Perception of Thematic Structure in Music Over Short and Long Timescales Available to Purchase

Thematic structure is valued highly in music theory and in many models of music cognition. However, strong empirical evidence supporting its perception is scarce. To provide a basis for experimental research, Hall and Pearce (2021) developed a probabilistic computational model of the cognitive processes underlying perception of thematic structure. The model hypothezises that thematic structures become perceptible due to the statistical regularities formed by repetition and variation of material. Two experiments were conducted to test this model. Experiment 1 focused on small timescales. Forty participants rated whether pairs of themes and repetitions (N = 100) came from the same composition, varying across model-based measures of variation and stylistic unpredictability. Significant effects on ratings were found for both measures, across pitch and rhythm representations. Experiment 2 focused on large-scale structures. Forty participants heard 40 two-minute-long melodies, varying in internal unpredictability, repetition, variation, and stylistic unpredictability. Participants identified whether moments in the stimuli were repetitions and rated each melody on its structural unity. Both tasks provided significant evidence for the effects of internal unpredictability on the perception of thematic structure, both for pitch and rhythm representations. These findings suggest perception of thematic structure depends on intra-opus unpredictability, as predicted by the model.