This paper compiles a core set of data categories to be used in combination with the Reference Annotation Framework. It shows also that the underlying objective − ensuring internal and external coherence − is a difficult issue, since reference annotation articulates various levels of linguistic description which all contribute to the identification or qualification of markables and referential links. 1. Current practice of reference annotation Reference annotation associates referring expressions – usually certain types of noun phrases – with information that enables their interpretation (e.g. antecedents). This type of knowledge is required for a variety of language processing applications, including information extraction and retrieval, machine translation and human-machine dialogue. Reference annotation in a broad sense, covering coreference, anaphora and reference, has already been the subject of substantial practical and theoretical work which suggest basic principles for coherent coding procedures and attempts to unify existing practices (Hirschman &

Abstract. This paper presents a statistical system for resolving bridging descriptions in French, a language for which current lexical resources have a very low coverage. The system is similar to that developed for English by [20], but it was enriched to integrate meronymic information extracted automatically from both web queries and raw text using syntactic patterns. Through various experiments on the DEDE corpus [8], we show that although still mediocre the performance of our system compare favorably to those obtained by [20] for English. In addition, our evaluation indicates that the different meronym extraction methods have a cumulative effect, but that the text pattern-based extraction method is more robust and leads to higher accuracy than the Web-based approach. 1