In 1871 the young Hermann Cohen, then only twenty-nine years old, published his Kants Theorie der Erfahrung.1 Though now nearly forgotten, this little book was revolutionary in its day. It is indeed no exaggeration to say that it marks “a turning point in the history of the interpretation of Kant.”2 It laid down the foundation not only for the philosophy of the Marburg school but also for much of our contemporary understanding of Kant.
Why, more specifically, was Cohen’s book so important? Part of its significance lay with its method of interpretation. In an age when Kant reception was very anachronistic, twisting Kantian texts to fit contemporary needs, Cohen strived for a proper historical understanding of Kant. A student of August Böckh and Adolf Trendelenburg, he applied the same philological and historical methods to Kant as his mentors had to classical texts. Thus in his preface he states that he intends to uncover “the historical Kant” and that he wants to interpret him in his own terms, contrary to the attempts of his contemporaries, who were bent on understanding Kant in their terms. And so, with all the love and care of a Talmudic scholar, Cohen read and re-read the Kantian texts, finding in their exact words depths and subtleties of meaning that Kant’s critics had simply ignored.3
Another part of the significance of Cohen’s book lay with his interpretation of Kant. Cohen emphasized aspects of the critical philosophy that had been much neglected in the neo-Kantian tradition, viz., the transcendental deduction of the concepts of understanding, the role of synthesis in constituting objects of experience, the place of empirical realism within transcendental idealism, and the principle behind Kant’s “new method of thought”. More important than all these particular points, though, was Cohen’s general conception of Kant’s aims and methods. He stressed the essentially epistemological purpose of Kant’s enquiry. Kant’s central question was not the quid facti?—“What are the causes of representations?”—but the quid juris?—“What is the justification of synthetic a priori judgements?” Cohen insisted upon the transcendental status of Kant’s enquiry, drawing attention to how it is not first-order, about objects in the world, but second-order, about the conditions for knowledge of such objects. Accordingly, he understood the a priori neither in psychological terms, as innate ideas, nor in metaphysical terms, as Platonic archetypes, but in epistemological terms, as the general conditions for knowledge of experience. This was a sharp break with the neo-Kantian tradition since its inception with Fries. We have already seen how that tradition was essentially psychologistic, interpreting Kant’s epistemology as a proto-psychology. Only at rare moments did Fischer and Liebmann question this interpretation, but then only to fall back into it. Though the young Cohen too was influenced by psychologism, he sharply distinguished it from epistemological questions. The psychological interpretations had pushed Kant’s epistemological intent into the background; it was Cohen’s great contribution to have pushed it again into the foreground.4
The great significance of Cohen’s book for the Marburg school lay in its ideas about philosophical method. A crucial part of Kant’s epistemological enterprise, Cohen taught, lay in his transcendental method. This method begins with the “fact of science”, viz., Newtonian physics, and then determines the conditions of its possibility. It does not attempt to prove this fact from higher principles or to secure it against sceptical doubt; rather, it accepts it as a given and then determines its underlying conditions. Philosophy is thus the analysis of the logic of scientific enquiry rather than the foundationalist enterprise of demonstrating knowledge from first principles. In an article published in 1912, Paul Natorp stressed the importance of this method and conception of philosophy for the Marburg school.5 Their germ lay forty years earlier in Cohen’s little book.
Given its great historical importance, it is worthwhile to investigate the origins of Cohen’s book. Such an investigation promises to shed some light on not only the genesis of Marburg neo-Kantianism but also on the profound turn in the neo-Kantian tradition, the turn away from psychology and towards epistemology proper. What led Cohen to his epistemological or transcendental interpretation of Kant? What made him break with psychologism? Prima facie these seem to be sterile questions about the history of scholarship. But it is important to see that much more was at stake. The proper interpretation of Kant was for Cohen an issue of the greatest philosophical and cultural importance. For he saw Kant’s philosophy, and specifically his interpretation of that philosophy, as the only solution to the cultural crises of his age. His neo-Kantian philosophy grew directly out of two major philosophical disputes of the mid-19th century: the materialism controversy and the Fischer–Trendelenburg debate. To understand his philosophy, we need to reconstruct how Cohen conceived it as a solution to the problems raised by these disputes.
There is another reason why we should re-examine Cohen’s book. Although it has often been the subject of scholarly analysis, it has also been misinterpreted. All too often the book has been interpreted anachronistically, as if it were an anticipation of Cohen’s later philosophy in his 1902 Logik der reinen Erkenntnis.6 But Cohen’s early work has to be understood in its own terms, not least because it is at odds in fundamental points with his later philosophy. Only when these points are fully taken into account will we have an accurate account of the origins of Marburg neo-Kantianism and Cohen’s own philosophical development.
Our purpose here, then, is to reconstruct the genesis of Cohen’s book, to restore its original purpose and meaning, removing the layers of anachronistic interpretation imposed upon it. This quickly proves to be a daunting undertaking. For there is a severe lack of documentation in trying to understand Cohen’s intellectual development. Tragically, too many materials have been lost or destroyed.7 As a result, we have a much more detailed knowledge of philosophers from the remote past than we do of Cohen. Incredibile sed verum: we can know much more about Leibniz than Cohen.
This is all the more a pity because, from the little we do know, Cohen’s intellectual development is fascinating. The story is dramatic. For Cohen came into his own only by breaking with his three great teachers: Heymann Steinthal, Friedrich Lange and Adolf Trendelenburg. Though he revered them all, he also rebelled against them. If a teacher is somehow a father, as the German word Doktorvater has it, then Cohen came into his own only through a treble parricide.
The story is also mysterious. For when we go back to Cohen’s early years we find that he was himself a staunch advocate of the very psychological approach to Kant that he would later repudiate. Indeed, Cohen championed this approach as late as 1870, the very year he wrote Kants Theorie der Erfahrung. Why did Cohen make such a volte face, and so suddenly?
Our story begins in May 1865 in Berlin. It was then and there that Cohen formed a friendship with Heymann Steinthal, who was a pioneer in the fledgling disciplines of linguistics and ethnology.8 Along with his friend and brother-in-law, Moritz Lazarus, Steinthal had founded and edited a journal devoted to these new disciplines, the Zeitschrift für Völkerpsychologie und Sprachwissenschaft.9 Seeing in Cohen a promising disciple, Steinthal decided to give the struggling young student the opportunity to prove himself by writing articles for his new journal. In commissioning articles Steinthal expected that Cohen would promote the intellectual cause behind the journal; and in this respect, he was not to be disappointed. The young Cohen soon became a champion of the new science of Völkerpsychologie.
Völkerpsychologie was the ancestor of modern anthropology or social psychology.10 Its subject matter was the Volksgeist, the mind or spirit (Geist) of a nation or people (Volk).11 Though it has vague and mystical connotations, Lazarus went to great pains to define the Volksgeist in clear and non-metaphysical terms as “what is common to the inner activity of all individuals of a nation”.12 The task of Völkerpsychologie was to examine how this Volksgeist appeared in, and affected the lives and psyches of, individuals. Its guiding assumption is that a community is an irreducible whole that affects the psychology of its individual members; in other words, we can fully understand the individual psyche only by seeing it within its historical and cultural context. While Völkerpsychologie has a Hegelian provenance and inspiration—it has an obvious debt to Hegel’s category of objective spirit—both Lazarus and Steinthal insisted that the method of Völkerpsychologie should be utterly empirical and thoroughly mechanical.13 They were indeed devotees of the psychology of Johann Friedrich Herbart, which was the dominant empirical psychology of their day. There are four salient characteristics of Herbart’s psychology, most of which reappear in Völkerpsychologie: an emphasis on mechanical explanation; an insistence that the mind consists in processes rather than powers; the analysis of complex psychic processes into more basic units, so that parts precede wholes; and the application of mathematics to explain the dynamics of mental events. It is noteworthy that, though Herbart advocated a mechanical and mathematical approach to the mind, he was no materialist himself, and he would often stress the sui generis status of the mental. Thus he found the concept of the soul indispensable in defining the subject matter of psychology.14
For better or worse, the young Cohen was very much a follower of this Herbartian psychology.15 His devotion to Herbart’s psychological programme appears in the first thesis of his doctoral promotion: “Omnem philosophiae progressum in psychologia constitutum esse.”16 Two of his early articles for the Zeitschrift für Völkerpsychologie apply Herbart’s psychology to the fields of religion and poetry.17 Though he appropriates Kantian concepts—the a priori, the unity of apperception, form and content—Cohen interprets them in Herbartian or Steinthalian terms. His interpretation of these concepts is flatly at odds with Kant’s transcendental psychology, the very psychology Cohen will defend in Kants Theorie der Erfahrung.
The contrast between Cohen’s early and later psychology is striking, and in three specific respects. First, the early Cohen understands the a priori in terms of temporal rather than logical priority, that is, an a priori representation is prior in time to others, and it determines how the later or a posteriori ones appear to consciousness.18 The later Cohen, however, will stress the importance of a strictly logical interpretation of the a priori, so that it is a necessary condition of knowledge of experience. Second, the early Cohen interprets the unity of apperception strictly in terms of a mechanism of association, so that earlier representations exert an attractive force on later ones, forming habits of association according to the degree that past sequences have been repeated in consciousness.19 The later Cohen, by contrast, will interpret the unity of apperception in transcendental rather than mechanical terms, so that “the affinity of appearances”, their universal and necessary form, is the precondition for their association. Third, the early Cohen accepts a distinction between the form and content of representations that is entirely non-Kantian. According to this early distinction, the content is the intentional object, what we represent, while the form is the feeling or tone of consciousness in general, which consists in nerve movements.20 If all these differences were not enough, there are passages in these early articles where Cohen adopts the language of Herbart’s alleged realism, writing as if objects are simply given to us and our representations resemble them.21 So in these passages we are very far from the transcendental idealism that Cohen will later advocate in Kants Theorie der Erfahrung.
