“Network Analysis and Spatial Stylometry in American Drama Studies” (NASSA)

Dr. Dennis Mischke, University of Potsdam (Potsdam, Germany), Dr. Michał Choiński Jagiellonian University (Kraków, Poland),Joanna Byszuk, Institute of Polish Language, Polish Academy of Sciences (Kraków, Poland),Mathias Göbel, Staats- und Universitätsbibliothek (Göttingen, Germany) Abstract:“Network Analysis and Spatial Stylometry in American Drama Studies” (NASSA)The transnational workgroup “NASSA” (NASSA, 2019) seeks to promote computational research methodologies in American Drama Studies. In particular, the workgroup develops a digital workflow that combines stylometry, network analysis and GIS mapping to conduct a comprehensive digital study of the Chadwyck-Healey American-Drama Corpus 1714-1915 (ADC-CHC 2006) – a collection of 1500 American dramatic texts from three centuries. The project aims at the accomplishment of three overall research goals. First (A), we want to account for a number of about 150 texts by anonymous authors by measuring their stylometric distance to texts by known playwrights of the same corpus, using the Burrows’ Delta method (Burrows, 2002). Second (B), we want to study potential correlations between the linguistic origin (such as regional dialects) of authors (cf. Choiński,​ Eder​ ​&​ ​Rybicki, 2018) and the non-fictional spatial settings of a selected number of dramatic texts in a method that we call spatial stylometry. We propose spatial stylometry as a cross-section between geospatial humanities and stylometry. In our approach we combine the automatic extraction of geonames via Named-Entity Recognition (Spacy, NLTK) and match these entities with open repositories of geonames (such as the OpenStreetMap database). In a further step, we finally visualize the annotated results on a map that displays detected stylometric signals related to the spatial positioning of particular dramas. With this workflow, we wish to inquire whether detected real-word settings of a selected number of dramatic texts cluster in meaningful ways, approaching questions such as: are texts by playwrights from distinct regions also set in these places? and how do detected similarities between texts correspond to authorial and metatextual information such as race, class or gender? In a third avenue (C), we seek to enhance our study with a network analysis and distant reading of scenic presence of characters (cf. Trilcke, Fischer 2018) to find potential correlations between the social configuration of individual texts, their spatial styles as well as racial, ethinic or gender-specific configurations.