Tracing Syntactic Change in the Scientific Genre: Two Universal Dependency-parsed Diachronic Corpora of Scientific English and German

Proceedings of the 13th Conference on Language Resources and Evaluation (LREC 2022), pages 4808–4816 Marseille, 20-25 June 2022 © European Language Resources Association (ELRA), licensed under CC-BY-NC-4.0 4808 Tracing Syntactic Change in the Scientific Genre: Two Universal Dependency-parsed Diachronic Corpora of Scientific English and German Marie-Pauline Krielke, Luigi Talamo, Mahmoud Fawzi, J¨org Knappen Saarland University Campus A2.2, 66123 Saarbr¨ucken, Germany {mariepauline.krielke, luigi.talamo}@uni-saarland.de maib00001@stud.uni-saarland.de j.knappen@mx.uni-saarland.de Abstract We present two comparable diachronic corpora of scientific English (RSC UD-Parsed 1.0) and German (DTAW UD- Parsed 1.0) from the Late Modern Period (17thc.–19thc.) annotated with Universal Dependencies. We describe several steps of data pre-processing and evaluate the resulting parsing accuracy showing how our pre-processing steps significantly improve output quality. As a sanity check for the representativity of our data, we conduct a case study comparing previously gained insights on grammatical change in the scientific genre with our data. Our results reflect the often reported trend of English scientific discourse towards heavy noun phrases and a simplification of the sentence structure (Halliday, 1988; Halliday and Martin, 1993; Biber and Gray, 2011; Biber and Gray, 2016). We also show that this trend applies to German scientific discourse as well. The presented corpora are valuable resources suitable for the contrastive analysis of syntactic di- achronic change in the scientific genre between 1650 and 1900. The presented pre-processing procedures and their evaluations are applicable to other languages and can be useful for a variety of Natural Language Processing tasks such as syntactic parsing. Keywords: universal dependencies, evaluation, English-German contrastive, diachronic linguistics, scientific language 1. Introduction In the past years, interest in diachronic linguistic stud- ies has grown. Thanks to new methods in corpus-based, computational diachronic studies such as diachronic word embeddings and deep neural networks to detect lexical semantic change, research in diachronic linguis- tics has advanced a great deal in understanding the evo- lution of the lexicon over time. However, efforts in unravelling grammatical change are less dynamic due to the lack of suitable and sound historical linguistic resources. For this reason, most previous studies on the diachronic syntactic development of scientific dis- course in German (M¨oslein, 1974; Beneˇs, 1981; Haber- mann, 2011) and English (Halliday, 1988; Halliday and Martin, 1993) are purely descriptive, or, if corpus- based, limited to English (Biber and Gray, 2011; Biber and Gray, 2016). In this paper, we present two Universal Dependency (UD) parsed corpora suitable to trace the syntactic de- velopment in the genre of scientific English and sci- entific German in the period between 1650 and 1900. This period, marked by the scientific revolution and a turn towards experimental science, is known as the beginning of modern science. To avoid pitfalls con- nected to the processing of historical language data, we propose several steps to make UD-parsing less error- prone. Familiar difficulties in working with historical data stretch from pre-processing to linguistic annota- tion. Due to not being digitally born, historical data require high pre-processing efforts including standard- ization of data formats, data cleanup (e.g. OCR errors) and meta-data derivation …

Content truncated to 500 words