author: Joshua B. Mailman

Society for Music Perception and Cognition, Toronto, 2000

ABSTRACT: Previous research on transpositional- and octave-equivalence had as primary variables their participant groups (musicians vs. non-musicians) and stimulus types (tonal vs. atonal). Moreover, atonal melodies were typically tested in contexts involving only short-term memory. The present study attempts to test atonal melodies in a long-term memory paradigm by training listeners to a set of atonal melodies in repeated sessions preceding the experimental trials. This investigation concerns interval -octave-equivalence, as opposed to the pitch-octave-equivalence commonly considered. Interval-octave equivalence (which relates, for instance, an ascending major third to an ascending major tenth and a descending minor sixth) is tested for the learned atonal melodies in tasks involving target/lure discrimination in addition to the tasks involving long-term-memory retrieval.

Stimuli consisted of melodies drawn from the 12-tone and atonal music literature rather than manufactured melodies typical of previous research on atonal melody perception. Each of the training sessions on three consecutive days also involved testing. A melody labeling phase on the first day asked participants to invent melody labels which they would later employ in the identification task of the final session. Familiar condition testing on the final day consisted not only of target/lure discrimination under random octave-scrambling + transposition but also of melody identification under conditions: random octave-scrambling + transposition; contour-preserving octave-scrambling + transposition; and transposition only.

Significantly above chance performance on the tasks lends support to the existence of interval-octave-equivalence in perception of atonal melodies by skilled musical listeners in a familiar condition. The musical instrument of participants also had a significant effect.