본문 바로가기
  • Home

A Study on a Case of Anti-Feminism in Neural Machine Translation and its implications for Web as Corpus

  • The Journal of Translation Studies
  • Abbr : JTS
  • 2021, 22(5), pp.299-326
  • DOI : 10.15749/jts.2021.22.5.011
  • Publisher : The Korean Association for Translation Studies
  • Research Area : Humanities > Interpretation and Translation Studies
  • Received : November 7, 2021
  • Accepted : December 12, 2021
  • Published : December 31, 2021

Jee, Yoon-Ju 1

1한국외대

Accredited

ABSTRACT

The purpose of this study is two-fold: (1) to show how neural machine translation (NMT) translates online newspapers about the anti-feminism phenomenon in South Korea and (2) to examine the co-occurrence of the results by using Web as Corpus, Sketch Engine. This study aims to examine the anti-feminism phenomenon in three commercial neural machine translations (Google Translate, Naver Papago, and Kakao I-NMT). Findings show that three neural machine translation engines are generally prone to anti-feminism translation. For example, they translate ‘여성 우월주의 (female supremacy)’ into ‘feminism’. And then, in Web Corpus using Sketch Engine tools, the co-occurrence around ST keyword ‘여성 우월주의’ at the negative label can be observed in the lexical distributions of ‘feminism’. As a result, this anti-feminism in neural machine translation is considered to reflect negativity toward feminism in online anti-feminist communities. This paper concludes with a brief discussion of ethical questions related to both machine translation and technology in translation studies.

Citation status

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