I am very grateful to Stefanie Buchenau, Eric Watkins, Brian Chance and Paul Franks for accepting the invitation to act as commentators during the launch of my Kant’s Reform of Metaphysics: The Critique of Pure Reason Reconsidered (De Boer Reference De Boer2020a), which took place on 29 October 2020 (online). I am also grateful for their praise of the book as well as for their willingness to share their thoughts and criticisms in writing. I very much appreciate, moreover, that their comments range from Wolff and Mendelssohn to the implications of my reading for the way Kant’s impact on German idealism should be assessed. Their pertinent and sometimes challenging questions will allow me to clarify aspects of my reading of Kant, to elaborate on topics not sufficiently treated in the book, and to indicate where, and why, I agree or disagree with my interlocutors. Finally, I would like to thank Henny Blomme for organizing both the book launch and this author-meets-critics section of Kantian Review. Before replying to the commentators, I briefly indicate what I take to be the main features of the book.
1. Overview
Kant’s Reform of Metaphysics challenges a widespread way of framing the Critique of Pure Reason, namely, the idea that Kant in this work first and foremost aimed to identify the conditions of possibility of experience and mathematical cognition. My book shifts the focus from Kant’s account of these conditions to his account of the conditions under which metaphysics – represented by the works of Leibniz, Wolff and their followers – can become a science. I argue that Kant aimed to determine these conditions in the Critique to prepare the ground for the elaboration of a comprehensive metaphysical system. While this system would not have suffered from the defects of the systems put forward by Wolff and Baumgarten, it would have been similar to the latter in important respects. In the Critique and later texts, Kant repeatedly announced his intention in this regard, but the plan never materialized. This contributed importantly to the tendency among commentators, from Kant’s own time onward, to regard the Critique as a self-standing investigation into the limits of human cognition.
Chapters 1 and 2 deal with Wolff’s metaphysics, its impact on eighteenth-century debates, and Kant’s critical engagement with Wolff in his early works up to the 1772 letter to Herz. Chapters 3 to 7 analyse key aspects of the Transcendental Analytic, including the Transcendental Deduction and the Schematism chapter, in light of the thesis that motivates the book as a whole. Chapter 8, finally, offers a reconstruction of the metaphysical system as outlined in the Architectonic. In what follows I specify the result of the shift I propose in three respects.
First, Kant’s well-known yet ambiguous suggestion in its 1787 Preface that the Critique carries out a revolution suggests that he took this work to break with the metaphysics of his predecessors completely. Yet this inference is at odds with the passages in which Kant presents his project as a reform or transformation of metaphysics. I argue that Kant did not intend his critique of pure reason to destroy Wolffian metaphysics, but rather tried to disentangle its rational core from the untenable assumptions that had prevented the discipline from developing into a science. I further argue that the Transcendental Analytic and the Transcendental Dialectic undertake a reform of the two main parts of metaphysics, namely, general and special metaphysics. In both parts, I claim, Kant identifies the relevant pure concepts in a systematic manner and determines the boundaries within which these concepts can be rightfully employed.
Second, I argue that Kant’s analysis of the conditions of possibility of experience is merely a stepping-stone in a larger and more complex argument. I take this larger argument – elaborated in the Transcendental Deduction – to be concerned with the question as to how the human mind can produce a priori cognitions of objects at all. The upshot of my interpretation is that the human mind can rightfully employ a priori categories only if they function as rules that determine how a manifold of sensible representations can be objectified. This entails, for Kant, that a priori judgements about the soul, the world as such and God, while necessary in some respects, do not amount to cognitions of objects and, moreover, that first-order metaphysics must restrict itself to a systematic treatment of the concepts and principles by which this objectifying activity is guided.Footnote 1
Third, as was mentioned above, I give much weight to Kant’s various presentations of the Critique as a propaedeutic. On this account, the Critique interrupts the work that is normally carried out by metaphysicians in order to determine how a metaphysical system could be elaborated as a science. This sustained methodological reflection, which is carried out throughout the first Critique, allows Kant to conceive of metaphysics as a discipline that is comprehensive in one respect and limited in another. On his account, I contend, metaphysics is comprehensive in that it deals with the sum total of a priori concepts and principles that our cognitions of objects presupposes, but limited in that it cannot use these concepts and principles to generate cognitions of particular objects. The final chapter of my book substantiates this reading by outlining the structure and main content of the various parts of Kant’s projected metaphysical system, which would have comprised a general metaphysics or ontology and, on the other hand, a rational physics, rational psychology, rational cosmology and rational theology. With reference to the Metaphysical Foundations of Natural Science, I argue that Kant in each case would have tried to ‘deduce’ the conceptual determinations that we actually find in these disciplines by adopting the table of categories as his guiding thread. In this regard, I consider Kant’s thoughts on the task of philosophy to be more continuous with German idealism than is commonly assumed.
