Towards a Lexicographic Approach to Lexical Transfer in Machine Translation (Illustrated by the German–Russian Language Pair)

  • Authors:
  • Igor Melčuk;Leo Wanner

  • Affiliations:
  • Department of Linguistics and Translation, University of Montreal, C.P. 6128 ``Centre-Ville'', Montreal, QC, Canada H3C 3J7 E-mail: melcuk@ling.umontreal.ca;Department of Computer Science, University of Stuttgart, Breitwiesenstr. 20–22, 70565 Stuttgart, Germany E-mail: wanner@informatik.uni-stuttgart.de

  • Venue:
  • Machine Translation
  • Year:
  • 2001

Quantified Score

Hi-index 0.00

Visualization

Abstract

The translation of lexical items is still a formidableobstacle in the field of Machine Translation. The present articleaddresses this problem from the perspective of modernlexicography, putting forth detailed monolingual lexica whichcontain highly elaborated descriptions of lexical items. Leavingaside the resolution of homonymy on the source side and theselection of the appropriate synonym on the target side, weconcentrate on the case of ``multiple lexical correspondences'',i.e., mismatches between monosemic lexical items. We proposeto use the lexical transfer methodology. Our proposal exploits afine-grained semantic decomposition of the meaning of both sourceand target lexical items as well as information on ``actantialvalency'' (semantic restrictions) and on collocations. The lexicaldescriptions are formalized so as to allow for automatic matchingthereof. We present a rule-based procedure that selects the mostappropriate translation equivalent of a source lexical item byusing matching. The proposal is illustrated drawing on threeGerman and three Russian monosemic speech-act nouns; their lexicalentries are given in an appendix. Two examples of lexical transferbetween these lexemes are analyzed in detail.