5. The Abyss of Judgment

[Space] is (like time) limitless (not infinite).

—Immanuel Kant, Opus Postumum

Wahrscheinlichkeiten fallen hier gar weg, wo es auf Urteile der reinen Vernunft enkommt.

—Immanuel Kant, Critique of Judgment

We still do not know what a frontier is, or even what its nature is, except to be of nature. And we could even say that just that is the frontier: not knowing. We trace frontiers in order to know, but we will know nothing of the frontier itself. Kant is less interested in knowledge per se, pace the neo-Kantians, than in its frontier, where knowledge fails.

If there were in Kant a faculty of the frontier, it would clearly be the faculty of judgment. The success of the operation might be disputed, but the aim of the third Critique seems clear enough: that of throwing a bridge over the abyss that has opened between the world of experience and the world of freedom, between speculative and practical reason.1 The abyss is, it would seem, again what we are here calling the frontier, in a peculiarly exacerbated or exasperated form. But judgment would be the faculty allowing it to be crossed, or at least allowing the two sides to be joined. Judgment would then be reason itself being rational, the pure faculty of relations that are both singular and analogous. But it will turn out, following a fractal logic that has been appearing regularly along our path, that the frontier, and therefore judgment itself, is merely the layered multiplication of its own collapsing structure.

Let’s put this suggestion to the test of reading. We must first note a certain complication in the treatment Kant reserves for judgment. For example, in the first Critique, just where he explains the discursive character of human knowledge, Kant draws from it a primordial place for judgment. The understanding is not an intuitive but discursive faculty, because, unlike mathematics, it gives rise to a knowledge by concepts that it does not construct. But all the understanding can do with concepts is to judge with them:

Now the understanding can make no other use of these concepts than that of judging by means of them. [. . .] Judgment is therefore the mediate cognition of an object, hence the representation of a representation of it. In every judgment there is a concept that holds of many, and that among this many also comprehends a given representation, which is then related immediately to the object. We can, however, trace all actions of the understanding back to judgments, so that the understanding in general can be represented as a faculty for judging [ein Vermögen zu urteilen]. For according to what has been said above it is a faculty for thinking. Thinking is cognition through concepts. (CPR, A68–69/B93–94)

To think is to know via concepts; for us, those concepts have a discursive use; so we use them in order to judge. What is proper to the act of judgment is the fact of identifying, for a concept, a case of that concept, a case that will be subsumed under that concept. Judgment always has to do with cases, whence, a little further along in the first Critique, a more detailed description of its operation, at the very beginning of the chapter entitled “Of Transcendental Judgment in General,” where “judgment” now translates Urteilskraft. As the understanding is the faculty of rules, judgment is the power of subsuming under those rules (i.e., of recognizing cases). General logic cannot contain rules for judgment because (according to a formula Wittgenstein would later make famous) there would have to be a further rule for the application of rules, and then a rule for the application of that rule, and so on ad infinitum. Judgment is an innate capacity that cannot be taught, the Mutterwitz that no knowledge can provide or do without.2

What distinguishes transcendental judgment is that it seems that it can give itself its cases a priori and thus find its object without exposing itself to the risk that Mutterwitz will fail. Given that it deals only with the relation to an object in general, transcendental logic as such presupposes that its case is given, without having to subsume it each time. As there is transcendental logic only insofar as a possible object in general is given, this logic cannot fail to presuppose that object. But as, according to a formula Kant constantly repeats, our knowledge is limited by the bounds of possible experience, even this apparently certain judgment is oriented toward empirical judgments as alone giving it possible content. Judgment as such, then, cannot avoid the singular case, and this is what will give it its peculiar status in Kant: For judgment is presented both as a faculty and as the “faculty” of any possible distinction between faculties, the faculty of their frontiers.3

According to the definitive introduction to the third Critique, judgment has no place in doctrine, in the achieved system of a metaphysics, but only in critique. So it seems there is something properly critical about judgment. From the point of view of doctrine, there are only two types of concepts—of nature, on the one hand; of freedom, on the other. In the system, judgment, situated between these two domains, can be attached to either of these two faculties as needed. Having repeated and justified the division of philosophy into two in its first section, the introduction pulls back to consider “The Domain of Philosophy in General,” “Vom Gebiete der Philosophie überhaupt.” This is where we encounter a famous passage that tries to draw frontiers in this domain, which contains all objects of any possible application of concepts:

Concepts, insofar as they are related to objects, regardless of whether a cognition of the latter is possible or not,4 have their field [ihr Feld], which is determined merely in accordance with the relation which their object has to our faculty of cognition [Erkenntnisvermögen] in general.5—The part of this field within which cognition is possible for us is a territory [Boden, a ground] (territorium) for these concepts and the requisite faculty of cognition. [So the territory is situated within the field, is smaller than it.] The part of the territory [following a concentric progression, then] in which these are legislative [gesetzgebend] is the domain [Gebiet, the same word as for the whole space of philosophy in general: as with all legislation, what irrupts in the center is also the absolute outside] (ditio) of these concepts and of the corresponding faculty of cognition. Thus empirical concepts do indeed have their territory in nature, as the set of all objects of sense, but no domain (only their residence [Aufenthalt], domicilium) [so we are to imagine that the part of the territory that does not fall within the domain is the residence or can give rise to a residence]; because they are, to be sure, lawfully generated, but are not legislative, rather the rules grounded on them are empirical, hence contingent. (CJ, 61–62)

Let’s take this distribution as literally as possible. Kant implicitly invites us to draw a diagram of it, which might look like this:

But we have immediately to complicate things, because Kant opens the following paragraph by saying that “our cognitive faculty as a whole has two domains, that of the concepts of nature and that of the concept of freedom; for it is a priori legislative through both” (CJ, 62). But the terrain or territory of these two domains is the same (the terrain of objects of possible experience), and so we must think that they can and even must share some objects, which would give the following diagram:

We shall see this still quite simple topography getting more and more complicated as Kant advances in his description. The fact that the two legislations bear on domains situated on the same terrain (domains that overlap at least somewhat, and perhaps completely) does not give rise to interferences at the level of legislation itself but to a mutual limitation (Kant uses the verb einschranken, which has become familiar to us) at the level of its effects. But the two domains remain separate to the extent that their objects are different in nature, or at least are considered in different lights or perspectives. Concepts of nature present their objects to intuition, as phenomena, whereas concepts of freedom present them as things in themselves, without that presentation taking place in intuition. Where there is theoretical knowledge, there is no thing in itself, and where there are things in themselves, there is no theoretical knowledge. Of such a “knowledge” (which if it claimed to be such would be, in us, finite beings, transcendental illusion), we must indeed have an Idea, which then takes its place in the field, but without giving rise to a terrain or, a fortiori, to a domain:

There is thus an unlimited [unbegrenztes: so we were wrong to draw a frontier around the field within the general domain, but if we erase it we are at a bit of a loss as to how to represent the relations between the field and the general domain] but also inaccessible field for our faculty of cognition as a whole, namely the field of the supersensible, in which we find no territory for ourselves, and thus cannot have on it a domain for theoretical cognition either for the concepts of the understanding or for those of reason, a field that we must certainly occupy with ideas for the sake of the theoretical as well as the practical use of reason, but for which, in relation to the laws from the concept of freedom, we can provide nothing but a practical reality, through which, accordingly, our theoretical cognition is not in the least extended to the supersensible. (CJ, 63)

This already complicates the description given by Kant because, where we thought we were dealing with a concentric arrangement of field, terrain, and domain, we now see that things are more complex. Previously, it appeared that theoretical and practical cognition each had their domain within the terrain; now, the domain of the practical is, if not identified, at least placed in a strange relation with the (now unlimited) field that surrounds the terrain. To understand what is happening here, we clearly need to introduce a dynamic element into the picture, which would allow practical legislation to take place from out of the field while still aiming at the domain inside the terrain apparently occupied by theoretical legislation. This is why the famous gulf or abyss, the unübersehbare Kluft, the cleft over which one cannot see, is not simply a division or accentuated division between two sides of the same type and also why the attempt to throw a bridge over it (the bridge of judgment) is not as easy to conceive as is sometimes thought, unless we imagine a one-way bridge:6

Now although there is an incalculable gulf fixed between the domain of the concept of nature, as the sensible, and the domain of the concept of freedom, as the supersensible, so that from the former to the latter (thus by means of the theoretical use of reason) no transition is possible, just as if there were so many different worlds, the first of which can have no influence on the second: yet the latter should [soll] have an influence on the former, namely the concept of freedom should [soll] make the end that is imposed by its laws real in the sensible world; and nature must [muß] consequently also be able to be conceived in such a way that the lawfulness of its form is at least in agreement with the possibility of the ends that are to be realized in it in accordance with the laws of freedom.—Thus there must [es muß] still be a ground of the unity of the supersensible that grounds nature with that which the concept of freedom contains practically, the concept of which, even if it does not suffice for cognition of it either theoretically or practically, and thus has no proper domain of its own, nevertheless makes possible the transition from the manner of thinking in accordance with the principles of the one to that in accordance with the principles of the other. (CJ, 63)

So if there is a possibility of transition from one domain to the other, this transition happens in one direction only (it is impossible to move from the theoretical to the practical, on pain of transcendental illusion), and this is done not by throwing a bridge over the abyss but by going under, through a tunnel; because the Kluft does have a bottom, it cuts into the common subsoil of the field, a sub-soil that is, however, super-sensible and that is shared by theoretical and practical cognition. This passage, or its possibility, is not itself of the order of theoretical knowledge. That it be possible to pass from the practical to the theoretical (that legislation according to freedom be compatible with legislation according to nature), that the moral law has effects that can be seen in the world, forms part of what makes that law the law that it is. It is duty’s duty to realize itself in the phenomenal world. The moral law must, ought, or has to be able to actualize itself in the sensible world, according to a pre- or archi-moral duty we already signaled as the archi-prescription of prescription (“do it!”): Duty owes it to itself (se doit) to make itself real. But this primary Sollen (which must precede the distinction between nature and freedom, because it is the principle of their separation and their communication), according to which one has to have effects in the world, provokes in return another duty that this time owes it to itself to be a Müssen, according to which the legality of natural legality is such that it must, by its form, lend itself to this realization of moral ends. Only this consequence gives any substance to the archi-duty of duty, because if nature did not respond, duty could owe itself to itself ad infinitum without that duty ever becoming possible to carry out. It is necessary, then, for the duty of duty to be realizable as it ought to be, failing which it would not be duty and there would be no duty. There ought to be the it is necessary that, because it is necessary that duty ought to be done. This originary connivance between necessity and obligation, nature and freedom, müssen and sollen, mechanism and ends, will make judgment (which has not yet been named in the introduction) into an essentially teleological judgment.

The sensible world is thus the end of duty, or because duty is obligatory from the end, it is the end of the end. What Kant has not yet called “judgment” will be the faculty of this doubled-up end or of this movement that carries duty toward its end and ensures that the moral law will find in the natural world the formal condition of its possible realization. Consequently, and contrary to what one might think on reading the “typic” in the second Critique, the form of law, the Gesetzmässigkeit that the moral law supposedly borrows from nature, is, rather, given to nature by the form of law of the moral law. Nature has legal form in order to respond to the moral law; natural necessity is thus dictated by the moral law, which gives a (supersensible) ground to its necessity but, at the same stroke (by digging the tunnel), undermines this ground. Because natural necessity owes itself to duty, it is no longer quite necessary or quite natural. That the (legal) form of nature must accord with the possibility of ends is its end. The end of nature is that law find its end in it. And in this way, even the archi-determinative judgment of the first Critique is already teleological, because it must presuppose that nature is such that it will provide it with cases that fit the law that the understanding is in a position to prescribe to it. This is precisely what allows for the possibility of the transcendental position itself. The understanding is thus legislative with respect to nature in a very strong sense, which involves a prescriptive or pre-scriptive element—the understanding prescribes itself a nature such that the laws it pronounces will find cases in it. Nature is in this way made for the judgment that will make a case, and this made for already gives its end to nature as able to receive practico-moral realizations. Even the subsumptive determinative judgment, then, is determined in return by its apparent other, which returns through the tunnel to pop up in the middle, like a foreign legislator.

Kant says this much later, in §76 of the Critique of Teleological Judgment. Reason demands the unification of the diverse. But the understanding, insofar as it is tied to sensibility and therefore subject to a transcendental logic (for the transcendental is less what rises away from and more what comes back down to the singular and the empirical), can only proceed from the universal to the particular, in which it cannot fail to find something contingent. This contingent element must be brought back under the law, which is done by postulating a purposiveness of the contingent as such:

But now since the particular, as such, contains something contingent with regard to the universal, but reason nevertheless still requires unity, hence lawfulness, in the connection of particular laws of nature (which lawfulness of the contingent is called purposiveness), and the a priori derivation of the particular laws from the universal, as far as what is contingent in the former is concerned, is impossible through the determination of the concept of the object, thus the concept of the purposiveness of nature in its products is a concept that is necessary for the human power of judgment in regard to nature. (CJ, 274)7

The apparently paradoxical consequence is that, for all its legality, nature must be fundamentally contingent. To be able to welcome the law, nature must not be necessary through and through. Natural necessity, seen from the point of view of its end (the law), is that it be contingent. The necessity of necessity is contingency. This chance of contingency (law’s only chance) is not knowable as such but the object of a nonknowledge that calls for what is called judgment. For us to be able to judge from the end, there must be contingency, because only contingency calls for the end as the only chance of its legality. Necessity has no end, it is blind. For the end, this blindness must blind itself again in the contingency of blind chance.

