ISO-DR-core plugs into ISO-dialogue acts for a cross-linguistic taxonomy of discourse markers

Abstract

The present paper proposes an interoperable taxonomy to represent the meaning of discourse markers based on ISO DR-core (ISO 24617-8) but with a plug-in to ISO-dialogue acts (ISO 24617-2). The proposed taxonomy encompasses two dimensions: the semantic, with values regarding the discourse relations signalled by discourse markers, and the pragmatic, with values concerning the communicative function realized by discourse markers. We present a proof of concept for this twodimensional taxonomy in a multilingual parallel dataset in three languages, English, European Portuguese and Bulgarian, comprising 165 textual segments with multiword discourse makers obtained from publicly available TED Talk transcripts. We show that the two-dimensional taxonomy can successfully annotate cross-linguistically the meaning of discourse markers and discuss linguistic evidence where extension of the proposed taxonomy can be relevant.