This paper grew out of my work as a Humboldt fellow at the Georg-August Universität in Göttingen. I would like to thank the Humboldt Foundation for its generous support. This project has also been generously supported by the FCAR of Quebec. I am very grateful indeed to Wolfgang Carl for generously finding the time to give me extensive comments and criticisms of this paper. They were invaluable. I would also like to thank the participants of his Colloquium for the stimulating discussion of an earlier version. Finally, I would like to thank the Philosophy Department of the Georg-August Universitat and Konrad Cramer in particular for graciously allowing me the use of an office and library facilities during a subsequent sabbatical.

¹ 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 Scbriften (the Academy Edition) in 29 volumes (Berlin: Walter de Gruyter and other publishers, 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 first edition of 1781 (A) and that of the second edition of 1787 (B).

² Adickes, 18: 269–70.

³ At least, he did not yet see the need to undertake a transcendental deduction to allay the worry articulated at B122-3 of the Critique of Pure Reason.

⁴ Let me here briefly situate my work to previous studies of the Duisburg Nachlaβ. There is not much. The most important study is in Wolfgang, Carl'sDer schweigende Kant (Göttingen: Vandenhoeck & Ruprecht, 1989). I owe a great deal to this book — beginning with the optimism that sense can be made out of the Duisburg Nachlaβ and that it can shed light on how Kant was thinking at the crucial time when he was developing the ideas behind the Critique of Pure Reason. My specific debts to Carl in this article I will acknowledge along the way. What distinguishes my work from his is that I take into account the history of Kant's earlier metaphysical commitments. Thus I read the Duisburg Nachlaβ in a backward-looking way, while he reads it in a forward-looking way. My reading brings out the pull that Leibniz continued to exercise on Kant in the mid-17705. Another full-length treatment of the Duisburg Nachlaβ in recent years is that ofGoogle ScholarPaul, Guyer in Kant and the Claims of Knowledge (Cambridge: Cambridge University Press, 1987).Google Scholar This study is important for recognizing the significance of the Duisburg Nachlaβ for understanding the evolution of Kant's ideas towards the first Critique. It brought the Duisburg Nachlaβ to the attention of English-speaking readers of Kant for the first time. But the reading is highly questionable and some of its significant difficulties have already been discussed at length and persuasively by Carl. In the interest of moving things along, I think that the constructive thing to do is to refer the reader to Carl's discussion. See Der schweigende Kant, pp. 82-103.

⁵ ‘Subject’ in this context should not be construed as ‘subject of a judgement’, but rather as ‘the knowing subject’. Context makes this perfectly clear. I will come back to the significance of Kant's remarks about the self and the knowing subject in this passage in due course.

⁶ Thus compare R4674 (17: 646. 1–4) and R4679 (17: 664. 2–17).

⁷ It should be noted, however, that Kant explicitly avoids appeal to the principle of sufficient reason in the proof of Proposition One, ‘Bodies consist of monads’. (See the scholium that follows the proof.) On the other hand, there is no reason to suppose that Kant is thereby rejecting this principle. Rather, he says his concern is to make his proof as convincing as possible, not only to Wolffian metaphysicians, committed to some version of the principle one way or another, but also to followers of Euler, who had serious doubts about metaphysics. (The stated purpose of the Physical Monadology was to prove to both parties that metaphysics and geometry are not in conflict, but somehow complementary disciplines, 1: 475-6). Kant's hope is to avoid controversy at this point in his dissertation.

⁸ For a detailed discussion of this argument in Kant's Physical Monadology, its presuppositions and implications, see chapter 3 in my book, Kant's Early Metaphysics and the Origins of the Critical Philosophy (Atascadero, CA: Ridgeview Publishing, 1995)Google Scholar.

⁹ Kant's position in R4674 does not so far imply that the objects of our sensible representations are without any monadic properties whatsoever. On that point Kant is perfectly silent. By contrast, in the relevant parallel passages in the Amphiboly section of the Critique of Pure Reason he says quite clearly that our concept of a thing's ‘inner determinations’ is an empty abstraction of the understanding. See, for example, A284/B340-1.

¹⁰ Here a qualification is needed. In the Physical Monadology, Kant departs from Leibniz (or, anyway, from Leibniz, as understood by the Wolffians) in these matters in one very important respect. The Wolffians were prepared to say that all representations of external relations are confused and obscure. They even included geometry in their indictment, because it presupposes the continuity of space. If space is continuous, then any given part of it is infinitely divisible. This led the Wolffians to conclude that geometry overlooks, obscures and even pre-empts the metaphysical truth, certifiable by reason freed from the senses, that all spatially extended, material things have simple substances as their sufficient reason. Kant objected to this line of thinking pretty much from the beginning. The purpose of the Physical Monadology is to show that metaphysics and geometry somehow belong together: both equally true, clear, distinct, etc. Hence, Kant would deny that all representations of external relations are clouded and obscured by the senses. On the other hand, he does say or intimate in the first paragraph of the preface, very much in the Wolffo-Leibnizian spirit, that those who trust only to experience never gain knowledge of sufficient reasons.

