CHAPTER ONE
VARIETIES OF SKEPTICISM
1. Kant’s preoccupation with skepticism is hardly news, being a staple component of almost all recent Anglophone interpretations of Kant for example. But this further point is much more controversial, contradicting most of those interpretations for example.
2. Again, Kant’s central preoccupation with reforming metaphysics is hardly news in itself; it has been emphasized especially in the German literature (notably by H. Heimsoeth and M. Heidegger), but also by others (e.g., H.J. de Vleeschauwer and D. P. Dryer). However, the further point made here is more controversial (sharply contradicting the anti-epistemological readings of Heimsoeth and Heidegger, for example).
3. It will seem less heretical—though not, I hope, merely orthodox—to scholars in the German tradition. For example, my devaluation of “veil of perception” skepticism agrees with H. Heimsoeth’s reading, and my accentuation of Pyrrhonism coheres to a significant extent with positions argued for by B. Erdmann, G. Tonelli, and N. Hinske.
4. Certain aspects of this part of my account will be at least roughly familiar from existing secondary literature. Others will be altogether unfamiliar. One example of the latter: an anti-skeptical motive and function which I shall be ascribing to the systematic, or “architectonic,” character of Kant’s reformed metaphysics.
CHAPTER TWO
“VEIL OF PERCEPTION” SKEPTICISM
1. Two recent examples—both in many ways excellent philosophers, incidentally—are Paul Guyer and Barry Stroud. For instance, Guyer claims in this vein that “Kant clearly conceived of the problem of knowledge in terms of methodological solipsism. That is, like thinkers from Descartes to Hume, he supposed that an answer to skepticism must lie in what reflection on his own thoughts can reveal to the individual thinker even on the supposition that nothing but his own consciousness exists” (P. Guyer, review of D. Henrich, Identität und Objektivität, in Journal of Philosophy, no. 76 [1979]; cf. his Kant and the Claims of Taste [Cambridge, Mass.: Harvard University Press, 1979], pp. 288–91; Kant and the Claims of Knowledge [Cambridge: Cambridge University Press, 1987], pp. 75, 207–8, 252–6, 279; such an assumption leads almost inevitably to Guyer’s position in Kant and the Claims of Knowledge that the Refutation of Idealism constitutes the real heart of the Critique). For Stroud’s similar picture, see B. Stroud, The Significance of Philosophical Scepticism (Oxford: Clarendon Press, 1984), ch. 4; and “Kant and Skepticism,” in The Skeptical Tradition, ed. M. F. Burnyeat (Berkeley: University of California Press, 1983).
2. H. Heimsoeth seems to me essentially correct when he argues concerning Kant that “the grounds for the formation of the concept of the thing in itself and the limitation of our objective cognition to appearances which were earliest and then remained constantly in effect are not ‘epistemological’ in an introspective-reflexive, phenomenological, or indeed psychological . . . sense of the word . . . In contrast to later idealistic lines of thought which begin from consciousness and consider that they thence prove the assumption of a thing in itself transcendent of consciousness to be a ‘non-thought,’ Kant sees no difficulty in the transition from being into consciousness, and particularly in the impact of existing substances on existing intelligences (which thereby have direct experience of properties of those external substances). For him the riddle lies solely in the synthetic a priori aspects of our cognition of reality” (“Metaphysische Motive in der Ausbildung des kritischen Idealismus,” Kant Studien, vol. 29 [1924], pp. 121–2).
3. It is of some symptomatic significance for my overall thesis in this essay—and also important for interpreting the texts correctly—that Kant hardly ever refers to “veil of perception” skepticism as “skepticism” simpliciter, instead reserving this designation for Humean skepticism and Pyrrhonian equipollence skepticism. (There are occasional exceptions to this rule, e.g., Mrongovius Metaphysics, AA 29[1,2]:927.)
4. Critique, Bxxxix.
5. AA 1:410, proposition 12.
6. AA 1:411.
7. There are a few minor exceptions. For example, there is a brief attempt to dispel this sort of skepticism in the Inaugural Dissertation of 1770, at AA 2:397, par. 11. Again, one can see from Herder’s Versuch über das Sein, written under the immediate influence of Kant’s metaphysics lectures in the early 1760s, that Kant had been discussing this sort of skepticism—as the problem of “egoism” or “idealism”—in his lectures at this period (Johann Gottfried Herder Werke, ed. U. Gaier et al. [Frankfurt am Main: Deutscher Klassiker Verlag, 1985–], 1:10–11, 17, 19, 21). Again, there are a few brief and undeveloped private notes by Kant plausibly assignable to the period in question which allude to this sort of skepticism (see esp. AA 17, no. 4445; also Reflexionen Kants zur Kritik der reinen Vernunft, ed. B. Erdmann [Leipzig: Fues’s Verlag, 1884], no. 1167; cf. nos. 1165–9). But not much more.
8. For example, in the Blomberg Logic from 1771 Kant writes of Pyrrho in a spirit of defense: “People . . . blamed Pyrrho for doubting the truth of all empirical judgments. However, this is a pure invention which has no basis at all” (AA 24:214).
9. AA 4:260–2. Cf. Critique, B5, B19–20, B127–8 (comparing A112–13, B168), A760–9 / B788–97; Critique of Practical Reason, AA 5:50–6.
10. For example, A. C. Ewing, Kant’s Treatment of Causality (London: Routledge and Kegan Paul, 1924), esp. pp. 6–15, 47; P. F. Strawson, The Bounds of Sense (London: Methuen, 1966), pp. 18–19, 88.
11. Hume, An Enquiry concerning the Human Understanding, sec. 12.
12. Thus in the Critique of Practical Reason, recounting his own debt to Hume’s skepticism concerning causation, Kant says: “I granted that when Hume took the objects of experience as things in themselves (as is almost always done), he was entirely correct in declaring the concept of cause to be deceptive and an illusion” (p. 53, emphasis added; H. E. Allison’s claim at Kant’s Transcendental Idealism [New Haven: Yale University Press, 1983], p. 18 that Kant here only means that Hume assumed the subject’s own mental states to be things in themselves seems impossibly forced). Cf. F. C. Beiser, German Idealism (Cambridge, Mass.: Harvard University Press, 2002), pp. 43–5.
13. Critique, A177 / B218. Occasional exceptions to this rule occur, e.g., in the Refutation of Idealism at B275–8, and later at A343 / B401: “inner experience.”
14. See ibid., A107: “There can be in us no modes of cognition . . . without that unity of consciousness which precedes all data of intuitions, and by relation to which representation of objects is alone possible” (cf. A110, 113, 115–16).
15. See esp. ibid., A92–4 / B125–6; cf. A96, 109–10, 115, 119–20; B130, 137, 145–7.
16. See, e.g., ibid., B233–4, A195 / B240, A201–2 / B246–7.
17. Ibid., Bxxxix–xli.
18. AA 18, nos. 5461, 5653–5, 5709, 6311–17, 6323.
19. It merely makes a brief appearance at Prolegomena, pp. 336–7, where Kant anticipates the Refutation of Idealism of the second-edition Critique.
20. Kant’s letters from the relevant period rarely touch on this sort of skepticism, and when they do, tend to be hastily dismissive of it (see, e.g., the letter to J. S. Beck of December 4, 1792, in Emmanuel Kant: Philosophical Correspondence 1759–99 [henceforth: Philosophical Correspondence], tr. A. Zweig [Chicago: The University of Chicago Press, 1986], p. 198). His Reflexionen concerning it from the period—especially AA 18, nos. 5461, 5653–5, 5709, 6311–17, and 6323—are more substantial, but hardly numerous in relation to his output, and hardly central (the lowly position and little weight that he accords it in his sketch “Von der Veranlassung der Kritik” in no. 6317 are revealing). His lectures on metaphysics from the period touch on it at a few points, but again not extensively or centrally (see, e.g., AA 28[2,1]:680, 725, 770–3). Finally, his formal writings virtually ignore it.
21. Kant’s own position in the Refutation of Idealism is in a way similar to the position I am advocating here, since he too thinks that one’s beliefs about one’s own subjective mental states are no more certain than one’s beliefs about the mind-external world (a stance which also emerges at other points in the Critique, e.g., in the Transcendental Deduction at A108, and later on at B291–4). However, he thinks them both certain—if not individually, then at least as groups—whereas I am inclined to see them both as vulnerable to skeptical attack.
22. The following are some points to note in this connection. Concerning the Fourth Paralogism, Kant’s strategy there of answering the “veil of perception” skeptic by means of a phenomenalistic reduction of objects to representations confronts at least two immediate problems: (1) There is the internal problem that—even in the first edition, let alone in the second—Kant’s official view is that phenomenalism is only true with the considerable qualification that “representation in itself does not produce its object insofar as its existence is concerned” (A92, cf. A104–5), but of course for the “veil of perception” skeptic it is above all precisely the existence of mind-external objects that is in question. (It seems to me likely that Kant lost faith in the Fourth Paralogism by the time of the second edition largely for just this reason.) (2) In order to be even remotely plausible as a thesis, a phenomenalistic reduction of objects to representations would need to reduce them to representations from different times, but—as Descartes had already hinted in the Second Meditation (“This proposition ‘I am,’ ‘I exist,’ whenever I utter it or conceive it in my mind, is true”; “I am, I exist; that is certain. For how long? For as long as I am experiencing”; “It is just not possible, when I see or . . . I think I see, that my conscious self should not be something” [Descartes: Philosophical Writings [Norwich: Thomas Nelson, 1971], pp. 67, 69, 74; emphasis added])—any “veil of perception” skeptic worth his salt will restrict the mental states concerning which he grants a person certain knowledge to the present time only.
Again, the Refutation of Idealism faces at least the following two problems: (3) In its central formulation it depends on doctrines borrowed from the First Analogy which seem very dubious—specifically, concerning the imperceptibility of time, and the consequent need for a perceptible analogon of permanent time in the form of substance. (4) To the extent that we search in the text for a more plausible line of argument which could dispense with those dubious assumptions, we encounter serious problems in connection with Kant’s highly ambiguous premise that “I am conscious of my own existence as determined in time.” In order to generate a plausible argument that such consciousness presupposes knowledge of something outside of one, we would need to construe this premise either as concerned with one’s knowing objective dates of one’s mental states (e.g., January 29, 2007 at 11 p.m.), in which case no sensible “veil of perception” skeptic will concede it, or as concerned with one’s knowing that one has had mental states in a certain order over a stretch of time, in which case the problem already indicated in (2) arises, namely that no “veil of perception” skeptic worth his salt will concede this claim either. By contrast, such a skeptic might well be prepared to concede that one knows that one is in such and such a mental state or states now (a third possible reading of the premise). But then, how is Kant going to argue plausibly from that concession to one’s knowledge of something outside oneself?
CHAPTER THREE
SKEPTICISM AND METAPHYSICS (A PUZZLE)
1. Cf. W. H. Walsh, Reason and Experience (Oxford: Clarendon Press, 1947), pp. 130, 238–9.
2. See M. Wundt, Die deutsche Schulmetaphysik des 17. Jahrhunderts (Hildesheim: Georg Olms, 1992), Die deutsche Schulphilosophie im Zeit-alter der Aufklärung (Hildesheim: Georg Olms, 1964).
3. Concerning this conception of metaphysics within the Wolffian tradition and Kant’s relation to it, see H.J. de Vleeschauwer, “Wie ich jetzt die Kritik der reinen Vernunft entwicklungsgeschichtlich lese,” Kant Studien, vol. 54, no. 4 (1963). For an example of the precritical Kant’s own commitment to such a two-part metaphysics, see his Notice concerning the Structure of Lectures in the Winter Semester 1765–1766, at AA 2:308–10. One anomaly should perhaps be mentioned: Wolff and Baumgarten had both included within special metaphysics an empirical as well as a rational psychology (Kant himself identifies this as an anomaly at AA 20:281; cf. Critique, A848–9 / B876–7).
4. Critique, A761 / B789: “The first step in matters of pure reason, marking its infancy, is dogmatic. The second step is skeptical . . . But a third step, such as can be taken only by fully matured judgment . . . is now necessary . . . This is . . . the criticism of reason.”
5. Prolegomena, p. 260. Cf. Critique, A760–8 / B788–96.
6. Philosophical Correspondence, p. 252.
7. “This is indeed puzzling. Can it be due to a lapse of memory? The letter is filled with complaints about Kant’s declining health and mental abilities” (L. W. Beck, Essays on Kant and Hume [New Haven: Yale University Press, 1978], p. 119).
8. Prize Essay, AA 20:319–20: “Another strange phenomenon in the end inevitably startled awake the Reason which was slumbering on the pillow of the cognition which it supposed itself to have extended beyond all bounds of possible experience by ideas, and that is the discovery that . . . those [a priori propositions] which overstep the bounds of possible experience . . . come into conflict, partly among themselves, partly with those [a priori propositions] which are concerned with knowledge of nature, and seem to destroy one another, but thereby to rob reason of all confidence in the theoretical field, and to introduce a boundless skepticism. Against this disaster there is, now, no remedy but to subject Reason itself, i.e., the faculty of a priori cognition generally, to an exact and detailed critique.” Cf. Critique, A407 / B433–4,
CHAPTER FOUR
KANT’S PYRRHONIAN CRISIS
1. It is characteristic of the Critique’s Antinomies that “thesis, as well as antithesis, can be shown by equally convincing, clear, and irresistible proofs” (Prolegomena, p.340).
2. In the Critique the first two Antinomies are concerned with the world as a whole, the third Antinomy with the human soul, and the fourth Antinomy with God (though Kant also tends to think of them all as concerned with the world as a whole).
3. Philosophical Correspondence, p.97; emphasis added. Lambert’s original letter making his proposal has survived (see ibid., pp. 46–7).
4. The First Antinomy—the world has a beginning in time and limits in space vs. it has no beginning and no limits—is not found in Kant’s own writings from the period in question, but its spatial version had already been published by Crusius in the 1740s (see L. W. Beck, Early German Philosophy [Cambridge, Mass.: Harvard University Press, 1969], pp. 399–400) and would therefore certainly have been familiar to Kant. Famously, an early version of the Second Antinomy—all composite substances are made up of simple parts vs. none are—is already found in Kant’s Physical Monadology of 1756 (where he sets philosophy’s denial of the infinite divisibility of space and mathematics’ affirmation of this into opposition to one another). A version of the Third Antinomy—there is a causality of freedom vs. there is only thoroughgoing causality in accordance with the laws of nature—had already been developed by Crusius in the 1740s (see Beck, Early German Philosophy, p. 400), and though it was at first countered by Kant with a compatibilist argument in the New Elucidation of 1755, it was subsequently taken much more seriously by him in The Only Possible Argument in Support of a Demonstration of God’s Existence of 1763, where he pruned back his initial claim of compatibility to a claim that the causality of freedom is not entirely emancipated from the laws of nature, and says that the nature of the causality of freedom is not properly understood (AA 2:110–11). Finally, concern about a version of the Fourth Antinomy—there is an absolutely necessary being vs. there is none (only a chain of contingent causes)—already lies behind Kant’s argument in the Universal Natural History and Theory of the Heavens of 1755 for nature being thoroughly explicable in terms of natural laws and for this being not only compatible with God’s design of the world, but actually indicative of it, and better in keeping with it than a restriction of the applicability of natural laws within nature (this is a sort of counterpart to Kant’s compatibilist argument concerning human freedom from the New Elucidation of the same year). Concerning the precritical Kant’s anticipations of the four Antinomies of the Critique, cf. N. Hinske, Kants Weg zur Transzendentalphilosophie (Stuttgart: Kohlhammer, 1970).
