| Abstract [eng] |
This article explores the use of large language models (LLMs) and digital tools for analysing the YIVO's administrative correspondence from 1931 to 1939. The study aims to explore the cultural and ethnolinguistic identity of Jewish communities during the interwar period, with a focus on the Yiddish-speaking diaspora in Central and Eastern Europe, specifically in territories that were part of the Austro-Hungarian and Russian empires before the First World War. By leveraging modern technologies, such as OCR and XML markup, a large corpus of multilingual archival documents was processed efficiently. The article is structured in three parts. The first presents a historical overview of the YIVO's efforts in documenting Jewish cultural life during the interwar period. Following this, the second section discusses the methodology, detailing the use of digital tools such as OCR, XML markup, and LLM-driven post-correction. Finally, the third part offers the results of a quantitative and qualitative analysis of the correspondence, focusing on its geographical scope, linguistic diversity, and thematic evolution. The analysis underscores the transformative potential of LLMs and other digital tools in historical research, enabling more efficient processing of archival documents and deeper insights into cultural patterns. The study concludes by advocating for a 'second archival turn' through mass digitisation and markup of archival content, which will significantly boost research in underrepresented languages and cultures. |
