The Origin and Validity of Root Concepts, the Place of General Logic Within Transcendental Logic and Kant’s Critique of Dogmatism: A Response to My Critics

1 Response to Filieri

I have already mentioned that the part of transcendental philosophy which is established within the pages of the Critique of Pure Reason concerns root concepts. With respect to these, transcendental philosophy needs, first, to identify them and track their origin, and second, to establish what kind of validity they have. In my account, the first task is carried out by metaphysical deductions, while the second by transcendental deductions (Gava 2023, pp. 5, 73). In his comments, Luigi Filieri raises some questions that are related to fundamentally three issues: the way in which I confine Kant’s investigation into the origin of root concepts within their metaphysical deductions; my reading of the transcendental expositions of space and time as their transcendental deductions; and my reconstruction of the second step of the transcendental deduction of the categories in the B-edition of the Critique.

Let me begin with the first issue. Filieri suggests that the question of the origin of root concepts might not be the only concern of metaphysical deductions, as I suggest. Rather, determining the origin of these concepts might be the task of an investigation that should be characterized as ‘transcendental’. Filieri finds indications that this might be the case respectively in Kant’s notion of the acquisitio originaria of space, time, and the categories and in his characterisation of the ‘transcendental topic’ in the Amphiboly of the Concepts of Reflection. The notion of an ‘original acquisition’ points toward a way of considering the origin of space, time and the categories which is different from determining how they originate in different faculties, as the metaphysical deductions do. Additionally, the task of a transcendental topic is exactly that of finding the ‘transcendental place’ of a certain representation either in sensibility or the understanding, where this seems connected to locating its origin.

One quick answer to Filieri’s first worry is this: since I regard metaphysical deductions as essential parts of transcendental philosophy, I do think that their investigation into the origin of root concepts is part of an inquiry that can be called ‘transcendental’ in a broader sense.^[2] Yet, I do not think that this answer is sufficient for responding to Filieri’s point. For his remarks invite me to say something more on the roles of Kant’s notion of an original acquisition and of his transcendental topic within transcendental philosophy.

As far as the former is concerned, it seems to me that even though the notion of an original acquisition of the root concepts is important within Kant’s transcendental philosophy, it does not belong to either the metaphysical or the transcendental deductions. The notion provides a clarification of Kant’s claim that these concepts are a priori. It explains that this is very different from saying that they are innate. They arise as soon as we start to use the faculties to which they belong without depending on the particular objects that set this use in motion. While this clarification is important within Kant’s transcendental philosophy because it is a way of distinguishing Kant’s view from Leibniz’s and Eberhard’s,^[3] it seems to me that it does not fit Kant’s account of the tasks of the metaphysical and transcendental deduction, respectively. One way to support this point is to note that Kant’s notion of original acquisition concerns an aspect that our representations of space and time, on the one hand, and the categories, on the other, have in common. All these representations are ‘originally acquired’. By contrast, the focus of the metaphysical and transcendental deductions on the origin and validity of these representations is intended to display something that is specific to either our sensible or intellectual a priori representations, respectively. In this sense, tracking the origin of the concept of space is instrumental to determining what makes it radically different from, say, the category of cause.

Let me now consider how Kant’s transcendental topic addresses the issue of the origin of our representation. I think there are two possible ways to account for this doctrine. On the one hand, one can take Kant’s description of the transcendental topic as an account of what he does in the Critique when he attributes some a priori representations to sensibility and some to the understanding, respectively. Understood in this way, Kant’s notion simply emphasizes one of the achievements of transcendental philosophy. This means that speaking of a transcendental topic that traces our representations back to their transcendental place is only another way of describing what the metaphysical deductions do when they track the origins of root concepts.

However, there is another possible way of reading Kant’s transcendental topic that links it more closely to problems and issues in the Amphiboly. In this chapter, Kant identifies the conditions to properly use what he calls ‘concepts of reflection’. These concepts are concepts that govern the comparison between other concepts. Kant believes that the conditions for properly using the concepts of reflection are different depending on whether the concepts that are being compared are concepts of objects that are given through sensibility or concepts of objects that are given through the pure understanding (see Gava 2019). Kant relates the notion of a transcendental topic to the task of identifying the right conditions for comparing concepts. Accordingly, identifying the transcendental place of a concept means determining whether the object of that concept is given to us through sensibility or the understanding, so that we can determine under which conditions we should use the concepts of reflection for comparing the concept in question with other concepts.^[4] This operation, which is linked to Kant’s notion of transcendental reflection, could be considered part of his transcendental philosophy. However, it does not specifically concern either root concepts or their metaphysical or transcendental deductions.

