A90/B122). What Kant actually provides in the Deduction is an argument that the application of the categories is a necessary condition of thinking about an object. If I want to think about an object, I have to think of it as a unity, plurality or a totality, a reality or a negation, standing in relations of inherence or causality, etc. On the assumption that there are objects that I think about, it follows from this that there are objects that I think of as unities, pluralities and totalities, realities and negations, substances and causes, etc. It follows that there are objects that I think instantiate the categories.
But it doesn’t follow that any objects instantiate the categories. So there is a logical gap in the Deduction. The Deduction fails to rule out the possibility that we are subject to what Anil Gomes has described as an “unavoidable error”, whereby we are “compelled, of necessity, to think of the world as containing persisting substances, capable of existing unperceived and standing to each other in causal relations”, in spite of the fact that “none of these judgements about the world [is] accurate” (Gomes 2014:11). The point can also be expressed in terms of a distinction drawn by James Van Cleve (1999:89): Kant is supposed to be proving that the categories must apply to objects, but what he actually proves is that we must apply the categories to objects. These are not the same thing.
How does radical subjectivism help us to address this concern about the Deduction? Well, consider the possibility that the Deduction supposedly fails to rule out. In effect, this is the possibility that our categorial framework fails to correspond to reality. It is the possibility that the objects that we think instantiate the categories do not instantiate the categories. But this presupposes that we can conceive of a reality with which to compare our categorial framework apart from reality according to that framework. While a bad subjectivist would accept this presupposition, a radical subjectivist would not. According to a radical subjectivist, we cannot conceive of the possibility that the objects that we think instantiate the categories do not instantiate the categories. Why? Because this would involve thinking of objects—and, therefore, thinking of objects that instantiate the categories—and then thinking that these same objects do not instantiate the categories. Schulting expresses this point by suggesting that
it is not germane to ask the question whether I can be mistaken about the application of the categories […].
In the case of the categories—which are a priori concepts after all—we are, as it were, immune to error through misapplication, to vary a well-known phrase in the philosophy of mind. Categories either apply or they do not; I cannot be mistaken about my application. (p. 153)
It should be noted that this isn’t the only way of explaining why there isn’t a serious logical gap in the Deduction. Another way is to suggest that there is more to the Deduction than a proof that the application of the categories is a necessary condition of thinking about an object. Philosophers such as Anil Gomes (2010), Hannah Ginsborg (2008), and Aaron Griffith (2010) take Kant to be defending a further claim, which is that the application of the categories is a necessary condition of the empirical intuition of an object. The basic idea is that if the application of the categories is a necessary condition of the empirical intuition of an object, then we have some sort of guarantee that when we think of the objects of empirical intuition as instantiating the categories, we are getting things right.
I find the radical subjectivism explanation more plausible for the following reason. In the A-Preface, Kant makes an extraordinary claim about the Deduction. He writes that insofar as the Deduction is supposed to “demonstrate and make comprehensible the objective validity of […] concepts a priori […] what is said at pages [A] 92–3 should even be sufficient by itself” (Axvi–xvii). Now what Kant actually does at A92–3 is to distinguish the conditions of the possibility of intuition from the conditions of the possibility of thinking about objects, and to raise the possibility that the categories are conditions of the possibility of thinking about objects. He writes that “the objective validity of the categories […] rests on the fact that through them alone is experience possible (as far as the form of thinking is concerned)”, and that the categories “are related necessarily and a priori to objects of experience, since only by means of them can any object of experience be thought at all” (A93/B126). This strongly suggests that, according to Kant, there is no need to show anything about the necessary conditions of empirical intuition to accomplish the aim of the Deduction. It is sufficient to prove that the application of the categories is a necessary condition of thinking about an object. By reading Kant as a radical subjectivist, Schulting can give an account of why Kant thinks that this is sufficient.