5. As Hinske notes—ibid., pp. 95–6—this only happened later. It began in the Inaugural Dissertation of 1770, where Kant for the first time brought together the first two Antinomies (the third and the fourth were added to the system still later).
6. For example, in the Physical Monadology of 1756 Kant mentions not only an early version of the Second Antinomy but also several other antinomies which do not belong to the canonical four: “How to reconcile metaphysics, which denies that space is infinitely divisible, with geometry, which asserts it with certitude? [This is Kant’s early version of the Second Antinomy.—M.N.F.] Geometry contends that empty space is necessary for free motion, metaphysics denies it. Geometry holds universal attraction or gravitation to be hardly explicable by mechanical causes, but shows clearly that it comes from forces inherent in bodies at rest and acting at a distance; metaphysics renounces this as an imaginary plaything” (AA 1:475–6). Another example of non-canonical antinomies is provided by Dreams of a Spirit Seer, Illustrated by Dreams of Metaphysics from 1766, namely in its reflections on the philosophy of mind. Concerning the whole question of further antinomies, cf. Hinske, Kants Weg zur Transzendentalphilosophie, pp. 95–6.
7. Sextus Empiricus, Outlines of Pyrrhonism (Cambridge, Mass.: Harvard University Press, 1976), bk. 1, sec. 12 (translation slightly amended): “The main basic principle of the skeptic system is that of opposing to every proposition/argument [logos] an equal proposition/argument; for we believe that as a consequence of this we end by ceasing to dogmatize.”
8. Critique, A388–9.
9. AA 2:306–7.
10. The term “zetetic” is one distinctively used by the Pyrrhonists to characterize themselves and this procedure of theirs (see, for example, Outlines of Pyrrhonism, bk. 1, sec. 7: “The skeptic school, then, is also called zetetic [zêtetikê] from its activity in investigation [to zêtein] and inquiry”). In a letter to Mendelssohn roughly contemporary with the Notice Kant advocates what is obviously the same thing as the “zetetic” method referred to in the latter as a method of treating metaphysics’ “pretended insights skeptically” (Philosophical Correspondence, p. 55; emphasis added). In some roughly contemporary notes written into his copy of his own work Observations on the Feeling of the Beautiful and the Sublime he states concerning metaphysics: “The doubt that I accept is . . . a doubt of postponement. Zetetici (zêtein), Seekers. I shall accentuate the grounds belonging to both sides” (AA 20:175). Again, if further evidence were needed, in the Blomberg Logic of 1771 Kant notes that the Pyrrhonists describe themselves as “zetetici” (pp. 36, 213), and in Reflexionen from the 1770s he uses the same adjective to characterize the method of balancing opposed arguments (AA 17, nos. 4454–5). Concerning the adjective’s reference to Pyrrhonism in the passage quoted in my main text, cf. G. Tonelli, “Kant und die antiken Skeptiker,” in Studien zu Kants philosophischer Entwicklung, ed. H. Heimsoeth et al. (Hildesheim: Georg Olms, 1967), pp. 94–6. In seeing the adjective as referring more specifically to the Pyrrhonists’ equipollence method I go beyond Tonelli, however (contrast ibid., p. 121, n. 98).
11. Observe in particular that the Notice still only envisaged applying the zetetic method to metaphysics as a means to discovering truth in the discipline, not yet as a means to destroying it, for the text says that the method’s catharsis will lead to a metaphysics which “becomes . . . in various parts dogmatic, that is decided” (AA 2:307).
12. To illustrate the fundamental role that the application of this method to metaphysics plays in Dreams of a Spirit Seer (since this may not be immediately obvious from the text): (1) That role is evident from Kant’s explicit comments there on his approach to metaphysical questions. For example, he writes, “I am sufficiently sure of myself not to fear any opponent, no matter how dreadful his weapons may be . . . so that I can in this case test argument against argument in refutation, for among men of learning such testing is really the art of demonstrating each other’s ignorance” (AA 2:328). (2) It is also evident from his practice in the text of actually arguing on both sides of metaphysical questions in order to motivate suspension of judgment. In particular, after opening with some remarks which call the very concept of “spirit” into question (p. 320 ff.)—incidentally, an opening move that is typical of Sextus Empiricus, who often prefaces an equipollence argument in this way—he goes on to develop an argument which is generally supportive of a view like Swedenborg’s concerning the relation between a world of spirits and the world of material phenomena (pp. 322–41), but then turns round and argues against any such view, instead offering a physiological explanation of its delusiveness (pp. 342–8), and finally concludes with the moral that we are ignorant about such matters (pp. 349–52). (3) It is because of his conviction that this method can always be deployed successfully against supersensuous metaphysics that he argues in the essay that the judgments of supersensuous metaphysics are irremediably subjective rather than intersubjective, and consequently nothing more than a kind of dream: “dreams of reason,” no better than the sensuous illusions, or “dreams of sensation,” of the seer Swedenborg when he claims direct acquaintance with a spiritual realm unrecognized by the rest of us (p. 342). For Kant’s thought here, to unpack it a little, is as follows. He is assuming that intersubjectivity in judgments is a necessary condition of their constituting objective cognitions rather than merely expressing a subject’s own mental states (“It seems to me that one ought, perhaps . . . to say: if different people have each of them their own world, then we may suppose that they are dreaming” [p. 342]; incidentally, this assumption subsequently survives to play an important role in Kant’s critical period). And he thinks that the equal balance of opposed arguments in supersensuous metaphysics which is uncovered by the zetetic or equipollence method ensures that judgments in the discipline are merely subjective rather than intersubjective, because it leaves, not rational arguments, but instead non-rational inclinations as the ultimate determinants of belief within the discipline, and these, unlike decisive rational arguments, are variable from individual to individual (pp. 349–51).
13. “Previously we moved like Democritus in empty space where the butterfly wings of metaphysics had carried us and we conversed there with spiritual forms. Now that the contracting power of self-knowledge has drawn together metaphysics’ silken wings, we find ourselves on the lowly ground of experience and the common understanding; fortunate! if we regard this ground as our appointed place, which we can never leave without paying the penalty” (ibid., p. 368).
14. Ibid., pp. 306–8, 342, 368–72 (empirical disciplines and mathematics); pp. 311, 334–5, 372 (first-order moral judgments); p. 310 (common logic). (Moral philosophy and more ambitious forms of logic are another matter—see pp. 308, 310–11.)
15. Ibid., p. 367.
16. This move had precedents among some of Kant’s skeptically minded modern predecessors. See Tonelli, “Kant und die antiken Skeptiker,” pp. 107–8.
17. Pace Tonelli, ibid., p. 110. Tonelli mistakenly takes a contrary view because he underestimates both (1) the distinctiveness of Dreams of a Spirit Seer in comparison with previous works by Kant and (2) the moderateness of orthodox Pyrrhonism as Kant interprets it.
18. More exactly, it is precisely the estimation belonging to what Kant considers the superior orthodox Pyrrhonism of Pyrrho himself, as opposed to what Kant considers the inferior excessive skepticism propounded by some later Pyrrhonists and Academic skeptics (Blomberg Logic, pp. 210, 213–16).
19. On the one hand: “People . . . blamed Pyrrho for doubting the truth of all empirical judgments and not placing faith in them. However, this is a pure invention which has no basis at all” (ibid., p. 214, emphasis added); “Pyrrho: that general dogmata (except for mathematics) were uncertain” (Herder Logic, AA 24:4, emphasis added); “That [Pyrrho] rejected each and every dogma is fundamentally mistaken. One who accepts no dogmas at all cannot teach morality at all” (Blomberg Logic, p. 214, emphasis added); for Pyrrho “there are certain, so to speak, eternal principles of our reason [i.e. logic and mathematics—M.N.F.], which cannot be disputed at all” (ibid., p. 214). On the other hand: “[Pyrrho] taught only that one ought not to accept and decide on the propositions of philosophy straight away but should begin by doubting them” (Philippi Logic, AA 24:330), and “it is beyond doubt that he particularly rejected many rational judgments” (Blomberg Logic, pp. 213–14).
20. Dreams of a Spirit Seer, pp. 318, 368 (the useful); 368, 373 (happiness); 338–9, 351, 362–3 (life).
21. In thus accounting for Kant’s skepticism about metaphysics in Dreams of a Spirit Seer as primarily Pyrrhonian in character and inspiration, I am somewhat at odds with K. Fischer, who argues at Immanuel Kant und seine Lehre (Heidelberg: Carl Winter, 1889), pt. 1, pp. 265–73 that the skepticism which underlies the essay is Humean skepticism. Favoring my reading over Fischer’s, note, in addition to the evidence already presented, also the following fact: Just a few years after Dreams of a Spirit Seer, in the Blomberg Logic of 1771, Kant gives an extended and highly sympathetic account of Pyrrhonism (pp. 4, 36, 207–18), whereas by contrast his discussion of Hume is brief, critical, poorly informed, and moreover interprets Hume as in effect merely a sort of inferior Pyrrhonist—in particular, as a skeptic who uses the method of balancing opposed arguments (which he does not) in order to arrive at an unrestricted and therefore excessive doubt (pp. 210–11, 217). However, my disagreement with Fischer is not as sharp as it may seem at this point, for I would allow that Hume’s influence is at work in the essay alongside Pyrrhonism’s (especially in connection with the essay’s thesis that particular causal connections and laws cannot be known a priori but only on the basis of experience—concerning which more later).
22. Critique, Aviii. Cf. B22–3: reason’s “dogmatic employment . . . lands us in dogmatic assertions to which other assertions, equally specious, can always be opposed—that is, in skepticism.”
23. Prolegomena, pp. 255, 271 (the historical allusion here is of course to ancient skepticism). Cf. Prize Essay, p. 263; and Kant’s lectures on metaphysics from the critical period, which contain numerous similar statements, e.g., at AA 28(1):362, 378; 28(2,1):618–21; 29(1,2):752, 779–80, 955–8.
24. For example, compare with the passage just quoted from the first edition of the Critique, A421–5 / B448–53 in the same work.
25. Passages such as the following might be cited in support of such a proposal: “There are just so many, neither more nor fewer” (ibid., A462 / B490).
26. See ibid., A395 / B417–18, note b; A701–2 / B729–30; A741 / B769. Cf. Kant’s encouragement of a broad “defensive” use of the equipollence method in relation to metaphysics at A776–82 / B804–10.
27. See Critique of Practical Reason, pp. 113 ff.; Critique of Judgment, AA 5:338 ff., 385 ff.; Religion within the Limits of Reason Alone, AA 6: 116 ff.
CHAPTER FIVE
HUMEAN SKEPTICISM
1. Critique, A760–8 / B788–96.
2. Hume, Enquiry, sec. 4, pt. 1. Thus Kant writes: “[Hume] demonstrated irrefutably that it was entirely impossible for reason to think such a connection a priori and by means of concepts . . . We cannot at all see how, in consequence of the existence of one thing, another must necessarily exist, and how therefore the conception of such a connection can be introduced a priori” (Prolegomena, p. 257; cf. Critique, A765–6 / B793–4).
3. Hume, Enquiry, sec. 7. Thus Kant writes: “[Hume] concluded that reason deceives itself entirely in this concept, that . . . it is in fact nothing but a mongrel of the imagination which, impregnated by experience, has brought certain representations under the law of association and substitutes a subjective necessity which arises thereby, that is, custom, for an objective necessity arising from insight” (Prolegomena, pp. 257–8). Cf. Critique, B127, A760 / B788; Critique of Practical Reason, pp. 50–1. Also, Volckmann Metaphysics, AA 28(1):403–4: “Hume . . . based a whole skeptical philosophy on the question: how do you arrive at the concept of cause? A cause is that which contains the real ground for something’s existence and is quite one with the real ground, for example the wind is the cause of the ship’s motion. How can it be, then, that when you posit the wind, at the same time something quite different follows as well, namely the motion of the ship? What sort of connection has the wind with the motion? According to him, all concepts of cause and effect are got from experience, and necessity, he says, is merely something imagined and a long habit. Hence he saw no other way, since he could not derive the relation from reason, than to assume the concept of the real ground to be an empirical concept.”
4. Hume, A Treatise of Human Nature (henceforth: Treatise), bk. 1, sec. 3. The Prolegomena’s opening discussion of Hume’s reminder refers to this third Humean view only obliquely, but its central importance for Kant is clear from later parts of the Prolegomena (in particular, secs. 27–30), and from corresponding passages in the Critique (B4–5, B19–20, A760 / B788, A765–6 / B793–4).
L. W. Beck and others have indeed seen this Humean view concerning the causal principle as the Humean catalyst in Kant’s awakening from his dogmatic slumber in metaphysics (Essays on Kant and Hume, pp. 117–20). However, I hope to show that that is an oversimplification.
5. Dreams of a Spirit Seer, p. 370: “The rules of reason concern only comparison with respect to identity and contradiction. But in the case of a cause something is assumed to have come from something else, so that one can find no connection with respect to identity. Similarly, if I want to postulate that the same thing is not a cause, a contradiction never results, because no contradiction arises if one postulates something but does away with something else. Therefore our fundamental concepts of things as causes . . . are absolutely arbitrary and can be neither proved nor disproved unless they are derived from experience.” It would be perverse to deny Hume’s role in influencing Kant to accept this doctrine in Dreams of a Spirit Seer, as Erdmann did—B. Erdmann, Reflexionen Kants zur Kritik der reinen Vernunft, pp. xxi ff., lff., and “Kant und Hume um 1762,” pts. 1 and 2, Archiv für Geschichte der Philosophie, vol. 1 (1887–8). (1) By the time Kant wrote the passage just quoted in 1766 Hume’s Enquiry had been available to him in German translation for over ten years; we have ample evidence of his reading Hume in the intervening period (e.g., Herder, who attended Kant’s lectures in the early 1760s, notes in his Briefe zu Beförderung der Humanität that Hume was one of the most frequently cited authors, and Hume is explicitly referred to in connection with practical philosophy in Kant’s Notice of 1765); and, as we have seen, the Prolegomena acknowledges the doctrine’s indebtedness to Hume about as prominently and explicitly as one could wish. (2) The case for Hume’s influence is, moreover, reinforced by the fact that Dreams of a Spirit Seer contains, immediately following the passage just quoted, a further passage on the subject of causation—specifically, concerning the causal influence of the will on the body—which is so similar in both argument and illustrative details to a passage from the Enquiry on the same subject that it seems virtually inconceivable that it was written without that passage in mind. Kant’s passage reads: “I know to be sure that thought and volition move my body, but I can never reduce this appearance as a simple experience to another through analysis, so that I can certainly recognize it but not understand it. That my will moves my arm is no more intelligible to me than if someone were to say that the will could also control the moon in its orbit; the difference is only this: that I experience the former but the latter has never occurred to my senses” (p. 370). Compare with this the following passage from the Enquiry: “Is there any principle in all nature more mysterious than the union of soul with body; by which a supposed spiritual substance acquires such an influence over a material one, that the most refined thought is able to actuate the grossest matter? Were we empowered, by a secret wish, to remove mountains, or control planets in their orbit; this extensive authority would not be more extraordinary, nor more beyond our comprehension” (sec. 7, pt. 1). Erdmann’s attempt—“Kant und Hume um 1762,” pt. 2, pp. 218–19—to dismiss the resemblance between these passages as merely coincidental is surely very implausible; and his claim in support of this dismissal that there is nothing that makes the hypothesis of an influence of Hume on Kant in Dreams of a Spirit Seer “secured elsewhere [anderweitig gesichert],” unless it presupposes an absurdly high standard for evidence, is simply false, due to the remarks on particular causal judgments in Dreams of a Spirit Seer and the Prolegomena which we have just considered, and also the presence of moral reflections in a Humean spirit at the end of Dreams of a Spirit Seer (pp. 372–3) shortly after Kant had explicitly praised Hume’s moral philosophy in the Notice (p. 311). (3) None of this need be to deny that influences other than Hume were also pushing Kant’s views on the subject of particular causal judgments in the direction of the position found in Dreams of a Spirit Seer—as has been plausibly suggested by L. W. Beck, and especially G. Tonelli (“Die Anfänge von Kants Kritik der Kausalbeziehungen und ihre Voraussetzungen im 18. Jahrhundert,” Kant Studien, vol. 57 [1966]).