The second issue on which Filieri insists concerns my reading of the metaphysical and transcendental expositions of the concepts of space and time as their metaphysical and transcendental deductions, respectively. In contrast to my proposal, Filieri argues that Kant’s choice to call these arguments ‘expositions’ is no coincidence and has to do with the specific nature of the representations that are being considered. Filieri concedes that it might be legitimate to speak of ‘concepts’ of space and time. However, since these concepts depend on space and time as forms of intuitions, they require a different treatment in comparison with the categories. This is especially true for the transcendental expositions of these concepts. Filieri argues that it is only in the case of the pure concepts of the understanding that the juridical question concerning their legitimacy arises, where this question is addressed by their transcendental deduction. By contrast, in the case of our representations of space and time, their legitimacy automatically follows from the fact that all intuitions must necessarily agree with them as forms. Accordingly, we only have a transcendental exposition of the concepts of space and time.

To respond to this second worry, let me emphasize that when I say that, for Kant, we have concepts of space and time, I mean ‘concept’ in a strict sense and fully accept the consequences of this claims, including the fact that these concepts are a product of the understanding. For example, we use the concepts of space and time when we form judgments on them, saying that ‘space is X’ or ‘time is Z’. This is compatible with saying that we also have intuitions of space and time and that these are more fundamental than our concepts, since the latter are derivative with respect to the former (Gava 2023, pp. 77, 83–5). The fact that we can have concepts of space and time is important to understand how we can (illegitimately) attribute temporal and spatial properties to objects that we cannot in fact intuit. For example, think of the First Antinomy. We attribute spatial and temporal properties to the totality of the world, saying that its spatial and temporal extension is either finite or infinite, but we cannot in fact intuit this totality. This is a way of using the concepts of space and time that is not warranted by our intuitions of them and is illegitimate.

The possibility of an illegitimate use of the concepts of space and time is relevant for the question whether the problem of the legitimacy of our root representations arises only for the categories or for the concepts of space and time, too. It might be true that the intuitions of space and time are immediately legitimated by the fact that they are the forms of our intuition. It might also be true that the possibility of wrongly applying them is ruled out by the simple fact that they are these forms. However, these considerations do not apply to the concepts of space and time. We can have illegitimate uses of these concepts, which renders the question of their legitimacy pressing. This is exactly Kant’s point when, at the beginning of the transcendental deduction of the categories, he refers back to his analysis of space in the Aesthetic and claims that what made a transcendental deduction of the concept of space ‘unavoidably necessary’ was our tendency to use that concept ‘beyond the conditions of sensible intuition’ (A 88/B 120–1). Therefore, there is a normative question linked to our use of the concepts of space and time and it seems appropriate to use the term ‘deduction’ in its juridical sense.

Let me now address the third issue that Filieri raises. In my reconstruction of the B-transcendental deduction, I argue that the role of its second step is to establish that to any synthesis of an empirical manifold of intuition there must correspond a synthesis of a pure manifold. The necessary agreement between these two syntheses explains how the categories can provide a priori cognition that constrains possible experience (Gava 2023, pp. 141, 145–9). Filieri worries that this interpretation of Kant’s argument blurs the distinction between cognition through the construction of concepts, which is mathematical, and cognition through concepts, which is philosophical. This has consequences for how we regard both empirical cognition and philosophical cognition. As far as the former is concerned, my proposal might end up endorsing a form of extreme ‘conceptual constructivism’, according to which every use of concepts, including empirical concepts, implies some form of mathematical or geometrical construction. As far as the latter is concerned, Kant’s claim that the methods of mathematics and philosophy are fundamentally different might require revision.

