This paper grew out of my work as a Humboldt fellow in the philosophy department at the Georg-August Universität in Göttingen. I am very grateful to the Humboldt Foundation for its generous support. I would also like to express my gratitude to Wolfgang Carl who was generous with his time and unstinting with his comments and criticisms. This paper benefited no end from his feedback. It also benefited from a lively discussion with members of his Colloquium. Finally I would like to express my gratitude to Emily Carson, Predrag Cicovacki, Julian Wuerth and anonymous referees of the Review for their valuable comments and criticisms.

¹ See Beck, , ‘Lambert and Hume in Kant's Development from 1769 to 1772’, in Essays on Kant and Hume (New Haven, CT: Yale University Press, 1978), pp. 101–11.Google Scholar

² All translations from the German, Latin or French of the authors quoted are my own. All citations from Kant's works refer to Kants gesammelte Schriften (the Academy Edition) in 29 vols (Berlin: Walter de Gruyter, 1902–)Google Scholar. I give the volume, page and line number. Following custom, citations from the Critique of Pure Reason refer to the pagination of the 1st edition of 1781 (A) and that of the 2nd edition of 1787 (B).

³ Kant would make this point repeatedly throughout the 1770s, both in his lectures on metaphysics and in his lectures on anthropology. See, for example, 28:1.175. 9–10; 25:1.243. 3–11; 25:1.473.10–14.

⁴ All references to Wolff, are to the Gesammelte Werke, ed. Jean, Ecole et al. (Hildesheim: Olms, 1965–xx).Google Scholar References are by work title and section number.

⁵ All references to Lambert's, works are to the Olms reprint of the Neues Organon (Leipzig: Johann Wendler, 1764)Google Scholar and to the Anlage zur Architectonik oder Theorie des Einfachen und des Ersten (Riga: Johann Friedrich Hartknoch, 1771). The references are, in the case of the latter, to work (with title abbreviation, Archi.), section and page number; or, in the case of the former work, to major division (title thereof abbreviated), section number and page number.

⁶ In a letter to Kant from Berlin, dated 9 July 1771, Marcus Herz writes: ‘I have only just begun to read Lambert's Architectonik and hence cannot judge it. I have only a little time to devote myself to studies outside of medicine’ (10:127.7–8). The remark suggests that Kant was pressing Herz to read the Architectonik and to let him know what he (Herz) thought of it. The year 1771 was when Kanter's associate in Riga, Johann Friedrich Hartknoch, published the work and thus made it available to the learned public. Was the object of Kant's curiosity Herz's opinion of the Architectonik, given that Herz was one of Kant's most distinguished students at this time? Or was the object of Kant's curiosity the Architectonik itself — assuming that he had not yet been able to get a hold of a copy? Unfortunately, it seems impossible to tell.

⁷ Laywine, , Kant's Early Metaphysics and the Origins of the Critical Philosophy, North American Kant Society Studies in Philosophy, 3 (Atascadero, CA: Ridgeview Publishing, 1995).Google Scholar

⁸ This has been stressed by Michael Friedman. See the Introduction to his Kant and the Exact Sciences (Cambridge, MA: Harvard University Press, 1992).

⁹ See Laywine, , Kant's Early Metaphysics, chapters 2 and 3.Google Scholar

¹⁰ See, for example, § 30 (in light of the remarks about the intellectual and the corporeal worlds in the preceding § 29) in the Architectonik, 1.1. See too the Organon in the Alethiologie, 1.48.

¹¹ In Kant's Early Metaphysics and the Origins of the Critical Philosophy, I argued that Kant became aware of, and concerned about, metaphysical subreption in his own philosophical writings sometime in the mid-1760s and that Dreams of a Spirit-Seer of 1766 was both a satire on philosophers (in particular on Kant himself) who made themselves susceptible to this mistake and a call to investigate the source of our knowledge of forces so as to prevent the mistake from doing further harm. It is interesting to bear in mind that Lambert was one of the people to whom Kant sent a copy of this work. Quite apart from the fact that Kant highly valued Lambert's judgement, had agreed in 1765 to collaborate with him on projects devoted to the reform of metaphysics and might therefore be expected to send Lambert a copy of anything he had written on this topic, one has to wonder whether he sent Lambert a copy of the Dreams, in part because he thought that Lambert himself was especially prone to the sorts of mistakes exposed in this work. Unfortunately, the evidence we have is suggestive, without being conclusive.

