Wittgenstein argued, roughly, that a language has to be learned, and to be learned it has to be used first. Therefore, no private language. But if language has to be learned what then are the a priori categories of understanding, concepts, and forms of pure intuition? Kant tells us in Critique of Pure Reason:
"Whereas all intuitions, as sensible, rest on affections, concepts rest on functions. By function I mean the unity of the act of ordering various representations under one common representation... But a concept is always, as regard its form, something universal which serves as a rule".
This sounds very Wittgensteinian, but Kant's concepts are also very private. In all three Critiques the action takes place entirely in the private mind of a private individual confronting her private manifold of sensation. It is interesting that the conventional relativization of Kant makes it worse. If the a priori are concepts and are acquired (on a longer time scale), then how are they acquired (and communicated)? Kant did not offer a working account of new concept formation, Pippin analyzes his attempts in Kant on Empirical Concepts, perhaps this was one reason why he made his a priori so absolute.
If concepts are functions, rules, and "unities of the acts of ordering" then can we have private "language of thought" made of them? Or is Wittgenstein right, and we can not? Can we reconcile Kant and Wittgenstein?