Patterns in Music: How Linguistic Corpus Analysis Tools Can Be Used to Illuminate Central Aspects of Jazz Improvisation

Abstract:It is well-known that improvisations by artist-level jazz improvisers contain repeated melodic and rhythmic patterns. In previous research, we argued that these patterns reflect an ongoing internal process in which the insertion of linked auditory and motor patterns underpin improvisation. This view was supported by an analysis of 48 solos by jazz great Charlie Parker. Here we provide an updated and expanded analysis that incorporates the principles of Open Science where both data set and analysis paradigm are immediately available to the reader. Specifically, three main limitations of our previous work are addressed by 1) using a much larger dataset—the Weimar Jazz Database (WJD), which contains 456 solos by 78 jazz artists, 2) providing a formatted version of the data set, and 3) demonstrating a free tool that allows the reader to nearly instantly verify our analysis using software originally designed for corpus linguistic analysis. This tool can also be used in jazz theory and improvisation classes, allowing students to investigate any type of melodic, rhythmic, or chordal patterns. We provide and validate several measures that demonstrate the ubiquitous use of patterns by different jazz artists and show the reader a number of ways to explore pattern distribution and contexts. Our analyses provide further support for the central role of patterns in real-time music creation as originally suggested by Pressing (1988).