|| Checking for direct PDF access through Ovid
Despite often being conceived as a spontaneous and creative mode of performance, improvisation is predicated on prior knowledge. What characterizes this knowledge, and how is it represented or recalled differently as compared with other modes of music making? Asking about knowledge and trying to distinguish improvisation as a distinct performance process can locate research questions within the theoretical frameworks of cognitive science, but it is not clear how to make such questions experimentally accessible. Differences arising from music–analytical versus cognitive conceptions of improvisation are explored to provide a theoretical framework compatible with experimentation. Experimental research could concern itself with how the embodied interface between performer and instrument, when manipulated, invokes different cognitive processes of music making, helping to describe the cognitive characteristics of various modes of music performance. Here, an experiment is reported that synthesizes previous techniques used to analyze improvisations with experimental strategies from the neuroscientific literature aimed at differentiating performance processes within a given improviser. Jazz pianists improvised monophonically over backing tracks in a familiar and unfamiliar key as well as with their right and left hands. Among other findings, in some of the less familiar performance situations, participants relied more on diatonic pitches and produced more predictable improvisations as measured by entropy and conditional entropy. The nature of the different underlying processes and knowledge at play under these different conditions is explored, and future research directions to better describe them are identified, including incorporating motor theories of perception.