In her impressive book, Kant’s Reform of Metaphysics: The Critique of Pure Reason Reconsidered (De Boer Reference De Boer2020; hereafter referenced simply by page number), Karin de Boer argues that Kant’s primary goal in the Critique of Pure Reason is not so much to overthrow the claims and arguments of traditional metaphysics, but rather to reform metaphysics so that it can be established as a science. To this end, it investigates the thought of some of Kant’s immediate predecessors, especially Wolff and Crusius, traces the development of Kant’s views throughout his pre-critical and critical periods, and offers interpretations of a number of major parts of the first Critique, including the Transcendental Deduction, the Schematism, the Appendix to the Transcendental Analytic and the Architectonic. The contextual and developmental approach that is on display throughout the book made it possible for De Boer to avoid the kind of ahistorical reading that can be found all too frequently in Kant scholarship, and to shed important light on Kant’s actual intent in the first Critique.
De Boer’s main thesis, that Kant’s project is to reform traditional metaphysics, entails both that some things must be preserved and that some things must be rejected in metaphysics as it was practised in Kant’s day. Specifically, according to De Boer, Kant took over from Wolff ‘the idea of metaphysics as a comprehensive system of the concepts and principles constitutive of any cognition of objects as well as the idea that such a system ought to be established by means of a strict method’ (p. 42). In this way, Kant would be able ‘to preserve what he took to be the rational core of the metaphysical disciplines devoted to the soul, the world as such, and God’, since those are crucial to supporting the moral improvement of human beings and to thwarting the dangers of scepticism, naturalism and, one might add, atheism (p. 3). He rejected, however, ‘the assumption that the treatment of these ideas and their determination amounts to the cognition of objects’ (ibid.), and one of his reasons for rejecting cognition of such objects is that it is based on the (according to Kant, mistaken) ‘assumption that sensibility and thought are nothing but two different ways to obtain knowledge of things’, an assumption that she labels ‘continuism’ because sensible intuitions and intellectual thoughts are on a continuum as they differ merely in degree rather than in kind, as Kant wants to maintain (p. 17).
In my comments, I would like to focus attention on five topics that are directly relevant to De Boer’s main thesis. They concern: (1) practical philosophy, (2) the Principle of Sufficient Reason, (3) the analyticity or syntheticity of the fundamental principles of metaphysics, (4) Kant’s argument for the indispensability of sensible schemata for cognition and (5) the role of the Transcendental Dialectic in Kant’s criticism of traditional metaphysics. All of these topics are related, in one way or another, to De Boer’s claim that Kant’s critical philosophy should be viewed as reforming metaphysics rather than bringing about a revolution in philosophy.
First, Kant presents the fundamental principle of morality that underlies his entire practical philosophy not only as entirely new, but also as explicitly opposed to Wolff’s. For example, in the Groundwork of the Metaphysics of Morals Kant complains that Wolff’s practical philosophy is based (1) on a generic notion of volition that allows for both empirical and a priori motives rather than on a more specific principle that requires a motive that is a priori, and (2) on an indeterminate concept of perfection. From Kant’s perspective, these points reveal that Wolff’s position is heteronomous, in stark contrast to Kant’s commitment to an autonomous principle of morality that is based on reason’s legislation of the moral law. Insofar as the metaphysics that Kant most wants to rehabilitate is not theoretical – it cannot be, given the restrictions he places on our theoretical cognition – but rather practical, his criticism of the most basic principles of Wolff’s practical philosophy can seem to stand in some tension with the claim that Kant is reforming Wolff’s position.