Let me start from empirical cognition. Indeed, I do think that when we synthetize an empirical manifold through the categories, we also use them to synthetize a pure manifold in a way that has many similarities with mathematical construction. That Kant proposes a view of this kind is clear from one example he uses in § 26 of the B-deduction:

Thus if, e.g., I make the empirical intuition of a house into perception through apprehension of its manifold, my ground is the necessary unity of space and of outer sensible intuition in general, and I as it were draw its shape in agreement with this synthetic unity of the manifold in space. This very same synthetic unity, however, if I abstract from the form of space, has its seat in the understanding, and is the category of the synthesis of the homogeneous in an intuition in general, i.e., the category of quantity, with which that synthesis of apprehension, i.e., the perception, must therefore be in thoroughgoing agreement. (B 162)

In the example, Kant suggests that when I empirically experience a house, I perform an operation on space as the form of outer intuition in that I draw the shape of the house. This operation is guided by the categories of quantity. Accordingly, it seems that when I experience an object with the shape of a house, I perform a construction that is similar to a construction of that shape in a purely geometrical space. Of course, there are also important differences between the two cases, since in the case of the empirical house, my construction is constrained by the empirical object. Nonetheless, there is a component of that experience that rests on a construction of a pure spatial manifold.

Having said that, I want to emphasize that my reading of the second step of the B-transcendental deduction only concerns the role of the categories in performing such constructions. I do not say anything on the question whether our use of empirical concepts (e.g. the concept of a house) requires a similar construction. It might be true that our use of empirical concepts in cognition requires a ‘constructive’ operation. After all, it is plausible that the spatial and temporal components of our empirical concepts depend on constraints that are set by the categories of quantity. This suggests that, in using those empirical concepts, we are also using the relevant categories in a way that determines a synthesis in both pure and empirical intuitions. However, this is not a view that I defend in the book. My claim concerns only the categories and their use in empirical cognition.

So, I do endorse a form of ‘cognitive constructivism’, as Filieri suggests, but this only concerns the use of the categories in empirical cognition and leaves the question whether a similar constructivism is required for empirical concepts open. What consequences does this position have for Kant’s distinction between mathematics and philosophy? I do not think that accepting this constructivism blurs the distinction in question. There is a fundamental difference between having empirical cognition and having a philosophical theory about how that cognition takes place. While having the empirical cognition in question does require a construction which is analogous to a mathematical procedure, developing a philosophical theory about that cognition does not need a similar construction. At best, it only requires a reflection on the role of construction in it. This is similar to what we do when we do philosophy of mathematics. For Kant, a philosophical account of mathematical cognition certainly requires an explanation of what construction is. However, this is different from actually providing constructions of concepts within philosophy.

2 Response to Dalla Rosa

If the part of transcendental philosophy that is established withing the Critique has to do with root concepts in general, Kant’s analyses of the categories and the ideas, as instances of root concepts, belong to transcendental logic, which is the part of transcendental philosophy that investigates the faculty of the understanding in a broad sense. In his comments, Davide Dalla Rosa invites me to elaborate on the relationship between general logic and transcendental logic, focusing in particular on the role that the former could play in the metaphysical deductions of the categories and the ideas, respectively.

Dalla Rosa’s reconstruction of my reading of the metaphysical deduction of the categories draws on three claims I make: the general claim that the metaphysical deductions contribute to establishing Kant’s distinction between faculties of cognition, the claim that the metaphysical deduction of the categories focuses on the understanding in the narrow sense, as the faculty of concepts, and, finally, the claim that the table of judgments from which Kant derives the categories is already part of transcendental logic. Dalla Rosa suggests that these claims commit me to the following assumptions. First, since the metaphysical deductions determine a distinction between faculties by identifying different root concepts, my analysis of the faculties prioritize their ‘products’ over their ‘activities’. Second, given that the table of judgments belongs to transcendental logic, general logic does not play any relevant role in Kant’s argument. Third, the fact that the metaphysical deduction of the categories both focuses on the understanding in the narrow sense and is a work in transcendental logic suggests a particular way of considering the division of labour between transcendental logic and general logic. The former considers faculties in their narrow sense by focusing on their products, while the latter analyses the activities of the understanding in the broad sense.

