Immanuel Kant's Critique of Pure Reason — trans. F. Max Müller

by which I know a priori (as necessary) that something is, while the practical use of reason is that by which I know a priori what ought to be. If then it is certain, beyond the possibility of doubt, that something is, or that something ought to be, though both are conditioned, then a certain definite condition of it may be either absolutely necessary or presupposed only as possible and contingent. In the former case, the condition is postulated i^per thesifi^)^ in the latter supposed {per Jiypothesin). As there are practical laws, which are absolutely necessary (the moral laws), it follows, if they necessarily presuppose [p. 634] any existence as the condition of the possibility of their obligatory power, that the existence of that condition must h^ postulated, because the conditioned, from which we infer that condition, has been recognised a priori as absolutely necessary. On a future occasion we shall show that the moral laws not only presuppose the existence of a Supreme Being, but that, as they are in other respects absolutely necessary, they postulate it by right, though of course practically only. For the present we leave this mode of argument untouched.

If we only speak of that which is, not of that which ought to be, the conditioned given to us in experience is always conceived as contingent, and the condition belonging to it can therefore not be known as absolutely necessary, but serves only as a relatively necessary, or rather needful, though in itself an a priori arbitrary supposition for a rational understanding of the conditioned. If, therefore, we wish to know in our theoretical knowledge the absolute necessity of a thing, this could only be done from concepts a priori, and never as of a cause in reference to an existence which is given in experience.

Transcejidental Dialectic 511

I call a theoretical knowledge speculative^ if it relates to an object, or such concepts of an object, which we can never reach in any experience. It is opposed to our knowledge of natter e, which relates to no other objects [p. 635] or predicates of them except those that can be given in a possible experience.

From something that happens (the empirically contingent) as an effect, to infer a cause, is a principle of natural, though not of speculative knowledge. For if we, no longer use it as a principle involving the condition of possible experience, and, leaving out everything that is empirical, try to apply it to the contingent in general, there does not remain the smallest justification of such a synthetical proposition, showing how from something which is, there can be a transition to something totally different, which we call cause ; nay, in such purely speculative application, the concepts both of cause and of the contingent lose all meaning, the objective reality of which would be made intelligible in the concrete.

If from the existence of things in the world we infer their cause, we are using reason not naturally, but speculatively. Naturally, reason refers not the things themselves (substances), but only that which happens, their states, as empirically contingent, to some cause ; but it could know speculatively only that a substance itself (matter) is contingent in its existence. And even if we were thinking only of the form of the world, the [p. 636] manner of its composition and the change of this composition, and tried to infer from this a cause totally different from the world, this would be again a judgment of speculative reason only; because the object here is not an object of any possible experience. In this case the principle of

512 Transcendental Dialectic