The book gives a new interpretation of the impact of Hume’s philosophy on Kant’s Critique of Pure Reason. The author claims that Hume did not interrupt the dogmatic slumber of Kant by attacking the causal principle, but rather by attacking metaphysics in general and the principle of sufficient reason in particular, which was used by dogmatic philosophers to conceive of God and to prove His existence. He concludes that ‘it was the problem of whether we can think God, not the question of whether the concept of cause is valid for experience, that first interrupted Kant’s dogmatic slumber’ (p. xix). Thus, both Hume and Kant belong to a philosophical tradition which attacks metaphysics in order to promote Enlightenment, or ‘the liberation of the mind, both public and individual, from theological authority’, and Kant’s dogmatic slumber turns out to be a ‘subjection to theological illusion, a lack of Enlightenment’ (p. 44). These far-reaching claims require clarification, and arguments for them are provided by the author in due course. First, he gives a detailed interpretation of Kant’s account of Hume and his objection in the preface to the Prolegomena (chapters 1 and 2). Second, he argues that the Enquiry contains an attack on the rationalist principle of sufficient reason and was understood by Kant in this way (chapters 3 and 4). Finally, in the last chapter he explores the so-called ‘hidden spine of the Critique’ – its ‘unity’ stemming from Kant’s joining ‘Hume’s attack on rational theology via an attack on the principle of sufficient reason’ (p. 145).

In order to assess Anderson’s interpretation of the preface to the Prolegomena, one has to consider the main topic of the preface. It is not ‘the struggle for Enlightenment’ (p. 45), but it is rather concerned with a particular branch of philosophy called metaphysics, as the very title of the Prolegomena suggests. According to Kant, the most important contribution to this issue in modern times is Hume’s attack on metaphysics by his analysis of the connection of cause and effect. However, this contribution has to be evaluated in two different ways. On the one side Hume was concerned with just ‘a single, although important metaphysical concept’, and this restriction explains for Kant the failure of his attack on metaphysics (4: 260). On the other side he emphasizes the correctness of Hume’s analysis by accepting the view that the connection of cause and effect cannot be ‘a priori and from concepts’ (4: 257). What Kant has in mind is that it cannot be a priori just because it is not based upon an analysis of concepts. Unfortunately, Anderson does not give an account of the first part of Kant’s evaluation, which would have required a closer examination of his views about metaphysics before Hume’s interruption. The claim that metaphysics is conceived as ‘a science of things beyond experience … presumably, God and immaterial spirits’ (pp. 76, 86) is vague and does not fit the subtle views about metaphysics of the pre-Critical Kant (cf. 1: 410, 2: 308–9, 368, 395).

As Anderson correctly points out, Kant’s account of Hume’s attack on metaphysics is divided into three steps: his proof that the connection of cause and effect is not a priori and analytic, and two conclusions which he draws from that (pp. 72, 90). According to Kant, the proof is correct, but the conclusions are wrong; thus, Hume makes a mistake by inferring the latter from the former. Anderson is mainly concerned with the proof, and he claims that for Kant it is a ‘challenge to the rational origin of the concept of cause’ (pp. 78, 80). However, this notion of origin is not clarified. According to Kant, Hume aims at an analysis of the origin of that concept and not of an explanation of its ‘indispensability in use’ (4: 259). Kant’s conception of origin is part of a general theory of representations which distinguishes between their semantic content, the empirical circumstances of their acquisition, the cognitive faculties or sources that are the origins of them, and their objective reality (cf. Reflection 4917). Kant believes that he was in agreement with Hume with regard to the semantic content of the concept of cause, but that he disagrees with him with regard to its origin. While Hume considers reason, conceived as a cognitive faculty of a priori cognitions based upon the analysis of concepts, as the origin of the concept of cause, Kant, as it is shown by his ‘complete solution of the Humean problem’, brings into play his notion of understanding as an a priori cognitive faculty essentially tied to experience (4: 312–13, cf. B87). Anderson’s notion of rational origin does not take account of Kant’s distinction between reason and understanding, and it is strange anyway that he never discusses what Kant considered as his way of solving Hume’s problem (cf. 4: 310–13).

Anderson’s main argument for his thesis that Hume’s objection was a challenge to the very idea of metaphysics, and was understood by Kant in this way, consists in his attempt to prove that the objection is concerned with the principle of sufficient reason. In this way we get the still missing ‘account of the objection of David Hume that lets us clearly see how it interrupted Kant’s dogmatic slumber’ (p. 41). This approach gives rise to the following questions. Did Hume attack the principle of sufficient reason? Did Hume’s objection, as understood by Kant, concern that principle? If the objection interrupted Kant’s dogmatic slumber, what significance did the principle have for the metaphysical theories of the still sleeping Kant? I will address these questions in turn.

