|
This study compared: (a) the effects of explicit training in typical attributes of Classical and Romantic styles with training that required subjects to induce style concepts from musical examples on abstraction and categorization; (b) the relevance of four musical dimensions (beat, phrases, texture and chord progression) for subjects' categorizations; and (c) the congruence of categorizations with predictions of exemplar- or prototype-theories of concept formation. Stimuli were piano excerpts varying systematically along the four dimensions. Six high school general music classes were randomly assigned to "prototype-plus" and "examples-only" conditions. The former subjects received word-descriptors about typical attributes of each style with selected training stimuli; the latter received no such information. Receiving corrective feedback, subjects classified training exemplars on 16 trials, with errors tracked for each trial and two posttests. Subjects' strategy reports and justifications helped interpret the error data. Directly presenting subjects with prototype information about ill-defined musical styles was more beneficial than induced category information. Only texture was reliably abstracted by the examples-only subjects. Texture and chord progression affected categorization more than beat and phrase attributes. Subjects in both conditions apparently relied on independently abstracted cues rather than holistic exemplar similarity.
_____________
|
| 7. desember1996 |