2.1 Land’s objections

Land suggests that according to my characterization of the critique of pure reason, what metaphysics needs in order to become a science, and what the critique provides, is an investigation that shows that a set of already established metaphysical cognitions can form a coherent whole. In contrast to this view, Land argues that what metaphysics needs is more fundamental, since the Critique must show that metaphysical cognitions are possible in the first place. Accordingly, the ‘core’ of the Critique concerns the ‘elements’ of metaphysics and should be located in the Transcendental Doctrine of Elements and not in the Transcendental Doctrine of Method.

This objection is an occasion to clarify my view. For when I claim that what the critique does for establishing that metaphysics can become a science is showing that it is capable of architectonic unity, I do not want to suggest that all the ‘elements’ (or the ‘materials’, to recall the word I used above) of metaphysics are available before Kant carries out the investigations he sets forth in the Critique. Indeed, I agree that metaphysics, in order to become a science, also needs that its ‘materials’ are established with proper arguments. However, I believe that this is not what characterizes the critique of pure reason as an independent discipline. In my view, the investigations that, within the Critique, show that those ‘materials’ are possible, do that by actually providing those materials. As such, they must be considered integral parts of metaphysics and belong to transcendental philosophy.

Land further claims that my account of the Critique is unable to account for the ‘how possible’ question that Kant puts at its core, namely: ‘how are synthetic judgments a priori possible?’ (B19). In turn, this has consequences for our understanding of Kant’s answer to Hume. The latter rests on showing that judgements based either on ‘relations of ideas’ or on experience do not exhaust the scope of possible valid judgements. Indeed, there is a third group of possible valid judgements, which are synthetic a priori. But to show this, Kant needs to respond to his ‘how possible’ question. Land suggests that neither transcendental philosophy nor the critique of pure reason as I describe them can accomplish that. In his view, I describe the former as simply ‘identifying’ metaphysical cognitions and the latter as showing that they can form a coherent whole.

Kant does believe that his predecessors and contemporaries made use of metaphysical synthetic a priori principles that are valid in his account, like the principle of causality. But this is different from saying that those principles constitute metaphysical cognitions that are ‘available’. What is needed in metaphysics is an adequate justification of those synthetic a priori principles, which shows that they can count as cognitions. This is exactly one of the tasks of transcendental philosophy, in my account. Indeed, the parts of transcendental philosophy that are established within the Critique concerns ‘root concepts’ (Stammbegriffe) for the cognition of objects. These are pure concepts that lie at the basis of synthetic a priori claims (Gava Reference Gava2023b: 5, 7, 70). In its analysis of these concepts, and in particular in those concerned with the objective validity of the representations of space and time and the categories, transcendental philosophy establishes ‘that’ and explains ‘how’ synthetic a priori principles are possible. In this way, Kant provides an answer to his ‘how possible’ question.Footnote ³

2.2 De Boer’s objections

De Boer agrees with me that we need to distinguish between transcendental philosophy and the critique of pure reason as two different disciplines that are established within the Critique, but she proposes a different way of drawing this distinction. In particular, she suggests that transcendental philosophy has a systematic end, namely that of developing the pure concepts and principles of reason in their ‘entire scope’, whereas the critique of pure reason has a normative end, which consists in determining the warranted use of these concepts and principles.

I distinguish between transcendental philosophy and the critique of pure reason in a different way. I characterize transcendental philosophy as an investigation into our pure concepts for the cognition of objects. The Critique only includes the parts of this discipline that are concerned with ‘root’ concepts, namely those pure concepts that lay at the basis of synthetic a priori claims. Transcendental philosophy must first identify these concepts and track their origin, which is the task of metaphysical deductions. It must also establish their validity, which is the task of transcendental deductions (Gava Reference Gava2023b: 5, 73). By contrast, the critique of pure reason is the discipline within the Critique that achieves its aim as the doctrine of method of metaphysics (Gava Reference Gava2023b: Ch. 2). As such, it must show that metaphysics is capable of architectonic unity. In my account, an essential tool for doing that is setting clear limits to the validity of the root concepts identified by transcendental philosophy, since this is instrumental to establishing that metaphysics can achieve systematic coherence (Gava Reference Gava2023b: 6–7, Introduction to Part III).

