Representation of multiword expressions in the Bulgarian integrated lexicon for language technology

The chapter introduces a representation model of multiword expressions from the perspective of integrated lexicons for Bulgarian. The lexicons considered are an inflectional one, a valency one, and a wordnet. We created a joint representation entry that incorporates morphology, valency potential and lexical semantics through synonym sets. The selected mechanism for displaying all the information is catena-based since the catena allows for better modeling of idiosyncratic elements and is tree-based. Also, a general typology of multiword expressions is proposed that focuses on fixedness and (dis)continuity. We believe that providing a unified representation of multiword expressions and common lexica would improve the performance of the various natural language processing applications.