Le 30 juin 2017
de 16h30 à 17h00
Le Patio (université de Strasbourg)
22 rue René Descartes, 67000 Strasbourg
amphithéâtre 6
Séance précomposée - Formal, Theoretical, and Computational Models in Popular Music Analysis
Pré-acte / Acte
Auteur : John Covach
Various datasets have been used to create corpus studies of popular music. Such sets are typically analyzed to reveal characteristics of the music that might not otherwise be possible without powerful computer processing. This approach contrasts strikingly with the ways scholars have traditionally come to generalizations about musical structure, style, and development, which build up a corpus on the basis of preferences that some might consider subjective, skewed, or incomplete. Large datasets seem to promise greater objectivity, though some scholars question if they produce results that provide useful or even valid insight. Covach 2015 argues that such studies may lead to misleading or false conclusions.
This paper argues that meaningful results are most often obtained when the songs included in the dataset (digital or analog) are determined by characteristics that scholars find significant and not, say, by the Billboard charts, frequency of streaming in the recent past, Rolling Stone lists of top songs, etc. It is clear that traditional historical and analytical lenses are created (sometimes tacitly) by the music-historical and music-aesthetic preferences of scholars, and that this may create blind spots. But without such sorting applied to the music studied, the answers any given dataset provides may not address questions scholars would find significant or meaningful, even in the most ecumenical sense. Each approach brings with it powerful tools; it may be that a blend of traditional and digital approaches will ultimately produce the most useful results.
