In this paper I investigate Kant's conceptions of the sources, limits and mutual interrelationships of our knowledge of logic and mathematics. Kant's view of the limitations of mathematical knowledge by forms of intuition, in contrast with the negative generality of logical knowledge, is not a radical difference from the analytic tradition. The large differences between the Aristotelian principles of inference available to Kant and the laws of higher-order quantificational generality set out by Frege and Russell obscure a more fundamental agreement on the nature of logic. The genuinely fundamental disagreement is over whether the notions of identity and difference underlying arithmetical judgment have a purely (sortal) conceptual basis. I end by discussing the relations of these conceptions of logic and of the epistemic bounds of mathematics to the notions of sameness and difference central to the semantical, epistemological and logical thought of Mohism.