The omission becomes important at the next step. 'Under the modest title of an explanation of the meaning of several names,' says his son, James Mill discusses 'some of the deepest and most intricate questions in all metaphysics.' A treatise on chemistry might almost as well be 'described as an explanation of the names, air, water, potass, sulphuric acid, and so forth.'[547] Why does the chapter come in this place and in this peculiar form? Probably because James Mill was partly conscious of the inadequacy of his previous chapters. The problems which he has been considering could not be adequately treated by regarding ideas as 'things' bound together by association. What, after all, is a proposition? What is meant by 'true' or 'false,' as distinguished from real and unreal? If an association actually _is_ a truth, what is the difference between right and wrong associations? Both are facts, and the very words 'right' and 'wrong,' that is, true and false, apply not to facts but to propositions.[548] The judgment is tested in some way by correspondence to the 'order of Nature,' or of our sensations and ideas. What precisely is meant by this order? So far as we have gone, it seems as if ideas might be combined in any order whatever, and the most various beliefs generated in different minds. Perhaps, however, the principle of association itself may reveal something as to the possible modes of coalescence. Mill makes contiguity an ultimate ground of association; and contiguity implies that things have certain relations expressible in terms of space and time and so forth. These primitive relations now come up for consideration, and should enable us to say more precisely what kind of order is possible. In fact, Mill now endeavours to analyse the meanings of such words as relation in general, time, space, number, likeness, personal identity and others. The effect of his analysis is that the principles, whatever they may be, which might be supposed to underlie association appear to be products of association. He begins by asking what is the meaning of 'relative terms.' Their peculiarity is that they 'always exist in pairs,' such as 'father and son,' 'high and low,' 'right and left.' 'If it is asked, Why do we give names in pairs? the general answer immediately suggests itself; it is because the things named present themselves in pairs, that is, are joined by association.'[549] J. S. Mill thinks that no part of the _Analysis_ is more valuable than the 'simple explanation' which follows. There is no 'mystical bond called a relation' between two things, but 'a very simple peculiarity in the concrete fact' marked by the names. In 'ordinary names of objects, the fact connoted by a name ... concerns one object only'; in the case of relative names, 'the fact connoted concerns two objects, and cannot be understood without thinking of them both.' A 'fact concerning an object' is a curiously awkward expression; but one point is clear.