Frederick Beiser has been rewriting the history of nineteenth-century German philosophy. For many years, in many courses and books, the history of nineteenth-century German philosophy, or even nineteenth-century philosophy tout court, has been the history of German Idealism with a few of its offspring – thus, the history of Fichte, Schelling, Schopenhauer and Hegel, together with their wayward children Kierkegaard, Marx and Nietzsche. On Beiser’s approach, this represents only one strand within a much more complex fabric, in which a wide range of philosophers within the burgeoning German university system were just as important in the post-Hegelian years – the greater part of the century, after all – as the ‘great outsiders’, as Beiser calls them (p. 9), Kierkegaard, Marx and Nietzsche. Beiser has provided a concise overview of his revisionary approach in his other book from 2014, After Hegel: German Philosophy 1840–1900 (Princeton), in which he discusses the ‘identity crisis in philosophy’; the ‘materialism controversy’; what he calls the ‘ignorabimus controversy’, which concerns the standing of philosophy in the face of the escalating prestige of natural science throughout the century, especially from mid-century on; the attempts of philosophers to deal with the equally impressive development of historiography during the nineteenth century (‘trials and tribulations of Clio’), and the controversy over the pessimism of Arthur Schopenhauer and Eduard von Hartmann. In more detail, he has chronicled the development of modern historiography and philosophy of history, actually beginning with Herder, Wilhelm von Humboldt and several other later eighteenth-century figures, in a volume as stout as that here under review, namely The German Historicist Tradition (Oxford, 2011), and studied the two immensely influential mid-century philosophers Friedrich August Trendelenburg and Rudolf Herman Lotze, the latter of whom also had a major impact on British and American philosophy in the last part of the century, in Late German Idealism: Trendelenburg and Lotze (Oxford, 2013). In The Genesis of Neo-Kantianism, Beiser tells the story of Neo-Kantianism from its earliest stirrings while German idealism still dominated the stage (think of the first Neo-Kantians as early mammals waiting in the underbrush for the extinction of the dinosaurs) through the early works of Hermann Cohen, Wilhelm Windelband and Alois Riehl, the leaders of what would become the three main schools of Neo-Kantianism at the turn of the twentieth century and until the debacle of the First World War, namely the Marburg, Southwestern and Berlin schools. And those who manage to absorb all of this work will be glad to know that Beiser is currently completing his full-fledged study of the pessimism controversy to redeem the pledge given in the final chapter of After Hegel.

The Genesis of Neo-Kantianism, as its subtitle indicates, chronicles the development of Neo-Kantianism from the turn of the nineteenth century to 1880, actually well into the 1880s in the case of the last two figures discussed, namely Windelband and Riehl. It might seem surprising that Neo-Kantianism is said to begin before Kant himself was even retired, but Beiser’s argument is that, while Kant’s own philosophy was being overshadowed by the innovations of the young Fichte and even younger Schelling beginning in 1794–5, before Kant had even published his final books, other young philosophers, not from Kant’s own circle in Königsberg but from elsewhere in Germany, were trying to refine his ‘critical’ idealism without succumbing to ‘absolute’ idealism, and thus already deserve the name of Neo-Kantians. This group of philosophers includes Jakob Friedrich Fries, Johann Friedrich Herbart and Friedrich Eduard Beneke, the ‘Lost Tradition’ that comprises part I of Beiser’s book. Part II discusses the ‘Coming of Age’ of Neo-Kantianism in the 1850s and 1860s, and covers Kuno Fischer (whose debate with Trendelenburg over the possible isomorphism of appearances and things in themselves is studied in more detail in Late German Idealism), Eduard Zeller (best remembered as the nineteenth century’s pre-eminent historian of ancient philosophy), Otto Liebmann (often thought to have initiated Neo-Kantianism with the refrain ‘Back to Kant!’ in 1865, but here taking his proper place in a movement already well under way), Jürgen Bona Meyer and Friedrich Albert Lange (who influenced Nietzsche among many others), and concludes with chapters on the Neo-Kantians’ resistance to Schopenhauerian pessimism and their attempts to assimilate Darwinism. Part III then covers the early work of Hermann Cohen from which not only the later work of Ernst Cassirer but also much post-1966 Anglophone work on Kant’s theoretical philosophy ultimately descends, Wilhelm Windelband’s introduction of the concept of ‘normativity’ into both theoretical and practical philosophy, and Alois Riehl’s attempt to present Kant as at least as much of a realist as he was an idealist. One might quibble with the amount of space lavished on a couple of the figures covered in part II, perhaps Zeller and Bona Meyer; but on the whole Beiser makes a powerful case that the philosophers he discusses, most of whom will be just names even to well-informed contemporary Kantians, identified problems of ongoing interest and importance in the interpretation and assessment of Kant’s philosophy and had interesting and important things to say about these issues.

