A fuzzy measure for explanatory coherence

In a series of articles, Paul Thagard has developed a connectionist's model for the evaluation of explanatory coherence for competing systems of hypot heses. He has successfully applied it to various examples from the history of science and common language reasoning. However, I will argue that his fo rmalism does not adequately represent explanatory relations between more th an two propositions. In this paper, I develop a generalization of Thagard's approach. It is not subject to the connectionist paradigm of neural nets, but is based on fuzzy logic: Explanatory coherence increases with the fuzzy truth value of the conjunction of explanans and explanandum and decreases with the value of the conjunction of explanans and the negation of the expl anandum.