2. Reply to Stefanie Buchenau
Stefanie Buchenau is appreciative of the main thesis I defend, and I am pleased that she considers my reading of Wolff and his impact on Kant convincing in many respects. Many of her comments and suggestions concern themes that might have been further developed or directions for further research. As she acknowledges, it was impossible to treat all relevant topics in sufficient detail, or at all, within the scope of a single monograph, so I am grateful for the opportunity to elaborate on some of them.
One of her worries concerns my prioritization of Kant’s engagement with metaphysics over what she takes to be his engagement with debates on logic, epistemology and the method to be employed in philosophy. As she sees it, these two approaches are intimately intertwined both in the debates at the time and in Kant’s works. It seems to me that this worry is, at least partly, due to a misunderstanding, in particular as regards the term ‘epistemology’. This term and the discipline to which it refers did not exist in Kant’s time. Its terrain was partly occupied by logic and empirical psychology and partly unoccupied. The term only came to denote a specific philosophical discipline – called Erkenntnistheorie – within neo-Kantianism.
Yet I had a more important reason to avoid the term ‘epistemology’. I take issue with readings of the Critique of Pure Reason that focus on epistemological questions – that can be traced back to neo-Kantianism as well – because they assume that Kant investigated the conditions of possibility of empirical and mathematical cognition either for their own sake or to provide the natural sciences with a solid foundation. I maintain, by contrast, that Kant’s analysis of the cognitive activities carried out by the human mind serve the purpose of getting clear on the cognitive activities involved in the generation of a priori cognitions of objects as well as the conditions under which this type of cognition is warranted. On my account, it is this twofold investigation that allows Kant to determine the conditions under which the type of a priori cognition aspired to in metaphysics is possible and, thus, to put the latter discipline on the path to scientificity. For this reason, I maintain that the term ‘epistemology’ covers over the intentions of the Transcendental Analytic and the Critique as a whole.
Apart from this, it would be hard to disagree with Buchenau that problems related to logic and the method to be employed in philosophy were highly relevant to Kant. In this regard as well, however, I would claim that Kant was primarily interested in these problems because of his sustained attempts to reform the metaphysics of his Wolffian predecessors: there is no doubt he took this to require a new methodology. As I see it, in sum, neither his reflections on method in texts like the Inquiry nor his analysis of the human mind in the Critique can adequately be captured by the term ‘epistemology’.
In the book, I also caution against projecting the contemporary meaning of the term ‘metaphysics’ onto eighteenth-century German philosophy. In this context, the term referred at least partly to an investigation into the ultimate principles of our cognition of things rather than these things themselves. As I point out in chapter 1, Wolff and Baumgarten considered the first and foundational part of metaphysics to treat the concepts and principles that are necessarily presupposed in all other disciplines. In his Metaphysics, Baumgarten defines metaphysics as the science of the first principles of human cognition (§1), a definition that Kant criticizes only because of its lack of precision (CPR, A843/B871). Various commentators have shown that this conception goes back to the Suarezian tradition (cf. De Boer Reference De Boer2020a: 25–6; except to avoid possible unclarity, references henceforth simply by page number). Clearly, Kant’s Critique radicalizes this strand of the tradition. I agree with Buchenau, therefore, that Kant’s gnoseological conception of metaphysics shares common ground with Wolff’s, and I do not take her remarks in this regard as a criticism.
While it is true that I could have paid more attention to Wolff’s conception of systematicity in the essays she mentions, the scope of the book did not allow me to offer a detailed comparison of Wolff and Kant’s positions in all respects. The focus is on those elements of Wolff’s philosophy that I held were most relevant to clarifying Kant’s implicit and explicit engagement with Wolff.
As regards a more specific point, Buchenau claims that Kant in the 1764 Inquiry, contrary to what I take to be the case, holds on to Wolff’s view that the first grounds of cognition, to be identified by philosophy, are common empirical notions rather than pure concepts. For reasons of space, the book devotes only a few pages to Kant’s development prior to 1770 and mentions the Inquiry only in a note (p. 34). Yet I take Buchenau’s point to suggest that I give too much weight to Kant’s critique of Wolff in texts from the 1760s. While I cannot further support my account in the context of this response, I believe that the term ‘empirical’ she uses does not adequately capture Wolff’s appeal to indubitable inner experiences and Kant’s endorsement of this element of Wolff’s method in the Inquiry. I discuss elsewhere how both Wolff and the early Kant held that philosophy ought to proceed by means of inner experience and why Kant became unsatisfied with this procedure during the second half of the 1760s (De Boer Reference De Boer, de Boer and Prunea-Bretonnet2021).
3. Reply to Eric Watkins
I am very pleased that Eric Watkins is sympathetic to the overall thesis of my book and convinced by many of its more specific arguments.
