Why they do not trouble him more. 114. It is Hume’s more logical version of Locke’s doctrine that first forces these questions to the front. In Locke himself they are kept back by inconsistencies, which we have already dwelt upon. For the real, absolutely void of intelligible qualities, because these are relative to the mind, he is perpetually substituting a real constituted by such qualities, only with a complexity which we cannot exhaust. By so doing, though at the cost of sacrificing the opposition between the real and the mental, he avoids the necessity of admitting that the system of the sciences is a mere language, well-or ill-constructed, but unaccountably and without reference to things. Finally, he so far forgets the opposition altogether as to find the reality of ‘moral and mathematical’ knowledge in their ‘bare ideality’ itself. (Book IV. chap. iv. sec. 6, &c.) Thus with him the divorce between knowledge and reality is never complete, and sometimes they appear in perfect fusion. A consideration of his doctrine of propositions will show finally how the case between them stands, as he left it. They re-appear in his doctrine of propositions. 115. In the Fourth Book of the Essay the same ground has to be thrice traversed under the several titles of ‘knowledge,’ ‘truth,’ and ‘propositions.’ Knowledge being the perception of agreement or disagreement between ideas, the proposition is the putting together or separation of words, as the signs of ideas, in affirmative or negative sentences (Book IV. chap. v. sec. 5), and truth--the expression of certainty [1]--consists in the correspondence between the conjunction or separation of the signs and the agreement or disagreement of the ideas. (Book IV. chap. v. sec. 2.) Thus, the question between the real and the mental affects all these. Does this or that perception of agreement between ideas represent an agreement in real existence? Is its certainty a real certainty? Does such or such a proposition, being a correct expression of an agreement between ideas, also through this express an agreement between things? Is its truth real, or merely verbal? [1] All knowledge is certain according to Locke (Cf. IV. chap. vi. sec. 13, ‘certainty is requisite to knowledge’), though the knowledge must be expressed before the term ‘certainty’ is naturally applied to it. (Book IV. chap. vi. sec. 3.) ‘Certainty of knowledge’ is thus a pleonastic phrase, which only seems not to be so because we conceive knowledge to have a relation to things which Locke’s definition denies it, and by ‘certainty,’ in distinction from this, understand its relation to the subject. ‘Certainty of truth’ is, in like manner, a pleonastic phrase, there being no difference between the definition of it (Book IV. chap. vi. sec. 3) and that of ‘truth’ simply, given in Book IV. chap. V. sec. 2. The knowledge expressed by a proposition, though certain, may not be real ...