The ability of six-year-olds, eight-year-olds, and adults to abstract visual patterns.

ed only a part of the prototype. The ratings of the 8-year-olds and of the adults indicated that they had abstracted more of the prototype. Only six different stimulus patterns were presented in the present study during acquisition, yet none of the groups of Ss tested demonstrated evidence of specific memory. The Ss had to reconstruct the stimulus patterns until they were correct, and thus it seems likely that all the Ss encoded the stimulus patterns. The present results seem incompatible with the specific memory capacities reported for both adults (Shepard 1967) and children (Kagan, Klein, Haith, & Morrison 1973). The stimuli in the Shepard and Kagan et al. studies portrayed This content downloaded from 157.55.39.36 on Sun, 17 Jul 2016 05:47:56 UTC All use subject to http://about.jstor.org/terms 632 Child Development real objects and were noticeably dissimilar. The stimuli employed in the present study were meaningless and very similar. Each stimulus in the studies by Shepard and by Kagan et al. may have been encoded and stored in memory as a different stimulus, whereas all of the stimulus patterns in the present study may have been perceived as inconsequential variations of the same prototypical stimulus pattern. In this respect it would be interesting to determine what variations in stimuli cause them to be classified as different stimuli and what variations are inconsequential, causing them to be integrated into a unified concept. The procedure used to generate the prototype and to predict relative recognition ratings differed from that used by Franks and Bransford. In the present study both procedures produced the same results and seem to be variations of the same method. The procedure used was favored since it seemed to offer insight into the nature of the abstraction processes the Ss had employed. The difference between the two procedures can best be expressed in terms of the competence-performance distinction (Chomsky 1965). Franks and Bransford's procedure seems to lie within the realm of competence theory. It describes the knowledge the Ss have presumably abstracted but bears no direct relationship with the psychological mechanisms involved in the utilization or abstraction of that knowledge. The procedure used in this study seems to lie within the realm of performance theory. It concerns psychological processes presumed to have accounted for the abstraction of the prototype. Both procedures hypothesize that the S should have knowledge of the prototype and of permissible transformations. However, the procedure used in this study hypothesizes how that knowledge was acquired (i.e., by observing the frequency of occurrence of geometric forms whose positions were determined by the admissible transformations).