cb 2023. This work is licensed under a Creative Commons “Attribution 4.0 International” license. The Location and Function of Formulaic Expressions in the Resolutions of the Dutch States General Marijn Koolen 1,2 and Rik Hoekstra 1,2 1 Huygens Institute 2 DHLab - KNAW Humanities Cluster 1 Introduction Formulaic expressions are commonly used in administrative documents to signal important aspects of a document (Koolen and Hoekstra, 2022, Kopaczyk, 2012, 2013). Medieval charters contain opening and closing formulas to signal that the document is a charter and what type of charter it is (Boonen, 2005, De Boor, 1975). Notarial deeds contain formulas based on notary manuals to make sure the transaction they confirm is unambiguous and follows protocol (Lemercier and Trivellato, 2022, Marques, 2018, Zomeño, 2007). In previous work, we developed techniques to automatically detect formulas in historic document collections, while dealing with orthographic variation introduced by historic spelling variation and change and errors introduced by OCR and HTR processes Koolen and Hoekstra (2022). In this paper, we investigate the nature of the formulas detected in the resolutions of the Dutch States General. The resolutions are transcripts of the decisions of the Dutch States General (SG) in the period 1576-1796, estimated one million in total. Each resolution consists of at least two parts, a proposition and a decision (Thomassen, 2019). Most resolutions contain no more than this, but some resolutions have a deliberation part between the proposition and the decision. The clerks of the States General used formulas to signal important elements of resolutions: the start of a proposition, the type of proposition, the start of a decision, the type of decision, a request for a committee to investigate a matter further, or a reference to a previous resolution. Our aim is to algorithmically extract and classify these formulas, to provide additional metadata per resolution. This process is complicated because formulas contain lots of variation, caused by, amongst others: • Common linguistic variations due to grammatical changes (single, plural, present tense, past tense etc), orthographic variability and OCR/HTR mistakes, • Variations in writing styles between clerks resulting in slightly different wording, • Changes in the entities (often dates, names of persons and locations) contained within a formula, 1