There are some aspects of Schulting’s ‘radical subjectivist’ interpretation that call for clarification. The first concerns its compatibility with what Rae Langton calls Kant’s “epistemic humility” (1998:2), which is to say, his claim that we cannot know anything about things as they are in themselves. Here’s the worry. If we read Kant as what Schulting calls a ‘bad’ subjectivist, there is at least a clear way to interpret his epistemic humility. He thinks that we can know the facts according to our framework, that is, we can know things as they appear to us, but he doesn’t think we can know whether our framework corresponds to reality, that is, we cannot know things as they are in themselves. But how should we interpret Kant’s epistemic humility if we read him as a radical subjectivist? According to this interpretation, Kant thinks that there is only one conceivable reality, i.e. reality according to our framework, and this is a reality of which we can have knowledge.
One way to deal with this worry would be to distinguish our conceptual, categorial framework from our sensible, spatiotemporal framework, and to suggest that Kant is only a radical subjectivist with respect to the former. We cannot think of an object without thinking of it as something that instantiates the categories, but we can perfectly well think of an object that isn’t spatiotemporal. Perhaps Kant’s epistemic humility should be interpreted as the claim that although we can know the facts according to our spatiotemporal framework, we cannot know whether or not this framework corresponds to reality. This would be compatible with his claiming that we can know that our categorial framework corresponds to reality. Another way to deal with the worry would be to suggest that while Kant doesn’t think that we can conceive of a reality apart from reality according to our categorial framework, he still thinks we can have some even weaker notion of such a reality. Perhaps he makes a distinction between a ‘positive’ conception of such a reality, which we cannot have, and a ‘negative’ conception of such a reality, which we can have.
It is not entirely clear to me which of these options Schulting would prefer to take. The fact that his initial statement of radical subjectivism refers almost exclusively to the role of the categories and the faculty of understanding creates the space for him to take the first option (pp. 10, 17). But some of his remarks about things in themselves in Chapter 9 of KRS suggest that he would prefer the second option (e.g. “strictly speaking, things in themselves cannot even be thought as such in any objectively significant, determinate sense”, p. 395). Perhaps he would plump for an entirely different way of reconciling Kant’s radical subjectivism with his epistemic humility. One of the things that makes it difficult to get clear about Schulting’s answer to this question is that he doesn’t spend a great deal of time in KRS focusing directly on the idea of radical subjectivism; responding to obvious objections and answering simple questions about how it relates to Kant’s other commitments. I would have found it especially helpful if Schulting had compared radical subjectivism with other more familiar forms of subjectivism such as ethical subjectivism or subjectivism about colour.
I would like to end this section by raising a second question about Schulting’s claim that Kant is a radical subjectivist. Is this itself a radical claim? According to Schulting,
To say that the above picture I painted of Kant’s subjectivism is not popular—neither among more or less orthodox Kantians, nor among those philosophers chiefly critical of Kant, such as Hegel—is an understatement. (p. 13)
But is this really true? At the beginning of The Bounds of Sense, Peter Strawson portrays Kant as investigating “the set of ideas which forms the limiting framework of all our thought about the world and experience of the world” (2006:15), and as arguing that “a certain minimum structure is essential to any conception of experience that we can make truly intelligible to ourselves” (2006:11). It seems to me that something close to the claim that Schulting defends is an implication of Strawson’s picture. Kant wants to argue that in order to make sense of the world, in order to render the world intelligible to ourselves, we have to make use of a particular set of ideas. What can be said of the possibility that these ideas don’t really capture the way the world is? Well, at least this much: it isn’t a possibility that we can really make sense of, or render intelligible to ourselves. It isn’t clear to me how this differs from the radical subjectivism that Schulting ascribes to Kant.
