HENRY E. ALLISON | Kant’s Transcendental Deduction. An Analytical-Historical Commentary | Oxford: Oxford University Press, 2015


By Henry Allison

I wish to thank Michael Friedman for both his extraordinarily rich discussion, which amounts to a commentary of his own on the B-Deduction, and his generous comments about the import of my own work. Given the detailed nature of his reading of the text, with much of which I am in agreement, I shall not here attempt to comment on his account as a whole, since that would require a reply of almost equal length to his. Instead, I shall focus first on what he suggests is our major point of disagreement, namely, the conceptual-non-conceptual issue, and then consider some of our more specific differences.  

In recent years, the contrast between conceptualist and non-conceptualist readings has become the great divide in interpretations of the Transcendental Deduction. And, as Friedman notes, the focal point of the dispute is the note that Kant attaches to the first paragraph of §26 of the B-Deduction. It reads:

Space represented as object (as is really required in geometry) contains more than the mere form of intuition, namely the comprehension [Zusammenfassung] of the manifold given in accordance with the form of sensibility in an intuitive representation, so the form of intuition merely gives the manifold, but the formal intuition gives unity of the representation.   In the Aesthetic I ascribed this unity merely to sensibility, only in order to note that it precedes all concepts, though to be sure it presupposes a synthesis, which does not belong to the senses but through which all concepts of space and time first become possible.  For since through it (as the understanding determines the sensibility) space or time are first given as intuitions, the unity of this a priori intuition belongs to space and time, and not to the concept of the understanding (§24).  

Friedman places me in the non-conceptualist camp on the basis of my reading of this note, while defining his position as a ‘moderate conceptualism’, thereby aligning his position with that of Béatrice Longuenesse. According to him, “the former [the conceptualists] maintain the absolute primacy of the understanding over sensibility, the latter [the non-conceptualists] emphasise the radically independent contribution of sensibility, and ‘moderate conceptualists,’ like myself and Longuenesse, attempt to steer a middle ground between them” (emphasis added). What makes me a non-conceptualist in Friedman’s view is that my position is closely aligned to that of Christian Onof and Dennis Schulting (Onof & Schulting 2015), who distinguish between two conceptions of unity in the note, namely, one that pertains to space and time as forms of intuition, which they refer to as ‘unicity’ and which essentially amounts to singularity, and another that pertains to space and time as formal intuitions; while he acknowledges only the latter unity, which is a product of the understanding by way of the transcendental synthesis of the imagination. In response, I shall first sketch my own reading of the note, explaining why I deem it necessary to distinguish between two conceptions of unity and then consider what I take to be the main differences in our readings of the note. Finally I shall use these results to address the conceptualist-non-conceptualist issue.

To begin with, my reading of the note and the claim that it is necessary to distinguish between the two senses of unity in it are rooted in my view of the overall structure of the B-Deduction. The starting point and the framework in which I have developed my reading of it is an earlier debate regarding the relationship between its two parts, which was brought to the centre of interpretive and critical work on the text by Dieter Henrich in his classic paper ‘The Proof-Structure of Kant’s Transcendental Deduction’ (Henrich 1982). My basic claim is that, contrary to the then dominant reading that the heavy lifting was all done in its first part (§15–§20), where Kant argues that the objective validity of the categories is demonstrated by showing that they are necessary conditions of the cognition of an object of a sensible intuition as such (überhaupt), following Henrich, I focused my attention on Kant’s remark in the pivotal §21 that through this “only the beginning of a deduction of the pure concepts of the understanding has been made” (B144). The reason why it is only a beginning Kant tells us is that, since he was concerned in the first part with the distinctive contribution of the understanding to cognition, he found it necessary to “abstract from the way in which the manifold is given, in order to attend only to the unity that is added to the intuition through the category”. Accordingly, in the second part in order to complete the deduction, “it will be shown from the way in which empirical intuition is given to intuition that its unity can be none other than the one the category prescribes to the manifold of a given intuition in general according to the preceding (§20)” (B144–5).  The crucial point for understanding the argument of the second part is that what will be added is precisely what the first part abstracted  from for methodological reasons, namely, “the way in which the manifold is given”. This, I take it, refers to space and time as forms of human sensibility, which determine the manner or way in which the manifold is received, and the point is that this manner imposes external conditions stemming from sensibility on the unifying activity of the understanding. Otherwise, I can see no reason why a second part of the Deduction, or at least one as complex as Kant provides, would be required.

