One thing seemed clear to me, although this already is controversial, namely that the two parts differ in level of abstraction. The difference is described in §21 (titled “Remark”), which marks the transition from the first part to the second:
In the above proposition, therefore, a beginning is made of a deduction of the pure intellectual concepts [Verstandesbegriffe], in which, since the categories arise merely in the intellect [Verstand], independent of sensibility, I must still abstract from the special [way] [Art] in which what is manifold in an empirical intuition is given, in order to look only upon the unity which advenes to the intuition through the intellect by means of the category. In the sequel (§26) it will be shown, from the special [way] [Art] in which an empirical intuition is given in sensibility, that its [i.e., the empirical intuition’s] unity is none other than that which, according to the previous §20, the category prescribes to a given intuition in general [überhaupt], and so, in that its [i.e., the category’s] a priori validity with respect to all objects of our senses is explained, the aim of the deduction will only then be fully achieved. (B144–5)
The initial sections abstract from the species [Art] of our faculty of sensible intuition, and discuss the categories as unificatory only in relation to the synthesis of what is manifold in a sensible intuition überhaupt, whereas the final sections show that the synthetic unity of our sensible intuition[1]This synthetic unity is to be distinguished from the immediate unity of time and space, insofar as each is the identical formal reality of every inner or outer sensible intuition, respectively. Longuenesse, out of confusion about this point, is led to suppose that the B Deduction “rewrites” the Transcendental Aesthetic. is that very unity supplied by the categories.
If this is correct, then the B Deduction, so to speak, begins before the beginning of the A Deduction. For the A Deduction proper begins precisely with a statement about our form of intuition:
Our representations may arise whence they will … they nevertheless all belong, as modifications of the mind, to inner sense, and as such all our cognitions are ultimately subject to the formal condition of the inner sense, namely to time, as that in which they must, collectively, be ordered, connected, and brought into relations. (A98–9)
This is directly contrary to Longuenesse’s view, that “Kant takes up the B Deduction where the A Deduction leaves off.”[2]Kant and the Capacity to Judge (Princeton University Press, 1998), p. 58. According to her, then, the two Deductions are “complimentary,” as “the B Deduction takes the initial steps of the demonstration in A for granted” (p. 59). But this ill matches Kant’s own description of the relationship between the A and B editions (cited by Longuenesse herself on p. 34): “that a variety [of things] which, although they do not belong essentially to the completeness of the whole, some readers would nevertheless be reluctant to do without, insofar as they could be otherwise useful in some other respect, had to be omitted or treated in an abridged way” (Bxlii). It is not certain, I should add, that this remark concerns the B Deduction at all, and not rather the B Paralogisms, which have clearly been abridged in just the way described. I am thus left open, however, to Longuenesse’s obvious objection:
It might seem, then, that the new contribution of the second part of the Deduction is the consideration now given, not to a sensible intuition in general, but to our sensible intuition. But why should this be necessary? If our intuition provides merely a particular case for a rule that has been proved universally (all sensible intuition is subordinate to the categories), it would seem superfluous to provide a specific proof for this case, and so to a mistake to regard such a proof as that by which the purpose of the deduction should be “fully realized.” (Kant and the Capacity to Judge, p. 213)
This indeed is exactly what I have always found so puzzling.
And yet, Longuenesse’s phrasing is suggestive. Our form of intuition is indeed a special case, an Art, of sensible intuition in general. But this is no ordinary special case. How it is extraordinary, I have often expressed (though not in print) by calling it degenerate.