I am now going to turn to Schulting’s discussion of the debate between conceptualists and non-conceptualists, which can be found in Chapters 5–7 of KRS. According to Schulting, non-conceptualists are right to deny that Kant regards the application of the categories as a necessary condition of empirical intuition (p. 226), while conceptualists are right to insist that Kant regards the application of the categories as a necessary condition of what Schulting calls “genuine perception” (p. 266). Where non-conceptualists and conceptualists alike go wrong is in assuming that empirical intuition is the same thing as genuine perception. This is what leads non-conceptualists to conclude that the application of the categories is not a necessary condition of genuine perception, and it is what leads conceptualists to conclude that the application of the categories is a necessary condition of empirical intuition.
This position, which can also be found in the work of other philosophers such as Clinton Tolley (e.g. Tolley, forthcoming), is one with which I have a great deal of sympathy. I agree with Schulting that there is good textual evidence that Kant does not regard the application of the categories as a necessary condition of empirical intuition, and I think Schulting does a good job of dismantling the textual arguments against this claim in Chapter 5 of KRS. I also agree that what Kant means by empirical intuition in the Deduction is a much thinner type of representation than is often assumed. Empirical intuition is not the product of any type of synthesis. Nor is it a conscious representation. It is to be carefully distinguished from what Kant himself calls ‘perception’, which is a conscious representation, that is, a product of synthesis (cf. B160–1). Finally, empirical intuition does not count as the cognition of an object.
However, this interpretation of empirical intuition is not without its difficulties, and this is where I think Schulting needs to do more work. There is a reason so many philosophers have interpreted intuition as a thicker type of representation. Take the famous Stufenleiter passage, for example. Here Kant famously defines an intuition as one of two types of cognition, where cognition is defined as an “objective perception” (A320/B376), and where a perception is defined in turn as a “representation with consciousness” (A320/B376). This passage suggests that an intuition is a thicker sort of representation than Schulting thinks it is. He needs to explain why it doesn’t tell against his interpretation.
It seems to me that there are at least three options. One is to argue that Kant has two conceptions of objectivity and two conceptions of consciousness, and that while there is a sense in which empirical intuition is conscious and objective, there is also a sense in which it is not. The second option is to argue that Kant has two conceptions of intuition, and that the conception of intuition to which he appeals in the Stufenleiter is not the same as the conception that plays such a crucial role in the Deduction. A third option is to argue that Kant has a single conception of intuition, and indeed a single conception of consciousness and objectivity, but thinks that intuition has different properties depending on whether it is considered in isolation from the faculty of understanding. If Schulting prefers the first option, as I do, then he needs to explain as clearly as possible how the consciousness and the objectivity that pertain to empirical intuition differ from the consciousness and the objectivity that do not.
Schulting generally identifies the objectivity that is not a feature of mere intuition with objective validity. He takes non-conceptualists to think that “the mere intuition of […] objects provides us with an objectively valid relation to objects” (p. 21). Now there are places where Kant identifies objective validity with truth (A125), but this raises obvious problems: Kant thinks that all judgements are objectively valid, but he doesn’t think that all judgements are true. On what Schulting recognises as the standard view, a judgement is objectively valid in Kant’s sense just in case it has a truth-value. But Schulting thinks that this must be false since there are judgements that have a truth-value in spite of the fact that they lack objective validity. According to Schulting, “the objective validity of judgement must not be seen in terms of the truth value of judgement, that is, that one’s judgement is either true or false” (p. 116). Objective validity is a “more fundamental feature of judgement than its truth value” (p. 153). A judgement of mine is objectively valid just in case “I am primordially connected with the object of my judgement” (p. 154), which is to say, my judgement is about an object, or has “object-intentionality” (p. 154).
The problem with this interpretation is that it isn’t at all clear that intuitions are not objective in this sense. Even if, as Colin McLear (2016) has argued, intuitions don’t have “intentional content” in the sense of representing an object as being a particular way, don’t they at least have some sort of ‘object-intentionality’? Doesn’t Kant think that they put us into some sort of basic connection with objects? Doesn’t he think that they represent objects to us? Take the famous savage looking at the house, whose representation Kant tells us is “mere intuition” (Log, AA 9:33). Doesn’t the savage’s representation have an intentional object? Doesn’t it put him into a basic connection with the house, in virtue of which he might someday be connected with it in a less basic way? If Schulting thinks that the application of the categories is a necessary condition of this type of objectivity, he should conclude with the conceptualists that the application of the categories is a necessary condition of empirical intuition.
