Deductive-Nomological versus Statistical Explanation
Carl G. Hempel, James H. Fetzer · 2001
Abstract This essay is concerned with the form and function of explanation in the sense in which it is sought, and often achieved, by empirical science. It does not propose to examine all aspects of scientific explanation; in particular, a closer study of historical explanation falls outside the purview of the present investigation. My main object is to propose, and to elaborate to some extent, a distinction of two basic modes of explanation—and similarly of prediction and retrodiction— which will be called the deductive and the inductive mode. The structure of deductive explanation and prediction conforms to what is now often called the covering-law model: it consists in the deduction of whatever is being explained or predicted from general laws in conjunction with information about particular facts. The logic of this procedure was examined in some earlier articles of mine, and especially in a study carried out in collaboration with P. Oppenheim.