The term “degenerate” is not, in this context, Kant’s.[3]It seems that Kant’s usual equivalent to Latin degenerare, in the few places he requires one, is ausarten, although he also uses entarten and degenerieren. None of these terms occur in the First Critique. Current German for “degenerate case” is Entartungsfall. The use of Art (instead of Gattung) for genus in these calques strikes me as imprecise, but the story is probably more complicated than I know. I am familiar with the term from mathematics, although even there I’m afraid I know little about its history. Still, the word itself says the right thing: a degenerate case (as opposed to a generic case) is a special case in which the characteristic marks of the genus are not present. A line, for example, can be regarded as a conic section (the section of a cone by one of its tangent planes), but, considered in itself, it has no determinate center. Given a whole family of conics, you easily specify the line as one of its members, but, given just a line, it is entirely unclear how to generalize it into a family of conics. And that, according to Kant, is our situation with respect to our form of sensible intuition: we know that it is only a special case of sensible intuition, in that it includes conditions not derivable from the concept of a sensible intuition — which is to say, not derivable from the concept of a discursive intellect. But we do not, and cannot ever, know of any other possible case. The human intellect
cannot even form for itself the least concept of another possible intellect, either of such a one as would itself intuit or of one which possessed, as its ground, although a sensible intuition, nevertheless one of another species [Art] than that in space and time. (B139)
The mere process of abstraction cannot, obviously, yield a contradiction, and so we can, without formal problem, think a more general concept by abstracting from the concept of an object for us — abstracting, that is, from the schematized category of modality. But this more general concept cannot be more generally schematized, i.e. cannot be constructed with any more general case: the transcendental schemata, in realizing the categories, i.e. lending them objective reality, also restrict them (A147/B187).
If, per absurdum according to Kant, our species of intuition were a generic case of the genus sensible intuition, then Longuenesse’s objection would be insuperable. The possibility of our form of sensible intuition would then have to be derivable from the possibility of sensible intuition in general, by adding to the genus some compossible specific differentia. Hence the category of modality would have to be objectively valid in a use general to all objects of sensible intuition überhaupt, and the First Postulate would have to read:
What accords with the formal conditions of sensible intuition and discursive intellection is possible.
The first part of the Deduction having shown, however, that whatever is manifold in the sensible intuition of a discursive intellect must meet the conditions of discursive intellection (since otherwise it could not count as manifoldly given to one intellect), the first condition in this postulate (“of sensible intuition”) would be redundant, and we could say instead:
What accords with the formal conditions of discursive intellection is possible.
But the formal conditions of discursive intellection are simply the categories, so the aim of the Deduction (to show that the categories are objectively valid) would have to have been, somehow, achieved without reference to the specific differentia of our form of sense.
I hope needless to say, this is all very different from what the Analytic of Principles actually shows. The First Postulate actually reads:
What accords [or rather, “convenes”: übereinkommt] with the formal conditions of experience (according to intuition and concepts), is possible. (A218/B265)
Where “the formal conditions of experience,” according to intuitions, are space and time. We have no concept of possibility which can refer (in a theoretical use[4]The practical is what “possible through freedom” (A805/B828): in practical intellection, in other words, the conditions of the will take the place of the conditions of intuition. See also Bxxvi n.) to any object not in space and time. This limitation derives ultimately from a similar limitation upon the supreme principle of all synthetic judgments:
Every object stands under the necessary conditions of the synthetic unity of what is manifold in intuition in a possible experience. (A158/B197; my emphasis)
But if we ask, in turn, where the supreme principle gets this limitation, we will be led back to the second part of the Deduction, which begins with §22, titled: “A category has no other use for cognition of things than its application to objects of experience” (B146). All this because our species of sensible intuition, according to Kant, is a degenerate case of its genus. Is and must be: it is analytic that the formal conditions of a (species of) sensible intuition are not derivable from the formal conditions of discursive intellection. An intellect which could derive the possible order of what is manifold in its object from a principle of its own, would be intuitive.
And this is why the Deduction must have two parts. The first part establishes the analytic proposition that an object manifoldly given, to a discursive (thinking) intellect, in a sensible intuition, must be somehow capable of synthesis in accordance with a rule of synthetic unity (a concept).
