That the concept of a purposiveness of nature belongs to transcendental principles can be sufficiently seen from the maxims of the Judgement, which lie at the basis of the investigation of nature a priori, and yet do not go further than the possibility of experience, and consequently of the cognition of nature—not indeed nature in general, but nature as determined through a variety of particular laws. These maxims present themselves in the course of this science often enough, though in a scattered way, as sentences of metaphysical wisdom, whose necessity we cannot demonstrate from concepts. “Nature takes the shortest way (lex parsimoniae); at the same time it makes no leaps, either in the course of its changes or in the juxtaposition of specifically different forms (lex continui in natura); its great variety in empirical laws is yet unity under a few principles (principia praeter necessitatem non sunt multiplicanda),” etc.
If we propose to set forth the origin of these fundamental propositions and try to do so by the psychological method, we violate their sense. For they do not tell us what happens, i.e. by what rule our cognitive powers actually operate, and how we judge, but how we ought to judge; and this logical objective necessity does not emerge if the principles are merely empirical. Hence that purposiveness of nature for our cognitive faculties and their use, which is plainly apparent from them, is a transcendental principle of judgements, and needs therefore also a Transcendental Deduction, by means of which the ground for so judging must be sought in the sources of cognition a priori.
We find in the grounds of the possibility of an experience in the very first place something necessary, viz. the universal laws without which nature in general (as an object of sense) cannot be thought; and these rest upon the Categories, applied to the formal conditions of all intuition possible for us, so far as it is also given a priori. Now under these laws the Judgement is determinant, for it has nothing to do but to subsume under given laws. For example, the Understanding says that every change has its cause (universal law of nature); the transcendental Judgement has nothing further to do than to supply a priori the condition of subsumption under the concept of the Understanding placed before it, i.e. the succession [in time] of the determinations of one and the same thing. For nature in general (as an object of possible experience) that law is cognised as absolutely necessary.—But now the objects of empirical cognition are determined in many other ways than by that formal time-condition, or, at least as far as we can judge a priori, are determinable. Hence specifically different natures can be causes in an infinite variety of ways, as well as in virtue of what they have in common as belonging to nature in general; and each of these modes must (in accordance with the concept of a cause in general) have its rule, which is a law and therefore brings necessity with it, although we do not at all comprehend this necessity, in virtue of the constitution and the limitations of our cognitive faculties. We must therefore think in nature, in respect of its merely empirical laws, a possibility of infinitely various empirical laws, which are, as far as our insight goes, contingent (cannot be cognised a priori), and in respect of which we judge nature, according to empirical laws and the possibility of the unity of experience (as a system according to empirical laws), to be contingent. But such a unity must be necessarily presupposed and assumed, for otherwise there would be no thoroughgoing connexion of empirical cognitions in a whole of experience. The universal laws of nature no doubt furnish such a connexion of things according to their kind as things of nature in general, but not specifically, as such particular beings of nature. Hence the Judgement must assume for its special use this principle a priori, that what in the particular (empirical) laws of nature is from the human point of view contingent, yet contains a unity of law in the combination of its manifold into an experience possible in itself—a unity not indeed to be fathomed by us, but yet thinkable. Consequently as the unity of law in a combination, which we cognise as contingent in itself, although in conformity with a necessary design (a need) of Understanding, is represented as the purposiveness of Objects (here of nature); so must the Judgement, which in respect of things under possible (not yet discovered) empirical laws is merely reflection, think of nature in respect of the latter according to a principle of purposiveness for our cognitive faculty, which then is expressed in the above maxims of the Judgement. This transcendental concept of a purposiveness of nature is neither a natural concept nor a concept of freedom, because it ascribes nothing to the Object (of nature), but only represents the peculiar way in which we must proceed in reflection upon the objects of nature in reference to a thoroughly connected experience, and is consequently a subjective principle (maxim) of the Judgement. Hence, as if it were a lucky chance favouring our design, we are rejoiced (properly speaking, relieved of a want), if we meet with such systematic unity under merely empirical laws; although we must necessarily assume that there is such a unity without our comprehending it or being able to prove it.
In order to convince ourselves of the correctness of this Deduction of the concept before us, and the necessity of assuming it as a transcendental principle of cognition, just consider the magnitude of the problem. The problem, which lies a priori in our Understanding, is to make a connected experience out of given perceptions of a nature containing at all events an infinite variety of empirical laws. The Understanding is, no doubt, in possession a priori of universal laws of nature, without which nature could not be an object of experience; but it needs in addition a certain order of nature in its particular rules, which can only be empirically known and which are, as regards the Understanding, contingent. These rules, without which we could not proceed from the universal analogy of a possible experience in general to the particular, must be thought by it as laws (i.e. as necessary), for otherwise they would not constitute an order of nature; although their necessity can never be cognised or comprehended by it. Although, therefore, the Understanding can determine nothing a priori in respect of Objects, it must, in order to trace out these empirical so-called laws, place at the basis of all reflection upon Objects an a priori principle, viz. that a cognisable order of nature is possible in accordance with these laws. The following propositions express some such principle. There is in nature a subordination of genera and species comprehensible by us. Each one approximates to some other according to a common principle, so that a transition from one to another and so on to a higher genus may be possible. Though it seems at the outset unavoidable for our Understanding to assume different kinds of causality for the specific differences of natural operations, yet these different kinds may stand under a small number of principles, with the investigation of which we have to busy ourselves. This harmony of nature with our cognitive faculty is presupposed a priori by the Judgement, on behalf of its reflection upon nature in accordance with its empirical laws; whilst the Understanding at the same time cognises it objectively as contingent, and it is only the Judgement that ascribes it to nature as a trancendental purposiveness (in relation to the cognitive faculty of the subject). For without this presupposition we should have no order of nature in accordance with empirical laws, and consequently no guiding thread for an experience ordered by these in all their variety, or for an investigation of them.
For it might easily be thought that, in spite of all the uniformity of natural things according to the universal laws, without which we should not have the form of an empirical cognition in general, the specific variety of the empirical laws of nature including their effects might yet be so great, that it would be impossible for our Understanding, to detect in nature a comprehensible order; to divide its products into genera and species, so as to use the principles which explain and make intelligible one for the explanation and comprehension of another; or out of such confused material (strictly we should say, so infinitely various and not to be measured by our faculty of comprehension) to make a connected experience.
The Judgement has therefore also in itself a principle a priori of the possibility of nature, but only in a subjective aspect; by which it prescribes, not to nature (autonomy), but to itself (heautonomy) a law for its reflection upon nature. This we might call the law of the specification of nature in respect of its empirical laws. The Judgement does not cognise this a priori