The articles for the Zeitschrift für Völkerpsychologie that expound a Herbartian psychology were published between 1866 and 1869. It would seem, then, that Cohen had more than a year to grow out of these doctrines before writing Kants Theorie der Erfahrung, which appeared in 1871. But the truth of the matter is much more puzzling. For Cohen finished his Kant book in the autumn of 1870,22 which would have given him much less time to reverse his thinking. Furthermore, some of his later articles for the Zeitschrift, which appeared in 1871, still advocate a psychological approach to the history of philosophy.23 Recognizing the discrepancy between the early and later Cohen and the short time span between them, Klaus Christian Köhnke has postulated an abrupt “revolution” (Umbruch) in Cohen’s thinking, one that took place in only one year.24 But even allowing for generous lags in publication dates, the revolution must have been even more swift and sudden. It took place probably in a matter of months. In any case, it is safe to assume that Cohen had a very creative and inspired summer in 1870! For he wrote his book very quickly, in the course of a single summer.25 But what happened in those heady months to reverse so drastically the course of his thinking?
Someone might object that we should not overproblematize or overdramatize, because the break in Cohen’s thinking is not so sharp and clean.26 It is not as if we suddenly have a psychological Cohen in the 1860s, and then an epistemological Cohen in 1870, who springs fully formed from a single inspiring summer. There are anticipations of the later Cohen in the early essays, and residues of the early Cohen in the later work. After all, it is not as if Cohen completely understood the full implications of his new epistemological approach in 1870, so that he abruptly changed course and repudiated Herbartian psychology. Rather, he gradually realized the results of his new thinking throughout the 1870s.
Though perfectly valid, these points do not make the problem go away. They should not be stretched to obscure the very real and very big differences between early and later Cohen, that is, Cohen before and after the summer of 1870. Two basic differences should not be underestimated. First, there is an important difference in emphasis. While prior to the summer of 1870 Cohen thinks that the psychological approach is the key to understanding Kant, after that summer he insists that the epistemological approach is crucial; the earlier psychology is pushed into the background, and it is indeed regarded as misleading for an understanding of Kant. Why? Second, in Kants Theorie der Erfahrung Cohen repudiates the basic tenets of Herbart’s psychology, more specifically its conceptions of the a priori, apperception, and the form–content dualism. The many passages criticizing Herbart’s psychological approach to Kant are indeed in striking contrast to his earlier work. The Herbartian empiricist and realist had become a Kantian transcendental idealist. Why?
To answer these questions we need to take a closer look at some of Cohen’s early essays. These will reveal to us a budding transcendental philosopher, one ready to burst out of the confines of Völkerpsychologie, and one on the threshold of his later interpretation of Kant.
Cohen’s debut article for the Steinthal-Lazarus journal was his ‘Die Platonische Ideenlehre psychologisch entwickelt’, which appeared in 1866.27 This article has been the source of some controversy among Cohen scholars, who quarrel about whether it should be placed close to, or far removed from, his later position in Kants Theorie der Erfahrung. Some have seen it as the source and inspiration for his later position.28 Others, however, have regarded it as the very antithesis of that position, and they have even dismissed it as “of negligible interest from the standpoint of critical idealism”.29 Which of these interpretations is correct? Only a close examination of the article will tell.
The main questions Cohen intends to answer in his article are “What is the Platonic idea?” and “By what psychic process did it arise?” (405). His strategy for answering them, which he explains in the very beginning of the article, reveals his approach and intentions. Cohen states that he wants to complete “the literary investigation” that Kant began in the ‘Dialektik’ of the first Kritik where he proposed “a milder interpretation” of Plato’s theory of ideas (405).30 According to that interpretation, the theory of ideas should be read not in a metaphysical or constitutive sense but in a regulative one, so that ideas are understood not as entities or substances but as goals or ideals for enquiry. The chief purpose of Cohen’s article is to defend this “milder” Kantian interpretation.
Though the regulative interpretation of Plato’s theory of ideas brings this article close to Cohen’s later views in Kants Theorie der Erfahrung, it also seems in another respect far removed from them. For Cohen still wants to give a “psychological” interpretation of Plato’s theory. The very title of the article—‘Platonische Ideenlehre psychologisch entwickelt’—shows that he is following the psychological programme of Völkerpsychologie–the very programme from which he will distance himself in his later work. The article originally appeared as part of a volume of the Zeitschrift für Völkerpsychologie devoted to the history of science, where the methods of Völkerpsychologie were to be applied to explain scientific discoveries. The aim of Cohen’s article was to extend those methods into the domain of classical studies.
What, more precisely, did it mean to explain Plato’s theory of ideas “psychologically”? Cohen tells us that he wants to identify the “psychic process” behind Plato’s theory. By this he means not the motives behind the theory but the facts it intends to signify. Cohen thinks that Plato meant to refer to the psychic process behind intellectual and artistic creation. Plato identified this process with the act of intuiting or seeing something, with how the philosopher or artist saw “in his mind’s eye” the object he intended to create. The idea was thus the artist’s or philosopher’s creative vision, his imagining what he intended to create (427). Cohen stresses that, in its original meaning, the idea signified more an activity than an entity. What Plato originally meant by the idea, we are told, is not an intelligible or substantial form but an act of “intellectual intuition” (intellektuelle Anschauung). He admits, however, that Plato sometimes wrote about his ideas as if they were things, and notes that in his later dialogues he fell into the habit of hypostasizing them. Rather than talking directly about the activities themselves, Plato would sometimes refer to them obliquely through their objects, so that “the original being of intuiting surreptitiously became the being of the intuited” (428). Because of his careless metaphoric language, and because of the influence of the Pythagoreans, Plato hypostasized his ideas, making them into a peculiar kind of entity (428–429, 440–441, 448).
The controversy surrounding this article arises from Cohen’s interpretation of Platonic ideas in terms of intellectual intuition. In some puzzling and often cited lines, Cohen writes that if Plato only kept true to the original meaning of his theory, he would have been “the ancestor of Fichte’s intellectual intuition, of transcendental idealism itself” (427). It is this striking equation of Fichte’s intellectual intuition with transcendental idealism that has proven so troubling. Some see it as the source of Cohen’s philosophy in Kants Theorie der Erfahrung, whereas others see it as a violation of that very philosophy. But, in either case, both parties to the dispute assume that Cohen, in attributing Fichtean intellectual intuition to transcendental idealism, is advocating a very metaphysical interpretation of Kant.
The shared premise behind both views is false. A closer examination of the texts reveals that Cohen is not advocating a metaphysical interpretation of Kant at all, and that he is really only following through with his original plan to defend Kant’s “milder” regulative interpretation of the theory of ideas. It is of the first importance to note precisely how Cohen understands his act of intellectual intuition. He explains that the highest form of intellectual intuition—that which would grasp the totality of all things—consists in the idea of a purpose. “The unity of ideas is the idea of the good”, we are told, where the good is the purpose behind all things, the ultimate reason for their creation (449). At this very point, Cohen admits, idealism loses its psychological character and appears to become teleological, even theological (450). So it now seems that we are on very metaphysical terrain indeed, and that Plato’s intellectual intuition is still very far removed from Kant, who had notoriously repudiated all intellectual intuition.31 But it is precisely here, Cohen argues, that Kant understood Plato better than Plato understood himself. Kant perceived the original meaning of Plato’s idealism, which gives the ideas not a constitutive but a regulative meaning. He recognized that Plato’s ideas should be read as “a regulative concept for reflective judgement” (451). Kant himself had famously discussed intellectual intuition in §77 of the Kritik der Urteilskraft, and stressed that it should be given a purely regulative meaning for our finite understanding. Cohen argues that this was the original meaning Plato intended to give to it all along, even though in his enthusiasm he was sometimes seduced to write about ideas as if they were things (450–454).
Now that we have seen the faulty common premise behind both interpretations—the assumption that Cohen understands intellectual intuition in a metaphysical or constitutive sense—we can cut through the controversy between them and place this article properly within Cohen’s intellectual development. Since Cohen understands intellectual intuition in a regulative sense, this article does indeed anticipate his later more critical position, which interprets Plato’s theory of ideas strictly methodologically. It is therefore a mistake to argue that this article is of no relevance to Cohen’s later transcendental idealism. However, it is no less an error to assume that this article foreshadows the later Cohen because it anticipates an alleged metaphysics involved in Kants Theorie der Erfahrung. That presupposes not only a mistaken account of the later work, which is not metaphysical at all, as we shall soon see,32 but also a false metaphysical interpretation of the early article.
The important but overlooked fact that Cohen is intent on a regulative reading of Kant in this early article shows us that, even in the 1860s, he was very much a Kantian philosopher, and indeed in just the ways he appears in Kants Theorie der Erfahrung. In 1868 Cohen was not Platonizing Kant; he was Kantianizing Plato. Realizing this helps us to bridge the gap between the Cohen of the 1860s and the Cohen of 1870. We can close it even more if we consider the next stage in his philosophical development.