In responding to Dalla Rosa, I wish first to show that my reading of the metaphysical deductions of the categories does not commit me to these assumptions. Let me start with the first assumption, namely the one according to which the analysis of the activity of the understanding plays little or no role in my reconstruction of Kant’s argument. It is true that I claim that Kant’s investigation into the origin of root concepts contributes to establishing his distinction between faculties. But this claim does not rule out that the investigation in question also considers the activities of said faculties. In fact, according to my reading of the metaphysical deduction of the categories, a fundamental step in it establishes that the only way in which we can attain cognition through concepts is by using them in judgments (Gava 2023: 92–6). Establishing this priority of judgments over concepts is fundamental to Kant’s strategy, which consists in deriving the categories, as root concepts, from the forms of judgments. But this priority also shows that the metaphysical deduction of the categories does not neglect the activities of the understanding. Rather, it fundamentally relies on them.

This brings me to the third assumption that Dalla Rosa attributes to me, according to which general logic considers the activities of the understanding in the broad sense, while transcendental logic would attain a distinction between different faculties by identifying their different products. The fact that the activity of judging does play a relevant role in Kant’s derivation of the categories already shows that transcendental logic considers the activities of the faculties, too. Similarly, it is not the case that general logic only focuses on the activities and neglect the products. Clearly, general logic can distinguish between types of representations that are the results of the activities of different faculties. Accordingly, it can identify properties of concepts in general or distinguish between different types of judgments and inferences. The fact that Kant insists that general logic abstracts from the content of our cognitions does not imply that it cannot considers different types of representations as the products of the activity of thinking. But this means that both general logic and transcendental logic consider the activity and the products of thought, but from different perspectives.

As we cannot draw a distinction between general and transcendental logic by saying that the former considers the activity of thinking while the latter focuses on its products, we cannot distinguish between them by claiming that general logic considers the understanding in the broad sense and transcendental logic investigates the faculties in the narrow sense. The understanding in the broad sense is the higher faculty of cognition and comprises the understanding in the narrow sense as the faculty of concepts, the faculty of judgment and reason in the narrow sense as the faculty of inference. It seems to me that both general logic and transcendental logic can make use of the distinction between the understanding in the broad sense and the faculties that are part of it. Accordingly, my claim that the metaphysical deduction of the categories, as an argument within Kant’s transcendental logic, focuses on the understanding in the narrow sense should not be taken to imply either that transcendental logic only considers the faculties in their narrow sense or that general logic is exclusively concerned with the broad sense of the understanding.

Let me now say something regarding the second assumption that I identified above. Dalla Rosa is right that I claim that the table of judgments in the metaphysical deduction of the categories already is part of transcendental logic. But this is not meant to argue that general logic is not relevant within Kant’s argument. Rather, it is meant as a partial response to those critics who criticize Kant because he assumes a classification of judgments that was customary in logic textbooks of his time without providing an argument for it (Gava 2023, pp. 98–100). Given that the table is already a result of an investigation within transcendental logic that draws some distinctions among types of judgments that are not made in general logic, it is not the case that Kant simply took a traditional table for granted.

Here, let me emphasize a point. When I claim that the table of judgments in the metaphysical deduction is a work within transcendental logic, I base my contention on the consideration that there are some distinctions between types of judgments that we make within transcendental logic but not within general logic. But this is compatible with saying that some other distinctions within transcendental logic simply reflect those of general logic. Characterizing general logic and transcendental logic in this way would allow us to clearly distinguish between the two, while maintaining that the former informs the latter in relevant ways. In saying this, I do not want to defend a particular way of constructing the relationship between general logic and transcendental logic. I simply want to show that my reconstruction of the metaphysical deduction of the categories is compatible with recognizing a role for general logic within it.

Now that I have clarified my position in relation to the assumptions that Dalla Rosa regards as implied by some of my claims, I can consider the issues on which he invites me to reflect. In general, these concern the role of general logic within transcendental logic. One first issue is related to one possible way in which general logic might inform Kant’s metaphysical deduction of the categories which would be compatible with my reconstruction of Kant’s argument. We saw that, according to Dalla Rosa, I regard the metaphysical deduction of the categories as only focusing on the products of the understanding in the narrow sense and as a failing to consider its activities. Dalla Rosa points out that my account of the metaphysical deduction of the ideas is different, since the latter does consider reason in the narrow sense as the faculty of inference. Therefore, it seems that the metaphysical deduction of the ideas is informed by general logic, insofar as the analysis of the activity of inferring belongs to the latter. Dalla Rosa suggests that a way to show the relevance of general logic for the metaphysical deduction of the categories is to claim that both the metaphysical deduction of the categories and the metaphysical deduction of the ideas equally depends on an investigation of the activity of the understanding in the broad sense as the faculty of rules, where this analysis is carried out by general logic.

