in the Inquiry, under three main heads: The general question of causation,[1] the particular question of causation,[2] and the idea of necessary connection.[3] Professor Huxley states that the evidence by which Hume supports his conclusion in the Inquiry, concerning the general question of causation, is not strictly relevant to the issue.[4] And Mr. Selby-Bigge asserts that Hume passes over this question in the Inquiry.[5] It seems to me, however, that he does not quite pass over the general question of causation in the later work. True, the evidence by which he supports his conclusion is almost wholly irrelevant.[6] But the same may be said of his reasoning in the Treatise, his chief argument there being a petitio principii.[7] Mr. Selby-Bigge also contends that the relation of 'contiguity' holds relatively a much less prominent place in the Inquiry than it did in the Treatise; and that in the later work the distinction between cause as a natural, and cause as a philosophical relation is no longer observed. In reply to the first point, it may be admitted that in "the account of the origin, in particular cases, of the idea of cause and effect," in the Inquiry, "'contiguity' practically drops out altogether." Yet this seems to have been occasioned chiefly by the somewhat different order of discussion in the later work; perhaps partly also by the evident confusion of the question of causation with that of the law of the uniformity of nature. According to the Treatise, inference is a transition of the mind from a present impression of sense or memory to a related idea,—related according to the two primary laws of association, resemblance and contiguity.[8] And the ultimate ground of the inference, consequently "the foundation of all conclusions from experience," is custom.[9] According to the Inquiry, the reason, or ground on which "we form an inference from one [instance] to another," and consequently, "the foundation of all conclusions from experience," seems to be the assumption that nature is uniform. 10 Thus far the infer-
