us to see that, however faultless it may be as a test of truth, it is of no avail in furnishing a foundation for the activities of science and of ordinary life. Only in the case of things seen so clearly as to make doubt of them impossible, does our test of truth find any application; but we do not see enough to get a fulcrum for the levers of physical science and ordinary life. We have, then, a test of truth that is entirely reliable, but one that we cannot make any use of except in the case of the very propositions from which we derived it. What are we to do?
We must try to find a test of beliefs. By "test of belief," I do not, of course, mean that we must find a test by means of which we can determine which of our beliefs are true; that would be another name for a test of truth. I mean that, inasmuch as such "beings as men in such a world as the one we live in" must base a large part of their reasonings on propositions not known to be true, the most that we can do is to endeavor to be consistent. If we learn the characteristics of the beliefs that we are obliged to assume without proof, we can generalize them. We can say that, since we have accepted the trustworthiness of memory and the uniformity of nature and the proposition, 'An hypothesis that explains facts, and at the same time fits in with everything else that we believe is true,' we will accept any other proposition without further proof that has the same characteristics.
What, then, are the characteristics of the belief in memory? We can name one at least: the assumption of the trustworthiness of memory is one that we have a natural tendency to make. I mean by this to state no more than that, when we begin to reflect, we find ourselves making it. And we can name another: experience leaves it alone, does not deprive us of it, does not clash with it. But this confirmation of experience must be taken in a negative sense only. Of positive verification of the trustworthiness of memory, we have none; if we had, we should be speculatively sure of it. The assumption of the trustworthiness of memory, then, has two characteristics: it is one that we have a natural tendency to make and that experience does not deprive us of. The characteristics of
