A Multiple Definitions Model of Classification Into Fuzzy Categories

This paper describes a new hypothesis, referred to as the multiple definitions model, concerning the mental representation of fuzzy concepts. The basic claim of the model is that such concepts are represented as a set of multiple definitions, where each definition is exact. Fuzziness results from the fact that using such concepts requires sampling multiple such exact definitions of the concept. The model was applied to concepts that can be defined as a range of values over a single dimension (such as middle-age), and tested using conjunctions and disjunctions of middle-age (e.g., "A person is middle-aged at both 50 and 63."). The model predicts that, controlling for the truths of individual ages, the truths of conjunctions involving ages that are close together will be judged higher than the truths of conjunctions involving ages farther apart, and that the opposite effect will occur for disjunctions (the distance effect). The results of two experiments confirmed this prediction. However, both experiments also found that conjunctions were judged truer than the less true of their component ages, and that disjunctions were judged less true than the truer of their component ages. The model does not predict this "minimax" effect. One possible explanation of the minimax effect was tested; another modeled. The overall conclusion is that the multiple definitions model is a viable contender to explain the distance effect. The minimax effect, however, is still in need of a satisfactory explanation.