¹¹ Thus consider bk II, ch. 23, of the Essay concerning Human Understanding, ed. Nidditch, (Oxford: Clarendon Press, 1990), 317Google Scholar (the last paragraph of this chapter). See too, in this regard, §3 of bk II, ch. 21.

¹² There seems to be another sense in which we know the self ‘before all predicates’ according to R3921. But it is not entirely obvious how this sense relates to that which I just spelled out. Throughout this reflection Kant treats predicates as part concepts (Teilbegriff, 17: 346. 12–15), that is, as the marks or constituent parts of general concepts such as Body, which he says we abstract from sensation somehow. We uncover these marks when we subject our general, abstract concepts to analysis. By implication, general concepts such as Body are just the sum of their marks or predicates, and judging is the act that shows this explicitly, as when we judge that ‘a body is something with extension, solidity, rest, motion, etc’ A further implication seems to be that predicates are typically prior to the concept they constitute: though we may start off in the order of discovery with a hazy concept of Body, having performed the relevant abstraction, subsequent analysis shows us that, in the order of reasons, this concept is nothing over and above its constituent predicates; were it not for these predicates, we could never have formed the concept in the first place. To say that we know the self ‘before all predicates’ presumably means-perhaps among other things-that our representation of the self is not the sum of a bunch of marks or part representations. Hence, the difference between the representation of the self and a concept such as Body is this: the latter consists entirely and exclusively in predicates, whereas no predicate can be found in the former. Perhaps for this very reason, the representation of the self is better suited than any other to give us the idea of a genuine, absolute subject. Note, however, that if this is how Kant is thinking about the matter, he seems to leave himself open to the following objection. It might be argued that the idea of the self will be shown to be a predicate, if analysis reveals that it is a part concept of some abstract, general concept. But to the extent that the idea of the self can be used as a predicate, so the objection would go, it cannot serve as the idea of an absolute subject, that is, as the idea of something that is subject only and never predicate. To block an objection such as this, not only would Kant have to show that the idea of the self cannot itself be subject to analysis and so cannot be reducible to any constituent part-concepts or predicates, he must also show that the idea of the self can never be a predicate revealed by analysis to be a part-concept of a whole, general concept. There is no evidence from R3921 how Kant would have dealt with such an objection.

¹³ He also says here that it (sc. the object) is the representation of the subject ‘but made universal’ (my emphasis). I want to postpone discussing the significance of the ‘universal’ for later, when I take up the so-called ‘functions of apperception’.

¹⁴ That I do stand in such relations is essential to his account of the so-called ‘functions of apperception’. See below.

¹⁵ Notice that he says explicitly in the Latin section of the Duisburg Nachlaβ -a rational psychology in outline - that we truly call the human soul ‘subject’ or substance (17: 672. 8-10).

¹⁶ Indeed, it is immediately relevant to the case of synthetic judgements, as treated in the Duisburg Nachlaβ, just because Kant says explicitly throughout these reflections that the subject of a synthetic judgement is always somethin g sensibly given to us (17: 644. 21-9-645. 5). As we saw earlier, he also says tha t we can represent sensibly given things only according to their relations. But that means we can never properly represent such things as subject and that we ca n do so, if at all, only derivatively, that is, by analogy with the one thing we can represent properly as a true subject, namely the self.

¹⁷ , Locke, Essay, 317.Google Scholar

¹⁸ Ibid., 234.

¹⁹ Ibid.

²⁰ For typical statements of this claim, see Leibniz's letter to Arnauld of 9 October 1687 in , Leibniz, Die philosophischen Schriften, ed. C. I., Gerhardt (Hildesheim: Olms 1978), vol. 2, p. 119Google Scholar ; his De modo dis-tinguendi phaenomena realia ab imaginariis in , Gerhardt, vol. 7, p. 322Google Scholar ; his Primae veritates in Opuscules et fragments inedits, ed. Louis, Couturat, (Hildesheim: Olms, 1988), p. 523Google Scholar ; and the beginning of Part Two of the Specimen dynamicum in , Leibniz, Die mathematischen Schriften, ed. C. I., Gerhardt (Hildesheim: Olms, vol. 6, 1962).Google Scholar This is, in fact, a claim that Leibniz makes all over the place; it does not ge t retooled until the correspondence wit h Des Bosses from 1709 to 1715. See , Gerhardt's edition of Die philosophischen Schriften, vol. 2, pp. 369–72Google Scholar.

