Theory Reduction: A Question of Fact or a Question of Value?

Abstract

Frequently the reduction of scientific theories is treated as a linguistic issue: One theory is said to be reduced to another if the objects referred to by the former are identified with entities in the domain of the latter, and the laws of the former are derived from the laws of the latter (plus whatever connecting principles, correspondence rules, or bridge laws are needed to link the vocabularies of the two theories). The objects of the reduced theory are thus shown to be nothing but objects (or combinations of objects) recognized by the reducing theory, and the concepts of the reduced theory are shown to be theoretically superfluous. The resulting ontological and conceptual economies may reasonably be construed semantically, for they demonstrate that the language of science requires fewer primitive terms than had been previously supposed. Positions taken by Quine and Goodman suggest, however, that evidential and linguistic arguments are in principle too weak to secure theoretical reduction.