I wonder whether Schulting is not too quick in his rejection of the more widely accepted interpretation of objective validity in terms of having a truth-value. Let’s look in a little more detail at Schulting’s argument against this interpretation. It runs as follows: (1) Analytic judgements have a truth-value. (2) Analytic judgements are not objectively valid. So (3) objective validity is not the same thing as having a truth-value (p. 118). Premise (1) is true. Analytic judgements such as the judgement that bodies are extended are either true or false (cf. e.g. A151/B190). But there is more of a question mark over premise (2). I am not aware of any compelling evidence that analytic judgements are not objectively valid, nor does Schulting provide any. He writes that “an analytic judgement is true or false solely on the basis of the principle of non-contradiction (B190/A151), without objective validity having anything to do with this” (p. 118; cf. also p. 152). He also writes:
Objective validity is not at issue in the context of determining the truth (or falsity) of an analytic judgement, since here the reference to an object (objective reality; B194/A155) is otiose (cf. A258–9); its truth can be determined solely through analysis of subject and predicate of such a judgement. (p. 118; again cf. p. 152)
Here Schulting begins by assuming that a judgement is objectively valid only if it is necessary to do more than analyse the constituent concepts in order to determine its truth-value. He concludes that since it is not necessary to do any more than this to determine the truth-value of an analytic judgement, analytic judgements are not objectively valid. But it is not clear to me what evidence there is to support Schulting’s initial assumption. One thing he might appeal to here is Kant’s claim in the B-Preface that a representation is objectively valid only if it is possible to prove its “real possibility” (Bxxvi), where proving real possibility requires something more than just the analysis of constituent concepts.
Here is an argument against premise (2) of Schulting’s argument. In the Prolegomena, Kant tells us that objective validity and necessary universal validity are “interchangeable concepts” (Prol, AA 4:298). This suggests that a judgement is objectively valid if and only if it is necessarily universally valid. A judgdment is necessarily universally valid in Kant’s sense if and only if we “intend that the judgment should also be valid at all times for us and for everyone else” (Prol, AA 4:298), which is to say, we “require that [we] should find it so at every time”, and that “everyone else should find it just as [we] do” (Prol, AA 4:299). Now suppose I judge that bodies are extended. This is an analytic judgement. And it is also, surely, necessarily universally valid. I require that I should find it so at every time, and that everyone else should as well. But if this judgement is necessarily universally valid, it must also be objectively valid. And this goes for analytic judgements in general. So Schulting is wrong to deny that analytic judgements are objectively valid.
He might object that the key premise of this argument is taken from a controversial passage of the Prolegomena, but there is also evidence in the First Critique that Kant equates objective validity and necessary universal validity. In §19 of the B-Deduction, he suggests that “a relation that is objectively valid” is one where “representations are combined in the object, i.e. regardless of any difference in the condition of the subject” (B142). Now when I judge that bodies are extended, the concepts of body and extension are indeed connected without regard to any difference in the condition of the subject. For Kant, this is enough to make the connection objectively valid.
If Schulting were to drop his opposition to the truth-value interpretation, he could give a far clearer response to my worry about the objectivity of intuition. He could say that there is a sense in which empirical intuition is an objective representation, and a sense in which it is not. It is not an objective representation in the sense that it has a truth-value, or represents an object as being a particular way. This distinguishes it from genuine perception. But once empirical intuition is distinguished from genuine perception, it is perfectly possible to agree with the non-conceptualists that the application of the categories is not a necessary condition of empirical intuition, while at the same time agreeing with the conceptualists that the application of the categories is a necessary condition of genuine perception.
