Truth criteria and the very project of a transcendental logic

This paper argues that Kant's idea for a new kind of logic is bound up with a very specific strategy for obtaining truth criteria, where he takes Christian Wolff to have failed. While the First Critique 's argument against any universal criterion for empirical truth has almost always been treated as extraneous to the main concerns of the Transcendental Analytic, I argue that Kant inserted it at an important juncture in the text to illustrate a signal difference between traditional logics and transcendental logic. Namely, while a criterion of truth as correspondence cannot be provided by traditional logics, since they cannot, in Kant's view, identify an object determinately and distinguish it from others, transcendental logic overcomes that particular barrier. The key to the improvement is to be found in how Kant modifies Wolffian order and “transcendental truth”, while still retaining them as central elements of his project. Once we have reconstructed Kant's strategy, we also gain a new perspective on Kant's puzzling assessments of his relation to Berkeley.