Nature, then, has an end, a purpose: that of having a purposiveness, which it has only in contingency. That there be this end comes under neither theoretical cognition (for if we could demonstrate it, we would reduce contingency to natural necessity; this is why Kant constantly repeats that the principle of judgment—namely, that we must judge as if it were made to be judged—is only subjective, or at least is not objective, while functioning just as if it were objective, because otherwise there would be no “subject”) nor practical cognition (there is no duty here) but underlies the possibility of both domains as the indispensable presupposition of their joining and separation. And yet, if this presupposition is what links the two domains beneath the abyss, it alone gives rise to the abyss. The theoretical and the practical are separated only because they are thus linked together by the purposive legality of the singular judgment. Kant’s problem in the third Critique is thus not to bring together two domains that ran the risk of being too far separated but precisely to think the frontier of these two domains, to think them together as separated. This is why, at the beginning of section 3 of the introduction, Kant repeats that we are not here dealing with a third domain but with what allows us to establish the legality of the other two domains, by tracing their frontiers:

The critique of the faculties of cognition with regard to what they can accomplish a priori has, strictly speaking, no domain with regard to objects, because it is not a doctrine, but only has to investigate whether and how a doctrine is possible through it given the way it is situated with respect to our faculties. Its field [Feld] extends to all the presumptions of that doctrine, in order to set it within its rightful limits [um sie in die Grenzen ihrer Rechtmässigkeit zu setzen]. However, what cannot enter into the division of philosophy can nevertheless enter as a major part into the critique of the pure faculty of cognition in general if, namely, it contains principles that are for themselves fit neither for theoretical nor for practical use. (CJ, §64)

The critical chance of thought (of reason insofar as it is finally, in the end, devoted to morality) would then be the frontier as a separation that allows for a passage. This chance depends on contingency, which alone bears witness to the possibility in nature of a causality other than that of mechanism, and which, extended to the system of nature as such, will tell the truth of mechanism itself, as we shall see. But then, the whole Kantian topography is turned upside down when it comes to specifying the nature and place of judgment. Kant began by establishing the apparent mutual independence of the domains of nature and freedom; he then deduced a quasi-practical quasi-necessity of a one-directional transition between the two, which alone ensured their true separation. Now he is seeking a faculty to take charge of this uniquely critical activity.

Only here does analogy intervene, and as we have had several occasions to note, analogy analogizes endlessly (or bottomlessly). The analogy is not merely the bridge (or the tunnel) that crosses the abyss, it is simultaneously the abyss itself, which is starting to provoke cracks and fissures everywhere in the fields, terrains, and domains delimited so far. Kant takes another look at the frontier between understanding and reason and finds between the two the power of judgment (here named for the first time), “about which one has cause to presume, by analogy [nach der Analogie]” that it could also (by analogy, then, with the faculties of the understanding and of reason) have the means to legislate, or rather (because what is proper to the frontier is that it is not a domain) not to have its own legislation but “a proper principle of its own for seeking laws [ein ihr eigenes Prinzip nach Gesetzen zu suchen].” Which would give judgment, qua bearer of its own principle, if not a domain (“even though it can claim no field of objects as its domain [wenn ihm gleich kein Feld der Gegenstände als sein Gebeit zustände]” says Kant, insouciantly muddying any remaining terminological clarity of his own description), at least a terrain.8

To seek laws, for what is proper to judgment as such (i.e., when it is not simply pressed into the service of theoretical knowledge—which, while being already teleological, as we saw, does not need to judge in a strong sense to the extent that the law is already given to it and it only has to subsume the case)9 is precisely to be seeking its own law. As we said, the whole Kantian system is teleological in that it is organized for the law. But we must not conclude from this, as is perhaps a common thing to do, that the law qua law is already given. To the contrary—and this is precisely why the general contingency of nature passes the understanding, which, for its part, believes in its necessity—purposiveness as legality of the contingent stretches out toward a law that has not yet come. Which is why judgment as such is essentially reflective (determinative judgment being the blind judgment of blind necessity) in search of its law. Qua reflective, judgment does not know how to judge but tries to judge. In judging, it calls for its law for which it can only wait. One can always draw a law from a judgment, but the law of judgment is that that law will never be certain. Judgment is made for a law it can never establish. Judgment itself thus has the structure of a purposiveness without purpose, a legality without law (CJ, §22), and, following the play of analogy, finds this structure in, or projects this structure into, some of the objects of its judgment. So the aesthetic judgment is “aesthetic,” and so on without end.

But this structure is that of transcendentality as such. The transcendental is a movement beyond that comes back. Experience is empirical for the transcendental—in view of the transcendental, viewed from the transcendental—as nature is contingent for the law. The transcendental comes back to the empirical as law comes back to nature. But this means that the transcendental is (no more than) the empirical that comes back (to itself). And this analogy between the transcendental and the law is, rather, an identity. The transcendental names the being-law of law as such.10 This (being-)law names a seeking after the law (still the Seefahrer or perhaps the astronaut), because we do not yet have the law and we will not have it, on pain of dogmatism: We only have the law of the law, which is its endless retreat. So the transcendental is not yet transcendental but what is nowadays called the quasi-transcendental. The transcendental can “be” transcendental only on condition of not being transcendental (for then it would be transcendent). What separates the transcendental from the transcendent is what dynamically separates the transcendental from the transcendent—namely, the frontier “itself.” Saying that this separation is dynamic is also a way of saying that it is not done but that it is doing, as the becoming-transcendent of the transcendental and the becoming transcendental of the transcendent, which is the very tracing of the frontier. Every frontier transcendentalizes to the extent that it separates, and it prevents the achievement of that transcendentalization to the extent that it joins. The transcendental is the abyssal demultiplication of the frontier, along the frontier.

This is why Kant, under the name “cosmopolitanism,” has to think at most an internationalism (the chance of cosmopolitanism lies this side of cosmopolitanism, it needs the frontier whose erasure it also projects) and why the end as (mortal) perpetual peace withdraws as a finality that must never reach its end. The chance of peace is not to be perpetual but perpetually in the process of putting off its perpetuity. Teleology thus understood puts an end to the end; it returns on itself. According to a logic that has been haunting us all along, the end of teleology is the end of teleology. “Zweckmässigkeit ohne Zweck” is no doubt naming nothing but that. “Purposiveness without purpose,” “finality without end” still has too much end in its finality. So as not to be held hostage by the end, perhaps we could try other formulations: ((finality (without end) without end)), or “((finality (without finality)) without end).”11

This could easily bring us back to Epicurus. As we have seen, every time Kant states a crucial problem for his philosophy, the ghost of Epicurus comes to prowl around. And if Epicurus, before, was opposed to Plato for the antinomy in the first Critique, and to the Stoics for the relation between virtue and happiness in the second, here he will be opposed to Spinoza. As always for what we are interested in here, the context is one of antinomy, as expounded in the dialectic of teleological judgment.

The determinative judgment cannot have an antinomy, says Kant, because it is not autonomous, does not lay down a law. However, when we are dealing with the reflective judgment—in which, as we saw, the law is not given, but where judgment gives itself the law—judgment has to follow maxims for guidance, and these maxims can enter into conflict among themselves, giving rise to the possibility of an antinomy and, therefore, a dialectic, which calls for critique. Exploring empirical nature beyond what is legislated determinatively by the categories of the understanding, we are led to invoke the maxim that all natural products must be possible according to mechanical causality but also, sometimes, especially when we come across organisms, the maxim according to which at least this part of nature (but perhaps nature as a whole, understood according to the analogy between it and the organisms it contains) can be explained only by appealing to purposive causality, to final causes. Kant’s presentation can give the impression that this antinomy is simply solved (by the end of §71, and even as early as §70, in the very statement of the antinomy, which on that reading would evaporate immediately) by insisting on the fact that these are merely maxims of investigation, regulative principles of our investigation and not constitutive principles of the objects of that investigation. There would be an obvious contradiction if one were to assert that an object were objectively caused both by natural mechanism and by final causality (a technic of nature) but no contradiction in appealing as a principle for judging nature (without claiming that this gives rise to objective knowledge), on the one hand to the maxim of mechanical causality (I have to judge according to that maxim as far as possible, for only this maxim puts me on the path to knowledge properly speaking, even if that knowledge is not determinative) and, on the other, to the maxim of final causality (when the first maxim no longer serves, I have to judge some natural objects—most notably organisms, and even, following their example, nature as a whole—according to a quite different principle, that of final causality).

Kant’s argument here is extremely delicate. If one were to maintain these two principles dogmatically (making them into determinative principles), there would indeed be an antinomy, because these two principles would be in contradiction. But as Kant points out, this contradiction would not be an antinomy of judgment, because by making these principles into principles determining objects, we would have provoked a contradiction in rational legislation itself with regard to nature, and that contradiction could not be resolved, because human reason cannot determine the objects of empirical nature. We do, of course, prescribe laws to nature, according to transcendental logic, but these laws cannot determine nature qua field of empirical investigation, in which, as we shall see in more detail, objects qua particular cases remain in part contingent with respect to the legislation of reason.

It cannot, therefore, suffice to insist that we are dealing only with maxims, regulative principles for the reflective judgment of empirical nature, in order to solve the antinomy of judgment, because it is only on the level of maxims that there is an antinomy of judgment (rather than of rational legislation itself). And yet Kant, at the end of a paragraph that presents itself as a mere preparation to the solution of the antinomy (§71: “Vorbereitung zur Auflösung obiger Antinomie”), can write the following:

All appearance of an antinomy between the maxims of that kind of explanation which is genuinely physical (mechanical) and that which is teleological (technical) therefore rests on confusing a fundamental principle of the reflecting with that of the determining power of judgment, and on confusing the autonomy of the former (which is valid merely subjectively for the use of our reason in regard to the particular laws of experience) with the heteronomy of the latter, which has to conform to the laws given by the understanding (whether general or particular). (CJ, 261)

But Kant has already explained that an antinomy of judgment can arise only if the maxims of judgment are only maxims of judgment (and not principles determinative of objects). The antinomy of judgment does not result from a confusion of autonomy and heteronomy but from the fact that judgment is autonomous. The antinomy cannot be solved by simply recalling the same antinomy. Any appearance of a legislative antinomy (an insoluble contradiction) is removed by recalling the maxim-status of the mechanistic and finalist principles. It still remains to resolve the properly judgmental antinomy that results—the resolution of the antinomy of judgment is prepared by removing any appearance of an antinomy in the law.

So where exactly do we find the antinomy that remains to be resolved? In the very presentation of the antinomy in §70, Kant seems to suggest that the coexistence of the two maxims is fundamentally peaceful. But this peace of judgment about nature does not result from the equal status of the two maxims but a subordination of the one to the other, in the name of knowledge:

For if I say that I must judge [beurteilen] the possibility of all events in material nature and hence all forms, as their products, in accordance with merely mechanical laws, I do not thereby say that they are possible only in accordance with such laws (to the exclusion of any other kind of causality); rather, that only indicates that I should [soll] always reflect on them in accordance with the principle of the mere mechanism of nature, and hence research the latter, so far as I can, because if it is not made the basis for research then there can be no proper cognition of nature [keine eigentliche Naturerkenntnis]. Now this is not an obstacle to the second maxim for searching after a principle and reflecting upon it which is quite different from explanation in accordance with the mechanisms of nature, namely the principle of final causes, on the proper occasion, namely in the case of some forms of nature (and, at their instance, even the whole of nature). (CJ, 259)

So the point is not to have both maxims available and choose one or other arbitrarily. In the perspective of knowledge, the principle of mechanical causality (in harmony with everything in the understanding that is legislative with respect to nature) comes first, and we stick with it as long as possible, moving to the other maxim only when the mechanistic explanation comes up short (for example, when faced with organisms). However—and this is where the antinomy will come into its own—the odd occasions on which one appeals to the purposive principle of judgment also serve as a pretext or a motive to judge similarly of nature as a whole (judging that nature is fundamentally teleological), which cannot fail to bring about a conflict between the two maxims, which henceforth are trying to apply to the same objects at the same time. The antinomy comes about as a result of abandoning any dogmatic claim to knowledge, which allows free competition between the two maxims of judgment.

We do not know if mechanical causality suffices to account for everything that is in nature, and we do not know this because natural particularities and their laws are, for us, empirical and therefore, at least in part, contingent. So we know that we cannot know nature through and through according to mechanical causality, despite the infinite field for judgment opened up by this principle in the scientific investigation of nature. We know that for us mechanical causality does not account for everything (for example, organisms). We could simply leave things at that (and there would be no antinomy). If our investigation of nature is driven by the search for scientific knowledge, we will follow the mechanistic maxim as far as possible, and the purposive maxim as regards organized beings, without further elaborating the “speculative” question of knowing whether (without making it a determinative principle) this latter maxim does not tend toward an inclusive understanding of nature as such, which would reverse the order of the hierarchy of maxims in favor of the maxim of purposive causality. That maxim tells us to judge as if nature involved a final causality, and for the purposes of the investigation of nature we can and even must remain with this subjective status of the maxim without making more of it. But we might also ask (while ruling out any theoretical knowledge in this respect) whether the invitation nature gives us to judge it according to the maxim of final causality does not take us further:

Now one could leave this question or problem for speculation entirely untouched and unsolved, for if we are satisfied with speculation within the boundaries [Grenzen] of the mere cognition of nature, the above maxims are sufficient for studying nature as far as human powers reach and for probing its most hidden secrets. [But to the extent that we are not satisfied with this, and do try to go beyond those limits] It must therefore be a certain presentiment [Ahnung] of our reason, or a hint as it were given to us by nature, that we could by means of that concept of final causes step beyond nature and even connect it to the highest point in the series of causes if we were to abandon research into nature (even though we have not gotten very far in that), or at least set it aside for a while, and attempt to discover first where that stranger [Fremdling] in natural science, namely the concept of natural ends, leads. (CJ, 261–62)

Suspending scientific research in this way, in the name of speculative reason, does not mean one is falling back into the dogmatism that consisted in taking a subjective principle for a determinative principle but rather opening a question, a problem, as Kant says, “a wide array of controversial problems.” And these controversies have indeed taken place in the tradition: Either it has been asserted that the technic of nature (its at least apparent operation according to final causality) is not intentional (and thus is not really a final causality but rather “identical with the mechanism of nature or dependent on one and the same ground, where, however, since in many products of nature this ground is often too deeply hidden for our research, we attempt to ascribe it to nature by analogy with a subjective principle, namely that of art, i.e., causality in accordance with ideas” [CJ, 181]) or that it is intentional and that there really is a final causality. Kant, running the risk of confusing the reader, calls the claim that the technic of nature is not really one the system of idealism (in the sense that the final causality operates only at the level of the idea) and realism the system according to which there really is such a final causality.