The clarifications on my positions that I provided above are here useful to respond to this proposal. First of all, I have suggested that my reconstruction of the metaphysical deduction of the categories does not neglect to consider how the understanding in the narrow sense ‘acts’. Accordingly, I have emphasized how the characterization of judgments as the functions through which we use concepts is essential in Kant’s argument. Dalla Rosa is right that I emphasize the pluralism in the argumentative procedures of the different metaphysical deductions. In this case, however, this pluralism does not imply that the metaphysical deduction of the categories fails to consider the activity of judgment as essential to the understanding in the narrow sense.

Furthermore, I have already suggested that my claim that the table of judgments is part of transcendental logic does not rule out that this table is essentially informed by general logic. This already makes room for recognizing an important contribution of general logic in Kant’s metaphysical deduction of the categories. Additionally, since I do not pair general logic with an investigation directed toward the understanding in the broad sense, I do not think that in order to see how general logic is relevant for the metaphysical deductions one need to add a layer of analysis that is only concerned with this sense of the understanding.

There is one last issue raised by Dalla Rosa that I want to discuss. One central claim in my book is that the Critique of Pure Reason is the particular doctrine of method of metaphysics. As such, the Critique identifies methodological rules that are specific to metaphysics and should be distinguished from the doctrine of method of general logic, which provides methodological rules that apply to any science. Dalla Rosa asks whether this sharp distinction between the doctrine of method of metaphysics and the doctrine of method of general logic has consequences for the possibility of identifying a role for general logic within transcendental logic, which is part of metaphysics in my reconstruction.

To answer this question, let me first point out that the distinction between these two doctrines of method does not rule out that the doctrine of method of general logic informs the doctrine of method of metaphysics. After all, since the former provides rules that apply to any science, while the latter identifies methodological procedures that only apply to metaphysics, it is plausible to think that in defining its specific rules, the doctrine of method of metaphysics takes the rules of the doctrine of method of general logic into consideration, while specifying them further. The problem with the doctrine of method of general logic is that the rules identified by it only provide minimal instructions regarding how we should proceed in science. This is the case because it is only when we take into account the nature of the cognitions belonging to a particular science that we can identify methodological rules that are informative (Gava 2023: 46–7). But this does not mean that particular doctrines of method, in identifying their methodological rules, do not need to respect the scarcely informative rules of the doctrine of method of general logic.

There is a further consideration that is relevant in relation with Dalla Rosa’s question. For the fact that the Critique of Pure Reason is the doctrine of method of metaphysics might be taken to have consequences for the ‘nature’ of the cognitions that are part of metaphysics. The Critique, as a particular doctrine of method, identifies rules that are object- and cognition-dependent. Namely, it identifies rules that are specific to the objects and cognitions that are part of metaphysics. One might wonder whether this means that metaphysical cognitions are not ‘general’, in the sense that they do not identify principles that generally apply to any cognition of objects. In turn, this might be used to claim that transcendental logic, as part of metaphysics, is not ‘general’ in nature, but identifies principles that only apply to a particular set of objects and cognitions. Furthermore, since general logic is ‘general’ in nature, this means that it does not play a key role within transcendental logic.

Another way to put this point would be to say that my account of the Critique of Pure Reason as a particular doctrine of method is incompatible with interpreting transcendental logic as a part of general logic (a reading along these lines has been developed by Tolley 2012). However, I do not think this is the case. The fact that the doctrine of method of a particular science identifies rules that are specific to the cognitions of that science has no implication on whether the cognitions belonging to that science are ‘general’ or ‘particular’ in nature. To see this, consider the possibility of developing a particular doctrine of method of general logic. This would be a discipline that identifies methodological rules that are specific to general logic. As such, it would single out procedures of investigation that are typical and specific to the investigations carried out in general logic in its endeavours to discover principles that apply to thinking in general. If the idea of a particular doctrine of method for a science that is ‘general’ in nature is not inconsistent in the case of general logic, the fact that there is a particular doctrine of method of metaphysics has no implication for the ‘general’ or ‘particular’ nature of the cognitions belonging to metaphysics. This means that my reading of the Critique is compatible with claiming that transcendental logic is ‘general’ in nature. Here, I am not arguing for this view. I simply want to emphasize that I can remain neutral on this issue.

