How a concept is formed: Prototype or contingency abstraction?

Abstract

‘Prototype’ models of concept formation have been popular in recent years, in spite of several misgivings. In this paper a new model, called the contingency model, is introduced which meets some of these misgivings, and suggests that the empirical success of prototype models in the past may be explained by the fact of sharing certain computational processes with the new model. The ecological and evolutionary foundations of the new model are explained, and an experiment is reported which compared the predictions of the two models. The contingency model fared better than a representative prototype model.