¹² That the early Kant assumed that the notions of cause, ground and force were all one and the same and that the notions of effect, consequence and acceleration were likewise all one and the same is plain throughout his early writings. For another example of this, see his Attempt to Introduce Negative Magnitudes into Philosophy of 1764 (2:203.1–6). Of interest too is the following passage from the lectures on metaphysics: ‘In the concept of force lies that of cause. The substance is regarded as subject, and this in turn as cause. Accidens is something real for this reason, namely because it inheres [sc. in something else, i.e., the subject] and does not exist for itself. Causality is the determining of something else, whereby it is posited according to universal rules. The concept of respectus or of the relation of the substance to the existence of the accidents, insofar as it contains the ground of the same, is force. All forces are divided into primitive or fundamental forces and into derivative or secondary forces. We are looking to reduce vires derivativae to primitive forces. All physics, as much that of bodies as that of minds [Geister], the latter of which is called psychology, comes to this: as far as possible to direct the different forces, which we learn about only through observation, back to fundamental forces’ (28:2, 1.564.7–19). This passage is from Metaphysics L2, which is thought to be relatively late — certainly post-dating the appearance of the second edition of the first Critique. Max Heinze argues that the lecture itself was probably held in the winter of 1790–1 (Vorlesungen Kants über Metaphysik aus drei Semestern, Nummer VI des XIV Bandes der Abhandlungen der philologisch-historischen Classe der Königl. Sächsischen Gesellschaft der Wissenschaften (Leipzig: bei S. Hirzel 1894), p. 504). If Heinze's dating is right, it seems improbable that Kant was expounding his own, current view of causality. But the passage does succinctly sum up the view that Kant had held up to 1770.

¹³ For the details on what follows, see Laywine, , Kant's Early Metaphysics, chapter 5.Google Scholar

¹⁴ Here one should compare Kant's letter to Mendelssohn of 8 April 1766 (10:71–2).Google Scholar I will return to this argument strategy, its weakness and philosophical significance for Kant's later development in the fifth section of this paper.

¹⁵ This replay in the Inaugural Dissertation of the lesson of Dreams of a Spirit-Seer is the source of a peculiar inner tension in the former work. For, in § 8 of the Inaugural Dissertation, Kant reckons the concept of cause, in spite of everything, among the pure concepts of general metaphysics, which issue not from experience, but from pure intellect. I shall have to return to this point later.

¹⁶ It should be noted that Kant refers once again to general metaphysics in § 8 of the Inaugural Dissertation. He says that metaphysics (or perhaps more precisely, general metaphysics itself) is that part of philosophy ‘containing the first principles of the use of pure intellect’ (2: 395. 16–17). He says, moreover, that this part of philosophy contains certain pure intellectual concepts, including that of possibility, existence, necessity, substance and cause. Again, it is not clear how his willingness here to reckon the concept of cause among the general metaphysical concepts, proceeding from pure intellect, is supposed to square with his claim in § 28 that the concept of force must be derived from experience (2:416.33–417.3).

¹⁷ The emphasis is mine.

¹⁸ The emphasis is mine.

¹⁹ Lambert himself is more direct about this in the Architectonik than in the Alethiologie section of the Organon. Thus compare the interesting discussion of Locke at the beginning of the Architectonik (§§ 7–10 of the first Hauptstück) with § 29 at the beginning of the Alethiologie. Be that as it may, Lambert's debt to Locke is perfectly clear in the development of the argument. Indeed, many sections quite obviously play back discussions from various parts of Locke's Essay, sometimes with elaborations by Lambert. Locke and Lambert agree on the basic significance of the simple concepts: they are the materials out of which all knowledge is made up. Compare, for example, Locke's, Essay, book 2, chapter 7, § 10Google Scholar with Lambert's, Organon, in the Dianologie, Hauptstück IX, § 653.Google Scholar Of course, Lambert is just as clear that, for his own purposes, he thinks he must go beyond Locke. Locke succeeded in producing an ‘anatomy’ of human understanding, that is, he methodically sought out and enumerated the different classes of concepts. But Lambert wants to understand how properly scientific knowledge can issue from the simple concepts, and hence how the simple concepts may be combined to form a systematic whole. Locke himself did not address this problem.

