Pure mathematics and pure natural science would not have needed, for the purpose of their own security and certainty, a deduction of the sort that we have hitherto accomplished for them both; for the first is supported by its own evidence, and the second, though arising from pure sources of the understanding, is nonetheless supported from experience and thoroughgoing confirmation by it – experience being a witness that natural science cannot fully renounce and dispense with, because, as philosophy, despite all its certainty it can never rival mathematics. Neither science had need of the aforementioned investigation for itself, but for another science, namely metaphysics.

Apart from concepts of nature, which always find application in experience, metaphysics is further concerned with pure concepts of reason that are never given in any possible experience whatsoever, hence with concepts whose objective reality (that they are not mere fantasies), and assertions whose truth or falsity, cannot be confirmed or exposed by any experience; and this part of metaphysics is moreover precisely that which forms its essential end, toward which all the rest is only a means – and so this science needs such a deduction for its own sake. The third question, now put before us, therefore concerns, as it were, the core and the characteristic feature of metaphysics, namely, the preoccupation of reason simply with itself, and that acquaintance with objects which is presumed to arise immediately from reason's brooding over its own concepts without its either needing mediation from experience for such an acquaintance, or being able to achieve such an acquaintance through experience at all.