444 CRITICAL NOTICES I

his philosophical fame reposes ". This is erroneous. The Inquiry has been far more widely known, and it consequently has had much more influence on the evolution of thought than the Treatise. We may conveniently divide the strictly philosophical part of Prof. Knight's work under three heads (1) the review of Hume's predecessors, (2) the exposition of Hume, (3) the criticism of Hume. The historical retrospect seems to me for the most part super- fluous, and not always accurate. It is inexact to represent matter and mind in their relative independence as the funda- mental existences recognised by Descartes, together exhausting reality. The primary fact for Descartes was in the ordo ad nos the Ego, in the ordo ad universiun God. The existence of matter depended from instant to instant on the divine causality, and was known to us merely as a corollary from the divine veracity. It is inexact also to say that Spinoza identified material and mental phenomena by " bringing in a tcrtium quid distinct from both, to which he affixes another name ". It is both inexact and contradictory to call the relation of Hobbes and Gassendi one of mutual indebtedness, and at the same time to add that " Gassendi's works were published earlier, and his theories wrought out in independence and isolation ". I fail to see that this historical retrospect deserves insertion through any conspicuous merit of its own, and I am unable to discover what value it has for the general plan of the book. It would have been better if the account of Hobbes, Spinoza and Leibniz had been omitted, so as to make room for a fuller exposition of Locke. Prof Knight declares that " in explaining Locke we virtually explain Hume," and yet makes no reference to the cardinal doctrine of Locke's epistemology as contained in bk. iv. of the Essay. " In some of our ideas there are certain relations, habitudes and connexions, so visibly included in the nature of the ideas themselves, that we cannot conceive them separable from them by any power whatsoever. And in these only are we capable of certain and universal knowledge." Here indeed we have a " virtual explanation " of Hume. But there is no hint of this in Prof. Knight's book. The impression produced by it is that there is nothing in Locke from beginning to end but " empirical psychology ". Closely allied with this imperfect appreciation of Hume's rela- tion to Locke is the imperfect appreciation which Prof. Knight displays of Hume's general attitude to philosophical questions, as distinguished from his treatment of special problems. No explicit account is given of Hume's doctrine of belief as distinguished from knowledge. Nothing is said of " that general maxim in the science of human nature," that wherever there is a close relation betwixt two ideas, " the mind is very apt to mistake them, and in all its discourses and reasonings to use the one for the other ". No reference is made to the discussion of Space and Time (in pt. ii.