explains, not only the facts known to the first man, but others equally certain, the last man's hypothesis is the more credible, although we can never say that it, in turn, may not have to give place to another. As a matter of fact, no one has ever believed anything on the ground of inductive reasoning, or, indeed, independently of all reasoning, that did not belong to one of these three classes. All that any one believes is either a necessary truth, or an ultimate belief, or an hypothesis accepted because it explains the facts and fits in easily with the rest of his beliefs.
What, then, it may be asked, is the practical outcome of this reasoning? It enables us to see the kind of evidence we ought to demand for anything we believe, and the kind of objection it is pertinent to make. If any one says, "I hold such and such a proposition to be a necessary truth," we are entitled to ask him to show it. It is absurd to say that a belief is a necessary truth whose contradictory I know I can conceive. To defend a belief as a necessary truth, we must show, not that its contradictory is hard to believe, not even that it cannot, as a matter of fact, be believed at all, but that it is in the literal sense of that much abused term unthinkable. To appeal to the testimony of consciousness to establish the necessity or self-evidence of a belief, which by the very appeal is shown not to be necessary or self-evident, is a degradation of philosophy, no matter who is guilty of it. But if I am asked to believe anything which is confessedly not a necessary truth, it must, as we have seen, belong to one or the other of the two remaining classes: those that I have a natural tendency to make and which experience does not interfere with, and hypotheses that explain facts and at the same time fit in naturally with the rest of the contents of our minds, and the only evidence I can demand for it, is that it shall have the characteristics of the class to which it is referred.
This theory will probably be objected to from a number of different points of view. It will be urged that it opens the door to unbounded credulity. Not so. The very prominence which it gives to the fact that inductive reasoning is only a
