The Institute for Natural Language Processing (IMS) at University of Stuttgart has an opening for a doctoral researcher (or a postdoc) in the context of project CAUTION  to work on the operationalization of concepts from the field of narratology and on tools for computer-aided analysis of a corpus of literary texts. CAUTION is a collaboration between literary studies and computational linguistics, headed by Janina Jacke (University of Göttingen) and Jonas Kuhn (University of Stuttgart) , which addresses the phenomenon of unreliable narration in fiction. The project aims to advance the representational means for capturing this phenomenon, to devise a framework for intersubjectively stable annotation in texts, and to develop computational tools for automatically detecting signals for unreliable narration in a corpus. The project is funded by DFG (the German Research Foundation) and is associated with the Priority Programme SPP 2207 Computational Literary Studies.
The successful candidate will work (i) on the detection of text properties signaling unreliable narration using data-driven techniques from natural language processing (NLP) and (ii) on capturing systematically what reasoning leads a reader to consider the narrator of a story to be unreliable, using a symbolic knowledge representation & reasoning framework. Specifically, the project explores how the interaction between story-internal knowledge and background knowledge from various sources can be formalized in a belief-desire-intention model for intelligent agents and how such a formalization can be integrated with the practice of text annotation. An important element of the project will be a close exchange between theoretical research in literary studies and research on algorithmic modeling techniques.
The candidate must have a Master’s degree in computational linguistics, computer science, digital humanities, or similar. Familiarity with the data-driven modeling paradigm in current NLP research, programming skills and experience in running and evaluating corpus-based modeling experiments are a prerequisite.