Anderson gives an answer to the first question by referring to Enquiry 4.11, where Hume offers a kind of résumé of the main points of his analysis of our ‘reasonings concerning matter of fact’ as it is ‘founded on the relation of Cause and Effect’ (4.4). He claims that we cannot ‘determine any single event, or infer any cause or effect, without the assistance of observation and experience’ (4.11). According to Anderson, Hume denies ‘that we can ever conceive a cause as a sufficient reason of any effect – that is, as a ground from which we could derive that effect a priori’ (p. 102). However, at 4.13 Hume says ‘something more general and more abstract’ which amounts to the denial ‘that we can have knowledge a priori of the principle of sufficient reason in its general form’. Anderson refers to Hume’s claims that, by reasoning a priori, we are neither able to apply the very notion of a connection of cause and effect as a relation between different objects nor to understand ‘the inseparable and inviolable connexion between them’ (4.13). These are two claims, and it has to be pointed out that, if one takes Kant’s summary of Hume’s objection as ‘evoking’ Enquiry 4.13, as Anderson does, this summary is concerned with the second claim. According to Kant, Hume’s question was why is it necessary that, if something is given, something else is given as well, and his ‘undisputable’ proof was that ‘it is wholly impossible for reason to think such a connection a priori and from concepts, because this connection contains necessity’ (4: 257). That the notion of cause entails the idea of a necessary connection is pointed out by Hume in the second of the claims which cannot be established by a priori reasoning from concepts. Anderson’s account invokes the principle of sufficient reason, mentioned neither by Hume nor by Kant, but does not emphasize the central role of the idea of a necessary connection between cause and effect in the considerations of both of them.

He discovers a ‘more explicit expression’ of Hume’s rejection of that principle in a footnote added to Enquiry 12.29 (p. 104). This footnote is the keystone for Anderson’s interpretation of Hume and his relation to Kant. It is supposed to show that Hume is rejecting the cosmological proof of a Supreme Being and ‘modern philosophical theology’ in general, and that the footnote, thereby, ‘gives Kant grounds for finding in Hume an anticipation of his own transcendental idealism’ (pp. 110, 113, 158). In order to assess the significance of the note for Kant’s philosophy, we have to take a look at the broader context of the footnote, which occurs at the end of the Enquiry. In 12.27 Hume points out that we have demonstrative knowledge of quantity and number, and he goes on to emphasize that we do not have this kind of knowledge with regard to matters of fact and existence. In particular, the non-existence of any object is ‘as clear and distinct an idea as its existence’ (12.28). He concludes that ‘it is only experience, which teaches us the nature and bounds of cause and effect, and enables us to infer the existence of one object from that of another’ (12.29). The footnote added to this remark says: ‘That impious maxim of the ancient philosophy, Ex nihilo, nihil fit, by which the creation of matter was excluded, ceases to be a maxim, according to this philosophy.’ Given the context of the note, it states an application of a general claim about the empirical nature of our knowledge of causal connections to the special case that ‘the will of supreme Being may create matter’. Because of that empirical nature, the maxim cannot be valid and is, therefore, incompatible with the ‘philosophy’ outlined in the main text. The maxim comes from Lucretius and is identified by Descartes with some version of the principle of causality (cf. AT VII, 135).

According to Anderson, in the footnote Hume is attacking Descartes’ view (p. 107), but if so then why should we consider Hume’s attack on the maxim, as it is understood by Descartes, as an attack on the principle of sufficient reason? Anderson claims that the footnote is Hume’s ‘most direct attack on the principle of sufficient reason’ (p. 145), and that Descartes considers the maxim as being ‘equivalent’ to that principle, because Leibniz and ‘Kant’s German contemporaries Wolff and Baumgarten’ understood the maxim in this way (p. 106). This is a rather strange way of interpreting Descartes, and it has to be pointed out that neither Wolff nor Baumgarten connected the principle to the maxim. According to them, the principle regards whatever it is that allows us to understand or to know something or other (cf. Wolff, Ontologia, §70; Baumgarten, Metaphysica, §20). As far as Kant is concerned, Anderson believes that he was ‘perfectly aware’ of their intimate connection, because he refutes their demonstrations of the principle of sufficient reason in the Nova Dilucidatio (pp. 119–20). But neither their demonstrations nor his refutation has anything to do with the ‘maxim of the ancient philosophy’. That it ‘ceases to be a maxim’ for Hume could not be understood, by Kant or by anybody else, as a refutation of the principle of sufficient reason.

