Abstract

Two experiments were performed to investigate the effects of category size and item discriminability on the rule extraction and storage of exemplar information in learning of rule-described categories. In both experiments MDS-solutions based on the MDS-choice model were obtained from the data of item identification task, and categorization data were obtained using the typical learning-transfer phase paradigm. In Experiment 1, category size was increased up to 16 items per category. Compared to the results of previous research, it seems that tendency to extract rules increases as the category size becomes large. In Experiment 2, discrimination cues to increase the discriminability among the items were given to the subjects. The results showed that a tendency to extract and use rules in learning categories and categorizing new items increase in the case of high discriminability among the items. It was suggested that many factors, conditions, or constraints can affect learning various kinds of categories including rule-described categories.