Counterfactual Reduction

Abstract

The view that explanations are arguments has been frequently challenged in the current philosophy of science. For instance, the assumption that in scientific changes earlier theories, or laws, are deductively explainable by, thus reducible to, their successors has appeared controversial and — in view of actual scientific developments — very open to criticism. Various arguments have been presented against this optimistic conception of scientific change, but in this paper I shall only consider the following type of objections. It has been argued that an old theory is often replaced by a new one, which would mean that the two theories are incompatible or that the old one is false if the new one is true, or that they are incommensurable. How could the new theory in such a case deductively explain its predecessor? Is there anything to be explained? If there exists an inference from the new theory to the old one and the former is true and the latter false, the inference presupposes connecting assumptions which are at variance with facts. But if the theories are incompatible, there may not even exist any explanatory argument — except in some approximate or limiting sense. If, on the other hand, they are incommensurable, but there is a connecting inference, it is only formal, that is, its conclusion is not really a law of the old theory — even though it is syntactically identical with one — since the scientific terms contained in the conclusion still represent concepts of the new theory and they are not identical with those of the old.