In the preface to second edition of the Critique of Pure Reason, Kant famously characterizes the project he is about to undertake as one of setting metaphysics on the ‘secure path of a science’ (Bxiv). The idea is that the Critique provides insights that will enable metaphysics to be transformed from a discipline in which there is no significant progress and which does not have the same intellectual standing as mathematics or physics into a discipline that does enjoy similar standing and that, like those sciences, delivers genuine knowledge. What are these insights and how does the Critique seek to bring about this transformation? In Kant’s Critique of Pure Reason and the Method of Metaphysics Gabriele, Gava argues that the key to answering these questions lies in appreciating that the Critique provides the so-called doctrine of method of metaphysics. In Gava’s view, the chief task of a doctrine of method is to explain how the form of a science is to be imparted to an area of inquiry. This task contrasts with that of a doctrine of elements, which lays out the cognitions that make up the body of the science. A doctrine of method, then, explains how these cognitions must relate to one another if they are to exhibit the form of a science. The central characteristic of a science, in Kant’s view, is a demanding kind of unity, which Gava calls architectonic unity. So a doctrine of method must explain how architectonic unity is to be imparted to a body of cognitions. Since this must take into account the specific subject matter of a given area of inquiry, each such area, including metaphysics, requires its own doctrine of method. Gava’s thesis, then, is that the overarching aim of the Critique of Pure Reason is to show how the various knowledge claims that are put forth in the course of metaphysical inquiries are to be unified in the demanding way required to achieve architectonic unity for metaphysics.

It is important to note that Gava views the Critique as a whole as giving the doctrine of method for metaphysics, not just the (comparatively short) part of it entitled Transcendental Doctrine of Method. He reads this part as, among other things, providing an account of the notion of architectonic unity. But it does not explain how such unity is to be implemented among the cognitions that prima facie belong to metaphysics. In Gava’s view, this task is confronted instead in various stretches of the Transcendental Doctrine of Elements, specifically in those stretches concerned with determining the limits of the kind of metaphysical cognition possible for human beings. Moreover, Gava holds that, even if the main task of the Critique as a whole is to develop a doctrine of method, this is compatible with saying that the Critique also seeks to establish substantive metaphysical cognitions; that is, some of the ‘elements’ of the science of metaphysics. This is because, as Gava sees it, to achieve its aim a doctrine of method for metaphysics must inter alia show how to remove possible conflicts among purported substantive metaphysical cognitions and must therefore be given in conjunction with a doctrine of elements.

The point here can be made clearer by drawing on Kant’s contrast between the critical, propaedeutic aspects of the Critique, on the one hand, and its doctrinal aspects, on the other (which Gava refers to as, respectively, the critique of pure reason and transcendental philosophy).Footnote ¹ Kant claims that while the Critique does not give the completed science of metaphysics (the ‘system’), it does offer both the entire outline of the system and substantial parts of it.Footnote ² This means that it is no mere propaedeutic to metaphysics, but, in addition to its properly propaedeutic function, also establishes some parts of the science itself. If we call this the doctrinal aspect of the Critique, we can put Gava’s point by saying that the critical aspect – the aspect that in Gava’s view is concerned with generating architectonic unity in metaphysics – could not do its job unless the Critique also had a doctrinal aspect.

The focus of the Critique, however, is on the critical aspect. Since Gava interprets this aspect as developing a doctrine of method, this is also the lens through which he reads Kant’s comments that the Critique is a ‘treatise on the method’ rather than substantive metaphysics.Footnote ³ On this view, it is a treatise on the method precisely in the sense that it offers a doctrine of method – which in turn is what is needed to turn metaphysics into a science because the main obstacle to scientific status is the lack of architectonic unity among putative metaphysical cognitions.

Gave supports this reading of the Critique through in-depth discussion of central arguments from all of its major parts, including the Transcendental Aesthetic, the Transcendental Analytic, and the Transcendental Dialectic. There are numerous insightful and original points here, many of which deserve separate critical engagement. However, I will use this contribution to take a bird’s-eye view and focus on the guiding idea of Gava’s reading. Specifically, I will raise three related objections to this idea, two of which concern the claim that the primary reason why metaphysics has not achieved the status of a science is its lack of architectonic unity. The third concerns the claim that Kant characterizes the Critique as a treatise on the method because he intends it to offer a doctrine of method in Gava’s sense.

