“Türkensteuer, Anschlagmodus, Gemeiner Pfennig”, what else?

A key challenge for the long-term editorial project Deutsche Reichstagsakten is the lack of a cross-edition subject index, alongside the complexity of the existing specialised indices. These factors hinder the searchability of texts and their comparability with sources from other European proto-parliamentary estates assemblies. While these indices are largely consistent with contemporaneous language, they are shaped by editorial selection and primarily designed from a print-focused perspective. In this article, we test and compare three Topic Modelling methods frequently used in Digital Humanities, particularly for modern parliamentary records. These methods facilitate the automatic identification of language-based word clusters, which can be labelled with new subject terms to create updated indices and critically evaluate on editorial practices. Using the minutes from the Imperial Diet Records of 1576, we applied and compared Gensim and Mallet, both based on Latent Dirichlet Allocation, as well as the transformer-based BERTopic. All three methods produced strong results for approximately 100 topics. These include not only content- and function-related topics—traditionally prioritised in editorial work—but also topics related to the functioning and procedural aspects of the Imperial Diet. Such findings highlight and support recent research efforts into the functioning and symbolic dimensions of these historical assemblies.