Periodic Reporting for period 1 - MIME (Musical Imitation in Medieval Europe (1110-1300))
Reporting period: 2023-09-01 to 2025-08-31
Summary of the context and overall objectives of the project
In the late Middle Ages, especially in the territories that we now call France, lyric poetry was made to be sung, and poets were most often music composers and singers. Initially restricted to religious environments, musical scores started to appear from the 13th century in secular sources, and medieval songbooks have transmitted thousands of melodies accompanying lyrics in European languages.
The rise and development of vernacular musical repertories, mostly linked to courtly cultural environments, have been profoundly shaped by processes of melodic imitation and reuse. On the one hand, the absence of copyright and a less pronounced sense for musical authorship set no limits to the legitimate reuse, recomposition, and performance of existing music, and on the other, musical imitation became a means for poets-performers to establish cultural and literary connections within a European lyric tradition. Studying these connections from a musicological and literary perspective provides us with a better understanding of cultural relationships and processes of exchange that would otherwise be completely lost to us.
The connections between secular, courtly, vernacular song and Latin sacred music composed for liturgical purposes must have been prolific, as suggested by scholars such as Jacques Challey and Hans Spanke. However, these connections are extremely difficult to identify because of the large amount of extant vernacular music and the almost countless examples of liturgical sources. What is more, melodies transmitted orally, transcribed and copied by hand, underwent significant modifications, making melodic comparisons even more complex.
The first step in understanding the system of cultural connections created through music in the late Middle Ages is to equip ourselves with the right tools. In this respect, adopting a digital approach appears the most compelling way forward, as it enables large-scale comparisons and analysis. The main objectives of the project are:
1) Enlarge the dataset of encoded melodies of the vernacular corpus of medieval French lyrics.
2) Integrate other existing musical datasets, especially those of liturgical music, converting them to a common encoding standard.
3) Create a set of analytic tools to compare a large number of melodies efficiently and flexibly.
4) Define and describe techniques of music imitation and reuse.
5) Analyse the first results and discuss the cultural connections both within and between secular and religious corpora.
The project serves as a proof of concept for a larger collaborative initiative. It demonstrates that computational strategies can advance our knowledge of cultural exchanges through music, while also providing the technical tools and a sustainable workflow (including encoding and analysis) to support future research. It demonstrates the necessity of an integrated approach, bringing together musicology, literature, philology, and digital humanities, to investigate artworks that develop across multiple media.
The rise and development of vernacular musical repertories, mostly linked to courtly cultural environments, have been profoundly shaped by processes of melodic imitation and reuse. On the one hand, the absence of copyright and a less pronounced sense for musical authorship set no limits to the legitimate reuse, recomposition, and performance of existing music, and on the other, musical imitation became a means for poets-performers to establish cultural and literary connections within a European lyric tradition. Studying these connections from a musicological and literary perspective provides us with a better understanding of cultural relationships and processes of exchange that would otherwise be completely lost to us.
The connections between secular, courtly, vernacular song and Latin sacred music composed for liturgical purposes must have been prolific, as suggested by scholars such as Jacques Challey and Hans Spanke. However, these connections are extremely difficult to identify because of the large amount of extant vernacular music and the almost countless examples of liturgical sources. What is more, melodies transmitted orally, transcribed and copied by hand, underwent significant modifications, making melodic comparisons even more complex.
The first step in understanding the system of cultural connections created through music in the late Middle Ages is to equip ourselves with the right tools. In this respect, adopting a digital approach appears the most compelling way forward, as it enables large-scale comparisons and analysis. The main objectives of the project are:
1) Enlarge the dataset of encoded melodies of the vernacular corpus of medieval French lyrics.
2) Integrate other existing musical datasets, especially those of liturgical music, converting them to a common encoding standard.
3) Create a set of analytic tools to compare a large number of melodies efficiently and flexibly.
4) Define and describe techniques of music imitation and reuse.
5) Analyse the first results and discuss the cultural connections both within and between secular and religious corpora.
The project serves as a proof of concept for a larger collaborative initiative. It demonstrates that computational strategies can advance our knowledge of cultural exchanges through music, while also providing the technical tools and a sustainable workflow (including encoding and analysis) to support future research. It demonstrates the necessity of an integrated approach, bringing together musicology, literature, philology, and digital humanities, to investigate artworks that develop across multiple media.
Work performed from the beginning of the project to the end of the period covered by the report and main results achieved so far
- The project significantly enlarged the existing corpus of encoded French melodies in the platform MedMel (https://medmel.seai.uniroma1.it(opens in new window)) completing the digital editions of MS Paris, BnF, fr. 20050 (~100 new transcriptions) and Città del Vaticano, BAV, Reg. 1490 (299 transcriptions). Furthermore, it has integrated into the MedMel dataset the entire corpus of the Cantigas de Santa Maria, 405 melodies edited by Andrew Casson, and over 60.000 liturgical pieces from the Cantus Database.
