Pas de documents, pas d’histoire

Can history be written without a close reading of documents, relying instead on regression analysis and predictions generated by artificial intelligence? While traditional historiography remains grounded in archival evidence, recent advancements in AI challenge us to explore new forms of historical inquiry, especially in contexts where documentation is fragmentary, lost, or inaccessible. Drawing on the digital infrastructure of the Portal to Jesuit Studies, the Institute for Advanced Jesuit Studies (Boston College) is developing a system of interconnected repositories populated with human-validated data. This ecosystem constitutes a validated corpus upon which data-centric AI models can be trained and subsequently deployed for regression analysis and predictive or reasoning modeling. A key application of this approach lies in archival reconstruction—predicting the existence of missing documents and unpreserved fonds or assisting archivists in reorganizing documents. This article presents the architecture of the evolving database system and initial AI experiments on the Portal’s materials. By embedding AI within transparent, traceable, and critically governed systems, the project seeks to open space for methodological innovation while maintaining the epistemological rigor that defines historical scholarship

Digital humanities · Documentation · Historiography · Key (lock) · Politics · Reading (process) · Style (visual arts) · Architecture · Computational and Text Analysis Methods · Digital and Traditional Archives Management · Digital Humanities and Scholarship