Finally, I want to make a few remarks about Schulting’s defence of the claim that the Deduction contains a “logical derivation” of the categories from self-consciousness (p. 58), which he pursues in Chapter 2 of KRS.
Here is a more familiar story about the derivation of the categories. Kant’s aim in the Metaphysical Deduction is to identify the categories. He thinks that he has found a clue that will prevent him from either missing out any concepts that belong on the list, or including any concepts that don’t belong. The clue concerns the tight connection between the categories and the possible ways of making judgements. All Kant needs to do to identify the categories is to list all of the basic types of judgement, and there will be a category corresponding to each type. As he puts it: “The functions of the understanding can […] all be found together if one can exhaustively exhibit the functions of unity in judgments” (A69/B94). Following this method, Kant first decides that judgements about objects can be classified in four ways, with respect to their quantity, their quality, their relation and their modality, and that there are three possible classifications in each case. With respect to quantity, a judgement may be universal, particular, or singular, with respect to modality, it may be problematic, assertoric, or apodictic, and so on. From the twelve forms of judgement that he has identified, Kant derives the twelve categories: totality from universal judgement, plurality from particular judgement, possibility from problematic judgement, and so on.
According to this familiar story, Kant has an answer to the question: ‘why are these particular concepts the categories?’ The answer is: ‘because they are the concepts corresponding to the forms of judgement’. But Kant doesn’t have an answer to the question: ‘but why are these the forms of judgement?’ He doesn’t give us any deeper insight into the fact that these particular concepts are the categories.
According to Schulting’s story, Kant does go on to give us a deeper insight into this fact. Broadly, the answer runs as follows: If we think about what a judgement consists in, we will see that it involves a special type of self-consciousness, viz. transcendental apperception. If we then think about what this special type of self-consciousness consists in, we will see that it involves each of the twelve concepts that Kant has already identified as the categories. A simple case: The self-consciousness in question is the representation ‘I think’, but this is, among other things, the ascription to myself of thoughts as predicates, and, therefore, the representation of my self as substance. We are now in a position to explain why the twelve types of judgement listed in the Metaphysical Deduction are the forms of judgement. We no longer have to take the fact that these are the forms of judgement as explanatorily basic. And our final answer to the question why the twelve concepts listed in the Metaphysical Deduction are the categories is that these concepts are all involved in the special type of consciousness, viz. transcendental apperception, that is itself involved in every judgement.
Schulting acknowledges that his story about the derivation of the categories “is not going to be the standard view any time soon” (p. 25). But it is a story to which he remains committed, in spite of the objections considered in Chapter 2. I would like to end by explaining why I think this story is so hard to accept.
The first point to note is that some of the passages where Kant indicates that, on his view, the categories can be derived from a single principle are just as compatible with the first, familiar story as they are with Schulting’s story. Here is one such passage, from the Prolegomena:
Nothing can be more desirable to a philosopher than to be able to derive, a priori from one principle, the multiplicity of concepts or basic principles that previously had exhibited themselves to him piecemeal in the use he had made of them in concreto, and in this way to be able to unite them all in one cognition. (AA 4:322, quoted by Schulting on p. 60)
Here is another key passage from the Metaphysical Deduction:
Transcendental philosophy has the advantage but also the obligation to seek its concepts in accordance with a principle, since they spring pure and unmixed from the understanding, as absolute unity, and must therefore be connected among themselves in accordance with a concept or idea. Such a connection, however, provides a rule by means of which the place of each pure concept of the understanding and the completeness of all of them together can be determined a priori, which would otherwise depend upon whim or chance. (A67/B92; quoted by Schulting on p. 60–1)
On the first story, what Kant is saying in these passages is that we can systematically identify the categories—that is, we can identify them in something other than a haphazard way—as soon as we realise that they correspond to the forms of judgement, which we have, of course, already identified. It is the principle that the categories correspond to the forms of judgement that is doing the work in this derivation. There is no reason to think that the principle to which Kant refers in each of these passages is the principle of apperception.
