Abstract

Mr. Clark argues that if Kant is to be interpreted as a realist his theory of space and time must be based on inductive argument. In support of this contention he suggests that a realist must be a nominalist and, hence, can legitimately advance only inductive arguments in support of his doctrines. But Mr. Clark goes further than this. He states that "if Kant is not to be taken as a realist, still his arguments, as written in the Critique, are as a matter of fact inductive." I am not sure in what sense Mr. Clark is using the term "realist" here. I fail to see that realism necessarily implies nominalism or, for that matter, that even an inductive argument is possible for a thoroughgoing nominalist. But this point is not particularly important for the discussion of Kant's views since Kant is an idealist as well as a realist. In my article I attempted to stress the realistic side of Kant's theory. But I did not mean to argue that his position can be construed as exclusively realistic. Kant characterizes his theory as maintaining at once the "empirical reality and transcendental ideality" of space and time. His formal idealism was expressly designed to provide an escape from nominalism and to make possible a non-inductive or "transcendental" method of proof. Kant would agree with Mr. Clark that if space and time are regarded as things-in-themselves we can have no a priori knowledge of them. Space and time are real for Kant, but only as forms of appearances. I did not mean to obscure the idealistic feature of space and time, but only to stress the fact that Kant did not regard them as psychologically ideal. Mr. Clark mistakenly concludes that Kant's doctrine must be either realistic or idealistic but not both.