²⁰ All references to Locke's Essay are to book number, chapter number and, where applicable, section number in the Nidditch edition (Oxford: Clarendon Press, 1975).

²¹ To the same feeling, Lambert traces our simple concept of moving forces and hence the ‘first principles of dynamics or the doctrine of forces' (Aleth. 1.18). Lambert's reflections on Newton and Sanderson (1. 53 ff.) seem to be a riff on Locke 2.2.2.

²² Thus compare book 2, chapter 5 of Locke's Essay with the Alethiologie section of Lambert's, Organon (1.20).Google Scholar

²³ Compare Essay (2.14. 3) and Organon (Aleth. 1.23).

²⁴ Compare Essay (22. 7. 7) and Organon (Aleth. 1.26).

²⁵ See Organon (Aleth. 1. 24). See Essay (2.7.7).

²⁶ I should now point out that Lambert's position is actually more complicated than I have hitherto represented it. For Lambert apparently wants to say that, though we acquire the simple concepts either by sensation or reflection, as Locke had already shown, they are nevertheless a priori, considered from a certain point of view — indeed, a priori in what Lambert calls the ‘strictest sense’. An idea is a priori in the strictest sense if ‘we in no way owe [it] to experience’ (Dia. 9. 639).

Lambert thinks we owe our simple concepts to experience (and therefore that they are a posteriori in the strictest sense) only inasmuch as they are enlivened by experience. Were it not for experience, he says, they would never become ‘lively’ (lebhaft) enough to intrude upon our thoughts, and hence we would never take notice of them. But once they have been acquired thus, the mind — as Locke himself points out — can do nothing to destroy or eliminate them. Just to that extent, Lambert will conclude oddly enough that they are also a priori in the strictest sense. Lambert arrives at this conclusion — so fundamentally at odds, it would seem, with his first premiss — by expressing Locke's insight in formal, logical terms.

On Locke's view, the simple concepts are just like the material atoms out of which all bodies are constituted: no creaturely power can either create or destroy the atoms; no finite understanding can create the simple concepts — or destroy them, once they are in its possession (2.2.2). Lambert himself apparently assumes that the mind can destroy a concept already in its possession, only by exposing a contradiction in it. But it can never expose a contradiction in any of its simple concepts, precisely because the simple concepts are simple. A contradiction can obtain just in case we have two concepts, that is, two formally inconsistent concepts. Precisely because no contradiction can ever arise in any given simple concept, Lambert infers that the very possibility of such a concept, once it is securely in our possession, depends only on our being able to entertain or grasp it in our mind. Lambert himself puts it as follows: ‘Accordingly, the mere representation constitutes the possibility of a simple concept, and the latter imposes itself on us together with the representation’ (Dia. 9.654). Assuming that the simple concepts are indeed already in our possession, Lambert then draws the obviously false conclusion that their possibility in no way depends on experience and hence that they are indeed a priori in the strictest sense.

The conclusion is obviously false, I say, because we are free to suppose (from Lambert's own premisses) that, without further assistance from experience, our simple concepts may perhaps gradually cease to be lively and thus fade from our awareness, so that we ultimately lose possession of them. Moreover, Lambert's reasoning apparently leads to the conclusion that sensations of colour, sound, smell, etc. can never be destroyed in the mind or that the mind will never lose possession of them. For, again, following Locke, Lambert thinks that they too are simple concepts (though not the sort useful to science — Aleth. 1.28), and hence that they too are free from contradiction, a priori in the strictest sense and thus intellectually indestructible. But from the fact that a concept or idea is free of inconsistency, it does not follow that it will remain in the mind's possession.

