Digital Hermeneutics of Historical Source: Formalization as Interpretation

The article is devoted to understanding the transformation of the methodology of historical research under the in-fluence of the “digital turn”, with a focus on the key discipline – source studies. The author examines how the rapid development of artificial intelligence (AI) technologies, especially machine learning, poses new challenges and opens up opportunities for historians, requiring a rethinking of traditional practices of working with historical sources. The central problem of the article is the process of transition from a historical source to formalized data (for example, through digital hermeneutics) and back to interpretation and new historical knowledge. The author empha-sizes that this process is not just a technical operation, but a deeply methodological act of abstraction (using the ex-ample of collected historical data, or capta), associated with the inevitable loss of nuances and requiring critical re-flection. The article examines methodological dilemmas of historical research, such as the opposition between source-oriented (ad fontes, striving for maximum completeness) and problem-oriented (ceteris paribus, focus on se-lected parameters) approaches, their evolution in the digital age, and related discussions. The “illusion of objectivity” of digital methods is critically analyzed. The author argues that algorithms, digital tools (DBMS, GIS, NLP), and interfaces are not neutral: they carry biases, limit focus, simplify complexity, and sometimes even predetermine research questions, which gives rise to risks: algorithm bias, data selectivity, the influence of visualization and interface design, and uncritical acceptance of AI results. As a promising solution that overcomes the key limitations of genera-tive models (hallucinations, “black box”, isolation from sources), the article considers in detail the RAG (Retrieval-Augmented Generation) approach, showing its advantages for a historian. The main conclusion of the article empha-sizes the indispensable role of the historian-researcher in the era of AI. RAG systems and similar tools act not as a replacement, but as powerful “digital magnifying glasses” or intelligent assistants, taking on the labor-intensive tasks of searching and primary data processing, freeing up the historian's time for critical interpretation, assessing the reliability of sources, building arguments, formulating research questions, and creating new historical knowledge.