On my reading, and I am sure that Friedman concurs, the argument of the second part of the B-Deduction contains two major steps. First, in §24 Kant connects the categories with space and time by means of the transcendental synthesis of the imagination, which he there characterises as “an effect of the understanding on sensibility and its first application (and at the same time the ground of all others) to the objects of the intuition that is possible for us” (B152). The result of this synthesis is the determination of space and time and it yields what Kant refers to in the note as formal intuitions. Then, in §26, appealing to the conclusion of the Aesthetic that space and time as forms of human sensibility are also forms of empirical intuition, Kant connects the categories with empirical intuition and perception by means of the empirical synthesis of apprehension, which is the second effect of the understanding on sensibility, which, in turn, connects them with appearances. The argument turns largely on the transitivity of the necessary condition of relation. For if the categories are necessary conditions of the determination of space and time via the transcendental synthesis of the imagination (§24) and this synthesis is a necessary condition of the empirical synthesis of apprehension §26), then the categories are also necessary conditions of what is apprehended in space and time, i.e. appearances.

Quite apart from the text of the note, the above account of the proof-structure of the B-Deduction is my first reason for attributing the two conceptions of unity to the note. As noted above, the crucial point for me is that the second part be seen as making a distinct and essential contribution to the argument; otherwise Kant would not have claimed that the first part is only a beginning. What it explicitly contributes to the account of cognition is the function of the imagination, which in the A-Deduction was operative from the beginning, even before the categories. And it is clear that the reason why Kant found it necessary to interject the imagination, which at one point he characterised as “a blind though indispensable function of the soul” (A78/B103), into his account is to mediate between sensibility and the understanding, which, in turn, is necessary because of what I term Kant’s discursivity thesis, namely, the claim that

[t]houghts without content are empty, intuitions without concepts are blind. It is thus just as necessary to make the mind’s concepts sensible (i.e., to add an object to them in intuition) as to make its intuitions understandable (i.e., to bring them under concepts). […] Only from their unification can cognition arise (A51/B75).    

The key question, however, is why Kant thought that the imagination is uniquely qualified to perform this indispensable role, which, as I show in the book, is a view that he only arrived at late in the ‘Silent Decade’, long after he had subscribed to the discursivity thesis. The answer, I argue, lies in a combination of the nature of the imagination and the nature of space and time as forms of sensibility, which were shown in the Aesthetic to be infinite, in the sense of all-inclusive, given wholes or totalities. Kant defines the imagination in §24 as “the faculty for representing an object even without its presence in intuition” (B151) and his use of Fettdruck indicates the importance that he attributed to this characterisation. Moreover, immediately after providing this definition, Kant links the imagination with sensibility by virtue of “the subjective condition under which alone it can give a corresponding intuition to the concepts of understanding” (B152). I take this subjective condition to be the nature of inner sense, the form of which is time, and as a result of which the elements of the manifold of an intuition are given successively. And since the task of the imagination in its productive capacity for Kant is to make possible a determinate representation of time, that is, of a stretch of time, such as is involved in the experience of an event, and since due to the nature of human sensibility this requires representing it as a delimited portion of a single-all inclusive time (an infinite given whole), it follows that the execution of this task requires the capacity to represent past and future times that are not ‘present’, as well as the single time of which they are delimited parts. Accordingly, it requires the imagination as defined above. 

 On my reading, the crucial point with respect to the unity issue is that the productive activity of the imagination is guided by rather than the source of the unity in the sense of  the singularity of time (and mutatis mutandis of space) as an a priori form of human sensibility. In short, it is precisely because time (and space) are unitary (in the terms of Onof and Schulting have unicity) or, as I prefer to characterise them, are infinite, i.e. all-encompassing, intuitive wholes, that cognition requires the unification of given representations in a single time (or space). Otherwise expressed, given representations, that is, the manifold of intuition, must be unified, i.e. brought together into a synthetic unity, by the imagination, if they are to be cognised together in a single time (or space). Moreover, this means that these representations are, as it were, subject to two masters; for while, on the one hand, they are subject to the unity imposed by the understanding through the transcendental synthesis of the imagination, on the other hand, they (as well as the understanding, if it is to have a real, as contrasted with a merely logical use) are also subject to the constraints on this unification that stem from the forms of sensibility.

