The Relevance of Unrealistic Models for Empirical Political Science

This article focuses on the possible relevance of unrealistic models for empirical political science. The author rejects both the covering law program, according to which there is no place for unrealistic assumptions, as well as the Friedman-instrumentalist position, which argues that the validity of a model's assumptions is irrelevant. Instead, a third view is being offered, according to which scholarly work based on unrealistic models can be a highly legitimate activity, while at the same time decriptively accurate assumptions is retained as an important ideal for construction of theoretical models.