Spoken Songs: Algorithms of Composition and Improvisation

• Subscribe to the project newsletter (in Latvian) for quarterly updates.

The project focuses on spoken songs, the oldest layer of Latvian traditional vocal music, characterized by narrow-range melodies shaped by the text and by improvisation. By combining algorithmic and qualitative methods, the project aims to understand principles of composition and improvisation that have long been neglected in scholarly research due to the separation and fragmentation of melody and text. An interdisciplinary team of researchers encode melodies and texts, conduct algorithmic analysis and qualitative case studies, and identify and map compositional and improvisational skills and techniques, focusing on the interaction between text and melody and on local specificities. The project is further developing Baltic and Nordic cooperation and strengthening the presence of digital musicology and folkloristics in Latvia’s research and education. Its innovations include the use of AI methods in analysis.

Project outputs will include seven scholarly articles, publicly accessible datasets and corpora, a thematic website, an EU project application, at least seven conference presentations, educational and outreach activities, and student involvement, thereby promoting heritage preservation and skills in digital humanities.

Project newsletters (in Latvian):

First-quarter newsletter (09.03.2026.)
Second-quarter newsletter (19.06.2026.)
Third-quarter newsletter (18.09.2026.)

Project No. lzp-2025/1-0252.