The second and most obvious reason why I deem it necessary to distinguish between two senses of unity in the note is that it seems necessary in order to avoid the apparent contradiction between the claim in the second sentence that the unity of space and time, which in the Aesthetic was ascribed “merely to sensibility”, presupposes a synthesis, which does not belong to the senses, and Kant’s claim in the third sentence that the unity belongs to space and time rather than to the concept of the understanding. For if Kant is referring to the same conception of unity in both sentences, then, given his distinction between sensibility and the understanding, if this unity does not belong to the senses, it must belong to the understanding, albeit in the guise of the transcendental synthesis of the imagination, which contradicts the claim in the third sentence.  On my reading the two unities pertain to space and time as forms of intuition and as formal intuitions. By a form of intuition I take Kant to mean a form in the sense of the manner or way in which “the manifold of intuition”, that is, the sensible data (impressions) that result from the mind being affected from without, are received by the mind. Otherwise expressed, it refers to our form or manner of intuiting, or receiving the given sensory data, which Kant suggests by referring to a form of intuition as a “form of receptivity” (A27/B43), and which in the case of inner sense means successively. Space and time, as forms of outer and inner sense respectively, are these forms. Conversely, by a formal intuition I take Kant to mean the formal structure of what is intuited, insofar as it has been determined by the transcendental synthesis of the imagination, per the account in §24. Kant states that it is through a formal intuition that space is represented as an object, citing geometry as an example. Moreover, as such, a form of intuition and a formal intuition are related as matter and form, which, as Kant makes clear in the Amphiboly chapter, means as the determinable and the determined (see A266/B322).

Since I take the latter point to be crucial, I regret not having focused sufficiently on it in either the book or previous discussions of the note, where I tended to contrast a form of intuition and a formal intuition respectively merely as indeterminate and determinate pure intuitions (see Allison 2004:115–16). The problem was that in so doing I neglected to emphasise that, while space and time as forms of sensibility are to be sure indeterminate, they are also determinable in virtue of their status as a priori forms of intuiting. Moreover, I regard this determinability as essential to the argument of the Deduction in both editions because it guarantees the possibility that what is given in intuition is capable of being determined by the understanding, thereby avoiding the spectre that 

[a]ppearances […] could be so constituted that the understanding would not find them in accord with the conditions of its unity, and everything would then lie in such confusion that, e.g., in the succession of appearances nothing would offer itself that would furnish a rule of synthesis and thus correspond to the concept of cause and effect, so that the concept would therefore be entirely empty, nugatory, and without significance (A90/B123).  

I call this possibility a spectre because I take the central task of the Deduction to be showing that appearances necessarily conform to the categories and I term this Kant’s non-contingency thesis. Moreover,  I argue for this thesis at some length because it has become the target of a highly influential line of criticism of the Deduction, which stems from Strawson and defenders of a Strawsonian style transcendental argument and finds its most vocal proponent in Paul Guyer. The basic charge is that all that Kant is entitled to claim is the conditional necessity that appearances must contain some indeterminate degree of uniformity, if cognition is to be possible; but that largely due to the pernicious influence of transcendental idealism, he inflates this into an absolute or unconditional necessity, as a result of which even empirical propositions are considered to be necessarily true (see e.g. Guyer 1987:342). In response to this line of objection I argue that Kant’s worry about contingency that is expressed in the above cited passage is well founded and is to be seen as his response to Hume’s challenge to the rational grounding of the belief in the uniformity of nature, as expressed, for example, in his remark that

it may […] be a subject worthy of curiosity to enquire what is the nature of that evidence, which assures us of any real existence and matter of fact beyond the present testimony of our senses, or the records of our memory. (Hume 1999:108)

For Hume the question was purely rhetorical because he denied that there was any such “evidence” and he considered the belief as the product of custom rather than reason. Kant obviously denied Hume’s explanation of the belief in the uniformity of nature, indeed that it should be considered a belief, but I argue that he also realised that Hume’s account pointed to a real problem and that its resolution requires providing a ground or warrant to justify projections beyond what is being presently perceived or recollected, which, in turn, requires attributing a necessary rather than a merely contingent uniformity to experience.  