Cohen’s growing distance from the psychological programme of Völkerpsychologie appears in his next major article for the Zeitschrift, ‘Mythologische Vorstellungen von Gott und Seele’, which was published in 1868.33 It now becomes perfectly plain that we cannot describe the young Cohen only as a budding anthropologist and psychologist along the lines of Völkerpsychologie. Of course, he is that, but he is also much more. For in the opening pages of the article he firmly and expressly casts himself in the role of a critical philosopher whose proper field is epistemology. That field is then explicitly distinguished from all empirical disciplines, including psychology and anthropology (400). The primary task of the critical philosopher, Cohen tells us, is to engage in “the identification and examination” of “the great ideas” of philosophy (400). Among these great ideas are God and the soul, which will be the subject of the article. Although Cohen intends to determine the psychological and ethnological origin of these ideas, he also distances himself from this enterprise, confessing that such an investigation is only an “excursion” (Streifzug) for him, a foray into “a field alien from his own” (399).
We learn from Cohen’s introductory disquisition exactly how he understands the method of epistemology, which turns out to be very far from that of anthropology or psychology. He calls the method of epistemology “deductive critique”, where it is “deductive” for the simple reason that “True thinking is deduction even in the realm of the inductive” (398). To claim that proper thinking is deduction, and that the business of epistemology is to examine such thinking, is to take epistemology outside the field of empirical enquiry. Cohen then goes on to explain in more precise terms the subject matter of deductive critique (398–399). It comprises two aspects of concepts: their “logical innerness” and their “metaphysical power”. Their logical innerness consists in the logical consistency of their elements; and their metaphysical power consists in their capacity to solve problems.
After explaining the method of deductive critique, Cohen then sharply distinguishes it from that of psychology. While deductive critique determines “the metaphysical validity” of concepts, psychology engages in an analysis of their elements and origins, the “mechanism” from which they arose in the mind and in history (399). Despite the difference between these methods, Cohen reassures us that they are complementary, and that deductive critique has much to gain from psychology. From the empirical analysis of concepts, deductive critique receives new materials and starting points for an investigation into their consistency and logical powers (399). Cohen even goes on to stress the great value of psychological analysis for philosophy: it shows us that ideas which appear eternal really arose from history, and that ideas which appear unanalysable grew from simpler primitive elements (400). Nevertheless, he is also perfectly clear about, and indeed insists upon, the limits of psychology: “psychological analysis decides nothing about the metaphysical validity of a concept” (399). Already here, then, we see the clear distinction between the Quid juris? and Quid facti? that will play a decisive role in Kants Theorie der Erfahrung.
Cohen’s statement that deduction is the proper business of thinking, his careful explanation of the difference between the deductive and psychological methods, and his confession that his essay is only an “excursion” for him, all make it plain that he is distancing himself from the Lazarus-Steinthal programme. Though he stressed how that programme could be very illuminating for the philosopher, he also made it clear that was not where his true interests lay. He was first and foremost a philosopher, and his discipline had very different methods and goals from empirical psychology and anthropology.
It is no less plain from this early article that Cohen saw Kant as his model philosopher, for he sees him as the source of the method of deductive critique (398). Already in 1868, then, Cohen identified himself strongly with Kant’s critical philosophy, and he distinguished its methods and concerns from those of psychology and anthropology. Thus the main idea for Kants Theorie der Erfahrung was already laid; the only remaining business was to develop it.
Thus Cohen’s writings in the late 1860s show us a philosopher already sympathetic to Kant, and one who interpreted him in non-psychological terms. These conclusions take us part of the way towards understanding the origins of Cohen’s first book; but they hardly take us the whole way. We still have to answer two basic questions: Why was Cohen sympathetic to Kant in the first place? And why did he interpret him non-psychologically? Unless we have some answers to these questions, our explanation has not even begun.
As elementary as they are, these questions permit no straightforward answer. There is simply not enough historical evidence to answer them with certainty. Cohen’s surviving correspondence is very scanty, and his early writings leave few clues about his philosophical leanings outside Völkerpsychologie. It is necessary to reconstruct, then, the path of Cohen’s development, using evidence from his historical context and later writings.
If we piece together information from these sources, it becomes clear that one event in particular played a decisive role in Cohen’s intellectual development, both in moving him towards Kant and in making him interpret Kant in transcendental rather than psychological terms. That event was the materialism controversy.
We know that the young Cohen, like most men of his generation, was deeply preoccupied with the chief issue raised by the pantheism controversy: the conflict between reason and faith. From his earliest days, Cohen had been searching for some way to reconcile a scientific naturalism with his religious heritage.34 It is noteworthy, however, that, initially, Cohen did not find the solution to this problem in Kant but in a very different place. In a short anonymous article he wrote in 1868, ‘Virchow und die Juden’, Cohen described the modern worldview as “the position of scientific materialism and ethical idealism”.35 Rather than shirking back from materialism, Cohen seemed to embrace it. It is important to see, however, that by “scientific materialism” he did not mean the materialism of Marx and Engels, still less that of Vogt, Moleschott, Czolbe or Büchner. Rather, what he had in mind was the materialism of Heinrich Heine, which was inspired by Spinoza’s pantheism.36 According to Heine, Spinoza’s single infinite substance, which manifests itself in all of nature, sanctifies the reality of matter, so that we have good reason to appreciate the things of this world, the joys of the senses, which are as divine as the most ethereal realm of the spirit.37 To the young Cohen, Heine’s pantheism seemed to promise the very synthesis of reason and faith he was so ardently seeking. After all, Spinoza had not only naturalized the divine: he had also divinized nature. Furthermore, Cohen saw Spinoza’s pantheism as the developed and refined form of Jewish monotheism.38
So, as late as 1868, Cohen was still very far from embracing all aspects of Kant’s philosophy. He had accepted Kant’s interpretation of Plato; he had endorsed his critical conception of philosophy; but he did not see the import of Kant’s philosophy for the resolution of the all-important conflict between reason and faith. Like so many thinkers of the Goethezeit, he had seen Spinozism as the solution to that conflict. Before he could become a full-fledged Kantian, then, he would first have to abandon that pantheism. Someone would have to convince him that Heine’s solution to the dilemma of reason versus faith does not work, and that ultimately only the Kantian solution is really viable. Who could this someone be?
We do not have to search long for an answer. It was the pied piper of Marburg, Friedrich Albert Lange. Like many young men in the late 1860s, Cohen had fallen under the spell of Lange’s Geschichte des Materialismus.39 Lange had shown him that Kant’s philosophy offers a middle path between a rational materialism and an irrational fideism, that it alone can save the autonomy of our moral and religious ideals without metaphysics, and that it alone can uphold the principles of mechanism and naturalism without jeopardizing morality and religion. But Lange’s arguments would have also taught Cohen the weaknesses in the traditional Spinozist position. For all the attractions of Spinoza’s divinized nature and naturalized divinity, his philosophy was still much too metaphysical, much too dogmatic. It just assumed that we could have some knowledge of the infinite, that we could have some knowledge of the thing-in-itself. Lange had shown all too well how the materialists were naive, how they simply presupposed that we have knowledge of a material reality independent of us, and how that presupposition did not square with the latest results of the natural sciences. Even if Spinoza’s refined and religious materialism was very different from the coarse and crass materialism of “Vogt und Consorten”, it still shared the same naive realism.
That the young Hermann Cohen was inspired by Lange’s book we have no reason to doubt. There is no direct evidence from the 1860s about its influence, no letters or fragments that tell us of its immediate impact. We do not even know when Cohen read Lange, though it must have been sometime after he wrote ‘Virchow und die Juden’, which was in 1868. Nevertheless, we have more than enough evidence from the 1870s that Cohen had read Lange’s book, and that it had a great impact upon him. There is a letter to Lange from November 1871, in which Cohen praises Lange’s “noble work on materialism” and thanks him for its “liberating” effect.40 Such, indeed, were Cohen’s debts to Lange that he dedicated the second edition of Kants Theorie der Erfahrung to his memory, and in 1876 he wrote two appreciative obituaries for him.41 Fittingly enough, in 1898 Cohen wrote a preface and introduction for the sixth edition of Lange’s Geschichte des Materialismus.42 So the evidence more than suggests that it was Lange who effected Cohen’s final conversion to the critical philosophy.