- The MedMel platform underwent major updates and improvements during the duration of the project, providing an optimal infrastructure for encoding and publishing the music of medieval vernacular lyrics.
- In light of the advancement of digital musicology in the domain of Optical Music Recognition (OMR) of medieval sources, the project has developed a tool to integrate MedMel as the endpoint of the Rodan + Neon.js OMR workflow. Rodan is a software developed at McGill University for semi-automatic music information retrieval from medieval manuscripts; the results can be adjusted with Neon.js to produce an XML/MEI file, which can now be imported into MedMel.
- Three digital searching tools have been developed for musical analysis and comparison. The first, the “Search Sequence” (https://medmel.seai.uniroma1.it/searchMelody.php(opens in new window)) allows the user to enter a melodic sequence and search for matches in the MedMel database of vernacular song; the second, the “Search Song Tool” (https://medmel.seai.uniroma1.it/searchSong.html(opens in new window)) allows to select a whole song and find similarities in the MedMel database and in the Cantus Database; the third, the “Many2Many Search Tool” replicates the logic of the Search Song Tool comparing all melodies with every other melody in a dataset, and it is developed in Python for computational performance; the results can then be displayed in the browser. All these tools allow the user to set multiple parameters, such as the percentage of accuracy, searching intervals or exact pitch, considering or ignoring the distribution of the notes on syllables, etc.
- The development and testing of different computational strategies showed that customised implementations of the Needleman-Wunsch algorithm offer optimal results in terms of relevance from a musicological perspective as well as computational performance.
- The analyses performed offer valuable insights into the composition techniques of medieval lyrics and their connections. The comparison highlighted the difference between common and uncommon melodic patterns, which helps us define the elusive principles regulating medieval secular music composition, i.e. its “musical grammar.” The results show how there is, indeed, a stark difference between liturgical melodies and vernacular song, beyond the evident differences of formal structures. While they are both transcribed with a similar notation system and are largely developed through diatonic movements, a vernacular song is statistically much more likely to find matches within the relatively small encoded corpus of vernacular lyrics (~1400 pieces) than within the Cantus DB (over 60.000). This shows that these traditions were very different expressions of European musical tastes during the Middle Ages. This could be explained by the fact that the liturgical tradition originated during a much earlier phase, and most of the sacred melodies sung in the 12th to 14th centuries were already centuries old when troubadours and trouvères composed their songs. Furthermore, even if some troubadours studied in monasteries and some were or became monks, monastic environments, which produced a large portion of extant liturgical sources, engendered rather closed musical traditions, which were less likely to strongly influenced courtly musical practices. The project has returned, on the other hand, a large number of musical similarities across vernacular repertories, made that more relevant given the limited data available, drawing the most promising path that future research should undertake.
- Important acquisitions regarding melodic reuse concerned the question of orality. Seminal works on oral cultures have focused on formularity as the characteristic element of oral composition and transmission, and musicologists, starting with Leo Treitler, have explored how this idea applies to liturgical traditions. This project has demonstrated how the as not yet well-defined category of formularity could be better described through more specific phenomena, and especially how, in medieval European vernacular song, formulaic elements mostly occur in sub-sets that have been transmitted almost entirely by written means, directly challenging a formularity–orality correlation.
- The MedMel platform underwent major updates and improvements during the duration of the project, providing an optimal infrastructure for encoding and publishing the music of medieval vernacular lyrics.
- In light of the advancement of digital musicology in the domain of Optical Music Recognition (OMR) of medieval sources, the project has developed a tool to integrate MedMel as the endpoint of the Rodan + Neon.js OMR workflow. Rodan is a software developed at McGill University for semi-automatic music information retrieval from medieval manuscripts; the results can be adjusted with Neon.js to produce an XML/MEI file, which can now be imported into MedMel.
- Three digital searching tools have been developed for musical analysis and comparison. The first, the “Search Sequence” (https://medmel.seai.uniroma1.it/searchMelody.php(opens in new window)) allows the user to enter a melodic sequence and search for matches in the MedMel database of vernacular song; the second, the “Search Song Tool” (https://medmel.seai.uniroma1.it/searchSong.html(opens in new window)) allows to select a whole song and find similarities in the MedMel database and in the Cantus Database; the third, the “Many2Many Search Tool” replicates the logic of the Search Song Tool comparing all melodies with every other melody in a dataset, and it is developed in Python for computational performance; the results can then be displayed in the browser. All these tools allow the user to set multiple parameters, such as the percentage of accuracy, searching intervals or exact pitch, considering or ignoring the distribution of the notes on syllables, etc.
