The principles of problematic induction
In C. D. Broad (ed.) Induction, Probability, and Causation, Dordrecht, 1968, pp. 86–126
Abstract
By Problematic Induction I mean any process of reasoning which starts from the premise that all, or a certain proportion of, observed S’s have had the characteristic P, and professes to assign a probability to the conclusion that all, or a certain proportion of, S’s will have this characteristic. It is assumed that no intrinsic or necessary connexion can be seen between the characteristics S and P. Where such a connexion can be seen, the fact that all observed S’s have been found to be P can hardly be called a logical premise; it is at most a psychological occasion which stimulates the observer to intuit the intrinsic connexion between S and P. The latter process is called Intuitive Induction by Mr. Johnson, and I do not propose to consider it here. It is generally admitted that, when intuitive induction is ruled out, premises of the kind which we are considering can lead only to conclusions in terms of probability. This has been argued independently by Mr. Keynes and myself, and I am going to assume that it is true.