We say birds are singing, but actually they are just wooing. The origin of music is considered to be a primitive communication method, and music may possess somewhat a similar structure to language. With the help of hippocampus in our brain, which works as a push-down stack, we can memorize the beginning part of a sentence and can expect a corresponding ending. This mechanism is represented by context-free grammar (CFG) which generates a hierarchical tree structure of a sentence, while other animals can speak only by regular grammar (RG). The same machinery works when we listen to music. When a music piece begins with a tonic chord, we expect a cadence which is typically a dominant followed by a tonic; this long distance dependency implies that a stable music structure should have a hierarchical tree in which another tree may be embedded. Thus far, we have implemented a music analyzer ATTA based on Generative Theory of Tonal Music (GTTM) as a more polished theory than Schenkerian method, which embodies such tree structure. we show the current stage of the analyzer, discuss the efficacy of the theory, and show a more sophisticated grammar than simple CFG for music analysis.
ABOUT SATOSHI TOJO:
Satoshi Tojo received degrees of Bachelor of Engineering, Master of Engineering, and Doctor of Engineering from University of Tokyo, Japan. He joined Mitsubishi Research Institute, Inc., Tokyo, Japan from 1983 to 1995, and JAIST (Japan Advanced Institute of Science and Technology), Ishikawa, Japan, as associate professor from 1995 to 2000 and professor from 2000.
His research interests have been in logic in artificial intelligence, including knowledge representation of artificial agent and formal semantics of natural language. Currently, he is interested in linguistic model of music.