3 Response to Walschots

Famously, Kant describes his critical philosophy as a remedy against dogmatism (B xxxv). He claims that Christian Wolff is the key defender of this approach (A 856/B 884) and specifies that rejecting Wolff’s dogmatism is not equivalent to rejecting his dogmatic procedure, which Kant sees as valuable in philosophy (B xxxvi). Chapter 8 of my book is dedicated to making sense of these contentions. In particular, I distinguish between three senses of dogmatism identified by Kant and claim that he regards Wolff as a dogmatist in two of these. I further clarify what Kant regards as problematic in these two forms of dogmatism and explain how Wolff can be a dogmatist in both of these senses. I also clarify what it means that philosophy can follow the dogmatic procedure once the Critique of Pure Reason has completed its job. In his comments, Michael Walschots focuses on my interpretation of the two senses of dogmatism that Kant attributes to Wolff, which I call dogmatism₁ and dogmatism_2. Walschots challenges both my reconstruction of their relationship and my claim that Kant’s critique of them operates at two levels.

I characterize dogmatism₁ as the attempt to establish metaphysical truths by the sole means of conceptual analysis (Gava 2023, pp. 235–7). This implies regarding metaphysics as entirely composed by conceptual truths. By contrast, dogmatism₂ is the use of synthetic a priori principles without an antecedent critique, where this critique is required to establish clear boundaries to their legitimate use (Gava 2023, pp. 237–40). I suggest that these two senses of dogmatism appear incompatible. Accordingly, at a first sight, it is hard to see how Wolff could be both a dogmatist₁ and a dogmatist₂. For, clearly, if Wolff, as a dogmatist₂, assumed synthetic a priori principles in his arguments, the results of those arguments could not be considered purely conceptual truths, as dogmatism₁ would require.^[5] Notwithstanding this apparent incompatibility between dogmatism₁ and a dogmatism₂, I argue that Kant can coherently regard Wolff as being both a dogmatist₁ and a dogmatist₂ at the same time when we see that his critique operates at two levels (Gava 2023, pp. 242–6). According to Kant, when Wolff assumes synthetic a priori principles without critique, he does so believing that those principles are only analytic. In this sense, while Wolff believes he is only establishing conceptual truths (according to Kant’s characterization of dogmatism₁), he actually assumes synthetic a priori principles without critique (according to Kant’s characterization of dogmatism₂). Kant’s critique of Wolff as a dogmatist₁ objects that one cannot establish metaphysical truths by conceptual analysis alone. This critique operates first and foremost at the metaphilosophical level, namely at the level of Wolff’s theory of method. By contrast, Kant’s critique of Wolff as a dogmatist₂ objects that he makes an illegitimate use of synthetic a priori principles. This critique operates at the methodological level, namely not at the level of what Wolff believes he does, but at the level of what he actually does.

Walschots advances two objections against my reconstruction of Kant’s critique of Wolff. First, he suggests that dogmatism₁ and dogmatism₂ are compatible even if we do not distinguish between two levels of critique as I do. Second, he argues that Kant’s critique of dogmatism₂ operates at the metaphilosophical level. Let me begin with the first objection. Walschots’s claim that dogmatism₁ and dogmatism₂ are compatible appears to rest on a characterization of them that is different from mine. In his view, to be a dogmatist₁, a philosopher needs to base her arguments in analyses of concepts and proceed by syllogisms. But this is compatible with assuming, as a dogmatist₂, some synthetic a priori principles as premises of those syllogisms. In this sense, Wolff can be both a dogmatist₁ and dogmatist₂ because, on the one hand, he begins his investigations with analytic definitions and uses syllogisms to establish further conclusions, while, on the other hand, he also assumes some self-evident principles that are synthetic a priori in Kant’s account. This description of Kant’s understanding of Wolff’s procedures of argument seems accurate to me. However, if we use my characterization of dogmatism₁ and dogmatism₂, it depicts Wolff as a dogmatist₂ and not as a dogmatist₁. Wolff is a dogmatist₂ because he assumes synthetic a priori principles without critique. However, the description does not depict him as a dogmatist₁ because, given that the conclusions of his syllogisms rest on synthetic a priori principles, those conclusions cannot be purely conceptual truths.

