In search of abstraction: The varying abstraction model of categorization. (original) (raw)

AI-generated Abstract

A classic question in cognitive psychology concerns what is stored as a consequence of learning a category, and hence what information people rely on when they make a categorization decision. It is generally assumed that learning a category involves the generation of a category representation and that assigning a novel object to a category involves the comparison of the object to that category representation. However, one of the most fundamental and unresolved issues in the categorization literature concerns the exact nature of this category representation. Although the debate on category learning and category representation has a very long history, in the past few decades it has centered on the question of whether people represent a category in terms of an abstracted summary or a set of specific examples. Early work argued for the prototype view of category learning. Under this view, on the basis of experience with the category examples, people abstract out the central tendency of a category. In other words, a category representation consists of a summary of all of the examples of the category, called the prototype (see, e.g., Posner & Keele, 1968; Reed, 1972; Smith & Minda, 2002). The initial success of this view has gradually declined in favor of the exemplar view, in which experience with examples of a category does not lead to the development of an abstracted prototype; instead, people simply store all of the examples they encounter. In other words, a category representation consists of all of the individual examples of the category, called the exemplars (Brooks, 1978; Estes, 1986; Medin & Schaffer, 1978; Nosofsky, 1986). The shift from the prototype to the exemplar view was motivated by several arguments, including evidence that exemplar models yield fits superior to those of prototype models in various experimental settings.