Danish verbs as knowledge probes in corpus-based terminology work.

Abstract

The aim of this article is to study the suitability of a number of Danish verbs as linguistic signals or rather knowledge probes for the retrieval of definitions in a Danish corpus on hydraulics. Moreover, this study will form the basis of a catalogue of knowledge probes for the retrieval of terminological information from machine-readable corpora in practical terminology work. The choice of one single subject area makes my study comparable with a terminologist’s work, which, I believe, is very important for the evaluation of the method. A valency theory called the Pronominal Approach will be implemented to provide a more focussed approach than would be possible with a simple list of verbs. The basis of my method is to examine whether the valency patterns of a number of verbs are suitable as knowledge probes and thus also as search patterns in corpus-based terminology work. Further, I shall discuss how and to what extent the Pronominal Approach offers an efficient and suitable search method for practical terminology work based on machine-readable corpora consisting of raw texts that are neither lemmatised nor tagged. Finally, I look briefly at some aspects related to verb diathesis, semantics, LSP readings, and other linguistic signals than valency patterns as well as the importance of applying authentic corpora examples.