Nominalisation and rhetorical structure

A study of a small corpus of pharmaceutical leaflets supports the hypothesis that a subset of Mann and Thompson’s Rhetorical Relations may hold within as well as between clauses, with one or more propositional arguments nominalised. This paper summarises the main results of this study and outlines a formal account of the phenomenon within a framework of discourse generation, whereby nominalisation of propositions is licensed as a by-product of explicit realisation of the relations as verbs or propositions. 5.1 Rhetorical Structure Rhetorical Structure Theory (Mann & Thompson 1987) claims that a small set of rhetorical relations (RRs) hold between adjacent text units in a discourse, and that recognition of these relations may enable the hearer/reader to recover information or semantic content which is not explicitly conveyed by the components of the text. The list of relations in (M & T op cit) is not claimed to be exhaustive, but the relations can be classified into two types: on the one hand informational or semantic such as Cause and Condition, on the other hand intentional (presentational, pragmatic) such as Evidence and Justify. A typical pair of text units is said to consist of a Nucleus and Satellite, the former being more “central to the writer’s purposes”1 A full list of RRs from (Mann & Thompson 1987) is shown in Table 5.1. 5.1.1 Background RST originally assumed that there is a single preferred relation holding between two adjacent units, even in cases of ambiguity. Moore and Pollack (1992) challenge this assumption, and argue in particular that it is not desirable for informational and intentional relations to be in competition. They propose that both types of relation may hold simultaneously, for example Evidence and Volitional-Cause in example (1): The nucleus+satellite pattern is in fact one of five “schemas” identified by M & T but these distinctions are not relevant to the issues discussed in this paper.