Periodic Reporting for period 2 - DIDONE (The Sources of Absolute Music: Mapping Emotions in Eighteenth-Century Italian Opera)
Período documentado: 2020-07-01 hasta 2021-12-31
The belief that “the end of music is to move human affections” (Descartes, Compendium musicae) has been a central issue in European musical thought since Plato. In fact, the history of opera is a permanent exploration of the capacity of action, words, and music to convey emotions. In the 18th century a new type of opera consolidated with the chief concern of expressing the character’s emotions as these changed throughout the drama. The key expressive medium was the aria, where a single, distinct passion was represented, like a concentrated pill of emotional meaning.
The ideal corpus to study how emotions were expressed through music are the 900 operas set to music by 300 composers on the 26 dramas by Pietro Metastasio (1698-1782). It contains a comprehensive catalogue of emotions in music, unique to scrutinise conventions that defined music expression for over a century, paving the way for the emergence of ‘absolute’ instrumental music, autonomous from any other art form.
DIDONE presents an innovative approach to unveil these conventions: the creation of a corpus of 3,000 digitised arias from 200 opera scores based on Metastasio’s 5 most popular dramas—Alessandro nell’Indie, Didone abbandonata, Demofoonte, Artaserse, Adriano in Siria—to be analysed using traditional methods and big data computational analysis. The comparative scrutiny of dozens of musical settings of the same librettos will reveal how composers correlated specific emotions with distinct musical features. The project has 4 main objectives that serve to respond to our main research question, i.e. How are emotions expressed through music?:
1. To identify the techniques used by eighteenth-century composers to express and/or arise human emotions, demonstrating the central role of opera in the construction of the European musical language that still dominates today’s music.
2. To create a new corpus of digital objects, namely 3,000 arias in standard digital format with relevant metadata, in order to enable both conventional and computational analysis, as well as its eventual performance.
3. To develop a cluster of related analytical models that permit the identification of narrative, poetical, and musical patterns related to the expression of emotions in opera.
4. To transfer the results within and beyond Academia.
The ideal corpus to study how emotions were expressed through music are the 900 operas set to music by 300 composers on the 26 dramas by Pietro Metastasio (1698-1782). It contains a comprehensive catalogue of emotions in music, unique to scrutinise conventions that defined music expression for over a century, paving the way for the emergence of ‘absolute’ instrumental music, autonomous from any other art form.
DIDONE presents an innovative approach to unveil these conventions: the creation of a corpus of 3,000 digitised arias from 200 opera scores based on Metastasio’s 5 most popular dramas—Alessandro nell’Indie, Didone abbandonata, Demofoonte, Artaserse, Adriano in Siria—to be analysed using traditional methods and big data computational analysis. The comparative scrutiny of dozens of musical settings of the same librettos will reveal how composers correlated specific emotions with distinct musical features. The project has 4 main objectives that serve to respond to our main research question, i.e. How are emotions expressed through music?:
1. To identify the techniques used by eighteenth-century composers to express and/or arise human emotions, demonstrating the central role of opera in the construction of the European musical language that still dominates today’s music.
2. To create a new corpus of digital objects, namely 3,000 arias in standard digital format with relevant metadata, in order to enable both conventional and computational analysis, as well as its eventual performance.
3. To develop a cluster of related analytical models that permit the identification of narrative, poetical, and musical patterns related to the expression of emotions in opera.
4. To transfer the results within and beyond Academia.
The work in DIDONE is organised in five main work packages:
Compilation and tabulation of data.
We have completed a comprehensive catalogue of all versions and sources for three dramas by Metastasio: Didone abbandonata, Alessandro nell’Indie, and Demofoonte (the latter has been published by Llorens, Rubiales & Usula), amounting to some 225 different operas. A catalogue of Artaserse and Adriano in Siria will follow, and we estimate to reach more than 400 versions in total. We have also catalogued all the arias thereof for which a musical source has survived, making a total of 2,322 so far; and the list increases as we receive new sources. All these data are being uploaded onto a database platform that is already hosted in our server.
Study and annotation of five librettos.
We have thoroughly studied several librettos by Metastasio to analyse their dramaturgy and thus determine the emotion(s) expressed in each aria. For this purpose we have developed a cognate theoretical framework of Descartes’ theory of human passions, which was at the foundations of Metastasio’s thought, to be published by Torrente & Domínguez.
Digital edition of 3,000 arias.
We have defined editorial criteria that work for a large corpus of music created by dozens of composers along 80 years, which comply with both the conventions of critical editing of music and the computational requirements of the analysis. We have also designed an ad-hoc workflow for the engraving, correction, annotation, and export in machine-readable format of the operatic arias in a 10-phase process that involves most members of the team. Around 1,000 arias, 25% of the final goal, have already been engraved, and we foresee to make all of them available in several digital formats by the beginning of 2024.
Conventional and computational analysis of the arias.
We have established an analytical model for the analysis of relevant features from 11 musical parameters. A computational method has been implemented for eight of these: key, scoring, tempo, melody, harmony, metre, form, and texture/density, while the rest are already in progress. We have also started to perform some exploratory statistical analyses on the data to validate the analytical results obtained through traditional, score-based exploration.
Interpretation of results and publication.
We have produced 2 journal articles, 2 book chapters, 5 publications in conference proceedings, and 2 musical editions (in preparation). Furthermore, we have participated in 23 international conferences. We were also organising 2 international conferences, initially planned for 2020 but postponed until 2022 due to the pandemic.
