The Importance of Syntagmatic Dimension in Multilingual Lexicography

Ruta Marcinkeviciene
Vytautas Magnus University
Kaunas, Lithuania
e-mail: ruta.marcinkeviciene@vdu.lt

The paper deals with the philosophy behind a software tool meant as a multilingual DB of nouns from 8 languages (Lithuanian, Russian, German, English, Polish, French, old Greek, Latin), called Muldi (Multi Lingual Database Interface). Abstract nouns (AN) form the bases for the DB since they belong to the category of those lexical items with fuzzy meanings, and it is only in connection with their contexts that these words acquire a concrete sense. Besides, abstract nouns cause more problems for translators, as their usage and especially metaphorical patterns are different in different languages.

The tool is designed so as to give a possibility to: view a list of AN in each of the languages, view a list of equivalents of a chosen word from the source language, view the monolingual concordances of the selected AN from monolingual corpora, view aligned concordances from parallel corpora, view noun record or descriptions based on concordances. The main difference of Muldi in comparison with other publicly known lexical DB lies in incorporation of monolingual and parallel concordances based on respective corpora.

It is argued here that patterns of usage as detected in corcordances, esp. collocational profiles, are of paramount importance for identification and presentation of translation equivalents (TEs) since most of translations of AN are caused by larger semantic units than isolated words. This observation is based on the analysis of the usage of the English noun MIND and its TEs in Lithuanian.

A comparison of two lists of TEs from the bilingual dictionary and the Lithuanian translation of George Orwell’s novel 1984 is made in order to show the discrepency between them. Since TEs are mostly influenced by contextual partners in general and metaphorical patterns of MIND in particular concordances are very helpful in providing the explanation for seemingly accidental TEs.

See previous, next abstract.