Identifying Copied Fragments in a 18th Century Dutch Chronicle

Proceedings of the 13th Conference on Language Resources and Evaluation (LREC 2022), pages 5865–5878 Marseille, 20-25 June 2022 © European Language Resources Association (ELRA), licensed under CC-BY-NC-4.0 5865 Identifying Copied Fragments in an 18th Century Dutch Chronicle Eleanor L. T. Smith1, Lianne Wilhelmus2, Erika Kuijpers2, Alie Lassche3, Roser Morante2 1 Faculty of Arts, University of Antwerp Eleanor.Smith@uantwerpen.be 3 Faculty of Humanities, Leiden University a.w.lassche@hum.leidenuniv.nl 2 Faculty of Humanities, Vrije Universiteit Amsterdam {erika.kuijpers, r.morantevallejo}@vu.nl; l.e.wilhelmus@student.vu.nl Abstract We apply computational stylometric techniques to an 18th century Dutch chronicle to determine which fragments of the manuscript represent the author’s own original work and which show signs of external source use through either direct copying or paraphrasing. Through stylometric methods the majority of text fragments in the chronicle can be correctly labelled as either the author’s own words, direct copies from sources or paraphrasing. Our results show that clustering text fragments based on stylometric measures is an effective methodology for authorship verification of this document; however, this approach is less effective when personal writing style is masked by author independent styles or when applied to paraphrased text. Keywords: authorship verification, stylometry, chronicles, Dutch 1. Introduction Recent studies have expanded our understanding of the production, circulation and marketing of news media in Early Modern Europe (c. 1500-1800) (Blair, 2010; Pettegree, 2014; Blair et al., 2021). Far less is known about the reception of news and information. How and when did the wider public change their media consump- tion in this period? An interesting source of information about reading habits and media use are local chronicles. These are hand-written narratives produced by in ma- jority middle class authors, who recorded events and phenomena they considered important and memorable (local politics, upheavals, climate, prices, crime, deaths) (Pollmann, 2016). To do so, authors made use of a wide array of sources and frequently copied older chronicles and history books, excerpts from official documents, local announcements, by-laws and broadsheets, as well as of newspapers and periodicals that became increas- ingly popular in the course of the eighteenth century. Some chroniclers clearly mention where they got their information, for instance from a newspaper, in church, at the local market or through official government pub- lications. More often authors did not explicitly quote their sources. There were no conventions concerning referencing, and copy rights still had to be conceived of. This paper explores whether computational techniques of authorship verification can be used to detect copied fragments in chronicles, and thus contribute to our un- derstanding of media use through the practice of copy- ing by early modern chroniclers. More specifically, we apply computational stylometric techniques to an 18th century Dutch chronicle to determine which frag- ments of the manuscript represent the author’s own orig- inal work and which show signs of external source use through either direct copying or paraphrasing. The chronicle is part of a recently digitized corpus1 that contains c. 320 early modern Dutch language chroni- cles written between 1500 and 1850 in what are now the Netherlands and …

Content truncated to 500 words