Reason with all its concepts and laws of the understanding, which suffice for empirical use, i.e., within the sensible world, finds in itself no satisfaction because ever-recurring questions deprive us of all hope of their complete solution. The transcendental ideas, which have that completion in view, are such problems of reason. But it sees clearly, that the sensuous world cannot contain this completion, neither consequently can all the concepts, which serve merely for understanding the world of sense, such as space and time, and whatever we have adduced under the name of pure concepts of the understanding. The sensuous world is nothing but a chain of appearances connected according to universal laws; it has therefore no subsistence by itself; it is not the thing in itself, and consequently must point to that which contains the basis of this experience, to beings which cannot be cognised merely as phenomena, but as things in themselves. In the cognition of them alone reason can hope to satisfy its desire of completeness in proceeding from the conditioned to its conditions.

We have above (§§ 33, 34) indicated the limits of reason with regard to all cognition of mere creations of thought. Now, since the transcendental ideas have urged us to approach them, and thus have led us, as it were, to the spot where the occupied space (viz., experience) touches the void (that of which we can know nothing, viz., noumena), we can determine the bounds of pure reason. For in all bounds there is something positive (e.g., a surface is the boundary of corporeal space, and is therefore itself a space, a line is a space, which is the boundary of the surface, a point the boundary of the line, but yet always a place in space), whereas limits contain mere negations. The limits pointed out in those paragraphs are not enough after we have discovered that beyond them there still lies something (though we can never cognise what it is in itself). For the question now is, What is the attitude of our reason in this connexion of what we know with what we do not, and never shall, know? This is an actual connexion of a known thing with one quite unknown (and which will always remain so), and though what is unknown should not become the least more known—which we cannot even hope—yet the notion of this connexion must be definite, and capable of being rendered distinct.

Kant's Prolegomena — Part 80