Throughout the work, Beiser argues that the Neo-Kantians were concerned with two of the great issues he identified as central to nineteenth-century German philosophy wie sie eigentlich gewesen war in After Hegel, namely the ‘identity crisis of philosophy’ that ‘began in the 1840s’ as a result of the simultaneous ‘decline of speculative idealism and the rise of the empirical sciences’, which thus included the methodological question of whether there is any other alternative to the discredited speculation of absolute idealism for a method in philosophy than the empirical methods of the natural sciences, especially psychology, and the question of whether philosophy even has any ongoing role in the face of the development of those sciences, or whether the sciences need philosophy as epistemology or as a ‘logic of the sciences’, as well as the substantive question of whether philosophy could resist the materialism of the natural sciences and leave room for ‘a moral or normative sphere above and beyond nature’ without lapsing into the kind of metaphysics that Kant had discredited but that had been so quickly revived in German idealism (pp. 6–7). With regard to the interpretation of Kant, these great debates took the form of debates over whether Kant’s method was psychological or logical or epistemological, over the necessity of things in themselves within Kant’s transcendental idealism, and whether Kant’s combination of transcendental idealism with empirical realism showed how to defend materialism, avoid it or both.

In part I, on the ‘Lost Tradition’, Beiser shows how Fries, Herbert and Beneke shared their advocacy of ‘Kant’s transcendental idealism in its original form, especially its limitation of knowledge to appearances’ but also retaining ‘its postulate of the thing-in-itself’; how ‘they upheld Kant’s regulative constraints upon teleology, which had been violated by Schelling and Hegel’; that they ‘banished appeals to intellectual intuition, which Kant had proscribed but which had been re-invoked by Fichte, Schelling, and the young Hegel’; that ‘they dwelled in “the bathos of experience,” insisting that all knowledge be limited to experience’; and that they ‘re-affirmed Kant’s dualisms’, which had come under such heavy attack by the subsequent idealists, ‘more specifically, those between understanding and sensibility, essence and existence, [and] practical and theoretical reason’ (p. 15). Beiser presents these thinkers as founding an ‘empiricist-psychological tradition’ in the interpretation of Kant’s methodology, in reaction to the ‘neo-rationalism of Schelling’s and Hegel’s metaphysics’ (p. 16), a psychological approach which would inform not only that of the middle generation of Neo-Kantian philosophy professors that he describes in part II but also the work of the great scientist Hermann von Helmholtz (pp. 196–206). According to Beiser, it was only with Cohen’s Kants Theorie der Erfahrung, first published in 1871, that there was a decisive turn from the psychological to the epistemological approach to Kant’s theory of synthesis and judgement that dominated not only Marburg Neo-Kantianism but also, one might add, the Anglophone approach to Kant’s theoretical philosophy from Strawson at least until the ‘cognitive science’ approach of Patricia Kitcher and Andrew Brook.

At the same time, Beiser makes it clear that these early Neo-Kantians, especially Herbart, did not accept the details of Kant’s faculty psychology, which they found Wolffian and old-fashioned. ‘The very idea of a faculty is for Herbart a hypostasis, the reification of an abstract concept’, and in particular ‘The tripartite theory that divides the soul into representing, desiring, and feeling’ was a particular ‘subject of Herbart’s wrath’ (p. 137). Throughout the book, while he makes it clear what he thinks of other interpretations of the Neo-Kantians, Beiser does not explicitly judge the views of the Neo-Kantians themselves; but in this case he calls Herbart’s critique of Kant’s psychology ‘trenchant’ (p. 141). Yet I think there might be something to be said for Kant’s tripartite division: after all, we can think either about how the world of objects beyond our own thoughts of them is (representation) or how we might like or think it ought to be (desire), and we can think about the objective content of either of those sorts of thoughts or about how those thoughts may affect us subjectively, i.e. how they may feel (feeling). (This may seem a stretch, but I am thinking of the argument of book II of Collingwood’s Principles of Art here.) Thus it is not clear to me that Kant’s faculty psychology is as outmoded as Herbart and following him Beiser seem to think. Later in the book, Beiser reports that Bona Meyer defended Kant’s tripartite faculty psychology along lines similar to those I have just suggested (pp. 340–1). Beiser does not take sides on this issue, but he has clearly suggested that both Herbart and Bona Meyer would be worth reading on it.

