Rhythm has received scant attention from modelers and experimenters in working memory, and the mechanisms by which we retain rhythms over the short-term are poorly understood. This is a substantial gap in our knowledge of how music may be learned. Furthermore, recent theories of musical origins and psychological research into differential processing of voice sounds have converged to point towards the possibility that “rhythm” may be derived from at least two processes; a vocally-derived component and a manual, drumming component. This background provokes questions into how we remember rhythm, and whether vocal and instrumental rhythms are treated differently by working memory.
This thesis reviews the foundation of these questions and then presents two behavioural experiments designed to investigate working memory for voice and non-voice rhythm. Participants performed two recognition memory experiments in which they were asked to judge the rhythm in pairs of recordings as either “same” or “different.” These recordings were drawn from field recordings of the Dyirbal of Australia, and on each trial the participants were presented with both voice and clapstick rhythm but were instructed to respond solely to one or the other. In the first experiment, musicians and non-musicians were tested at two interstimulus intervals, short (500ms) and long (12500ms). Musician performance was found to be superior in the clapstick task, but not in the voice task. Furthermore, both musicians and non-musicians performed worse in the voice task at the long duration compared to the short duration, while performance for the clapstick rhythms was not significantly decremented by the increased duration. Reaction time analysis showed an effect of duration, condition (different or same), and a small interaction between musical training and task (clapstick or voice). In the second experiment, musicians were tested with the same stimuli at the long duration but with additional interference tasks designed to suppress the temporal encoding. In one condition participants sub-vocally repeated the word “the” at a steady tempo, and in the other condition they tapped the finger of their dominant hand at a steady tempo, each while performing the recognition memory tasks as before. Finger-tapping caused a significant increase in accuracy for both voice and clapstick tasks, while sub-vocal articulation did not significantly affect performance for either.
The results of these two experiments are interpreted as providing support for differential processing of voice and non-voice rhythms, as the voice rhythms suffered substantial performance loss at the longer duration and clapstick did not. Further, the lack of interference caused by the sub-vocal articulation task indicates that the articulatory loop is not providing the rehearsal process for recognition memory for rhythm, and supports a component of working memory that is invested with rhythm and/or temporal processing.