It is neither possible nor even necessary to present the basic thrust of my defence of Kant against this line of attack here, since  Friedman does not raise this objection to Kant’s procedure. But I thought it necessary to at least mention it because it helps to explain why I think it important for Kant to insist that the forms of sensibility, particularly time, provide a pure, i.e. a priori content for the categories, apart from which they would be empty forms of thought. Simply put, the point is that, given the discursivity thesis, Kant must assume some such support from the side of sensibility, if he is to deny that the conformity of appearances to the requirements of the understanding is not merely contingent or, as his critics claim, merely conditionally necessary. Moreover, it is for this reason that I disagree with Friedman, when he suggests that, in contrast to pure sensible concepts, “concepts of the understanding can be meaningfully thought or represented independently of their (spatio-temporal) schemata”. Indeed, setting aside the practical use of these concepts, particularly causality, which Kant explores in the second Critique, there is little room for doubt on this point, since Kant tells us in no uncertain terms that

[w]ithout schemata […] the categories are only functions of the understanding for concepts, but do not represent any object. This significance comes to them from sensibility, which realizes the understanding at the same time as it restricts it. (A147/B187)

And to this I only wish to add that, inasmuch as Kant defines the schemata of the categories as transcendental determinations of time (A139/B187), he must consider time qua form of intuition determinable, since he could not say this about time qua formal intuition because it is already determined.

Moreover, as suggested above, as determinable an intuitive whole has for Kant a kind of unity that is distinct from the unity attributed to a formal intuition. The latter is a synthetic unity in the sense of being the product of a unifying activity through which many representations are combined or grasped together in a single consciousness as one, and, like all such unities, it is grounded in the synthetic unity of apperception. And though Kant usually contrasts a synthetic with an analytical unity, by which he understands a one that is instantiated in a many and attributes such unity to both concepts and the unity of apperception, the unity of an intuitive whole clearly cannot be analytical in that sense. Rather, an intuitive whole is a unity in the sense of being a totality. The situation is complicated, however, by the fact that Kant distinguishes between two kinds of totality, only one of which is applicable to an intuitive whole. Thus, in his remark on the thesis of the Second Antinomy he contrasts space as a totum with a substantial whole or a composite on the grounds that its parts are possible only in the whole (A438/B466); while in an earlier Reflexion he had termed both space and time as tota analytica, by which he understood wholes, the parts of which are only possible with reference to that whole, and he contrasted this with a totum syntheticum in which the part-whole relation is reversed (Refl 3789, AA 17:293). Although the conception of a totum analyticum is closely related to the conception of unity as singularity (Einzelnheit) or, as Onof and Schulting render it, unicity, to which Kant refers obliquely in B136n, I do not regard them as equivalent. The difference stems from the fact that the sensible world (or material universe) for Kant is also singular and, therefore, might be considered as having unicity, even though it is a compositum in the terms of the Second Antinomy or a totum syntheticum in the terms of the Reflexion.  

Evidently in response to my noting Kant’s reference to geometry in this context, Friedman criticises me for a “restrictive reading of ‘formal intuition’”. Instead, he advocates a more expansive reading of space and time as formal intuitions “to be the two all-encompassing intuitive wholes of the Aesthetic”. And on the basis of this together with the distinction between two species of a priori concepts, namely, pure concepts of the understanding or categories and pure sensible, i.e. geometrical concepts such as that of a straight line, a triangle and a circle, he argues against me that the note involves only one conception of unity, that of formal intuitions, whether of particular figures or space and time as all-inclusive imaginative wholes, rather than the two that Onof, Schulting and I, albeit in somewhat different ways, affirm. Moreover, it is clear that this is central Friedman’s to his reading of the B-Deduction as a whole. Indeed, he notes that “it is of the utmost importance, on my reading, that Kant explicitly moves in §24 from concepts of particular figures to representations of space and time themselves as given intuitive wholes”. Nevertheless, despite the impressive comprehensiveness of Friedman’s reading, which ties together in fresh and illuminating ways the note, the B-Deduction as a whole, the Analogies, and the Metaphysical Foundations of Natural Science, I find his account of formal intuition somewhat too expansive. The problem as I see it is that the conception of an all-encompassing intuitive whole is indeterminate, since it entails that however large a region of space or period of time one may intuit, it is bounded by more of the same ad infinitum, which is also how I take Kant’s claim in the intuition arguments in the Aesthetic that space and time are represented as infinite given magnitudes. Accordingly, I find it difficult to see how this could be considered the product of the transcendental synthesis of the imagination, since the explicit function of this synthesis for Kant is to determine sensibility.