6. Dreams of a Spirit Seer, pp. 370–1; Philosophical Correspondence, pp. 55–7.
7. See, for example, the Physical Monadology of 1756, at AA 1:475, and the Inquiry concerning the Distinctness of the Principles of Natural Theology and Morality from 1764, at AA 2:286.
8. Further Humean influences on Dreams of a Spirit Seer are likely as well. In addition to those already mentioned in my main text and in note 5, plausible candidates are passages which suggest that all of our concepts must be empirical (pp. 322, 367–8) and passages which confine our non-mathematical, non-logical, non-moral knowledge to experience (pp. 351–2, 368 ff.).
9. Prolegomena, p. 260.
10. See esp. Dreams of a Spirit Seer, p. 370.
11. See, e.g., Critique, A127, B165, A216 / B263.
12. Cf. ibid., B127, A609 / B637, A621–2 / B649–50, A635–6 / B663–4.
13. Concerning the Metaphysical/Transcendental Deduction serving as an answer to the second of the three Humean views, cf. Critique, B127–8, 168.
14. Ibid., A760–1 / B788–9.
15. AA 2:392, 395–6.
16. The significance of this letter has been best recognized by some of the Francophone literature on Kant, in particular work by H. J. de Vleeschauwer and B. Longuenesse.
17. AA 2:392. Cf. the similar definition of sensibility at Critique, A19 / B33.
18. In the letter Kant seems to think that an early version of the Critique’s Metaphysical Deduction holds the key to solving this problem. That would turn out not to be quite right in the end, though it did provide the key for solving a closely related problem in which he was about to get interested (Hume’s problem concerning the very existence of a priori concepts).
19. For Kant’s statement of these two problems in the letter to Herz, see Philosophical Correspondence, pp. 71–2. Cf. AA 17, nos. 4470, 4473, the former of which approximately sums up these two worries thus: “The primary question is: how concepts of things can arise in us which have not become known to us through any appearance of the things, or propositions which no experience has taught us.”
20. I here agree with the scholarly consensus (B. Erdmann, N. Kemp Smith, R. P. Wolff, L. W. Beck, et al.) that Kant was first made aware of Hume’s view concerning the causal principle—which is clearly stated only in the Treatise, effectively inaccessible to the precritical Kant with his poor to non-existent grasp of English because as yet untranslated into German, not in the Enquiry, already available to him in German translation since the 1750s—by the 1772 German translation of this work of Beattie’s, which contains a clear exposition of the view in question.
21. For example, already in the Herder Metaphysics from the early 1760s Kant associates such a position with Aristotle and Locke (AA 28[1]:60; 28[2,1]:851–2, 952 ff.). In subsequent metaphysics lectures he returns to the subject of this position repeatedly, associating it mainly with Aristotle (ambiguously), Locke, and Epicurus. Hume was therefore not the first or even the main figure in whom Kant found such a position.
22. For example, already in the Universal Natural History and Theory of the Heavens of 1755 Kant had argued that “man . . . has all his concepts and representations from impressions which the universe rouses in his soul through the body” (AA 1:355). Again, in the Herder Metaphysics from the early 1760s he asserts that “all our concepts begin from sensory feelings,” and that all our unanalyzable concepts are sensory (AA 28[1]:60; 28[2,1]:953). Again, in Dreams of a Spirit Seer from 1766 he remarks that “it is, of course, impossible to form any concept of that which deviates from common empirical concepts and which no experience can explain, even analogically,” and he writes of “empirical concepts, upon which all our judgments must at all times be based” (pp. 322, 367–8).
On the other hand, the Herder Metaphysics also argues that logical possibility and impossibility are examples of concepts which are not sensory (AA 28[2,1]:936–8). And similarly, in Dreams of a Spirit Seer Kant qualifies the remarks just quoted to allow for exceptions (p. 320, footnote). Moreover, exceptions soon afterwards play a major role in the Inaugural Dissertation.
23. For a vacillation between these two worries in a Reflexion from the 1770s, see AA 17, no. 4470.
24. Critique, A84 / B116–17. Cf. the discussion of Hume at B127; also B168.
25. Ibid., A85 / B117.
26. Kant’s considered position, not only in the critical writings, but already in the Reflexionen of the 1770s, is that there are concepts which violate the Humean principle (see, e.g., AA 17, nos. 3927, 3930). This had indeed also been his considered position even earlier, for example in the Inaugural Dissertation.
27. Thus, in the Critique Kant would write that “the very concept of a cause so manifestly contains the concept of a necessity of connection with an effect . . . that the concept would be altogether lost if we attempted to derive it, as Hume has done, from a repeated association of that which happens with that which precedes, and from a custom of connecting representations, a custom originating in this repeated association, and constituting therefore a merely subjective necessity” (B5, emphasis added). Cf. the following remark in the Mrongovius Metaphysics of 1783: “Locke said all concepts are borrowed from experience. Leibniz no. We also have some through pure reason. That is easy to decide, e.g., can experience very well provide the concept of cause and effect . . .?” (p. 781).
28. Thus in the Critique Kant would write that “the concept of a cause involves the character of necessity, which no experience can yield” (A112; cf. A91 / B123–4).
Note that this confirmed rather than overturning Kant’s predominant pre-existing intuitions concerning the matter. Already in the Herder Metaphysics from the early 1760s, for example, he had noted: “Cognition of [cause and effect] is merely a judgment. I do not see that lightning sets a house on fire” (AA 28[2,1]:928). And in the Inaugural Dissertation he had argued, similarly, that the concept of a cause was not derivable from sensation (AA 2:395).
29. Thus in a passage of the Von Schön Metaphysics from the critical period Kant characterizes such a priori concepts as, unlike such a priori concepts as God, those “to which an object in our experience corresponds,” those which “have . . . objective reality, they can be measured off from the object of experience,” those “which we really need in order to understand the objects which present themselves” (AA 28[1]:470; cf. Critique, A720–1 / B748–9).
30. Thus in the Mrongovius Metaphysics Kant treats quite respectfully a position—which he finds Aristotle and Locke embracing inconsistently, but praises Epicurus for embracing consistently—that “if all concepts are borrowed from experience, then they can accept nothing except what rests on experience. But God is in no experience. Therefore we cannot say anything about him” (p. 763).
31. Hume, Enquiry, sec. 4, pt. 1.
32. Ibid., sec. 12, pt. 3.
33. For reasons which have been pointed out by L. W. Beck and which are discussed in the next note, it can be questioned whether this gives a completely accurate account of Hume’s view of the causal principle in the Treatise. It does, however, give an accurate account of Hume’s considered view of the causal principle as distilled from the Treatise and the Enquiry together. And, more importantly for our present purposes, it gives an accurate account of the view of the causal principle which Kant—aware of the Treatise’s general conclusion concerning the causal principle, but drawing his understanding of the “fork” which drove Hume to that conclusion rather from the Enquiry—would inevitably have taken to be Hume’s.
34. My implication here that Kant identified Hume’s relations of ideas with his own analytic judgments, understanding truth in virtue of the law of contradiction to be definitive of both, is conventional wisdom. It has, however, been challenged by L. W. Beck, at Essays on Kant and Hume, pp. 82 ff. Beck argues that, although if one goes by the opening passages of Enquiry, sec. 4, where Hume explains relations of ideas, “relations of ideas . . . seem to be judged in what Kant calls analytic judgments,” nevertheless “Kant was correct in not taking this to mean that relations of ideas . . . are equivalent to the relation expressed in an analytical judgment.” As far as I can see, Beck’s only evidence for his implication here that Kant did not believe relations of ideas and analytic judgments to be equivalent is the fact that Kant does not explicitly cite the opening paragraphs of Enquiry, sec. 4 and point out the equivalence. However, that is very weak evidence indeed. And it seems still weaker when one remembers that, as Beck himself correctly notes, “again and again Kant writes as if Hume had distinguished between analytic and synthetic judgments and had categorically denied the possibility of a priori synthetic judgments.” For if Kant did not see relations of ideas—to all appearances defined at Enquiry, sec. 4, pt. 1 as principles whose negations violate the law of contradiction, just as Kant defines analytic judgments at Critique, B12, 15–16, A151 / B190–1, Prolegomena, p. 267, and elsewhere—as Hume’s version of analytic judgments, then what did Kant see as Hume’s version of analytic judgments?! Beck has been misled into his implausible implication that Kant did not regard Humean relations of ideas and his own analytic judgments as equivalent through a more plausible thesis that if Kant had read Hume’s Treatise then he would have had reason to distinguish them. Thus Beck writes: “The relations of ideas in the Enquiry correspond to the ‘necessary and unalterable’ philosophical relations of the Treatise. But the necessary and unalterable relations are not analytical in the sense that . . . the denial of such a relation involves a contradiction. By ‘contradiction’ Hume does not mean merely an assertion like ‘A is not A.’ He means also ‘A is not B’ where an A that is not a B is ‘inconceivable’ or ‘unimaginable’ . . . [Hume’s] relations of ideas are not . . . analytic propositions as Kant understood the term, though this cannot be seen by anyone who, like Kant, reads only the Enquiry and not also the Treatise. Had Kant read Hume’s Treatise, he would have found Hume tacitly admitting a class of . . . relations of ideas which are not testable by the logical law of contradiction.” This more plausible thesis of Beck’s still strikes me as inaccurate. Is the situation not rather that the Enquiry really does understand its relations of ideas to be based on the law of contradiction in the same way as Kant’s analytic judgments and that this marks a change of position from the Treatise’s account of its—consequently, only approximately equivalent—necessary and unalterable relations? However, the more important point to note here is that, that question aside, inferring from this more plausible thesis to the conclusion that Kant really did understand Humean relations of ideas and his own analytic judgments to be non-equivalent is a non sequitur. Moreover, it is a non sequitur the implausibility of whose conclusion is sufficiently shown by two facts which Beck acknowledges in his very statement of his more plausible thesis: (1) Kant had not read the Treatise and (2) the Enquiry, which he had read, gives no indication that Hume takes as his criterion of relations of ideas anything other than the violation of the law of contradiction by their negations. On the question of whether Kant regarded relations of ideas and analytic judgments as equivalent, then, conventional wisdom proves for once to be wisdom.
35. For example, in the Prolegomena he states that “all analytic judgments depend wholly on the principle of contradiction, and are in their nature a priori cognitions” (p. 267). I say “of course” not only because of the intrinsic force of the point, but also because Kant really already believed principles true in virtue of the law of contradiction to be knowable a priori, as can be seen, for example, from a passage on causal knowledge at Dreams of a Spirit Seer, p.370.
36. Critique, B19. That Hume’s “fork,” as illustrated in his treatment of the causal principle, did in this way lead Kant to this central puzzle of the critical philosophy is reflected in the fact that the Critique, after its first formulation of the puzzle, immediately goes on to say: “Among philosophers, David Hume came nearest to envisaging this problem . . . He occupied himself . . . with the synthetic proposition regarding the connection of an effect with its cause (principium causalitatis) and he believed himself to have shown that such an a priori proposition is entirely impossible. If we accept his conclusions, then all that we call metaphysics is a mere delusion” (B19–20).
37. In a passage from his 1794–5 lectures on metaphysics, Kant reportedly went as far as to say the following: “Metaphysics contributes nothing to the extension of empirical principles, to the science of empirical physics: the cognition is quite unnecessary in connection with physics, in that the principles of metaphysics are quite set aside and recognized appearances are the basis, and the fundamental principles derived from them are sufficient for explaining everything. Experience confirms, e.g., the rational propositions: in all changes substance never vanishes but only the form of things, or: each change has its cause; so much so that one simply accepts them without investigating their basis, and one already becomes certain through experience of their truth under all circumstances” (AA 29[1,2]:947–8).
38. Hence, in a passage from the Dohna Metaphysics of 1792–3 Kant gives the causal principle as his paradigmatic example of a priority in a synthetic principle outside of mathematics: “The situation with physical axioms is similar [to that with mathematical propositions]; we know them, are convinced of them, without recourse to experience, e.g., all effects have a cause” (AA 28[2,1]:620). And in the Critique he gives as a reason for the unacceptability of Hume’s account of the causal principle the circumstance that “the concept of a cause so manifestly contains the concept . . . of the strict universality [and consequently—since Kant takes strict universality to be a “criterion,” i.e., at least a sufficient condition, of a priority; see Critique, B3–4—the a priority] of the rule [‘every alteration must have a cause’]” (B4–5, emphasis added). Cf. Prolegomena, p. 275, where Kant argues that we can take it as given that we have synthetic a priori knowledge of the principles of pure physics, which include the causal principle.
39. The causal principle was not, though, for Kant, the only, or even the most compelling, example of synthetic a priori knowledge that could be appealed to in this way in order to refute the now-refined objection; mathematics afforded examples of at least equal force. An analogous point applies in connection with the objections concerning concepts: in Kant’s view, mathematics contains clear examples of a priori concepts which both genuinely exist and refer.
40. Prolegomena, pp. 258–60.
41. AA 12:371.
42. Critique, A50: “When [concepts] contain sensation . . . they are empirical. When there is no mingling of sensation with the representation, they are pure.” Cf. AA 17, no. 3965: “A concept which cannot be regarded as an impression of the senses is pure.” The slightly broader definition of a concept’s “a priority” which I give here in the main text follows from these definitions of a concept’s “purity” together with Kant’s standard explanation of the distinction between “purity” and “a priority” at Critique, B3. Kant does not himself usually distinguish between pure concepts and merely a priori concepts. But he does sometimes imply such a distinction—e.g., at Critique, A95 he refers to “pure a priori concepts,” and evidently does not intend this expression to be merely pleonastic. The concepts of the understanding in their purely logical meanings certainly count as pure concepts. However, in their schematized meanings—when, for example, the category of reality gets schematized in terms of sensation—they may only be a priori concepts.