Yet, for all the importance Lange’s book had for him, Cohen had his reservations about it. He made two basic criticisms, both decisive for developing his own transcendental interpretation of Kant. First, he took issue with Lange’s psychological and physiological interpretation of Kantian epistemology. The problem with that interpretation is that it placed the Kantian subject firmly within the realm of phenomena and nature. To Cohen, such a conception of the a priori showed that Lange had not sufficiently liberated himself from the very materialism he denounced. Rather than seeing the a priori in terms of epistemological conditions, Lange had hypostasized them, as if they were physiological conditions, making them just another kind of phenomenon in the natural world. That seemed to open the back door to materialism, allowing it to enter into the most sacred chambers of transcendental philosophy itself. Second, Cohen took exception to Lange’s sceptical attitude towards Kant’s practical philosophy, which Lange regarded as the weakest side of Kant’s system. Lange regarded the noumenal realm beyond the sphere of phenomena as spooky and mysterious, as a hypostasis of our rational activities. For the young Cohen, however, that attitude was simply unacceptable, characteristic of the very materialism Lange attempted to refute. Lange had failed to appreciate one of the central teachings of the critical philosophy: that morality and religion could be based on practical reason. It was to counter Lange’s scepticism about morality and religion, Cohen later said, that he wrote Kants Begründung der Ethik.43
Lange’s shortcomings in understanding Kant, Cohen believed, were ultimately traceable to one underlying cause: his failure to grasp the transcendental dimension of Kant’s philosophy. Such is his assessment of Lange, in the first edition of Kants Theorie der Erfahrung, where he criticizes him for misunderstanding the epistemological purpose and status of Kant’s discourse (207–208). In his later 1875 obituary in the Preussische Jahrbücher he expanded this point by saying that critical idealism is not a metaphysics but a method, a procedure whose task is to determine the validity of knowledge claims rather than any truths about the existence of things.44 The idea, the stock-in-trade of idealism, is not a thing but “an epistemological symbol of value” (ein erkenntnistheoretisches Werthzeichen).45 Lange would have had a more adequate understanding of idealism, Cohen maintained, if he had only appreciated another thinker, indeed the very father of the idealist tradition: namely, Plato.46 Plato’s realm of ideas, if it is only properly understood, is not about a special kind of transcendent thing, as Lange interpreted it, but about a realm of value and validity completely distinct from that of existence. It is only by appreciating the sui generis status of this realm, he taught, that we secure the place of value in the world and find its source in practical reason.
It was Cohen’s response to the materialism controversy, then, that had fully and finally converted him to Kant, and that had even pushed him in the direction of an anti-psychological interpretation of his philosophy. But the materialism controversy alone did not spawn Kants Theorie der Erfahrung. It needed a more direct stimulus and occasion. This was provided by the Trendelenburg–Fischer dispute, of which Cohen became a major contributor.
It was impossible for the young Cohen to ignore the dispute which took place right on his doorstep, so to speak. For Trendelenburg had been his teacher in Berlin,47 and he was so exactly when the dispute was at its very height in the mid 1860s. Now that Lange had converted him to Kant, he could hardly be indifferent to the controversy, which concerned such major issues of Kant interpretation. A contribution to the dispute would also be the perfect opportunity to show his intellectual mettle. And so Cohen wrote an article on the dispute, his ‘Zur Kontroverse zwischen Trendelenburg und Kuno Fischer’, which appeared in 1871 in the Zeitschrift für Völkerpsychologie und Sprachwissenschaft.48 This sketched in nuce the position he took in Kants Theorie der Erfahrung, which was his chief contribution to the dispute.49
By July 1870 Cohen had already formulated the essence of his position. Though he felt greatly indebted to Trendelenburg, he decided to take issue with his former teacher. His resolve and stance could not be more explicit: “I must attack Trendelenburg. He has completely misunderstood Kant in my view. So much so, that Fischer was right to wonder how little he has been understood by his colleagues.”50 There is some irony in the fact that Cohen chose to show the internal coherence of Kant’s philosophy against Trendelenburg, for this was to apply his master’s own methods against him. It was a central doctrine of the “Trendelenburg school” to interpret each philosopher sympathetically and from within, to show how his thought formed a coherent whole from a few governing ideas.51 That was just the lesson that Trendelenburg had failed to apply to his own interpretation of Kant, Cohen believed. Trendelenburg would now be getting a dose of his own medicine.
Cohen’s article ‘Zur Kontroverse zwischen Trendelenburg und Kuno Fischer’ begins by acknowledging the great philosophical importance of the dispute. We are told roundly that its issues concern all philosophy (249). They raise the fundamental problem of whether “the nature of things depends on the conditions of our mind? or the laws of nature can and must confirm our thinking?” In other words, at stake is the great dispute between idealism and realism. Having paid tribute to the issues at hand, Cohen then makes a decisive methodological claim about how to resolve them. All thinking about the dispute can go forward, he says, only if one accepts Kant’s distinction between psychology and metaphysics (250). This distinction is between matters of psychological fact and matters of a priori principle, which can be settled only through deductive reasoning.52 This was Cohen’s way of saying that the dispute was fundamentally epistemological rather than psychological.
Cohen analyses the dispute into two central questions: 1) Has Trendelenburg proven that there is a gap in Kant’s proofs for the exclusive subjectivity of space and time?, and 2) Has Trendelenburg shown that Fischer has taken up non-Kantian elements in his exposition of the Kantian theory of space and time? (251). Cohen complains that the first question, which is the really important one, has been obscured by all the dust raised by the dispute. Rather than examining the logic of Trendelenburg’s arguments, Fischer has been more concerned to defend his interpretation of Kant. But this is just as well, Cohen says, because a complete answer to the first question would require nothing less than a full examination of Kant’s theory of experience and the proofs of the Transcendental Aesthetic. Because this goes beyond the limits of a single article, he has decided to leave aside the first question, which he will treat in a more comprehensive monograph, what eventually became Kants Theorie der Erfahrung. The article will deal with only the second question and another subsidiary one: Whether Fischer has proven against Trendelenburg that there is no gap in Kant’s reasoning?
Although Cohen, true to his word, devotes much of his article to these lesser questions, he still reveals the main lines of the position he later takes in Kants Theorie der Erfahrung. Regarding the first question, Cohen denies that Trendelenburg has shown that there is a gap in Kant’s proofs for the exclusive subjectivity of space and time. The problem with Trendelenburg’s interpretation, he argues, is that it begins with some very dogmatic assumptions—assumptions that Kant decisively rejects. Trendelenburg assumes that the a priori is a psychological faculty, and that the world as we experience it, as it appears to our senses, consists in things-in-themselves (255). He has his own concept of objectivity, which is essentially that of transcendental realism, according to which we know an object when our representations correspond to the thing-in-itself, which is given in experience. But it is just this concept of objectivity, Cohen argues, that Kant intends to overthrow with his transcendental idealism. The central argument of Kants Theorie der Erfahrung will consist in showing that Kant’s Copernican Revolution involves a very different conception of objectivity, one according to which objectivity is constituted by our a priori forms of intuition and understanding.
It is noteworthy that, regarding the all-important first question, Cohen takes a firm and explicit stance with Fischer against Trendelenburg. It is sometimes said that this article sides more with Trendelenburg against Fischer because the young Cohen was reluctant to take issue with his venerable teacher.53 But the very opposite is the case. Although Cohen does side with Trendelenburg regarding the lesser second question, he stands against him on the more important first one. He states firmly his agreement with Fischer that not a single sentence of Kant’s philosophy would be true if Kant had not refuted the third possibility. Fischer was indeed right: the entire Kantian system, the central doctrines of transcendental idealism, would collapse if the a priori forms of subjectivity were also true of things-in-themselves (260). Though Fischer was correct in substance, he did not establish clear and convincing arguments for his interpretation. If he is to be a good advocate of Kant, Fischer must prove and not simply proclaim his interpretation (260). Cohen sees himself as taking on the mantle of an advocate of Kant, confident that he can succeed where Fischer has plainly failed. That task he will undertake in his forthcoming book.
We will leave aside here Cohen’s response to the lesser questions and his long stance on behalf of Trendelenburg against Fischer. These concern matters that are peripheral to the main issues of the dispute, as Cohen himself acknowledged. Noteworthy, however, is Cohen’s concluding disquisition on method in the history of philosophy (290–296). Here he distances himself from the methods of both Fischer and Trendelenburg. He agrees with Trendelenburg’s critique of Fischer’s “free reconstructive method”, which introduces terms and ideas into a philosopher even when there is no textual evidence for them; but he also cannot accept Trendelenburg’s view that the only interpretation of a philosopher should be “an intricate mosaic” where the scholar reconstructs a philosopher’s meaning from a mass of fragments. A mosaic, no matter how finely reconstructed, is still not a single cohesive whole. The best interpretation of a philosopher is one that understands him according to his own “basic thought” (Grundgedanke) (292–293).
The crucial question remains, however, how one finds this basic thought? It is in his answer to this question that Cohen surprises us and reveals how much his thinking was still in flux. Having made a firm distinction between metaphysical and psychological questions at the beginning of the article, Cohen blurs it at the close. For it turns out that the means for discovering a philosopher’s basic thought lies in the realm of psychology after all (293). The basic thought is a “psychic process”, and the historian of philosophy has to discover it through “the method of psychology”, which will show the process through which it arose. This closing passage reveals, therefore, Cohen’s lingering loyalty to psychology and the methods of Völkerpsychologie even in the midst of his discovery of the transcendental dimension in Kant.
Cohen was true to his word when he announced his plans to write a monograph on the Trendelenburg–Fischer dispute. For that is exactly what he did in the summer of 1870, finishing it even before his article appeared in 1871. Cohen seems to have been inspired in the summer of 1870, for he made rapid progress in writing his book. Sometime in mid-July 1870 he reported: “I am getting on with my work. The whole thing is alive in my head; it is only a matter of writing it down.”54 A few weeks later, on August 2, he told his friend Herman Lewandowsky that he had already completed his exposition of the Kantian theory of space and that he was now working on the transcendental logic.55 By October 3, 1870, the book was as good as done, for Cohen told Lewandowsky that he had completed his defence of Kant, which was all he intended the book to be.56 The publication of the manuscript seems to have been delayed a full year, however, because the preface is dated October 31, 1871.