De Boer has two main reservations against my way of characterizing transcendental philosophy and the critique of pure reason. First, she disagrees with my inclusion of a normative dimension within transcendental philosophy. Second, she believes that the way in which I separate the investigations that establish the validity of root concepts from those that limit it is artificial.

Let me begin with the first reservation. If we keep the focus on root concepts, de Boer’s way of distinguishing between transcendental philosophy and the critique suggests that the former has only to do with their systematic presentation, rather than with their justification. This seems to me implausible. For if transcendental philosophy is a part of metaphysics that investigates root concepts, it is reasonable to think that it offers arguments to the effect that those root concepts are valid. Unless we have reasons to assume that those arguments are different from those establishing the validity of root concepts within the pages of the Critique of Pure Reason, the most straightforward reading of the latter arguments is to take them as presenting parts of transcendental philosophy.

Here, I wish to add a clarification regarding the link between transcendental philosophy and the issue of systematicity on which de Boer insists, for she suggests that this issue is not central for the critique of pure reason, but only for transcendental philosophy. The issue of systematicity is important for transcendental philosophy because it is itself a system. But this is different from saying that the investigations carried out by transcendental philosophy take its own systematicity into consideration. It is plausible to think that while transcendental philosophy is responsible for developing the cognitions that form the system of transcendental philosophy, it is the task of another discipline to show that these in fact constitute a system. Indeed, I do believe that some arguments within transcendental philosophy directly consider the issue of systematicity (I will come back to this issue later). However, what I want to emphasize here is that the fact that transcendental philosophy is a system does not mean that there is no need for another discipline that takes into consideration its systematicity. I believe that this distinction between simply being a system and considering the systematicity of a discipline is central to the passages in the Introduction of the Critique where Kant discusses the relationship between transcendental philosophy and the critique of pure reason (as for example at A13/B27). It is the specific task of doctrines of method to consider whether a science possesses architectonic unity. The fact that a science has this unity does not make the former superfluous.

Let me now consider de Boer’s second reservation, namely the one that targets the division of labour that I propose between transcendental philosophy and the critique of pure reason as far as their investigations into the validity of root concepts are concerned. In this respect, de Boer criticizes my interpretation of transcendental deductions as arguments that are only concerned with establishing positive results regarding this validity. In her view, the identification of clear limits for the use of root concepts is a central aim of transcendental deductions that cannot be described as secondary. To support her interpretation, de Boer challenges my reconstruction of the transcendental exposition of the concept of space, which I read as its transcendental deduction. While I take the exposition to establish that the spatial properties of objects that we do cognize are properties that objects have only as they are intuited by us, I claim that the exposition does not rule out that objects might have spatial properties independently of our intuition (Gava Reference Gava2023b: 134). In contrast to my reconstruction, de Boer argues that the transcendental exposition already rules out this possibility. Additionally, de Boer draws on the passage in which Kant speaks of the ‘unavoidable necessity’ of the transcendental deduction of the concept of space (A87-8/B119-21). Since this necessity is rooted in the need to avoid an unwarranted use of the concept, de Boer argues, it is clear that setting limits to the valid use of a concept is a central aim of transcendental deductions.

Here, I do not have enough space to consider what the transcendental exposition of the concept of space establishes and what it leaves undetermined (for my reconstruction, see Gava Reference Gava2023b: 129–34). I will focus on de Boer’s reading of Kant’s considerations concerning the ‘unavoidable necessity’ of the transcendental deduction of the concept of space. While I agree with her that the passage proves that the need of setting limits for the concept of space is the main reason for providing this deduction within the Critique, the passage also suggests that setting limits of this kind is not essential for transcendental deductions as such. This becomes clear if we consider the context of the passage. Kant argues that the only deduction that is appropriate in the case of the categories is transcendental and not empirical, given that these concepts are a priori and that we need an answer to a normative question, not a descriptive question (A85-7/B117-19). Kant refers to the ‘unavoidable necessity’ of the transcendental deduction of the categories and the concept of space and links it to the need of setting limits to these only after he has already established what a transcendental deduction is and why this is the proper approach in the case of the categories. This suggests that, at least in principle, we can make sense of the concept of a transcendental deduction even for concepts whose validity does not need to be limited. But of course, the need of preventing an unwarranted use of a concept can make providing a transcendental deduction of it more pressing, if the latter is instrumental for setting limits to that concept, as I suggest.