43. For Kant’s sensitivity to such a possibility, and perception of the need to address it, see for example Prize Essay, p. 263, where he interprets the more extreme forms of skepticism which have called into question “even . . . principles of the cognition of the sensible and . . . experience itself” as “perhaps a challenge to the dogmatists . . . to prove those a priori principles on which rests even the possibility of experience, and since they could not do so, to depict experience as doubtful to them as well” (the anacoluthon here is Kant’s own).
CHAPTER SIX
KANT’S REFORMED METAPHYSICS
1. Prolegomena, p. 371; cf. pp. 255–7, 271–5, 327, 365–71, 383; and Critique, A849–51 / B877–9. This fundamental assumption of Kant’s is somewhat obscured by the many passages in which he makes less standard uses of the term “metaphysics,” and has consequently not always been recognized by the secondary literature. For example, De Vleeschauwer evidently overlooks it when he argues that the critical philosophy saves a metaphysics concerning the subject matter of traditional special metaphysics—“Wie ich jetzt die Kritik der reinen Vernunft entwicklungsgeschichtlich lese,” p. 363.
2. “Metaphysics actually exists, if not as a science, yet as a natural disposition,” which “considered by itself alone . . . is dialectical and illusory” (Critique, B21; Prolegomena, p. 365).
3. “The term ‘metaphysics,’ in its strict sense, is commonly reserved for the metaphysics of speculative reason [i.e., for theoretical metaphysics—M.N.F.]” (Critique, A842 / B870; cf. Prolegomena, pp. 278–9; Prize Essay, p. 261).
4. There are a few passages in the Critique which are potentially very misleading in this connection. At A334–5 / B391–2, and again in a footnote at B395, Kant seems to envisage metaphysics as a science of precisely such items. However, he is here using the word science either in an inverted-commas sense or in the sense of a genuine science but one that treats the illusions which occur in connection with such items—cf. A408 / B435: “pretended science”; A841 / B869: metaphysics is “the science which exhibits in systematic connection the whole body (true as well as illusory) of philosophical cognition arising out of pure reason” (emphasis added). Again, at A845–7 / B873–5 he gives a very broad definition of metaphysics which includes cognition of such items. However, this definition is immediately preceded by the passage just quoted from A841 / B869 in which he conceives the discipline as one which contains “true as well as illusory” cognition.
5. See ibid., B395, note a. Kant also makes this point repeatedly in the lectures on metaphysics—see, e.g., AA 28(i):301–2, 381–3; 28(2,1):618–20, 774–5, 821–2; 29(1,2):947–8.
6. Critique, Bxix. Cf. A471 / B499, A701–2 / B729–30. Also Prize Essay, pp. 260, 296–301, 316–17.
7. Critique, A50 / B74: “Neither concepts without an intuition in some way corresponding to them nor intuitions without concepts can yield a cognition”; “Thoughts without content are empty, intuitions without concepts are blind.”
8. This is obvious enough where empirical knowledge is concerned. But according to Kant it is also true of mathematical synthetic a priori knowledge, whose basis in pure intuition requires confirmation by sensible intuitions if it is to constitute real knowledge, and of such non-supersensuous metaphysical synthetic a priori knowledge as knowledge of the causal principle, which, although lacking any proof in pure intuition, and although not derived from sensible intuitions, nonetheless requires confirmation by sensible intuitions in order to constitute real knowledge. See esp. Critique, B147, A156–8 / B195–7; cf. A239–40 / B298–9.
9. Concerning this deeper reason, see esp. Prize Essay, pp. 273–7, 296. For Kant this deeper reason (supersensuous metaphysics’ lack of the required foundation in intuitions) in effect explains the former one (its incurable vulnerability to equipollence problems).
10. Critique, Bxix–xxii, Bxxx–xxxi, A470–1 / B498–9, A634 / B662.
11. See ibid., A508 / B536 ff., A616–20 / B644–8, A642–702 / B670–730.
12. See ibid., Axx–xxi, Bxliii, A841 / B869, A845–7 / B873–5. Note that “physiology” here does not imply a restriction to the physical but instead bears the broad etymological sense of a science of nature (physis).
13. Cf. Metaphysical Foundations of Natural Science, AA 4:469–70, where Kant provides a fuller account than in the Critique of how the foundation of the “metaphysics of nature” treated in the Transcendental Analytic relates to the rest of the discipline which it supports.
14. Hence, for example, the Prolegomena describes both the causal principle and the principle of the permanence of substance as “metaphysical” (pp. 273, 378). Cf. Von Schön Metaphysics, p. 468: “Nature is the object of all possible experience. Here again I can have principles which are not borrowed from experience and these belong to metaphysics, e.g., that all changes have their cause.”
15. This helps to explain why Kant announces at Critique, A247 / B303 that “the . . . name of an Ontology [i.e., general ontology—M.N.F.] . . . must give place to the . . . title of an . . . Analytic of the pure understanding [i.e., Kant’s metaphysics of nature—M.N.F.].”
It is interesting to note that Kant’s project of reducing metaphysics to general ontology (or something like it) had some earlier precedents in Germany, for example in Stahl (see Wundt, Die deutsche Schulmetaphysik des 17. Jahrhunderts, pp. 171, 219–20, 223–7).
16. This motive behind Kant’s conception of the discipline is visible, for example, in a passage from the Von Schön Metaphysics where, after distinguishing between two types of a priori concept, those to which an object corresponds in experience and those to which it does not, Kant observes: “The aggregate of these pure concepts of the understanding to which an object in experience does correspond constitutes transcendental philosophy or ontology [i.e., metaphysics of nature—M.N.F.]” (p. 470).
17. Concerning a priority serving as a distinctive mark of metaphysics for the critical philosophy, see, e.g., Prolegomena, pp. 265–6. Concerning supersensuousness having done so for traditional metaphysics, and also concerning the distinction between a priority and supersensuousness, see esp. Prize Essay, pp. 316–19.
18. It deserves emphasis, because the point is commonly overlooked, that the term “a priori” means very different things for Kant depending on whether he is applying it to concepts or to principles. For one thing, when applied to concepts it signifies a certain sort of independence from sensations, whereas when applied to principles it signifies a certain sort of independence from experience (where experience is far from being the same as sensation).
19. Some textual evidence for this was presented in an earlier note (ch. 5, n.42).
20. Mrongovius Metaphysics, p.762: “All concepts of the understanding would mean nothing if the senses did not supply any objects and examples. If I, for example, explained as much as you like what a substance was, but did not know how to give any example, then all would be in vain . . . But it is quite true that the concepts of the understanding are not borrowed from the senses.”
21. Prize Essay, p. 317: “We . . . can distinguish the cognition a priori . . . which, though grounded a priori, can nevertheless find the objects for its concepts in experience, from that . . . whose object . . . lies beyond all bounds of experience.” Cf. p. 260: in considering that supersensuousness knowledge of which has traditionally been metaphysics’ goal, “We count as sensuous, though, not only that whose representation is considered in relation to the senses, but also in relation to the understanding, just so long as the understanding’s pure concepts are thought of in their application to objects of the senses, and hence for the purpose of a possible experience.” Also p. 319, where Kant accuses traditional metaphysics of failing to mark this distinction between different types of a priori concepts. Also, Von Schön Metaphysics, p. 470.
22. Von Schön Metaphysics, p. 470: “The aggregate of these pure concepts of the understanding to which an object in experience does correspond constitutes transcendental philosophy or ontology [i.e., metaphysics of nature—M.N.F.].”
23. Critique, B2–3.
24. Mrongovius Metaphysics, p.768: “There are two sorts of uses of the pure understanding: the immanent is namely that in which a priori cognitions have their objects in experience. In just the same way there are also a priori axioms which concern objects of experience—that is pure reason, but in its immanent use . . . And there is the transcendent use, namely that in which a priori cognitions do not have their objects in experience.” Cf. Prize Essay, pp. 317–19, where Kant again draws this distinction, and accuses traditional metaphysics of failing to mark it. At Critique, A764–5 / B792–3 Kant takes Hume to task for failing to mark the distinction as well: “We suppose ourselves to be able to pass a priori beyond our concept, and so to extend our cognition. This we attempt to do either through the pure understanding, in respect of that which is at least capable of being an object of experience, or through pure reason, in respect of such properties of things as can never be met with in experience. [Hume] did not distinguish these two kinds of judgments, as he yet ought to have done, but straightway proceeded to treat this self-increment of concepts, and, as we may say, this spontaneous generation on the part of our understanding and of our reason, without impregnation by experience, as being impossible.”
25. See Critique, B3–5.
26. Kant points out at Critique, A184–5 / B228 that he was the first to notice that the synthetic a priori principles involved “are valid only in relation to possible experience, and can therefore be proved . . . through a deduction of the possibility of experience.”
27. Thus Kant writes that “its principles are merely rules for the exposition of appearances; and the proud name of an Ontology that presumptuously claims to supply . . . synthetic a priori cognitions of things in general [i.e., things in themselves—M.N.F.] . . . must, therefore, give place to the modest title of a mere Analytic of pure understanding” (ibid., A247 / B303).
28. Concerning this, see esp. Critique, A832–51 / B860–79; Von Schön Metaphysics, pp. 463–5; Prize Essay, p. 321
29. See esp. Critique, A726–37 / B754–65. Kant’s early rejection of this mathematical model as unsuited to metaphysics is most conspicuous in his Inquiry concerning the Distinctness of the Principles of Natural Theology and Morality from 1764.
30. Only “roughly” due to a modest exception which Kant explains at Critique, A71–2 / B97 concerning “infinite” judgments.
31. Note that Kant extends such systematicity beyond his “metaphysics of nature” as well, envisaging in addition a systematicity among the three transcendent ideas of the world as a whole, the soul, and God established by their one-to-one correlation with the three traditional forms of syllogism (see Critique, A333–8 / B390–6; Prolegomena, p. 330).
32. See Critique, Bxix, A841 / B869, A849–51 / B877–9. Cf. AA 17:259, 261, 495–6, 552–3, 558, 613.
33. Critique, A841–2 / B869–70. Cf. Prolegomena, p. 279, where Kant distinguishes from metaphysics this “transcendental philosophy”—an expression which he here uses in this sense (as, e.g., at Critique, A65–6 / B90–1), not, as often elsewhere, in the sense of metaphysics of nature (e.g., at Critique, A13–15 / B27–9;AA 29[1,2]:949, 970). Also, a 1783 letter to Garve in which Kant writes that “it is not at all metaphysics that the Critique is doing but a whole new science, never before attempted, namely, the critique of an a priori judging reason” (Philosophical Correspondence, p. 102). Also, Philosophical Encyclopedia, at AA 29(1,1):11–12.
CHAPTER SEVEN
DEFENSES AGAINST HUMEAN SKEPTICISM
1. See Critique, B19–20; Prolegomena, pp. 272–3.
2. Thus in connection with the puzzle of how a priori concepts can refer, Kant draws a distinction between two sides of the section of the Critique which addresses the puzzle, the Transcendental Deduction, implying that one side proves that the a priori concepts treated there refer, while the other side explains the possibility of their being employed by us to refer: “The one [side] . . . is intended to expound . . . the objective validity of [the pure understanding’s] a priori concepts . . . The other [side] seeks to investigate the pure understanding itself, its possibility . . . and so deals with it in its subjective aspect” (Critique, Axvi–xvii). Recognizing Kant’s dual strategy is therefore essential for properly understanding this famous distinction between an “objective” and a “subjective” side of the Transcendental Deduction (though it should be noted that he here runs together with the distinction between these two strategies certain further distinctions).
Similarly, in relation to the puzzle of synthetic a priori knowledge, Kant promises both an explanation of the possibility of such knowledge and a proof of the truth of particular cases of it: the Critique “enables us to explain how there can be cognition a priori; and in addition, to furnish satisfactory proofs of the laws which form the a priori basis of nature” (Critique, Bxix).
Kant’s distinction at Prolegomena, pp. 274–6 between the partly “synthetical” method of the Critique and the exclusively “analytical” method of the Prolegomena is again in large part a distinction between the proving that and the explaining the possibility sides of his strategy respectively. Thus he says there that the analytical method, in contrast to the synthetical, does not answer the question “whether [synthetic a priori cognition] is possible . . . but how it is possible” (p. 275; cf. p. 276).
3. By contrast, the corresponding material in the Prolegomena develops only the “analytical,” or explaining the possibility, side of the strategy, due to the work’s more popular function (see pp. 274–5).
Again by contrast, but this time for more principled reasons, in the Transcendental Aesthetic of the Critique, the a priori intuitions/concepts of space and time and the a priori concepts and synthetic a priori principles of mathematics are considered by Kant to be so secure that they can (and probably also must) dispense with any special proof of reference or truth, and instead receive only an explanation of their possibility as cognition (see Critique, A149 / B188–9, A261 / B316–17, A732–3 / B760–1). The main reason for this asymmetry to the advantage of mathematics over metaphysics lies, according to Kant, in the role of a priori intuition as a sphere for expounding and guaranteeing the validity of the former but not of the latter (see, e.g., ibid., A87–8 / B120, A732–3 / B760–1; Mrongovius Metaphysics, p. 765).
4. For example, Kant says of the proving that or “objective” side of the Transcendental Deduction’s defense of the a priori concepts of the understanding: “The objective validity of the categories as a priori concepts rests . . . on the fact that . . . through them alone does experience become possible” (Critique, A93 / B126; cf. A217 / B264 on the Analogies of Experience).
5. Cf. R.C.S. Walker, Kant (London: Routledge and Kegan Paul, 1978), p. 10. The particular type of experience involved varies from argument to argument. For example, in the Transcendental Deduction it is experience simpliciter, in the First Analogy experience of objective succession or co-existence, in the Second Analogy experience of objective events, and in the Third Analogy experience of the co-existence of objects in space.
6. Kant’s summary of the argument of the Third Analogy is one of his clearer expressions of this whole model: “Substances must stand . . . in dynamical community if their co-existence is to be cognized in any possible experience. Now, in respect to the objects of experience, everything without which the experience of these objects would not itself be possible is necessary. It is therefore necessary that all substances in the [field of] appearance, so far as they co-exist, should stand in thoroughgoing community of interaction” (Critique, A212–13 / B259–60).
7. Much of the voluminous secondary literature on “transcendental arguments” has in mind a significantly different type of argument: one concerned with conditions of the possibility—not of experience, or empirical knowledge but—of meaning. This type of argument does not play any central role in Kant. However, as we shall see in chapter 12, something like it does play an implicit role in his views about formal logic.
8. Revealing this motive, Kant says in the Mrongovius Metaphysics of 1783, for example: “Must there not be certain synthetic judgments a priori through which the synthetic judgments a posteriori are possible? And they were certainly true because they provide the foundation of experience, and experience is true . . . We will show that [the a priori axioms on which the possibility of experience rests] are certain because experience is certain and experience rests upon them . . . The a priori proposition which precedes all experience is certain, for what is more certain than experience?, and experience is certain only to the extent that the a priori proposition permits” (pp. 794, 799, 805).
9. Transcendental idealism was originally motivated by reflection on, and applied to, the former subject matter (space, time, and mathematics), namely in the Inaugural Dissertation of 1770. It was only later extended to cover the latter subject matter as well (the a priori concepts of the understanding and the metaphysical synthetic a priori principles associated with them).