At the time of the pre-Critical Kant several versions of the principle were discussed, and he made an attempt to demonstrate just one version of the principle of sufficient reason in the Nova Dilucidatio – the cause of the existence of contingent things. It was published in 1755, but in the early 1760s he rejected the very possibility of such a proof by distinguishing between logical and real grounds (2: 202–4). This distinction may be due to his reading of Hume’s Enquiry, although Kant mentions Crusius in this context and not Hume. In the Prolegomena he refers to the principle as ‘obviously synthetic’ (4: 270), and in the Critique he claims that there can be no proof of it, except by the way of the Second Analogy (B264–5). He never mentions Hume at these places, and he never points out that he was the philosopher who challenged that principle for the first time. Why should we attribute this view to Kant? According to Anderson, the principle ‘applies to things or objects in general, merely as they are conceived, not to events or objects of experience’ (p. 103). The dogmatic use of the principle is ‘its application, for purposes of knowledge, to things in themselves as opposed to appearances’, and Hume’s critique of that use led ‘Kant to transcendental idealism, or the restriction of human knowledge to appearances as against things in themselves’ (p. 159). Thus, Kant’s transcendental idealism is just a ‘conclusion from Hume’s attack on metaphysics’ (p. 162).

Yet, this argument is not convincing for several reasons. First, the distinction between appearances and things in themselves is drawn by Kant in the Inaugural Dissertation, published in 1770. There is no hint at all in this book that his views have something to do with Hume’s attack on metaphysics, and Kant himself makes use of the distinction in order to define his metaphysical project (2: 395). Anderson understands the distinction according to the so-called two-worlds interpretation, as opposed to the two-aspect interpretation, but he does not give any argument for his view, not much favoured today, because it does not fit Kant’s notion of the world in the Dissertation nor his later remarks (cf., for instance, 22: 26). Second, transcendental idealism entails without doubt a restriction of human knowledge to appearances, but the way Kant makes this restriction can only be grasped by taking into account his distinction, not mentioned by Anderson, between ‘a noumenon in the negative sense’ and ‘a noumenon in the positive sense’ (B307–8). Our knowledge of appearances is constituted by concepts of the pure understanding which are supposed to refer to noumena in the negative sense (cf. 5: 136). Thus, transcendental idealism does not relinquish the epistemological significance of the notion of noumenon, as Anderson seems to suggest, but restricts its application to ‘the unity of intuitions in space and time’ (B308). This kind of restriction has no correspondence in Hume’s theory and is completely different from Hume’s restriction of human knowledge to empirical knowledge. Thus, even if his attack on metaphysics is based upon an attack on the principle of sufficient reason, transcendental idealism, as understood by Anderson, cannot be a conclusion of that attack.

By way of concluding, I want to point out that it is of course the case that Hume’s interruption of Kant’s dogmatic slumber has something to do with the former’s attack on metaphysics, but for Kant this was due to the metaphysical significance of the concept of cause, while for Anderson it is the principle of sufficient reason which is at bottom of Hume’s attack on metaphysics. That his attack on this principle should be what first interrupted Kant’s dogmatic slumber presupposes that Kant considered it a keystone of metaphysics. In 1755 he made the attempt of proving a particular version of that principle valid for the cause of contingently existing things. In the early 1760s he rejected this view, and there is no evidence that his pre-Critical conception of metaphysics was essentially connected with and based upon the principle of sufficient reason. Anderson emphasizes that the principle in this ‘unrestricted’ version applied to things beyond experience (cf. pp. xiv–xv), but this holds as well for the metaphysical concepts of pure reason in the Inaugural Dissertation (2: 395), and in any case the principle is not mentioned there at all. Anderson claims that ‘Hume roused Kant by challenging the unrestricted principle of sufficient reason, since it is this principle that is used in theology, to prove the existence of God’ (p. xv), but the theological use of the principle was not at the very centre of the pre-Critical notion of metaphysics. Kant’s own proof of the existence of God is a so-called ‘ontological proof’ in the Der einzig mögliche Beweisgrund, published in 1763, and does not invoke that principle. Anderson claims that this text contains ‘an attack on theology’ (p. 86), and that it is ‘clearly influenced by Hume’s critique of theology’ (pp. 161–2), but he gives no arguments for these claims. Thus, there is no evidence that Kant was roused by Hume through his alleged challenge to the principle of sufficient reason.

The story told by Kant about the interruption of his dogmatic slumber in the preface of the Prolegomena leads easily to the view that Hume was in some way his ‘predecessor’ (4: 260). But only two pages later we are told by Kant that he considered his own approach as an alternative to Hume’s way of proceeding (4: 262). Later on he compares the account of the connection between cause and experience, given by Hume, with his own account and points out that the latter is ‘a completely reversed type of connection that never occurred to Hume’ (4: 313). Thus, even in the Prolegomena Kant presents Hume not only as a ‘predecessor’, but as a real and fundamental alternative – a view which is very much emphasized in the second edition of the Critique (B127–8). Regrettably, the simple picture of Kant ‘as a devoted heir to Hume’ (p. xx) or of Hume as ‘the forerunner of the Critique’ (p. 158) neither accounts for the differing assessments of the Critical Kant, nor does it allow for an adequate understanding of the variety of problems the pre-Critical Kant attempted to solve.