Watkins begins his pertinent comments by addressing an issue that I largely bracket in the book, Kant’s practical philosophy. As Watkins points out, Kant opposed main tenets of Wolff’s practical philosophy, namely, (1) its lack of a clear-cut distinction between empirical and a priori motives and (2) an indeterminate concept of perfection. Both tenets can be regarded as examples of what I call ‘continuism’, i.e. the assumption that there is no gap between sensible and intellectual cognitions and, thus, that philosophy can move freely from the former to the latter. I maintain in chapter 2 that Kant at least from 1770 onward came to see this assumption as the main source of the problems afflicting Wolffian and early post-Wolffian metaphysics. I fully agree with Watkins that Kant considered Wolff’s practical philosophy to be defective for similar reasons. However, Kant’s criticism of the Wolffian tradition does not mean he wanted to abolish the latter. As Watkins acknowledges, I argue that, for Kant, the term ‘critique’ refers to the act of ‘sifting the wheat from the chaff’ (p. 2). In the book, I seek to show what this means with respect to metaphysics conceived as the theoretical part of a comprehensive metaphysical system. I hope that the results of this analysis will inspire others to pay closer attention to the positive and negative aspects of Kant’s engagement with Wolffian practical philosophy.
Turning to Kant’s theoretical philosophy, Watkins points to another element which I do not give its due in Kant’s Reform of Metaphysics, namely, Kant’s thoughts on the principle of sufficient reason, its role in Wolffian metaphysics and its relationship to the supreme principle of pure reason of which he endorses at least one version in the Transcendental Dialectic. I do not elaborate on these issues because, as was mentioned above, the focus of the book is on the Transcendental Analytic.Footnote 2 However, it goes without saying that I fully agree with what Watkins calls his ‘reformist view on this point’, namely, the view that Kant opposed certain employments of the PSR while deeming other employments indispensable. In line with my overall argument, I hold that Kant took issue with the PSR insofar as metaphysics relied on this principle to generate cognitions of quasi-objects that cannot be sensibly intuited, but that he endorsed the PSR as a principle that allows the human mind to achieve the largest possible unity among given cognitions, i.e. to generate the ideas of the soul, the world as such and God.
Given this criterion, not mentioned by Watkins, I would disagree with a possible implication of his remark, which draws on A498/B526, that Kant endorsed the use of the PSR even as regards things in themselves. As can be seen from the context, Kant here refers to things in themselves as things insofar as they can be thought, that is, as noumena. I take this to mean that a reformed version of Wolffian cosmology ought to restrict itself to the world as such insofar as it can be thought by purely intellectual means (cf. pp. 241–2). In this case, the use of the PSR would fully abstract from the question as to whether such noumena can be known (cf. A498/B526).
Watkins’ third question concerns my take on the analyticity or syntheticity of the principles treated in metaphysics. I agree with his – and Lanier Anderson’s – view that, on Kant’s definition of analyticity, Leibnizians would conceive of these principles as analytic and that Wolff would concur with this conception. However, I disagree with Watkins’ assumption that Kant, for his part, likewise took the principles treated in Wolffian metaphysics to be analytic. I rather take the Critique to claim that Wolff did not realize that all judgements that do more than provide definitions of things – by means of conceptual analysis – are of a very different kind and cannot be explained on a traditional account of judgement. Seen from Kant’s perspective, these judgements are synthetic. As he puts it in the Introduction:
In metaphysics, even if one regards it as a science that has thus far merely been sought …, synthetic a priori cognitions ought to be contained (sollen enthalten sein), and its aim is not at all to merely analyze concepts that we make of things a priori so as to clarify them analytically, but we want to amplify our cognition a priori; to this end we must make use of such principles that add something to the given concept that was not contained in it, and through synthetic a priori judgments go so far beyond that experience itself cannot follow us that far, e.g., in the proposition ‘The world must have a first beginning,’ and others besides, and thus metaphysics, at least as far as its end is concerned, consists of purely synthetic a priori propositions. (B18)Footnote 3
As I see it, this passage pertains to Wolff and Kant’s conception of metaphysics alike and as yet abstracts from the question as to which synthetic a priori judgements are warranted.
Moreover, on my reading, Kant’s claim that the core principles treated in metaphysics are both synthetic and a priori is just a first step. The second and much more important step consists in asking to what extent these principles – qua synthetic a priori judgements – are objectively valid, i.e. under which conditions they are warranted. Focusing on those judgements that employ the categories as predicates, Kant seeks to answer this question in both the Transcendental Deduction and the Schematism chapter. Thus, I hold that the question as to how radically Kant departed from Wolff cannot be answered by looking at his account of the distinction between analyticity and syntheticity alone.
