Analysing symbolic music with probabilistic grammars
- Submitting institution
-
The University of Lancaster
- Unit of assessment
- 32 - Art and Design: History, Practice and Theory
- Output identifier
- 236021402
- Type
- C - Chapter in book
- DOI
-
10.1007/978-3-319-25931-4_7
- Book title
- Computational music analysis
- Publisher
- Springer
- ISBN
- 9783319259291
- Open access status
- -
- Month of publication
- October
- Year of publication
- 2015
- URL
-
http://link.springer.com/chapter/10.1007/978-3-319-25931-4_7
- Supplementary information
-
-
- Request cross-referral to
- 33 - Music, Drama, Dance, Performing Arts, Film and Screen Studies
- Output has been delayed by COVID-19
- No
- COVID-19 affected output statement
- -
- Forensic science
- No
- Criminology
- No
- Interdisciplinary
- Yes
- Number of additional authors
-
2
- Research group(s)
-
-
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- The chapter was published in the only substantial book on computational music analysis to date. It draws on the authors’ prior work in computational implementation of Schenkerian analysis (Marsden, 2010), and in musical application of probabilistic grammars guided by information theory (Abdallah & Gold, 2014). It contributes to the long competing traditions of music theory which treat musical structure as essentially linear (such as classical theories of harmonic progression and Meyer’s and Narmour’s ‘implication-realization’ models) and those which treat it as hierarchical (most notably Schenker). The chapter recruits methods from a strand of research rooted in machine learning, which attempts to derive the ‘rules’ which govern a language or repertoire directly from the utterances or pieces in that repertoire. This is the first application of these techniques in aiming to resolve the primacy of linear or hierarchical musical structure, through application of advanced computer-science techniques to two simple musical repertoires. The outcome – that the best hierarchical models were somewhat more efficient than the best linear ones on one repertoire but not on the other – did not resolve the question but did prove the method. For wider music theory, the research demonstrates how a computational approach can produce musical insights for high-level music-theoretic questions, bringing with it a degree of objectivity and insulation from preconceptions not typical of music theory, and it provides a model for future application of similar methods to more complex musical repertoires. The objectives and outline of the research arose through discussion between Marsden and Gold. The technical research was conducted by Abdullah under the guidance of Gold. Marsden provided the music-theoretic contextualisation. All authors were involved in preparation of the text and discussion of the conclusions.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -