This is not an isolated instance, it illustrates the method of science in all theorising. An even more { 61} striking illustration than that we have just given is the case of the hypothesis of the luminiferous æther--a supposed existence, a fiction, that has served a useful, even an indispensable service in the history of modern physics. To many physicists, even to Lord Kelvin, the hypothesis seemed so surely established that its nonexistence hardly seemed thinkable, yet all the experiments designed to detect its presence have been uniformly negative in result, and it now seems not even necessary as a hypothesis, and likely to disappear. The æther was not only not discovered, it was not even suspected to exist, as in the case of the unknown planet Neptune--it was logically constructed. It was required to support the theory of the undulatory nature of light and to fulfil the possibility of light propagation in space. It was therefore a postulate, called forth by a need--so far we may adopt the pragmatist account. But what was the nature of the need, and what was the method by which the postulate was called forth? It is in answering this question that the pragmatist criterion fails. The need was intellectual in the purely logical meaning of the term, and it was met by a purely logical construction. The need was a practical human need only in so far as the intellect working by logical process is a human endowment but not in any personal sense such as is conveyed by the term psychological. Willingness or unwillingness to believe, desire, aversion, interest were all irrelevant. Given the intellect, the logical necessity was the only need that called forth by logical process the "truth-claiming" hypothesis of the æther. But even so, the pragmatist will urge, is its truth anything else but its usefulness as shown in the practical consequences of believing it? Was it not true while it was useful, and is it not only now false, if it is false, if it is actually discovered not to be useful? { 62} The reply is that no mathematician or physicist would recognise the possibility of working with a conception of truth that simply identified truth with utility, and for this reason that he can only conceive reality as a system whose truth is symbolised in an equation. It is the system that determines and characterises the postulate, and not the postulate advanced at a venture, tried and verified, that constitutes the system. The mathematician begins by placing symbols to represent the unknown factors in his equation, and proceeds by means of his known factors to determine their value. The æther is at first a pure fiction constructed to supply an unknown existence recognised as a defect. Its truth cannot mean that it works for it cannot but work, having been constructed purely for that purpose. Its truth means that it corresponds to some actual existence at present unknown. To prove its truth the physicist does not appeal to its value as a hypothesis, but devises experiments by which, if it does exist, its existence will be demonstrated. In this actual case the experiments have had a uniformly negative result, and therefore the truth of the hypothesis is made doubtful or denied. The hypothesis continues to work as well as it ever did, and physicists will probably long continue to use it, but it has failed to establish its truth claim. The result is the modern Principle of Relativity, which, as we have already said, has produced a revolution in modern physics. The abolition of the æther would have been impossible if the physicist had been content with the utility of his hypothesis and had not experimented to prove its truth. The relation between truth and utility is thus proved to be that it is useful to know what is true.