As is often the case in Kant, there is some degree of convergence between the history of philosophy and the a priori deductions of reason: Reason alone suggests a matrix of four possible ways of approaching the problem of the objective purposiveness of nature—and it turns out that these four possibilities have indeed been realized in the texts of the tradition. Kant, very satisfied with this convergence, which means that the dogmatic resources of reason have all been exploited, henceforth leaving room only for critique (according to the schema laid out in the first preface to the first Critique), adds a footnote:

One sees from this that in most speculative matters of pure reason the philosophical schools have usually tried all of the solutions that are possible for a certain question concerning dogmatic assertions. Thus for the sake of the purposiveness of nature either lifeless matter or a lifeless God as well as living matter or a living God have been tried. Nothing is left for us except, if need be, to give up all these objective assertions and to weigh our judgment critically, merely in relation to our cognitive faculty, in order to provide its principle with the non-dogmatic but adequate validity of a maxim for the reliable use of reason. (CJ, 263n)

If we clearly must not conclude too rapidly that, in order to orient himself in thinking, Kant begins by appealing to the tradition, we must also beware of accepting too readily that critique has been able to overcome that tradition. As we shall see, Kant has the greatest difficulty establishing the status of his “maxims” here (the whole drama of this antinomy) and, by the same token, separating the properly critical moment of his thinking from what the tradition presents dogmatically. As we shall not cease verifying, it is contingency that plays a crucial role here, and we shall see that, in spite of appearances, it is again from Epicurus that Kant will have the greatest difficulty taking his distance.

Each of Kant’s two ways of approaching the problem in turn divides into two. On the side of idealism, we have on the one hand, as we saw earlier, Epicurus (and Democritus), and on the other Spinoza:

The idealism of purposiveness (I always mean objective purposiveness here) is now either that of the accidentality or of the fatality of the determination of nature in the purposive form of its products. The first principle concerns the relation of matter to the physical ground of its form, namely the laws of motion; the second concerns the hyperphysical ground of matter and the whole of nature. The system of accidentality, which is ascribed to Epicurus or Democritus, is, if taken literally, so obviously absurd that it need not detain us; by contrast, the system of fatality (of which Spinoza is made the author, although it is to all appearance much older), which appeals to something supersensible, to which our insight therefore does not reach, is not so easy to refute, since its concept of the original being is not intelligible at all [nicht zu verstehen ist]. (CJ, 263)

This “absurdity” of the Epicurean system (Democritus never reappears in the rest of the text and the question of any differences there may be between Democritus and Epicurus, which was to occupy a young German philosopher a few decades later, is not even raised), which is here dismissed in the same way as in the “Universal History” text, is the object a little later of a fuller explanation as to its absurdity:

The systems that contend for the idealism of final causes in nature [i.e., those who argue that there are no such causes objectively speaking] concede to its principle a causality according to laws of motion (through which natural things purposively exist [i.e., there is an appearance of purposive causality, which is really only a mechanical causality]), but they deny intentionality to it, i.e., they deny that nature is intentionally determined to its purposive production, or, in other words, that an end is the cause. This is Epicurus’s kind of explanation, on which the difference between a technique of nature and mere mechanism is completely denied [according to Epicurus there is only mechanical causality that, in some cases, produces effects which misleadingly seem to have been produced by final causality], and blind chance is assumed to be the explanation not only of the correspondence of generated products with our concepts of ends, hence of technique, but even of the determination of the causes of this generation in accordance with laws of motion, hence of their mechanism, and thus nothing is explained, not even the illusion in our teleological judgments, and hence the putative idealism in them is not demonstrated at all. (CJ, 264)

As always, the rejection of Epicurus turns out to be more complicated than Kant’s initial gesture would suggest. For what is the basis of the refutation here? Not, as one might have expected, the fact that Epicurus simply denies all final causality in the name of mere mechanism (which would be enough to motivate rejection on the grounds of dogmatism, as in the first Critique, where, it will be remembered, the ambiguity in the presentation of Epicurus depended on the possibility of making his [anti-]theses into mere hypotheses leading to an enlargement of knowledge) but indeed the way he understands mechanical causality itself when he writes it off to chance. According to Epicurus as read by Kant, the only causality is causality by chance, which immediately removes all possibility of understanding anything at all. The appearance of final causality is then “explained” by the reality of a mechanical causality, which is not even a causality because it rests entirely on chance. Epicurus, called in to show that the appearance of finality in nature is not the result of an intention, shows too much from Kant’s point of view by refusing even the principle of nonintentional mechanical causality, which makes the question of causality in general fall even below the opposition between intentionality and nonintentionality into pure chance, which explains nothing and is absurd by virtue of invoking absurdity (chance) as the principle for explaining anything at all.

So chance remains in an excessive and eccentric position with respect to the needs of Kant’s demonstration. To fill this square in his matrix (the nonintentional/idealist square), Kant did not need a doctrine as radical as that of Epicurus, which, as he presents it, is not even a mechanistic dogmatism. Unless—and this is the suspicion that will henceforth guide our reading of the antinomy—in Kant’s logic, mechanism without final causality will always be doomed to the abysses of chance, in which case, as always, Epicurus, excessive with respect to the position Kant assigns to him in the demonstration he is attempting here, will not really have been dismissed and will continue to prowl around the antinomy as what we might call its negative resource.

Recall Kant’s procedure thus far. The antinomy of judgment results from the possibility for reason to appeal, in its consideration of nature, to two incompatible maxims: This antinomy is an antinomy of judgment—rather than a contradiction in the legislation of reason itself—only to the extent that we are in fact dealing with two maxims (regulative principles). In the perspective of the investigation of nature (in the optic of natural science), the coexistence of these two maxims is not a problem so long as we give priority to the maxim of mechanical explanation and follow it as far as possible, appealing to the finalist maxim only when the former no longer works (i.e., as we shall see, at the moment of contingency). But, following the (natural) movement of speculation, one can suspend this priority of the mechanical maxim and give priority to the finalist maxim, on the grounds that natural contingencies (as seen from the point of view of mechanism) may give us a clue as to a more inclusive natural purposiveness to which our reason cannot be indifferent and which would indeed be the question of transcendental philosophy. In the tradition, there are four main ways of dealing with this problem of at least apparent natural purposiveness, but the Epicurean way is absurd, for not only does it attempt to reduce purposiveness to mechanism (which in itself is not so different from what the naturalist must do to attain a true knowledge of nature), it also reduces mechanism to chance. It remains to be shown that that is the fate of all mechanistic explanation as such (and thus of any knowledge of nature in the theoretical sense and, in the end, of any determinative judgment) if it is not subordinated to the finalist maxim, which by the same token becomes much more than a simple “subjective” maxim. And this disturbs all the oppositions on the basis of which Kant is trying to think here, beginning with the oppositions between the constitutive and the regulative,12 the determinative and the reflective, the contingent and the necessary, among others. In this perspective, the excessive position of Epicurus will give us something like a negative truth of Kant’s thinking.

Kant thus disposes of the four traditional ways of dealing with the question of natural purposiveness: Epicureanism is “absurd,” as we have just seen; Spinozism is unintelligible, in that it ends up positing a “blind necessity” (CJ, 265) that is the counterpart of the “blind chance” (CJ, 264) of Epicureanism; hylozoism is unsatisfactory because not only is the attribution of life to matter a simple category error but it also relies on a vicious circle (we suspect on the basis of organisms that nature as a whole is perhaps also organized in terms of final causality; hylozoism explains the purposiveness of the parts by simply presupposing the purposiveness of the whole, which is what we were supposed to be explaining); and theism is dogmatic in that we would first have to prove the impossibility of an exhaustive explanation according to mechanical causality before having grounds to posit in determinative fashion the existence of a superior intelligence outside nature. Having thus disposed of the four traditional ways of dealing with the problem, Kant needs to motivate the properly critical explanation of the antinomy (and even the bare presentation of the antinomy, because we are still waiting to find out exactly how this antinomy really is an antinomy), and this he attempts in §74. To have grounds for a dogmatic use (for a determinative judgment, then) of the concept of a thing as a natural end (namely the concept of a thing as part of a nature subject to final causality), we would have to be able to establish the objective reality (“objective,” as always in Kant, according to the conditions of possibility of an experience for us) of this concept. One might be tempted to think that this concept must be objective, because it is undeniable that it comes to us in a context of empirical conditions, which must be subject to the laws of the understanding in general, which would then provide the necessary principle of subsumption. But we cannot abstract this concept from experience (i.e., produce it qua concept by a process of abstraction from the singular conditions in which it shows up in given empirical situations), because it is a concept possible only as a rational principle of the judgment we bring to bear on the object. Which amounts to saying that we can never see a natural end as such (in a moment we shall see that all we ever see is a contingency), but we can judge a given object as being the result of a causality other than a mechanistic causality. But judging in this way, according to a rational principle indeed, but one the objective validity of which remains undemonstrable, we cannot be proceeding by subsumption according to a determinative judgment. So the principle remains regulative for a reflective judgment and thus susceptible to a treatment that is merely (but radically) critical (i.e., one that, given the position of judgment as we saw it in the introduction, cannot ever advance beyond critique to doctrine).

Because the judgment brought to bear according to this principle is necessarily brought to bear on the basis of a natural datum (what we judge to be a natural end—an organized being—is indeed found in nature), it seems to involve a contradiction, which will be the formal marker of the impossibility in which we find ourselves of making it into an objectively established concept giving rise to determinative judgments (and therefore objective knowledge). For to the very extent that the object in question is found in nature, it must necessarily involve necessity entailed by laws, whereas qua natural end it involves contingency with respect to those same laws. So the antinomy properly speaking is here: The point is not that we can invoke one or other of two maxims of investigation of which the one relays the other when the first fails, because the concept that guides one of those maxims (at least) (i.e., the one that guides the purposive maxim) is apparently split by two contradictory suppositions, of necessity and of contingency. This antinomy (in the very concept of natural end) can be resolved, says Kant, only by placing “end” and “nature” on radically different planes, which has as a consequence that the concept of natural end can never be objectively established:

As a concept of a natural product it includes natural necessity and yet at the same time a contingency of the form of the object (in relation to mere laws of nature) in one and the same thing as an end; consequently, if there is not to be a contradiction here, it must contain a basis for the possibility of this thing in nature and yet at the same time a basis of the possibility of this nature itself and its relation to something that is not empirically cognizable nature (supersensible) and thus is not cognizable at all for us, in order to be judged in accordance with another kind of causality than that of the mechanism of nature, if its possibility is to be determined. (CJ, 267)

The contingency with respect to mechanical causality revealed by the form of organized beings, then, refers us not only to a different maxim to guide our research but refers us beyond nature as such, or at least toward a thinking of the whole of nature as grounded in something nonnatural and unknowable. The impossibility of getting to the end of our determinative-subsumptive judgments according to mechanical causality, the fact that we stumble over contingencies (namely, objects the organization of which is indubitable and yet escapes our mechanistic explanations), does not allow us to posit a purposive causality as objective (because we understand it not at all) but obliges us to think of the possibility of nature in general, the sphere of mechanical causality, as not itself falling, qua totality, under the law of that same mechanical causality that, in general, holds sway within it.

So the antinomy resides in the fact that, faced with organized beings, we find ourselves unable not to appeal simultaneously to both causalities, and its resolution consists in referring purposive causality (which had appeared in nature in the form of organized beings, which are contingent with respect to mechanical causality) to the unknowable (and hence super-natural) realm of the supersensible. The concept of natural end will, then, remain forever problematic and will never be able to motivate dogmatic responses, be they positive or negative. The concept of purposive causality does indeed exist objectively (as in human art and technology), as does that of a mechanical natural causality; what cannot attain to such objectivity is the concept of a natural purposive causality, for such a concept cannot be drawn from experience (by abstraction, as we were just saying) or posited as necessary for an experience to be possible. Such a concept will remain forever problematic because even if, per impossibile, I managed to show that some natural objects really were the intentional productions of a divine artistic understanding, I would in so doing simply withdraw them from the domain of nature, whereas the problem was that of knowing now to understand them as natural products. We can preserve the naturality of nature only by invoking in a merely problematic way a concept the (positive or negative) objectivity of which will never be established. Nature needs this strange concept of a nature other than simply natural nature, this phantom nature of which all we know is that we need it. I will attempt to show that once again this structure is that of transcendentality itself.

The fact that this concept is fundamentally problematic, then, does not above all mean that we could do without it. The problematic concept of natural end is in a certain sense necessary. How? In that it is necessary for necessity. We saw earlier that the study of nature in view of knowledge must lean toward the maxim of judgment according to mechanical causality, a judgment that is perfectly in accord with the necessary, determining, character of the law of nature (even if that law be, in the event, an empirical one). But the movement of Kant’s thought, which consisted initially in “suspending” this investigation to answer the speculative, transcendental call of the other maxim, leads to the point where, at least when faced with organized beings, it is impossible for us not to invoke the two causalities simultaneously and to refer the second to the unknowable supersensible. But this (“subjective,” regulative, problematic) reference of purposive causality beyond nature as such turns out, after the suspension of scientific investigation, to be indispensable to the progress of that investigation and even to its possibility as such.