10. Thus, in the Prolegomena Kant says that he realized that “they were not derived from experience, as Hume had tried, but sprang from the pure understanding” (p. 260). And in the Critique he begins the Transcendental Analytic with a declaration that the task of transcendental philosophy is “to investigate the possibility of concepts a priori by looking for them in the understanding alone, as their birthplace,” to “follow up the pure concepts to their first seeds and dispositions in the human understanding” (A66 / B90–1).
The Metaphysical Deduction’s tracing of the twelve a priori concepts of the understanding back to the twelve logical forms of judgment, which immediately follows these remarks in the Critique, is meant to establish the central thesis of their origin in the understanding. Thus Kant says later in the text: “In the metaphysical deduction the a priori origin of the categories has been proved through their complete agreement with the general logical functions of thought” (B159).
Ironically, Kant had already possessed this solution to Hume’s problem, in its essentials, before he took the problem itself deeply to heart in or shortly after 1772. Thus, the Inaugural Dissertation of 1770 already asserts the origin of these concepts in the understanding or intellect, as opposed to sensibility (AA 2:395), and that assertion is in effect repeated by the letter to Herz of 1772, which in addition already sketches an early version of the project of their Metaphysical Deduction (Philosophical Correspondence, p. 73). This chronological antecedence of the solution to its problem in Kant’s development may seem strange, but is not really so surprising on reflection; one factor which quite often makes a problem come to seem interesting to an inquirer is a sense that he has the resources for solving it.
11. See Critique, A243 / B301; cf. Volckmann Metaphysics, p. 404. This is Kant’s considered solution to Hume’s problem concerning the origin of the concept of causal necessity. It should be noted, however, that Kant sometimes muddies these relatively clear waters by also equating causal necessity with the necessity of the causal principle (Critique, B4–5, A766–7 / B794–5), or even with the necessity of a particular causal law (A91 / B123–4).
12. Critique, A93 / B126: “through them alone does experience become possible.” This statement should now be read in a causal or quasi-causal sense, and while keeping in mind Kant’s position that “the a priori conditions of a possible experience in general are at the same time conditions of the possibility of objects of experience” (ibid., A111; cf. A158 / B197; also Dohna Metaphysics, p. 654).
13. Thus at Critique, A92–3 / B124–6 Kant writes: “There are only two possible ways in which synthetic representations and their objects can establish connection, obtain necessary relation to one another, and, as it were, meet one another. Either the object alone must make the representation possible, or the representation alone must make the object possible [i.e., the two possibilities which Kant had already recognized in the letter to Herz—M.N.F.]. In the former case, this relation is only empirical, and the representation is never possible a priori. This is true of appearances, as regards that in them which belongs to sensation. In the latter case, representation in itself does not produce its object in so far as existence is concerned, for we are not here speaking of its causality by means of the will [i.e., as in the case of the sort of divine archetypal intellect envisaged by the letter to Herz—M.N.F.]. Nonetheless, the representation is a priori determinant of the object, if it be the case that only through the representation is it possible to cognize anything as an object . . . Concepts of objects in general . . . underlie all empirical cognition as its a priori conditions. The objective validity of the categories as a priori concepts rests, therefore, on the fact that, so far as the form of thought is concerned, through them alone does experience become possible. They relate of necessity and a priori to objects of experience, for the reason that only by means of them can any object whatsoever of experience be thought” (cf. A95–7, 128–30; Bxvii–xviii, 166–7). See also Von Schön Metaphysics, p. 474: “How can I have a priori concepts of things before they have presented themselves to me? The categories are concerned solely with objects of possible experience and the categories are valid precisely because they contain the conditions under which experience first becomes possible. Consequently, what contains the ground of the possibility of experience also contains the ground of the possibility of the object.”
14. Accordingly, Kant writes towards the end of the Transcendental Deduction, as he is about to embark on the Schematism chapter, that “how [the a priori concepts of the understanding] make experience possible . . . will be shown more fully in the following chapter” (Critique, B167).
15. See ibid., A137–47 / B176–87, A240–6 / B300–2.
16. See ibid., A144 / B183, A243 / B301.
17. For classic statements of this thesis, see ibid., A113–14, 125–6, Bxvi–xviii. Note that the two motives behind the thesis oftranscendental idealism which I have in effect just identified—explaining the reference of a priori concepts and explaining the possibility of synthetic a priori knowledge—are complemented by a third: explaining the necessity of synthetic a priori knowledge. This third motive was actually the one that first induced Kant to develop the thesis, namely in the Inaugural Dissertation (the other two motives only coming into play later). During the critical period Kant often tends to run this third problem concerning necessity together with the problem of how synthetic a priori knowledge is possible—something that is easy for him to do due to the intimacy of the connection that he sees between necessity and a priority (see ibid., B3–4). However, it is really a distinct problem—an ontological problem in contrast to an epistemological one.
CHAPTER EIGHT
DEFENSES AGAINST PYRRHONIAN SKEPTICISM
1. Critique, B19.
2. Ibid., B22.
3. See ibid., A423–4 / B451–2, A740 / B768.
4. I shall here simplify what could easily become a long story in its own right. Kant’s strategies for resolving the Antinomies are rather unsettled even within the Critique, and that is even truer of the whole history of his attempts to resolve them (including such earlier works as the Inaugural Dissertation of 1770 and such later ones as On a Discovery of 1790).
5. Critique, A506 / B534. In the “Mathematical” Antinomies each side’s proof of its principle takes the “apagogical” form of (1) a disproof, by reductio ad absurdum, of the principle asserted on the other side, from which (2) an inference to the truth of the principle to be proved is then drawn in light of the two principles’ apparent exhaustion of the logical possibilities. Kant in each case accepts that step (1) is compelling, but (as we are about to see) not that step (2) is.
6. Kant sometimes—for instance, prominently in the Prolegomena, and at Critique, A740 / B768—develops his diagnosis of these Antinomies in terms of a more specific claim that the concept of the whole spatio-temporal world as it occurs in them is actually self-contradictory.
7. See Critique, A502–6 / B530–4, A740 / B768, A792–3 / B820–1; Prolegomena, pp. 341–2.
8. See Critique, A529–37 / B557–65, A559–65 / B587–93; Prolegomena, pp. 343–7.
9. Critique, B22–3.
10. Ibid., B22. Cf. A476–81 / B504–9, A761 / B789.
11. Cf. ibid., A761–4 / B789–92, A768–9 / B796–7.
12. Ibid., Axv. Another route to the same interpretive conclusion is to note that because these conditional propositions involve necessity, they must in Kant’s eyes be a priori (ibid., B3–4), and that he believes a priori knowledge quite generally to be apodeictically certain: “Everything that is to be cognized a priori is thereby announced as apodeictically certain” (Prolegomena, p. 369; cf. Critique, A775 / B803, A822–3 / B850–1.
13. Recall Blomberg Logic, p. 214: “People . . . blamed Pyrrho for doubting the truth of all empirical judgments and not placing faith in them. However, this is a pure invention which has no basis at all.”
14. Recall Kant’s implication at ibid., p. 214 that for Pyrrho there are “certain, so to speak, eternal principles of our reason [i.e., logic and mathematics—M.N.F.], which cannot be disputed at all.”
15. Here again, the fact that Kant clearly understands these explanations to be a priori points to the same interpretive conclusion.
16. Critique, Aix.
17. Prolegomena, p. 371.
18. Ibid., p. 383. Cf. Critique, A12 / B26: transcendental philosophy provides “a touchstone of the value, or lack of value, of all a priori cognition.”
19. In Kant’s view, once expelled, some of them will in fact go on to win their wars, though by means of practical justifications, and with a status short of knowledge, which radically distinguish them from metaphysical principles properly so called (the “postulates” of practical reason). Others, accordingly, will perish (their contraries). And still others will in a sense war on indefinitely, though the critical philosopher, recognizing that such questions are not rationally decidable, will refrain from pursuing them any further (Volckmann Metaphysics, p. 379: the critical philosophy leads to “complete rest” in connection with metaphysical disputes, “since our reason ceases to ask questions when she sees that this or that is no object for her”).
20. Critique, Axx, xiv; emphasis added. Cf. Bxxiii–xxiv, B89–92. Also, Metaphysical Foundations of Natural Science, AA 4:473: “In all that is called metaphysics the absolute completeness of the sciences may be hoped for . . .; and therefore just as in the metaphysics of nature in general, so . . . also the completeness of the metaphysics of corporeal nature may be confidently expected.”
21. Prize Essay, p. 321: “Metaphysics distinguishes itself quite especially among all sciences by being the only one which can be presented completely fully, so that nothing more remains for posterity to add or [with which] to extend metaphysics in its content, indeed so that if the absolute whole does not emerge systematically from the idea of metaphysics, its idea can be considered as having not been correctly understood.” Prolegomena, p. 365: “A critique of reason must itself exhibit the whole stock of a priori concepts, their division in accordance with their various sources—sensibility, understanding, and reason—and further, a complete table of them and the analysis of all these concepts together with everything which can be derived from them, and then especially the possibility of synthetic a priori cognition by means of the deduction of these concepts, the fundamental principles of their use, and finally the bounds of their use, but it must do all this in a complete system.”
22. See esp. Critique, A832–3 / B860–1. Also A403: “in order to show the systematic interconnection of . . ., and so to show that we have them in their completeness . . .” Kant is especially explicit about this strategy in relation to the a priori concepts of the understanding. For example, at Critique, A64–5 / B89 he responds to the challenge of demonstrating their completeness with the following declaration of strategy: “When a science is an aggregate brought into existence in a merely experimental manner, such completeness can never be guaranteed by any kind of mere estimate. It is possible only by means of an idea of the totality of the a priori cognition yielded by the understanding; such an idea can furnish an exact classification of the concepts which compose that totality, exhibiting their interconnection in a system” (cf. A66–7 / B91–2, A80–1 / B106–7; and esp. Prolegomena, p. 322).
Incidentally, Kant’s possession of this strategy vis-à-vis the a priori concepts of the understanding has often been overlooked by the secondary literature. For example, Koerner accuses Kant of failing to demonstrate the uniqueness of his categorial schema, and thereby overlooks both the fact that Kant was acutely aware of the need to do so, and the fact that he had this strategy for doing so (S. Koerner, “The Impossibility of Transcendental Deductions,” The Monist, no. 51 [1967]).
23. To illustrate this point with the help of a homely example, a child might be able to show that he had an entire system of Lego pieces by constructing, say, a toy helicopter from them (like the one pictured on the box), but an inference from that fact to his having a complete collection of all Lego pieces would by no means be trivial; indeed in this case it would be massively mistaken.
24. “In order to know whether a type of cognition is completed one must already have a concept of the whole; and so any science must constitute a system, a whole in accordance with an already discovered idea . . . The whole of our pure cognition a priori or of our pure rational cognition is metaphysics, or it is a system of pure rational cognition from mere concepts . . . There lies in human cognition the wherewithal for a system, a system of pure concepts, i.e., for metaphysics: there is hence a system of pure rational cognition a priori . . . If I use my reason merely a priori through concepts, then experience does not guarantee my cognition, since a mere rational concept tolerates no admixture of experience and so here there is more danger that I go wrong, since no experience supports me. Such a science will often come upon things which tolerate no application in experience, and so we lose an important touchstone of truth; we can no longer test our judgment by experience and it is easy to see that consequently the use of pure reason becomes very slippery [i.e., the equipollence problem—M.N.F.]. Given this danger which pure reason constantly runs of going astray, we see very easily that it is necessary to examine in advance (1) whether there really is, then, such a pure rational faculty [i.e., it is necessary to explain how such concepts can refer and such principles constitute knowledge—M.N.F.], (2) whether this reason provides us with any genuine cognition [i.e., it is necessary to prove that specific such concepts refer and specific such principles are true—M.N.F.], and (3) whether we can determine the scope and boundaries of pure reason . . . Without an ability to determine whether pure reason can judge, without knowing its boundaries and scope, nothing protects us from error; we in that case fall into illusions and chimeras without recognizing them as such. And that is precisely why people could challenge all its propositions [i.e., the equipollence problem—M.N.F.] before there was a critique of reason” (Von Schön Metaphysics, pp. 463–5).
25. Critique, A832 / B860: “By an architectonic I understand the art of constructing systems.”
26. Kant certainly also has additional motives for wanting to demonstrate the entire systematicity of his new metaphysics besides the one under discussion here. For example, Lambert had argued that the consistency and harmony characteristic of an entire system made it a criterion of truth (see Beck, Early German Philosophy, p. 407). And there is a strong element of such a position in Kant as well. Thus, we find him arguing in the Prize Essay of 1791 that “when a system is so constituted that, first, each principle in it is independently provable, and second, if one had any concern about its correctness, still even as a mere hypothesis it leads inevitably to all the remaining principles of the system as consequences, then nothing more can be required in order to recognize its truth” (p. 311; it is of course the second point that is pertinent here; cf. Critique, A65 / B90, A647 / B675; also, Kant’s lectures on metaphysics, which in places almost sound as though they are advancing a coherence theory of truth; also, Kant’s remark in his 1799 open letter on Fichte’s Wissenschaftslehre that so far from having meant the Critique to be only a propaedeutic rather than “the actual system of [transcendental] philosophy . . . I took the completeness of pure philosophy within the Critique of Pure Reason to be the best indication of the truth of my work” [Philosophical Correspondence, p. 254]). Again—though largely because of the two motives already mentioned—Kant stipulates that entire systematicity is required for scientificness: “systematic unity is what first raises ordinary cognition to the rank of science” (Critique, A832 / B860; cf. Von Schön Metaphysics, p. 463: “A system of a type of cognition is called a science . . . In order to know whether a type of cognition is complete one must already have a concept of the whole; and so any science must constitute a system”).
27. Critique, A64–81 / B89–107; Prolegomena, pp. 322–6; Prize Essay, pp. 271–2; Volckmann Metaphysics, p.400.
28. Critique, A148 / B187, A161 / B200. Kant gives a helpful tabular summary of all these correspondences at Prolegomena, pp. 302–3.
29. Critique, A243 / B301. Cf. Volckmann Metaphysics, p. 404.
30. Critique, Bxiv.
CHAPTER NINE
SOME RELATIVELY EASY PROBLEMS
1. “The proud name of an Ontology [i.e., general ontology—M.N.F.] . . . must . . . give place to the modest title of a mere Analytic of the pure understanding [i.e., Kant’s metaphysics of nature—M.N.F.]” (Critique, A247 / B303). In the Prize Essay Kant simply calls it ontology: “Ontology is that science (as part of metaphysics) which is constituted by a system of all concepts and fundamental principles of the understanding, but only so far as they refer to objects which can be given to the senses, and can therefore be validated by experience” (p. 260). He generally does so in the lectures on metaphysics as well.
2. Strictly speaking, one should say here, more cautiously, “the apparent etymologically derived force.” For it is not clear that the word bore this sense in its original application to Aristotle’s work, rather than, more mundanely, “after the Physics” (i.e., after the latter work of Aristotle’s in some cataloguing scheme). In the lectures on metaphysics, Kant shows that he is fully aware of this alternative way of interpreting the expression, but consistently rejects it in favor of the interpretation assumed here (AA 29[1]:174, 381–2, 468; 28[2,1]:616;29[1,2]:773, 947).