More generally, as I also note in the book, I am rather sceptical of readings – including Longuenesse’s – that tend to frame the core issues at stake in the Critique of Pure Reason in terms of Kant’s account of general logic (cf. pp. 140–1, 186). Drawing on this account, as I do on various occasions myself as well, can be very illuminating, but this should not obscure the fact that Kant took his transcendental logic to pursue very different ends.
Watkins further asks why a Wolffian should accept that any theoretical cognition of objects rests on non-intellectual conditions, called schemata, and, consequently, that purely intellectual cognitions of the soul, the world as such and God are impossible. Evidently, this is a crucial point. In fact, I consider the whole Critique of Pure Reason, in particular the Transcendental Deduction and the Schematism chapter, to develop a set of intricate arguments to support this one claim (see in particular chapters 5 and 6).
Put briefly, I take these arguments to hinge on the question as to what it means to produce an object of cognition at all. Unlike Watkins, I attribute to Kant the view that philosophy cannot answer this question by considering whether an a priori judgement refers to a mind-independent matter of fact. Kant writes in this regard:
It is clear, however, that, since we have to do only with the manifold of our representations, and the X which corresponds to them (the object), because it should be something distinct from all of our representations, is nothing for us, the unity which makes necessary the object (Gegenstand) can be nothing other than the formal unity of consciousness in the synthesis of the manifold of the representations. (A105, translation modified)
I consider Kant to establish here and elsewhere that (1) producing an object of cognition is nothing but consciously positing a manifold of representations as a rule-bound unity, (2) that this act rests on the synthetic activity that the transcendental imagination carries out on a manifold of representations given in inner sense, and (3) that this activity must be guided by transcendental schemata qua sensible representation of the purely intellectual rules contained in the categories. If a Wolffian would accept these results, I take Kant to hold, he or she would have to admit that judgements such as ‘the soul is immortal’ or ‘God exists’ do not amount to cognition proper. However, Kant’s criterion does not require metaphysics to abolish judgements that function as basic principles in other disciplines, such as his reformed version of the classical principle ‘all events have a cause’ and the principles of the pure understanding more generally. I hope this also clarifies my account of critique as a matter of sifting the wheat (the principles to be treated in ontology) from the chaff (the judgements put forward in former special metaphysics).
Watkins finally asks how my analysis can be squared with the fact that Kant deemed it necessary to complement the Transcendental Analytic with the Transcendental Dialectic. Unlike Watkins and many other commentators, I hold that Kant develops his critique of former general metaphysics, or ontology, in the Transcendental Analytic, in particular in the Transcendental Deduction and the Schematism chapter. This critique establishes that the principles treated in this discipline are valid with regard to objects of possible experience but not with regard to anything whatsoever, i.e. things or objects as such. For this reason, I disagree with Watkins’ claim that the Transcendental Analytic does not address the class of objects in general, i.e. the class of objects treated in former general metaphysics. As I see it, and have argued especially in chapter 5, what Kant has to say about the conditions of possibility of experience in the Transcendental Deduction and elsewhere in the Transcendental Analytic is intended not to explain how we obtain empirical cognitions, but to establish the limits within which we can obtain a priori cognitions of objects at all.Footnote 4 In this regard, it treats the same domain as former ontology. However, Kant does not present his investigation as an investigation concerned with the determinations proper to objects as such because the term ‘object as such’ obfuscates his critical distinction between possible objects of experience and quasi-objects of reason that can only be thought.
Seen in this way, the Transcendental Dialectic is indeed not necessary to demonstrate that things such as the soul and God cannot be known. However, I take the Transcendental Dialectic to complement the Transcendental Analytic in an important way. By treating the main questions of former special metaphysics in an allegedly exhaustive manner, Kant can illustrate and further support his main claim concerning the narrow limits within which a priori cognition of objects is possible. More specifically, as Kant points out, readers who are not convinced by the results of the Transcendental Analytic might be more convinced by his analysis of the antinomies generated within former cosmology and his solution, i.e. the necessity to distinguish claims about appearances from claims about things that can only be thought (A506/B534; cf. Prol., 4: 338–40, in Kant Reference Kant2002).
4. Reply to Brian Chance
Brian Chance’s questions concern my interpretation and translation of a passage from Kant’s 1799 open letter to Fichte, the role of sensibility in the various parts of Kant’s projected metaphysical system and the difference between Kant’s projected metaphysical system and the Wolffian tradition he criticizes. Since I consider the second question to be the most important and most challenging one, I will start with this.
Chance rightly maintains that the role Kant grants to pure sensibility in the various parts of metaphysics is difficult to pin down, and this with regard to both his published works and the parts of the metaphysical system he merely sketched. While I think that the tension Chance identifies can partly be attributed to Kant’s own remarks, I admit that I struggled with the issue while writing the book. Even so, it seems to me that Kant’s position is not necessarily incoherent. In what follows I try to explain why.