One has to wonder, in light of all this, what could be Lambert's motivations for saying these things. It seems that either he wants to split the difference between Locke and Leibniz on the question whether we have any innate ideas or he wants to avoid having to take sides. Thus he claims that ‘it cannot be entirely excluded that the [simple] concepts themselves should be able, in and of themselves [given their freedom from inconsistency], to be present in the soul, before they are brought to the soul's attention by sensation’ (Aleth. 1.6).

²⁷ See Cicovacki, , ‘Kant's Letter to Markus Herz’, Kant-Studien, 82 (1991), 349–60.Google Scholar Cicovacki's point is intended to mediate a debate between Wolfgang Carl and Lewis White Beck. It is quite true, as Carl argues, that the Herz-letter problem concerns the agreement of intellectual concepts and things accessible by experience. (See Carl, , ‘Kant's First Draft of the Deduction of the Categories’, in Förster, E. (ed.) Kant's Transcendental Deductions (Stanford, CA: Stanford University Press, 1989), pp. 3–20.)Google Scholar But it is not the case, as he also argues, that the Herz letter concerns only this agreement. It likewise concerns the agreement of intellectual concepts and things in themselves, as Beck rightly argues, though not exclusively so — as Beck wrongly argues. The problem is much more general, as Cicovacki argues: to give an account of the intellect and to understand how its concepts relate a priori to any object whatsoever.

²⁸ Cicovacki says the table of categories was supposed to explain how the categories can relate a priori to any object. I argue the other way around: Kant believes that the problem to solve is that of drawing up the table of categories and that explaining the possibility of the agreement between the categories and their objects will help solve that problem. I discuss the historical and philosophical implications of my reading in the conclusion of this paper.

²⁹ As spelled out at the beginning of the third section of this paper.

³⁰ Hume, , Enquiries, ed. Nidditch, P. H. (Oxford: Clarendon Press, 1975), 72.Google Scholar

³¹ Ibid., 73.

³² Newton, , Principia, trans. Motte, and Cajori, , vol. 2 (Berkeley: University of California Press, 1962), 547.Google Scholar

³³ Ibid., pp. xxv–xxvii.

³⁴ I doubt very much that these Reflexionen, or most others, can be dated so precisely. In many cases, it is doubtful whether Reflexionen can be dated at all — they are thought fragments without any surrounding context, and one has to suppose that the forensic evidence itself is frequently inconclusive. Adickes himself has to admit all too often that the dating of a given Reflexion is up for grabs. I am inclined to believe that the Reflexionen I propose to discuss here are from the end of the 1760s or the beginning of the 1770s partly for purely Benno-Erdmannesque reasons: Kant can be seen in these Reflexionen to address a problem that he explicitly addressed in published works from this time, notably Dreams of a Spirit-Seer and the Inaugural Dissertation. As the bulk of the Nachlaβ attests, it was Kant's habit to try to clarify his ideas on paper. Thus one would expect him to have left behind a number of Reflexionen from this time in which he tried to sort out the nature and status of the concept of ground and consequence. This expectation alone inclines me to date the ‘kappa-phase’ Reflexionen of interest to me here sometime, say, between 1766 and 1770. But, again, an expectation is not of itself conclusive evidence. Of more importance to me than the dating of these Reflexionen is just the fact that Kant can be seen here to experiment with the idea that the concept of ground and consequence is somehow radically subjective. That that idea crossed his mind is, I believe, important ultimately for understanding why he raises the famous question in the letter to Herz.

³⁵ See also R3971, which reads as follows: ‘True is that which holds for every cognition, and thus not merely under the condition of reason, which judges mediately. Accordingly, that is true whereof the contrary contradicts itself, because the contrary cannot even once present itself to the senses. But that something be without a ground indicates a merely subjective condition and the object could — even without such a condition — present itself to the senses.'

