Distant Listening

The Development of Harmony over Three Centuries (1700–2000)

  • Funded by the Swiss National Science Foundation (SNSF)
  • Duration: early 2019 — early 2023 (four years)
  • Principal investigator: Prof. Dr. Martin Rohrmeier

The Digital Humanities are a vibrant field on the rise, witnessing a steady growth of research initiatives. One of the central potentials of computational approaches for Digital Musicology is the power to analyze amounts of musical sources that have been entirely unachievable for traditional methods. However, this also brings with it a core challenge in terms of bridging approaches from different disciplines, both on the level of theory and methodology. Hence, this proposal makes a step towards integrating music theory, musicology, and computational music modelling in an overarching project.

This project focuses on the analysis and historical development of harmony between ca. 1700 and 2000. Using corpus-based methodologies, we aim to model harmonic structures of common-practice tonality, extended tonality of the 19th century, and new forms of harmonic usage in Pop/Rock/Jazz. The proposed research plan involves three main pillars:

First, it aims to create and publish a large ground-truth corpus of instrumental and vocal music with expert harmonic annotations along with tools enabling their access, analysis, and visualization.

Second, we will work on developing a robust computational model of harmonic analysis and inference based on generative Bayesian modeling, aiming to advance the state of the art which is still not sufficiently high to be of practical use for subsequent music-theoretical purposes.

Third, we will analyze the annotated ground-truth corpus as well as additional musical sources, seeking to advance the understanding of how harmonic principles developed over the entire timespan by approaching our dataset from a bird’s eye perspective (“distant listening”). In particular, we will combine information-theoretical measures and inferred music-theoretical structures with music-theoretically informed analysis to statistically examine, visualize, and interpret large-scale trends and developments in the light of existing musicological literature, seeking to achieve a cross-disciplinary, cross-methodological enrichment of our approach.