Connectionism and Conceptual Structure

Abstract

The prototype model of concepts was first proposed formally by Eleanor Rosch and Carolyn Mervis. This model was the most widely accepted view of the structure of concepts for many years. Since the 1980s, however, there have been many criticisms of the model's ability to represent the full range of our cognitive capacities. It now seems clear that a more sophisticated model is required. In this article, the author shows how connectionist models can answer some of the problems of prototype and exemplar theories, but only by transforming many aspects of the original model. Additional criticisms of prototype theory that also apply to connectionist models are then considered. These criticisms have deeper implications for our theories of conceptual structure. Before examining prototype and connectionist models, however, the author first describes what led to the prototype view and the rejection of the earlier view.