Exemplar-based retrieval enables an illusion of category simplicity: Implications for theories of category coherency
Randall Kenneth Jamieson, Lee R. Brooks · 2007
We apply an exemplar model of memory, Minerva2 (Hintzman, 1986), to examine peoples' false belief that family resemblance categories have simple defining features (e.g., that all tables have four legs). In the model, participants encode training exemplars. When asked whether a studied category had a simple defining feature, representations of the queried categories are retrieved and, then, are used to infer category structure. We show that parallel retrieval from a store of studied exemplars produces systematic distortions in category representation and that these distortions enable an illusion of category simplicity.