The crucial passage for Schulting’s purposes is the one in the Deduction where Kant writes that the “principles of the objective determination of all representations insofar as cognition can come from them […] are all derived from the principle of the transcendental unity of apperception” (B142). Presumably, on Schulting’s view, we should interpret this as the claim that the categories are derived from transcendental apperception. He assumes that the “principles of the objective determination of all representations insofar as cognition can come from them” just are the categories.
But it seems to me that there is another reasonable interpretation of the crucial passage. According to this interpretation, what Kant means by the principle of the transcendental unity of apperception is what he describes in an earlier section of the Deduction as “the supreme principle of all use of the understanding”, which is the principle that “all of manifold of intuition stand under conditions of the original synthetic unity of apperception” (B136). Now it is possible to derive other more specific rules from this general principle by specifying the conditions of the original synthetic unity of apperception. For example, if we have worked out that the concept of a cause is a condition of the original synthetic unity of apperception, we can derive from our general principle the more specific rule that all of the manifold of intuition must stand under the concept of a cause. I suggest that these more specific rules are what Kant is talking about when he refers in the crucial passage to “principles of the objective determination of all representations insofar as cognition can come from them”. These rules about the application of the categories to the manifold of intuition—not the categories themselves—are what Kant thinks are derived from the principle of the transcendental unity of apperception.
The second point to note is that it would not be so peculiar if Kant took the fact that the forms of judgement are what they are as explanatorily basic. After all, this is precisely his attitude towards the forms of sensibility. According to Kant, it is clear that there are precisely two forms of sensibility, space and time, but we can have no insight into why these are the forms of sensibility. Consider the following passage from §21 of the B-Deduction:
But for the peculiarity of our understanding, that it is able to bring about the unity of apperception a priori only by means of the categories and only through precisely this kind and number of them, a further ground may be offered just as little as one can be offered for why we have precisely these and no other functions for judgment or for why space and time are the sole forms of our possible intuition. (B145–6)
Schulting discusses this passage in both KDA (pp. 10–11) and KRS (p. 90). He notes that Kant’s claim that no “further ground” of the “kind and number” of the categories can be offered is perfectly compatible with the claim that some such ground has been offered. But he has less to say about the analogy with sensibility. That space and time are the forms of our sensibility is something that, according to Kant, we can know, but not explain. An analogous claim about the forms of judgement appears to conflict with Schulting’s story.
The third point to make is that it is extremely hard to find any clear evidence in the text of the claims that Schulting attributes to Kant—in particular, the claim that the specific categories of unity, plurality and totality, possibility and necessity, inherence and causality, etc., are bound up in the way that Schulting suggests with transcendental apperception, the representation ‘I think’. One of the remarkable aspects about the text of the Deduction is quite how little reference is made to specific categories. The category of Causality is mentioned a handful of times (A92/B125, A112, B168, B163), the category of Substance twice (B129, B149). The category of Unity is explicitly mentioned once, and then only to contrast it with the unity involved in transcendental apperception:
This unity, which precedes all concepts of combination a priori, is not the former category of unity (§10); for all categories are grounded on logical functions in judgments, but in these combination, thus the unity of given concepts, is already thought. (B131)
On none of the occasions just mentioned does Kant suggest that the specific category in question is part of the representation that constitutes transcendental apperception.