I believe that a first step toward a solution of the problem consists in distinguishing between, on the one hand, the concepts and principles treated within a particular part of metaphysics and, on the other hand, the actual use of these concepts and principles in attempts of the human mind to turn its representations into a coherent set of cognitions of objects. I argue in Kant’s Reform of Metaphysics that Kant from the 1770 Inaugural Dissertation onward held that metaphysics cannot become a science unless it becomes a purely intellectual discipline, i.e. completely abstracts from the meaning of categories such as substance and causality insofar as this meaning rests on time qua form of intuition. This is what I call ‘strand 1’ of Kant’s critique of the metaphysical tradition. According to this strand, metaphysics cannot rely on schematized versions of pure concepts, as Kant takes it to have done unknowingly throughout its history. This diagnosis entails that metaphysics itself cannot proceed by synthesizing a manifold of representations so as to generate particular objects of cognition, i.e. unities that the mind consciously posits as rule-bound unities and that presuppose transcendental unity of apperception. What I call ‘strand 2’ of Kant’s critique follows from the first and comes down to the claim that the noumena generated by former special metaphysics do not amount to objects of cognition (see pp. 64–5).
While I take Kant’s position in this regard to be consistent, Chance is right to point out that it is at least potentially at odds with both other elements of the Critique and elements of texts such as the Metaphysical Foundations of Natural Science. More specifically, he claims that the chapter devoted to the principles of the pure understanding, which is part of the Analytic of Principles, ‘uses’ schematized categories to establish these principles qua principles of possible experience. Given my distinction between treatment and use, however, Kant cannot be said to actually use schematized categories: using a category means adopting it as a rule that guides the unification of a set of successive representations. So I would disagree with Chance’s use of the term ‘use’ here and defend the reading I develop in the book, namely, that for Kant transcendental philosophy qua first part of metaphysics can and must be purely intellectual and, thus, can merely provide an account of the concepts and principles that are used by other disciplines to generate cognitions of objects. This view chimes well with Kant’s repeated but often-ignored assertion that philosophy as such is rational cognition from concepts (A713/B741; cf. p. 219).
On this account, even Kant’s repeated claim that the first part of his projected metaphysics, like the Critique itself, ought to include a discussion of space and time does not pose a real problem: treating the concepts of space and time as a priori forms of intuition does not mean that the intellect actually relies on these forms to carry out the peculiar – purely intellectual – cognitive activity that is proper to philosophy.Footnote 5
As regards the relevant parts of former special metaphysics – rational physics and rational psychology – the case is somewhat more complicated. As Chance points out, rational physics rests on, or presupposes, synthetic a priori principles treated in transcendental philosophy or ontology, including the principle of causality insofar as it pertains to appearances. Since, for Chance, synthetic a priori principles ‘involve pure sensibility’, i.e. presuppose time, he claims that rational physics must involve pure sensibility as well. Here as well, however, I believe that my distinction between treatment and use suffices to maintain that rational physics, qua part of Kant’s projected metaphysics, can and must be purely intellectual. Even this discipline does not actually use the synthetic a priori principles treated in transcendental philosophy or ontology: it presupposes them only to establish a set of concepts and principles that are constitutive of a particular domain, i.e. the objects of outer sense. In principle, the same point applies to rational psychology, i.e. the part of metaphysics concerned with the object of inner sense.
Chance further points out that, as I admit, Kant maintains in the Metaphysical Foundations that rational physics presupposes the empirical concept of matter. So how can this discipline nevertheless be purely intellectual? In chapter 8, I answer this question as follows:
The concept of matter need not affect the purity of rational physics, I hold, since this discipline itself does not apply the concepts it treats to material things: it merely treats the pure concepts and a priori principles by means of which scientists can actually do so. In accordance with the Architectonic, Kant considers the metaphysics of corporeal nature elaborated in the Metaphysical Foundations to treat ‘all determinations of the general concept of matter as such’ (4: 475–76). While these determinations are suitable to be used with regard to material things, their content is obtained – ideally – by considering the concept of matter in view of the conceptual determinations treated in the ontology alone. (p. 235)
Seen in this way, the purely intellectual nature of rational physics is not threatened by the fact that the method by means of which it generates concepts and principles involves an empirical concept per se, since this method itself – guided by the table of categories – can still be purely intellectual. As regards the concepts and principles that constitute the content of rational physics, I claim in the book that the Metaphysical Foundations testifies to a tension that Kant ‘may not have been able to resolve: a pull toward concreteness required even in the foundational part of physics … and a pull toward purity required to remedy the defects of previous metaphysical systems’ (p. 235).
