Agnès SÁNDOR (Xerox Research Centre Europe, Meylan)
Modeling metadiscourse conveying the author's rhetorical strategy in biomedical research abstracts
2007, Vol. XII-2, pp. 97-108
The importance of the role of metadiscourse is increasingly recognized for natural language processing applications like text-mining and information extraction. Thus the detection of metadiscourse has recently been identified as a task in several domains, including the processing of scientific literature. We have developed a natural language processing system that detects and highlights in biomedical research abstracts a particular kind of metadiscourse that conveys the author’s rhetorical strategy. In this paper we describe the model of rhetorical metadiscourse underlying the system. Our model, combining and extending previous discourse analysis methods and models, is based on both conceptual and syntactic analyses of metadiscourse. We argue that this model is effective for automatic processing.