LostMa: The Lost Manuscripts of Medieval Europe: Modelling the Transmission of Texts

Why have some cultural productions survived to this day while others have been lost? LostMa aims to understand how human cultures are formed and evolve, through the transmission of written artefacts. The project focuses primarily on European medieval chivalric literature and combines complexity science, artificial intelligence, and philological expertise.

Presentation

LostMa aims to understand how human cultures are formed and evolve, through the transmission of written artefacts. This project seeks to establish to what degree the transmission (and subsequent preservation or loss) of written artefacts, texts and ideas deviates from chance, and if so, why. To this end, LostMa will analyse how manuscript texts have been copied, transformed, or destroyed, following the example of the evolution of living organisms or linguistic variants, through processes of innovation/mutation, fixation or extinction.

Hence, the project aims not only to understand the processes by which texts are transmitted but also to grasp the extent to which humans are the actors in the transmission of their own culture and the extent to which the survival of texts or the constitution of cultural canons are due to chance.

While this notion may seem provocative to researchers in the human sciences, evolutionary biologists have long since discovered the role of random drift in the survival or extinction of genetic traits and species.

To investigate this question, the project will attempt a paradigm shift in philological methods, combining artificial intelligence, complexity science, and philological expertise. Stochastic birth and death processes and multi-agent computer simulations will be used to simulate the text transmission process.

The case study will focus on chivalric literature in the European context. Deep learning methods will support large-scale data collection on a corpus of 4,000 documents in Romance, Germanic, and Celtic languages, focusing on the full text of around 1,000 Old French manuscripts. The data will provide observable values to compare with simulation results, measuring deviations from chance, making inferences about unobservable values such as loss/survival rates of works and manuscripts, and understanding the dynamics at work behind the transmission of texts.