3.1 Land’s objections

In two passages of the Critique of Pure Reason, Kant describes the Critique respectively as a treatise on method (Bxxii) and as a doctrine of method (A82-3/B108-9). I suggest that these passages are indications that the Critique is the doctrine of method of metaphysics. Land argues that when one carefully reads the passage in which the Critique is equated to a treatise on method, one immediately sees that my suggestion is implausible. For the passage presents the idea of the Copernican turn in metaphysics. Accordingly, the term ‘treatise on method’ points toward Kant’s argument that the change of approach symbolized by the Copernican revolution is what metaphysics needs in order to become a science. But this puts the idea of faculty analysis at the centre of the Critique as a treatise on method, since this revolution involves accounting for our a priori cognitions by focusing on how the subject and its faculties actively contribute to them.

Land is absolutely right that Kant is presenting his Copernican experiment in the passage in question. As I will show, the passage also contains elements that hint at the way in which the Critique considers the architectonic unity of metaphysics. Before I do that, I would like to present some considerations that build on the fact that Kant describes the Critique as a treatise on the method, even if we assume that the method it proposes is faculty analysis. For it is one thing to say that the Critique performs faculty analysis, another to say that it presents a theory or treatise according to which faculty analysis is the appropriate procedure in metaphysics. So, what does the Critique do in this respect? It seems to me that it does both. It certainly provides arguments that perform faculty analysis. In particular, metaphysical deductions appear to fit this description. They trace the ‘origins’ of different root concepts and, by doing that, they are able to both distinguish between different faculties and clarify their role in cognition.

When the metaphysical deductions perform faculty analysis, they are not thereby presenting a theory or treatise about faculty analysis. So, perhaps, the critique of pure reason, as a specific discipline that is different from transcendental philosophy, is responsible for presenting this theory. If we can draw this distinction between the ways in which faculty analysis is relevant for transcendental philosophy and the critique of pure reason, respectively, we can stress the following. First, faculty analysis, as a particular procedure of investigation, is distinctive of transcendental philosophy as part of metaphysics. Second, the critique of pure reason, as a ‘treatise on method’, would not itself be a work of faculty analysis. Rather, it would be a ‘second-order’ investigation, to use de Boer’s way of describing the critique, which shows that the appropriate procedure in metaphysics is the one displayed in the ‘examples’ of faculty analysis contained within the Critique, namely in those arguments that present parts of transcendental philosophy. But third, this confirms that the critique, as a doctrine of method, fundamentally relies on transcendental philosophy. It relies on it because the way in which the critique displays the correct procedure in metaphysics rests on providing actual examples of that procedure that belong to transcendental philosophy.

In my reconstruction of the achievements of transcendental philosophy within the Critique, I argue that it only establishes ‘positive’ results regarding the nature, origin, and validity of root concepts, where faculty analysis is essential to these investigations (Gava Reference Gava2023b: 6, 12, 74). By contrast, I ascribe the role of setting limits to our use of root concepts to the critique of pure reason. In this sense, the critique draws the ‘negative’ consequences of the analysis of root concepts carried out by transcendental philosophy by showing that root concepts cannot be used beyond possible experience. This picture is compatible with what I have just said on the critique as a ‘theory’ of method that highlights the merits of faculty analysis. As we saw, the critique is a doctrine of method which considers the possibility of metaphysics as a science. As such, it provides a second-order investigation of transcendental philosophy. This investigation is first of all directed towards establishing clear limits to cognition by showing that the understanding of the root concepts that was established and legitimated within transcendental philosophy is the only legitimate understanding of these concepts (Gava Reference Gava2023b: Ch. 5).

