Use Case: Narrative Techniques and Subgenres in the German Novel
Partners: Uni Würzburg, TU Darmstadt
The use case demonstrates how large collections of literary texts can be used to analyse the historic development of narrative techniques, subsequently enabling assertions on corresponding literary categories. Futhermore its goal is to provide a set of best practices and example workflows in the form of plain language tutorials that allow humanities scholars to use and adapt state-of-the-art tools according to their own research questions.
Data basis for the use case's development is a corpus comprising about 2000 German novels, primarily from the 18th and 19th century, and partly coded in TEI format. To further assess the transferability and robustness of proposed solutions, another corpus of about 200 French crime novels from the 19th and 20th century will be used.
Its technical foundation is the Darmstadt Knowledge Processing Software Repository (DKPro), an Apache UIMA based framework, featuring a unified approach to varying language processing tools, thereby simplifying the task of building processing pipelines. We access the framework using Python, a scripting language prevalent in academia.
Easy to use programming recipies provide starting points on numerous levels. Here we follow the typical setup of natural language processing pipelines, beginning with default pre-processing steps and gradually progressing up to more advanced topics such as automatically identifying textual entailments and semantic roles. The recipies are currently under development and available in the form of a living document.
Dr. Steffen Pielström
Stefan Pernes, MA