²¹ See , Gerhardt's edition of Die philosophischen Schriften, vol. 5, pp. 202–3.Google Scholar

²² It is surprising that these themes in Leibniz's thought do not figure prominently in the Nouveaux essais, namely in the chapter on substance. They do come up in the more narrowly specific case of personal identity. See , Gerhardt'sDie philosophischen Schriften, vol. 5, pp. 229–36.Google Scholar Be that as it may, one of the classic statements of these themes can be found in §§1-19 of the so-called ‘Monadology’ in vol. 6 of , Gerhardt'sDie philosophischen Schriften, pp. 607–23Google Scholar.

²³ The argument sketched in this paragraph does not appear in the Duisburg Nachla. But if we imagine Kant as advancing it, it may help us situate the ‘original-of-all-objects’ claim in its historical and philosophical context. It is reasonable to imagine Kant as quite possibly having some such argument in mind at the time of the Duisburg Nachlaβ, not only because t i follows naturally from the account of sensibility in the Inaugural Dissertation, but also because there is an explicit statement of it in the Politz lectures on metaphysics, LI (28: 1. 209-10). These lectures could have been written at the time of the Duisburg Nachlaβ - in any case no more than five years later. For the dating of the Pölitz lectures, see Max, Heinze, Vorlesungen Kants tiber Metaphysik (Leipzig: S. Hirzel, 1894), p. 516Google Scholar.

²⁴ This is the diagnosis, at any rate, that Kant would later give of the motivation behind Leibniz's monadology. Thus see the Amphiboly section of the Critique of Pure Reason (B330).

²⁵ , Carl, Der schweigende Kant (Göttingen: Vandenhoeck & Ruprecht, 1989), pp. 91–3.Google Scholar For further evidence see the Latin interlude of R4684 - a brief rational psychology in which Kant explicitly characterizes the soul as a single, simple, immaterial substance and cites this fact to justify calling it ‘subject’ (17: 672. 8-10).

²⁶ I will show how much the two claims have in common in my notes as I discuss the new one - in the interest of speeding the exposition.

²⁷ An alternative construal would be that this is the relation of ground to consequence as exemplified among the mind's representations themselves, as when the one follows as an inference or consequence from a few others, serving as the premises or grounds of a syllogism.

²⁸ The claim, ‘The mind is the model for all real synthesis’, differs from the claim, ‘I am the original of all objects’, in just this one respect. It makes explicit reference in the passage it occurs in to the functions of apperception, the mind's original relations. There is no mention of these relations in the passage in which the original-of-all-objects claim appears.

²⁹ Der schweigende Kant, p. 91. Carl's favourite proof text is in R4676 at 17: 656. 2-6. See also R4677 at 17: 659. 18-20 and at 17: 658. 5-12.

³⁰ That Kant takes such assumptions to be at the basis of ordinary empirical judgements such as this is plain from his discussion of his own examples in R4679. See 17: 664. 2-17.

³¹ We must also wonder whether a mere three apperceptive relations would be enough to account for all empirical concepts.

³² See too R4675: 17: 648. 23-7.

³³ See, for example, R4675: 17: 23–7, a passage I will discuss later.

³⁴ I discuss the significance of these ideas in the larger context of the history of Kant's intellectual development in ‘Kant on Sensibility and the Understanding in the 1770s’, Canadian Journal of Philosophy, vol. 33, 4 (December 2003), 443–82CrossRefGoogle Scholar.

³⁵ See too R4677: 17: 657. 25-8; R4681: 17: 668. 12-13.

³⁶ This point follows from everything we have seen up to now. But, as a proof text, consider the following from R4677: ‘Appearance is made objective by the following means: when it is brought under as contained under a title of self-perception (sc. under a function of apperception); and, thus the original relations of apprehension (sc. the functions of apperception) are the conditions (stricken: not only) of the perception of the (added: real) relations in appearance; and, precisely to the extent that we can say, “appearance belongs thereunder”, is appearance determined from the universal and represented objectively, i.e., thought’ (17: 658. 4–9). See too the passage from R4674 in which Kant says that ‘I am the original of all objects’ (17: 646. 7-14).

