Abstract
Ever since the formulation of Rhetorical Structure Theory (RST) by Mann and Thompson, researchers have debated about what is the ‘right’ number of relations. One proposal is based on the discourse markers (connectives) signalling the presence of a particular relationship. In this paper, I discuss the adequacy of such a proposal, in the light of two different corpus studies: a study of conversations, and a study of newspaper articles. The two corpora were analysed in terms of rhetorical relations, and later coded for external signals of those relations. The conclusion in both studies is that a high number of relations (between 60 and 70% of the total, on average) are not signalled. A comparison between the two corpora suggests that genre-specific factors may affect which relations are signalled, and which are not.
Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
