[On the relation between prototype abstraction and exemplar retention in the memory of melodies]

This study examined the relation between prototype abstraction and exemplar retention by using memory of melodies. Acquisition lists of melodies were constructed by applying a set of transformation rules to a fixed prototype melody (cf. Welker, 1982). Immediately after the presentation of acquisition list, recognition task was performed. The recognition rating for prototype was used as the indicator of the prototype abstraction. The discriminability of old melodies from new melodies (d') was used as the indicator of exemplar retention. In Exp. 1 with 30 undergraduates as subjects, the transformational distance of acquisition list was varied and it was demonstrated that as the transformational distance decreased, prototype abstraction was facilitated but exemplar retention was not changed. In Exp. 2 with 90 undergraduates as subjects, the number of presentation of each melody and the kinds of melodies in acquisition list were varied. Results demonstrated that as the number of presentation of each melody increased and also as the kinds of melodies decreased, the exemplar retention improved but prototype abstraction was not changed. It was argued that the systems underlying the prototype abstraction and exemplar retention were functionally independent of each other.