In machine learning, algorithms are increasingly used for understanding, processing and producing human language. The computer learns, for example, linguistic patterns that people typically employ when giving positive, neutral or negative valuations, and can then classify valuations automatically.

The machine learning turn is not only relevant to disciplines like computer sciences, engineering and social and economic sciences, but also for the humanities, notably for the field of digital humanities where literature and functional texts are subjected to automated analyses.

What needs to be borne in mind, however, is that these algorithmic methods follow their own research logic and premises, since they have been developed for entirely different ends, such as for commercial marketing purposes. So, how do these research epistemologies influence the humanities? And what kind of societal implications can we expect if algorithms classify valuations or translate texts?

In order to investigate these implications, the project team will focus on valuation practices as the central element of Web 2.0. They will carry out exemplary analyses of digital valuation practices in travel blogs and social reading, applying the methods of machine learning and deep learning, but also using the humanistic approaches of corpus linguistics and the literary approach of distant reading. This will allow entirely new findings about valuation practices in Web 2.0. What is of particular interest, however, is that the project will be able to reveal and reflect on hidden research epistemologies and to discuss elementary questions of digital society and science.

Project title: Research Epistemologies in Text-based Digital Humanities: Analyses of Valuation Practices after the Machine Learning Turn
Project duration: 1 December 2018 – 31 May 2020
Funding: SNSF Digital Lives
Project leader: Prof. Noah Bubenhofer (Winterthur) and Dr. Berenike Herrmann (Basel)

Digital Linguistics (in German only)

Digital Humanities Lab