Computational-Model-Based Analysis of Context Effects on Harmonic Expectancy |
| |
Authors: | Satoshi Morimoto Gerard B. Remijn Yoshitaka Nakajima |
| |
Affiliation: | 1. Department of Design, Graduate School of Design, Kyushu University, Fukuoka, Japan.; 2. Research Center for Applied Perceptual Science, Kyushu University, Fukuoka, Japan.; 3. Department of Human Science, Faculty of Design, Kyushu University, Fukuoka, Japan.; University of Salamanca- Institute for Neuroscience of Castille and Leon and Medical School, SPAIN, |
| |
Abstract: | Expectancy for an upcoming musical chord, harmonic expectancy, is supposedly based on automatic activation of tonal knowledge. Since previous studies implicitly relied on interpretations based on Western music theory, the underlying computational processes involved in harmonic expectancy and how it relates to tonality need further clarification. In particular, short chord sequences which cannot lead to unique keys are difficult to interpret in music theory. In this study, we examined effects of preceding chords on harmonic expectancy from a computational perspective, using stochastic modeling. We conducted a behavioral experiment, in which participants listened to short chord sequences and evaluated the subjective relatedness of the last chord to the preceding ones. Based on these judgments, we built stochastic models of the computational process underlying harmonic expectancy. Following this, we compared the explanatory power of the models. Our results imply that, even when listening to short chord sequences, internally constructed and updated tonal assumptions determine the expectancy of the upcoming chord. |
| |
Keywords: | |
|
|