3. Kant makes this point at Prize Essay, pp. 315–16.
4. Hence, for example, we read in the Prolegomena: “As concerns the sources of metaphysical cognition, its very concept implies that they cannot be empirical. Its principles (including not only its basic propositions but also its basic concepts) must never be derived from experience. It must not be physical but metaphysical knowledge, i.e., knowledge lying beyond experience. It can therefore have for its basis neither external experience, . . . nor internal . . . It is therefore a priori cognition” (pp. 265–6).
5. “No one attempts to establish a science unless he has an idea upon which to base it. But in the working out of the science the schema, nay even the definition, which, at the start, he first gave of the science, is very seldom adequate to his idea. For this idea lies hidden in reason, like a germ in which the parts are still undeveloped and barely recognisable . . . Consequently, . . . we must not explain and determine [sciences] according to the description which their founder gives of them, but in conformity with the idea which, out of the natural unity of the parts that we have assembled, we find to be grounded in reason itself. For we shall then find that its founder, and often even his latest successors, are groping for an idea which they have never succeeded in making clear to themselves, and that consequently they have not been in a position to determine the proper content, the articulation . . ., and limits of the science” (Critique, A834 / B862; cf. Prize Essay, p. 342, where Kant applies this model to metaphysics specifically).
6. Consider especially T. S. Kuhn’s observations on the role of “paradigm shifts” in the development of sciences in his important and influential book The Structure of Scientific Revolutions (Chicago: The University of Chicago Press, 1970).
CHAPTER TEN
A METAPHYSICS OF MORALS?
1. Critique, Bxliii, A841–2 / B869–70, / A850 / B878.
2. Ibid., A842 / B870; cf. Prolegomena, pp. 278–9.
3. See Critique, A14–15 / B28–9, A54–5 / B79. I say that this is questionable because it is difficult to see why empirical concepts would have to be any more essentially or deeply involved in the “metaphysics of morals” as Kant officially conceives it than in his “metaphysics of nature.”
4. See, for example, ibid., Bxiv, Bxxx, A744–5 / B772–3, A805 / B833, A828–9 / B856–7; Prolegomena, pp. 278–9, 371.
5. For example, we read at Critique, A14 / B28 that “the highest principles and fundamental concepts of morality are a priori Erkenntnisse,” and at Bxxi that “when all progress in the field of the supersensible has . . . been denied to speculative reason, it is still open to us to inquire whether, in the practical Erkenntnis of reason, data may not be found sufficient to determine reason’s transcendent concept of the unconditioned, and so enable us . . ., though only from a practical point of view, to pass beyond the limits of all possible experience.” Cf. A633–4 / B661–2. Also, Prize Essay, pp. 281, 310.
6. See Critique, Bxix–xxii, Bxxx–xxxi, A744–5 / B772–3, A823 / B851, A828 / B856. Cf. Pölitz Metaphysics, AA 28(1):304.
7. For example, in the important passages at Critique, A50 / B74 where Kant presents his fundamental principle that all knowledge requires both concepts and intuitions, the word that he uses for knowledge is throughout Erkenntnis. (There is therefore some ground for N. Kemp Smith’s usual practice of translating both words as knowledge. However, this conflation of the two words is for the most part extremely confusing, and has therefore been avoided in the present essay, where only Wissen has been translated as knowledge, but Erkenntnis instead as cognition.)
8. Kant’s clearest account of this generic sense of the word Erkenntnis occurs at Critique of Judgment, AA 5:467 ff., where he explains that “Erkennbare things are threefold in kind: things of opinion (opinabile), things of fact (scibile) [i.e., knowable, scio being Latin for to know—M.N.F.], and things of faith [Glaubenssachen] (mere credibile).”
Kant also uses the word Erkenntnis in certain further senses which are less relevant here. For example, in one of its senses it is not restricted to judgments at all, but also includes such truth-valueless items as concepts and intuitions (see, e.g., Critique, A320 / B376–7).
9. Logic, AA 9:70; cf. Critique, A822 / B850.
10. Logic, p. 66; cf. Critique, A820 / B848, A822 / B850.
11. Logic, p. 66;cf. Critique, A822 / B850. Two further points to note: (1) Kant also says that whereas Wissen is certain, Glaube is uncertain—but since he does not seem to stick to the latter half of this position consistently, I shall bracket it here. (2) He distinguishes Glaube from mere Meinung, which is both objectively and subjectively insufficient.
12. “Even after reason has failed in all its ambitious attempts to pass beyond the limits of all experience, there is still enough left to satisfy us, so far as our practical standpoint is concerned. No one, indeed, will be able to boast that he knows [wisse] that there is a God, and a future life; if he knows this, he is the very man for whom I have long [and vainly] sought. All knowledge [Wissen], if it concerns an object of mere reason, can be communicated; and I might therefore hope that under his instruction my own knowledge would be extended in this wonderful fashion. No, my conviction is not logical, but moral certainty; and since it rests on subjective grounds (of the moral sentiment), I must not even say, ‘It is morally certain that there is a God, etc.’ but ‘I am morally certain, etc.’ In other words, belief in a God and in another world is so interwoven with my moral sentiment that as there is little danger of my losing the latter, there is equally little cause for fear that the former can ever be taken from me” (Critique, A828–9 / B856–7; note that moral sentiment had played a fundamental role in Kant’s moral theory in Dreams of a Spirit Seer, and that it still plays a significant, though far more qualified, role in the Metaphysics of Morals of (1797). Cf. Pölitz Metaphysics, p. 304.
13. See Dreams of a Spirit Seer, pp. 372–3.
14. See Critique, Bxxviii, B29, A548 / B576, A800–2 / B828–30, A807 / B835, A841 / B869; Groundwork of the Metaphysics of Morals (1785), at AA 4:442–3; Metaphysics of Morals 1797), at AA 6:375–7. Kant had indeed already officially taken this step from a sentimentalist to a cognitivist theory in the Inaugural Dissertation of 1770, at AA 2:396.
15. Thus at Critique, A54–5 / B79 Kant represents pure morals as analogous to general logic. The analogy lies mainly in the facts that (1) the categorical imperative is an a priori principle which is valid for all reason as such, not merely for all human reason (see Groundwork of the Metaphysics of Morals, pp. 411–12), i.e., just like general logic, and (2) the categorical imperative appeals to the criterion of avoiding contradiction (in a maxim under universalization), i.e., just as the logical law of contradiction does.
The analogy does have certain limits in Kant’s view, in particular because he conceives the categorical imperative to be synthetic a priori, like transcendental rather than general logic (see ibid., p. 420). However, these are not limits that make it any less cognitive in nature.
16. It is true that Kant is reluctant to say that logical insight itself is either objective or a matter of Wissen: “General logic . . . abstracts from all content of cognition, that is, from any reference of it to the object, and treats only the logical form in the relation of any bit of cognition to another” (Critique, B79). However, he could hardly appeal to this fact in order to support his position at Critique, A828–9 / B856–7. For there is a world of difference between denying that morality is Wissen because it is merely a matter of sentiment and doing so because it is like logic—roughly, the difference between saying that it is less than Wissen and saying that it is more than Wissen. Accordingly, if he were to appeal to this fact in order to justify denying that morality, and hence whatever else rests upon it, is Wissen, and therefore excluding them from metaphysics “in its strict sense,” the appropriate response would be that this fact would not show morality, and hence whatever else rests upon it, to fall short of metaphysics “in its strict sense,” but instead to be metaphysics in an even stricter sense than mere Wissen can be.
17. Perhaps because of this problem, Kant elsewhere at about this time seems to cast around for alternative ways of justifying his denial to morality—and hence to whatever else is based upon it as well—of the status of genuine Wissen(schaft). But without greater success. For instance, in the Mrongovius Metaphysics of 1783 he writes concerning moral principles: “Not every natural use of reason can be turned into a scientific [wissenschaftlichen] one . . . For example, in morality one cannot represent the rule in abstracto, but merely in concreto. It is, for example, hardly possible to prove a priori or to show in abstracto the impermissibility of lies; in concreto it can indeed be done” (p. 782). His idea here seems to be that since the categorical imperative is a principle for testing individual maxims (or intentions), it can only show the moral impermissibility of general classes of maxims via showing the moral impermissibility of individual ones (e.g., it can only show the moral impermissibility of intending to lie in general via showing the moral impermissibility of my intending to lie to Mary in order to prevent her from disliking me, the moral impermissibility of Smith’s intending to lie to Jones in order to avoid paying him, and so on). However, it is unclear why Kant thinks that such a characteristic of cognition should disqualify it from counting as Wissen(schaft). Moreover, claiming that it did so would carry some very awkward consequences for him, such as, for example, that empirically established laws of nature cannot count as Wissen(schaft) either (since they too are established, not directly, but via particular instances). So here again he seems in the end to be left without any good rationale for denying that morality—and hence whatever else is based upon it as well—is Wissen(schaft), and so for denying that it belongs to metaphysics “in its strict sense.”
18. In the Groundwork of the Metaphysics of Morals the word Erkenntnis predominates throughout. There is one reference to moral Wissen (p. 457), but it is ambiguous, and in effect soon taken back (p. 462). There are also a couple of references to a Wissenschaft of morals (pp. 388, 405), but they are not prominent or emphatic.
19. There is an exception at Critique of Practical Reason, p. 91, where the word Wissenschaft is used, but the text goes on to back away from this, implying that instead only a love of science is involved (pp. 108–9; cf. 150–1).
20. Ibid., p. 163 .
21. Ibid., p. 4: “Freedom is . . . the only one of the ideas of speculative reason of which we know [wissen] the possibility a priori, without though understanding it, because it is the condition of the moral law, which we know [wissen].”
22. Critique of Judgment, p. 468: “And, what is very remarkable, there is one rational idea . . . which also comes under things of fact. This is the idea of freedom, whose reality . . . may be exhibited by means of practical laws of pure reason, and conformably to this, in actual actions . . . This is the only one of all the ideas of pure reason whose object is a thing of fact and to be reckoned under the scibilia.”
23. Ibid., pp. 469–71.
24. Ibid., pp. 469–71. It is interesting to note that this asymmetry in the treatment of the postulates reverses one that predominated in the Critique, where transcendent freedom was said to be only tenuously required for morality, but God and immortality strictly so (A803–13 / B831–41; Kant already changed his mind by the time of writing the preface of the second edition, however—see Bxxvii–Bxxx).
25. It is not clear that Kant consistently adheres to this new position in the writings of the 1790s. The Prize Essay of 1791 seems closer to the position of the Critique. The Metaphysics of Morals (1797) does characterize morality as Wissen and does envisage a Wissenschaft concerned with morality (AA 6:375, 410–13), but does not contain an account of the postulates.
26. Cf. Critique, A551 / B579, including note a.
CHAPTER ELEVEN
FAILURES OF SELF-REFLECTION
1. This is clear, first, from the necessity which they involve (or, equivalently, their assertion that such and such is a condition of the possibility of experience). For Kant famously holds that a proposition’s necessity is a sufficient condition of its a priority (Critique, B3: “If we have a proposition which in being thought is thought as necessary, it is an a priori judgment”). Second, it is also clear from the fact that Kant evidently regards all of the (essential) contents of the Critique as a priori (ibid., Axv).
2. For example, A. C. Ewing, J. Bennett, and R.C.S. Walker have taken these propositions to be analytic, whereas L. W. Beck takes them to be synthetic.
3. See, for example, Prolegomena, pp. 294–7, 308, 312–13, 316–17, 320.
4. This point must not be confused with the quite different—and entirely fallacious—thought that if some proposition q is analytically implied by another proposition p, then q must itself be analytic. Counterexample: Smith is a bachelor, therefore Smith is unmarried.
5. See Critique, A202 / B247, B289, A259 / B315; Prolegomena, pp. 312–13.
6. See, for example, Prolegomena, 276–8: “Metaphysics stands or falls with the solution to this problem [How are synthetic propositions a priori possible?]; its very existence depends upon it. Let anyone make metaphysical assertions with ever so much plausibility, let him overwhelm us with conclusions; but if he has not first been able to answer this question satisfactorily, I have the right to say: this is all vain, baseless philosophy and false wisdom . . . All metaphysicians are . . . solemnly and legally suspended from their occupations until they shall have satisfactorily answered the question: How are synthetic cognitions a priori possible? For the answer contains the only credentials which they must show when they have anything to offer us in the name of pure reason. But if they do not possess these credentials, they can expect nothing else of reasonable people, who have been deceived so often, than to be dismissed without further ado.” Also, Critique, A3–4 / B7–8, A209–10 / B255, A233 / B285–6.
7. Any move to stop this infinite regress by resorting to circularity at some point would obviously involve vicious circularity.
8. To my best knowledge, I invented this criticism (some ten or fifteen years ago). It has been circulating in drafts of this essay for many years and was in particular present in the draft submitted to the press in 2005. I was recently shocked and saddened to discover an article published by another author in 2006 which contains what seems to be an abbreviated and garbled version of it, presented without the proper attribution. (The article also shows several further signs of familiarity with this essay, again unacknowledged.)
9. For example, at Prolegomena, p. 276, n. 6 he says that in giving these explanations “we often use nothing but synthetic propositions, as in mathematical analysis,” and his very inclusion of them in metaphysics and in the Critique marks them as a priori.
10. Consider again Kant’s strictures at Prolegomena, pp. 276–8 (as recently quoted in note 6); and also Critique, A3–4 / B7–8, A209–10 / B255, A233 / B285–6.
11. For this central doctrine, see, e.g., Critique, Bxx, xxv–xxvi, A42–3 / B59–60, A252–60 / B310–15. The latter of the two grounds for the doctrine mentioned here (Kant’s theory of schematism) is especially conspicuous in the Phenomena and Noumena chapter and at A286–9 / B342–6.
12. For complaints similar to these ones, cf. Beck, “Toward a Meta-Critique of Pure Reason,” in his Essays on Kant and Hume, and Walsh, Reason and Experience, ch. 9.
13. Critique, Axiv. Kant’s suggestion here of a Cartesian self-transparency of the mind is indefensible within the framework of the critical philosophy, at least if understood as a doctrine about the mind and its activities in themselves, as it would have to be in order to explain our knowledge of the thesis of transcendental idealism.
14. Ibid., Bxvi–xxii.
15. Ibid., Bxviii, note a.
16. Ibid., Bxvi–xvii.
17. Ibid., Axv; cf. A769–82 / B797–810.
18. Ibid., Bxviii, note a; cf. A769–82 / B797–810.
19. Ibid., Bxxii, note a. Cf. Prize Essay, p. 268.
20. Kant seems to imply here that our success in referring by means of the a priori concepts of his new “metaphysics of nature” is more certain than our knowledge of its synthetic a priori principles.
21. Critique, Bxviii, including note a; Bxx; A506–7 / B534–5.
22. The Antinomies would still be available as arguments in its support. However, it seems unlikely that they could plausibly bear the burden of proof by themselves. Indeed, Kant’s tendency to cast them in a merely supporting role in this connection amounts to acknowledging as much.
23. Critique, A69–70 / B94–5. Kant says the following at Mrongovius Metaphysics, p. 803 concerning the analogous case of the a priori concepts of the understanding: “In every system a gap reveals itself if even the slightest thing is missing, because an idea of the whole is the basis there. When we have a sand heap we do not see if a few grains are missing, but when we have a pyramid, then straightaway. Just this occurs when we sketch a system of the categories as well.” Cf. R. Brandt, Die Urteilstafel (Hamburg: Felix Meiner, 1991), p. 59.