This last proposition is, as I have said, itself analytic, notwithstanding it makes synthetical unity the condition of all thought; for it says nothing more than that all my representations in any given intuition whatever must stand under the condition under which alone I can credit them, as my representations, to the identical self, and thus can comprehend them, as conjoined in one apperception, through the universal expression I think. (B138)
This is indeed the key step in the Deduction, insofar as it shows that an unsynthesizably manifold content of sensible intuition, which in itself is something perfectly conceivable, is not conceivable as given to a discursive intellect. But sensible intuition, as treated at this step, is an ens rationis. I cannot coherently deny that I think (because denial that … is an act of discursive intellection), and it follows from that that I cannot coherently deny that an object is given to me in sensible intuition. But this is not sufficient basis to assert that an object is so given to me: not because some step is missing from the proof, but because I cannot assert anything about sensible intuition unless sensible intuition is possible, and it is only of our sensible intuition that we can know that. So there must be something else, beyond I think, that I cannot coherently deny. What that other thing is may be gathered from the immediate conclusions of the cogito argument, as it occurs in the Second Meditation: Imo certo ego eram, si quid mihi persuasi; fallat quantum potest, nunquam tamen efficiet, ut nihil sim quamdiu me aliquid esse cogitabo (AT 7:25). If I have thought, then I certainly was; and: if I am made to think, in the future, then I certainly will be. A res dubitans is more than just a thinking thing. The reasons to doubt are in the past, and their consequence — the doubt itself — is in the future (prudentiæ est nunquam illis plane confidere qui nos vel semel deceperunt [Med. 1, AT 7:18]). Our intellect knows itself, then, not by means of, and therefore not as subject of, intellection alone, but also by an additional faculty (a faculty which, as Descartes says, “insofar as it differs from the power [vis] of intelligizing, is not required for me myself, that is, for the essence of my mind” [Med. 6, AT 7:73]). This additional faculty is “the faculty of representing an object even without its being, at present [auch ohen dessen Gegenwart[5]I translate this somewhat oddly in order to emphasize that Gegenwart, like English “present” (and unlike, say, Anwesenheit), may be opposed to “past” and “future.” I agree with Heidegger in emphasizing this, although I think my interpretation otherwise quite different from his.], in intuition” (B151) — that is, the imagination. That kind of synthesis, the synthesis of the successive, is the only kind we know to be possible, and we know it is possible because, if we try to deny it, we find that we have now tried to deny it, or because to try and deny we must hope to deny it later.
The second part of the B Deduction, then (a part not missing in the A Deduction, but just imperfectly separated out), is not only the part that actually demonstrates the objective validity of the categories; it is also the point at which, if you follow the argument, your intellect cognizes itself. In doing that, it knows itself only through, and therefore only as conjoined with, this other faculty, imagination, which is not of its own essence:
The ego cogito [Das, Ich denke] expresses the act [Actus], of determining my existence. The existence is thus already given thereby, but the special [way] [Art] in which I should determine it … is not thereby given. To that belongs self-intuition, which has, lying as its ground, an a priori given form, i.e. time. (B157 n.)
This self-cognition is the act of transcendental reflection: the act in which the intellect, reflecting on the sources of its own cognition, itself determines the limits of its own use (A238/B297; cf. A260/B316).[6]In the Amphiboly, Kant (following Baumgarten, Metaphysica §626) gives Überlegung as the equivalent of reflexio, but in Phenomena and Noumena he apparently uses nachsinnen in the same sense. In doing so, it establishes a transcendental topic, that is, a doctrine of the place of objects (objects of intellection) as such, by drawing a limit between the object of experience, on the one hand, and an entirely empty place, on the other (A259–60/B315). The degeneracy of the case of experience (the case of imagination as the faculty of synthesis) is what makes such an empty space representable: an empirical topic could never succeed in representing all experience as surrounded by an empty place, because of the ideality of space (A427–9/B455–7; see also B147), and formal-logical dichotomy of course always presupposes the formal reality of the divided concept, hence of both concepts resulting from the division.
Note that transcendental reflection so understood does not involve any faculty other than discursive intellection and our species of sensible intuition. There is no intellectual intuition, no transcendental psychology, no Wesensschau, no ontological Vorbegriff of being as temporal. There is only the mismatch, already noticed by Descartes, between the way I think myself and the way an object is thereby determined.