This second-order investigation of the results of transcendental philosophy might have a second dimension that I do not discuss in my book. The latter consists in emphasizing the merits of the new approach to metaphysics exemplified by transcendental philosophy, where this approach fundamentally relies on faculty analysis. I believe that this aspect of the critique can form a coherent component of my characterization of it. The main goal of the critique is to establish that metaphysics can achieve architectonic unity. With this goal in view, the critique draws on the results of transcendental philosophy, first of all, as I show, in order to determine strict limits to our cognition. But the critique can also draw attention to the procedure that allowed transcendental philosophy to achieve its positive results, exactly because those procedures have been successful in establishing valid metaphysical cognitions.Footnote ⁴ Both these ways of assessing the results of transcendental philosophy are instrumental to establishing whether metaphysics can achieve architectonic unity.

Let me emphasize that Kant himself links the critique’s assessment of the merits of faculty analysis to the question of the architectonic unity of metaphysics. In the passage where he describes the critique as a treatise on method, he suggests that its analysis of those merits is performed with the aim of cataloguing ‘the entire outline of the science of metaphysics, both in respect of its boundaries and in respect of its entire internal structure’ (Bxxii; see Gava Reference Gava2023b: 59). This confirms that faculty analysis is first of all a procedure carried out by transcendental philosophy. The critique, in its evaluation of the achievements of transcendental philosophy, highlights the merits of faculty analysis. In doing that, it does not itself perform faculty analysis. The evaluation of those merits is accomplished having the main goal of the critique in view, which is determining whether metaphysics can achieve architectonic unity.

3.2 De Boer’s objections

As we saw, I grant that the Critique of Pure Reason contains investigations that proceed through faculty analysis. However, these investigations primarily belong to transcendental philosophy in my account. De Boer worries that this way of marginalizing faculty analysis by limiting its scope of application within metaphysics prevents us from seeing how this procedure essentially contributes to the main purpose of the Critique, namely establishing that metaphysics can become a science. It is only through faculty analysis that the Critique shows that we can obtain a priori metaphysical cognitions of objects, since these cognitions rest on the act of unifying a manifold of successive representations according to the categories. But since this act is performed on a material that we receive through intuition, this way of establishing the validity of those cognitions rules out that we can have cognitions of objects of pure reason, like God. Both these results are important for establishing the possibility of metaphysics as a science. While the first result is essential to determining what metaphysical cognitions are in our reach, the second prevents us from falling prey to transcendental illusion.

In response to de Boer’s worry, I offer three considerations. First, I agree with de Boer that Kant’s account of the act of unifying a manifold of successive representations according to the categories is essential for the way in which he establishes the possibility of metaphysics as a science. In my view, however, this account determines that metaphysical cognitions are possible essentially by providing an actual instance of those cognitions. This is simply another way to put my claim that the Critique contains parts of transcendental philosophy and that those parts identify root concepts while they trace their origin and establish their validity. In my view, the main reason to insist on the fact that faculty analysis primarily characterizes the investigations of transcendental philosophy is fundamentally this: if we focused on these investigations to determine what the critique of pure reason is, we would not have a clear way of distinguishing between transcendental philosophy and the critique. For when we ask how transcendental philosophy identifies and validates its root concepts, we would have to make reference to the same procedure of faculty analysis that we ascribe to the critique.

Second, I also agree with de Boer when she insists on the role of faculty analysis for establishing limits to our cognition. Recognizing this relevance of faculty analysis does not necessarily imply that the critique of pure reason is itself responsible for carrying it out. Rather, as we saw, I suggest that the ‘negative side’ of the critique of pure reason shows that the understanding of the root concepts that was established and legitimated within transcendental philosophy is the only legitimate understanding of these concepts when they are used for cognition (Gava Reference Gava2023b: Ch. 5). Accordingly, faculty analysis is essential for establishing limits to cognition, but only insofar as the investigations on the origin and validity of root concepts carried out by transcendental philosophy are so essential. In drawing those limits, it seems to me, the critique does not itself engage in faculty analysis. Rather, it draws some consequences from an actual exercise of faculty analysis (belonging to transcendental philosophy) which is included in the Critique.