24. Cf. Critique, A304 / B360–1.
25. See Brandt, Die Urteilstafel. Brandt also notes that there seems to be an intended correlation of I, II, and III with, respectively, the subject in a subject-predicate judgment, the copula, and the predicate. However, this idea seems much less promising. For example, how could it warrant the inclusion of hypothetical and disjunctive judgments under III?
26. Critique, A68–9 / B93–4.
27. The suggestion that this is all that Kant has to say in answer to the problem of horizontal exhaustiveness would be vigorously disputed by some interpreters, in particular by Klaus Reich and his followers. In his influential book The Completeness of Kant’s Table of Judgments (Stanford: Stanford University Press, 1992; originally published in German in 1932) Reich argues that Kant in the Critique implicitly has a far more elaborate strategy for establishing the horizontal exhaustiveness of his four groups of forms of judgment, namely one of showing that the existence of all and only these groups is implied by the very “definition” of a “judgment” (pp. 47, 102). However, as far as I can see, Kant actually has no such strategy, at least not in the Critique. This seems clear from the following facts (some further criticisms can be found in work by L. Krüger, R. Brandt, and M. Wolff): (1) As Reich himself concedes (p. 108), Kant nowhere explicitly advances such a strategy in the Critique. (2) Reich explains its absence from the work in terms of Kant’s statement at Critique, A13–14 / B27–8 that, although the work will completely enumerate all of the fundamental concepts which contribute to a priori knowledge, it will not exhaustively analyze them or review those which can be derived from them (pp. 108–9). This explanation is unconvincing, however. For in the passage in question Kant is merely saying that the work will exclude certain analyses which are not crucial for its position, not, as Reich’s explanation would require, analyses which are crucial for it. (3) Just before the section in the Critique which presents the logical forms of judgment, Kant says that the section will show that giving “an exhaustive statement of the functions of unity in judgments . . . can quite easily be done” (A69 / B94). So Kant clearly thinks that his demonstration of their exhaustion is easy and evident in this section of the text, not a difficult and hidden demonstration such as Reich hypothesizes. (4) Kant in the Critique presents the fact that we have “just these and no other functions of judgment” as a paradigmatic example of a fact not “capable of further explanation” (B145–6). However, if Reich’s conception of the analytic derivability of all and only these functions of judgment from the very concept of a “judgment” were Kant’s, then the fact in question would be further explicable. (5) Kant seems to preclude Reich’s strategy as one that would be positively unacceptable to him in the following passage of the Critique: “No concept given a priori . . . can, strictly speaking, be defined. For I can never be certain that the clear representation of a given concept, which as given may still be confused, has been completely effected, unless I know that it is adequate to its object. But since the concept of it may, as given, include many representations, which we overlook in our analysis, although we are constantly making use of them in our application of the concept, the completeness of the analysis of my concept is always in doubt . . . Instead of the term definition, I prefer to use the term exposition, as being a more guarded term, which the critic can accept as being up to a certain point valid, though still entertaining doubts as to the completeness of the analysis . . . Philosophical definitions are never more than expositions of given concepts . . . [which] can be obtained only by analysis (the completeness of which is never apodeictically certain)” (A728–30 / B756–8). One problem which this passage poses for Reich’s interpretation concerns his use of the notion of a “definition” of judgment, for the passage shows that this would be unacceptable to Kant. Another, and more serious, problem is that the passage seems clearly to reject the idea, central to Reich’s whole strategy, that the analysis of an a priori concept such as “judgment” can establish the exhaustiveness of the elements resulting from the analysis. In sum, Reich’s strategy is not Kant’s.
M. Wolff, Die Vollständigkeit der kantischen Urteilstafel (Frankfurt am Main: Vittorio Klostermann, 1995) develops a very sophisticated variant of Reich’s approach. However, it seems vulnerable to similar objections. (1) Wolff rightly faults Reich for locating the argument which he attributes to Kant outside of the Critique. But Wolff’s own attempt to excavate it from the short stretch of text at A67–70 / B92–5 involves such a massive interpolation of tacit premises and steps of inference as to be hardly less exegetically implausible (recall that Kant says at A69 / B94 that the relevant section of the Critique will show that giving a demonstration of the exhaustion of the functions of judgment “can quite easily be done”). (2) Wolff’s interpretation relies on a very implausible reading of Kant’s assertion at B145–6 that the fact that we have just these and no other forms of judgment is not “capable of further explanation.” On Wolff’s reading this in effect merely means not “capable of further explanation than the ambitious one I have already given” (Die Vollständigkeit, pp. 180–1). (3) In addition, some of the details of the argument that Wolff attributes to Kant are, to say the least, very surprising. For example, on Wolff’s interpretation “I Quantity” turns out to be concerned with differences in predicate-concepts (ibid., pp. 143–4), whereas it is clearly for Kant rather concerned with differences in subject-concepts (see, e.g., Kant’s Logic, p. 102).
28. Cf. Reich, The Completeness of Kant’s Table of Judgments, pp. 103 ff., which already develops the germ of the following interpretation.
29. Reflexion no. 5854 from the 1780s states: “There are therefore three logical functions [elementary and not derivative] under a certain head, and hence three categories also: two of the functions manifest the unity of consciousness as regards two opposites, but the third function mutually connects the consciousness again. No more kinds of the unity of consciousness can be thought. For there is (a) one consciousness which combines a manifold, (b) another consciousness which combines in an opposite way, and so (c) is the combination of (a) and (b)” (AA 18, no. 5854; cf. Critique, B110–11).
30. Thus in the Critique itself Kant writes in a similar spirit to Reflexion no. 5854 concerning the analogous case of the a priori concepts of the understanding, but adds this point about non-analyzability due to elementariness, or primitiveness: “It is significant that in each class the number of the categories is always the same, namely, three. Further, it may be observed that the third category in each class always arises from the combination of the second category with the first . . . It must not be supposed, however, that the third category is therefore merely a derivative, and not a primary, concept of the pure understanding. For the combination of the first and second concepts, in order that the third may be produced, requires a special act of the understanding, which is not identical with that which is exercised in the case of the first and the second” (B110–11).
31. Cf. Critique, A71 / B96, and B111 on the corresponding a priori concepts. (Since Kant correlates the universal form of judgment with unity and the singular with totality, he would himself express what I have here put in terms of totality rather in terms of unity. However, it seems very plausible to say that he gets the correlations the wrong way round. And in any case the difference does not affect my main point here much.)
32. Cf. ibid., B111 on the corresponding a priori concepts.
33. Concerning the non-analyzability here, cf. Walker, Kant, pp. 26–7.
34. Cf. Critique, A73–4 / B98–9.
35. In order to see the obstacles in the way of analysis here, it is important to realize that for Kant (1) disjunctive judgments are aut judgments rather than vel judgments, “p or q” here in effect meaning (p v q) & -(p & q), and (2) hypothetical judgments are not merely judgments of material implication (p ⊃ q as logically equivalent to -(p & -q)) but instead include a robust notion of consequence (Konsequenz).
36. See Critique, A74–5 / B100.
37. Ibid., B111: “Necessity is just the existence which is given through possibility itself.” Cf. B4, where Kant defines strict universality as a matter of thinking something “in such a manner that no exception is allowed as possible,” and says that in this sense of strict universality it and necessity “are inseparable from one another.”
38. Note in particular that the reality expressed here is arguably not quite the same as that expressed by an assertoric judgment, in that here it can be ascribed to merely possible cases (e.g., the necessity that the internal angles of a triangle sum to 180 degrees consists in the fact that all possible triangles—whether they are actual or merely possible—really have internal angles that sum to 180 degrees).
39. Notice, incidentally, that this interpretation of Kant’s strategy for establishing vertical exhaustiveness enables one to see it as the original inspiration of the dialectical method subsequently developed by Fichte and Hegel, who likewise employed it in the service of demonstrating entire systematicity and thence complete collection. In their dialectical method Fichte and Hegel reiterated Kant’s basic model of two opposites leading to a non-analyzable synthesis at successive levels, and then used circularity in the resulting chain as the criterion that an entire system had been achieved.
40. Critique, A74 / B99–100; cf. A219 / B266, A233–5 / B286–7.
41. Of course, an alternative rationale for including the modal forms of judgment might be found—for example, their fundamental role in the distinctive forms of inference treated by modal logic.
42. Critique, A126, A132 / B171.
43. Ibid., A106, A141–2 / B180–1.
44. Ibid., A132 / B171.
45. Ibid., A132–3 / B171–2.
46. The reduction would be: p vel q = (-p ⊃ q) & (-q ⊃ p). Such a reduction would only “arguably” be possible, because it would require accepting Kant’s apparent assumption that conjunction can somehow be taken for granted.
47. Cf. Brandt, Die Urteilstafel, p. 88. Brandt notes that Lambert had already included copulative judgments among the fundamental bases of inference (p. 102, n.5).
48. Critique, Bviii.
49. M. Wolff in his very thought-provoking book Die Vollständigkeit der kantischen Urteilstafel attempts to forestall this sort of objection by arguing (a) that for Kant general logic is restricted to propositions which only employ general terms, (b) that Kant himself distinguishes between such a general logic and special logics, such as that devoted to forms of inference peculiar to mathematics, and that Frege’s logic counts as an example of the latter rather than the former, and (c) that if Kant’s general logic omits certain forms of inference which Frege’s logic recognizes, the converse is also true. However, I think that one should be skeptical of this defense. Ad (a), even if Fregean quantificational logic is not restricted to propositions which employ only general terms (because it admits proper names), it still recognizes many forms of inference which do only employ general terms but which are not accounted for by Kant. Ad (b), while it is certainly true that Frege’s logic is especially adapted and suited to the needs of mathematics, it also captures many forms of inference involving multiple occurrences of quantification within a proposition which are implicitly used and recognized as valid in everyday discourse but which Kant’s general logic omits. Ad (c), the arguable fact that Kantian logic’s omission of forms of inference which are included by Fregean logic is complemented by a converse omission does nothing to diminish the former omission, which seems sufficient by itself to show that Kant’s table of forms of judgment is incomplete.
CHAPTER TWELVE
THE PYRRHONIST’S REVENGE
1. The representative of the Kantian reading of Pyrrhonism with whom Hegel directly took issue was the neo-Kantian self-proclaimed “skeptic” G. E. Schulze rather than Kant himself.
2. My qualification “virtually” flags the fact that Hegel makes one exception: in its highest form, ancient Pyrrhonism did not, according to him, attack the position of true philosophy. This exception need not concern us here, however.
3. See M. F. Burnyeat, “Can the Skeptic Live His Skepticism?” in The Skeptical Tradition. For an adjudication of this interpretive dispute in which I basically favor the Hegel-Burnyeat reading of Pyrrhonism over the Kant-Frede reading, see my “Hegelian vs. Kantian Interpretations of Pyrrhonism: Revolution or Reaction?” in Kritisches Jahrbuch der Philosophie, vol. 10 (2005).
4. Blomberg Logic, pp. 210, 213–16.
5. Ibid., p. 216.
6. Diogenes Laertius, Lives of Eminent Philosophers (Cambridge, Mass.: Harvard University Press, 1979), vol. 2, ch. 9, sec. 102: “The dogmatists answer [the Pyrrhonian skeptics] by declaring that the skeptics themselves do apprehend and dogmatize; for when they are thought to be refuting their hardest they do apprehend, for at the very same time they are asseverating and dogmatizing. Thus even when they declare that they determine nothing, and that to every argument there is an opposite argument, they are actually determining these very points and dogmatizing.”
7. Sextus Empiricus, Outlines of Pyrrhonism, bk. 1, chs. 1, 7, 11, 33. Of course, one might still raise questions about the tenability of this stance—as, for example, Burnyeat does in “Can the Skeptic Live His Skepticism?” But they will not be Kant’s questions, and so I shall not pursue them here.
8. I say “a fortiori” here both because judgments of objective experience have traditionally seemed, and perhaps really are, easier to attack than judgments of subjective experience—this was Hegel’s own reason for thinking in “a fortiori” terms—and because the former seem to presuppose the latter.
9. “No skeptic denies that there are in man intuitions, concepts, or ideas, or that these are distinct from one another. This is a matter of fact” (G. E. Schulze, Aenesidemus, oder über die Fundamente der von dem Herrn Prof. Reinhold in Jena gelieferten Elementarphilosophie, nebst einer Verteidigung gegen die Anmaßungen der Vernunftkritik [published anonymously and without specification of place of publication, 1792], pp. 100–1; cf. p. 45).
10. Thus Hegel writes that Pyrrhonism “through its turning against knowledge in general . . . finds itself, because it here opposes one thinking to another and combats the ‘is’ of philosophical thinking, driven likewise to overcome the ‘is’ of its own thinking” (G.W.F. Hegel, Jenaer Schriften [Frankfurt am Main: Suhrkamp, 1977], p. 248; cf. p. 254).
11. See, for instance, the Pyrrhonian “tropes of Aenesidemus” (as preserved in Sextus Empiricus, Outlines of Pyrrhonism and elsewhere).
12. Sextus Empiricus, Against the Logicians (Cambridge, Mass.: Harvard University Press, 1983), bk. 1, sec. 369.
13. Cf. Sextus’s attack on the existence of sense-impressions which follows immediately afterwards in Against the Logicians. This additional material both provides further examples of counterarguments against judgments of subjective experience and also addresses a possible objection to my interpretation here, namely that Sextus’s invocation of Democritus was intended in a less radical spirit than I imply, not as an attack on judgments of subjective experience but merely as an attack on judgments about the sensibly apparent objective world.
14. See, for example, Outlines of Pyrrhonism, bk. 1, ch. 10.
15. For example, Galen reports on the existence of a group of extreme Pyrrhonists who questioned even their own affections and whether or not we think (C. G. Kühn, Galeni Opera Omnia [Leipzig: C. Knobloch, 1821–33], 8:711, cf. 14:628; K. Deichgräber, Die Griechische Empirikerschule [Berlin: Weidmann, 1930], p. 133).
The evidence which I have just cited from Sextus and Galen seems to me to constitute a much better basis for a claim that the ancient Pyrrhonists levelled an equipollence attack even against subjective experience than the evidence to which Hegel himself tends to point in The Relation of Skepticism to Philosophy, namely the tropes of Aenesidemus.
16. According to Sextus, the skeptic does not refrain from all avowal, “for the skeptic assents to the feelings necessitated in sense-impression, and he would not, for example, say when feeling hot or cold ‘I think that I do not feel hot or cold’ ”; the appearance or sense-impression, “since [it] lies in feeling and involuntary affection . . . is not open to question” (Outlines of Pyrrhonism, bk. 1, chs. 7, 11; Bury’s translation slightly amended).
Similarly, the core of Descartes’s case against skepticism in the Meditations and elsewhere lies in an argument to the effect that one’s current subjective experience necessitates belief in and knowledge of it—that it is necessarily the case that if one is in some mental state S having a character C then one believes and knows this (Bernard Williams has aptly called this Descartes’s principle of evidence)—so that one’s current subjective experience is immune to skeptical doubt.
