KNOWLEDGE over time No one was more perturbed by Hume’s sceptical challenge than Immanuel Kant, and no one worked harder to meet the challenge and to re-establish the function of the intellect in the ordering of our perceptions Just as Hume started his argument with a contrast between matters of fact and relations of ideas, Kant begins his response by making distinctions between diVerent kinds of propositions But instead of a single distinction, he has a pair of distinctions to make, one epistemological and one logical First, he distinguishes between two modes of knowledge: knowledge derived from experience, which he calls a posteriori knowledge, and knowledge independent of all experience, which he calls a priori knowledge Next, he makes a distinction between two kinds of judgement, analytic and synthetic He explains how to decide to which kind a judgement of the form ‘A is B’ belongs: Either the predicate B belongs to the subject A, as something that is contained (though covertly) in the concept A, or it lies quite outside the concept A even though it is attached to it In the Wrst case, I call the judgement analytic, in the second synthetic (A, 6) He gives as an example of an analytic judgement ‘all bodies are extended’, and as an example of a synthetic judgement ‘all bodies are heavy’ Extension, he explains, is part of the concept ‘body’, whereas weight is not Kant’s distinction between analytic and synthetic propositions is not wholly satisfactory It is clearly intended to be universally applicable to propositions of all kinds, yet not all propositions are structured in the simple subject–predicate form he uses in his deWnition The notion of ‘containing’ is metaphorical and although the distinction is clearly intended to be a logical one, Kant sometimes speaks of it as if it were a matter of psychology Some later philosophers tried to tighten up the distinction, and others tried to break it down; but it retained a permanent place in subsequent philosophical discussion What is the relation between the epistemological distinction a priori/a posteriori and the logical distinction analytic/synthetic? The two distinctions are made on diVerent bases, and they not, according to Kant, coincide in their application All analytic propositions are a priori, but not all a priori propositions are analytic There is no contradiction in the notion of a synthetic a priori proposition, and indeed there are many examples of such 157