TESTING THE PROTOTYPE THEORY OF CONCEPTS

Four experiments were designed to test two predictions of Prototype Theory. The first prediction was that when the defining (necessary) features of a concept are only partially matched by an instance, then characteristic (nonnecessary) features of concepts can affect categorization. The test of this prediction was rendered problematic as successive experiments failed to identify clearly necessary features for a range of concepts. The second hypothesis related to the independence of features in determining similarity. Most versions of Prototype Theory assume a linear combination of feature matches, which would predict that the effect of changing a feature on category membership should be greatest when the probability of categorization is closest to 50% (i.e. at the category border). The results showed that, contrary to this prediction, the effect of changing a feature was greatest when other features were all positive, and so categorization probability was at a maximum. The results support either a logistic combination rule for assessing similarity on the basis of feature match (Medin & Shaffer, 1978), or an exponential generalization function relating similarity to prototype to the sum of matching features (Shepard, 1987). (C) 1995 Academic Press, Inc.