The logical status of the principle of induction

Abstract

For the purposes of the following discussion, let me assume that there is a statement which may be called the principle of induction. I also want to assume that the proposition has the form, similar objects behave in similar ways. That is certainly a very rough formulation. It could be made a little better by saying, similar objects behave in similar ways under similar circumstances. The process of trying to make the principle suffi ciently precise is probably and endless one. For that reason I do not want to begin it. The problem of induction arises when a philosopher holds three proposi tions. They are: (1) The principle of induction states a fact about the world, (2) The principle is empirically untestable, and (3) The principle is a necessary premise of every inductive argument. It is from the last proposition, that the principle is a necessary premise of every inductive argument, that the problem of induction derives its philosophical importance. The problem itself, however, arises from the first two propositions, that the principle says something about the world but is empirically untestable. Everyone, I think, knows the usual Humean formulation of the problem. If the principle of induction says something about the world, then it can be proved or established only by an inductive argument based on empirical evidence. But, since the principle is a premise of every inductive argument, no inductive argument can establish its truth. Actually, the problem of induction needs one more supposition which is provided by the context of traditional western philosophy stemming from Descartes. Descartes' method was to doubt every proposition until he could find an indubitable proof for it. And, in order to maintain him self in a state of doubt, Descartes decided to regard as false any statement he had not deductively proved from incontrovertible premises. The problem of induction comes from applying the Cartesian attitude to the principle of induction. In asking that acceptance of the principle be justified, philosophers in effect look upon it as initially doubtful. Some