Computational Analysis of Coherence Relations in Dutch

The NWO-programme Modelling textual organisation: coherence and cohesion studies the organisation of text into structural units by means of coherence (discourse relations between clausal and larger textual units) and cohesion (lexico-semantic relations between words in textual units). The programme is organised around two related PhD-projects, focussing on coherence and cohesion, respectively. Both projects rely on the availability of an annotated corpus of Dutch texts. Below, we propose a modification of the first project (on coherence). The original project concentrated on an in depth study of a theoretical issue with practical implications (the structure of discourse relations). In the modified proposal, we emphasize computational issues related to discourse annotation and discourse structure. In particular, we propose a project which studies coherence in Dutch texts, and the mechanisms that are used to establish coherence. We are interested in the following computational research questions: