There is no doubt that the second-edition version of the Transcendental Deduction of the Pure Concepts of the Understanding (henceforth: B-Deduction) is among the most important, but also most difficult, chapters of the Critique of Pure Reason. Footnote ¹ Despite this, an up-to-date commentary, which draws on the advances Kant scholarship has made in recent decades and guides the non-specialist reader through the text paragraph by paragraph, has – until very recently – been lacking.Footnote ² While there are numerous works that can be used as companions to some or all of the first Critique and that address the B-Deduction in more or less detail, these do not usually comment in detail on every portion of the text.Footnote ³ Much less do the many introductory guidebooks to the Critique.Footnote ⁴ Mario Caimi’s Kant’s B Deduction seeks to fill this gap. It is explicitly intended as a commentary and addresses each section of the text, extensively quoting and commenting on individual passages and occasionally proceeding sentence by sentence. It does not set out to defend the interpretation presented and largely refrains from engaging with secondary literature. Almost all of it thus constitutes a running commentary on the primary text. The B-Deduction is covered in its entirety, including notoriously obscure passages such as the discussion of self-affection in §24 or the footnote at B160. This makes it fit to be read alongside the primary text. And while it is often not as clear as one might hope, it succeeds in providing orientation to the reader, especially the non-specialist.

I cannot summarize Caimi’s rich and detailed reading of the Deduction in its entirety here. Many of the things Caimi says strike me as on target and relatively uncontroversial. There are also, however, a number of claims that amount to distinctive exegetical commitments bound to invite controversy. A good way of conveying a sense of Caimi’s reading is to discuss some of these commitments. I will focus on two in particular, both of which concern central aspects of the Deduction.

The first commitment concerns Kant’s method. According to Caimi, Kant’s method in the Critique as a whole is conceptual analysis, carried out by means of a ‘method of isolation and progressive synthesis’ (p. 13). This means that components of the target concept are isolated, individually brought to clarity and distinctness and subsequently joined together again to yield the target concept, which is now clearly and distinctly understood. Bringing individual components to clarity and distinctness may itself involve what Caimi calls synthetic steps, which consist in the discovery of ‘new’ components. These new components are required by a complete analysis of the target concept but up to that point had not been represented clearly. On Caimi’s view, the Critique as a whole gives an analysis of the concept <pure reason>, the Transcendental Aesthetic of <sensibility>, and the Transcendental Analytic of <understanding>. The analysis of <understanding> in the Analytic in turn proceeds by means of isolation and progressive synthesis. In the Transcendental Deduction, the concept to be analysed is ‘the vague concept of synthetical thought related to a manifold, that is, the indeterminate concept of “combination in general”’ (p. 14).

Caimi argues that from this starting point Kant is led directly to the concept of apperception. This concept, in turn, is ‘synthetically enriched’ in the course of the B-Deduction, leading to progressively more determinate formulations of what synthetic unity of apperception in a finite intellect comes to. Thus the first half of the B-Deduction (§§15–20) articulates what synthetic unity of apperception for a discursive mind in general involves, while the second half (§§22–7) enriches this concept by spelling out what synthetic unity of apperception of a specifically temporal manifold amounts to.

This view of Kant’s method will strike many readers as problematic. One reason for concern here (though not the only one) is that Kant appears to be at pains to distinguish his method from conceptual analysis. Thus, when he explains why the first substantive section of the Transcendental Analytic – the one that contains the Deduction – is entitled Analytic of Concepts, he says:

Whatever it is to dissect a faculty (or, related, to analyse the pure use of a faculty), Kant seems to think that it is not the same as analysing a concept. Perhaps Caimi has this passage in mind when he points out that the concept to be analysed in the Analytic is the concept of a cognitive faculty, namely, the understanding (pp. 12–13). But, for one thing, the passage appears to be making a different point. For another, crucial junctures of Kant’s actual procedure in the Analytic do not look like conceptual analysis. For instance, it is hard to see how the Table of Judgements, which Kant presents in the so-called Metaphysical Deduction, could be generated by such a process. It is equally hard to see how the move from ‘the categories apply to objects of intuition in general’ to ‘the categories apply to objects of spatio-temporal intuition’, which Kant undertakes to vindicate in the second half of the B-Deduction, could be based on conceptual analysis.Footnote ⁵

