(The following is a brief note in response to this post.)
Rorty once suggested that the peculiar fate of transcendental argumentation is its independence from and even its opposition to transcendental philosophy. Since Davidson we have been rightfully suspicious of the distinction between content and schema that seems to be central to Kantian philosophy and which falls on its own terms. Instead, however, of the idealist separation of form and content, the minimal, irreducible difference on which transcendental argumentation turns is between what there is and what can be said about it (which holds for any recognizable transcendental argument from Kant to Wittgenstein, Strawson, and Putnam). But the price that transcendental argumentation must pay is truth as correspondence. In fact, any strictly transcendental argument must surrender the prima facie objective validity of any reference other than self-reference, where the latter functions as the essential logical form of transcendental argumentation (“you cannot reject X without presupposing X”) as well as the ultimate purchase of such arguments (which result in knowledge about but not knowledge of). Perhaps against himself – and against his absolute idealist critics – what Kant demonstrated was that we lack knowledge of our own subjectivity and, indeed, criticism consists in nothing other than the fact that subjectivity can always be called into question. But such questioning proceeds hypothetically (“if you say Y, then you must presuppose X”) and negatively, i.e., transcendental philosophy must reject any particular fact as epistemically basic since all such facts are subject to constitutive rules governing the possibility of their interpretation, viz., qua facts, but which themselves say nothing about the world. All transcendentalism is therefore a structuralism that insists on a tripartite distinction of language, thought, and world founded on the excess of each to the others.