SHERBET - Stemmatology for the HEbRew BiblE Transmission – SHERBET

The SHERBET project aims at improving the understanding of the genealogical lineage of the Hebrew Bible through computational stemmatology methods, by applying the latest advances in applied mathematics and natural language processing to reconstruct the stemmata of the Hebrew Bible manuscripts and more specifically of its oldest witnesses, the Dead Sea scrolls (DSS). To reach this goal, we will (1) objectively validate the relevance of using stemmatology algorithms to reconstruct biblical tradition lineage by generating synthetic datasets representative of the DSS acting as a ground truth for the algorithms; (2) provide a comprehensive benchmarking survey of algorithms using this dataset; (3) design our own probability-based model specific to the constraints of textual transmission and the complexity of human language by modeling scribal behavior, and compare it to the state of the art; (4) apply the best performing algorithm to the real data and draw conclusions regarding scribal behaviors and the textual fluidity of biblical texts transmission.
This multidisciplinary project will be carried by different teams, representing each involved field: Pr. ( J.-S.) F. Rey and Pr. J. Elfassi from the Ecritures laboratory (University of Lorraine) as well as E. Reymond (Yale) will provide their textual criticism and biblical expertise, Pr. J. Istas from the Jean Kunzmann laboratory (University Grenoble Alpes) and Pr. M. Clausel from the Institut Élie Cartan de Lorraine (University of Lorraine) will provide their expertise in the modelization of complex systems and applied mathematics, and Dr. M. Amblard and Dr. G. Guibon from the LORIA laboratory (University of Lorraine) will provide their expertise on linguistics and Natural Language Processing methods. Additional input regarding software development and algorithms implementation will be provided by Pr. L. Gonnord from the LCIS laboratory (Institut polytechnique de Grenoble).