Second, with respect to theoretical philosophy, Kant is typically thought of as being highly critical of all (or nearly all) metaphysical claims, but especially of those put forward by his rationalist predecessors, such as Wolff. For example, it is often claimed that Kant rejects the Principle of Sufficient Reason as dogmatic, which is consequential insofar as this principle is, for Leibniz and Wolff, one of the ‘primary truths’ that they put to use at many crucial junctures in their philosophical systems. Specifically, the Principle of Sufficient Reason is typically viewed either as unjustified – in his pre-critical period Kant explicitly rejected Wolff’s derivation of it from the principle of contradiction as confused – or as inapplicable to things in themselves, since applying it to things in themselves would (allegedly) violate Kant’s doctrine of epistemic humility. Indeed, many hold that Kant views with suspicion a close cousin of the Principle of Sufficient Reason, namely Kant’s own Supreme Principle of Pure Reason: if the conditioned exists, then so too does the totality of its conditions and thus the unconditioned. In fact, their view is that this principle is precisely the source of the transcendental illusion that leads us to make dogmatic claims about the objects of special metaphysics. Now, I do not accept these claims as I think that Kant is much more sympathetic to the Supreme Principle as a principle that can be justified on the basis of the notions involved in the principle (and Kant’s analysis of our faculty of reason). For there is significant textual evidence that Kant takes the Supreme Principle to apply to things in themselves. For example, in the Antinomy of Pure Reason, Kant claims:
If the conditioned as well as its condition are things in themselves, then when the first is given not only is the regress to the second given as a problem, but the latter is thereby really already given along with it; and, because this holds for all members of the series, then the complete series of conditions, and hence the unconditioned is thereby simultaneously given, or rather it is presupposed by the fact that the conditioned, which is possible only through that series, is given. (A498/B526; trans. throughout, Kant Reference Kant1998)
In fact, one can even make a case that it applies to appearances, though much would need to be said to articulate and defend this idea. However, viewed in this light, it seems that Kant is more of a reformer than a critic of Wolff on this point (though there are crucial differences between the Principle of Sufficient Reason and the Supreme Principle, given that the latter asserts the existence of the unconditioned, which is incompatible with the Principle of Sufficient Reason). Given this, is De Boer open to accepting the Supreme Principle and agreeing with my reformist view on this point? Or does she think that Kant’s rejection of continuism somehow entails that Kant must reject Leibniz’s and Wolff’s Principle of Sufficient Reason and view his own Supreme Principle as the source of metaphysical error?
Third, the reason that is often given for rejecting claims like the Principle of Sufficient Reason is that they are supposed to be analytic, in contrast with the synthetic a priori claims that Kant believes would have to constitute the core of metaphysics. Lanier Anderson (Reference Anderson2014) has recently argued that Kant’s ‘master argument’ against traditional metaphysics of the specifically Wolffian variety turns on the alleged poverty of the notion of containment that is fundamental to the analytic judgements that Wolff limits himself to. Insofar as Kant breaks with Wolff by distinguishing between analytic and synthetic judgements and then maintaining that the most crucial claims of metaphysics are synthetic, one might think, as Anderson does, that Kant has broken radically with Wolff. Whether one agrees that this line of thought represents Kant’s ‘master argument’ against the claims of traditional metaphysics or is rather a more limited metaphilosophical narrative that describes Kant’s first-order results from a certain distance, it does require some kind of response, especially given that De Boer (rightly) takes Wolff’s views seriously rather than dismisses all of his metaphysical claims as meaningless.
Now, De Boer does indirectly discuss these issues in chapter 3 and chapter 8, where she claims that ‘the pure understanding can establish metaphysics as a science by isolating its cognitions from the surreptitious impact of pure sensibility’ (p. 70) and ‘none of the metaphysical disciplines can consist of synthetic a priori judgments about things’ (p. 253). However, it is one thing to say that if one removes all sensible elements from the understanding’s concepts so as to form ideas of reason, this allows us to see the limits of sensibility more clearly, quite another to claim that the pure understanding can establish metaphysics as a science that must consist exclusively of analytic judgements such that Kant’s position would be a continuation of Wolff’s. While Kant may have thought that the fundamental principles of logic would have to be analytic, the metaphysics of nature that Kant envisions as one of the two parts of his system of philosophy would have to consist of synthetic a priori principles. Consider, for example, a passage from the Architectonic, where Kant explains how he is conceiving of philosophy: ‘Now the legislation of human reason (philosophy) has two objects, nature and freedom, and thus contains the law of nature as well as the moral law, initially in two separate systems, but ultimately in a single philosophical system’ (A840/B868). Though this passage does not specify that the laws of nature and the moral law are synthetic a priori principles, it does make clear that the system of philosophy that Kant is ultimately interested in consists of a priori laws of nature and an a priori moral law, and there are other passages that make it clear that these laws must be synthetic rather than analytic. So while I am sympathetic to seeing Kant as more of a reformer of Leibniz and Wolff on many points, it is not clear that it is possible to do so on this point.
