ERC Synergy MiDRASH Awarded to EPHE and partner institutions
MiDRASH – Migrations of Textual and Scribal Traditions via Large-Scale Computational Analysis of Medieval Manuscripts in Hebrew Script
The European Research Council has awarded an ERC Synergy Grant to the EPHE and its partner institutions in Israel for a six years computational-humanities project on Medieval Hebrew manuscripts starting in 2023 with a budget of over 10 million Euro. eScriptorium and kraken play a central role in it.
The National Library of Israel has already digitized some 100,000 manuscripts in Hebrew (as well as Aramaic, Judeo-Arabic, and other Jewish languages) located in libraries around the globe as part of its Ktiv project. However, making their contents accessible to researchers and laypersons for detailed study remains a monumental challenge.
In this project, we aim to achieve the following goals:
• Convert the writing appearing in the images of a large fraction of those manuscripts into searchable text — including as much as possible of interlinear insertions and marginal comments. This will require significant improvements in the state-of-the-art of computerized page segmentation and handwritten-text recognition.
• Have the computer compare the multitude of texts to find quotations, paraphrases, borrowings, allusions, and other intertextual relations. These can be made available to people browsing or studying the texts.
• Train computer algorithms to analyze handwriting so as to determine where and when each manuscript was written — down to the century and the country. This will involve synergetic feedback between human paleographers and modern neural-network machine learning, teaching the computer to detect the features that have been identified by experts throughout the manuscripts, identifying outliers, and explaining machine classifications in human-understandable terms.
• Train computer algorithms to analyze the texts for linguistic (lexical, morphological, and syntactic) features that can be used in text searches and which can help place literary works in their historical context.
• With the above resources in hand, several case studies will be undertaken to map out the migration and evolution of texts and ideas within medieval Jewish communities.
The results of this project — data as well as algorithms — will be made available to the public.
PI Nachum Dershowitz, Tel-Aviv University
PI Judith Olszowy-Schlanger, EPHE-PSL
PI Avi Shmidman, Bar Ilan University
cPI Daniel Stökl Ben Ezra, EPHE-PSL
with the partner institutions National Library of Israel and Haifa University.