The first edition of Kants Theorie der Erfahrung is a modest book, only 270 pages long, less than half the size of its more ambitious second edition. It is very much a different book than its successor, which introduces many new themes. Still, the first edition is not rendered obsolete by the second, and it is indeed in one respect the better version, for its lesser size means that its main message emerges much more clearly. The first edition is also much more revealing about the original context and intention behind the work. For purely historical reasons, our focus now will be upon the first edition.57
The preface reveals much about Cohen’s aims and methods. We learn that his main aim is “to present the historical Kant”—“Kant as he is present in his sources” (den urkundlich vorhandenen Kant)—so that he could defend him against his critics (iv). Some of the most common objections against the critical philosophy can be easily removed, he believes, simply by the citation of the appropriate texts. As if he were a disciple of Ranke, Cohen declares that the scholar only fully understands his sources through a surrender to the material, by casting aside his own biases and entering into the world of the author (vii). To encourage that practice, and to ward off tiresome and superficial objections, it is best, Cohen advises, to proceed according to one popular notion: that Kant is a genius (vii). Cohen does not think that this is literally true, but he still believes that it is a useful regulative assumption in Kant interpretation, because it makes us think twice about our criticisms of Kant, which all too often prove superficial.
Though he stresses the need to restore the historical Kant, Cohen also insists on understanding Kant in the light of enduring systematic issues or philosophical problems. Historical reconstruction ultimately has to be guided by philosophical issues. The scholar should have an understanding of these issues, and he needs to work out his own response to them, Cohen maintains, if he is to fully understand the text before him. For what one reads into a text crucially depends on one’s stance towards these problems. Cohen says that the Fischer–Trendelenburg dispute has taught him that one’s ultimate philosophical commitments also determine one’s exegesis, and that there is no such thing as a completely neutral reading of the text. “One cannot deliver any judgement on Kant without betraying in every line the world in one’s own head.” (v)
There is a tension here in Cohen’s account of the philosophical and historical aspects of interpretation: the philosophical demands, or at least permits, prior commitments; but the historical insists upon laying such commitments aside and delving into the mental world of the author. Exactly how these philosophical and historical aspects hold together Cohen does not explain. It would be a mistake, however, to see Cohen’s emphasis on the philosophical as a covert or implicit license for a completely non-historical reading of Kant.58
If there were a single phrase to summarize the significance of Kants Theorie der Erfahrung, it would have to be Cohen’s discovery of the transcendental. The concept of the transcendental is the key to Cohen’s interpretation of Kant, and indeed the guiding thread for the later interpretation of the Marburg school. The transcendental is for Cohen co-extensive with the a priori, which consists in what he calls “the formal conditions for the possibility of experience”. It therefore comprises the formal intuitions of sensibility, space and time, as well as the pure concepts of the understanding. Cohen understands the a priori neither in terms of innate ideas—that would be too psychologistic—nor in terms of Platonic archetypes—that would be too metaphysical. Rather, 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.
True to its title, Kants Theorie der Erfahrung concerns what Cohen calls “Kant’s theory of experience”. Such a theory is a general account of the formal conditions of experience, that is, those conditions that hold regardless of the particular content of experience. Cohen maintains that there is a single central principle to explain all these formal conditions, which happens to be the main thesis behind Kant’s Copernican Revolution: namely, “that we know a priori of objects only what we put into them”.59 Kant himself had emphasized the decisive role of this principle in the prefaces to the Kritik, and even called it the guiding idea behind “our new method of thought”.60 Cohen emphasizes these passages, repeating them constantly, and making them the basis for his reconstruction of Kant’s theory of experience. All the mistakes of previous interpretations of Kant, he argues, come from failing to appreciate the decisive role of this principle.
This principle means first and foremost, Cohen explains, that the transcendental subject is the source of the a priori, that it generates or creates the formal conditions of experience. This is the basic principle behind the Copernican Revolution, because it implies, as Kant famously put it, “that concepts do not conform to objects but that objects conform to concepts.”61 In other words, we should think about knowledge not as the correspondence of representations with an object that exists independently of them, but as the conformity of representations to standards or norms imposed by the subject itself. This means that the subject determines the very standards of objectivity, that it creates, as Kant called it, “the concept of an object”, so that there is no need for an object outside our representations to which they have to conform.
It was one of the great merits of Cohen’s little tract that it went back to the basics of the critical philosophy, that it had stressed the importance of the meaning of the Copernican Revolution itself. By going back to Kant’s original starting point, Cohen thought he could see through the errors of other interpretations, not least the problem behind Trendelenburg’s interpretation of Kant. Once we recognize that the subject creates the very conditions of objectivity, we see that it is pointless to talk about the extra or “third” possibility of representations corresponding to things-in-themselves. We can now see that the thing-in-itself is not only superfluous to explain the possibility of knowledge, but also that it is completely contrary to the spirit of the Copernican Revolution, which demands that we explain objectivity entirely in immanent terms, that is, in terms of the conformity of representations with the formal conditions of experience created by us rather than their correspondence with a thing-in-itself beyond us. It follows from this, Cohen concludes, that the dualism between the subject and object now falls within the formal conditions of experience, inside the transcendental perspective itself (35–37, 52, 54). We must not conceive of the a priori as an innate subjective faculty that precedes experience, applying concepts and intuitions to given objects, as Trendelenburg did, because these objects are created and constituted by the a priori conditions of our knowing them (48–49, 63–64, 72–73).
Cohen’s conception of the a priori or transcendental is not as simple as it seems, and it is necessary to be on guard against misinterpretations. One of these conceives the a priori or transcendental entirely in normative terms, so that it consists in nothing more than rules, laws or norms, so that the subject behind them disappears. The subject seems dispensable because, on Cohen’s own showing, the dualism between subject and object falls under these norms. For these reasons Cohen’s transcendental idealism has been described as “an idealism without a subject”.62 While this interpretation is perhaps correct for the later Cohen, who thematizes “pure thought” in his 1902 Logik der reinen Erkenntnis, it is incorrect for the young Cohen, the author of Kants Theorie der Erfahrung. Although the young Cohen understands the a priori in normative terms, and although he stresses how the dualism between subject and object falls within experience, he still does not intend to eliminate the transcendental subject. On the contrary, he insists on retaining it, seeing it as the fundamental ground of all experience. Throughout Kants Theorie der Erfahrung, in both the first and second editions,63 Cohen insists on the ineliminable role of the transcendental subject, which is the ground of the formal conditions of experience. Thus he states that subjectivity is the sole ground of experience (49), that the subject knows experience a priori by constructing it (11), that the starting point of all knowledge lies within ourselves (35), and that the only objectivity is that produced or created by the subject (54). It is the empirical subject who falls within experience, not the transcendental subject who creates or posits the formal conditions of experience itself.64
Another misinterpretation reads Cohen’s early text in more metaphysical terms. It maintains that Kants Theorie der Erfahrung is the product of Cohen’s mystical Platonism. According to this interpretation, Cohen attributes vast creative powers to the transcendental subject, so that it produces not only the form but also the matter of experience, and so that the given intuitions of sensibility utterly disappear.65 Cohen so radicalizes Kant’s new method of thought that he makes the object of experience, in both its form and content, into an a priori construction, having no features beyond what we create in it. Kant’s paradigm of a priori knowledge—we know only what we create—then becomes the paradigm of all knowledge, even empirical knowledge. The source of this interpretation of Kant, we are told, is Cohen’s early interpretation of Plato in his 1866 essay ‘Die Platonische Ideenlehre psychologish entwickelt’.66 Supposedly, Cohen read Plato’s theory of ideas in terms of Fichte’s intellectual intuition, which he saw as the essence of transcendental idealism itself.
This interpretation seems plausible enough, especially when it appears to have historical evidence to support it.67 But this interpretation too is deeply anachronistic, turning Cohen’s early text into an anticipation of his later work. Starting with his 1896 Einleitung mit kritischem Nachtrag zur Geschichte des Materialismus, Cohen had indeed moved toward an interpretation of Kant that eliminates the dualism between understanding and sensibility and that completely removes the given content of experience.68 But Cohen stoutly resists this interpretation in his earlier work, which is much closer to the letter of Kantian doctrine. In his earlier work, Cohen insists that the a priori consists in the formal conditions of experience alone, and that the creative power of the subject does not extend to the matter of experience, which has to be given (100, 101). Although Cohen stresses the creative role of the subject in creating the object of its experience, he limits that role to its formal constitution; the subject creates the concept of the object, not the object itself. Cohen is indeed explicit that the subject has a passive sensibility, so that its object is given to it (163), and that it does not create the existence of the object (127–128). In classic Kantian fashion he stresses the interdependence of the form and content of experience, so that it is a mistake for the rationalists to think that matter is created and for the empiricists to assume that form is given (4). The central fallacy of rationalism is that it reduces the object of experience down to a mere construction or noumenon (210). Rather than seeing the a priori construction of experience as an intellectual intuition, Cohen stresses the role of inner sense, for which content must be given (163). Throughout his book Cohen emphasizes Kant’s opposition to an intellectual intuition, which would be an understanding that creates objects in the act of knowing them (149, 150, 159, 238). He notes that Kant attempted to undercut all need to postulate an intellectual intuition by locating the creative powers involved in mathematical construction in human sensibility rather than reason (243). Far from wanting to affirm intellectual intuition, Kant saw it as characteristic of material idealism, which he utterly repudiated (149, 150, 159, 243–244).