Because, as Kant says in §75, the fact of making the concept of natural purposiveness merely a principle of my judgment (the result of the particular, limited, constitution of human rationality) merely confirms the necessity of this concept that escapes natural necessity: “It is in fact indispensable for us to subject nature to the concept of an intention if we would even merely conduct research among its organized products by means of continued observation; and this concept is thus already an absolutely necessary maxim for the use of our reason in experience” (CJ, 269; emphasis added). This necessity as regards organized beings taken piecemeal, as it were, cannot fail, as we have seen, to lead us to thinking about nature as a whole. Now Kant, having established—contrary to the apparent movement of the beginning of the antinomy, which seemed to concede priority when it came to knowledge to the maxim of mechanical causality—that we must necessarily invoke the purposive maxim for all knowledge of organized beings, wants to move on to the totality without giving up on the question of knowledge (the necessity of the purposive maxim for knowledge of some beings must lead us to try it out on nature as a whole) but does not want to suggest that the problem is still essentially situated on the same level. Whence a paragraph that has to start out on the terrain of knowledge in order to jump beyond it:

It is obvious that once we have adopted such a guideline [ein solcher Leitfaden] for studying nature and found it to be reliable we must [müßen] also at least attempt to apply this maxim of the power of judgment to the whole of nature, since by means of it we have been able to discover many laws of nature which, given the limitation of our insights into the inner mechanisms of nature, would otherwise remain hidden from us. [Do not, however, imagine that that is the principal reason for performing that extension:] But with regard to the latter use this maxim of the power of judgment is certainly useful, but not indispensable, because nature as a whole is not given to us as organized (in the strictest sense of the term adduced above). By contrast, this maxim of the reflecting power of judgment is essential [essentially necessary, wesentlich notwendig] for those products of nature which must be judged only as intentionally formed thus and not otherwise, in order to obtain even an experiential cognition of their internal constitution; because even the thought of them as organized things is impossible without associating the thought of a generation with an intention [mit Absicht]. (CJ, 269)

So the finalist maxim leads us, on the basis of its grounding at the level of knowledge, to ask the question of the totality of nature. This question cannot, strictly speaking, be situated at the level of empirical knowledge (the maxim can always serve as a maxim for knowledge but does not essentially enter into the thought of nature as totality). But at least this much follows: The purposive maxim has turned out to be necessary as regards a part of nature (namely, organized beings). And this necessity of the purposive maxim also makes necessary the thought of a certain contingency of nature (thought of as mechanism). But if nature is therefore not thinkable (by us) as entirely necessary, then nature qua totality must itself be considered to be contingent and must therefore, still for us, be thought of as the work of a super-natural intelligence:

Hence natural things which we find possible only as ends constitute the best proof of the contingency of the world-whole [der vornehmsten Beweis für die Zufälligkeit des Weltganzen], and are the only basis for proof valid for both common understanding as well as for philosophers of the dependence of these things on and their origin in a being that exists outside of the world and is (on account of that purposive form) intelligent [verständigen]; thus teleology cannot find a complete answer for its inquiries except in a theology. (CJ, 269)

So by dint of explicating what is merely a maxim of the faculty of reflective judgment, Kant has sent us off beyond nature as such, toward God. But, according to a movement that is none other than the becoming-transcendent of the transcendental, this movement runs the risk of carrying us much too far. The thought of a purposive causality in nature can only refer us beyond nature to God, but it will never demonstrate the existence of God, which remains, then, here at least, a purely subjective necessity of our limited faculty of judgment. We can never establish the existence of God in this way because, as Kant recalls, we can never see natural ends as such, but only certain objects the thought of which we cannot even formulate without judging according to final causality.

The whole teleology, which thus finds its end in a theology, is radically subjective. What does this mean here? That human reason, which is, ex hypothesi, limited compared to the reason of other possible rational beings, cannot do without the purposive maxim (or, therefore, its theological consequence) and that the limit of reason that makes the maxim necessary for us simultaneously makes impossible for us any objective proof in this regard. Such an objective proof (which would establish that final causality is constitutive of the object and not merely regulative for the judgment we are in a position to bring to bear on the object) would be equivalent to showing that the maxim in question was valid not only for us but also for superior intelligences. And one senses the impossibility of such a demonstration: To the very extent that human reason is limited, any conception of a reason that could go beyond those limits is itself limited by the very limits it would be claiming to go beyond, according to the very structure of transcendentality, the structure of the Grenze—we transgress the frontiers of human reason only to the extent that we are transgressing frontiers, which thus by definition hold us short of what such a transgression wanted to achieve. One can cross the frontiers of human reason only the better to draw them. Any “objective” proof is thereby circumscribed, and everything that is not susceptible of objective proof (including the very objectivity of that objectivity, as we shall see in a moment) never will be. So it is not simply that we are condemned to appeal to the purposive maxim because of a negative limit from which superior intelligences would not suffer, because the very idea of such intelligences is positively produced only by the structure of that same limit (and so the infinite is finite). The limit produces both what is limited and what is beyond it, but what is beyond it is beyond it by the very fact of this limit. Reason thus produces itself both as understanding (at home inside the limit, grounding an “objectivity” entirely subordinate to the limit) and as reason (the limit itself as a bilateral frontier folding its outside back toward its inside, whence it cannot fail to externalize itself, to turn itself back out). This is literally what Kant says in the “Remark” that is §76:

Reason is a faculty of principles, and in its most extreme demand it reaches to the unconditioned, while understanding, in contrast, is always at its service only under a certain condition, which must be given. Without concepts of the understanding, however, which must be given objective reality, reason cannot judge at all objectively (synthetically), and by itself it contains, as theoretical reason, absolutely no constitutive principles, but only regulative ones. One soon learns that where the understanding cannot follow, reason becomes excessive, displaying itself in well-grounded ideas (as regulative principles) but not in objectively valid concepts; the understanding, however, which cannot keep up with it, but which would yet be necessary for validity for objects, restricts the validity of those ideas of reason solely to the subject, although still universally for all members of this species, i.e., understanding restricts the validity of those ideas to the condition which, given the nature of our (human) cognitive faculty or even the concept that we can form of the capacity of a finite rational being in general, we cannot and must not conceive otherwise, but without asserting that the basis for such a judgment lies in the object. (CJ, 271–72)

One senses that all of Kant’s frontiers are beginning to tremble here. On the basis of the apparently anodyne antinomy of the faculty of judgment (judge according to mechanism or purposiveness), Kant, having apparently defused the antinomy by appealing to the merely subjective character of the maxims in question, has, by exploring further the very subjectivity of the purposive maxim, exceeded nature as a whole to ground it in a principle of subjectivity that is so radical that it removes any “objective” definition of objectivity. The objectivity of the objective is overtaken by a principle that can never achieve objectivity (while dealing it a radical blow) and whose supposed “subjectivity” de jure precedes the very opposition of subjective and objective. The understanding (and thereby any use of the faculty of determinative judgment) is in the service of a reason that is merely the exposure of that understanding to its own frontier, where its finitude is, as it were, denounced before a principle of excess that is none other than the supposed regulative principle of reflective judgment itself. Qua rational (excessive), the reflective judgment (always, it will be remembered, in search of its law) “grounds” the understanding in an “objectivity” that is strangled by the subjective finitude of a reason that is only its own turning back toward an absolute exteriority. It will follow that henceforth Kant will have the greatest difficulty maintaining clear frontiers between the subjective and the objective, the constitutive and the regulative, the dogmatic and the problematic, the understanding and reason, the critical and the doctrinal.

This consequence can, clearly enough, give rise to two apparently opposite movements. On the one hand, Kant can register a certain satisfaction of human reason, grounded on the resemblance between what is objective and what is subjective but necessary to the subject insofar as it can have objects. He says in §75:

Now if this proposition, grounded on an indispensably necessary maxim of our power of judgment, is completely sufficient for every speculative as well as practical use of our reason in every human respect, I would like to know [so möchte ich wohl wissen] what we lose by being unable to prove it valid for higher beings, on purely objective grounds (which unfortunately exceed our capacity). (CJ, 270; Paul Guyer adds a question mark here not found in the German text)

But this implied rhetorical question (“I would like to know”) and the satisfaction it might indicate if we take it to be a rhetorical question also invite a more literal reading:13 I would like to know, I would really like to know, but everything I have said shows that unhappily I can never know, only judge. Human reason, just now self-satisfied, is here somewhat needy and miserable. What we lack will never be the object of theoretical knowledge for us, just because we lack it. But just this indeterminate lack is the principle of judgment as reflective judgment—that is, as characterized by lack (of the law for the case). The self-sufficiency and self-satisfaction of human reason are thus grounded in its radical insufficiency as understanding, its deficit of objectivity that thus becomes the reason of (human) reason itself.

Human reason is in this way characterized by a certain necessity of contingency. The nonobjectivity of the principle of purposive causality, the absence of the law for some cases at least, the contingency in and of nature are necessary for our understanding. This curious necessity drives Kant to give “examples” in this remark to §76, “examples, which are certainly too important as well as too difficult for them to be immediately pressed upon the reader as proven propositions, but which will still provide material to think over and can serve to elucidate what is our proper concern here.”14

And indeed, two examples will prepare the way for what we have just called the necessity of contingency. First “example”: It is necessary for human understanding to make a distinction between possibility and actuality (Wirklichkeit). We can represent a thing to ourselves (by its concept) without that thing’s being actually given to us. The actuality of things comes to us solely via our sensibility, according to the first Critique’s intuitus derivativus (B72). But we can imagine (in the modality of the possible, then) an understanding that would give itself objects directly by thinking them, without it being necessary for it to go via sensibility. For such an understanding, the distinction between possibility and actuality would be irrelevant, whereas for us (who need such a distinction), such an understanding is thinkable only according to one of the terms of that same opposition. An understanding (impossible for us) that would not recognize the meaning of possible can be represented (possible for us). One of terms of this opposition (necessary for us) allows us to think its nonpertinence. But this radicalization of the possible is only possible. The possible (as opposed to the actual) remains necessary for us so we can represent to ourselves, in the mode of the possible, its disappearance into an identity with the actual. The possibility of representing the end of the possible thus confirms the necessity of the possible. The very fact of the separation of sensibility and understanding allows us to think their nonseparation.

This separation of the possible and the actual for us is none other than the separation between the necessary and the contingent. The absolutely necessary (where the separation of the possible and the actual would lose all relevance) is only possible and thus contingent. The contingent is necessary for us to be able to represent its disappearance in the absolutely necessary.

Second “example.” Just as we must (necessarily) admit necessity (in the mode of the contingent), we are conscious of the moral law, and we must thus presuppose causality through freedom:

Now since here, however, the objective necessity of the action, as duty, is opposed to that which it, as an occurrence, would have if its ground lay in nature and not in freedom (i.e., in the causality of reason), and the action which is morally absolutely necessary [die moralisch-schlechtin-notwendige Handlung] can be regarded physically as entirely contingent (i.e., what necessarily should [sollte] happen often does not), it is clear that it depends only on the subjective constitution of our practical faculty that the moral laws must [müßen] be represented as commands (and the actions which are in accord with them as duties), and that reason expresses this necessity not through a be [Sein] (happening [Geschehen]) but through a should-be [Sein-Sollen]. (CJ, 273)

So there is necessity and necessity. The morally necessary is physically contingent (what ought necessarily to happen must not necessarily happen, and even most often does not happen, might even never happen, is totally contingent, necessarily contingent). Why is this? Because reason is not mankind’s only faculty, it is related to sensibility and thereby to nature (to a possible nature). The thing that gives nature as an object of possible experience, and as object of such an experience, as phenomenon, what gives rise to theoretical knowledge and to the necessity of Sein or Geschehen—namely, sensibility—by the same token withdraws Sein from the intelligible world and gives it the Sollen. And this is another version of the paradox of the possible and the actual that we were just reading. The fact of sensibility here gives the separation of possible and actual as the separation between Sein and Sollen:

Which would not be the case if reason without sensibility (as the subjective condition of its application to objects of nature) were considered, as far as its causality is concerned, as a cause in an intelligible world, corresponding completely with the moral law, where there would be no distinction between what should be done and what is done, between a practical law concerning that which is possible through us and the theoretical law concerning that which is actual through us. (CJ, 273)

Just as it is possible for us to represent to ourselves a world in which the modality of the possible would not have a place, it is equally possible—and in fact we ought, this is duty itself—to represent to ourselves a world from which duty would be absent, in which Sollen would merge with Sein, in which the law would not take the form of a command that constrains us. It is possible for us to do so, and in fact we ought to do so; this means also that we are free to do so. That is what freedom is. If freedom qua “formal condition” of this intelligible world, in which Sollen and Sein merge, remains for us an excessive concept (überschwenglicher Begriff) that can determine nothing as to objective reality, this is because “freedom” has meaning for us (doubly for us: once as our freedom itself, a second time as our concept or Idea of that freedom) only to the extent that it does not dissolve in a purely intelligible world (that would by this very fact be unintelligible for us) in which causality by freedom would merge with causality by necessity and would find its end in that merging. Which is, of course, the meaning of what Kant is trying to think under the name of regulative principle, which brings with it the whole issue of the Ideas of Reason.

So, for us, freedom will consist exemplarily (recall that we are dealing with “examples” in §76) in thinking nonobjectively the end of freedom in the (un)intelligible world. To give any meaning to freedom, we must, then, be in part constrained by a mechanical necessity (in sensible nature) and thus equally constrained by the moral law that presents itself to us as imperative and command. And this free thought of freedom—which is free only by dint of constituting nothing objectively, by dint of its ought-to-be-realized in a world in which, if in fact it were realized, it would not be free—which, says Kant, by making of freedom a mere regulative principle “does not determine the constitution of freedom, as a form of causality, objectively,” this thought in no way devalues freedom by refusing it any objective status but “indeed does so with no less validity than if it did determine freedom objectively [mit nicht minderer Gültigkeit, als ob diese geschähe].” Being able to represent to ourselves a world in which freedom and necessity would merge like actuality and possibility, then, we can only represent such a world to ourselves, and the only chance we have of representing it is by hanging on tight to the very distinction the disappearance of which that representation seems to promise. Our freedom depends on our possibility of representing its end and so depends on our having of freedom a final idea. According to a schema that Kant is about to refine and aggravate, but which we shall see to be the schema of the Ideas as such, the Idea of freedom is an Idea only to the extent that it projects its own end (i.e., the moment at which it would become concept rather than Idea, constitutive rather than regulative principle, or at least the moment at which these distinctions would disappear in what they made possible). Kant says that it has no less value than if it were the case; but if it were the case it would have no value at all, because “being the case” here is precisely geschähe, and as we have seen, geschehen is on the side of Sein, not of Sein-Sollen. The case as such cannot be free.

Freedom considered as the formal condition of an intelligible world is, then, inaccessible to us as an objective concept and must remain forever problematic. But the freedom of freedom consists in precisely this problematic status. And it is precisely this structure (which is, I repeat, the same as that of transcendentality as such) that Kant will formalize under the name of purposiveness. For the two “examples” he has just given serve to elucidate a thinking of purposiveness that thinks together lawfulness and contingency even before the distinctions between concept and Idea, theory and practice, etc.

