Causality Beyond Its Limit?

This is an informal piece of writing for PHILOS 151B taught by Prof. Dewitt at UCLA.

Autonomy and self-legislation are at the very core of Kant's practical philosophy, as they are the very source of our dignity as rational beings. However, both of these are formulated in terms of the will causing/determining itself, despite that Kant himself has already pointed out that the concept of causality "has application and so too significance strictly speaking only in reference to appearances. (KpV 5:49)" in KrV and KpV. Still, Kant insists on the validity of such concept for our will in practical reason, and tries to explicate how that is possible in the deduction in KpV.

However, as it often has been, his explanation from 5:49-50 is not that easy to parse and may even seem crazy at time -- he first admits that the such validity cannot be shown in the same way as in theoretical reason, and makes the move to claim that practical reason needs only to put "the determining ground of the causality of the human being ... in pure reason ... and accordingly use[s] the concept of cause itself." He seems to be saying that we first act according to autonomy, and then we will be able show the significance/validity of the concept of causality by pointing to that part of the principle of autonomy in us.

But this still sounds circular to me: To make the concept of causality for will valid, we have to first act according to autonomy. However, to act according to autonomy, which is a rule that involves "self-causality" (if that is a valid concept), doesn't it presume us to understand what the "causality" of will is? For this argument to be valid, Kant has to presuppose that we can act from morality without understanding what the moral law is. The only way I can think of to explain how this is possible is to say that our will (somehow, and uncognizably) is inherently autonomous causa noumenon, and when we cognize its "action" from the side of noumenon we can "abstract" (5:49) its patterns of operation and call that pattern "self-causality/autonomy". This interpretation still seems unsatisfactory as it leaves the causa noumenon unknown, but in fact such "gap" corresponds to his notion of the fact of reason, which cannot be known from any antecedent date of reason. (5:31) That is to say, from antecedent data not only can we not know if there is a moral law, but we can not even know what the concept of moral law is/means.