loading
x
From Handwritten Document to Computable Object: Critical Approaches to Automatic Text Recognition for the Digital Humanities (DHSI 2026)

Description

This course will provide hands-on experience in the automatic transcription of handwritten historical documents from a variety of periods and in different languages. We adopt a critical approach to existing technologies (Transkribus, eScriptorium, and LLM-based workflows) and introduce a diversity of workflows. Using examples from Canadian and international digital collections, participants will explore how transcription and layout models can be created from scratch, as well as how existing ones can be retrained using data from the community. We also examine the importance of transcription choices on research, including when we work with models that do not directly reflect our own considerations. At the end of the course, participants will be able to assess the viability of including automatic transcription methods in individual and group research projects and they will have gained an appreciation for how automatically transcribed documents are applicable to a variety of forms of computational research. The course is intended for all levels, including those with little prior knowledge of digital scholarship. We encourage participants to bring their own research interests in handwritten historical documents and digital archives to  the course.

Instructor(s)

David Joseph Wrisley is Professor of Digital Humanities at NYU Abu Dhabi (UAE). His doctorate is in Romance Languages and Literatures from Princeton, with a concentration in comparative medieval literature. He is interested in computational methods in the humanities, particularly in multilingual and non-Anglophone environments. His current research focuses on developing handwritten text recognition (HTR) models (in Latin, Arabic and medieval French), as well as on uses of AI for the study of historical sources. For more than twenty years now, he has been involved in interdisciplinary and collaborative endeavors in the Arab world. Before joining the faculty at NYU Abu Dhabi, he taught at the American University of Beirut (Lebanon) from 2002 to 2016. Currently he co-directs the Paris Bible Project and the OpenGulf research group. Along with Estelle Guéville, he is the co-author of Medieval Manuscripts and the Computational Humanities: Big Data, Scribes, and the “Paris Bible.” (2026).

 

Estelle Guéville is a PhD candidate in medieval studies at Yale and has worked as a curator and a researcher in multiple cultural institutions around the world, including the Louvre Abu Dhabi. Her research focuses on premodern manuscript cultures, gender studies, and computational approaches to textuality. Her doctoral work examines scribal practices and women’s roles in textual production and transmission, advancing data-driven approaches to medieval book cultures. With David J. Wrisley, she is the co-author of Medieval Manuscripts and the Computational Humanities: Big Data, Scribes, and the “Paris Bible” (2026) which explores how AI-driven and computational methods can renew questions of authorship, textuality, and scribal attribution.

Ce contenu a été mis à jour le 27 February 2026 à 13h32.