Recall the abyssally analogical structure of §76, which condenses the very structure of Kant’s thought, at least as it is expounded in the introduction to the third Critique. The examples given by Kant are linked by operators of analogy that have a cumulative effect:

Just as [So wie] in the theoretical consideration of nature reason must assume the idea of an unconditioned necessity of its primordial ground, so, in the case of the practical, it also presupposes [so setzt sie auf] its own unconditioned (in regard to nature) causality, i.e., freedom, because it is aware of its moral command [. . .] Likewise [Ebenso] [. . .] (CJ, 273–74; emphasis added)

The first two examples have opened the way to what will bring the series to an end. They have allowed us to think what one might call a certain free possibility that can only be thought in the evanescent perspective of its own disappearance, a perspective called “Idea.” This thought will now find a formulation that, in the context of a judgment supposed to give a certain mediation between theory and practice, will in fact enfold and transgress them at one and the same time. Ebenso, then:

Likewise, as far as the case before us is concerned [was unsern vorhabende Fall befrifft: that is, the case of judgment (which is always judgment of a case) insofar as it is divided in its antinomy between mechanism and purposiveness], it may be conceded that we would find no distinction between a natural mechanism and a technique of nature, i.e., a connection to ends [Zweckverknüpfung] in it, if our understanding were not of the sort that must go from the universal to the particular, and the power of judgment can thus cognize no purposiveness in the particular, and hence make no determining judgments, without having a universal law under which it can subsume the particular. But now since the particular, as such, contains something contingent [etwas Zufälliges] with regard to the universal, but reason nevertheless still requires unity, hence lawfulness, in the connection of particular laws of nature (which lawfulness of the contingent is called purposiveness [emphasis added]), and the a priori derivation of the particular laws from the universal, as far as what is contingent in the former is concerned, is impossible through the determination of the concept of the object, thus the concept of the purposiveness of nature in its products is a concept that is necessary for the human power of judgment in regard to nature but does not pertain to the determination of the objects themselves, thus a subjective principle of reason for the power of judgment which, as regulative (not constitutive), is just as necessarily valid for our human power of judgment as if it were an objective principle. (CJ, 274)

Everything trembles here. This regulative (and therefore “subjective”) concept, which alone can provide for a lawfulness of the contingent (but without such lawfulness the contingent would remain contingent, natural necessity would be partial and incomplete, depressing for reason), which cannot by definition achieve necessity, is “just as necessarily valid [. . .] as if it were an objective principle.” And even more necessary than necessity, because without this regulative principle, judgments of objective necessity would have no meaning, would not converge toward the unity demanded by reason, which is given only in teleology. Judgment would thus be the place from which proceed necessity and contingency, mechanism and freedom, archi-contingent necessity of contingency, archi-necessary contingency of necessity. This preobjective condition of objectivity, this place where all the key terms of Kantian thought can be exchanged and analogized, is clearly enough the blind origin of critique itself, but thereby also its essential resource, without which there would be no critique, properly speaking, but merely preparation or propaedeutic for doctrine. This point, reached via a thinking of teleology, must also precede all teleology and, therefore, all theology too.

Can we satisfactorily call this point, which precedes Kant’s distinction between speculative and practical, “finitude”? Logically no, because we have seen it to precede the distinction between finite and infinite. “Finitude,” especially when thought in the limiting form or privileged example of human mortality, can give the condition described by Kant a certain “existential” pathos that has had its day but that remains dependent on a thinking of God (of a God who is essentially, always already, dead), which falls short of Kant’s thinking in this regard. It would not be difficult to show that Kant’s thinking about God is merely thinking the same blind spot as fundamentally analogical, according to the ana-logics we follow in the Appendix around the distinction between Grenze and Schranke, because it is precisely around God that Kant pushes furthest with his analogism, and this thinking again passes through the problem of cosmopolitanism.

In the “Methodology of Teleological Judgment,” which Kant places as an appendix to the third Critique, and having reinforced the subordination of mechanism to teleology that we saw happening little by little in the antinomy, Kant, via the question of what the ultimate end of nature considered as a teleological system can be, comes to pose again the question of war and peace. Man is indeed the ultimate end of nature but only insofar as “he is the only being on earth who forms a concept of ends for himself” (CJ, 295). The ultimate end would be, then, that there be ends, postultimate ends, the end that there be always more ends. Since our reading of Aristotle in Chapter 1, we immediately suspect that this structure of the postultimate will bring us back to the archi-originary. What comes at the end is the possibility of seeing that there must be an end and that that end is the possibility of setting oneself still further ends. Clearly, it is not enough to say in this context simply that the ultimate end is man himself, because man is, at least in part, part of nature considered as mechanism. So the ultimate end cannot be man as such but something in man, and this ultimate end can be either an end that nature can bring about through its beneficence (i.e., happiness) or else something that man can realize by exploiting nature (i.e., culture). Obviously the first of these cannot be the one, because man comes up with such a diversity of ideas as to what constitutes happiness (a constant refrain in Kant’s moral philosophy) and also, on the one hand, it is not in man’s nature to be satisfied with what nature gives him and, on the other, nature spares him none of the trials and tribulations suffered equally by all other animals. And this is not all. Man also has natural dispositions that are “absurd.” What are these dispositions? Exactly those that earlier Kant wrote off to natural providence—namely, the oppressive and warlike tendency that makes man (considered as a natural being) work toward the destruction of his species. So here is a paradox: as ultimate end of nature, man, being of nature, is not the ultimate end (because if he were, he would be remaining in nature viewed as mechanism). If so-called natural man were the end of nature, nature would not even be a teleological system but pure mechanism, and thus would have no ultimate end. For man to be the ultimate end of nature considered as a teleological system of ends (and he must be, otherwise nature would be merely a mechanism without end or meaning, even for scientific exploration), it must be that what in man constitutes this end is a capacity to go beyond nature as such. What man must do to be the final end of nature is to project another end beyond nature, and that end is no longer in nature as its ultimate end but outside nature as its final end (no longer as letzter Zweck, then, but as Endzweck, according to Kant’s distinction). Man is the final end only because he can give himself ends after the ultimate end, and that can only be done by culture, because culture is precisely the development of the capacity for ends in a rational being. Properly speaking, then, the ultimate end of nature is not man but culture: “Only culture can be the ultimate end that one has cause to ascribe to nature in regard to the human species” (CJ, 299). The ultimate end of nature is thus the end of nature in culture, its apparent other, or else the becoming-culture of nature on its human frontier.

Culture is not a simple thing, and Kant distinguishes between the culture of skill (Geschicklichkeit) and the culture of discipline (Zucht [Disziplin]). These two determinations of culture are situated, we might say, on two sides of a frontier. Although skill is part of culture rather than nature, it is nonetheless on the side of nature, as a condition of possibility of the realization of ends of all sorts (“the foremost subjective condition of aptitude for the promotion of ends in general,” as Kant puts it [CJ, 299]), but it cannot of itself guide the will in the choice of ends to be realized; the culture of discipline will take care of that by behaving in a negative way toward what in man is natural—namely, passions and desires:

[It] consists in the liberation of the will from the despotism of desires [Begierden], by which we are made, attached as we are to certain things of nature, incapable of choosing for ourselves, while we turn into fetters the drives [Triebe] that nature has given us merely for guidance [die uns die Natur nur statt Leitfäden beigegeben hat] in order not to neglect or even injure the determination of the animality in us, while yet we are free enough to tighten or loosen them, to lengthen or shorten them, as the ends of reason require. (CJ, 299)

Everything we followed in Kant’s political writings belongs to the culture of skill, which, while remaining on the side of nature, never stops leaving nature behind (for, as we saw, politics is nothing but the leaving-behind of nature, never accomplished by mankind). All the negativity we saw at work in natural providence is summed up here in a striking formula: The culture of skill gives rise to what Kant calls a “splendid misery [glänzende Elend, brilliant distress],” because not only must there be inequality, and therefore oppression, among men for the culture of skill to develop, says Kant clearly with Rousseau in mind, but the development made possible gives rise to an increase of misery (Plagen: torments) on both sides (the oppressors as much as the oppressed) either through external violence or internal insatiability. This brilliant distress is nonetheless linked to the end of nature that is reached here, even if this end is not ours—which is where we come back to cosmopolitanism and its inner inhibition:

The formal condition under which alone nature can attain this its final aim [Endabsicht] is that constitution in the relations of human beings with one another in which the abuse of reciprocally conflicting freedom is opposed by lawful power in a whole, which is called civil society; for only in this can the greatest development of the natural predispositions occur. For this, however, even if humans were clever [klug] enough to discover it and wise enough to subject themselves willingly to its coercion, a cosmopolitan whole, i.e., a system of all states that are at risk of detrimentally affecting each other, is required. (CJ, 299–300)

This is, then, the argument that we followed in the “Universal History” text. But just as we were able to see in Perpetual Peace, this cosmopolitan perspective must not be realized, and this nonrealization, in its very violence, forms part of the realization of the end of nature (soon not to be ours)—namely, the development of culture as aptitude for ends. Culture (of skill, still) is no other than that brilliant distress that would be overcome by the cosmopolitanism that must not come about:

In its absence, and given the obstacles that ambition, love of power, and greed, especially on the part of those who are in power, oppose to even the possibility of such a design, war (partly of the kind in which states split apart and divide themselves into smaller ones, partly of the kind in which smaller ones unite with each other and strive to form a larger whole) is inevitable, which, even though it is an unintentional effort of humans (aroused by unbridled passions), is a deeply hidden but perhaps intentional effort of supreme wisdom if not to establish then at least to prepare the way for the lawfulness together with the freedom of the states and by means of that the unity of a morally grounded system of them, and which, in spite of the most horrible tribulations which it inflicts upon the human race, is nevertheless one more incentive (while the hope for a peaceful state of happiness among nations recedes ever further) for developing to their highest degree all the talents that serve for culture. (CJ, 300; emphasis added)

This culture of skill, which is on the side of nature, then, develops while remaining in nature (animality, warfare). Kant must now distinguish it from the other culture, the culture of discipline, which we said was on the other side of the natural frontier. But he will have the greatest difficulty in operating this distinction, and for good reason. On the one hand, as we saw, our natural drives (purposive for us qua animal species) encourage the development of culture while making more difficult (erschweren) the development of man as a rational being; on the other hand, then, there is the other culture, the culture of discipline, which must take us beyond nature. This culture, qua culture, must still take off from nature as its ultimate end (letzter Zweck). Where the culture of skill is answerable to an Endabsicht of nature, a final intention or aim, the culture of discipline will answer to a “purposive effort [ein zweckmässiges Streben],” of nature still, in view of an Endzweck yet to come. This culture of discipline must oppose all the penchants that, as we just saw, contribute to the culture of skill, because the “end of nature [der Zweck der Natur]” remains that of getting mankind over the natural frontier.

So there must be a cut. Remaining in nature (or, as we would say, on the frontier of nature), we will never arrive at the end (i.e., the end of all the other ends, the end even of the final end [culture], which remains fundamentally natural). All these ends must have their end, but for it to be their final end, their Endzweck, it must be cut from them. If this final end of ends were merely the end of nature, it would never make it out. So all previous ends (including the ultimate end, the letzter Zweck) must be finalized by a further end, which is their true end only by not exactly being their end, only by not belonging to them. Which is what Kant says, literally, at the beginning of §84: “A final end is that end which needs no other as the condition of its possibility [Endzweck ist derjenige Zweck, der keines andern als Bedingung seiner Möglichkeit bedarf]” (CJ, 301). The Endzweck does not need that of which it is the end to be its end. And reciprocally, that of which this end is the end (nature) cannot bring it about, because in that case the end would be dependent on and conditioned by the very thing (nature) that knows nothing unconditioned. Everything in nature is conditioned, so the end of ends cannot be found in it.

Kant considers this end to be human, still, in the sense that man is supposedly the only natural being to act by final causality all the while representing to himself the final finality of this finality as being necessary, unconditioned, and thus outside nature. But the humanity thus identified is again divided according to the fractal frontier demultiplication of nature that is our privileged “object.” For if, on the one hand, one can always say that man is the final goal of nature, to the extent that natural affairs must have some such end if they are not to fall back into the blind necessity or contingency of the natural mechanism, on the other hand, one must say that, qua final end, precisely, man becomes cut from all teleology and all nature:

The being of this sort is the human being, though considered as noumenon: the only natural being in which we can nevertheless cognize, on the basis of its own constitution, a supersensible faculty (freedom) and even the law of the causality together with the object that it can set for itself as the highest end (the highest good in the world).

Now of the human being (and thus of every rational being in the world), as a moral being, it cannot be further asked why (quem in finem) it exists. His existence contains the highest end itself, to which, as far as he is capable, he can subject the whole of nature, or against which at least he need not hold himself to be subjected by any influence from nature. (CJ, 302)

Nature in its totality, then, is teleologically subordinate to a natural ultimate end (mankind) and again to a nonnatural final end (mankind again). But qua end cut from that of which it is the end, qua end without finality (Zweck ohne Zweckmässigkeit, we might say), man (noumenal man, certainly) is no longer human. Just as the end of nature (in nature) is not quite natural, so the end of man (in man) is not quite human. Teleology arrives at its end only by cutting itself from its end to which it is, however, endlessly tending.

Of course Kant appeals to God to solve the problem that is persisting here and that is none other than the insistence of the frontier, now pushed inside as the frontier between phenomenal man and noumenal man. But as we shall see, God, called in to put an end to the end without end of interrupted and interminable teleology, will turn out to be just another name for the frontier, Deus terminus.

Having dismissed physical theology (because we can conclude nothing as to the supreme cause of nature while remaining within nature), Kant turns to ethical theology. He is trying to explain how we get across the frontier between moral teleology (for which we are still beings in the world, and for which the point is to understand the relation between our ends and the end of those ends in the world—i.e., reply to the question of knowing, given that we are beings in the world among other beings, how we can bring about that end in that context) and a moral theology. Moral theology necessarily appears, because moral teleology cannot help but ask whether we need to go looking beyond the world for a principle that can explain the relation in the world between the moral and the natural.