Third, perhaps it is true that my claim that the Critique simply ‘includes’ investigations that proceed through faculty analysis is too restrictive. Perhaps, when the critique draws the ‘negative consequences’ for our cognition of the investigations carried out by transcendental philosophy, it emphasizes ‘positive consequences’ too. However, this operation of the critique can be understood along the lines of what I have suggested in my response to Land. Namely, the critique would not itself be responsible for faculty analysis. Rather, it would be responsible for providing a theory of faculty analysis, which shows that the appropriate procedure in metaphysics is the one displayed in the ‘examples’ of faculty analysis contained within the Critique.

4.1 Land’s objections

Land offers four considerations that speak against my claim that the worldly concept of philosophy is the idea that can provide architectonic unity to metaphysics. First, he points out that Kant regards the theoretical and the practical parts of metaphysics themselves as forming two systems. This suggests that the theoretical part of metaphysics can attain architectonic unity without considering the highest good as the final end of reason. This would mean that its architectonic unity does not depend on the worldly concept of philosophy. Second, Land claims that the worldly concept of philosophy does not in fact play a key role in the Architectonic chapter. Third, he draws on the passage in the Introduction where Kant discusses the relationship between the critique of pure reason and transcendental philosophy. In the passage, Kant explicitly refers to the role of the critique in providing an architectonic plan of transcendental philosophy, but this plan clearly does not depend on the worldly concept of philosophy. Fourth, and finally, Land points out that the architectonic unity of metaphysics ultimately depends on the systematicity of reason. Therefore, it is reason itself that secures this unity and not the worldly concept of philosophy. A consequence of this is that faculty analysis is central for the project of establishing the architectonic unity of metaphysics.

I begin with Land’s second point. In my view, his contention that the worldly concept of philosophy is not central in the Architectonic chapter is difficult to sustain. In Section 1 above, we saw that Kant’s reference to reason’s needs at the beginning of the Transcendental Doctrine of Method already hints at the worldly concept of philosophy as essential for the plan of metaphysics. But Kant also clearly links the architectonic unity of metaphysics to this concept in the Architectonic chapter. Let me just quote two passages where this is evident. The first passage is located immediately before Kant presents his sketch of the system of metaphysics. He writes that in providing this sketch, he wants to ‘determine more precisely what philosophy, in accordance with this worldly concept, prescribes for systematic unity from the standpoint of ends’ (A839-40/B867-8; translation altered). The second passage comes immediately after the sketch. It reads: ‘The original idea of a philosophy of pure reason itself prescribes this division; it is therefore architectonic, in conformity with its essential ends, and not merely technical, in accordance with contingently perceived affinities and, as it were, established by good luck; and for that very reason it is unchangeable and legislative’ (A847/B875). Given these passages, it seems to me very difficult to question either that Kant considered the systematic sketch of metaphysics that he provides in the Architectonic as drawn on the basis of the worldly concept or that the sketch in question offers a grasp of the architectonic unity of metaphysics.

According to land’s first point, theoretical philosophy itself forms a system that can be attained independently of the worldly concept of philosophy. At best, this proves that the worldly concept is not relevant for the architectonic unity of the theoretical part of metaphysics. Even if we concede this claim,Footnote ⁵ the question regarding what idea can provide architectonic unity to metaphysics as a whole would remain open. Clearly, in order to provide unity at this level, the idea in question would need to determine the relationship between the practical and the theoretical parts of metaphysics. As I see it, it is only the worldly concept that is able to determine this relationship, since it establishes that the theoretical part of metaphysics should be subordinated to the practical. This is the case because the final end of reason, which is central to the worldly concept, is a topic for practical philosophy.

