A top-down approach to discourse-level annotation
Résumé
In this presentation, we present a top-down approach to discourse-level annotation. After situating our project within the context of discourse corpus-linguistics, and existing annotated resources, we expose the principles of the top-down approach which is based on the assumption that the text as a whole is considered as a functional and semantic unit. Instead of giving an exhaustive view of the corpus annotation project, we have chosen to exemplify our method with the detection of a specific discourse structure that is associated with a wide range of textual patterns: the enumerative structure. A review of previous studies will enable us to clarify the specificities of our approach, and to give the rationale for this choice. A presentation of our NLP-assisted annotation method, which involves automatic tagging and pre-marking, allows us to explain our framework and the annotation campaign. Finally a number of examples of annotated structures will be commented and discussed.