A music theory ontology

Abstract

Many existing music ontologies have focused on expressing metadata related to performances or recordings, aiding with recommendations of songs or artists, and studying the psychological affects of music. These music ontologies provide a foundation for describing many practical aspects related to music. We believe further primitives are needed in order to represent written music and provide a foundation for performing analysis of music. We are motivated by questions related to analyzing music that might inform composers or musicians. Informational elements may include possible underlying chords from a set of notes, as well as summaries of key signatures or scales used in a given song. In order to leverage Semantic Web technologies to answer such questions, we present our Music Theory Ontology that expands on existing work by including theoretical concepts that were absent from previous music ontologies. We further describe a methodology for using the ontology to infer new knowledge. We demonstrate this capability by inferring the notes in various scales and chords, and evaluate the ontology in terms of competency question answering.