This brings us to the second prong of Friedman’s single-unity reading of the note: the distinction between the pure concepts of the understanding and pure sensible concepts. He claims that, in addition to the expansive reading of ‘formal intuition’, the recognition that the conceptual component at issue in the note pertains to pure sensible (my emphasis) concepts rather than to pure concepts of the understanding makes it possible to avoid the apparent contradiction between the second and third sentences, without having recourse to the assumption that Kant is appealing to two conceptions of unity. As noted above, the problem is that, while the second sentence in the note states that the unity of space and time, which in the Aesthetic was ascribed merely to sensibility, is actually the result of a synthesis, specifically the transcendental synthesis of the imagination, which does not belong to the senses, the third sentence maintains that the unity belongs to space and time rather than to the concept of the understanding. Thus, Kant seems to be saying that the unity both does and does not belong to space and time, which does, indeed, seem to involve a contradiction.  

I must admit, however, that I find Friedman’s argument, which, while presupposing his entire complex account, is contained in a single dense paragraph, difficult to follow. If I understand him correctly, the second sentence makes two claims: (1) that the synthesis referred to is the transcendental synthesis of the imagination and (2) that it precedes all concepts (with ‘all’ emphasised) because “the mathematical constructions at issue mediate the application of the categories (now schematised) to the objects of a possible experience”. The first claim is clearly correct; but I find the second perplexing because I fail to see how the mathematical constructions to which Friedman refers can mediate the application of the categories to experience. In my reading of Kant this mediation is the result of the schematisation of the categories. Is the new claim that a second mediation is required in order to link the categories to objects of possible experience? If so, I fail not only to see why this is so, but also how it helps to unify the two sentences.  

Rather than conflicting with it, the third sentence on Friedman’s reading supposedly spells out the implications of the second, which would obviate the worry that they are contradictory. And it likewise seems to involve two claims regarding the unity resulting from the transcendental synthesis of the imagination: (1) that this unity belongs to space and time rather than to the understanding and (2) that rather than depending on the understanding it depends on “the spatio-temporal forms of sensibility, whether of particular spatio-temporal objects (geometrical figures, temporal intervals) or space and time themselves as all-encompassing orders of simultaneity and succession”. I agree that the unity referred to in the third sentence belongs to space and time rather than to the understanding. Indeed, that is the whole point of the two unities thesis, which affirms a distinctive contribution of sensibility. Nevertheless, I find Friedman’s claim that this unity “depends on neither the logical forms of judgement nor the hierarchical ordering of concepts in terms of intension and extension” puzzling. Although this is certainly correct, it seems somewhat beside the point, since I cannot think of anyone who claimed that it did. Moreover, though Friedman says that it “depends on the spatio-temporal forms of intuition”, I suspect that he must mean the conceptually determined formal intuitions, rather than the indeterminate but determinable forms of intuition to which I attribute it. And I also suspect that the same must be said about his immediately following claim that “[t]hese two structural features of space and time themselves, as intuitive wholes, are independent contributions of our faculty of sensibility”. Again, I agree with this claim, but not with what I take to be its underlying assumption, namely, that these intuitive wholes are products of the transcendental synthesis of the imagination; for in that case I do not see how they could be considered independent contributions of our faculty of sensibility; unless, that is, the imagination is tied completely to sensibility, which is hardly what Friedman intends to say.

