MACHINE TRANSLATABILITY OF RUSSIAN SCIENTIFIC TEXTS

Abstract

The problem of the identification of text parameters having negative impact on the
quality of machine translation is considered in the article. Definition and classification of
translatability indicators is given. Linguistic phenomena in the scientific texts in Russian
are considered at the graphic, lexical and syntagmatic levels. The special attention is paid
to textual features causing problems for machine translation at the syntagmatic level, including
both universal features and ones specific to Russian. A classification of formal machine
translatability indicators for Russian scientific texts is compiled on the basis of language
features analysis. The distinguished classes are matched against text translation
problems associated with the corresponding indicator. The obtained results can be applied
in translation practice and when developing tools for computational support of translation.