As regards rational psychology, finally, Chance notes that Kant takes this discipline to presuppose not only the non-sensible concept of a thinking being but also an empirical concept, namely, the ‘empirical inner representation “I think”’ (A848/B876). In this regard, the two disciplines can indeed be said to be analogous. However, drawing a concept ‘from experience’, as Kant puts it here, does not entail that the resulting concept must remain affected by the context of its discovery. As he puts it already in the Introduction, ‘even among our experiences cognitions are mixed in that they must have their origin a priori’ (A2/B5). Whereas it is unlikely that the concept of matter can be turned into a non-sensible concept, this is no problem as regards the concept of a thinking being or even the ‘I think’ qua concept from which rational psychology ‘is to develop its entire wisdom’ (A343/B401; cf. De Boer Reference De Boer2020a: 239). Moreover, Kant’s claim that we must generate the objects of rational physics and rational psychology by drawing from outer and inner experience (A848/B876) does not mean that the concepts by means of which we subsequently determine these objects must stem from experience or involve pure sensibility. As I argue in chapter 8, Kant thought it was possible, at least in principle, to generate the sum total of these determinations a priori, i.e. by relying on the table of categories.
In sum, while I agree to some extent with Chance’s identification of a tension between the various elements of Kant’s account, I believe that my distinctions between (a) treatment and use and (b) method and content – the latter of which I could have spelled out more – go some way toward solving the problem. Kant himself may not have been able to do so at the time he wrote the Metaphysical Foundations. I would like to suggest that versions of the problem continued to preoccupy Kant’s mind until his death – as can be seen from the notes and drafts known as the Opus postumum (which I do not discuss in the book).
Whereas Chance’s first question, to which I now turn, seems to be concerned with a detail, it is a significant one in that commentators until today assume, based on the contested passage of Kant’s 1799 open letter to Fichte, that Kant changed his mind as to the merely preparatory status of the Critique and that, according to his final judgement, the Critique itself contains his system. Evidently, this assumption makes it easier to ignore Kant’s earlier claims to the contrary. I challenge this position by proposing an alternative translation of the passage. It is true that I do not deal in any detail with the debate that prompted Kant’s letter. Yet I do not see how Chance’s remarks on this point support his defence of Zweig’s translation. Regardless of what was said by Kant’s contemporaries on this much-debated issue, there can be no doubt, in my view, that Kant in the passage merely makes a claim about something he had lauded in the Critique of Pure Reason – albeit that it is not very clear where he actually does so in this work.
Chance also challenges my alternative reading by proposing that derselben refers to ‘the Critique of Pure Reason’ rather than ‘the completed whole of pure philosophy’, i.e. Kant’s projected metaphysical system. Seen in his way, the sentence makes a claim about the truth of the content of the Critique itself. I agree that this is a possibility, though on both readings Kant’s use of the term ‘truth’ remains puzzling. Even so, my point concerning ‘gepriesen habe in x’, which was my main point, is not dependent on Chance’s proposal concerning the antecedent of derselben. For this reason, I still maintain that the 1799 open letter, while somewhat unclear, does not undermine the coherence of Kant’s various accounts of the relationship between the Critique of Pure Reason and the metaphysical system outlined in the Architectonic.
Finally, Chance suggests that the gap between the Wolffian metaphysics Kant targeted and the reformed metaphysics he aimed to establish might be larger than I take it to be, to the point that someone like Mendelssohn might not have recognized the latter as an improved version of the former. I find this question difficult to answer because my whole book precisely seeks to determine in which respects Wolff and Baumgarten’s metaphysics and, on the other hand, the metaphysical system Kant outlines in the Architectonic are continuous and discontinuous. I am convinced that Mendelssohn would have recognized the continuity between Wolffian metaphysics and what I call Kant’s first-order metaphysical system had the latter been published. What Mendelssohn and his contemporaries had a hard time understanding, by contrast, was the second-order investigation into the very possibility of metaphysics that Kant thought had to be carried out prior to the elaboration of a first-order metaphysical system. I take my book to elucidate how and why these two strands are interwoven in the Critique (see e.g. pp. 91–6).
5. Reply to Paul Franks
Many of Paul Franks’ astute questions are concerned with the tension between Kant’s stated aim to reform the metaphysics of his predecessors and elements of his critical project that suggest a more radical departure from the tradition. Franks also suggests that Kant after the 1781 publication of the Critique of Pure Reason may have changed his mind to a larger extent than I acknowledge, in particular in relation to the table of categories. A third theme Franks addresses concerns the continuity or discontinuity between Kant’s conception of a metaphysical system and the conceptions put forward in early post-Kantian philosophy.