As always when in comes to morality, Kant systematically invokes here what is obvious, in the sense of what anyone at all can understand, “even the most common understanding” (CJ, 308) “the commonest judgment of healthy human reason” (CJ, 309), “even the most common human reason” (CJ, 314). What even this common understanding can understand is that nature would have no end were it not for mankind and that it is qua moral being that mankind can lay claim to that place. What is man qua moral being? A rational being “under moral laws” (i.e., rational in the sense of a type of reason capable of giving itself unconditional laws):

The moral laws, however, have the unique property that they prescribe something to reason as an end without a condition, thus do exactly what the concept of a final end requires; and the existence of such a reason, which in the relation to ends can be the supreme law for itself, in other words, the existence of rational beings under moral laws, can alone be conceived of as the final end of the existence of a world. If, on the contrary, this is not the case, then there is either no end at all for the existence of a world in its cause, or it is grounded in an end without a final end. (CJ, 315)

For there to be an Endzweck, there must be a reason able to give itself an end according to the structure of an Endzweck (i.e., an unconditioned end). The unconditioned end is that there be an unconditioned end. The point is not to have such ends realized (impossible, and this impossibility is a positive condition of possibility for moral laws and, as we shall see, for God) but to be able to give them to oneself as ends to-be-realized. Reason giving itself its unconditional law, reason being unto itself its own supreme law, this end cut from all ends, or rather reason able to think itself as its own supreme law, is the only possible Endzweck for any world at all. To sum up a little brutally: The end is that one can think the end. The end is the end to the extent that it manages to think itself as the (unconditioned) end and is thus the end of everything. And the end, in this thinking the end, gives itself an end that is unrealizable (according to natural causality), an end that must not be realized but that nonetheless remains to be realized. To safeguard the tendentiality of this to-be-realized we are obliged to postulate a moral author of the world—namely, God.

God occupies in this reasoning a place that is extremely difficult to pin down and that Kant will further complicate in the paragraphs to come. The final end can only be mankind able to set himself moral (unconditioned) ends. But as these ends, if they are to be moral, must be capable of realization in the natural world in which there reigns only mechanical causality—that is, as the practical necessity in which we find ourselves to realize in the world what is prescribed by moral laws “is not congruent with [stimmt . . . nicht . . . zusammen] the theoretical concept of the physical possibility of producing it if we do not connect our freedom with any other causality (as a means) than that of nature” (CJ, 316)—we must admit, subjectively, the existence of God as moral author of the world, for this accord to be possible.

This “moral proof” of the existence of God is valid only for the finite beings we are (i.e., beings who cannot find the agreement between their freedom and natural causality). It “proves” the existence of God without demonstrating it, but it must (müßen) be admitted in order to maintain the to-be-realized of the moral law. If we were content to stay on the nature side of the frontier, the moral law could never apply to any case; but if we passed too rapidly to the other side, there would be no cases (contingencies) at all. For the moral law to be prescriptive, what is prescribes must be perpetually to be realized, and this, as we saw, is what gives rise to the teleological structure. And for this teleological structure to maintain itself in its purposiveness before its end, without being realized, then we must admit the moral proof of God.

Just like the principle of natural purposiveness itself, which, recall, was as necessary as necessity, necessary with a necessity before all necessity, so God is necessary for the moral law, which is the only truly final end of all the structures of purposiveness we have gone through. On the level of practical necessity, strictly speaking, God is, indeed, less necessary than the moral law as such, but, just as the prenecessary necessity of natural purposiveness was, in the end, necessary for natural necessity itself, the necessity of God, less necessary than strict moral necessity, is necessary (for us) for this practical necessity to have a chance—like peace—in the perpetually relaunched structure of its to-be-realized.15 The whole problem of Kant’s thinking of God will consist in keeping him in this exiguous but capital place. And in fact, by giving a “proof” in this way, even a “subjective” proof, Kant has no doubt already proved too much, which is why he devotes the next paragraph to a limitation (Beschränkung) of this moral proof, which always runs the risk of going too far in the direction of theoreticist objectivity, in the very crossing of the frontier between teleology and theology.

Kant repeatedly takes up this argument in these paragraphs. Here in §86: in following the principle of theoretical reflective judgment, we can sufficiently establish that the possibility of a world containing purposive products of nature entails the thought of a final end. But the concept of a final end is a concept of our practical reason alone:

No use of this concept is possible except solely for practical reason in accordance with moral laws; and the final end of creation is that constitution of the world which corresponds only to that which we can give as determined in accordance with laws, namely the final end of our pure practical reason, insofar as it is to be practical [so fern sie praktisch sein soll].—Now in virtue of the moral law, which imposes this final end upon us, we have a basis for assuming, from a practical point of view, that is, in order to apply our powers to realize it, its possibility, its realizability [Ausführbarkeit], hence also a nature of things corresponding to that end (since without the accession of nature to a condition that does not stand within our own power its realization would be impossible). Thus we have a moral ground for also conceiving of a final end of creation for a world. (CJ, 320)

So this reasoning gives us reasons to believe in a final end of creation. But to go from that to the conclusion that there must be a God, two more inferences are needed: first, that the author of the world must be intelligent (which we have already assumed for the purposive products of nature) and secondly that he must be moral (otherwise this is not God but a demon or a demiurge). These inferences can in no way give rise to an objective proof but operate merely for our finite reason: We cannot claim to say that such a being exists but only that we have no way of understanding, without supposing such an author, how to refer the moral law, and thus its final end, to its objects. For us, the abyss to be crossed between moral causality and natural causality can be bridged only on condition that we suppose, purely for the use of our practical reason, the actuality or effectivity, the Wirklichkeit, but not the objective existence, of such a moral author-legislator.

God, then, is an idea that is necessary for us on the level of practical reason and in the mode of reflective judgment. If we attribute qualities to him (and we have seen in §86 that we must indeed attribute to him the qualities of omniscience, omnipotence, absolute goodness, etc.), we must above all not do so in a theoretical manner. What is proper to God is precisely to put an end to nature while remaining radically cut from it; so he must above all not be the object of a theoretical knowledge of the type we can have of nature. The fact remains that we must attribute qualities to him if the idea of God is not to remain completely indeterminate and thus useless. How are we to do so? Precisely by analogy. We think the qualities of the Supreme Being only by analogy, and we think these qualities by analogy: We do not know (erkennen) these qualities, but we think (denken) them. So we think them by analogy with a knowledge we do not have; thought is the analogy of knowledge.

Kant’s difficulty here is patent and comes from the fact that he is trying to think together the frontier between speculative and practical reason, between knowledge and thought, and between constitutive and regulative principle. The difficulty, which the appeal to analogy is supposed to solve but which it complicates infinitely, consists in the fact that under the name of God, Kant is advancing a thought of relation as such, whereas he has at his disposal only a language of possible objects.

Thus just as we name a cause after the concept that we have of its effect (though only with regard to its relation [ihrer Relation] to the latter), without thereby meaning to determine its internal constitution intrinsically by means of the properties that are all that we know about such causes and which must be given to us by experience [. . .] so we must assume something that contains the ground of the possibility and the practical reality, i.e., the realizability, of a necessary moral final end; but given the constitution of the effect that is expected from it, we can conceive of it as a wise being ruling the world according to moral laws, and in accordance with the constitution of our cognitive faculties we must conceive of it as a cause that is distinct from nature, only in order to express the relation [Verhältnis] of this being transcending all of our cognitive faculties to the object of our practical reason, without on that account theoretically attributing to it the only kind of causality of this sort that is known to us, namely an intelligence and a will, indeed without meaning to make an objective distinction between the causality that we conceive in this being itself with respect to what is a final cause for us and its causality with respect to nature (and its purposive determinations in general); rather, we can assume this distinction only as subjectively necessary, for the constitution of our cognitive faculty, and as valid for the reflecting but not the objectively determining power of judgment. (CJ, 321–22)

All of this, then, on the side of possible knowledge. From the theoretical point of view, we know nothing of God objectively, but from the practical point of view, the same principle, here regulative for a reflective judgment, indeed becomes constitutive for a determinative judgment, but solely on the practical level:

But when it comes to the practical sphere, such a regulative principle (for prudence or wisdom)—namely, to act in conformity with something, as an end, which given the constitution of our cognitive faculties can only be conceived by us as possible in a certain way—is at the same time constitutive, i.e.,practically determining. (CJ, 322)

Kant’s problem is obvious. Up until now there was, not an identity, of course, but something like a proportionality that managed the relations between the couples theoretical-practical, concept-Idea, constitutive-regulative, determinative-reflective. Thus, for example, the whole slow drift of the antinomy of judgment toward the primacy of the teleological, and in the end the practical was doubtless programmed since the first Critique by the very distinction between concept and Idea of Reason. It was indeed the radically reflective nature of some judgments that led us to theology, with all the consequences that result for teleology and its internal cut. Here, however, as we arrive apparently toward the end, we discover (this is exactly the sense of what earlier we said was necessary for necessity) that this regulative side turns constitutive. This is again and always the problem of the applicability of the moral law and the to-be-realized of the final end. As the moral law is (for us) only moral and the law insofar as it is applicable (i.e., determinative with regard to its cases, as in morality we are dealing with constitutive principles), the figure of God has the job of making possible this conversion of the regulative into the constitutive while preventing any attempt to cross back over the frontier and to enter the domain of theoretical knowledge on such a basis.

This is why, having spent a paragraph congratulating himself as to the advantages of this way of conceiving of God as cut from all objectivity, Kant has to confront the question of exactly what has been proved by the moral proof thus constructed and duly limited. In other words, in what way are we to hold for true what is proved in the proof? Literally, what is the kind of holding-for-true (“der Art des Führwahrhaltens”) of this proof?16

Every proof, says Kant, following his usual analytic method, must not only persuade (überreden) but convince (überzeugen). Or rather, he adds immediately, already complicating the description in a way that is like the formalization of the structure of the to-be-realized, convince or tend to convince, to work on conviction, “auf Überzeugung wirke.” This means that proof must be based on objective grounds, because otherwise the understanding is simply seduced and not truly convinced. But a proof that tries to convince can in turn be of two sorts: either it seeks for what the object is in itself or else what it is for us; the first corresponds to the determinative judgment, the second to the reflective judgment. In this second case, the proof absolutely does not work on conviction if it is based on theoretical principles. If, however, it is based on a practical rational principle, it can still lay claim to a true moral conviction. But between conviction properly speaking and mere persuasion, we find proofs that work on conviction, that tend to convince, in the sense that they are on the way to conviction: “A proof, however, tends to conviction without actually convincing if it is merely led onto the path to conviction, i.e., if it contains only objective grounds for that, which, although they are not sufficient for certainty, are nevertheless of the sort that do not serve merely as subjective grounds of the judgment for conviction” (CJ, 327).

Kant will now show how the four degrees of holding-for-true that can be identified (namely: rigorous proof by syllogisms; proof by analogy; plausible opinion; hypothesis) cannot be attained by what claims to be a proof of the existence of God. But among the four possibilities thus dismissed, one poses some problems. And, as if by chance, this is proof by analogy.

As for a rigorous proof of the existence of God, Kant refers us to the first Critique, where he demonstrates that, given that by definition we cannot have an intuition of a being beyond nature, the concept of God remains forever problematic. But the question of analogy merits a longer treatment and is further complicated by a long note that will detain us in a moment.

One can, of course, think of one of two heterogeneous things, even on the very point of their heterogeneity [eben in dem Punkte ihrer Ungleichartigkeit], by means of an analogy* with the other; but from that respect in which they are heterogeneous we cannot draw an inference [Schliessen, conclude] by means of the analogy, i.e., transfer this characteristic of the specific difference from the one to the other. (CJ, 328–29)17

Analogy serves to think but not to conclude. The note called for by the word Analogie undertakes to clarify this assertion. In it, we read that analogy is the identity of a cause-and-effect relation despite a difference of objects. And before following the main text that will discuss the possibility of analogy between man and God as to technical activity, the note takes as a first example an analogy between man and beast with respect to this same technical activity:

Thus, in comparing the artistic actions of animals with those of human beings, we conceive of the ground of the former, which we do not know, through the ground of similar effects in humans (reason), which we do know, and thus as an analogue of reason, and by that we also mean to indicate that the ground of the artistic capacity in animals, designated as instinct, is in fact specifically different from reason, but yet has a similar relation to the effect (comparing, say, construction by beavers with that by humans). (CJ, 328n)

We are thinking these two (supposedly) heterogeneous things (human reason and animal instinct) according to the analogy of their effects and thus placing the analogy precisely at the point of heterogeneity between man and animal. This allows me to think animal instinct (which I do not know), but Kant will show that it does not allow me to conclude, by transferring the specific mark of mankind (reason) onto the beaver, that animals act rationally.

But, as the analogy must be good for something, I can nonetheless, in spite of the impossibility of concluding by analogy that animals are rational, conclude something by analogy:

Yet from the comparison of the similar mode of operation in the animals (the ground for which we cannot immediately perceive) to that of humans (of which we are immediately aware) we can quite properly conclude in accordance with the analogy that the animals also act in accordance with representations (and are not, as Descartes would have it, machines), and that in spite of their specific difference, they are still of the same genus as human beings (as living beings). The principle that authorizes such a conclusion lies in the fact that we have the same ground for counting animals, with respect to the determination in question, as members of the same genus with human beings, as humans, insofar as we compare them with one another externally, on the basis of their actions. There is par ratio. (CJ, 328n; tr. mod.)

That was the first exemplary analogy, then: An analogy to show what an analogy is, or rather what it is to conclude by analogy. In analogy, one can permit oneself to draw certain conclusions to the extent that there is paritas rationis—and we ought to be able to draw some conclusions about analogy as such from this example. The second analogy that follows will allow us to draw the right conclusions about analogy. This second analogy, in which the beast is replaced by God, has an analogical relation with the first. Ebenso, says Kant:

Likewise, in the comparison of the purposive products of the causality of the supreme world-cause in the world with the artworks of human beings, I can conceive of the former in an analogy to an understanding, but I cannot conclude to this property in the world-cause by means of the analogy; because here the principle of the possibility of such a conclusion is precisely what is missing, namely the paritas rationis, for counting the highest being as part of one and the same species along with human beings (with regard to their respective causalities). The causality of the being in the world, which is always sensibly conditioned (even its causality through understanding) cannot be transferred to a being that has no generic concept in common with the former except that of a thing in general. (CJ, 328; tr. mod.)