- The development and testing of different computational strategies showed that customised implementations of the Needleman-Wunsch algorithm offer optimal results in terms of relevance from a musicological perspective as well as computational performance.
- The analyses performed offer valuable insights into the composition techniques of medieval lyrics and their connections. The comparison highlighted the difference between common and uncommon melodic patterns, which helps us define the elusive principles regulating medieval secular music composition, i.e. its “musical grammar.” The results show how there is, indeed, a stark difference between liturgical melodies and vernacular song, beyond the evident differences of formal structures. While they are both transcribed with a similar notation system and are largely developed through diatonic movements, a vernacular song is statistically much more likely to find matches within the relatively small encoded corpus of vernacular lyrics (~1400 pieces) than within the Cantus DB (over 60.000). This shows that these traditions were very different expressions of European musical tastes during the Middle Ages. This could be explained by the fact that the liturgical tradition originated during a much earlier phase, and most of the sacred melodies sung in the 12th to 14th centuries were already centuries old when troubadours and trouvères composed their songs. Furthermore, even if some troubadours studied in monasteries and some were or became monks, monastic environments, which produced a large portion of extant liturgical sources, engendered rather closed musical traditions, which were less likely to strongly influenced courtly musical practices. The project has returned, on the other hand, a large number of musical similarities across vernacular repertories, made that more relevant given the limited data available, drawing the most promising path that future research should undertake.
- Important acquisitions regarding melodic reuse concerned the question of orality. Seminal works on oral cultures have focused on formularity as the characteristic element of oral composition and transmission, and musicologists, starting with Leo Treitler, have explored how this idea applies to liturgical traditions. This project has demonstrated how the as not yet well-defined category of formularity could be better described through more specific phenomena, and especially how, in medieval European vernacular song, formulaic elements mostly occur in sub-sets that have been transmitted almost entirely by written means, directly challenging a formularity–orality correlation.
Progress beyond the state of the art and expected potential impact (including the socio-economic impact and the wider societal implications of the project so far)
This project has pioneered a computational approach to musical reuse and techniques of composition in medieval vernacular lyric. It created tools that redefine scholarly approaches to medieval music, enabling, for the first time, large-scale melodic comparison and proposing new methodologies in the domain of editorial practices, music analysis, and interpretation of the interplay between text and music.
MIME constitutes a valuable proof of concept: retracing melodic connections allows us to re-envision aspects of past musical cultures. Only through systematic comparison can we distinguish genuine melodic reuse from commonly used patterns.
To bring to fruition the full potential that this project has created, a large project with a broader scope will need to:
1) Encode all melodies from the vernacular tradition, especially providing a new digital edition of the trouvères, which count more than 4000 monophonic transmitted melodies.
2) Conduct systematic comparisons on medieval European song corpora, including not only Occitan and French, but also Galician-Portuguese and German lyrics. This could be further extended to Latin conducti, Italian laude, polyphony, etc.
3) The results gathered will allow us to describe techniques of composition, and potentially distinguish regional, authorial, genric, and notator’s characteristics on a systematic basis.
4) By networking connections and common musical features with historical and geographical data concerning places and dates of compositions, the author’s biographical information, the sources’ transmission, and information of places of artistic production, it will be possible to leverage the tools and methodologies developed in the course of the MIME project.
All these steps will be implemented in the context of the ERC Starting Grant MUSICA FRANCA (2026–2031). This MSCPF has established the theoretical and technical foundations, providing the tools, methods, and conceptual frameworks to design the project, as well as the practical conditions to prepare the application.
MIME constitutes a valuable proof of concept: retracing melodic connections allows us to re-envision aspects of past musical cultures. Only through systematic comparison can we distinguish genuine melodic reuse from commonly used patterns.
To bring to fruition the full potential that this project has created, a large project with a broader scope will need to:
1) Encode all melodies from the vernacular tradition, especially providing a new digital edition of the trouvères, which count more than 4000 monophonic transmitted melodies.
2) Conduct systematic comparisons on medieval European song corpora, including not only Occitan and French, but also Galician-Portuguese and German lyrics. This could be further extended to Latin conducti, Italian laude, polyphony, etc.
3) The results gathered will allow us to describe techniques of composition, and potentially distinguish regional, authorial, genric, and notator’s characteristics on a systematic basis.
4) By networking connections and common musical features with historical and geographical data concerning places and dates of compositions, the author’s biographical information, the sources’ transmission, and information of places of artistic production, it will be possible to leverage the tools and methodologies developed in the course of the MIME project.
All these steps will be implemented in the context of the ERC Starting Grant MUSICA FRANCA (2026–2031). This MSCPF has established the theoretical and technical foundations, providing the tools, methods, and conceptual frameworks to design the project, as well as the practical conditions to prepare the application.