Treating `Free Word Order' in Machine Translation

Abstract

In `free word order' languages, every sentence is embedded in its specific context. Among others, the order of constituents is determined by the categories `theme', `rheme' and `contrastive focus'. This paper shows how to recognise and to translate these categories automatically on a sentential basis, so that sentence embedding can be achieved without having to refer to the context. Modifier classes, which are traditionally neglected in linguistic description, are fully covered by the proposed method. (Coling 94, Kyoto, Vol. I, pages 69-75)