Let me begin with the second set of questions. I completely agree with Franks that there is a tension between, on the one hand, Kant’s apparent conviction that the table of categories could be used as a guiding thread to develop the sum total of determinations of any concept and, on the other hand, the way he actually employed this guiding thread in the three Critiques and other writings. Arguably, Kant thought that doing so meant nothing but considering one’s subject matter from the standpoints of quantity, quality, relation and modality. Yet even on this minimal interpretation it seems that Kant was not always able to navigate convincingly between the strict a priori method he envisioned and the manifold of actual determinations put forward in the metaphysical systems of his predecessors. In my conclusion, I address this tension as follows:
[P]ure reason, in its capacity as architect, must impose order on material, handed down by the history of metaphysics, theology, and the sciences, that tends to resist its resolution into a limited number of pure intuitions and root concepts. Arguably, it is this intractable history that confronts Kant’s own critical project with a boundary it cannot go beyond. (p. 258)
As Franks mentions, a clear example of the discrepancy between theory and practice in this regard can be found in Kant’s remarks on rational theology at the end of the Ideal and in various lecture series from the 1780s. However, Kant in neither of these two contexts intended to provide a building plan for rational theology qua final part of his projected metaphysical system. It is quite comprehensible that he in his lectures stayed close to Baumgarten’s own discussion. Thus, we cannot infer from the material we possess that it would have been impossible for Kant to elaborate a rational theology – albeit a very minimal version – along the lines of the Metaphysical Foundations of Natural Science.
I do not quite follow Franks’ suggestion that the variation in Kant’s use of the table of categories is due to changes in his conception of judgement or the metaphysical deduction. As I see it, Kant did not change his innovative understanding of the role of concepts and judgements between 1781 and 1787. Already in the first edition, he conceived of the act of judging as an act that establishes unity among representations in such a way that an object of cognition ensues, something which would be impossible without transcendental apperception (A106–7) and transcendental imagination (A118). In the second version of the transcendental deduction, Kant only explains more explicitly how this account deviates from the traditional conception of judgement as a representation of the relation between subject and predicate (B140–1), i.e. a conception that completely disregards the cognitive activities through which the human mind turns representations into objects of cognition.
For reasons of space, my book deals with the much-debated issue of Kant’s metaphysical deduction of the categories only marginally. In line with All or Nothing, Franks suggests that Kant’s conception of God in terms of the principle of thoroughgoing determination plays a role in the metaphysical deduction (see Franks Reference Franks2005: 70). I doubt, however, that the connection between these two accounts goes beyond the idea that philosophy always needs to account for a possible set of conceptual determinations exhaustively, which means that it must presuppose the totality of such determinations as given. Seen in this way, the metaphysical deduction and Kant’s reformed account of the idea of God merely draw on a common source, namely, Kant’s conception of systematicity. Yet this conception did not require Kant – either in 1781 or later on – to draw on the idea of God to derive the logical forms of judgement and, hence, the categories. More generally, I am not inclined to assume that the results of the Transcendental Analytic – and the first Critique as such – depend on Kant’s various attempts to conceive of the idea of God in terms that harmonize with the critical strand of the work. In this regard, I would emphasize even more than Franks does in his comments, first, that Kant analyses our capacity to generate a priori cognitions of objects from a purely human standpoint and, second, that the human mind cannot obtain such cognitions without the pure forms of intuition.
Let me now turn to Franks’ first set of questions. Franks addresses Kant’s view on the human standpoint in the context of his comments on the idea of a Copernican revolution and his suggestion that I downplay the revolutionary elements of Kant’s first Critique. This issue might partly be dealt with by pointing out that Kant’s use of the term ‘revolution’ is often ambiguous. The passage to which Franks refers (see my p. 49) stems from a letter to Lambert dated 1766 and mentions an awaited revolution in the sciences rather than metaphysics per se. In the Preface to the second edition of the Critique, the term ‘revolution’ is not connected to metaphysics in a straightforward manner either. Kant proposes that metaphysics follow the example of mathematics and physics as regards the ‘change in the way of thinking that has been so advantageous to them’ (Bxvi), leaving open both whether this change must concern all aspects of metaphysics and whether it must occur all at once or gradually.
In my view, the change Kant addresses does not concern all aspects of his work in the same way. Although I do not deal with the relevant parts of the second Preface in any detail, I believe that Kant’s Reform of Metaphysics can account for the tension Franks addresses. As was mentioned in my response to Brian Chance, I distinguish between two strands of the Critique, namely, its positive contributions to first-order metaphysics and, on the other hand, its second-order reflection on the very possibility of a priori cognition of objects and, thus, metaphysics. Seen in this way, Kant’s account of space and time, the categories, the principles of the pure understanding and the ideas of reason is continuous with Wolffian metaphysics and differs from the latter primarily because of his attempt at strict derivations, i.e. at treating the a priori elements of all cognition in a systematic way.Footnote 6
By contrast, Kant’s second-order analysis of the subjective conditions under which metaphysics can actually become a science requires a ‘change in the way of thinking’ comparable to the one he attributes to mathematics and physics. In my view, Kant’s reference to the revolution enacted by Copernicus is concerned with this strand alone. It is in this respect that Kant – in the wake of Locke and Hume – investigates how the human mind generates objects of cognition and conceives of these objects as elements of a single totality. As I stress in the book, Kant departs from his Anglophone predecessors by focusing on the activities involved in the generation of a priori cognitions of objects: the ‘main question’, for Kant, is ‘what and how much the understanding and reason can know free of all experience’ (Axvii; cf. De Boer Reference De Boer2020a: 6).