My first objection can be put by saying that the primary reason why metaphysics lacks scientific status pertains, not to the doctrine of method, but to the doctrine of elements. More precisely, as I read Kant, he takes the reason to be that it is not sufficiently clear that the elements needed for a science of metaphysics – certain kinds of synthetic a priori judgements – are available. It is not clear that these are available because there is doubt that such judgements are even possible. This appears to be what Kant has in mind when he frames the guiding question of the Critique as ‘How are synthetic judgments a priori possible?’Footnote ⁴ The idea is this: if there are metaphysical cognitions (and so, for the doctrine of elements of metaphysics to do its job of expounding these), they are, to a significant extent, synthetic judgements a priori. But as long as we cannot explain how such judgements are possible, we are not entitled to assume that there are such judgements with genuinely metaphysical content (this lack of entitlement does not affect the synthetic a priori judgements in mathematics and physics, according to Kant). And this means that we are not entitled to assume that even the elements for a scientific metaphysics are available.Footnote ⁵

If this is right, then contrary to what Gava claims, it will not be possible to turn metaphysics into a science by implementing the relevant doctrine of method. Gava’s account presupposes that the elements for such a science are available and that the main obstacle to attaining scientific status consists in the fact that not all putative elements cohere with one another. In other words, there is some chaff mixed in with the metaphysical wheat, and the task of the Critique (and of the doctrine of method of metaphysics, as Gava understands it) is to sort out the one from the other. This will then create the unity required for scientific status. But my point here is that the primary obstacle to attaining the status of a science for metaphysics is that, for all we know, there may simply be no wheat at all among the chaff. So the problem that has thus far prevented metaphysics from attaining scientific status is significantly more serious.

In support of this point, let us turn briefly to the issue of the Critique’s guiding question. Why does Kant frame this as a how-possible question? As I understand Gava’s account, it makes such a question look unmotivated, and this speaks against it. For on his account, the main task of the two disciplines that make up the Critique (i.e., transcendental philosophy and the critique of pure reason) is, respectively, to identify the metaphysical synthetic a priori cognitions that there are (along with arguing that they are indeed cognitions, i.e., objectively valid) and to show how to unify these into a system by arguing that the apparent lack of coherence among them is merely apparent; for example, by offering a solution to the antinomies of pure reason. But if this is what it takes to make metaphysics into a science, then why does it have to be explained how synthetic judgements a priori are so much as possible? It would seem more appropriate for the guiding question of the project to be something along the lines of ‘Which putative metaphysical cognitions actually belong to metaphysics?’ or ‘Can putative metaphysical cognitions be unified in the way required for a science?’

As is well known, Kant associates scepticism about the possibility of synthetic a priori judgements with Hume and often portrays Hume’s doubts as an understandable (if ultimately incorrect) reaction to the sort of overreach of which he accuses traditional Rationalist philosophers, who purport to establish (what are in fact) synthetic a priori truths about experience-transcendent objects merely ‘from concepts’ – for instance, by mistakenly construing such judgements as analytic judgements (‘dogmatism’).Footnote ⁶ How does one counter such scepticism? Kant’s strategy seems to be roughly this: Humean scepticism rests on a failure to consider all relevant alternatives for making the possibility of synthetic a priori judgements intelligible. To counter the doubt, therefore, what one must do is show that there is another relevant alternative; that is, another candidate option for making synthetic a priori judgements intelligible – and that this is one that actually succeeds in doing so.

We can describe the situation in terms of Hume’s Fork. Kant appears to see Hume’s scepticism about metaphysics as supported by an argument by elimination, which features Hume’s Fork as a premise. Thus, if there are metaphysical judgements (of the relevant sort), these are either relations of ideas or matters of fact; but, Hume argues, they are neither; so there are no such judgements. Obviously, this argument can be blocked by showing that Hume’s Fork is not an exhaustive list of alternatives. But how does one show this? By sketching an alternative that succeeds at the task at which the other two alternatives fail (even by Kant’s own lights), viz. making intelligible that purported metaphysical judgements do indeed have the properties they are alleged to have (and must have if they are to be genuinely metaphysical). In other words, this alternative must be able to explain how it is possible for certain judgements that are not analytic truths to exhibit strict universality as well as a type of necessity that is not merely logical. And the account Kant offers in the Transcendental Aesthetic and Transcendental Analytic is intended to accomplish just this – which is why it can be characterized as being addressed to a how-possible question regarding synthetic a priori judgements.

This suggests that the question about the possibility of synthetic a priori judgements – the question that, according to Kant, structures the entire Critique – concerns the elements of metaphysics rather than its method. It concerns whether there are any cognitions at all that qualify as genuinely metaphysical – whether there is anything at all to be unified here, any matter to be organized so as to bear the form of a science called ‘metaphysics’. This is the main obstacle to be confronted if metaphysis is to become a science. While Gava is of course right that Kant requires architectonic unity for metaphysics to become a science, the question of how to generate such unity among the cognitions purporting to belong to metaphysics is not the concern that drives the Critique.