It seems clear that it is because of claims like this that Friedman places me in the non-conceptualist camp in the ongoing debate in the literature regarding not merely the note at issue, but Kant’s view in the Transcendental Deduction as a whole in both versions. Unfortunately, in so doing he ignored my labelling my view as ‘proto-conceptual’. But since I did this only in passing, I cannot really blame him for neglecting it, and I shall use his neglect as an occasion for attempting to clarify my view. Moreover, I assume that the reason for this neglect is that rather than introducing this conception in the context of my interpretation of the note, which I now realise I should have done, I only did so belatedly in connection with my analysis of Kant’s account of the empirical synthesis of apprehension in §26, which corresponds to the first two parts of the three-fold synthesis in the A-Deduction. In any event, by proto-conceptualism I understand the view that what is given in sensible intuition must be capable of being brought to concepts, if cognition is to be possible. Accordingly, it denies that Kant is committed to the view that everything that is so given must actually be brought to concepts or, equivalently, to the unity of apperception, which is an objection that has been raised against Kant’s account, while dismissing the worry that what is given in sensibility might not conform to the conditions of the understanding (the non-contingency thesis). And I might add that this accords with my claim that what is given in sensibility must be determinable rather than, as Friedman seems to suggest, already determined by the transcendental synthesis of the imagination, which, as noted above, I likewise regret not having made sufficiently clear in the book. To be sure, this position affirms a certain independence of sensibility from the understanding; though it is hardly the “radical independence” that Friedman cites as the defining characteristic of a non-conceptualist reading. In fact, I am not clear what a radical independence of sensibility would mean in a Kantian context in which intuitions without concepts are blind.

Besides labelling me a non-conceptualist, Friedman raises other objections to my account. The first, which he suggests is more of a general complaint than a specific objection, concerns the way in which I set up the problem with which the Deduction deals, rather than a specific objection to my interpretation of it. It involves my view that some aspects of Kant’s account are misleading because of his tendency to appeal to examples of causal necessity, when at issue is a broader notion of necessity, which for Kant pertains to judgement as such, including empirical judgements, and which I have labeled ‘normative necessity’. Since my appeal to this conception of necessity, which admittedly is not found explicitly in Kant, was questioned by both Alison Laywine (indeed it is the focal point of her comments) and Lucy Allais and I addressed the issue in my reply to their comments, I shall here simply note that I introduced it to characterise the kind of necessity that Kant attributes to judgement as such in §19 and to judgements of experience in the Prolegomena. Moreover, I did not claim that Kant actually conflated these kinds of necessity, which would have been a substantive objection, but merely that his fondness for causal examples may have contributed to the misunderstandings of position to which I call attention. Also part of  Friedman’s complaint, however, is that I failed to appreciate that Kant links the notion of necessary connection with all of the dynamical categories and not merely causality. And to that I plead guilty; though with the proviso that I also attribute a normative necessity to judgements involving only the mathematical categories insofar as they purport to be objectively valid, even though they do not involve necessary connection in even the extended sense.

Friedman’s most substantial objection to my account of the B-Deduction, however, seems to be that I fail to find “a clear and coherent argumentative sequence in the principal argument of §26, which extends from the premise (or explication) concerning apprehension and perception (B160) to the conclusion concerning experience (B161)”.  And he attributes this to my failure to recognise “the way in which the footnote contributes to the argument”. Again, if recognising the way in which the footnote contributes to the argument means understanding the relation between space and time as forms of intuitions and as formal intuitions in the same way as he does, then I must once again plead guilty. But in my defence I believe that I have found a coherent, though admittedly not terribly clear, argument in the body of §26. It presupposes the synthesis of apprehension, which Kant describes as “the composition of the manifold in an empirical intuition, through which perception, i.e., empirical consciousness of it (as appearance), becomes possible” (B160), and it proceeds from perception to experience, defined as “cognition by means of connected perceptions” (B161) (see pp. 413–26). On my reading, it consists of six steps, the first five of which link the categories with perception via the synthesis of apprehension and the last connects them with experience, defined as “cognition through connected perceptions” (B161). The five steps are motored by the claim that the empirical synthesis of apprehension, through which perception is generated, must conform to the transcendental synthesis of the imagination because in determining space and time as forms of intuition the latter also determines what is intuited, i.e. perceived, in space and time. And, as noted above, I think that the argument turns on the transitivity of the necessary condition of relation, i.e. if A is a necessary condition of B and B is a necessary condition of C, then A is likewise a necessary condition of C.  

