THE CHILD IN ALL OF US: ADULT ACCESS TO OVERALL SIMILARITY IN THE CLASSIFICATION OF SEPARABLE STIMULI

A common view is that adults\u27 access to the overall similarity relations among separable stimuli is secondary and difficult. The paradox is that children have easy access to those similarity relations. Accordingly, six experiments test the hypothesis that adults have direct, nonderived access to overall similarity in the classification of separable stimuli (size and brightness). In two experiments, classification by an overall similarity rule shows a smaller performance decrement over a five-fold increase in speed than classification by a dimensional rule. In addition, high speed similarity classification occurs more accurately than high speed dimensional classification. These results suggest that adult access to overall similarity cannot be a subsequent, secondary process or a recombination of dimensional information. Four experiments test the hypothesis that untrained subjects have spontaneous access to similarity relations and that this access will appear when normal adult classification is disrupted. Accordingly, spontaneous adult classifications are obtained under conditions of speed, a concurrent task, and with an instructional set that emphasized impressionistic responding. All three manipulations are successful in eliciting overall similarity classification. However, in a second classification task, speed does not increase overall similarity responding and this exception is discussed. Finally, adult individual differences in the spontaneous use of similarity in classification are associated with differences in classification tempo. A similarity mode of classification seems to be a spontaneously available part of the adult\u27s cognitive repertory. Based on the results, the asymmetry of the integrality/separability distinction is discussed. Separable dimensions apparently offer two easily available classification modes; integral dimensions only one. A response to overall similarity relations may be primary and primitive for both. The results suggest an identical developmental asymmetry, with adults having two easily available classification responses to separable dimensions and children having only effortful access to dimensional information. The analogy between adult similarity classification under stress and child-like classification is discussed. Finally, we discuss the possibility of an impressionistic, implicit mode of classification in adults, one very different from their unconstrained bias to be explicit, strategic and analytical in their classification