If the various aspects of Kant’s achievement are to be understood in terms of reform or revolution, then clearly the second-order investigation into the human mind that Kant calls transcendental critique (A12/B26) represents the most revolutionary aspect of his work. Moreover, unlike the first-order strand, it is primarily this strand that hinges on the idea that ‘objects must conform to our cognition’ rather than the other way round (Bxvi). I therefore completely agree with Franks’ view that Kant associated his own turn to the subject with Copernicus’ achievement not because of the latter’s heliocentrism, but because of his hypothesis that those who observe the universe from the standpoint of the earth are in movement, i.e. are active rather than passive (cf. Bxvii).
Franks further points to developments after 1781 that, in his view, qualify as revolutionary. I agree that Kant developed new insights in the 1780s and 1790s, some of which may have impelled him to reinterpret earlier achievements. In this case as well, however, I see continuity rather than discontinuity. As I point out in chapter 1, Kant’s explicit prioritization of practical reason in the Critique of Practical Reason can be traced back to Wolff’s and Crusius’ views (pp. 22, 38–9), so it is a position that Kant was most likely familiar with from early on. While the theme is absent from the 1781 Preface and Introduction, it is addressed in various earlier texts, including a letter to Herz from 1773 (see pp. 61, 248), as well as in the Transcendental Dialectic (A318–19/B375–6) and, more conspicuously, the Canon (cf. A797–801/B825–9).Footnote 7 More generally, I hold that Kant often highlights certain ideas at the cost of others in function of the specific purpose of a particular section, chapter or work. For this reason, we cannot infer from his silence on certain aspects that he was not aware of them or had changed his mind on the subject.Footnote 8
Franks mentions these examples not only to address what he sees as my underestimation of Kant’s radicality but also to suggest that ‘the more radical thrust of his developing thoughts’ after 1781 may well have affected his original plan to elaborate a reformed version of Wolff’s metaphysics. As was mentioned above, I consider the second-order investigation into the human mind carried out in the first Critique to be radically new. Since this strand takes pride of place both in the text itself and in its reception, it indeed makes sense to conceive of the Critique as a whole – despite its original motivation – as a revolutionary moment in the history of modern philosophy. Yet given both my response so far and my discussion in chapter 8 of texts related to Kant’s projected metaphysical system from the 1780s and 1790s, I have no reason to assume that insights Kant developed after 1781 had an impact on his understanding of the structure, content and limits of such a system or on the method by means of which this content should be generated.Footnote 9 Possibly, Kant became aware of obstacles he had not clearly foreseen (see pp. 229–30). Apart from this, however, I am inclined to think that Kant did not accomplish his intended reform of Wolffian metaphysics because his temporary interruption of first-order metaphysics had opened up a completely new, and more exciting, terrain for philosophy as well as because of his increased interest in practical and political matters. I do not think, in other words, that he came to believe that the system ought to be elaborated in a very different manner.
On my account, the fact that Kant did not go on to publish a metaphysical system was detrimental to the reception of the first Critique from early post-Kantian philosophy onward, since, as was mentioned above, the text came to be regarded as a self-enclosed system.Footnote 10 As Franks points out, the perceived shortcomings of what Kant did achieve in the Critique of Pure Reason – in the metaphysical and transcendental deductions in particular – led to sophisticated attempts on the part of Maimon, Fichte, Schelling and Hegel to generate a particular sum total of conceptual determinations by purely a priori means. Even though their methods differ considerably from anything Kant could have fathomed, I hope that my account of Kant’s projected ‘system of pure reason’ (A11/B25) sheds light on affinities between Wolff, Kant and the German idealists that are barred from view on the hypothesis that his critical philosophy revolves around the subject and its limited capacity to achieve empirical cognition of the world.
In conclusion, I would like to express my hope that students and scholars will further test this well-entrenched hypothesis against the available texts and their historical context, find that it does not fit the facts very well, explore alternative hypotheses and develop readings that go well beyond the scope of Kant’s Reform of Metaphysics.
————
Once again, I am very grateful to Stefanie Buchenau, Eric Watkins, Brian Chance and Paul Franks for sharing their reflections, and I hope that our inspiring dialogue will be extended beyond these pages and involve other interlocutors as well.
