On the Vindication of Induction

A practical or pragmatic justification ("vindication") applies to actions. The action concerned in the case of induction is the making of predictions, and -philosophically of prime importance-the adoption of such rules of procedure as will make the predictions maximally successful. Clearly all ordinary cases of the justification of actions utilize, and in this sense presuppose, inductions. (A prospector looking for gold presupposes that there is gold to be found somewhere in the area of his search. If he had no reason whatever for expecting gold there, we should call his belief irrational, a product of wishful thinking. Inductive evidence for the presence of gold is required.) When philosophers ask for a ground of induction in general the answer cannot be inductive evidence. This would be plainly circular or lead to an infinite regress. Hence we are here dealing with a limiting or degenerate case of justification. It is true that one important component of the meaning of such words as "reasonable" or "rational" is indeed the employment of inductive procedures. Many analytic philosophers rest their case right here and consider the quest for a justification of induction as a pseudo problem because, in their view, this quest comes down to asking "is it reasonable to be reasonable ?" But-because of an ambiguity of the term "reasonable"-it is precisely here where the distinction between validation and vindication is helpful and points the way to one more step that can be taken in the overall justification of inductive inference. The ascription of a certain degree of probability (i.e. degree of confirmation or evidential support) can be validated in the light of the available evidence and a rule of induction (or a definition of "degree of confirmation"). But when we ask for a justification of the given rule of induction, or for a given definition of "degree of confirmation," the justification amounts to a vindication. We need not be satisfied with saying that this rule (or the corresponding definition) is rational or reasonable because it conforms with the very meanings of "reasonability" or "rationality". We can ask with what right these honorific terms are applied to the procedure of inductive inference. True, only philsosophers who have been through the treadmill of (Cartesian or Humean) doubts will raise this question. Admittedly, these doubts are of a rather extraordinary sort. They do not occur in the business of every day life, nor do they occur in the ordinary pursuits of science. It is clearly useful and enlightening to distinguish ordinary empirical doubt from philosophical doubt. To have * Received November, 1960. This is a reply to Professor Daniel Kading's discussion "Concerning Mr. Feigl's 'Vindication' of Induction," in this journal, Vol. 27, No. 4, pp. 405-407, October 1960.