Kant claims that persons, by virtue of their rationality, can and should make a causal difference in the world of appearances. But how is this possible given that he also holds that the world of appearances is causally closed and that reason is not natural in the sense of the transcendental analytic? Causal closure and the non-naturalness of reason seem to leave only two options: causal inertness of reason, or systematic overdetermination. Yet both these options are obviously unattractive for our purposes, so that we need to enquire whether there is room for avoiding them in Kant's framework. I shall discuss first an (in the end unsuccessful) attempt to read Kant along the lines of Davidson's anomalous monism, and then present an alternative view which draws on the transcendental idealist reading of the relationship between essential properties and natural causal capacities.