Towards a NAvigator Tool for Dutch Verbaal-Archives: Leveraging Nineteenth-Century Archival Logic for Keyword Search

Recent advances in machine learning and Automated Text Recognition (ATR) have encouraged efforts to digitize historical archives. However, processing text with an ATR-engine is not (necessarily) making archives accessible to researchers and the broader public; additional datafication steps are needed. This paper adds to this conversation by introducing and evaluating a simple navigation tool, the Dutch National Archives navigator, or in short NAvigator, that capitalizes on the structure of a specific historical archival management method, the verbaalstelsel-1823 (‘verbal system’), which was common in Dutch government archives between the early nineteenth century and 1950. The NAvigator provides structured access to the information within such archives by recreating the historical workflow of search using the archives of the Dutch Ministry of Colonies (1850-1900) as a case study. The paper outlines the steps towards the development of the NAvigator and the challenges posed along the way, reports the results in terms of information retention throughout the search process, and discusses the potential for future research based on the intermediate output of the tool. As an early result, it finds that 84.7 percent of glossary entries are successfully connected to their corresponding documents.