Laurence Bonjour takes a large step forward in helping to provide a rational justification for induction.

As Hume formulated the problem, any attempt to validate induction is snared by one or another horn of a dilemma. It would have to be either an a priori argument or an empirical one. It cannot be the former, because all demonstrative reasoning depends on the avoidance of contradiction, and it is not a contradiction to argue that inductive reasoning could fail due to changes in the course of nature. Nor can the argument be empirical, because it would be begging the question for induction to rely on itself for its justification.

Two contemporary lines of defense have been constructed. Bonjour argues that both are inadequate.

The first, the 'pragmatic' defense, was originally developed by Hans Reichenbach. It holds that while Hume's argument is technically correct, we can at least say that if there is in fact a relationship between repetitively observed phenenomena, induction will find it over the long run. This position does not satisfy Bonjour, who wants more support for induction than The Best We Can Do Must Be Good Enough.

The second response to Hume's argument is supplied by the philosophers of ordinary language. Bonjour concentrates on P.F. Strawson's version. Strawson contends that believing in accordance with strong evidence is believing reasonably, and believing in accordance with inductive standards is believing in accordance with strong evidence: therefore belief based on induction is reasonable belief. But Bonjour finds a flaw in this approach, which he traces to an equivocation in the phrase 'believing in accordance with strong evidence'. In the first clause, that phrase must mean something like 'believing in accordance with evidence that actually establishes a strong likelihood that the belief is true', while in the second clause it means something like 'believing when the evidence is strong according to generally accepted standards'.

Other attempts to defeat Hume's argument have also failed, according to Bonjour. So what solution can we hope to devise?

His suggestion is that we develop a synthetic a priori argument for induction. He defends at some length the possiblity of such an argument against the empiricist dictum that the synthetic a priori is impossible. His principal point is that the empiricist position on this point is inconsistent, for it itself is either a synthetic a priori or a circular argument. Bonjour rebuts specific empiricist attempts to refute this charge, and he concludes that we are justified in adopting the synthetic a priori strategy if we can make it work.

He only outlines how such an argument for induction can be developed, however, conceding that a good deal more work needs to be done. In simple terms, he suggests that the inductive explanation is the best there is for the regularities we observe, that these observed regularities reflect corresponding objective reality in the world. Bonjour concludes briefly by commenting on some oddities and objections about this position.

Contents

Index