It might be objected that Gava’s reading derives support from the antinomies of pure reason and that the account I just sketched is on shakier ground here. But this is, to say the least, not obvious. Kant appears to regard the existence of antinomies as raising a challenge for the very possibility of rational cognition from concepts, not just for the less fundamental problem of ordering an extant body of purported cognitions of this sort in the way required for a science, as Gava proposes.Footnote ⁷ If this is right, then the antinomies, too, support the idea that Kant’s overarching concern in the Critique is one belonging to a doctrine of elements rather than a doctrine of method.

The second problem I wish to raise for Gava is related to the first and concerns his interpretation of Kant’s claim that the Critique is a ‘treatise on the method’ (Bxxii) of metaphysics rather than the system of metaphysics itself. Again, on Gava’s interpretation, what Kant signals here is that the Critique gives the doctrine of method for metaphysics. But the context of the passage from the B-Preface suggests otherwise, since Kant here articulates the idea of the so-called Copernican Turn for metaphysics, and it is clear that this is what the talk of method in this passage refers to. This suggests that the Critique is a treatise on method because it presents an ‘attempt to transform the procedure previously followed in metaphysics’ (Bxxii). ‘Procedure’ (Verfahren) is clearly a synonym of ‘method’ here. According to this passage, then, the Critique is a treatise on method because it articulates a new way of proceeding in metaphysics. The new way consists in starting with an account of the capacity for a priori cognition, which puts the metaphysician in a position to explain how synthetic judgements a priori are possible and, in addition, serves as a basis for identifying those metaphysical cognitions of which we are capable. This account includes, as one of its central elements, the Two Stems doctrine, fully appreciating which enables the metaphysician to avoid the sorts of amphibolies that Kant sees at the root of Leibniz’s metaphysical doctrines and thereby refrain from overstepping the narrow boundaries within which alone theoretical rational cognition from concepts is possible for us. If this is right, the benefits of adopting the new procedure are to be found primarily in the doctrine of elements of metaphysics, in the sense that the new procedure first makes it possible to explain the possibility of synthetic judgements and thereby establish that there are genuine elements for a science of metaphysics. Here, too, it looks as if Gava’s thesis that the Critique is primarily concerned to articulate the architectonic unity of metaphysics and that this captures the motivation for Kant’s remark in the B-Preface that the Critique is ‘a treatise on the method’ leaves out an important part of that motivation. If I am right about the context of the remark, the method at issue is the kind of faculty analysis demanded by the Copernican procedure. To the extent that the Critique is devoted to an analysis of our cognitive faculties, then, its primary concern is not with the architectonic unity required for metaphysics. Instead, Kant seeks to pioneer the method demanded by taking up the Copernican perspective on metaphysical knowledge.

This last point brings me to my third objection, which relates to Gava’s account of the architectonic unity metaphysics must exhibit if it is to be a science. It is part of Gava’s view that there are two reasons why the metaphysical cognitions set forth in the Transcendental Doctrine of Elements of the Critique do not yet exhibit architectonic unity: the first is that they include incompatible claims, most notably the antinomies of pure reason; the second is that they are not organized in accordance with an idea of the whole of the putative science, which Kant requires for architectonic unity and which grounds knowledge of the articulation of the science as well as its extent and can thus also ground claims to completeness. Call such an idea the governing idea of a science. Gava argues that for Kant the governing idea of metaphysics is the ‘cosmopolitan concept’ of philosophy, according to which metaphysics is ‘the science of the relation of all cognition to the essential ends of human reason’ (A839/B867). While Gava’s discussion of the cosmopolitan concept of philosophy is illuminating and must surely form part of an account of Kant’s conception of metaphysics, there are at least four reasons for doubting Gava’s thesis that it is this concept that functions as the governing idea of metaphysics. First, Kant is explicit that metaphysics as a whole divides into two parts, the metaphysics of nature and the metaphysics of freedom, each of which by itself already exhibits the unity of a system (that is, architectonic unity) but which ultimately belong together in a single system.Footnote ⁸ Now, the question how exactly theoretical and practical philosophy are related for Kant is a vexed one.Footnote ⁹ But my point here is just that if the metaphysics of nature – the ostensible topic of the Critique – already exhibits architectonic unity even in isolation from the metaphysics of freedom, then the cosmopolitan concept of philosophy, focused as it is on the concerns of the metaphysics of freedom, is not a plausible candidate for being the idea that confers unity on the former.