As one would expect, Kants Theorie der Erfahrung contains a critique of the psychological interpretation of Kant. Cohen’s concept of the a priori and transcendental dictate nothing less. The psychological interpretation had construed Kant’s concept of the a priori in terms of the psychic or physiological conditions of experience, and so it had failed to grasp Kant’s epistemological intention: explaining the conditions for our knowledge of objects. Sure enough, Cohen criticizes Herbart, Lange and Meyer for their basic misunderstanding of Kant’s epistemological enterprise. Cohen’s critique of these authors is notable, not least because he had himself followed them so closely in his early articles for the Lazarus-Steinthal journal. Indeed, shortly before his book appeared he had given a largely positive review of Meyer’s Kants Psychologie.69 There can be no doubt, however, that Cohen’s position in Kants Theorie der Erfahrung amounts to a break with some of his own previous psychology. Notably, Meyer’s work now comes in for a severe bashing (123).
That said, it is also noteworthy that Cohen did not completely turn his back on his psychological past. His old allegiances and habits lingered on, and they surface from time to time in Kants Theorie der Erfahrung. This is most evident when Cohen argues that Kant’s theory of the a priori does not exclude using the methods of psychology to discover the conditions of experience (100, 164). Fischer had argued that Kant’s a priori concepts had to be determined by a priori means, by some kind of logical reflection and deduction, because they would lose their universality and necessity if they were found in experience.70 But Cohen finds that argument fallacious on the grounds that, although the validity of a synthetic a priori principle is independent of experience, we still learn its existence from experience. Following a line of thinking worked out by Fries and Meyer, Cohen maintains that psychology is still the best means to discover a priori concepts. His early psychology resurfaces in other important forms, namely, when Cohen continues to affirm some fundamental doctrines of Herbart’s psychology, such as the primacy of mechanical explanation and the rejection of mental faculties.71
The first third of Kants Theorie der Erfahrung is devoted to an analysis of Kant’s arguments in the Transcendental Aesthetic. Such a focus is entirely what we expect, given that the Fischer–Trendelenburg controversy revolved around that portion of text, and given that the psychological interpretations had concerned Kant’s account of sense perception. It is all the more remarkable and interesting, therefore, that most of the book in fact concentrates on the Transcendental Analytic, which had been virtually neglected in the controversies of the 1860s. It was one of the great merits of Cohen’s work that it saw the central role of the Transcendental Deduction in Kant’s enterprise, and that it placed it once again in the centre of scholarly attention. This was entirely strategic, because the Deduction demonstrated so clearly Cohen’s central thesis about the a priori. Kant had stated unequivocally in the beginning of the Analytic that his guiding question was not the quid facti?—What are the origins of the concepts of the understanding?—but the quid juris?—What right or justification do we have for these concepts? Cohen thinks that the quid juris? was the decisive and leading question of the critical philosophy as a whole, which applies to the Aesthetic no less than the Analytic.
In giving such central importance to the Transcendental Deduction, Cohen had to take issue with no less a figure than Schopenhauer, who had notoriously claimed that the entire Analytic was modelled around the Aesthetic and only an afterthought to it.72 Cohen suggested the exact opposite: that Kant first formulated his central conclusions for the pure concepts of the understanding and only later applied them to the forms of space and time (109). Still, he did not insist on this point, which he considered only a historical hypothesis, and he focused instead on Schopenhauer’s misunderstandings of Kant’s project. Schopenhauer’s chief chief mistake was that he assumed, much like Trendelenburg, that the object in space is simply given to us, and that the categories are later applied to it (179–181). What he failed to see is that for an object even to be given to us, the pure concepts of the understanding must already have done their work. We cannot perceive a determinate object in space and time prior to the application of the categories, which are necessary conditions for it to be even perceived as such an object (142, 181). It is of the first importance to see that the a priori forms of intuition of the Aesthetic are really only an abstraction for the purpose of isolating the role of sensibility in transcendental enquiry; in reality understanding and sensibility are completely intermeshed, and no object even appears to sensibility without the co-operation of understanding (90–91).
It is striking that Cohen’s account of the Transcendental Deduction, unlike more modern interpretations, stresses the fundamental role of the activity of synthesis, that is, the creative role of the understanding in forming the conditions of experience (183).73 The fundamental principle behind the Deduction, he tells us repeatedly, is Kant’s new method of thought: that we know a priori of things only what we create in them (112, 127–128). Cohen stresses the leading role of this principle, advising his reader always to keep it in mind: “It is this thought that one must always keep before one’s eye. From this thought everything begins in Kant.” (112). On his account, the categories apply to experience because they are instruments of the unity of apperception, which synthesizes the manifold of sensations into comprehensible wholes. “Transcendental apperception, with its levers, the categories, supplies the transcendental affinity of appearances by which we grasp appearances.” (135). It is thus ironic to find Cohen, the opponent of psychological interpretations, emphasizing a concept that some modern scholars find too psychologistic because it allegedly rests upon upon “the imaginary subject of transcendental psychology”.74
While Cohen’s interpretation is more textually accurate, it seems that he can advance it only at the cost of inconsistency, for the concept of synthesis seems essentially psychological, involving reference to a mental activity. But in fairness to Cohen it must be said that synthesis falls within the purview of his own concept of the transcendental, which is less narrow than modern accounts. While the transcendental is indeed second-order for him, concerning knowledge of objects rather than objects themselves, he thinks that it should include more than the truth conditions for judgements. The transcendental also involves the activity of cognition itself, of which judgements are only the product. Any account of the conditions of knowledge would have to take account of this activity, its limits and powers. It never escaped Cohen, as it did some modern interpreters, that an account of what we know cannot be separated from an account of how we know. Since it takes into account the activity of cognition, Cohen’s concept of the transcendental is broader than the merely logical.
The more serious issue concerning Cohen’s account of the transcendental deduction concerns his attitude towards scepticism. Arguably, the deduction is Kant’s response to Humean scepticism, not least because Kant himself tells us about the importance of Hume for his awakening from his “dogmatic slumbers”. But Cohen does not reconstruct Kant’s answer to Hume, because he does not think of the deduction as a response to sceptical doubt. The deduction for him begins with “the fact of science”, that is, the acceptance of mathematical physics as a datum; it then explains how that fact is possible, specifying the conditions for a mathematical knowledge of nature. But it does not attempt to demonstrate that these conditions obtain, or to defend mathematical physics against sceptical doubts. In the second edition Cohen defended this interpretation by playing down the significance of Hume.75 The Scottish philosopher, we are told, had little understanding of mathematics and how it is used in natural science; and his scepticism of causality, which would have been of concern only to the pre-critical Kant, targeted an antiquated scholastic conception of little relevance to modern science. We need to ask, however, whether Cohen’s interpretation of Kant, by playing down Kant’s response to Humean scepticism, deprives it of some of its philosophical interest. On Cohen’s account, Kant, by beginning with the fact of science, simply begs the question against Hume.
When Cohen wrote his article on the Trendelenburg–Fischer controversy in 1870, it will be recalled, he did not attempt to answer the main issue behind it, that is, whether Trendelenburg had demonstrated a gap in Kant’s reasoning about the exclusive subjectivity of space and time. He promised an answer to this question in “a more comprehensive monograph”, which turned out to be Kants Theorie der Erfahrung. But what, exactly, was Cohen’s answer to this question? How, more specifically, did he attempt to resolve the controversy between Fischer and Trendelenburg? Only when we have an answer to this question will we understand Cohen’s motivations and argument in Kants Theorie der Erfahrung. Now that we have a better general idea of the contents of that work, we are in a better position to answer these questions.
The heart of Cohen’s strategy against Trendelenburg in Kants Theorie der Erfahrung is to show that his reasoning is based on a false dilemma. According to that dilemma, either we accept transcendental realism or we surrender to a complete subjectivism. Trendelenburg had argued that if the forms of space and time are not valid of things-in-themselves, we are locked inside the circle of consciousness, so that we know only our representations. But Cohen points out that there is a middle path between the horns of this dilemma. That via media is laid down by Kant’s fundamental principle: “that we know a priori of things only what we place into them.” This principle means that we, as transcendental subjects, create our own standards of objectivity, so that we must not conceive of objectivity as a thing-in-itself, that is, as an object that exists independent of consciousness and to which it must conform. Since our self-created and self-imposed standards are constitutive of our experience, we do not per impossibile have to go outside our experience to see if our representations conform to things-in-themselves. Trendelenburg’s fundamental error is that he had failed to grasp the full implications of Kant’s fundamental principle, namely, that standards of objectivity are not given but posited. He could not bring himself to renounce a transcendental realist conception of objectivity, according to which our representations have to conform to a thing-in-itself.
We might well ask ourselves, though, whether Cohen’s argument against Trendelenburg is effective as it stands. It works well enough against Trendelenburg’s stronger claim that some kind of transcendental realism is necessary for objective knowledge. Cohen shows us that Kant’s concept of the a priori undercuts such a necessity because it shows how objectivity is still possible within experience and how it is possible to have an empirical realism with a transcendental framework. However, it does not work well against Trendelenburg’s weaker claim that some kind of transcendental realism is still possibile. For it is still open for Trendelenburg to argue that empirical realism, though sufficient for objectivity, does not rule out the possibility that a priori forms still apply to things-in-themselves. Though we could explain the entire structure of objectivity in terms of a priori principles, it is still possible that this structure conforms to an objective world outside it.
