The current article discusses the main problems of human and machine translations, as well as introduces a new lexical description in machine translation for faster and more accurate translation. The new method uses so-called field indicators or the Field Indices to facilitate the MT search engine for words by marking these words with special components creating a semantic field, and allowing the MT devices to search for the word according to its usage in the text. The Field Index system covers the semantic description of the following main spheres: scientific field, public or social field and humanitarian field. These three spheres contain subfields that usually mingle with the parent index through a dash, with the parent index being with the first one. The scientific field includes such aspects as geography, mathematics, chemistry, physics, economy, medicine, etc., with related subfields like diseases and biological terms (for medicine), names of drugs (as a separate filed), finance and accounting (as part of economy), etc. Applied Sciences Index contains miscellaneous words used by other subfields of the same scientific sphere: e.g., computer, telephone, function and many others, and plays a crucial part in distinguishing polysemantic words such as mouse (hardware), root (in mathematics), etc. The public or social field contains subfields that narrow the meaning of words to a specific one and includes aspects as art, agriculture, law, education, religion, housing utilities, time, transportation, people, etc.i+iT Ss the widest semantic field containing a lot of subfields specifying words that belong to such groups as colours, architecture, games, music, sport, etc. (for art index), clothing, beverages, food and production (for agriculture index). Notions of time, people, professions and terms for religion and items used in household are present in this group due to its wide usage within the society. The humanitarian sphere deals mostly with terms used in languages, literature, manuscripts and libraries. These subfields help identify polysemantic words between nations and languages, book titles and ordinary words and phrases, and literary styles (documentaries, fairy tales, dramas, etc.). In turn, these can be further defined as prose or a poem. All formulae proposed in the project consider the presence of the Field Indices and its position at the end of the description of the word. Depending on the target language, the translated version should be identical with the source following this very principle.
Read full abstract