this: experience is only what has actually been; and, after never so many repetitions of a thing, there still remains the peril of venturing upon the untrodden land of future possibility. 'What has been, will be,' justifies the inference that water will assuage thirst in after times. We can give no reason or evidence for this uniformity."[1] So I think, and since we do not see it to be either true or probable, it is an assumption, a thing we believe in primarily because of the constitution of our emotional nature, and not because it appeals to us on intellectual grounds.
"This assumption," continues Bain, "is an ample justification of the inductive operation, as a process of real inference. Without it, we can do nothing; with it we can do anything."[2] It is certainly true that without this assumption we can do nothing; but it is not true that it alone is sufficient to enable us to make a conquest of nature, or, indeed, any approximation to it. So far from admitting that with it we can do anything, I assert that with it alone we can do nothing. With an absolutely certain knowledge of nothing but some of our present states of consciousness, we make a leap to the past; assuming to know the present and the past, we make a leap to the uniformity of nature; and, I hold, assuming to know the uniformity of nature, we make a leap to conclusions apparently directly deduced from it.
Perhaps the best way to bring out the truth of this contention is to examine briefly John Stuart Mill's theory of induction. Mill held that valid inductions terminate in proof; that the object of inductive reasoning is to obtain for its conclusions a kind of evidence that for a pure intellect would justify belief. In opposition to Whewell and Jevons, he denied that induction is a process of finding hypotheses to explain facts, insisting that only when a further step has been taken, when it has been shown not only that a given hypothesis explains the facts but that no other can, is an induction valid. From his point of view, induction may be briefly described as a process of finding hypotheses to explain facts, and proving that no other hypothesis
