Moses Mendelssohn described Kant's approach in the Critique of Pure Reason as ‘all destroying’, and it has traditionally been thought that what Kant destroyed was the possibility of substantive metaphysical inquiry for human beings. While the concepts of space, time, substance, and causation all receive a reconceptualization within his own transcendental idealist system, that system is also supposed to offer a strict delimitation on the possibility of insight into the nature of God, freedom, and the soul. These latter concepts come together as the pinnacle of traditional ontological speculation, and were shown by Kant to relate to the noumenal domain, a domain of things in themselves that he claimed was inherently unknowable. Thus metaphysics appears to be either denied, or reformulated in terms of something else – epistemology. As Kant put it, ‘the proud name of an ontology . . . must give way to the modest one of a mere analytic of the pure understanding’ (A247/B303). Henry Allison's Kant's Transcendental Idealism, first published in 1983, presented an account of Kant's project not merely as claiming that a metaphysical world lay beyond our reach, but rather as attacking the very idea of such a world, and with it the very idea of the meaningfulness of inquiry about it. In Allison's memorable phrase, Kant presented not an alternative ontology, but an alternative to ontology. On this reading, Kant's goal was to disabuse us of the idea of the meaningfulness of metaphysical discourse per se, and to replace it with an account of the meaningfulness of philosophical terms defined relative to human epistemic practices.

It is remarkable that this interpretative position evolved around the middle of the twentieth century, when the replacement of metaphysical speculation with the philosophy of language was still considered a live research project. In more recent years however, the anglophone world of Kant scholarship has started to see things very differently. Just as David Lewis and Saul Kripke in the late twentieth century made metaphysical speculation respectable again, so too have readings of Kant's philosophy as metaphysical speculation become respectable again. On this approach, Kant was a profoundly metaphysical thinker: concerned for instance with the theory of powers that might be behind his account of causation and freedom, and with definite metaphysical commitments of his own, sometimes perhaps esoterically maintained behind a veil of hostility to metaphysics. With this interpretative approach in hand, it becomes perfectly appropriate to speculate as to just what those metaphysical commitments were. Kant scholarship is currently in the curious position whereby an ordinary conference will include a healthy proportion of papers about Kant's views on the constitution, number, and powers of things in themselves – just the type of speculation Kant was supposed to have been encouraging us to forsake forever.

Christopher Insole's erudite work contributes to this revitalization of the metaphysical Kant. Insole hopes to fill a ‘gap in Kant scholarship’; specifically, the lack of a ‘sustained treatment from Kant's pre-Critical position on the status of space, time, and causation, to his Critical position, which keeps as its central focus Kant's changing conception of the relationship between divine action and human freedom’ (4). The core interpretative claim, then, is that a theological problem motivated some of Kant's Critical philosophy.

The book's central contention is that Kant ‘has intrinsically theological difficulties of his own that generate aspects of his Critical philosophy’ (5). This is a difficult claim to evaluate, since it is surely also the case that Kant was motivated to reconcile Leibnizian and Newtonian approaches to causation, to account for the empirical applicability of mathematics, to explain the possibility of self-consciousness and ordinary perception, by worries about the infinite divisibility of matter, etc. It is a challenge to isolate the influence a particular problem might have had upon the development of Kant's transcendental idealism. Kant saw that theory – remarkably – as one that could accommodate the controversies relating to all these topics. On the one hand, it seems implausible that Kant was unaware that transcendental idealism could provide a resolution of some theological issues; on the other hand, it seems equally implausible to attribute to a particular theological problem any prominent position among that list of competing influences.

Perhaps this worry is pressing because of the nature of transcendental idealism as a philosophical strategy. It seems clear that part of the strategy with regard to freedom is to show that it is beyond the limits of what is cognizable. This is a strategy that tries to secure the possibility of freedom at the cost of being able to prove its actuality in any theoretical way. It is also unable to speak in favour of any particular metaphysical conception of freedom over any other. It would be odd, then, if it emerged that Kant's theory was motivated by his commitment to one particular conception. The more traditional interpretation would have it that Kant considered all sorts of positions on a range of topics in his pre-Critical career, concluded that none could be securely established, and developed his Critical position to accommodate this fact about certain areas of metaphysical discourse.

Insole's scrupulous work must be considered a success, however, in providing a coherent case for thinking that theological questions were at least as prominent among Kant's concerns in his Pre-Critical period as any of the more familiar philosophical challenges. Insole's work negotiates a wide range of Kant's primary sources from both the Pre-Critical and Critical period. He moves ably through a diverse range of classic secondary literature but also draws heavily on recent work on Kant's metaphysics, notably on (as-of-then) unpublished work by Andrew Chignell, Desmond Hogan, Patrick Kain, and Nick Stang. The work is impressive not least for its ability to tie some of this recent work together into a single coherent narrative.

