EXPLANATORY COHERENCE IN SOCIAL EXPLANATIONS - A PARALLEL DISTRIBUTED-PROCESSING ACCOUNT

This article studies the impact of explanatory coherence on the evaluation of explanations. Tested were 4 principles of R Thagard's (1989) model for evaluating the coherence of explanations. Study 1 showed that Ss preferred explanations that accounted for more data (breadth) and that were simpler (simplicity or parsimony). Study 2 demonstrated that Ss thought an explanation was stronger when it could, in turn, be explained. Study 3 showed that the evaluation of explanations is comparative, affected by the availability of good alternatives. These results were then successfully simulated using Thagard's connectionist implementation of his model of explanatory coherence. The data and the simulation, taken together, strongly support the model. Two issues are then discussed: (a) the role of explanatory coherence in social explanation and (b) the relevance of parallel constraint satisfaction processes to social reasoning.