The second reason why I am unconvinced that the cosmopolitan concept of philosophy serves as the governing idea of metaphysics is that this concept is not as central to the Architectonic chapter of the Transcendental Doctrine of Method as one would expect it to be if Gava’s account is right. This chapter is the place in the text where Kant gives his official account of the governing idea of metaphysics (though the idea is mentioned in several other places, notably in the Introduction). What Kant does here is, first, to expound the idea of metaphysics and, second, set out what he calls the schema of this idea, which is needed for knowing how the science governed by this idea is articulated into parts. But what functions as the idea of metaphysics in this account is the notion of metaphysics as rational cognition from concepts.Footnote ¹⁰ This is the idea that is then ‘schematized’ in the Architectonic chapter to generate the articulation of the metaphysics. The cosmopolitan concept of philosophy appears to play a role subordinate to the idea of rational cognition from concepts here.Footnote ¹¹

The third reason that speaks against the cosmopolitan concept proposal derives from a passage in the Introduction, A13-4/B27-8, that is important for Gava because it offers textual support for the idea that the main task of the Critique is to develop the doctrine of method for metaphysics. Kant says here that the critique of pure reason ‘is to outline the entire plan [for transcendental philosophy, i.e., metaphysics] architectonically, i.e., from principles, with a full guarantee for the completeness and certainty of all the components that comprise this edifice’. What Kant here calls a ‘plan’ for a science seems to correspond to the idea-cum-schema he discusses in the Architectonic chapter. And the fact that he tasks the entire work with the job of outlining this plan speaks against the suggestion that the idea on which it is built is the cosmopolitan concept of philosophy, given that this concept has only a subordinate role to play in this plan. This is further confirmed by Kant’s suggestion, in the continuation of the passage, that transcendental philosophy (i.e., metaphysics) does not include the metaphysics of freedom, but is instead ‘a philosophy of pure, merely speculative reason’ (A15/B29).

Finally, the fourth reason why I am unconvinced by the cosmopolitan concept proposal is that Kant identifies the faculty of reason (in the wide sense) itself as the source of the architectonic unity of metaphysics. This is most evident in his notorious claim that the Table of Categories is complete. Whether or not this claim is convincing or adequately supported, it is clear that Kant thinks the completeness of the table is guaranteed by the fact that it has architectonic unity: it derives from a single idea of the whole, a clear grasp of which puts one in a position to know how the table is to be articulated and when it is complete:

Kant says here that it is the faculty of understanding itself that serves as the principle from which the articulation of the system of concepts that the understanding originally contains can be derived. This principle therefore is a principle of architectonic unity, a point Kant makes explicit in the following passage:

Notice how this passage ascribes to the pure understanding all the hallmarks of architectonic unity: a whole of parts that are connected in such a way that it can be grasped from the idea of this whole that each part in fact belongs to it; that these are all the parts; and that the order (‘articulation’) of the parts, their relations to one another, is the correct order.Footnote ¹² But if the pure understanding itself is the principle that provides for the architectonic unity of the kind of cognition at issue here, i.e., metaphysics, then it looks as if the cosmopolitan concept of philosophy is not needed for this task – or else it would have to be shown that this concept, too, has its seat in the pure understanding itself. Kant confirms this point in the Architectonic chapter, when he says:

If metaphysics is to present pure cognition a priori in its systematic unity, and if this unity is one that pertains to such cognition in virtue of the fact that the cognition has its origin in a particular faculty, then what ultimately accounts for the architectonic unity of metaphysics is the faculty that generates metaphysical cognition.

In closing, I wish to connect the point I just made that the faculty of pure understanding itself provides for the architectonic unity of metaphysics, to my earlier suggestion that Kant calls the Critique a ‘treatise on the method’ because in it he proposes, and implements, the Copernican Turn; that is, an approach to metaphysics that is guided by an account of the nature of the mind which provides the basis for determining, on principled grounds, what kind of metaphysical knowledge we are, and are not, capable of. We can now see that the emphasis on method in this sense (that is, the emphasis on proceeding ‘faculty-first’) is tightly linked also to the issue of achieving architectonic unity for metaphysics. Kant appears to hold that the correct idea of metaphysics – the idea that governs this science and grounds its unity – can be grasped only by proceeding ‘faculty-first’. If this is right, a Copernican method is required for securing both the elements of metaphysics and its unity, for achieving the task of a doctrine of elements as well as the task of a doctrine of method. Put differently, Kant’s focus in putting metaphysics on the secure path of a science does not lie on developing a doctrine of method for metaphysics, in the technical sense of this term that Gava so clearly articulates. Instead, his primary aim is to devise, and apply, a new method for generating the elements of metaphysics – one that will also enable the transcendental philosopher to develop a metaphysical doctrine of method. But the latter achievement is possible only as a consequence of the first; that is, only because of the successful implementation of the new ‘faculty-first’ method that is the main goal of the Critique.