Compilation and tabulation of data.
We have completed a comprehensive catalogue of all versions and sources for three dramas by Metastasio: Didone abbandonata, Alessandro nell’Indie, and Demofoonte (the latter has been published by Llorens, Rubiales & Usula), amounting to some 225 different operas. A catalogue of Artaserse and Adriano in Siria will follow, and we estimate to reach more than 400 versions in total. We have also catalogued all the arias thereof for which a musical source has survived, making a total of 2,322 so far; and the list increases as we receive new sources. All these data are being uploaded onto a database platform that is already hosted in our server.
Study and annotation of five librettos.
We have thoroughly studied several librettos by Metastasio to analyse their dramaturgy and thus determine the emotion(s) expressed in each aria. For this purpose we have developed a cognate theoretical framework of Descartes’ theory of human passions, which was at the foundations of Metastasio’s thought, to be published by Torrente & Domínguez.
Digital edition of 3,000 arias.
We have defined editorial criteria that work for a large corpus of music created by dozens of composers along 80 years, which comply with both the conventions of critical editing of music and the computational requirements of the analysis. We have also designed an ad-hoc workflow for the engraving, correction, annotation, and export in machine-readable format of the operatic arias in a 10-phase process that involves most members of the team. Around 1,000 arias, 25% of the final goal, have already been engraved, and we foresee to make all of them available in several digital formats by the beginning of 2024.
Conventional and computational analysis of the arias.
We have established an analytical model for the analysis of relevant features from 11 musical parameters. A computational method has been implemented for eight of these: key, scoring, tempo, melody, harmony, metre, form, and texture/density, while the rest are already in progress. We have also started to perform some exploratory statistical analyses on the data to validate the analytical results obtained through traditional, score-based exploration.
Interpretation of results and publication.
We have produced 2 journal articles, 2 book chapters, 5 publications in conference proceedings, and 2 musical editions (in preparation). Furthermore, we have participated in 23 international conferences. We were also organising 2 international conferences, initially planned for 2020 but postponed until 2022 due to the pandemic.
We have found that the number of productions based on Metastasio’s librettos Didone abbandonata, Demofoonte, and Alessandro nell’Indie is significantly larger than what is reflected in reference works, and accordingly we have published a catalogue of Demofoonte. Based on such data, we also have a novel catalogue of 2,322 arias, each annotated with significant metadata, and we expect to reach more than 4,000. All data are uploaded onto a database that will be made public by the end of the project.
As regards the emotional labelling of the arias, we apply Descartes’ theory of human passions, which was at the core of Metastasio’s dramatic conception, after a detailed dramaturgical analysis of the librettos. This has implied a step forward in respect of traditional emotional analysis is music, usually based on contemporary listeners’ ratings.
Our editorial criteria guarantee coherence among the corpus of arias being engraved as well as the readability by a computer. By the end of the project we will have a corpus of ca. 3,000 annotated arias, with the relevant numerical and qualitative data extracted from them. We will make public large amounts of data of various kinds (numerical and qualitative musical features, scores, historical and performance-related data, etc.) and in various formats (musicXML, PDF, MuseScore, CSV, etc.), creating the opportunity for further research on carefully extracted and curated sets of data.
In the analysis of the arias —the most complex task in the design of the project—, we strongly focus on the extraction of complex metrics and qualitative data from the digital scores, not only retrieving simple lists, but also performing statistical operations that enable valid and solid comparisons among the various pieces. We have developed various algorithms and are building a Python library that will be launched for general use in due course.
A major outcome of the DIDONE project will be the modern premiere of several musical settings of operas based on Metastasio’s dramas. These include Nitteti by Conforto (Madrid, Auditorio Nacional, May 2022), and Achille in Sciro by Corselli (Madrid, Teatro Real, February 2023). In the long term we aim to design an opera festival that will present in parallel several Metastasian operas in order to let the general public enjoy the same drama dressed in various musical settings by both canonic and neglected composers.
As regards the emotional labelling of the arias, we apply Descartes’ theory of human passions, which was at the core of Metastasio’s dramatic conception, after a detailed dramaturgical analysis of the librettos. This has implied a step forward in respect of traditional emotional analysis is music, usually based on contemporary listeners’ ratings.
Our editorial criteria guarantee coherence among the corpus of arias being engraved as well as the readability by a computer. By the end of the project we will have a corpus of ca. 3,000 annotated arias, with the relevant numerical and qualitative data extracted from them. We will make public large amounts of data of various kinds (numerical and qualitative musical features, scores, historical and performance-related data, etc.) and in various formats (musicXML, PDF, MuseScore, CSV, etc.), creating the opportunity for further research on carefully extracted and curated sets of data.
In the analysis of the arias —the most complex task in the design of the project—, we strongly focus on the extraction of complex metrics and qualitative data from the digital scores, not only retrieving simple lists, but also performing statistical operations that enable valid and solid comparisons among the various pieces. We have developed various algorithms and are building a Python library that will be launched for general use in due course.
A major outcome of the DIDONE project will be the modern premiere of several musical settings of operas based on Metastasio’s dramas. These include Nitteti by Conforto (Madrid, Auditorio Nacional, May 2022), and Achille in Sciro by Corselli (Madrid, Teatro Real, February 2023). In the long term we aim to design an opera festival that will present in parallel several Metastasian operas in order to let the general public enjoy the same drama dressed in various musical settings by both canonic and neglected composers.