The chapters on Fries and Herbart also include discussions of their responses to Kant’s practical philosophy. Beiser mentions Fries’s endorsement of the charge of ‘barrenness and emptiness of the universal law formula’ of the categorical imperative, which critique Fries found in a 1799 work by Gustav Hugo (but which had been made as early as 1786 by Herman Andreas Pistorius and was of course to be made famous by Hegel’s 1802 essay on natural law), and its need to be replaced or supplemented by something else, in Fries’s view by Kant’s own ‘end-in-itself’ formula (p. 58). The charge that the universal law formula is ‘formal’ and ‘empty’ with the ensuing debate over whether the necessary supplement can be found within Kant’s own work or only outside of it was to be a constant not only throughout the history of Neo-Kantianism but throughout nineteenth-century philosophy more broadly, for example in British idealism in F. H. Bradley, T. H. Green and Edward Caird. Beiser also reports Fries’s attempt to revise Kant’s distinction between Recht and Tugend by arguing that freedom in the external use of choice would require coercive enforcement only in a state of nature, not in a moral condition (p. 57); this discussion could be brought to bear on contemporary debates about the relation between Recht and morality. But what I found most interesting in this area was Beiser’s discussion of Herbart’s ‘aesthetic foundation of ethics’ (p. 125), his argument that ‘reason is not the source of moral obligation’ but that ‘obligation has to be determined on the basis of particular situations’ and morals ‘based on aesthetics’ (p. 126). Herbart also associates this claim with the ‘empty formalism’ charge, with the claim that ‘there are all kinds of universalizable imperatives, depending on the will from which one begins’ (p. 127), on the ground that one needs some kind of preference for an action in addition to the mere form of universalizability. On the surface, nothing could seem further from the spirit of Kant’s moral philosophy. But Beiser shows throughout the work that the Neo-Kantians were torn over Kant’s treatment of freedom, sometimes accepting the position of the Groundwork that the moral law is the causal law of the (noumenal) will, which makes immorality inexplicable (as Johann August Heinrich Ulrich had pointed out in his Eleutheriologie as early as 1788), sometimes inclining towards the Religion’s distinction between Wille and Willkür, which separates the moral law coming from pure practical reason and the decision whether or not to make it one’s supreme maxim; and if the latter distinguishes what it is to be (practically) rational from the decision to be rational, then it would seem that there must be room for something like a preference for being rational in the explanation of moral success, something which might not unreasonably be considered more like a feeling (if phenomenal) or desire (phenomenal or noumenal) than a pure cognition of rationality as such. If this is so, then Herbart might not have been completely off the mark – another interesting issue about which Beiser’s work can start us thinking.

Throughout part II of the book, the central issues are whether Kant’s transcendental method is really a form of psychology; whether the posit of things in themselves is a necessary part of Kant’s transcendental idealism or an incoherent excrescence; and whether transcendental idealism makes room for materialism or undermines it. On first issue, the answer tended to be affirmative; on the last issue, the prevailing tendency was to follow Kant’s own argument especially from the Paralogism of Pure Reason as revised in the second edition of the first Critique by holding that materialism might be an acceptable position for the practice of natural science (empirical realism) but not an ultimate truth about reality because the spatio-temporal framework it presupposes is not an ultimate truth (transcendental idealism). One of the more interesting positions that Beiser reports in this regard is that of Lange, who addressed the issue not by appeal to transcendental idealism but by arguing that ‘the progress of science consists … in replacing the concept of a thing with that of relations, so that the concept of force now replaces [that of] the atom’, and who ‘took his critique of hypostasis a step further, applying it to the concept of force itself’, as ‘really only a personification of the mathematical formulae that physicists use to describe and predict phenomena’ (p. 378). This is an important anticipation of Cassirer’s position in Substance and Function (1910), which might well be regarded as that philosopher’s most enduring work.

But the central debate in part II is that over the status of the thing in itself. On the one hand, many of Beiser’s figures, for example, Otto Liebmann, the author of Kant und die Epigonen (which we can now see not to have commenced Neo-Kantianism in 1865), held that the thing in itself is an incoherent addition to Kant’s model of cognition of objects, undermined by his own stricture against knowledge of anything other than the appearances informed by our own forms of intuition and conceptualization, space, time and the categories. But on the other hand, Lange recognized that ‘the thing-in-itself … is not so easily eliminable after all: that we are bound to postulate its existence as soon as we consider that our knowing faculties are finite, and that these faculties are only one way in which an independent reality is perceived’ (p. 381). Another way of putting this point is by saying that, if our forms of intuition and conceptualization have to be applied to manifolds other than themselves, then those manifolds must come from somewhere outside of them, and outside of the structure of our own minds – hence things in themselves. Whether or not the conception of such things can be invoked in moral philosophy, to solve the problem of freedom of the will or to ensure the possibility of the highest good, is another matter, on which few of the Neo-Kantians were eager to defend Kant; but it might indeed not be so easy to eliminate them from Kant’s theory of knowledge.

The central issues in part III are the transition from a psychological to an epistemological approach to transcendental philosophy and the continuing debate over the necessity of things in themselves to transcendental idealism. Both Cohen and Windelband are portrayed as converts to a purely epistemological approach. In Cohen’s case, the changeover, which Beiser describes as happening quite quickly in the year or two preceding the publication of Kants Theorie der Erfahrung, is connected with his adoption of the ‘logic of the sciences’ approach to the identity crisis of philosophy, in particular with his adoption of the position that transcendental philosophy analyses the presuppositions of natural science – an approach, it could be argued, more reminiscent of Kant’s ‘analytic’ method in the Prolegomena than of his ‘synthetic’ method in the first Critique. For Cohen, according to Beiser, ‘the a priori is strictly epistemological, consisting in the general conditions for knowledge of experience, where “experience” consists in the world as it is understood through mathematical physics’ (p. 483) – and you need both clauses to see how Cohen’s approach follows the analytic method of the Prolegomena as well as anticipating, for example, Henry Allison’s conception of ‘epistemic conditions’. In Windelband’s case, the new approach is closely connected with the idea that philosophy in general is normative rather than descriptive, an approach Windelband introduced in his lecture upon the centennial of the Critique of Pure Reason in 1881 (p. 495). Following Kant’s account of the role of the concept of an object in the Transcendental Deduction, Windelband holds that concepts of objects are ‘determinate rules for uniting representations’ (p. 496), and correspondingly conceives of moral and aesthetic concepts as norms for conduct and feeling (p. 497) rather than descriptions of anything given independently of those rules. Both approaches are connected to an eliminativist approach to things in themselves, for if concepts are conceived as rules for thinking rather than representations of given realities, they do not seem to need much other than themselves to represent. According to Beiser, Cohen ‘maintains that the thing-in-itself is simply a regulative or limiting concept of the understanding and … read in a constitutive sense, it is only an hypostasis of this concept’ (p. 490), and likewise for Windelband, ‘the thing-in-itself is a complete impossibility’ while ‘the idea of a thing-in-itself is simply the hypostasis of the concept of an object, which really derives from the synthesizing function of the understanding’ (p. 524). Both approaches may be seen as paralleling that of Charles Sanders Peirce, for whom the truth is simply what would be known at the end of enquiry, without any reference to an ontologically independent object, and even more that of the Kant interpretation of Wilfrid Sellars, who transforms the thing in itself into the ultimate even if only asymptotically approachable object of scientific knowledge. Riehl, on the other hand, ‘maintains that Kant’s philosophy is a form of realism insofar as it affirms the existence of things-in-themselves, which we know to be the ground of appearances, and insofar as it affirms the existence of a given manifold of sensation’ (p. 533). In other words, if there is appearance, there must be something that appears (as Kant insisted in the second edition Preface to the Critique).

