219. It is clear that the process described in these passages supposes ‘ab initio’ the conversion of a feeling into a conception; in other words, the substitution of the definite individuality of a thing, thought of under attributes, for the mere singleness in time of a feeling that occurs after another and before a third. The ‘finding of resemblances and differences among objects that often occur to us’ implies that each object is distinguished as one and abiding from manifold occurrences, in the way of related feelings, in which it is presented to us, and that these accordingly are regarded as representing permanent relations or qualities of the object. Thus from being related feelings, whether more or less ‘vivacious,’ they have become, in the proper sense, ideas of relation. The difficulty about the use of general names, as Hume puts it, really arises just from the extent to which this process of determination by ideas of relation, and with it the removal of the object of thought from simple feeling, is supposed to have gone. It is because the idea is so complex in its individuality, and because this qualification is not understood to be the work of thought, by comparison and contrast accumulating attributes on an object which it itself constitutes, but is regarded as given ready-made in an impression (_i.e._ a feeling), that the question arises whether a general proposition is really possible or no. To all intents and purposes Hume decides that it is not. The mind is so tied down to the particular collection of qualities which is given to it or which it ‘finds,’ that it cannot present one of them to itself without presenting all. Having never found a triangle that is not equilateral or isosceles or scalene, we cannot imagine one, for ideas can only be copies of impressions, and the imagination, though it has a certain freedom in combining what it finds, can invent nothing that it does not find. Thus the idea, represented by a general name and of which an assertion, general in form, is made, must always have a multitude of other qualities besides those common to it with the other individuals to which the name is applicable. If any of these, however, were included in the predicate of the proposition, the sleeping custom, which determines the mind to pass from the idea present to it to the others to which the name has been applied, would be awakened, and it would be seen at once that the predicate is not true of them. When I make a general statement about ‘the horse,’ there must be present to my mind some particular horse of my acquaintance, but if on the strength of this I asserted that ‘the horse is a grey-haired animal,’ the custom of applying the name without reference to colour would return upon me and correct me--as it would not if the predicate were ‘four-footed.’ He virtually yields the point in regard to the _predicate_ of propositions.