The fact is, of course, that an idea can be correct or incorrect only if by the word idea we mean not a datum but an opinion; and the abstract relation of correctness, by virtue of which any opinion is true, is easily stated. An opinion is true if what it is talking about is constituted as the opinion asserts it to be constituted. To test this correctness may be difficult or even impossible in particular cases; in the end we may be reduced to believing on instinct that our fundamental opinions are true; for instance, that we are living through time, and that the past and future are not, as a consistent idealism would assert, mere notions in the present. But what renders such instinctive opinions true, if they are true, is the fact affirmed being as it is affirmed to be. It is not a question of similarity or derivation between a passive datum and a hidden object; it is a question of identity between the fact asserted and the fact existing. If an opinion could not freely leap to its object, no matter how distant or hypothetical, and assert something of that chosen object, an opinion could not be so much as wrong; for it would not be an opinion about anything.
Psychologists, however, are not concerned with what an opinion asserts logically, but only with what it is existentially; they are asking what existential relations surround an idea when it is called true which are absent when it is called false. Their problem is frankly insoluble; for it requires us to discover what makes up the indicative force of an idea which by hypothesis is a passive datum; as if a grammarian should inquire how a noun in the accusative case could be a verb in the indicative mood.
It was not idly that William James dedicated his book on Pragmatism to the memory of John Stuart Mill. The principle of psychological empiricism is to look for the elements employed in thinking, and to conclude that thought is nothing but those elements arranged in a certain order. It is true that since the days of Mill analysis has somewhat extended the inventory of these elements, so as to include among simples, besides the data of the five senses, such things as feelings of relation, sensations of movement, vague ill-focused images, and perhaps even telepathic and instinctive intuitions. But some series or group of these immediate data, kept in their crude immediacy, must according to this method furnish the whole answer to our question: the supposed power of an idea to have an object beyond itself, or to be true of any other fact, must be merely a name for a certain position which the given element occupies in relation to other elements in the routine of experience. Knowledge and truth must be forms of contiguity and succession.