I turn now to Land’s third point. It is true that in the passage at A13-4/B27-8, where Kant considers how the critique of pure reason provides an architectonic plan of transcendental philosophy, he does not refer to the worldly concept of philosophy. It is also true that at A15/B29 he excludes moral principles from transcendental philosophy. But this has no consequence for the question whether the worldly concept of philosophy is relevant for the architectonic unity of metaphysics or not. In the passages in question, Kant is considering the relationship between the critique of pure reason and transcendental philosophy, which is only one part of metaphysics. Kant claims that the critique of pure reason allows us to see transcendental philosophy as a system. The fact that Kant does not mention the worldly concept of philosophy in this context proves at best that the worldly concept of philosophy is not relevant for the systematicity of transcendental philosophy. It does not prove either that the worldly concept of philosophy is irrelevant for the architectonic unity of metaphysics as a whole or that the critique of pure reason is only concerned with the architectonic unity of transcendental philosophy.

Land’s last point has it that the architectonic unity of metaphysics ultimately depends on the systematicity of reason, where this makes the worldly concept of philosophy superfluous. One quick answer to this objection is the following: as we saw, the worldly concept of philosophy puts at the centre of the architectonic unity of metaphysics the essential ends, and ultimately the final end, of reason. These ends point towards the teleological nature of reason, which, given its very nature, aims to realize some states of affairs that would result from its perfect and complete application (e.g., complete rational knowledge or the highest good; see Gava Reference Gava2023b: 30–1). In turn, these ends are essential for defining what reason as a system is. So, it seems that the fact that the architectonic unity of metaphysics depends on reason considered as a system is not at all in conflict with the claim that the worldly concept of philosophy is fundamental to it.Footnote ⁶

There is another aspect of Land’s last point that merits consideration. He takes Kant’s metaphysical deduction of the categories as an example of how the architectonic unity of metaphysics is achieved. Since the metaphysical deduction does not rely on the worldly concept of philosophy, it does not seem that this concept is relevant for achieving that unity. Here, let me recall my response to de Boer’s objection according to which the systematicity of cognition is a problem for transcendental philosophy, not for the critique. I argued that the issue of systematicity is important for transcendental philosophy because it is a system, while it is the task of the critique to show that metaphysics, including transcendental philosophy, possesses architectonic unity. In that context, I conceded that some arguments within transcendental philosophy directly consider the issue of systematicity. This is the case with the metaphysical deduction of the categories, where Kant wants to show that his list of the categories is complete. However, the question regarding how the architectonic ordering of the categories is achieved within transcendental logic has no consequences for how the critique shows that metaphysics as a whole is capable of architectonic unity. Clearly, the argument for the completeness of the list of the categories has no bearing on questions such as these: how is the theoretical part of metaphysics related to the practical part? Is it possible to see them as constituting a unity? What is the description of metaphysics that allows us to see this unity? These questions are not considered within transcendental philosophy or any other doctrinal part of metaphysics. They are addressed by the critique of pure reason and the worldly concept of philosophy is key for answering them.

4.2 La Rocca’s objections

In his comments, La Rocca displays an approach to the Critique that is very close to mine. This gives me the opportunity to close my response with a more constructive perspective, one that is not limited to defending my view, but takes advantage of some of La Rocca’s points to develop a line of research that in many respects we share. While agreeing with my approach on many points, La Rocca advances fundamentally four objections to my account of the architectonic unity of metaphysics. First, he challenges my claim that metaphysics cannot attain architectonic unity when it is construed according to the ‘school concept’ (Schulbegriff) of philosophy. Second, he suggests that I do not differentiate enough between the school concept and the worldly concept of philosophy, since I still characterize the latter as mainly concerned with obtaining knowledge. Third, he builds on Kant’s claim that philosophy cannot be learned in order to put into question my contention that the idea of metaphysics can in principle be realized. Fourth, he argues that one of the central tasks of the critique as a doctrine of method has to do with establishing a new ‘way of thinking’. In turn, this is connected to the idea of philosophy according to the worldly concept and has consequences for how we should approach Kant’s notion of belief (Glaube) as a free taking-to-be-true (Fürwahrhalten).

