Abstract
Musical sight-reading requires decoding visual notation into coordinated motor actions, making it an invaluable paradigm for studying the perceptual and motor representations underlying perception-action coupling. Two experiments examined the impact of tonal modulations on sight-reading, having pianists perform melodies varying in the tonal distance of their modulation (no modulation, close modulation, mid modulation, and far modulation). Experiment 1 presented melodies in a random order, whereas Experiment 2 presenting melodies blocked by condition (no modulation melodies first versus modulating melodies first). In both studies, analyses of performance errors revealed increased errors from the initial key to the subsequent key. Additionally, both experiments found gradated tonal distance effects, with far modulations producing the largest difference in error rate between initial and subsequent keys, no modulations producing the least difference, and close and mid modulations falling in between the two. Finally, both experiments observed a spike in error rates, with errors peaking at the transition point from the initial to the subsequent key. Of note is that Experiment 1 showed this (albeit non-significant) pattern for the no modulation melodies, suggesting that pianists developed expectations for key modulation irrespective of its occurrence. Experiment 2 confirmed this hypothesis, demonstrating no change in error rates for no modulation melodies when pianists performed these melodies prior to experiencing the modulating melodies. Together, these results support a perception-action account of piano performance, and suggest intriguing new directions for research on real-time music performance.