Although they raise some questions with which I attempt to deal in the book but cannot consider here, given Kant’s claim about the role of the transcendental synthesis of the imagination in the determination of time in §24 and the distinction between a form of intuition and a formal intuition in the note, I find the first five steps to be in order. Accordingly, I consider the most problematic aspect of the argument to be the move from perception to experience in the sixth step. Simply put, the problem, as I see it, is that, though, given the transitivity of the necessary condition of relation, if the categories are conditions of perception, then they must also be conditions of experience as defined above, it does not follow that they are also conditions of experience as distinct from mere perception, which I take it is what Kant needs to show. And given the distinction between the mathematical and the dynamical categories, which Kant introduces in the second edition to match the distinction between the mathematical and dynamical principles, the problem could also be expressed by claiming that the Deduction is incomplete, since though the first five steps establish that the mathematical categories (quantity and quality) are necessary conditions of perception, this does not suffice to show that the dynamical categories (substance, causality, and coexistence) are necessary conditions of experience, which is presumably the task of the last step. In short, confining ourselves to the text at hand, something seems to be missing. Moreover, I gather that it is largely because of my adherence to this line of thinking that Friedman charges me with a failure to appreciate fully the force and complexity of Kant’s argument, which on his reading evidently requires an appeal to the Analogies, the Refutation of Idealism, and even the Metaphysical Foundations of Natural Science, as well as the argument of §26 to understand. And though it does not bear on the issue, I must note that I find this somewhat amusing, since I am usually criticised for being too kind to Kant, rather than, as Friedman evidently thinks, not kind enough. 

In defence of my procedure I shall briefly make two points. First, in my role as commentator on the Transcendental Deduction I take a granular approach, by which I mean that I try to provide a close reading of the argument as found in the text. This includes connecting it with preceding portions of the Critique, which it presupposes, particularly the Aesthetic and the Metaphysical Deduction, and in the case of §26 the preceding portions of the B-Deduction; but, whenever possible, I endeavour to avoid having to appeal to what comes later in order to evaluate its cogency. Moreover, I believe that this approach is particularly called for in the case of the Transcendental Deduction because everything that follows it in the Analytic presupposes it. Second, although I do not develop the point to the extent that I perhaps should have, I do attempt to fill in what appears to be a gap between the fifth and sixth step of Kant’s argument on the basis of materials that are already at hand (pp. 425–6). Assuming that the argument up to that point is sound and, therefore, that at least the mathematical categories, as conditions of perception, are necessary conditions of the possibility of experience, I argue that by appealing to the principle of the synthetic unity of apperception (established in §17) Kant can claim that the application to appearances of the mathematical and the dynamical categories reciprocally entail each other, which means that step 6, which affirms the latter, follows from step 5, which affirms the former. This is because the first five steps show that without the mathematical categories there would be no perceptions to unify in a single consciousness; while it follows from the principle of the synthetic unity of apperception that without the dynamical (or at least the relational) categories, there could be no unification of these perceptions in a single consciousness. I expect that this may not satisfy Friedman, since he wants to say so much more, but I do think that it is in large part compatible with his more expansive and illuminating account of that challenging note to §26.


Allison, H. (2004), Kant’s Transcendental Idealism. An Interpretation and Defense, revised edition (New Haven: Yale University Press).

Guyer, P. (1987), Kant and the Claims of Knowledge (Cambridge: Cambridge University Press).

Henrich, D. (1982), ‘The Proof-Structure of Kant’s Transcendental Deduction’, in R. Walker (ed.), Kant on Pure Reason (Oxford: Oxford University Press), pp. 66–81.

Hume, D. (1999), An Enquiry concerning Human Understanding, edited by T. Beauchamp (Oxford: Oxford University Press).

Onof, C. & D. Schulting (2015), Space as Form of Intuition and as Formal Intuition. On the Note to B160 in Kant’s Critique of Pure Reason, the Philosophical Review 124 (1): 1–58.

© Henry E. Allison, 2019.

Henry Allison is Emeritus Professor of Philosophy at the University of San Diego and Boston University. He has published widely on Kant, Spinoza and German Idealism and is best known for the standard work of Kant scholarship Kant’s Transcendental Idealism. An Interpretation and Defense (Yale UP 1983, 2004). His next major work will be published by Cambridge University Press in late 2019, under the title of Kant’s Conception of Freedom. A Developmental and Critical Analysis.