Here we come to the doctrine which for him and his school gave the key to all psychological problems. It was James Mill's real merit, according to his son, that he carried the principle of association of ideas further than it had been carried by Hartley or other predecessors.[514] The importance of the doctrine, indeed, is implied in the very statement of the problem. If it be true, or so far as it is true, that our consciousness reveals to us simply a series of 'sensations' and 'ideas,' the question must be how they are combined. 'Thought succeeds thought, idea follows idea incessantly,'[515] says Mill; and this phrase assumes 'thoughts' and 'ideas' to be separable atoms. How, then, do they come to coalesce into an apparently continuous stream? The mind is a stream of 'ideas.' If the stream is composed of drops, we must, of course, consider the drops as composing the stream. The question is, What laws can we assign which will determine the process of composition? The phrase 'association' admittedly expresses some general and very familiar truths. Innumerable connections may be established when there is no assignable ground of connection in the ideas themselves other than the fact of a previous contact. One idea not only calls up the other, but in some way generates a belief in an independent connection. We hear thunder, for example, and think of lightning. The two ideas are entirely distinct and separate, for they are due to different senses. Yet we not only think of lightning when we hear thunder, but we have no doubt that there is a causal connection. We believe in this connection, again, though no further explanation can be given of the fact. Thunder and lightning have occurred together, and we infer that they will, and even must, occur together. When we examine our whole structure of belief, we find such 'arbitrary' associations pervade it in every direction. Language itself is learned simply by association. There is no connection whatever between the sound of the word 'man' and the 'ideas' which the word excites, beyond the fact that the sound has been previously heard when the ideas were excited. Here, then, is a phenomenon to be explained or generalised. We have in countless cases a certain connection established for which no further reason can be assigned than the fact of its previous occurrence. On such a ground, we believe that fire burns, that bread is wholesome, that stones fall; and but for such beliefs could know nothing of the outside world. 'Contingent' truth, therefore, or truth derived from mere contact, pervades, if it does not constitute, the whole fabric of our whole knowledge. To prove that all our knowledge is derived from experience is, according to Mill, to prove that in some sense or other association of ideas lies at the base of all intellectual processes. When Locke introduced a chapter upon 'Association of Ideas' into the fourth edition of his essay, he treated it as the exceptional case.