MORPHOLOGICAL ANALYSIS OF THE MECHANICALLY TRANSLATED GEORGIAN LANGUAGE,

Abstract

The article points out that one of the difficulties in morphological analysis in machine translation from Georgian is that prefixation as well as suffixation characterize this language. Putting the bases plus prefixes in the dictionary increases the extent of the lexicon. It is proposed to separate the prefixes. Here new problems arise: the spelling of a prefix sometimes coincides with the beginning of some base; a word may have 1-3 prefixes; the base or another prefix may follow a prefix. The base or prefix and a number of prefixes is thus analyzed: if a word form and a dictionary unit are congruent the word contains a base, otherwise a prefix. If a word has several prefixes the operation continues until the part of the base obtained after chopping off the prefix differs from a dictionary unit. The morphological analysis of Georgian is also complicated because, despite its agglutinative nature, homonymous affixes do occur, i.e., the same affix expresses several meanings. The way to differentiate their meanings is to find their place in the chain of affixes. In cases where morphological analysis cannot determine the meaning of a homonymous affix the question is settled by syntax.

Document Details

Document Type
Technical Report
Publication Date
Mar 27, 1967
Accession Number
AD0653954

Entities

People

  • L. Thamarashvili

Organizations

  • National Air and Space Intelligence Center

Tags

DTIC Thesaurus Topics

  • Applied Computer Science
  • Computational Linguistics
  • Computational Science
  • Computer Languages
  • Dictionaries
  • Formal Languages
  • Language
  • Linguistics
  • Machine Translation
  • Natural Language Processing
  • Translations
  • Words (Language)

Readers

  • Computational Linguistics
  • Educational Psychology

Technology Areas

  • AI & ML
  • AI & ML - Machine Translation