³⁶ That Kant might well have been tempted to treat the concept of ground and consequence as something like a principium convenientiae is suggested by a remark he makes in a Reflexion from the set we have been considering. In the first paragraph of this Reflexion, Kant has been presenting the familiar reasons for thinking that the concept of ground and consequence is subjective. Then he goes on to say in the second paragraph of the Reflexion, in rather more general terms: ‘If I say, then, that a principle is subjective, i.e., that it contains conditiones under which alone we can judge by means of our reason in accordance with empirical laws, this does not mean that our reason must accept this law [sc. the subjective principle?] as pertaining to objects; for it does not apply to them …’ (R3954).

³⁷ Consider, in light of this, R3954 from the set of Reflexionen we were considering earlier. Here too Kant explicitly concludes, from the subjective nature of the concept of ground and consequence, that the concept leads to or is the occasion of error — the erroneous supposition that the concept relates to objects. The role of imagination is not explicitly mentioned. The language suggests that Kant may be thinking of the concept of ground and consequence here as some kind of principium convenientiae.

³⁸ It might be thought that Kant is assuming a certain distinction, namely that between the concept of ground proper and the concept of a force. Thus the passage in which he insists that the relevant concept be drawn from experience seems to turn on forces in the first instance, while the passage in which he claims that the relevant concept is part of metaphysics and has its source in the intellect turns on the relation of ground and consequence. Hence, it might be thought that the tension, such as it is, dissolves, or is attenuated by, this distinction: Kant is not talking about precisely the same thing in the two passages. I am not satisfied, however, that such a move will help in the final analysis. For even in the passage directly concerned with forces and the source of our concept of the same, Kant says explicitly that the relation between a force and the acceleration it produces is to be conceived as that between a ground and its consequence. This remark is in line with his inclination in the early writings to assimilate grounds and forces. If our concept of a force is the concept of a ground, Kant is saying at 2: 379.11–22 that our concept of ground must be derived — at least some of time — from experience. That it would make sense for him to say in the context of this passage that the concept of ground must always be derived from experience is plain, because the very considerations that move Kant to say this about the concept of force apply just as readily to the concept of ground, as he conceives it at this time.

³⁹ Thus, on the one hand, Herz writes as follows, explicitly and selfconsciously arguing in Hume's vein: ‘But if I concede, once the concept of a ground (eines Grundes) has been given that the principle, “nothing is without a sufficient reason (ohne zureichenden Grund),” is an eternal truth, then I must grant nevertheless, as soon as I consider the way in which we arrive at this concept [that of a ground] that experience seems to me to be the only path which leads us to it [that of a ground]. It will not be unknown to you which philosopher I owe this thought to, although I am perfectly secured against the dangerous consequences for our knowledge, which this sharp-thinking writer derives therefrom, by the far more philosophical reasons (Gründe) advanced by a Mendelssohn. This much seems certain: the manifold observation of two occurrences, the one following the other, is the only thing that gives us occasion, according to the rules of perception, to expect these two occurrences as always coming together; and, it is the only thing that gives us occasion to name the one occurrence, which comes first in time, ground, and the other, which comes later in time, consequence. The concept of time, which has smuggled itself into these two concepts [those of ground and consequence], and which is peculiar to these two concepts, as to all empirical cognitions, has so intimately united itself with them in our representation that we are, however, incapable of thinking of ground and consequence without it, even when it comes to pure intellectual cognitions, where the condition of space and time does not at all prevail.’ See Betrachtungen aus der spekulativen Weltweisheit (Königsberg: Johann Jakob Kanter, 1771), pp. 140–1. By the same token, Herz does not hesitate (any more than Kant himself does) to appeal to a seemingly pure intellectual notion of reciprocal influence in his account of the form of the intellectual world. To be sure, he is careful to distinguish this notion of reciprocal influence from that of ground and consequence. But his reason for insisting on the distinction is not at all that the latter concept is ultimately empirical in origin (the relevant consideration is that reciprocal influence is, by definition, reciprocal, and not unilateral as for causal relations). In any case, one might have supposed that if Kant and Herz were willing to concede that the concept of ground and consequence is empirical in origin that they should also have conceded that the concept of reciprocal influence is too — and for the same reason: the intellect does not discern a necessary connection between any two things, which it otherwise conceives of as reciprocally influencing each other. See Herz, , Betrachtungen, pp. 24–5Google Scholar and 76–7.

