In his first critique Kant shows why ideas in contrast to categories do not refer to adequate empirical objects. How can he claim at the same time that ideas be indispensable? Kant's answer to this problem is ambiguous. In this article I demonstrate that neither his attempt of a transcendental deduction of ideas nor his confirmation that ideas are useful for systematical unity of knowledge is convincing. The only solution is: a foundation of ideas by empirical knowledge in relation to (...) its systematical form. Finally I will point out that all this is not possible without the aid of practical reason. (shrink)

In this paper, I explain how Kant's views can be reconciled, and I argue that the relevance of transcendental idealism here is that it shows that determinism is known to be true, not in accordance with the familiar correspondence notion of truth, but only in accordance with a weaker notion of truth, Kant's empirical notion of truth, which is a kind of coherence notion of truth. (edited).