Theoretical Explanation and Unification

Abstract

Reducing explanatory power to unifying power is a strategy for solving some of the long-standing problems associated with explanatory relevance. These problems stem from the fact that a derivation of a description of a phenomenon from accepted laws and theory components is not necessarily an explanation of that phenomenon. The laws and theory components must also be relevant to explaining the phenomenon. But what other considerations determine the explanatory relevance of laws and theory components? Friedman and Kitcher, for example, claim that the explanatory character of laws and other theory components is due to their unifying power, which can be defined, they claim, in formal, logical terms. An account of explanatory relevance in terms of unifying power is intended to avoid the epistemological and metaphysical issues that confront causal accounts of explanatory relevance and the context-dependency of pragmatic accounts. As I shall argue below, different types of theory components may be used to unify a body of knowledge in different ways. These facts, I claim, seriously undermine Kitcher's program for defining explanatory relevance in terms of unifying power, and also of characterizing a single property that