⁴⁰ This much is clear, in any case, from the correspondence between them.

⁴¹ Naturally it will turn out by 1781 that the concept of ground and conse quence and all the other categories are indeed subjective in some important sense. For one way of putting the question of the transcendental deduction of the categories is precisely to ask, as Kant puts it in the Critique of Pure Reason, ‘how subjective conditions of thought are supposed to have objective validity’ (A89/B122). Arriving at the statement of this question will have presupposed, however, that Kant had thought through more deeply the sense in which he was prepared to say of the categories that they are ‘subjective’ — in particular, it will have presupposed that he had come to see them as nothing more than pure forms of thought or logical functions of judgement and therefore as not necessarily relating to any object whatsoever. I believe that Kant came to this insight some time later. I will have to discuss this development in due course.

⁴² Notice too that the scope of the question is as general as possible. The question is how the concepts at issue can relate a priori to any object at all. That the scope of the question would have to be so general should be clear, once again, in light of the history of Kant's treatment of the concept of ground and consequence. As Kant thinks through the implications of the Humean arguments he had deployed selectively in 1766, he has to concede that the concept of ground and consequence is peculiarly subjective. Thus whereas he raised a localized doubt in 1766 about whether the concept of ground and consequence can be used reasonably and legitimately with respect to the intelligibles, he had to see subsequently that the same doubt must be raised about whether the concept can be used thus with respect to the sensibles as well. As Cicovacki argues, it is not that the nature of the objects (whether sensible or intelligible) raises any special problem; the source of the problem lies entirely in the concept itself, and whether it is indeed peculiarly subjective, as the implications of the Humean argument suggest, and hence more reasonably understood to be a creature of the imagination than the lawful offspring of pure understanding. Having himself finally recognized the concept of ground and consequence as a product of the understanding, Kant's concerns about this concept had to extend to all the others. Hence, the question in its most general form is how can any pure concept of the understanding relate a priori to any object at all.

⁴³ It should be pointed out that Kant's account in the Critique of Practical Reason of what the transcendental deduction is supposed to accomplish resonates sympathetically with what we have just seen. Thus after discussing in the former work what he takes to be the philosophical significance of Hume's ideas about causality and how he came to deal with the implications of these ideas, Kant writes as follows: ‘And so it transpired in fact, so that I was able to prove the concept of cause not only in accordance with its objective reality with respect to objects of experience, but also to deduce it as a priori concept on account of the necessity of the connection, which it carries with it, i.e., I was able to establish its possibility from pure understanding without empirical sources’ (5:53.27–32). What did Kant accomplish? He showed, to be sure, that the concept of cause relates a priori to objects, though only to objects of experience. How is that significant? He was able to establish what Hume left very much in doubt, namely the claim of the concept of cause to issue not from any empirical source, in particular, not from the imagination, but straight from pure understanding itself. Stephen Engstrom has recently emphasized that justifying this claim was indeed the central claim of the transcendental deduction in the Critique of Pure Reason. See his paper, ‘The Transcendental Deduction and Scepticism’, Journal of the History of Philosophy, 32 (1994), 359–80.

⁴⁴ The less anachronistic, more sophisticated thing to say would be, as Cicovacki says, to think that Kant undertook to draw up a list of intellectual concepts, at roughly the time he wrote to Herz in 1772, precisely as a way to answer the centre-piece question, how can intellectual concepts relate to an object a priori. Still, the effect of reading the 1772 letter to Herz is to make the project of drawing up a list of intellectual concepts a means to answering the centre-piece question.

⁴⁵ The Leibniz passages are quoted from Leibniz, , Die philosophischen Schriften, ed. Gerhardt, (Hildesheim: Olms, 1978)Google Scholar, with references to volume and page number.

⁴⁶ Here I am suggesting Leibniz as a conjectural paradigm for Kant's subsequent attempts to answer what I have been calling the Herz-question. Locke may also have been a paradigm in a different way. This is something Predrag Cicovacki has emphasized to me. See his ‘Locke on mathematical knowledge’, Journal of the History of Philosophy, 28/4 (October 1990), 511–24.