Cohen could counter even Trendelenburg’s weaker claim, however, provided that he made one very bold move: namely, eliminating things-in-themselves entirely. In that case there simply would be no extra reality for a priori principles to apply to. Cohen does not hesitate to take this more drastic step. In his penultimate chapter he maintains that the thing-in-itself is simply a regulative or limiting concept of the understanding (252), and that, read in a constitutive sense, it is only an hypostasis of this concept (258). It is only when we take into account Cohen’s position on things-in-themselves, then, that his reply to Trendelenburg becomes fully effective.
Whatever the ultimate merits of Cohen’s argument, it is striking that he had charged Trendelenburg with the same basic failing as Lange: neither had understood the transcendental dimension of Kant’s discourse. While Lange conceived the a priori in a naturalistic fashion, as if it consisted in physiological and psychological laws, Trendelenburg had understood it in a subjectivist fashion, as if it consisted in innate faculties which precede experience, and which are applied to an already given object. He did not seem to realize that the a priori makes the object of experience possible in the first place. It is not as if experience is given, and then we apply a priori forms to it; on the contrary, these forms constitute experience. Once we grasp the distinctive transcendental stature of the a priori, Cohen argued, we can easily see that it is neither naturalistic nor subjectivist; rather, it is that which makes knowledge of nature and subjectivity possible.
Ultimately, Cohen took issue with Trendelenburg’s interpretation of Kant for the same reason that he had contested Lange’s. Neither of his teachers had fully responded to the challenge of materialism; both had left open a foothold for materialism to re-enter the critical philosophy. By conceiving the a priori in psychological terms, Lange had allowed it to be explained naturalistically, as if it were like any other event in nature. And by insisting upon a transcendental realism to guarantee objectivity, Trendelenburg had reinstated matter itself with all its old nasty characteristics (viz. independent existence in space).
So, finally, we can now fully appreciate the motives for Cohen’s discovery of the transcendental: by seeing the a priori in its sui generis transcendental terms, Cohen could shut the backdoors to materialism unwittingly left open by both his teachers. The full purport of Cohen’s anti-materialist animus appears explicitly at the very end of Kants Theorie der Erfahrung:
Through the discovery of the apriority of space and time, of the apodictic in sensibility, the motives of all material idealism, and of all materialism, are destroyed … To dissolve the variety of things into the difference of ideas—that is the secret of idealism. The history of human thinking reveals this secret, and therefore itself as the history of idealism. (270)
If my account of Cohen’s path towards the transcendental is correct, it leaves us with a rather rich irony. It now turns out that Cohen’s motives were ultimately metaphysical, intended to undercut the last vestiges of materialism still clinging to Lange’s and Trendelenburg’s interpretation of Kant. But the very logic of Cohen’s concept of the transcendental excludes the metaphysical no less than the psychological. After all, the transcendental deals not with things—whether material or mental—but only our knowledge of them. Thus for decidedly metaphysical motives Cohen developed a decidedly non-metaphysical interpretation of Kant.
This is not a criticism of Cohen, still less a refutation; but it should make us mistrust the anti-metaphysical façade behind his transcendental idealism, which tends to conceal his own metaphysical allegiances. It was only in his later years, I believe, that Cohen revealed, finally, fully and frankly, the deepest motives behind his allegiance to Kant. The spirit of Kant’s philosophy and Judaism, it turns out, are really one and the same.76 But what exactly that means is the subject for another occasion.
1 Hermann Cohen, Kants Theorie der Erfahrung (Berlin: Ferdinand Dümmler, 1871). This book has been reprinted as Volume 1.3 of Hermann Cohen Werke, ed. Helmut Holzhey (Hildesheim: Olms Verlag, 1987f).
2 Andrea Poma, The Critical Philosophy of Hermann Cohen (Albany, NY: State University of New York, 1997), p. 6.
3 Recently Klaus Christian Köhnke and Reinhardt Brandt have questioned Cohen’s claim to provide an exact historical interpretation of Kantian texts. See Köhnke Entstehung und Aufstieg des Neukantianismus (Frankfurt: Suhrkamp, 1986), pp. 273–275; and Brandt, ‘Hermeneutik und Seinslehre bei Hermann Cohen’, in Philosophisches Denken-Politisches Wirken: Hermann-Cohen-Kolloquium Marburg 1992, eds Reinhardt Brandt and Franz Orlik (Hildesheim: Olms, 1993), pp. 37–54. Both Köhnke and Brandt interpret Cohen’s early work as if it were the first formulation of his later thought, which was indeed much less historical in its approach to Kant’s texts. But the early work did strive for a more accurate historical interpretation of Kant, and it must not be reduced to an anticipation of Cohen’s later views. Cohen’s reputation for careful scholarship, at least in his first work, deserves full restoration.
4 This point was stressed long ago by Ernst Cassirer, ‘Hermann Cohen und die Erneuerung der kantischen Philosophie’, Kant-Studien 17 (1912), 252–273, esp. 254–255.
5 Paul Natorp, ‘Kant und die Marburger Schule’, Kant-Studien 17 (1912), 193–221.
6 See the interpretations mentioned in note 3. See also Manfred Kühn, ‘Interpreting Kant Correctly: On the Kant of the Neo-Kantians’, in Neo-Kantianism in Contemporary Philosophy, eds Rudolf Makkreel and Sebastian Luft (Bloomington, IN: Indiana University Press, 2010), pp. 113–131, esp. 115–121.
7 The Cohen family archives were destroyed in 1941 following the deportation of Cohen’s widow, Martha Cohen neé Lewandowski, to Theresienstadt. The few surviving documents are in the Hermann Cohen Archiv in Zürich under the supervision of Helmut Holzhey.
8 On Cohen’s early years as a collaborator with Steinthal, see Dieter Adelmann, ‘H. Steinthal und Hermann Cohen’, in Hermann Cohen’s Philosophy of Religion, eds Stéphane Moses and Hartwig Wiedebach (Hildesheim: Olms, 1997), pp. 1–33; Ulrich Sieg, ‘Hermann Cohen und die Völkerpsychologie’, Aschkenas XIII (2004), 461–483; and Klaus Köhnke, ‘»Unser junger Freund« – Hermann Cohen und die Völkerpsychologie’, in Hermann Cohen und die Erkenntnistheorie, eds Wolfgang Marx and Ernst Orth (Würzburg: Königshausen & Neumann, 2001), pp. 62–77.
9 Zeitschrift für Völkerpsychologie und Sprachwissenschaft (Berlin: Dümmler, 1860–1890), 20 vols.
10 For the influence of Völkerpsychologie on modern anthropology and social psychology, see Ivan Kalmar, ‘The Völkerpsychologie of Lazarus and Steinthal and the Modern Concept of Culture’, Journal of the History of Ideas 48 (1987), 671–690.
11 Moritz Lazarus, ‘Ueber den Begriff und die Möglichkeit einer Völkerpsychologie’, Deutsches Museum I (1851), 112–126, pp. 112–113.
12 Moritz Lazarus, ‘Einleitende Gedanken über Völkerpsychologie’, Zeitschrift für Völkerpsychologie und Sprachwissenschaft I (1860), 1–73, p. 35.
13 See Moritz Lazarus, ‘Einige synthetische Gedanken für Völkerpsychologie’, Zeitschrift für Völkerpsychologie und Sprachwissenschaft III (1865), 1–94, esp. 85.
14 See Johann Friedrich Herbart, Psychologie als empirische Wissenschaft §31, in Sämtliche Werke, eds Karl Kehrbach and Otto Flügel (Langensalza: Hermann Beyer & Söhne, 1887–1912), V, 253.
15 See Cohen’s review of Jürgen Bona Meyer’s Kants Psychologie, Zeitschrift für Völkerpsychologie und Sprachwissenschaft VII (1871), 320–330, pp. 324, 328–329. For a more detailed account of Cohen’s early psychology, see Winrich de Schmidt, Psychologie und Transzendentalphilosophie: Zur Psychologie Rezeption bei Hermann Cohen und Paul Natorp (Bonn: Bouvier, 1976), pp. 19–31.
16 Hermann Cohen, Schriften zur Philosophie und Zeitgeschichte, eds Ernst Cassirer and Albert Görland (Berlin: Akademie Verlag, 1923), I, 29.
17 See Cohen, ‘Mythologische Vorstellungen von Gott und Seele, psychologisch entwickelt’, Zeitschrift für Völkerpsychologie und Sprachwissenschaft V (1868), 396–434 and VI (1869), 113–131; and ‘Die dichterische Phantasie und der Mechanismus des Bewusstseins’, Zeitschrift für Völkerpsychologie und Sprachwissenschaft VI (1869), 171–263. This article also appeared independently, Die dichterische Phantasie und der Mechanismus des Bewusstseins (Berlin: Dümmler Verlag, 1869). All references to it will be to this independent publication.
18 See Cohen, Die dichterische Phantasie, pp. 27–28.
19 See Cohen, Die dichterische Phantasie, pp. 28–29; and Cohen, ‘Mythologische Vorstellungen’, pp. 406–407.
20 See Cohen, Die dichterische Phantasie, p. 53; and Cohen, ‘Mythologische Vorstellungen’, pp. 420–421. Kinkel’s claim that Cohen already goes in a Kantian direction because of this distinction seems to me to come from a failure to examine its precise meaning. See Walter Kinkel, Hermann Cohen: Eine Einführung in sein Werk (Stuttgart: Strecker und Schröder, 1924), p. 44.
