Research Projects
d-Prose
The corpus “d-Prose 1870-1920” is a heterogeneous collection of 2511 German-language literary prose texts published between 1870 and 1920. It contains narrative texts from trivial and high literature with a minimum length of 1000 words. The corpus was created from manually and semi-automatically selected texts from the German-speaking area.
→ Corpus: d-Prose
Event
In EvENT (“Evaluating Events in Narrative Theory”) we are developing, in cooperation with the Language Technology Group at the University of Hamburg (Prof. Dr. Chris Biemann), an approach for formalizing and analyzing events and eventfulness in narratives. The approach is based on natural language processing (NLP) and narratological concepts from literary studies.
KatKit
The aim of KatKit (KATegories toolKIT) is to support humanities thinking through mathematically grounded methods in order to reduce the risk of unintended inconsistencies and inconsistent models, and to facilitate the operationalization of humanities concepts.
For the toolkit, methods of applied category theory are adapted and further developed for use in the digital humanities. The interdisciplinary approach pursued here combines thinking in the humanities, computer science and philosophy of science, whose common points of contact are structures and patterns that can be described in category theory. Ultimately, the aim is to integrate the results into existing tool chains and workflows of the Digital Humanities and to be able to use them with existing software (e.g. the annotation software CATMA).
→ More about KatKit
Conflicts
In literary studies, the concept of conflict is used in a variety of contexts. These include the analysis of conflicts between characters, the study of how epochs deal with societal conflicts, and the exploration of poetological concerns. In the fortext lab, we focus on (1) the operationalization of conflict types using a category-theoretical approach and (2) heuristics for measuring conflict e.g. through adapted methods of word embedding-based sentiment analysis.
PLANS
PLANS (“Unitizing Plot to Advance Analysis of Narrative Structure”) is a follow-up project of EvENT (“Evaluating Events in Narrative Theory”), focusing on the exploration and design of plot concepts in narrative theory. Based on linguistic and narratological approaches, we are modeling units of action that enable the exploration of narrative structures and the construction of texts.
→ More about PLANS
Szenen
“Szenen” is a joint research project with the Universities of Würzburg and Cologne, which targets the automatic detection of scenes in narrative texts. Based on our definition of scenes (Gius et al. 2019), we formulated annotation guidelines and conducted a manual scene annotation of several narrative texts of high and trivial literature. In addition, we organized a Shared Task on scene recognition in 2021, where participants were asked to develop models for automated scene segmentation.