In my book, I argue that the ‘school concept’ of philosophy can at best provide ‘technical unity’ to metaphysics, but that, in fact, it cannot even provide this unity insofar as the philosopher who proceeds according to this understanding of philosophy does not care about setting limits to rational cognition (Gava Reference Gava2023b: 36–7). According to this concept, metaphysics is understood as a ‘a system of cognition that is sought only as a science without having as its end anything more than the systematic unity of this knowledge’ (A838/B866). In my view, this concept of philosophy can at best provide ‘technical unity’ to metaphysics because it leaves out from our description of this discipline something that is fundamental to its very nature, namely the orientation provided by the essential ends of reason. In contrast to this approach, La Rocca argues that Kant stresses in various passages that metaphysics according to the school concept is systematic. He takes this as an indication that it is capable of a sort of architectonic unity.

I agree with La Rocca that Kant describes metaphysics understood according to its school concept as a discipline that aims at systematicity. However, this does not mean either that the systematicity displayed by it would yield architectonic unity or that it is able to actually attain systematicity. The systematicity provided by the school concept would be based on a description of metaphysics that leaves out something that is fundamental to it, namely the reference to reason’s essential ends. But this is enough to doubt that the systematic ordering in question would be the ‘correct’ one, which provides architectonic unity to metaphysics. Furthermore, even if metaphysics construed according to the school concept certainly aims at systematicity, it is doubtful that it can attain it, exactly because it is not interested in setting limits to metaphysics, where these limits are conditions for reaching systematic coherence, in my account.

La Rocca’s first objection gathers additional force from its being instrumental to his second objection, according to which I do not differentiate enough between the school concept and the worldly concept of philosophy. I characterize the worldly concept of philosophy as the idea according to which metaphysics can attain architectonic unity and become a science. This commits me to an understanding of this concept where the aim of philosophy is still mainly concerned with obtaining a body of knowledge. However, Kant presents the worldly concept through its personification in the ideal of the philosopher as a ‘teacher of wisdom’, which points towards a totally different understanding of philosophy. The central aim is no longer to attain a body of cognition, but rather to provide guidance regarding how we can become virtuous. La Rocca suggests that my denial that metaphysics according to the school concept can attain architectonic unity brings me to my claim that this unity is achieved through the worldly concept. As a consequence, however, I neglect the different perspective in philosophy introduced by the latter concept.

La Rocca is right to emphasize that Kant links the worldly concept of philosophy to a completely different understanding of the role of the philosopher as a teacher of wisdom. However, Kant also links it to metaphysics considered as a science. This connection is explicit in the passage in which Kant introduces the worldly concept. He writes: ‘[f]rom this point of view philosophy is the science of the relation of all cognition to the essential ends of human reason’ (A839/B867, emphasis added). So there is a sense in which taking the standpoint of reason’s ends is essential to metaphysics as a science. This is a point where La Rocca’s objection is an occasion to bring the discussion further. For I think that Kant is here discussing two different ways in which reason’s ends, and reason’s final end in particular, are relevant for metaphysics and philosophy. On the one hand, the realizability of the highest good is an important question within metaphysics as a science. Since this end is fundamental for our reason, metaphysics provides an answer concerning whether we can consider it realizable. But the highest good is primarily an end that we ought to pursue as moral agents, especially as far as virtue, as one of its components, is concerned. In this respect, philosophy should not only establish that the end is realizable but should also provide practical guidance concerning how we can become virtuous. Stressing that considering the highest good is essential for attaining architectonic unity in the science of metaphysics is not in conflict with the claim that the ultimate role of philosophy is teaching how we can become virtuous, where this is not a task for metaphysics as a science (I make a proposal along these lines in Gava Reference Gava, Dunphy and Lovat2023a. See also Gava Reference Gava2023b: 39–40).

In the Architectonic, Kant claims that philosophy cannot be learned. We can only learn to philosophize (A837/B865). First of all, this is due to the fact that philosophy, understood as the ultimate system of metaphysics, is nowhere to be found. Additionally, Kant treats the ideal of the philosopher as a teacher of wisdom as a model one cannot ever claim to have realized (A839/B867). Insofar as that perfect ‘teacher’ cannot be realized, the ‘teachings’ that she would provide cannot concretely be ‘learned’. Accordingly, in his third objection, La Rocca insists that the impossibility to learn philosophy in both these senses is a sign that the idea of philosophy or metaphysics cannot be realized.

