Abstract

Given the greater success of the context-dependent account, the philosophical conception of the role of explanation in science should be re-examined. Explanation should be seen as a part of applied science rather than pure science. Explanatory power should be seen as a pragmatic virtue of theories rather than a virtue with the same status as consistency or empirical adequacy. It would seem that to argue from the structure of explanations which we accept to the fact that science has a certain structure or that scientific knowledge has a certain status is illegitimate. This can be seen in relation to several issues in the philosophy of the sciences, most notably, the debate over the status of theoretical entities and the question of the nature of the social sciences. ;Salmon attempts to resolve the difficulties in Hempel's account by treating explanations as collections of statistically and causally relevant information. Since explanations can only include information which is both statistically and causally relevant on this theory, the problem of irrelevant information appears to be eliminated. Salmon accounts for the asymmetry of explanation by claiming that all scientific explanation is causal. Part of his characterization of causal relations includes that causes are temporally prior to effects and so the asymmetry of explanation results from the asymmetry of time. However, it is not completely clear that all cases of asymmetry can be accounted for in this way. Also Salmon's theory demands a fundamental role for causation in science in a way which is open to dispute. ;The shortcomings of these two accounts suggest that the sort of abstraction from context which they attempt is not possible if we are to give an accurate account of explanation. A theory of explanation can be given which takes contextual elements into account through an examination of why-questions and their answers. Such a theory takes seriously a comment made by almost everyone who has written on explanation. It takes seriously the suggestion that explanations are answers to why-questions. Such a theory provides solutions to the problems of irrelevance and asymmetry by describing the relationship of explanatory relevance as determined by the interests of the questioner. It also appears that the context plays a role in the evaluation of explanations. The background knowledge and interests which are elements of the context determine which questions are appropriate and how good the answers to those questions are. ;Carl G. Hempel and Wesley Salmon have both presented theories of scientific explanation which are designed to capture the formal, objective features of scientific explanation. In order to do this they attempt to describe these features by abstracting from the different contexts in which explanations occur. Hempel's account treats explanations as having the logical structure of arguments . He has difficulties formally describing the sentences which should appear in arguments which are explanations. He also runs into problems since, while irrelevant information has no effect on the validity of arguments, it appears to damage the value of an explanation. Finally, there is the problem of asymmetry. Though we can deduce the distance an object falls from Galileo's law of free fall and the time it took the object to fall, it does not appear that we can explain the distance the object fell in this way