Fourth, having suggested that Kant does not follow (and is right not to follow) Wolff in conceiving of philosophy as consisting in analytic principles, I now want to ask with what right Kant distances himself from Wolff’s position on a related point. In chapter 6, De Boer claims that ‘it is only in the Schematism Chapter that Kant fully develops the argument that supports his critique of post-Leibnizian metaphysics in the Transcendental Analytic’ (p. 163). In the course of this chapter, De Boer makes a number of interesting interpretative moves that deserve serious scholarly attention. For example, she holds that when the categories are used to think things as such, they are ‘nothing but deschematized pure concepts’ (p. 164). However, I would like to focus on one of the other central claims made in this chapter, namely that ‘any a priori cognition of objects rests on non-intellectual conditions he calls schemata’ which ‘occur wherever the human mind unifies a given manifold, except in the purported a priori judgments about things as such, the soul, the world as such, and God’ (p. 164). The point I would like to focus on is simply why a rationalist like Wolff should accept such a non-intellectual condition. In short, why is Kant not simply begging the question against rationalists like Wolff? Presumably, De Boer is indirectly acknowledging this point when she says: ‘Kant rejects a core assumption of Wolffian metaphysics, namely the assumption that the intellect can obtain a priori cognitions of things all by itself’ (p. 169). That is fair enough as a matter of interpretation, but one might wonder both why Kant is justified in rejecting that assumption and, if he does reject an assumption that is, as she notes, crucial to rationalist metaphysics, how he can still be reforming their metaphysics rather than rejecting it.
As her argument is developed further in this chapter, De Boer notes, picking up on Kantian phrases, that without the schemata our purely intellectual concepts would be empty and would not relate to an object. But what exactly does it mean to call a concept empty and precisely what kind of relation to an object would be missing if a concept does in fact refer to an object (because the object has the property that the concept represents it as having)? For example, if God happens to exist, then it can seem as if (1) the claim that ‘God exists’ is true, (2) its truth is based on the concept referring to God and (3) in some non-trivial sense the concept of God is not empty precisely because a relation of reference obtains between the concept (as it is used in the judgement) and the object. Now, the judgement that ‘God exists’ naturally does not amount to cognition insofar as cognition is a special mental state that requires the satisfaction of further conditions. Two distinct issues seem pressing. First, why should Wolff think that only cognition counts? Second, how can Kant rule out the possibility of true metaphysical judgements that are based on conceptual analysis without begging the question against the Wolffian?
Fifth, as is well-known, in the Transcendental Dialectic Kant offers an extensive analysis of pure reason so that he can determine whether it is able to generate cognition of the objects of special metaphysics. This is at least how Kant sets up the Transcendental Dialectic: ‘Does reason in itself, i.e. pure reason, contain a priori synthetic principles and rules, and in what might such principles consist?’ (A306/B363). He then answers this question by analysing and rejecting arguments that would establish cognition of the soul, the world as a totality and God. What this way of setting up the Transcendental Dialectic suggests is that, if the Transcendental Analytic had already established that we could not cognize things in themselves in general (which would thus have to include the soul, the world as a totality and God, since they are specific kinds of things in themselves), then Kant would be fully justified in immediately answering this question negatively and moving on. The first Critique would have been half as long, but, to my mind, also half as interesting.
However, if the Transcendental Dialectic contributes to the overall argument of the first Critique by showing that pure reason cannot generate cognition of the objects of special metaphysics, then one wonders what is supposed to establish that we cannot have cognition of the objects of general metaphysics. That is, if the Transcendental Dialectic shows both that Kant sees the need for an argument ruling out the possibility that we can cognize the soul, the world as a totality and God, and that he provides such an argument by ruling out all the main types of arguments that attempt to establish such specific claims to cognition, then it would seem, for analogous reasons, that Kant should also see the need to provide an argument that rules out the possibility that we can cognize things in themselves in general. But where are we supposed to find such an argument? Note that it cannot be in the Transcendental Analytic (e.g. the Transcendental Deduction), because if it were offered there, then the Transcendental Dialectic would not be needed. Instead, what the Transcendental Analytic shows is that the categories, which one might use to attempt to form cognition of the objects of general metaphysics, are valid for objects of experience, but that leaves unaddressed the objects of general metaphysics, namely the class of objects in general. It is tempting to think that there is a complex division of philosophical labour between the Transcendental Analytic and the Transcendental Dialectic that leads to the encompassing conclusion that Kant wants to be able to draw regarding both general and special metaphysics, but specifying the division of labour in a way that is consistent with Kant’s actual arguments is not a simple or straightforward matter. Given the centrality of this point to understanding the overall argument of the first Critique, it would be good to know where De Boer stands on this issue.
However the above questions are answered, the fact is that De Boer’s book represents a significant scholarly achievement. It advances an overall thesis with which I have great sympathy, she offers a range of interpretations that are both historically sensitive and philosophically sophisticated, and she provides convincing arguments in support of her main claims. It is an impressive book.