From Text to Meaning. Historical Humanities in the Age of AI and Interdisciplinarity

This outlines how computation can serve historical research when historians and data scientists define questions, data and methods together. It first shows why well-structured sources, such as library catalogues, are indispensable starting points for historians engaging in computational research, and how processing full texts can feed back into this knowledge infrastructure by adding new structured layers of data. It will then discuss different approaches that turn text into word counts, embeddings, topic models and text reuse networks, each producing patterns that invite historical interpretation. Next, it sketches how the use of large language models promises richer, context aware analyses across languages and national borders. Because any large-scale study depends on knowing what an archive captures and what it leaves out, it also emphasises corpus-level control for coverage and bias. Taken together, these elements push historical research and Digital Humanities beyond the discussion of ‘mixed methods’ and toward shared data ecosystems that will be indispensable to the future of the historical humanities.