³⁷ Note that this association between the universal and thinking objectively, that is, relating empirical concepts to an object, was already in place in the surrounding context of the claim I examined earlier, ‘I am the original of all objects.’ Thus Kant writes: ‘in exposition, conditions can be recognized a priori in the subject under which A (sc. a concept) relates to an object, namely something real’ (17: 646. 7-9). This makes clear that the problem addressed by the original-of-all-objects claim is precisely to explain how empirical concepts can relate to objects. The solution depends on these ‘conditions’ in the subject (presumably the functions of apperception) which are not further discussed in this passage except in so far as Kant adds the following now familiar claim: ‘This object can be represented only according to its representations and is nothing other than the subjective representation (of the subject) itself, but made universal, for I am the original of all objects’ (17: 646. 10-13 - emphasis is mine). Making the representation of the subject universal is precisely what is required to make our concepts relate to objects.

³⁸ See R4674: ‘X is thus the determinable (object), which I think through A, and B is its determination (added: or the way to determine it) … With respect to an inhering representation or thought as such (sc. thought of an object whatever it may be?), X (NB: context suggests that ‘B’ is probably what Kant intended to write) is the function of thought as such in the subject. There the (added: real) concept A is together with everything determined: 1. through the subject, 2. with respect to succession through the ground, 3. with respect to coexistence through composition’ (17: 645. 31-646. 4).

³⁹ My emphasis.

⁴⁰ Note that Kant himself seems to back off from identifying the functions of apperception directly with the titles of the understanding in one passage in R4675 (17: 648. 21-7). He says that the ‘three relations in the mind’ have to be ‘converted into concepts of the understanding’. For a fuller discussion of the significance of this claim, see , Carl, Her schweigende Kant, pp. 94–9Google Scholar.

⁴¹ See my ‘Kant in reply to Lambert on the disputed ancestry of general metaphysical concepts’, Kantian Review, 5 (December 2001).

⁴² The same implication can be found in Descartes. Along the way to the conclusion in the Second Meditation that the mind is better known than the body, the meditator observes that his most fundamental, universal ideas - for example, substance, unity and number - are applicable to all things, mental and bodily alike, and that they arise not from his thought of bodily things (as he might once have imagined), but rathe r from the thought of himself as a thinking thing. Had Descartes been interested in setting up a general metaphysics, he would have used this idea as his point of departure and proceeded once more to meditate on his nature.

⁴³ But Kant would certainly deny that sensible things are the only things we can think about. We can also think about intelligible things - in the first instance the mind or soul as it is in itself. See in particular the Latin rational psychology in R4684 (17: 672–3). Such thought depends once again on the fundamental concepts, which are themselves nothing more than a discursive representation of the soul's self-reflection. There are passages in LI, the Politz metaphysics lectures from some time between 1775 and 1779, that make the same points just as explicitly. Our transcript reports Kant as saying, for example, that ‘when we speak of the soul a priori, we will say nothing more about it than we can derive from the concept of the I and in so far as we can apply the transcendental concepts to the I’ (28: 1. 266. 1–4; cf. 28: 1. 269. 35-8). The ‘transcendental concepts’ mentioned here are nothing other than concepts of general metaphysics. A couple of pages earlier, the transcript reports Kant as laying out his programme for rational psychology as a whole. He had just announced that his programme would have three parts (28: 1. 263. 17–264. 7). The first part would be an investigation of the soul ‘absolutely in and for itself’; it would proceed ‘from merely pure concepts of reason alone’ (28: 1. 263. 18–19). Kant put it this way a few lines later: ‘We consider the soul in the first part (sc. of the rational psychological programme) absolutely and hence from transcendental concepts of ontology’ (28: 1. 264. 9-11). ‘Ontology’ is just another word for general metaphysics. This clearly indicates that, at the time of the Pölitz metaphysics between 1775 and 1779, Kant continued to think, as he apparently had in the Duisburg Nachlaβ, that general metaphysical concepts apply a priori to the soul. His general metaphysics was richer now than it had been. In the Duisburg Nachlaβ, he had only Substance, Ground and Consequence and Composite Whole. But, as Max Heinze reports, the list of ontological concepts in the Politz metaphysics had expanded to include all the categories from the table that would later appear in the first Critique - with some hesitation about the status of the modal concepts (28:1.186.11-26). But alas, the section on ontology from LI, the transcript of the Pölitz metaphysics, has been lost. All that is now known about it comes from Max Heinze's report in his important study, Vorlesungen Kants über Metaphysik (Leipzig: S. Hirzel, 1894).Google Scholar Heinze's report on the ontology of LI is reproduced in the first part of vol. 28 in the Academy Edition, pp. 185-91. If Kant really hoped to show that the fundamental concepts are necessary for thought of anything thinkable and therefore sufficiently universal to earn a place in general metaphysics, he would have to give some argument that these concepts are necessary for thinking about other intelligible things as well, namely God and the simple substances underlying all matter. Such an argument is not to be found in the Duisburg Nachlaβ, yet the resources are there for producing one. For Kant could always make thought of these other intelligible things reducible one way or another to thought of the thinking self.