In the analogy between man and beast, then, we think these two heterogeneous things at the point of their heterogeneity, without for all that being able to cross the frontier of that heterogeneity to a conclusion that would posit that beasts are rational. But we can conclude that animals and man both act according to representations (instinctually for the former, rationally for the latter) for we have equal reason to think so in both cases. But the analogy between man and God, which also allows us to think two heterogeneous beings at the point of their heterogeneity (the nature of their causality), does not have an intermediate concept that would allow for a conclusion of the same type. The man-God analogy is thus only partially analogous with the man-beast analogy: It allows us to think that there is an analogy between man and God but not that there must be the possibility of concluding by analogy any objective characterization of the nature of God. Which is as much as to say that the man-God analogy is only analogically analogical, because here we are lacking precisely the paritas rationis that gives rise to the analogy in the first place. Or rather we might say that the man-God analogy, being only analogically analogical, capitalizes the being-analogy of the analogy.18

This much in the note. In the main text, there are also two analogies, and the second is the analogy between man and God. But this time the first, exemplary, analogy, the analogy of analogy itself, directly following from the definition of analogy we read a moment ago, is directly political:

But from that respect in which they are heterogeneous we cannot conclude by means of the analogy, i.e., transfer this characteristic of the specific difference from the one to the other. Thus, in analogy with the law of the equality of effect and countereffect in the mutual attraction and repulsion of bodies, I can also conceive of the community of the members of a commonwealth in accordance with rules of justice, but I cannot transfer the specific determinations of the former (the material attraction and repulsion) to the latter and attribute them to the citizens in order to conceive of a system which is called a state. (CJ, 329; tr. mod.)

As we saw earlier, State organization is thinkable only analogically: The right that holds together the members of the community must not be force but can be thought only by analogy with force. It is striking in Kant’s definition of analogy that the thing we want to think and the place where we must not conclude coincide. “That respect in which they are heterogeneous” is exactly what we are interested in, the very reason why we are resorting to the analogy at all: I can get a sense of right only by appealing to the force that it precisely is not, that it above all is not. Analogy is thus fundamentally inhibited and frustrating: It allows what it by the same token makes impossible (i.e., a thinking of heterogeneity as such). This is what will come to the fore in the second analogy given in the main text, in which, as in the note, we are dealing with God. Here too there is an analogy between the two analogies:

Likewise, we can very well conceive of the causality of the original being with regard to the things in the world, in analogy with an intelligence as the ground of the forms of certain products that we call artworks, as natural ends (for this occurs only for the sake of the theoretical or the practical use we have to make of this concept by our cognitive faculty, with regard to the natural things in the world, in accordance with a certain principle); but from the fact that among beings in the world the cause of an effect that is judged as artistic has to be attributed to intelligence we can by no means infer by an analogy that the very same causality that we perceive in humans must also pertain to the being who is entirely distinct from nature in regard to nature itself; because this touches the very dissimilarity in their effects that must be conceived between a sensibly conditioned cause and the supersensible original being, and thus cannot be transferred from the former to the latter.—In the very fact that I am to conceive of the divine causality only in analogy with an intelligence (a faculty with which we are not acquainted in any being other than the sensibly conditioned human being) lies the prohibition against attributing this intelligence in a strict sense to the supersensible original being. (CJ, 329)

It is precisely because it is analogical that it cannot be proper: The best proof that we do not know what we are talking about when we talk about God is that we are talking about him analogically. To say that we can only think of God as if he were an understanding is to say that God is precisely not an understanding. We home in on the point of heterogeneity between God and us, what makes God God, by making use of analogy, which teaches us that at this point of heterogeneity we understand nothing. The road of analogy is, therefore, constitutively blocked, it leaves us stuck at the very point where we thought we ought to be able to touch the goal. Faced with the possibility that the reader might feel a little disappointed here, Kant adds a note:

One does not thereby lose anything in the representation of the relation of this being to the world as far as either theoretical or practical consequences are concerned. To wish to investigate what it is in itself is a pretension that is as purposeless as it is futile. (CJ, 329n)

Knowing nothing of God as such, then, is of no importance and even of no interest. The analogical approach, which cuts us from God just where it seemed to promise to reveal Him, is by that very fact perfectly adequate, leaves no regrets, because “God” here is precisely the name only for the pure relation to the world of an agency the logic of which is dictated by a teleology constitutively cut from its end. The Endzweck, recall, is an Endzweck only if it depends not at all on that whose end it is, only if it is cut from it. Analogy answers perfectly to this structure of the cut, espouses the cut by inscribing it right on the point of heterogeneity. Nothing is lacking here, for the cut has nothing negative about it but is the positive condition of Kant’s thinking of God as pure relation. By precisely not allowing us to conclude the objective existence of God, the analogy turns out to be perfectly adequate to the mode of being of God, which cannot by definition be that of a knowable object.

So, among the four possibilities of the Fürwahrhalten discussed in §90 (logical proof, analogy, opinion, and hypothesis), none of which allows one to conclude that God exists, analogy is set apart from the rest, because where the other attempts are simply inappropriate (no logical proof in the absence of any intuition of God; no plausible opinion because there is no possible continuity between this-worldly reasons and other-worldly aims; no hypothesis, because it would presuppose that what it puts forward is at least possible, whereas we have no reason at all to admit the possibility of God as object), analogy gives the exact measure of the thinking of God of which we are capable: God is perfectly analogical.

Teleology, then, which cannot not end up in a theology, does so according to its own proper structure (i.e., being cut from its end), which alone makes it a teleology. Cut from its end, teleology does not lose its guiding thread (which, as we have seen as we venture on this interminable reading, has no end, precisely, and just that is its finality) but prolongs it indefinitely in analogy, which describes exactly this relation to an end from which one is radically cut. As always in Kant, we must not see in analogy an ersatz theoretical knowledge—this is what §90 has just shown: Nothing is lacking in our representation of the relation of God to the world so long as this representation remains merely analogical, and if nothing is lacking, that is because this analogical representation is, precisely, not a substitute or surrogate for some missing theoretical knowledge. Analogy is not a substitute for theoretical knowledge but always has a practical import. Which is why, in the following paragraph, Kant comes to the problem of the relation between this mode of Fürwahrhalten and practical belief.

If we take up again the question of knowable objects, no longer on the side of the object (“What can we know of God?”) but on our side (“What can we know of God?”), we discover three sorts of knowable things: things of opinion (Sachen der Meinung [opinabile]), facts (Tatsachen [scibile]), and things of belief (Glaubensachen [mere credibile]). I can have an opinion only with respect to an empirical possibility that I might one day verify (so I can be of the opinion that there is intelligent life on other planets, because it is in itself possible that one day I could go and check this out) but never with respect to an Idea of Reason (because that would make it into mere fiction). Facts are where I can prove an objective reality (where there is a corresponding intuition). But there is one exception among facts and among the Ideas of Reason, and this is the Idea of freedom, which makes sense only to the extent that it can be made manifest by real actions in the world of experience. The factuality of freedom (none other than the to-be-realized of the moral law) is part of its Idea, which, as we saw for the law, must comprise its own realizability as a determining feature. Freedom is a fact and can therefore be known, even if it remains an Idea (because, qua concept, it is überschwenglich for theoretical reason, for the understanding). But the idea of freedom, when it is thought according to the moral law, entails other Ideas that go beyond the domain of theoretical concepts, and these Ideas will be the object of belief but a belief that must be practical. This means that freedom can be bought under the moral law only if we admit the Idea of the supreme Good to be realized in the world:

Of this sort is the highest good to be achieved in the world through freedom, the objective reality of the concept of which cannot be proved adequately in any experience possible for us, and hence for the theoretical use of reason, but the use of which is nevertheless commanded by practical pure reason for the best possible realization of that end, and which must thus be assumed to be possible. (CJ, 333)

And, says Kant, if we admit the supreme Good as an object of belief, we must also admit, for the same reasons, the conditions of its realizability (namely, the immortality of the soul and the existence of God). And these are really the only true objects of belief (because everything we believe on the basis of another’s testimony, for example, really implies a probative experience somewhere and can thus in principle become a known object),19 and belief becomes essentially a matter for practical reason, which, in order to be practical, has to give itself the structure of its own to-be-realized, according to the interrupted teleology we are following here.

It remains, however, to clarify the exact relation between the moral law and the final end it commands us to bring about, for which we have to admit God as merely an object of practical belief. For the cut that alone gives meaning to the Idea of the final end (which, recall again, is final only to the extent that it is cut from everything of which it is the end) intervenes again between the moral law strictly conceived and what that same law commands; or, better, the cut intervenes within the moral law itself as its very structure, that in a moment Kant is going to describe as the structure of a promise. But—and this is Kant’s problem—as such, the moral law is precisely not a matter of ends, on pain of falling into what the Groundwork and the second Critique call the hypothetical or technical imperative. The moral law is expressed as categorical imperative and is thus unconditioned and unconditional. So how are we to articulate it with what Kant has just demonstrated? We must try to articulate it with God and the final end, because otherwise the whole teleology would again fall back into mechanism and theoreticism: The teleology depends entirely on the positing of the final end, and the coherence of this final end depends on its relation with freedom thought under the aegis of the moral law. If the moral law did not command us to realize the supreme Good as final end, there would be no freedom other than contingency (i.e., we would be subject to blind chance or blind necessity). Kant attempts this articulation first of all by situating the cut between the necessity of the law as such, on the one hand, and the partial contingency of the final end, on the other:

Consequently, the concept of God acquires the distinction of counting as a matter of faith in our holding-for-true only through its relation to the object of our duty, as the condition of the possibility of attaining the final end of that duty; by contrast, the very same concept can still not make its object valid as a fact, because although the necessity of duty for the practical reason is quite clear, still the attainment of its final end, insofar as it is not entirely in our own power, is assumed only for the sake of the practical use of reason, and is thus not practically necessary like duty itself. (CJ, 334–35)

Duty is necessary (i.e., known a priori by reason) and is thus entirely within our power. The final end, however, escapes us at least in part and cannot therefore attain necessity. However, there must be this final end in its very nonnecessity, and, following a structure that is becoming familiar to us, this is necessary for necessity. Just as natural necessity needed teleology to safeguard its necessity (the necessity of necessity), so practical necessity, to the extent that it entails a necessity of realization (which is why the Idea of freedom can lay claim to its status as a fact and save the teleology) needs this nonnecessary final end if its necessity is to be properly practical. As is often the case, it is in a note that Kant tries to untangle these very complex relations:

The final end, the promotion of which is imposed upon us by the moral law, is not the ground of duty; for this lies in the moral law, which, as a formal practical principle, guides us categorically, regardless of the object of the faculty of desire (the matter of the will), hence regardless of any end. This formal property of my actions, in which alone their inner moral worth consists, is entirely in our power; and I can perfectly well abstract from the possibility or unrealizability of the ends that I am obliged to promote in accordance with that law (because only the external value of my actions consists in them) as something that is never completely in my power, in order to see only that which is my own doing. Yet the aim [die Absicht] of promoting the final end of all rational beings (happiness, insofar as it is consistent with duty) is still imposed precisely by the law of duty. (CJ, 335n)

So here is the crux of the problem: The “formal” character of the moral law, which alone guarantees its necessity, allows for the possibility of abstracting from ends and their realization or realizability in general. But I ought not follow this possibility opened by the law, because that same law imposes on me the duty not to ignore ends. Which means that the morality of the moral law (what is too rapidly called its “formalism”) is immediately counteracted by a movement toward the end, and this movement also forms part of the morality of the moral law. The categorical imperative orders me to act according to its purely formal law, but it also orders me to act (i.e., to pursue an end and make something happen). The maxim of my action must survive the formal test of the law, but my action is moral only if I do it, if, then, I realize some end; and this end, whatever it be, must also contribute to the pursuit of the final end. As to this incursion of moral action into the world of nature, we have known since the Antinomies of Pure Reason that theoretical reason does not understand it at all:

But speculative reason cannot understand the realizability of this final end at all (either on the part of our own physical capacity or on the part of the cooperation of nature); rather, on the basis of such causes, as far as we can rationally judge, it must hold the assumption of such a success of our good conduct in mere nature (inside and outside of us), without the assumption of God and immortality, to be unfounded and empty even if well-intended, and, if it could be completely certain of this judgment, it would regard the moral law itself as a mere deception of our reason in a practical respect. But since speculative reason is fully convinced that the latter can never happen, but by contrast those ideas the object of which lies beyond nature can be conceived without contradiction, it must acknowledge those ideas as real for its own practical law and the task that is thereby imposed, thus as real in a moral respect, in order not to contradict itself. (CJ, 335n)

While not understanding this at all, theoretical reason is nonetheless necessitated to believe it, on pain of contradiction. Practical reason ought not be, must not be, has not to be a mystification of theoretical reason, and theoretical reason has to believe it, had better believe it, to the extent that the whole teleological structure of reason—the rationality of reason itself—depends on it. And this is indeed the structure of the Endzweck, itself a rigorous consequence of teleology: All teleology must finalize its ends, however complex and numerous they be, by appeal to a final end, which must logically be cut from that of which it is the end. It is obvious that this end, which alone gives meaning to that of which it is the end, structurally speaking, escapes the comprehension of everything it finalizes. The “practical belief” that results as the only rational possibility of still orienting oneself in this structure takes the form of the kind of belief one can have in a promise. Better believe that the promise will be kept and fulfilled, but the promise as such excludes all certainty as to its realization (which is why it is hard to make a clear distinction between promise and threat)20 and thus represents the constitutive frustration, the productive inhibition of theoretical reason, which thus finds itself grounded in a slightly vacant confidence. Kant again resorts to a note to clarify his position:

Faith (as habitus, not as actus) is reason’s moral way of thinking in the affirmation of that which is inaccessible for theoretical cognition. It is thus the constant fundamental principle of the mind to assume as true that which it is necessary to presuppose as a condition for the possibility of the highest moral final end, on account of the obligation to that, although we can have no insight into its possibility or into its impossibility.*

* [Kant’s note:] It is a matter of trusting the promise of the moral law: not a promise that is contained in the moral law, but one that I put into it, and indeed on a morally adequate basis. For a final end cannot be commanded by any law of reason without reason simultaneously promising its attainability, even if uncertainly, and hereby also justifying the affirmation of the only conditions under which our reason can conceive this. (CJ, 335–36 and n)

Teleology comes to an end here. Which means that it interrupts itself rather than fulfills itself—or perhaps that it fulfills its interruption. Just as it was necessary for peace that peace not be realized, it is necessary for practical reason that the moral law—the telos of the whole Kantian system, which gives meaning to all rational activity as such (even in the theoretical domain)—be merely promised. Or rather, for the moral law itself makes no promises, does not contain this promise, as Kant says, it is necessary that I promise myself the law, that I promise myself the realization of what the law commands—namely, to make the law. I promise myself that the law, which is inaccessible as such, forever, forever unrealizable, remains nonetheless practicable as something always to come; or rather, for the law cannot really be some thing, that what I call the law is only the practice of the path that I am tracing with the end of the thread in my hand, a Seefahrer in spite of everything, and that is thus not even a path but just the minimal movement that carries me forward just enough for me to believe in this forward movement, not only in uncertainty as to the end to be reached but in the certainty, not only that I will never reach the end, but that I will never know what the end is, because there is no end beyond what I am doing here and now in promising myself to go on, or at least (because in the belief the promise gives me I do not know which way is onward, merely that I must go on) to take a step in nature, that uncrossable frontier, toward the singular, violent encounter on the occasion of which, the chance or opportunity, the Zweckmässigkeit of which I will again promise (as I am already promising without promising), promise myself the law that I will no more know then than now, as I take another step, holding in my hand the end of the thread I have always been dragging behind me, this point of free heterogeneity that therefore I am.

