본문 바로가기
  • Home

Analysis of Neural Machine Translation Meaning Errors in Chinese Color Terms “黑” -Focus on translation of a sentence with the adjective “黑”

  • Cross-Cultural Studies
  • 2022, 66(), pp.69-99
  • DOI : 10.21049/ccs.2022.66..69
  • Publisher : Center for Cross Culture Studies
  • Research Area : Humanities > Literature
  • Received : May 4, 2022
  • Accepted : June 2, 2022
  • Published : June 30, 2022

Lee, Seo-yi 1 Han, Yong Su 1

1동국대학교

Accredited

ABSTRACT

In this paper, neural machine translation semantic errors in sentences with the Chinese color term “黑” were analyzed. A total of 21 words, such as “黑暗”, “黑沉沉”, and “黑咕隆咚”, were studied as adjectives of “黑”. As a semantic feature of these words, first, black color itself was described. Second, it referred to a dark state because there was no light. Third, it was used as a metaphorical expression that generally symbolized the evil mind and negative things of humans. Then 43 Chinese example sentences using the adjective “黑'” were translated into Korean using neural machine translation. The first feature of the translation of the color term '黑' type adjective was that the basic meaning of “黑” used describing an object, space, and the dark state of light was relatively natural. The second feature was that when an adjective was used in a metaphorical sense, especially when it represented a person's psychology or behavior, there were many errors in translation. The third feature was that the length of the sentence did not significantly affect the results of translation. It was found that semantic errors were frequent in short-form translations, which had less information. Machine translation continues to expand its use from text translation to real-time translation. Therefore, continuous research on the error of machine translation is expected to contribute to both the development of related technologies and the effectiveness of their use.

Citation status

* References for papers published after 2022 are currently being built.