Verber med fælles syntaktiske konstruktioner og semantisk slægtskab

Abstract

This paper deals with the issue of observable relationships between the syntactic behaviour and semantics of selected verb groups. The aim is to investigate the feasibility
of exploiting syntactic similarity for effective uncovering of semantic kinship. The investigation is inspired by the demand for large-scale lexical resources combining
morphological, syntactic and semantic descriptions on the one hand, and on the other by the fact that semantic networks, like WordNet usually do not at all, or only to a very
limited degree provide syntactic information. Both aspects show that there is a need for effective means that combine syntactic and semantic description methods and resources
containing the two information types.
The pilot project presented here discusses verbs sharing selected syntactic descriptions; they are extracted from the Danish Lexicon for Language Technology Applications (STO)
and grouped together according to semantic labels, which in the first run are intuitivelys elected. The idea is to derive underlying semantic information systematically from the
syntactic surface structures described. Although this approach starts from the ‘wrong side’ compared to other, well-known analyzes of the semantics-syntax relationship, it
might be exploited in prediction of verb senses.