As I have already suggested, Beiser plays his own cards close to his vest, and does not explicitly choose between a psychological or epistemological approach to transcendental philosophy and an eliminativist or non-eliminativist interpretation of transcendental idealism. I think that Kant makes his position on the latter clear in the Prolegomena, the second edition of the Critique, and subsequent sketches of the Refutation of Idealism: he thought he had stripped spatial and temporal properties from the objects that of course exist independently of our representations of them, just as previous philosophers had stripped secondary qualities such as colour and smell from them, in any but a dispositional sense, although for opposite reasons: previous philosophers had relocated secondary qualities from objects to our representations because of their apparent contingency, while Kant relocated spatio-temporality precisely because of its alleged necessity, which he thought could not be explained except by the (Vichian) thought that we ourselves make these properties. But he never for a moment thought he had provided any reason to doubt the existence of objects other than our representations of them, which is why it took him so long even to concede the necessity of a refutation of idealism. He always thought that if there are appearances, there must be something that appears, or that ‘affects’ us with an empirical manifold – although he would have spared us all a lot of agita over ‘noumenal causality’ if he had put ‘ground and consequence’ in the table of relational categories, as in correspondence with the logical function of hypothetical judgement he should have, and ‘cause and effect’ only in the list of schematized categories, for then he could have spoken of things in themselves as ‘grounding’ our manifolds of intuition without having had to speak of them as ‘causing’ our manifolds, thereby upsetting everyone by using a spatio-temporal conception to describe the relation of allegedly non-spatio-temporal objects to our sensibility.

More interesting, perhaps, is the issue of psychological versus epistemological approaches to the method of transcendental philosophy. If I have a single complaint about Beiser’s otherwise monumental work, it is that while describing the emergence of this distinction it does not question its validity. But I think it does need to be questioned. I think that something like Patricia Kitcher’s distinction between the analysis of cognitive tasks and theorizing about how such tasks are carried out in Kant’s Transcendental Psychology (1990) shows that transcendental philosophy can and indeed must have both a normative and a descriptive part, or a high-level analysis of conditions of the possibility of knowledge and more concrete theory about how such tasks are actually accomplished. But further, I think that careful analysis of some of Kant’s most important transcendental proofs (yes, I know he does not use the expression ‘transcendental argument’, but he does discuss ‘transcendental proof’ in the Doctrine of Method) shows that they do depend on certain key assumptions which are not themselves norms but more like statements of fact, and statements of fact that might be more plausibly classified as psychological than anything else. What I have in mind are things like the premise of the Transcendental Deduction that I can attach the ‘I think’ to any of my representations or that of the Analogies of Experience that the temporal order of successive representations can always be varied in imagination because time itself cannot be directly perceived. In the latter case, for example, the norm that we ought to find a cause for any perceived event is justified by the fact that we cannot perceive temporal order directly and therefore must infer it from causal laws. If this example is indicative, then Kant’s transcendental proofs conclude with epistemological norms but begin with certain psychological assumptions. The case of practical rather than theoretical philosophy may be different, for there Kant seems to infer a fact of (noumenal) psychology, that we are free, from a norm, ‘ought implies can’, although perhaps we should instead infer limitations on the applicability of that norm, that is, constraints on ascription of responsibility, from inescapable facts about human (empirical) psychology. But either way, it seems that we need to think about both epistemological and practical norms, as the later Neo-Kantians did, and psychological facts, as the earlier ones did.

That Beiser’s work stimulated me to think about this issue is only another example of its extraordinary accomplishment. I am not aware of any other work that has covered this rich and fascinating period in the history of modern philosophy with equal breadth and depth. While I might have wanted Beiser to have revealed a little more of his own view of the philosophical merits of the positions he has discussed, from what I know of the various figures he treats, which is certainly not as much as he does, I could find no errors in what he says. For outright criticism, I would have to confine myself to the history of architecture: Alois Riehl’s house in the Berlin suburb of Neubabelsberg was not designed in 1906–7 by the ‘up-and-coming Walter Gropius’ (p. 532), but by the 20-year-old and totally unknown Ludwig Mies van der Rohe. But even Homer nods! Otherwise, The Genesis of Neo-Kantianism is to be recommended without reservation to every serious student of Kant.