Following an introductory chapter, chapter 2 begins the inquiry by arguing that for Kant, God's perfect freedom is compatible with his being unable to do other than the good. Chapter 3 discusses the extent to which Kant's pre-Critical conception of divine freedom is compatible with Christian orthodoxy, and takes issue with Rae Langton's reading. Chapter 4 begins to set out the animating problem of the book, which concerns how human freedom is compatible with the possibility of our being created by God. Chapter 5 sets out the challenge of reconciling this reading within the strictures of transcendental idealism. Chapter 6 discusses the coherence (or lack of it) of Kant's apparent notion of atemporal causation, in the context of the account of the possibility of evil actions. Chapter 7 argues that Kant did in fact believe in God. Chapter 8 considers the sense in which God is the creator of human beings qua noumenal substances even if he is not the creator of space and time. Chapter 9 goes into more detail on the nature of the causal relation between God and human beings, arguing that Kant rejects occasionalism and concurrence in favour of a ‘mere conservation’ model. Chapter 10 concludes with some reflections on whether Kant's own opposition to concurrentism represents an insurmountable hurdle for his relevance to Christian theology.

Insole's approach highlights one way to consider the plausibility of certain metaphysical interpretations of Kant's Critical philosophy. One can view the position as one that denies that we are able to apply concepts such as substance, cause, etc. to the noumenal domain, on the grounds that we cannot cognize that domain. Such a view is compatible with allowing less ambitious types of epistemic claim about that domain, such that one might believe or speculate (or hope perhaps) that the domain is nevertheless constituted in accordance with those concepts. On the other hand, one might take Kant's claims as involving a stronger position, namely that the noumenal domain is inherently inapt to be constituted in accordance with such concepts, and that these concepts only have a ‘sense and significance’ (Sinn und Bedeutung) with regard to the phenomenal realm that we can know. On the stronger reading, any kind of claim about ‘noumenal substances’ or ‘noumenal causation’ just lacks content.

The stronger reading has a particular motivation. Kant thought that he could secure knowledge of necessary truths regarding causation (for example). Such truths required ‘strict’ and not ‘comparative’ universality, i.e. they must apply to the entire set of possible things that fall under the concept. If the concept of causation applies to both phenomenal and noumenal things, and if the noumenal is unknowable, then Kant can at best claim to have secured truths about a subset of the things that fall under that concept. Kant cannot then have thought that the concept of causation applied in any way to the noumenal domain, since he thought that we could secure knowledge of the entire set of things that fall under that concept. This problem doesn't seem to be ameliorated by appealing to phenomenal and noumenal variants of the same concept, if they are indeed supposed to be variants of the same concept. If they are not the same concept, then there is no reason to use the word ‘causation’ with regard to the noumenal. An advocate of the stronger reading might then worry whether Kant thought such metaphysical and theological considerations are really meaningful uses of these words at all.

Nevertheless, the stronger reading can seem just too strong, and Insole's account is best understood as assuming the plausibility of the weaker approach, whereby Kant held that some use of these concepts with regard to the noumenal realm was at least minimally meaningful. Insole addresses the motivations for this approach in chapter 5, claiming that talk of God's creation of noumenal substances (which then in turn act and cause things to come about) is permissible within the strictures of transcendental idealism. Insole maintains – as many other Kant scholars now do – that Kant allows for the categories of causation and substance to have an application to things in themselves at least in thought, even if we cannot have a cognition that those categories in fact apply.

Insole's central thesis is not an uncontroversial one. For one thing, it is not always easy to grasp just what exactly is the problematic nature of the ‘problem’ that Insole claims motivated Kant's intellectual development. At a first pass it is this: human beings cannot be both created by God and entirely self-determining in the way that a robust conception of freedom requires. Put this way, the problem doesn't seem that problematic – had God created us as self-determining beings, then our freedom would not be incompatible with our created status. Rather, the problem is supposed to be that seeing ourselves as created involves seeing all our subsequent actions as determined by something ‘alien’ to ourselves. In so far as freedom is understood as requiring independence from alien determining causes, then our created status poses a challenge to our freedom. One might still wonder though whether this is a problem, since it is surely only so far as something more than our mere existence was determined by an external cause that our freedom is threatened. So long as none of our subsequent individual choices are themselves determined by the same cause that brought about our existence, those individual choices might still satisfy the requirement of being genuinely free actions, despite being the actions of created beings.

It is also not clear from Kant's published writings that this ‘problem’ concerned him throughout his life. Kant frequently went on the record stating that the very idea of freedom challenges human comprehensibility. There is a lot of evidence that he thought that the challenge was that of explaining the very idea of a non-physical yet causal power of freedom that somehow spontaneously intervened into the physical domain. He held that transcendental idealism could accommodate this idea, but only by providing a principled dividing line between the comprehensible and the incomprehensible, and by placing freedom beyond that line. Kant rarely presents this incomprehensibility as being especially pressing with regard to the question of freedom's reconciliation with our divine createdness, however. One might think that an application of Ockham's razor might be warranted: if it is sufficient to explain the trajectory of Kant's intellectual development by appeal to his concern with the very idea of freedom, then attributing to him an additional concern with freedom's relation to the fact of our createdness does not seem to be necessary.

