In a series of experiments and reanalyses of previous research, we tested the hypothesis that categories that are primarily represented by extrinsic features (i.e., those that are relations between two or more entities) would yield more graded structures than would categories primarily represented by intrinsic features (i.e., those features true of an item considered in isolation). These predictions were confirmed, Extrinsically represented categories showed (1) less agreement across subjects on membership judgments, (2)more graded membership in a membership judgment task, and (3) smaller differences between gradients of typicality and of membership judgments. Over the last 15 years, research into natural language 1979), and some "exemplar" theories (e.g., Medin & category representation has been strongly influenced by Schaffer, 1978). All of these models contrasted with the the work of Rosch and her colleagues (e.g., Rosch, 1973, earlier "classical" view that categories could be repre-1975; Rosch & Mervis, 1975). The essential finding in sented by a set of necessary and sufficient features that this research was that many natural language categories described all and only category members (see Smith & show graded membership functions. Some members are Medin, 1981). reliably rated as better examples of a category than are Theories designed to explain category "fuzziness," others (Rips, Shoben, & Smith, 1973; Rosch, 1973), and however, have both conceptual and empirical problems. some members are produced and verified as category Conceptually, probabilistic and prototype theories require members more often and more quickly than are others more complex representations than does the classical the-(Battig & Montague, 1969; Glass, Holyoak, & O'Dell, ory. Prototype theories require a prototype and some rule 1974; Rips et al., 1973; Rosch, 1973; Wilkins, 1971). that permits items that are discrepant from the prototype In addition, such categories include "borderline" mem-to be category members. Probabilistic models require a bers, that is, items over which subjects disagree with each set of features, a set of weightings reflecting the imporother regarding category membership, and over which tance of each feature to the category, and some "weighted they are inconsistent from day to day in assigning mem-feature combination rule" to determine category membership (McCloskey & Glucksberg, 1978). bership. These demonstrations of graded membership generated Empirically, these models have had difficulty explainmany theories proposing that category representations are ing natural language category phenomena other than fuzfuzzy. In other words, no membership criteria exist that ziness. Various investigators have demonstrated that cateclearly discriminate all members from all nonmembers. gory structure is influenced by context (e.g., Anderson Such theories included Rosch's own "prototype" ap

In this study, the authors examined the relations between 3 psychological variables-fatalism, self-confidence, and intellectual resources-and the subsequent development of illness and disability 20 years later in an adult sample. Results indicated that greater fatalism, assessed in 1974, predicted greater difficulty in everyday cognitive tasks as well as illness in 1994. Higher self-confidence in 1974 was associated with lesser degrees of cognitive and fine motor difficulty in 1994. Greater intellectual resources in 1974 (a combination of intellectual flexibility and education) predicted less cognitive and gross motor difficulty as well as lesser degrees of illness in 1994. Some of these relations were stronger for older than for middle-aged individuals. Results are discussed in the context of models of the disablement process.

Using randomized stimulus onset asynchrony (SOAs), the authors traced the time course of Stroop interference and facilitation in normal participants and participants with schizophrenia. Unlike earlier findings using blocked SOAs, singular peaks in interference, facilitation, or both occurred at particular SOAs. The peaks of normal participants and participants with schizophrenia differed. Findings are congruent with a model of Stroop performance that posits individual differences in processing speeds of target and nontarget stimulus dimensions, coupled with critical points in response selection. Participants with schizophrenia also showed more overall interference than normal control participants. A second experiment that added a temporal gap between the distractor word and target color obliterated Stroop effects only for individuals with schizophrenia. These findings provide a new empirical basis for models of Stroop effects. They are also consistent with hypotheses about the importance of the prefrontal cortex for working memory and prefrontal dysfunction in schizophrenia.

No abstract