a state of absolute uncertainty concerning all matters that come before us as it is to escape from our shadows. How do we in practice untie the knot? We do not untie it, we cut it. As we arbitrarily, from the point of view of the pure intellect, assume the trustworthiness of memory and the uniformity of nature, so we arbitrarily assume that the evidence is all in, and that known causes have brought about the effects we are studying—assume it until we discover facts that prove the assumption false, when we invent by hypothesis an unknown cause and endow it with attributes suitable for our purposes. Instead, then, of being able to do anything, as Bain says, assuming the uniformity of nature and the trustworthiness of memory, we can do nothing. To make these assumptions weapons of real efficiency we must arm ourselves with another. How shall we state it?
It will not do to say that we assume that any hypothesis that explains the facts we are considering is true, because it sometimes happens that more than one hypothesis will equally well explain them. By "equally well" I mean that, if the cause supposed in either of the hypotheses exists and acts as we suppose it does, the facts would be as we find them. Nor is it true that we have an inclination to believe any hypothesis that explains the facts. On the contrary, an hypothesis may seem to us absurd, as in the case of a possible supernatural agency represented as the cause of a violent death, although we may readily enough admit that if the cause supposed by it existed, it might produce the violent death. In what consists the absurdity of such an hypothesis? In its lack of harmony with our other beliefs. In order that an hypothesis may win our belief, it must not only explain the facts it was invented to explain, but it must fit in snugly with the rest of our beliefs. Three hundred years ago, a belief in witches was quite general; many facts were supposed to owe their existence to their agency. Why have such explanations been abandoned? Not because they will not account for the facts, but because other beliefs have gradually grown up in the minds of men, with which the belief in witches is inconsistent. That an hypothesis may