That such an attribution might not be necessary does not entail that it is not nevertheless warranted: as mentioned, Kant surely had many different concerns in mind, irrespective of what emerged in publication as his primary concerns. Insole's central thesis requires some evidence that Kant thought that the fact of our createdness did threaten the possibility of freedom, and that transcendental idealism was designed for the purposes of (among other things) resolving that threat. Insole finds just these concerns aired in some metaphysics lecture notes from the 1770s, where Kant states that if the human being is a ‘derived being’, i.e. something whose existence is brought about by something else, then ‘it appears to be quite probable that it is also determined by this cause in all its thoughts and actions’ (Kant, Lectures on Metaphysics, 28: 268, quoted p. 75). This seems clear evidence in favour of the hypothesis. Insole provides other references from Kant's notes that, he suggests, show Kant's ‘struggles to reconcile transcendental freedom with our createdness’ (75, n. 20); however, I could not find in these references any specific discussion of human beings’ createdness as the source of the tension. Rather, those passages seemed to me to show Kant's struggle with the very idea of freedom as a spontaneous cause that somehow begins outside the chain of physically caused events and yet can bring about effects in the physical world. More explicit support is found in the Critique of Practical Reason, and Insole discusses this in chapters 5 and 8. Overall, though, his argument would be strengthened by more textual evidence that it was the specific combination of our divine createdness with the possibility of freedom, rather than merely the problem of freedom itself, that so troubled Kant throughout his life.

I also have concerns regarding Insole's use of the published supporting passage, from the Critique of Practical Reason at 5: 100–3. Here Kant seems to claims that if space and time were properties of things in themselves (as the ‘transcendental realist’ maintains) then a problem regarding God's influence on human action would arise. Kant seems to have the following argument in mind: if I were entirely identical with my spatiotemporal self (i.e. if I did not have a distinct non-spatiotemporal noumenal self) then the causality that brought about my existence would have to be of the same univocal kind that is operational in all my spatiotemporal actions. If this were the case, however, the causality in accordance with which I was created would be just one link in the same causal chain in accordance with which I perform all my actions. Thus I could see all my actions as determined by that first action of my creation. If, however, I see the ‘causality’ in accordance with which I was created as different in kind from that through which my spatiotemporal physical interactions are conducted, then the causal chain is severed, and the fact of my creation does not entail that my spatiotemporal actions are determined by alien factors.

This is a difficult section to interpret, yet it is hard to square what it at least appears to say with Insole's central thesis. On any reading of Kant's transcendental idealism, the causality of our freedom is different in kind from the causality of phenomenal objects in the physical world (including our phenomenal selves). Kant's argument here seems to hinge on the fact that, while it would be a problem if the causality of God's creation were the same as that of spatiotemporal causality, it is not a problem just because God's creative power refers to a domain that refers only to the noumenal side of our being. However, it is this noumenal side of our being that is also the source of our power for free actions. Insole claims that Kant resolves what might appear to be a problem here by a practical turn towards the notion of a Kingdom of Ends (178ff.). The worry about this move is not that Kant does not mention the Kingdom of Ends or anything like it as the resolution of the problem of creation and causation at this point in the second Critique (Insole acknowledges the reconstructive character of the claim). It is rather that Kant seems to take the fact of God's creation of noumenal human beings as a sufficient resolution of the need for our free actions to be undetermined by alien causes. Insole's central claim is that Kant was worried by the possibility that our power of freedom might itself be undermined by the fact that we are the product of an act of divine creation. In the second Critique, however, Kant seems to make this exact claim as part of a solution to the problem rather than as the source of the problem.

Kimberly Brewer and Eric Watkins (‘A difficulty still awaits: Kant, Spinoza, and the threat of theological determinism’, Kant-Studien, 103 (2012), 163–187) argue that Kant's move here is philosophically defensible just for the reason mentioned earlier: the fact that the existence of a power freely to perform actions was causally determined by God does not entail that the actions subsequently performed in accordance with that power are themselves causally determined (ibid., 178). My transcendental power to do x (or otherwise) is not undermined by the fact that the power itself was brought into being in an act of creation. I imagine that Brewer and Watkins's work emerged too late for Insole to take account of it in this book, which is a pity, since it engages with the heart of the work. Despite its opposing his specific thesis, such work at least vindicates Insole's broader claim that Kant thought the theological context relevant to the evaluation of the claims of transcendental idealism.

The difficulty of adjudicating this point reflects one of this work's merits, however, which is the degree of sophistication with which it engages with its topics. Although its central claim is contentious, the book touches on a range of other key topics in Kant's philosophy, and always perceptively. This work will then reward any serious study. It is a very well-written and diligently researched piece of work, one that brings the more recent metaphysical interpretations of Kant's Critical philosophy into direct contact with traditional theological concerns.