A fourth related point is that while there may be a plausible story about how some of the categories—say, Substance, Reality, Unity—are derivable from the ‘I think’, the story looks strained in the case of the other categories. Take the category of Necessity, which Schulting discusses in connection with Andrew Stephenson’s (2014) objections. It is hard to see how the ‘I think’, the self-consciousness that Kant thinks is involved in every judgement, already somehow involves the category of Necessity. Schulting tries to make the story work by pointing out that the concept of necessity is involved in the claim that the ‘I think’ must be able to accompany all of my representations. Stephenson’s worry here is that this involves a “confusion of levels”. As I understand it, his thought is this: Schulting is supposed to be trying to locate a reference to necessity in transcendental apperception, which is to say, in the representation ‘I think’. What he actually locates is a reference to necessity in a particular claim about transcendental apperception, namely, the claim that the representation ‘I think’ must be able to accompany all my representations. But this is something of which we become conscious when we do transcendental philosophy. It is not an aspect of the self-consciousness that is involved in every judgement. The worry is that it is only by conflating these two things that Schulting is able to make it look as if he has extracted the concept of necessity from the fairly meagre looking resource that is the representation ‘I think’. In Chapter 2 of KRS, Schulting responds to Stephenson’s objection:
When interpretatively reconstructing the argument, presented in TD, about this thinking subject and its constitutive characteristics (i.e. the logical functions of thought with which the categories correspond), we should be careful not to artificially introduce a distinction between two levels of discourse (theory and object) on pain of losing sight of the self-reflexive aspect of Kant’s reasoning. (p. 81)
Schulting is surely correct that there is a reflexive aspect of Kant’s reasoning. What is going on in the Deduction is that the faculty of understanding is both the subject and the object of the investigation. It is both the thing doing the investigating and the thing being investigated. Perhaps it follows from this that there is no difference between what is recognised about the object of investigation and what is recognised by the object of investigation. If the understanding qua subject of investigation recognises that the ‘I think’ must be able to accompany all of my representations, then this is also something that is recognised by the understanding qua object of investigation.
But I don’t think this helps Schulting. It isn’t the faculty of transcendental apperception, which is to say, the mere representation ‘I think’, that is both the subject and the object of the investigation in the Deduction, but rather the faculty of understanding. And the faculty of transcendental apperception is not identical to the faculty of understanding. So from the fact that something is represented about the faculty of transcendental apperception in the Deduction, it doesn’t follow that this is also something represented by the faculty of transcendental apperception.
Acknowledgements: I would like to thank Ralf Bader, Anil Gomes, Colin McLear and Andrew Stephenson for helpful comments.
Invited: 16 August 2017; received: 7 October 2017.
Ginsborg, H. (2008), ‘Was Kant a Nonconceptualist?’, Philosophical Studies 137(1): 65–77.
Gomes, A. (2010), ‘Is Kant’s Transcendental Deduction of the Categories Fit for Purpose?’, Kantian Review 15(2): 18–137.
——— (2014), ‘Kant on Perception: Naïve Realism, Non-Conceptualism, and the B-Deduction’, Philosophical Quarterly 64(254): 1–19.
Griffith, A.M. (2010), ‘Perception and the Categories: A Conceptualist Reading of Kant’s Critique of Pure Reason’, European Journal of Philosophy 20(2): 193–222.
Langton, R. (1998), Kantian Humility: Our Ignorance of Things in Themselves (Oxford: Oxford University Press).
McLear, C. (2016), ‘Kant on Perceptual Content’, Mind 125(497): 95–144.
Schulting, D. (2012), Kant’s Deduction and Apperception: Explaining the Categories (Basingstoke/New York: Palgrave Macmillan).
Stephenson, A. (2014), ‘A Deduction from Apperception?’, Studi Kantiani 27: 77–86.
Strawson, P.F. (2006), The Bounds of Sense: An Essay on Kant’s Critique of Pure Reason (Abingdon: Routledge).
Tolley, C. (forthcoming), ‘The Meaning of ‘Perception in Kant and his Historical Context’, forthcoming in Akten des XII. internationalen Kant-Kongresses.
Van Cleve, J. (1999), Problems from Kant (Oxford: Oxford University Press).
© Robert B. Watt, 2017.
Robert Watt is a Teaching Associate in Philosophy at the University of Cambridge. His research focuses on Kant’s metaphysics and epistemology, especially the Transcendental Deduction of the Categories. He has articles forthcoming in KANTIAN REVIEW, Archiv für Geschichte der Philosophie, and the European Journal of Philosophy.