Brought this far by teleology, encouraged by those natural contingencies that call for teleological judgment, that encourage me to judge nature as a whole to be the work of an intelligent and moral God, I find myself now abandoned by teleology, abandoned to the pure teleology without telos of the promise I make myself as to the moral law, while the world around me can break up into scattered raw matter:

The moral proof (which of course proves the existence of God only in a practical respect although one that is also indispensable for reason) would thus always remain in force even if we found in the world no material for physical teleology at all or only ambiguous material for it. We can conceive of rational beings who see themselves surrounded by a nature that gives no clear trace of organization but reveals only effects of a mere mechanism of raw matter, and who on that account, and given the alterability of some merely contingently purposive forms and relations, seem to have no ground to infer an intelligent author, in which case there would also be no suggestion of a physical teleology; nevertheless, reason, which in this case gets no guidance from concepts of nature, would still find in the concept of freedom and the moral ideas that are grounded upon that a practically sufficient ground for postulating the concept of an original being in accordance with these, i.e., as the concept of a divinity, and for postulating nature (even our own existence) as a final end in accordance with that concept and its laws, and of course with respect to the indispensable command of practical reason. (CJ, 341–42)

If, then, we have no need of natural teleology in order to think the being-teleological of teleology as such, if freedom alone would suffice, one might imagine that one would think that teleology better, more purely, without the encouragement or “confirmation” given by natural teleology, which in the end has nothing to do with the final end, cut from the world of which it is the end. Reason as such has no use for the physico-teleological proof of the existence of God. It only serves to make a finite and failing reason more receptive to the real proof, the moral proof, which as such can very well do without it.21

Seen from the vantage of reason itself, the natural contingencies that demand to be judged teleologically really are contingent. Which is as much as to say that Epicurus could have been right about natural mechanism without that changing anything for reason. But which is also as much as to say that one cannot see from the point of view of reason, that we are not there yet, that teleology in general makes no sense unless one has not reached its telos, a telos that one never could reach. The whole teleological structure, including the projected telos that determines it while remaining cut from it, depends on our being always somewhere in the middle, depends on the telos remaining for us a promise, thus uncertain, and that therefore we have no certainty at all as to the teleological nature of teleology. Which is also what is called freedom. Here, now, point de téléologie, point de liberté.

Can we identify this point with Kant’s thinking of the transcendental imagination and ultimately with time itself? No more than with finitude, to the extent that such a reading, despite its obvious power, remains within a reading of Kant that misrecognizes the radicality of judgment, even within the first Critique, by having undue faith in what phenomenology can put up against a “scholastic” logic.

Rather, we must try to think this “point” (which is clearly not a point) according to the fractal logic of the frontier and the radical exteriority it allows us to think. As spectral residue of a nature that never was before this residue, the frontier, unsublatable trace of exteriority and therefore of any identity at all (even that of time, of which Heidegger’s conception would be the most powerful attempt to hold together identity and difference—but the frontier is also the trace of a spacing unsublatable by time itself), can only relaunch the scattering play of ana-logy or ananalogy without ever promising to gather it into a logos speakable as such. Ana-logos precedes logos. Whence the violence inscribed in the frontier as thought by Kant in the failing perspective of perpetual peace, and whence too a certain stupid literality of everything we have suggested here. The stupid resistance of the frontier to thinking is clearly enough the chance of thinking, which survives and even thrives only by not quite being able to think what it thinks. In this way the essence of peace, be it in the literal sense or, analogically, peace in philosophy, excludes its perpetuity. Nothing is thought in the peace of the graveyard; true peace is agitated and threatened, perpetually, here too, on this final frontier.

---

1. As Derrida points out, the bridge is a figure of analogy, an analogy for analogy itself, just as analogy is a bridge that, being thrown over the abyss, also throws us into a kind of mise en abyme: “The abyss calls for analogy—the active recourse of the whole Critique—but analogy plunges endlessly into the abyss as soon as a certain art is needed to describe analogically the play of analogy” (The Truth in Painting, 36).

2. This is why much later, and equally famously, Kant will say that philosophy cannot be learned (except historically) and that at best we can learn to philosophize (CPR, A837/B865). One can become erudite without making up the failure of judgment that Kant calls simply Dummheit, stupidity, and this would be the basis for his objections against the historians of philosophy we read earlier.

3. Jean-François Lyotard provides perhaps the most sustained reading of this position of judgment in Kant (see especially Leçons sur l’analytique du sublime [Paris: Galilée, 1991]). However, I am wary of the temptation to seek in the Analytic of the Sublime the place of the political in Kant. I shall rather be attempting what might appear to be a rehabilitation of the teleological judgment, but as radically cut from its telos. This does not simply return us to the famous “purposiveness without purpose” or “finality without end” of the aesthetic judgment (as brilliantly expounded by Derrida in The Truth in Painting [83–118]) but rather to something of the order of the end of the end, a purposiveness without purposiveness or a finality without finality.

4. The term Erkenntnis does not necessarily imply knowledge of a theoretical kind but refers to a relation to the object in general. But there is some hesitation in Kant on this matter, marked here by the fact that philosophy in general, according to the title of this section of the introduction, has a “domain,” whereas what is more properly called a domain (that of theoretical knowledge) will later be circumscribed within that general domain.

5. So the field is the zone within the general domain where there is an as yet indeterminate relation between concept and object; that there be this relation constitutes a kind of cognition prior to knowledge. Kant does not say whether the general domain of philosophy extends still further out, but one might imagine that general logic would occupy that part of the general domain that falls outside the field. We would need to read this passage against the preface to the Groundwork, where the first great division of philosophy separates formal philosophy (logic), which does not deal with objects, from material philosophy, which then divides again into natural philosophy and moral philosophy, each of which contains a pure part and an empirical part. Note that the version given in the third Critique abandons an arborescent model (a common trunk divides into branches that divide again) in favor of a territorial model (frontiers have to be drawn in space). This territorial language is of course widespread in Kant, at least from the first Critique onward.

6. At Dartford, near London, traffic crosses the Thames in one direction through a tunnel and in the other over a bridge. This might give a working representation of the relations in Kant between theory and practice.

7. See too the opening of §64: “In order to see that a thing is possible only as an end, i.e., that the causality of its origin must be sought not in the mechanism of nature, but in a cause whose productive capacity is determined by concepts, it is necessary that its form not be possible in accordance with mere natural laws, i.e., ones that can be cognized by us through the understanding, applied to objects of the senses, alone; rather even empirical cognition of their cause and effect presupposes concepts of reason. Since reason must be able to cognize the necessity in every form of natural product if it would understand the conditions connected with its generation, the contingency of their form with respect to all empirical laws of nature in relation to reason is itself a ground for regarding their causality as if it were possible only through reason; but this is then the capacity for acting in accordance with ends (a will); and the object which is represented as possible only on this basis is represented as possible only as an end.” §67 will extend this logic to nature as a whole qua system, as we shall see in detail later.

8. Recall that in the initial description of the general domain of philosophy, there was some terrain with no domain, what Kant called the “residence,” the Aufenthalt, the place for concepts that were lawful without being legislative, giving rise to rules that were empirical and thus contingent. This is the part of the general domain that judgment will occupy as its proper (non)domain.

9. This is no doubt why, in §74, Kant identifies determinative judgment with dogmatic procedure, reflective judgment with critical procedure: “We deal with a concept dogmatically (even if it is supposed to be empirically conditioned) if we consider it as contained under another concept of the object, which constitutes a principle of reason, and determine it in accordance with the latter. But we deal with it merely critically if we consider it only in relation to our cognitive faculties, hence in relation to the subjective conditions for thinking it, without undertaking to decide anything about its object. The dogmatic treatment of a concept is thus that which is lawful for the determining, the critical that which is lawful merely for the reflecting power of judgment” (CJ, 266). As Lyotard has most clearly brought out, critique is only really critical, only really itself, only comes to itself in reflective judgment. We might add that given the becoming-teleological even of the determinative judgment, critique entails a reflective judgment on the issue of the determinative judgment itself.

10. We must beware of this formulation, which is unduly ontological: law is not being and thus is not to be taken in terms of a “law-being,” even if it can only be understood in relation to being. The reading of teleology we are attempting here tries to situate it between a being and an “oughting” of the law. The law of the law is not ought, and ought not to be.

11. Perhaps we could exploit other semantic valences of the word Zweckmässigkeit, which also implies suitability, fitness, a being-opportune that conjures up a thought of chance and the event, the kairos, the right moment.

12. See Opus Postumum, 57, on the “transition” to Physics: “Regulative principles that are at the same time constitutive.”

13. According to a procedure brilliantly exploited by Paul de Man in Allegories of Reading (New Haven, CT: Yale University Press, 1979), 9–12.

14. These examples will, then, function like regulative principles for a reflective judgment in order to demonstrate the necessity of regulative principles and reflective judgments. Qua examples, these examples are contingent, and they are examples of the contingent—which is nonetheless necessary. Clearly, this use of “examples” goes beyond the description of the exemplarity of the example given in the first Critique, where examples are called the Gängelwagen, what Kemp Smith memorably translated as the “go-cart” (Guyer and Wood have “leading-strings,” Pluhar “walker”) of judgment (A134/B173), a kind of provisional prosthetic to help with a lack of natural aptitude. Here, however, these “examples” could never become “proven propositions,” and so the prosthesis is originary and irreducible.

15. Again, this structure seems to be misrecognized by Hegel in his critique of Kant. In putting the maxim of the action of giving money to the poor through the test of the categorical imperative, the point is not at all, as Hegel pretends to believe in the Natural Law essay, to project the end of poverty, thus eliminating the very thing that gave rise to the possibility of moral action, but to inscribe this singular act in the perspective of a to-be-realized that must precisely never arrive at its realization (its end).

16. Guyer translates “Art des Führwahrhaltens” as “kind of affirmation” and relates it to the analytic philosophy notion of “propositional attitude.”

17. Translation modified. Given the rather dramatic nature of Kant’s claims, here and in the following quotations I prefer “heterogeneous” to “dissimilar” for ungleichartig and “conclude” to “infer” or “make an inference” for Schliessen and make cognate changes as appropriate.

18. In the various versions of his lectures on Logic, Kant gives much less rich explanations of analogy, always in the context of induction and always saying that induction and analogy are merely “crutches” for the understanding, but crutches we cannot do without. In the so-called Dohna-Wundlacken version of the Logic, however (a transcript of the 1792 course, here a session from August 20), we find the following (the text includes several conjectural readings by its editor, Kowalewski): “According to the inference by analogy, if 2 things agree under as many determinations as I have become acquainted with, then I infer that they agree also in the other determinations. I infer, then, from some determinations, which I cognize, that the others belong to the thing too. This is an inference of a provisional judgment. One reserves the right to change it. No logician has yet developed analogy and induction properly. This field still lies open.” In the margin, Kant writes, “An inference according to analogy, that animals have souls—But . . . why? If wasps have drunk from the forbidden honey and now wish to sting, then they fall, but they have [. . .] Faculty of desire. This holds for plants, too, the fly-trap. They seek air and sun and water” (in Lectures on Logic, tr. J. Michael Young [Cambridge: Cambridge University Press, 2004], 503–4).

19. This distinction is placed into question in recent work by Derrida. See, for example, Demeure: Maurice Blanchot (Paris: Galilée, 1998).

20. As Derrida has conclusively shown, any promise, in order to be a promise, must bring with it the necessary possibility of its nonrealization and thus the constitutive threat that it not be kept.

21. Kant’s argument is supposed to provide a proof of the existence of God. This proof draws its strength and its rationality from its very theoretical insufficiency. This is a practical proof in the sense that in the end it proves nothing and is entirely held in the open structure of the promise. In this sense, God depends on the absence of any proof of his existence, which comes down to the thought that believing in God (in this Kantian way) and not believing in God come to the same thing. Given Kant’s proof, we cannot (not) believe in God. The frontier, on which we are still putting in time, persists as this very (im)possibility. In this sense, God is just a name for the quasi-transcendental as general condition of everything, pure Deus terminus.