21 See Cohen, Die dichterische Phantasie, pp. 29, 49.
22 This is plain from Cohen’s October 3, 1870 letter to Hermann Lewandowsky, Briefe, eds Bertha and Bruno Strauss (Berlin: Schocken, 1939), p. 28.
23 See Cohen, Review of Jürgen Bona Meyer, Kants Psychologie and Cohen, ‘Zur Kontroverse zwischen Trendelenburg und Kuno Fischer’, Zeitschrift für Völkerpsychologie und Sprachwissenschaft VII (1871), 249–271, p. 292.
24 Köhnke, Entstehung und Aufstieg, p. 282.
25 See Cohen to Hermann Lewandowsky, August 2, 1870, Briefe, pp. 24–25.
26 This point is made, perfectly correctly, by Ulrich Sieg, Aufstieg und Niedergang des Marburger Neukantianismus (Würzburg: Königshausen & Neumann, 1994), p. 110, n.225.
27 Cohen, ‘Die Platonische Ideenlehre psychologisch entwickelt’, Zeitschrift für Völkerpsychologie und Sprachwissenschaft IV (1866), 403–464.
28 Köhnke, Entstehung und Aufstieg, p. 280.
29 Poma, Critical Philosophy, p. 22.
30 Cohen explicitly cites KrV A 313–314, A 582–583.
31 Kant repudiated intellectual intuition in a constitutive sense and as a method of philosophy in his ‘Von einem neuerdings erhobenen vornehmen Ton in der Philosophie’, Schriften VIII, 387–406. He also denounced Fichte’s Wissenschaftslehre in the August 29, 1799, issue of the Allgemeine Literatur Zeitung, Schriften XII, 370–371. See also Kant’s April 5, 1798, letter to J.H. Tieftrunk, XII, 240–241.
33 Cohen, ‘Mythologische Vorstellungen von Gott und Seele’, Zeitschrift für Völkerpsychologie und Sprachwissenschaften V (1868), 396–434 and VI (1869), 113–131.
34 On Cohen’s early concern with this issue, and on his early education, see Michael Zank, The Idea of Atonement in the Philosophy of Hermann Cohen (Providence, RI: Brown Judaic Studies, 2000), pp. 48–107.
35 ‘Virchow und die Juden’, in Jüdische Schriften, ed. Bruno Strauß (Berlin: Schwetschke, 1924), II, 457–462, p. 461. First published in Die Zukunft, Demokratische Zeitung August 14, 1868.
36 See Cohen’s early essay ‘Heinrich Heine und das Judentum’, in Jüdische Schriften, II, 2–44. This article was first published in Die Gegenwart, I (1867).
37 Heinrich Heine, Zur Geschichte der Philosophie und Religion in Deutschland, in Sämtliche Schriften, ed. Klaus Briegleb (Munich: Hanser, 1976), V, 568–571. Cohen cites these passages in ‘Heinrich Heine und das Judentum’.
38 Cohen, ‘Heinrich Heine und das Judentum’, pp. 6, 9, 18.
40 See Cohen to Lange, November 16, 1871, Briefe, pp. 34–35.
41 See Cohen, ‘Friedrich Albert Lange’, Philosophische Monatshefte 12 (1876), 46–47; and Cohen, ‘Friedrich Albert Lange’, Preussische Jahrbücher 37 (1876), 353–381.
42 Lange, Geschichte des Materialismus, Sechste (wohlfeile und vollständige) Auflage (Leipzig: Baedeker, 1898).
43 See Cohen’s ‘Biographisches Vorwort des Herausgebers’, to the sixth edition of Lange’s Geschichte des Materialismus, p. xi.
44 Cohen, ‘Friedrich Albert Lange’, Preussische Jahrbücher, p. 374.
45 Cohen, ‘Friedrich Albert Lange’, p. 373.
46 Cohen, ‘Friedrich Albert Lange’, pp. 370–377.
47 Regarding the young Cohen’s relationship to Trendelenburg, see Ulrich Sieg, ‘Hermann Cohen und die Völkerpsychologie’, Aschkenas XIII (2004), 461–483, pp. 466–467. Little research has been done on Trendelenburg’s influence on Cohen. For a brief discussion, see Gerhard Lehmann, Geschichte der Philosophie: Die Philosophie des neunzehnten Jahrhunderts (Berlin: de Gruyter, 1953), VII, 75–76; and for suggestions, see Peter Schulthess’s introduction to Herman Cohen Werke (Hildesheim: Olms, 2005) V, 10* n.5, 25 *n.56, and 27* n.64. See also the article by Ernst Wolfgang Orth, ‘Trendelenburg und die Wissenschaft als Kulturfaktum’, in Hermann Cohen und die Erkentnistheorie, pp. 49–61.
48 Cohen, ‘Zur Kontroverse zwischen Trendelenburg und Kuno Fischer’, 249–271.
49 Cohen himself saw his book in these terms. In his article (p. 252) he announced his plans to write “eine umfänglichere Monographie” on the main issue raised by the dispute, an obvious reference to his later book.
50 See the letter quoted by Kinkel, Hermann Cohen, p. 46.
51 The importance of this approach for Cohen was first pointed out by Lange, Geschichte des Materialismus II, 130 n.35. Lange saw Cohen’s work as an application of “Aristoteles-Philologie” to Kant.
52 Cohen was using the term “metaphysics” in one of its peculiar Kantian senses, according to which it means a priori knowledge of the basic principles of reason. See KrV B 879.
53 See Poma, Critical Philosophy, pp. 4–5; and Ulrich Sieg, Aufstieg und Niedergang, p. 111.
54 Kinkel, Hermann Cohen, p. 46. Kinkel does not cite the source of this quotation. His source appears to have been lost.
55 Cohen to Herman Lewandowsky, August 2, 1870, Briefe, pp. 24–25.
56 Cohen to Herman Lewandowsky, August 2, 1870, Briefe, p. 28.
57 Unless otherwise noted, all references in parentheses are to the first edition: Hermann Cohen, Kants Theorie der Erfahrung (Berlin: Ferdinand Dümmler, 1871).
58 Pace Köhnke, Entstehung und Aufstieg, pp. 273–275; and Brandt, ‘Hermeneutik und Seinslehre bei Hermann Cohen’, pp. 37–54.
59 KrV B xviii.
60 KrV B xviii.
61 KrV B xvii.
62 See Manfred Brelage, ‘Transzendentalphilosophie und konkrete Subjektivität’, in Studien zur Transzendental Philosophie (Berlin: de Gruyter, 1965), p. 97; and Siegfried Marck, ‘Die Lehre vom erkennenden Subjekt in der Marburger Schule’, Logos IV, (1913), 364–366. For a fuller discussion, see Poma, Critical Philosophy, pp. 61–64.
63 There are indeed anticipations of Cohen’s later position in the second edition, where Cohen stresses the primacy of thinking over sensibility in the concept of the infinitesimal. See Cohen, Kants Theorie der Erfahrung (1885), pp. 423–424, 433–434. However, Cohen is still critical of rationalist attempts to reduce the content of sensation to the understanding. See Cohen, Kants Theorie der Erfahrung (1885), pp. 40–41. On the difference between these editions, see Poma, Critical Philosophy, pp. 37–53.
64 Though Cohen makes no clear distinction between the transcendental and empirical subjects, it is the presupposition of his argument. Only such a distinction could explain how the subject could both fall within experience and create its formal conditions.
65 See Köhnke, Entstehung und Aufstieg, pp. 273–301. Sieg, Aufstieg und Niedergang, pp. 111–112, follows Köhnke’s interpretation.
66 Köhnke, Entstehung und Aufstieg, p. 280.
67 We have already seen in Section 3 that the concept of intellectual intuition in this early essay should be read in regulative rather than constitutive terms. Hence it does not provide any real evidence for Köhnke’s reading.
68 It is a mistake to see the starting point of Cohen’s later views in his 1883 Das Princip der Infinitesimal-Methode und seine Geschichte (Berlin: Dümmler, 1883), as Sieg has done (Aufstieg und Niedergang, p. 141). Though there are indeed indications of later developments, Cohen continues to affirm the necessity of a dualism between understanding and sensibility. See, for example, p. 18, §24. On this whole question, see Helmut Holzhey’s introduction to Einleitung mit kritischem Nachtrag in Hermann Cohen Werke Band V, 7*–18*.
69 See note 15.
70 See Cohen, Kants Theorie der Erfahrung, Chapter 8, Section 4.
71 See Cohen, Kants Theorie der Erfahrung, Chapter 8, Section 4.
72 See Schopenhauer, ‘Kritik der kantischen Philosophie’, Anhang zu Die Welt als Wille und Vorstellung, Dritte Auflage, Werke (Stuttgart: Insel, 1968), I, 561–715, pp. 580–582.
73 Cohen himself would later object to the Kantian concept on the grounds that it presupposed a given manifold. See Cohen, Logik der reinen Erkenntnis, p. 24.
74 See P.F. Strawson, The Bounds of Sense (London: Methuen, 1966), pp. 32, 97.
75 Cohen, Kants Theorie der Erfahrung (1885), pp. 51–58.
76 ‘Innere Beziehungen der kantischen Philosophie zum Judentum’, in Achtundzwanzigster Bericht der Lehranstalt für die Wissenschaft des Judentums in Berlin (Berlin: Meyer und Müller, 1910), 39–61. Reprinted in Werke XV, 309–345.