Discovering the line of fifths in a large historical corpus


An increasing number of corpus studies relies on pitch-class distributions in order to infer characteristics of musical pieces under a historical perspective (Albrecht & Shanahan, 2013; Albrecht & Huron, 2014; Quinn & White, 2017; Weiß, Mauch, & Dixon, 2018; Yust, 2019; Harasim, Moss, Ramirez, & Rohrmeier, 2021).

This contribution shows that the line of fifths (LOF; Temperley, 2000) is the fundamental underlying tonal space in a large historical corpus (ca. 1360-1940) of Western classical pieces in MusicXML format. Modeling the pieces’ pitch-class distributions as vectors in a high-dimensional simplicial space and visualizing them via Principal Component Analysis reveals that the distance to the center of the LOF as well as the distinction between the natural (F, C, G, D, A, E, B) and the altered tonal pitch-classes (e.g., Abb, Db, F#, C##) are the most important factors for the dispersion of the data. These findings are robust with respect to different dimensionality reduction methods. Moreover, we introduce the concept of pitch-class coevolution and demonstrate that the LOF also underlies striking changes in the usage of pitch-classes between different historical periods.

Any empirical study is based on certain implicit or explicit modeling assumptions, some of which are given by the encoding of a corpus, e.g. whether enharmonic equivalence is assumed (e.g. MIDI-encoding) or not (e.g. MusicXML encoding). Relying on pitch-class distributions without assuming enharmonic equivalence, our findings emphasize the structural importance of the LOF for the organization of the pitch-class content of tonal music across a large historical timespan."

Mar 7, 2021 6:40 PM — 7:00 PM
The Ohio State University [online]
Fabian C. Moss
Fabian C. Moss
Research Fellow in Cultural Analytics

Fabian C. Moss is a Research Fellow in Cultural Analytics at University of Amsterdam (UvA). He was born in Cologne, Germany, and studied Mathematics and Educational Studies at University of Cologne, and Music Education (Major Piano) and Musicology at Hochschule für Musik und Tanz, Köln. He obtained is PhD in Digital Humanities from École Polytechnique Fédérale de Lausanne (EPFL). Working with large symbolic datasets of musical scores and harmonic annotations, he is primarily interested in Computational Music Analysis, Music Theory, Music Cognition, and their mutual relationship.