ABSTRACT
Distinguishing features (and their complement, shared features) have played a role in accounting for a number of primary empirical phenomena regarding normal subjects’ performance in semantic tasks, such as typicality judgements (Rosch & Mervis, 1975), similarity judgements (Tversky’s contrast rule, 1977), and semantic verification (Smith, Shoben, & Rips, 1974). Distinguishing features correspond to the aspects of people’s knowledge that discriminate among or between similar concepts. For example, a feature such as <moos> is highly distinguishing, whereas <eats> is not. Distinguishing features may be crucial to understanding category-specific deficits because most tasks used to establish the existence of these deficits, such as picture naming, word to picture matching, and definitions to words, require distinguishing a specific concept from among similar ones. That is, to properly identify something as a zebra rather than a horse requires preserved knowledge of the features that distinguish them. Therefore, given the probabilistic nature of brain damage, the likelihood that a patient can discriminate a concept from among similar ones is related to the number or proportion of a concept’s features that are distinguishing.
