Deduction and Historical Explanation

Abstract

Recently Professor William Dray counted it as an advantage of one account of historical explanation over another that on the former the explanation focused precisely on what was to be while the latter, which rested on statistical information, allowed the individual event to be explained room to rattle around in the sense that the statistical law given as an explanation is compatible with both the occurrence and non-occurrence of what is thus explained.' The authors of the two theories are Professors Scriven and Hempel.2 The irony of Dray's congratulating Scriven on being more faithful than Hempel to this aim of the hypothetico-deductive model has probably struck many. The situation is remarkable partly because Dray has become known as one of the best critics of that model; and more especially, I think, because Dray has been the most concerned to argue that explanations of human actions in terms of the reasons for which they are done are essential to history, and to try to give an adequate account of them. One who insists on the importance of reasons as opposed to causes in historical explanations might be expected to find sources of intelligibility other than that which accrues when what is to be explained is deducible from the explanation. But there does not seem to be any non-deductive way to stop the rattling. Another aspect of the irony of Hempel's giving up deduction as necessary while Dray and Scriven try to save it is only apparent. Hempel still believes in the logical symmetry between explanation and prediction and takes it as a necessary condition of any rationally acceptable answer to Why? that it provide information which constitutes good grounds for the belief that X