First of all, let me emphasize that I agree with La Rocca that philosophy, understood as a doctrine of wisdom that provides guidance for how to become virtuous, is unrealizable. This is due to the fact that this doctrine would require us to become examples of perfect morality, which is not possible for us.Footnote ⁷ But this does not mean that metaphysics as a science is not realizable. I believe that claiming that metaphysics as a science is realizable is compatible with Kant’s claim that philosophy cannot be learned. This can be appreciated when we consider that the claim is directed not only to philosophy understood as the ‘archetype’ that is not realized yet but also towards the different systems of philosophy that have been proposed historically by various philosophers. With respect to these, Kant claims that we can certainly learn them ‘historically’ (A837/B865; see also A836/B864), namely by memorizing a set of claims without putting them under rational scrutiny. This would be an inadequate way of learning philosophy. What we should do is learn to exercise our rational capacities in evaluating those claims. But this already means learning to philosophize. We can make a similar evaluation when we consider the ‘correct’ system of metaphysics, if it were ever realized. We could say that it would be wrong to learn it ‘historically’ and that we should approach it by exercising our rational capacities in evaluating its cognitions. This would mean learning it by learning to philosophize.

In his last objection, La Rocca argues that method has not only to do with attaining systematicity as a condition of science. Kant also relates the concept of method to the notion of a ‘way of thinking’ (Denkungsart). La Rocca provides a very interesting standpoint on this notion. He suggests that establishing a new ‘way of thinking’ is key to the worldly concept of philosophy. This new ‘way of thinking’ does not have the attainment of knowledge as its main goal. The goal becomes attaining ‘wisdom’, as a practical stance oriented towards the realization of reason’s practical ends. La Rocca uses the concept of Habitus introduced in the third Critique to clarify what this new way of thinking is. The latter should be taken as a practical attitude or stance that guide our actions in our pursuit of reason’s ends. La Rocca suggests that seeing Kant’s discussion of belief from this perspective weakens the kind of commitment that belief requires. Kant’s argument for the rationality of the beliefs in God and immortality in the Canon are not meant to secure a place for those beliefs in the edifice of metaphysics. Rather, their goal is to support the new way of thinking indicated by the Critique.

I thank La Rocca for having pointed out this characterization of the notion of a way of thinking in Kant, which I find very interesting. What La Rocca’s analysis shows is that we can consider Kant’s justification of the beliefs in God and immortality from two perspectives. One perspective, which is the one that I examine in the book, considers these beliefs from the standpoint of the system of metaphysics that Kant wants to build and determines whether they can be rationally justified and are coherent with other cognitions in that system. The other perspective is the one of the philosopher as a teacher of wisdom. The aim of philosophy seen through this lens is putting us in the condition to practically pursue the highest good in our actions, where this might require developing the right ‘stance’, which, perhaps, also requires having certain beliefs.

Let me point out that belief can be understood as an attitude that is essentially ‘practical’ from both these perspectives. From the perspective of the system of metaphysics, this means providing an account of belief according to which this form of taking-to-be-true cannot count as knowledge and essentially depends on ends that we pursue in practice. Having this account of belief at our disposal, we can then establish that the only justified taking-to-be-true regarding God and immortality is belief. However, in doing this, we are not actually promoting the adoption of those beliefs as essential to developing the right stance for pursuing the highest good. This could be what the perspective of the philosopher as a teacher of wisdom does. But finally, I do not think that the possibility of approaching the beliefs in God and immortality from this second perspective has any bearing on the question of how ‘strong’ our commitment towards the existence of God and the immortality to the soul should be. After all, it is possible that developing the right stance for our practical pursuit of the highest good requires a very strong commitment. In the Jäsche Logic, we read that moral belief ‘is often firmer than all knowledge’ (9: 72).