Let me now focus on the second objection. Walschots argues that Kant’s critique of Wolff as a dogmatist₂ operates at the metaphilosophical level, namely at the level of Wolff’s beliefs and theory of method. Additionally, he suggests that my claim that the critique operates at the methodological level is hard to follow. Walschots’s worry seems to be based on the following considerations. In my view, Kant regards Wolff as assuming synthetic a priori principles while believing that they are analytic. Accordingly, Kant’s criticism clearly concerns what Wolff believes. Kant criticises Wolff for wrongly believing that some principles he assumes are analytic, where they are instead synthetic a priori and accordingly require a proper critique. Walschots supports his reading with an analysis of Kant’s critique of Wolff’s description of the principle of contradiction. Kant believes that Wolff’s account of the principle makes the principles synthetic. In Walschots’s view, Kant’s critique targets Wolff as a dogmatist₂ because he assumes a synthetic a priori principle while believing that it is analytic.

While Walschots claims that we do not need to distinguish between two levels of critique in order to make sense of Kant’s criticism of Wolff as a dogmatist₂, let me first note that Walschots’s own account of this criticism rests on a distinction between what Wolff believes and what Wolff does. Wolff believes he assumes analytic principles but he is wrong because he in fact assumes synthetic a priori ones. So, it seems that we cannot get rid of the distinction between a metaphilosophical level and a methodological level of Kant’s analysis even in Walschots’s reconstruction of it. So why does Walschots insist that the critique operates at the metaphilosophical level against my view? Since, for Walschots, the main problem of Wolff’s way of proceeding as a dogmatist₂ has to do with his beliefs, it appears that Kant is mainly concerned with the fact that these beliefs are false in his account. Accordingly, the main upshot of Kant’s critique would be to show that the principles that for Wolff are analytic are in fact synthetic a priori.

I believe this way of describing Kant’s critique of Wolff as a dogmatist₂ fails to grasp Kant’s main concern with his use of synthetic a priori principles. For Kant is not really worried that Wolff’s beliefs on the principles he uses are wrong. Rather, he is worried that Wolff’s use of those principles is unjustified and leads him to unwarranted assertions. In this sense, Kant’s critique is not directed at Wolff’s self-understanding as a philosopher. His aim is not to correct Wolff’s beliefs in a way that they reflect what he actually does. Rather, the critique is directed against what Wolff actually does, which is using synthetic a priori principles without a proper account of the conditions of their rightful application. Walschots is right that Kant attributes to Wolff a false belief concerning the synthetic a priori principles he assumes. However, the upshot of this attribution is not primarily to criticise Wolff for that belief, but rather to provide an explanation for Wolff’s uncritical assumption of those principles. Accordingly, as I argue, the target of Kant’s critique of Wolff as a dogmatist₂ is the method he actually follows, not his theory of method or his believes concerning what that method consists of.

How should we read Kant’s critique of Wolff’s description of the principle of contradiction? Certainly, one layer of this critique concerns the fact that Wolff provides a wrong account of that principle, given that the principle is analytic for Kant. This criticism, however, does not target Wolff as either a dogmatist₁ or a dogmatist₂. The point is not that Wolff cannot establish metaphysical truths since he only proceeds by conceptual analysis. The point is not that Wolff assumes synthetic a priori principles without critique either. Rather, Kant simply complains that Wolff provides an incorrect description of the principle.

There might be a second layer in Kant’s critique of Wolff’s description of the principle of contradiction, which is closer to Walschots’s account. According to the latter layer, Kant does indeed target this description as an instance of dogmatism₂. Namely, Kant targets Wolff’s use of the principle because it results in an application of a synthetic a priori principle without an antecedent critique. But, again, Kant’s main problem in this case is not with Wolff’s beliefs regarding the principle. The problem is not that Wolff believes the principle is analytic while his formulation of it makes it synthetic. Rather, the problem lies in his use of the principle, which, given that the principle is synthetic, could result in unwarranted conclusions.

4 Conclusions

Let me conclude this response by thanking again Luigi Filieri, Davide Dalla Rosa and Michael Walschots for their insightful and useful comments. I hope that I was successful in taking them up on their invitation to further reflect on some of the implications of my interpretation.