The unity of consciousness

We now turn to the jungle of the Transcendental Deduction of the Categories. Kant's justified discontent with this desperately ill-written chapter led him to rewrite it for the second edition; but the later version, though it helpfully shifts some emphases, is only marginally clearer than the earlier. Attempts have been made to canonize the whole text of the Transcendental Deduction, but the results have been derisible. Far more usefully, Vaihinger and Kemp Smith have explained the unclarity of the first version of the Deduction by the conjecture that it is a collage made up of jottings written down over the years. However this fares as a biographical hypothesis, it is at least backed by some fairly sound critical exegesis. And yet, even if the Transcendental Deduction is reshuffled in accordance with this ‘patchwork theory’, the result, though it shows something like a linear development from primitive to sophisticated attacks on the same problems, is still dreadfully confused. The Deduction is not a patchwork but a botch. Since it contains some good things, however, it is not a negligible botch.

My aim in §§28–31 will be to lay bare certain arguments which are central to the Transcendental Deduction and from which something can be learned. While relating my remarks as closely as possible to the Kantian text, I shall leave many exegetical problems unsolved and all the exegetical debates unadjudicated.

The Transcendental Deduction can usefully be seen against the background of a possible empiricist philosopher who takes over Hume's basic epistemological doctrines and develops them into the account of objectivity-concepts which Hume himself failed to develop only because he could not see that there is clear water between the Scylla of Berkeley and the Charybdis of Locke. This empiricist argues that the world might at any moment become such that the objects in it cease to obey causal laws, or even such that it will cease to contain durable objects at all. His conclusion that objectivity and causality are at any moment liable to collapse is derived from a pair of premisses. (1) There is no logical connection between the content of experience at one time and its content at another time.