Early modern Dutch political historiography has long privileged Holland at the expense of the provincial periphery, leaving the political voices of petitioners and communities in regions such as Overijssel effectively silenced. This presentation introduces a ML/AI-driven pipeline — combining Handwritten Text Recognition (HTR), Named Entity Recognition (NER), and Wikidata-linked toponym resolution — developed to make the Resolutions of the States of Overijssel (1578–1795), comprising approximately 50,000 handwritten pages, accessible for historical analysis. The pipeline, realised in collaboration with Collectie Overijssel and the HAICu project, enables systematic identification of persons, dates, and places across the corpus, transforming manuscript sources into structured, spatially and temporally queryable data. Initial results — including the identification of meeting dates and linked place names — already permit preliminary analysis of the spatial and temporal patterns of provincial political participation, with fuller findings to follow as enrichment progresses. The complete dataset will be made publicly available by spring 2028, coinciding with the province's 500th anniversary, and the methodology is deliberately designed to be transferable to comparable provincial corpora across the Dutch Republic.
Romein et al. (Wed,) studied this question.