Another reason for concern is that, at least on the face of it, the method of conceptual analysis seems ill-equipped to yield answers to ‘how possible?’ questions. Famously, Kant’s aim in the Critique is not (or not just) to argue that there are synthetic a priori judgements, but (also) to explain how they are possible. In the same vein, it is difficult to see how conceptual analysis can serve to answer the quid iuris question that guides the Deduction: the question whether the application of categories to objects is legitimate.Footnote ⁶

This is not to say that Caimi has no resources for addressing these concerns. But since they are concerns that will likely occur to many readers, even those reading the Deduction for the first time, it would be helpful for them at least to be mentioned, even if a commentary of the sort Caimi has written is not the place to address them in detail.

The second distinctive exegetical commitment I wish to discuss concerns the relation between the two halves of the B-Deduction. Following Dieter Henrich’s classic proposal, it is now widely agreed that the Deduction has a ‘two-steps-in-one’ structure, but there is controversy over just what each of these steps is.Footnote ⁷ The challenge posed by Kant’s text is to explain why the second step does not consist in the following simple inference (as it manifestly does not): the categories are valid of objects of intuition in general; spatiotemporal intuition is a species of intuition in general; so the categories are valid of objects of spatiotemporal intuition. Caimi’s answer is as follows: the categories are objectively valid if and only if they are instantiated, that is, if and only if instances of the categories actually exist. But in Caimi’s view, the simple inference at best shows that it is possible for the categories to be instantiated. So additional work is needed, namely, the work of establishing the existence of objects. Here is how Caimi puts the point:

This account raises a number of problems. First, for Kant, knowledge of existence is dependent on sensation and therefore empirical (see A225–6/B272–4). This conflicts with the a priori status Kant claims for the central propositions involving the categories, that is, the pure principles of the understanding. It also conflicts with Caimi’s own account of Kant’s method, discussed above, since the view that existence claims are synthetic (and so cannot rest on conceptual analysis) is clearly fundamental for the Critical Kant. Second, the relevant passages in the second half of the Deduction do not seem to be concerned with existence claims. Caimi maintains that Kant’s discussion in §26 of the synthesis of apprehension, which pertains to empirical intuition, is intended to establish the existence of objects. But the text does not bear this out. Kant opens §26 as follows: ‘Now the possibility of cognizing a priori through categories whatever objects may come before our senses … is to be explained’ (B159–60, my emphasis).

The passage does not talk about objects actually before my senses; instead it talks about whatever objects may come before my senses. So the claim at issue does not seem to be:

1. (1) There are objects that are cognized through the categories.

   Rather, it appears to be something along the lines of:

2. (2) For any objects, if they can be given to the senses, they can be cognized through the categories.

Caimi is right, of course, that empirical intuition is an important topic in §26. But the point of discussing it is not to underwrite existence claims, which would require an appeal to actual empirical intuition. Instead, Kant seeks to establish a claim about any possible empirical intuition, namely, the claim that it involves a synthesis in accordance with the categories.Footnote ⁸

The third problem concerns the assumption underlying Caimi’s reading of §26, to the effect that establishing the objective validity of the pure concepts of the understanding requires establishing that they are instantiated. While this is a complex topic, there are good reasons for thinking that the categories are objectively valid just in case instances of the categories are really possible (in Kant’s technical sense of that term).Footnote ⁹ Showing this, however, is distinct from demonstrating the existence of such instances.

I hope that my brief discussion of two distinctive exegetical commitments Caimi embraces has conveyed a sense of their problematic nature. Although a commentary is not the place to offer detailed defences of such commitments, it would be consistent with the character of such a work to flag them as such and refer the reader to relevant literature. Since Caimi’s commentary is aimed inter alia at non-specialist readers, this would be especially desirable. While Kant’s B Deduction does a good job of guiding the reader through large portions of the Deduction, its failure to signal the controversial nature of some important aspects of the reading being presented must count as a shortcoming.