17. Descartes has a closely related but distinct principle which might also be supposed to pose problems for skepticism about one’s own current subjective experience: the principle that it is necessarily the case that if one believes oneself to be in a mental state S having a character C then one is in such a state and one’s belief constitutes knowledge. (Bernard Williams has aptly called this Descartes’s principle of incorrigibility; together with the principle of evidence, it constitutes what is sometimes known as Descartes’s conception of the mind’s self-transparency.) Someone might hope to counter a skeptic inclined to skepticism about his own current subjective experience by appealing to this principle, encouraging him to infer: “Since I at least believe that I am currently having subjective experiences X, Y, and Z, and since in addition the principle of incorrigibility is true, I must really currently be having subjective experiences X, Y, and Z.” However, this sort of response to the skeptic again fails to be compelling in the end, and for reasons similar to the first two reasons which just thwarted the argument from the principle of evidence: First, the principle of incorrigibility assumed here is again just the sort of dogmatic principle that the Pyrrhonian skeptic is adept at undermining through counterargument. One attractive step in this direction would be to point to the phenomenon of self-deception, for example. Second, the other essential premise, “I believe that I am currently having subjective experiences X, Y, and Z,” is again flagrantly question-begging in relation to such a skeptic. For belief is itself a subjective experience, and so precisely the sort of thing whose current existence in himself he is proposing to call into question.
18. Critique, B131. Cf. Kant’s assertion, quoted earlier, that when one is concerned with “reason itself and its pure thinking . . . to obtain complete cognition of these, there is no need to go far afield, since I come upon them in my own self” (ibid., Axiv).
19. Cf. ibid., B132.
20. AA 7:135: “The field of our sensible intuitions and sensations of which we are not conscious although we can without doubt infer that we have them, i.e., of obscure representations in man (and likewise in animals), is immeasurable, but on the other hand, those which are clear contain only infinitesimally few of their points which lie open to consciousness. On the great map of our mind only a few places are illuminated, so to speak.”
21. Thus, enumerating a set of propositions which he considers valid beyond all doubt, Schulze writes in his Aenesidemus that “the touchstone of all that is true is general logic, and every reasoning about matters of fact can lay claim to correctness only to the extent that it conforms to the laws of logic” (p. 45).
22. Hegel accuses Bouterwek of being unfaithful to his own principle of only going as far in his philosophizing as the skeptic would allow “to the extent that he [Bouterwek] erects . . . on the basis that doubting is itself a thinking, the whole system of laws of thought, as logic. For, on the contrary, the consistent skeptic denies the concept of a [logical] law altogether” (Jenaer Schriften, p. 141).
23. Hegel’s early interest in Bardili’s work is shown by a letter from Schelling to Fichte of May 24, 1801, and also by Bardili’s appearance in Hegel’s 1801 essay The Difference Between the Fichtean and Schellingian Systems of Philosophy.
24. Critique, A54 / B78.
25. Ibid., Bviii.
26. In K. L. Reinhold, Beiträge zur leichteren Übersicht des Zustandes der Philosophie beim Anfange des 19. Jahrhunderts (Hamburg: F. Perthes, 1801–3).
27. Ibid., p. 83.
28. “The Kantian school has itself demanded an appropriate metaphysics for everything that is supposed to be scientific in each kind of human knowledge, in which the connection of this piece of knowledge with its a priori grounds should be developed. Is only logic alone to do without such a metaphysics, and yet be, and be called, a science, indeed quite pure science?” (ibid., p. 85).
29. Thus, Bardili points out that “Kant thought he had recognized that the dogmatists and skeptics had remained standing with their quarrels only on the territory of metaphysics; and that on the territory of logic on the contrary an eternal peace has always prevailed” (ibid., p. 88). And he asks rhetorically in response to this view of Kant’s “whether it is really true that skepticism has only called into question the objective in human cognition, or has not on the contrary also more than once dared to attack the validity of the laws of our understanding themselves” (ibid., pp. 88–9).
30. C. G. Bardili sought to realize such a project in his Grundriß der ersten Logik, gereinigt von den Irrtümern bisheriger Logiken überhaupt, der Kantischen insbesonders [Outline of the First Logic Purified of the Errors of Previous Logics in General and of the Kantian in Particular] (Stuttgart: F.C. Löflund, 1800). Hegel sought to realize such a project in his Science of Logic with its—somewhat, and surely not coincidentally, similar—“total reconstruction” of logic (Science of Logic [New York: Humanities Press, 1976], p. 51), which, he says, has its “justification” in his Phenomenology of Spirit (ibid., p. 48).
31. Outlines of Pyrrhonism, bk. 2, secs. 134–203; Against the Logicians, bk. 2, secs. 300–481.
32. Cicero, Academica (Cambridge, Mass.: Harvard University Press, 1979), bk. 2, secs. 95–8. The word “lie” here should be understood in the broad sense “speak falsely” (a sense which Cicero’s mentior, like the Greek pseudomai which stands behind it, can readily bear). A simpler version of the paradox that works at least as well is: “This statement is false.”
It should be noted that on certain interpretations the laws of bivalence and excluded middle may be significantly different. However, I shall disregard the potential difference between them here.
33. Among the more serious proposals which the skeptic might exploit in this way, there have, for example, been various suggestions that the law of bivalence or excluded middle should be given up in order to solve perceived difficulties in mathematics (the intuitionists), in quantum mechanics, and in logic itself. For instance, in relation to logic, it has been proposed in response to Russell’s Paradox (in a manner strikingly reminiscent of the Academic skeptic’s argument) that Russell’s paradoxical sentence which says that the class of all non-self-membered classes is a member of itself should be coped with by instituting a three-valued logic in which the law of bivalence or excluded middle disappears and assigning this Russellian sentence the middle truth-value. Again, Hilary Putnam has proposed in response to problems arising within quantum mechanics a logic which, while retaining the law of excluded middle, gives up the distributive law: p & (q v r) ⊃ (p & q) v (p & r) (H. Putnam, “The Logic of Quantum Mechanics,” in his Mathematics, Matter and Method: Philosophical Papers, Volume 1 [Cambridge: Cambridge University Press, 1975]). Among the more hypothetical proposals which a skeptic might exploit are Wittgenstein’s suggestion in the Remarks on the Foundations of Mathematics that one could respond to Russell’s Paradox by dispensing with the law of contradiction in relation to the paradoxical Russellian sentence (L. Wittgenstein, Remarks on the Foundations of Mathematics [Oxford: Blackwell, 1978], pt. 4, par. 59), and Wittgenstein’s slightly less dramatic suggestion in the Philosophical Investigations that one could adopt a logic in which double negations were treated either as meaningless or else as equivalent to single negations, so that, for example, the law of double-negation elimination disappeared (Philosophical Investigations [Oxford: Blackwell, 1976], par. 554).
34. It might be objected here that applying the equipollence method to logic in such a manner would require holding at least one logical law firm: the law of contradiction. For does not the equipollence method presuppose the necessity of avoiding contradictions? However, this objection is dubious. For one thing, it seems open to the skeptic to reject particular contradictions that come along without doing so on the basis of a commitment to a general law proscribing contradictions.
35. The question of the epistemological security of logical principles has in general received rather scandalously little attention from philosophers, who have tended, instead, to show indecent haste in attempting to reduce other sorts of principles to logical ones, on the assumption that the latter were certain and that their certainty would thereby transfer to the former as well—as, for example, in Kant’s explanation of analyticity in terms of the law of contradiction, and Frege’s attempt to reduce arithmetic to logic.
36. Critique, A52 / B76. Cf Logic, p. 13: logic is a “science of the necessary laws of thinking without which no use of the understanding and of reason takes place at all.”
37. Critique, Bxxvi. Cf. Prize Essay, p. 325.
38. Critique, A291 / B348, emphasis added (Kant’s formulation in this sentence is of course paradoxical, but only in an easily remediable way).
It is therefore a mistake to say—as Walsh does at Reason and Experience, pp. 98–9, 218, and as several more recent commentators do too—that for Kant logic is purely prescriptive. It is prescriptive for him, but it is so in virtue of conformity to it being an essential feature of thought, so that failure to conform to it entails ceasing to think. It is true that he says in his published Logic that in logic the question is “not how we think, but how we ought to think” (p. 14). But the context of this remark shows clearly that what he means by it is that logic is not about how we contingently happen to think. Again, ascribing this position to him does naturally prompt the question how he can accommodate the sorts of deviations from logic which seem not only to occur in practice but also to be required, at least as possibilities, if the conception of the discipline as prescriptive is to make sense. However, this should be raised as a real (and interesting) question, not a rhetorical one.
39. As Béatrice Longuenesse has pointed out to me, there are two passages in the Critique which might be thought to imply a further epistemological defense or justification of classical logic: at A117 Kant writes that “the possibility of the logical form of all cognition is necessarily conditioned by relation to . . . apperception as a faculty” (cf. B131), and at B133–4 he gives an argument to the conclusion that “the synthetic unity of apperception is . . . that highest point to which we must ascribe all employment of the understanding, even the whole of logic.” As far as I can see, though, his point at A117 is merely that general logic presupposes the transcendental unity of apperception in the sense that, like all representations, those involved in logic must be ascribable to the self, and the point of his argument at B133–4 is merely that the general concepts which logic employs need to be abstracted from syntheses of representations of a sort that interdepend with such self-ascribability. In other words, there is no defense or justification of classical logic intended here. If there were, it would presumably just be that the transcendental unity of apperception, as essentially involving thought (cf. B157), essentially involves conformity to classical logic (i.e., it would merely be a variant of the defense which I have already attributed to Kant). Certainly, Kant’s idea cannot coherently be that we can justify classical logic by deducing it from the transcendental unity of apperception.
40. Pace some recent claims by H. Putnam and his followers that it was.
41. Metaphysics, 1005b–1007a.
42. I have of course chosen these examples in part because they are all cases of rather disciplined deviations from and rejections of the law. But these are the right sorts of cases to consider. It is easy enough to make cases of merely random deviations and rejections look like examples of non-belief and non-thought. But then, merely random utterances of any kind tend to look like that, whether they violate classical logical principles or not.
43. For such an objection, see J. Barnes, “The Law of Contradiction,” Philosophical Quarterly, vol. 19, no. 77 (1969), pp. 308–9.
44. Metaphysics, 1005b26–32.
45. Barnes embraces this line of argument, and elaborates on it at some length. But he too begs the original question, namely in assuming that belief and disbelief are contraries, so that it is necessarily true that, as he puts it, (x) ((xD:(P)) ⊃ (-xB:(P))), i.e., if anyone disbelieves (“D”) a proposition P then he does not believe (“B”) proposition P (“The Law of Contradiction,” p. 304).
46. Metaphysics, 1006a11–1007a35.
47. In other words, this would at best only constitute an argument for what Putnam has called the minimal principle of contradiction: not every statement is both true and false (H. Putnam, “There is at least one a priori truth,” in his Realism and Reason: Philosophical Papers, Volume 3 [Cambridge: Cambridge University Press, 1983]). Incidentally, Putnam’s own argument for such a principle is quite different from the one in question here. Note, in particular, that he would not be sympathetic to the appeal to analyticity made in the reconstruction suggested here.
48. G. F. Meier, Vernunftlehre (Halle: Johann Justinus Gebauer, 1752): logic is “a science that treats the rules that one must observe in order to think rationally” (p. 5; cf. p. 7); concerning the avoidance of contradictions, “It is contrary to the nature of our understanding to think otherwise. Who can imagine as possible that God exists and also doesn’t exist, that all men can err and some cannot err?” (p. 548); “When one at the same time affirms and denies one and the very same thing of one and the very same thing, one thinks nothing” (p. 567); “A simply impossible cognition [such as dry snow or wooden iron] is . . . no cognition at all, but it only seems to be a cognition” (pp. 131–2; cf. p. 484).
49. Metaphysics, 1005a15–1006a15.
50. Thus at Critique, A59–60 / B83–4 Kant says that general logic furnishes criteria for truth, in the sense of necessary conditions for truth. But his ground for saying this is not an Aristotelian (or Fregean) assumption that logical laws hold for all being as such. Rather, it is that “whatever contradicts [widerspricht] these rules is false [falsch]. For the understanding would thereby be made to contradict its own general rules of thought, and so to war against itself. These criteria, however, concern only the form of truth, that is, of thought in general” (in interpreting this passage, it should be noted that widerspricht literally means speaks against, and that in German falsch often means not false but wrong in a broader sense). In other words, because conformity to general logic is essential for thinking, failure to conform to it undermines the very existence of the acts of thought whose correspondence to objects constitutes truth on Kant’s standard correspondence theory of truth. Cf. A150 / B189–90.
51. This assumption—which can in the end be traced back to Aristotle’s position in Metaphysics, bk. theta that actuality is prior to potentiality—is perhaps most explicit in Kant’s precritical essay The Only Possible Argument in Support of a Demonstration of God’s Existence (1763), where it motivates positing God as the actual ground of all possibilities. At Critique, A579–83 / B607–11 Kant rejects that theological conclusion, but not the assumption, which on the contrary now motivates an empirical analogue of that conclusion, namely a principle that all empirical possibilities are grounded in the single reality of experience—a principle in terms of which Kant now seeks to explain the earlier theological conclusion as a sort of derivative illusion. The assumption in question also lies behind Kant’s central critical-period project of explaining synthetic a priori necessities in terms of the actual fact of mind-imposition (i.e., in terms of transcendental idealism). Indeed, it is visible in his critical-period explanations of both of the only other two types of propositional necessity which he recognizes besides logical necessity: synthetic a priori necessity consists in the actual fact of mind-imposition, while analytic necessity (insofar as distinguishable from logical necessity) consists in the actual fact of the containment of a predicate-concept by a subject-concept (cf. ibid., B17: it is a matter of “what we actually think in [the given concept]”).
52. See Critique, A54 / B78; Logic, p. 14.
53. In order to solve this problem while still retaining a version of the explanation, Kant would have had to arrive at conclusions which would at the very least have greatly surprised him. He recognizes just three types of necessity that might be of relevance here: logical necessity, the necessity of synthetic a priori principles, and the necessity of analytic principles (which he sometimes explains in terms of logic but sometimes in terms of the containment of a predicate-concept by a subject-concept). Obviously, the solution to his problem here could not be that the residual necessity in question was logical necessity. But nor would he be at all inclined to classify it as synthetic a priori necessity (since, for one thing, that would deprive general logic of its validity for all thought, rather than just human(-like) thought, given that he always explains theoretical synthetic a priori necessity in terms of the imposition of the principles involved by minds like ours). And so the only remotely attractive tack for him to take would be to say that it is instead an analytic necessity, grounded in the containment of a predicate-concept by a subject-concept—in other words, to say, roughly, that it is a necessity constituted by the fact that the subject-concept “thought” implicitly contains the predicate-concept “conforms to classical logical principles.” But saying this would require conceding to analyticity (in the sense of the containment of a predicate-concept by a subject-concept) a sort of primacy over logic which would at the very least greatly surprise Kant (who commonly, rather, explains analyticity in terms of logic, specifically in terms of the logical law of contradiction—see, e.g., Critique, A151 / B190–1).
54. Wittgenstein, Remarks on the Foundations of Mathematics, pt. 1, par. 131.
55. Wittgenstein, Philosophical Investigations, pp. 226e–7e.
56. Wittgenstein, Remarks on the Foundations of Mathematics, pt. 4, par. 59.
57. In his Encyclopedia Hegel calls classical logic “the logic of mere Understanding” (par. 82).