The topic of this paper is Kant’s account of propositional knowledge (Wissen), its relationship to his account of cognition (Erkenntnis), and its role in motivating the famous prohibition on knowledge of things-in-themselves. I offer a reading of the prohibition that is motivated not so much by Kant’s anti-soporific encounter with Hume as by his new view of the distinction between “real” and “logical” modality, a view that developed out of his reflection on the rationalist tradition in which he was trained. In brief: at some point in the 1770’s, Kant comes to hold that a necessary condition on knowing a proposition is that one be able to prove that all the items it refers to have a certain modal status. Propositions about things-in-themselves, in turns out, cannot meet this condition. I conclude by suggesting that we might best interpret Kant’s Modal Condition, at least in case of empirical knowledge, as a coherentist constraint that makes essential reference to the laws of nature.