Digital Edition and Text Modelling

The focus of the working group “Digital Edition and Text Modelling” is on how to represent, model, present and analyze texts, not only in terms of their language, but also - perhaps especially - in terms of their content.  

Members of the WG are involved in various edition projects at the ACDH-CH. The Austrian Academy has a long history of editing the works of Arthur Schnitzler (1862–1931) which began in 1976. This theme is continued in current projects including an edition of his early works and an edition of his correspondence with fellow writers. Other subjects of our editoral work include the writer and essayist Karl Kraus, a contemporary of Schnitzler, and the playwright and novelist Elfriede Jelinek.

The working group also explores methods of applying well-established encoding methods from digital edition practice (TEI) to non-textual materials, such as multimedia content and historical information, and work on methods for bridging TEI text encoding and the Semantic Web. Related to this, we offer an interconnected database of people, places, works, organisations and events (“Personen der Moderne Basis”, PMB) focused on the turn of the 20th century, related inter alia to Schnitzler and his environment, and used heavily in the relevant edition projects.

Beyond the edition projects in a narrower sense, the working group is interested in the dynamic processes that shape text transmission and, more generally, knowledge and information dissemination. Network science and machine learning are very useful tools in this regard. An ongoing topic within the group is how we might represent texts that are passed down in multiple copies, especially from ancient and medieval times. Here, we can make use of graph representations of the text, which allow for a clear visual indication of where the text has been particularly stable or unstable during its transmission. We can also focus on the temporal evolution of citations within a group of texts, or model the diffusion of news as a complex system of interactions among agents. The common ground in all of these cases is to focus on the relations between the texts to extract common patterns or relevant information from them. We can also rely increasingly on natural language processing tools to provide crucial information to enrich digital editions.