which bas its origin solely in the reflective judgement. For we cannot
ascribe to the products of nature anything like a reference of
nature in them to ends, but we can only make use of this concept to
reflect upon them in respect of the nexus of phenomena in nature-a
nexus given according to empirical laws. Furthermore, this concept
is entirely different from practical finality (in human art or even
morals), though it is doubtless thought after this analogy.

       V. The Principle of the formal finality of Nature is a

              transcendental Principle of Judgement.

  A transcendental principle is one through which we represent a
priori the universal condition under which alone things can become
objects of our cognition generally. A principle, on the other band, is
called metaphysical where it represents a priori the condition under
which alone objects whose concept has to be given empirically may
become further determined a priori. Thus the principle of the
cognition of bodies as substances, and as changeable substances, is
transcendental where the statement is that their change must have a
cause: but it is metaphysical where it asserts that their change
must have an external cause. For, in the first case, bodies need
only be thought through ontological predicates (pure concepts of
understanding) e.g., as substance, to enable the proposition to be
cognized a priori; whereas, in the second case, the empirical
concept of a body (as a movable thing in space) must be introduced
to support the proposition, although, once this is done, it may be
seen quite a priori that the latter predicate (movement only by
means of an external cause) applies to body. In this way, as I shall
show presently, the principle of the finality of nature (in the
multiplicity of its empirical laws) is a transcendental principle. For
the concept of objects, regarded as standing under this principle,
is only the pure concept of objects of possible empirical cognition
generally, and involves nothing empirical. On the other band, the
principle of practical finality, implied in the idea of the
determination of a free will, would be a metaphysical principle,
because the concept of a faculty of desire, as will, has to be given
empirically, i.e., is not included among transcendental predicates.
But both these principles are, none the less, not empirical, but a
priori principles; because no further experience is required for the
synthesis of the predicate with the empirical concept of the subject
of their judgements, but it may be apprehended quite a priori.

  That the concept of a finality of nature belongs to transcendental
principles is abundantly evident from the maxims of judgement upon
which we rely a priori in the investigation of nature, and which yet
have to do with no more than the possibility of experience, and
consequently of the knowledge of nature-but of nature not merely in
a general way, but as determined by a manifold of particular laws.
These maxims crop up frequently enough in the course of this
science, though only in a scattered way. They are aphorisms of
metaphysical wisdom, making their appearance in a number of rules
the necessity of which cannot be demonstrated from concepts. "Nature
takes the shortest way (lex parsimoniae); yet it makes no leap, either
in the sequence of its changes, or in the juxtaposition of
specifically different forms (lex continui in natura); its vast
variety in empirical laws is for all that, unity under a few
principles (principia praeter necessitatem non sunt multiplicanda)";
and so forth.

  If we propose to assign the origin of these elementary rules, and
attempt to do so on psychological lines, we go straight in the teeth
of their sense. For they tell us, not what happens, i.e., according to
what rule our powers of judgement actually discharge their
functions, and how we judge, but how we ought to judge; and we
cannot get this logical objective necessity where the principles are
merely empirical. Hence the finality of nature for our cognitive
faculties and their employment, which manifestly radiates from them,
is a transcendental principle of judgements, and so needs also a
transcendental deduction, by means of which the ground for this mode
of judging must be traced to the a priori sources of knowledge.

  Now, looking at the grounds of the possibility of an experience, the
first thing, of course, that meets us is something necessary-namely,
the universal laws apart from which nature in general (as an object of
sense) cannot be thought. These rest on the categories, applied to the
formal conditions of all intuition possible for us, so far as it is
also given a priori. Under these laws, judgement is determinant; for
it bas nothing else to do than to subsume under given laws. For
instance, understanding says: all change has its cause (universal
law of nature); transcendental judgement has nothing further to do
than to furnish a priori the condition of subsumption under the
concept of understanding placed before it: this we get in the
succession of the determinations of one and the same thing. Now for
nature in general, as an object of possible experience, that law is
cognized as absolutely necessary. But besides this formal
time-condition, the objects of empirical cognition are determined, or,
so far as we can judge a priori, are determinable, in divers ways,
so that specifically differentiated natures, over and above what
they have in common as things of nature in general, are further
capable of being causes in an infinite variety of ways; and each of
these modes must, on the concept of a cause in general, have its rule,
which is a law, and, consequently, imports necessity: although owing
to the constitution and limitations of our faculties of cognition we
may entirely fail to see this necessity. Accordingly, in respect of
nature's merely empirical laws, we must think in nature a
possibility of an endless multiplicity of empirical laws, which yet
are contingent so far as our insight goes, i.e., cannot be cognized
a priori. In respect of these we estimate the unity of nature
according to empirical laws, and the possibility of the unity of