This study aims to analyze the error patterns in Vietnamese–Korean machine translation (MT) outputs and the MTPE performance of Vietnamese learners of Korean in order to provide foundational data for Korean translation education tailored to Viet...

http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
https://www.riss.kr/link?id=T17551724
서울 : 한국외국어대학교 대학원, 2026
학위논문(석사) -- 한국외국어대학교 대학원 , 국어국문학과 , 2026. 8
2026
한국어
기계번역 ; 기계번역 포스트에디팅(MTPE) ; 의미 층위 오류 ; 베트남인 한국어 학습자 ; MTPE 수행 양상 ; 한국어 번역 교육
895.1 판사항(22)
서울
A Study on the MTPE Performance of Vietnamese Learners of Korean : Focusing on Semantic-Level Errors in Vietnamese–Korean Machine Translation
v, 129 p. : 삽도 ; 26 cm
한국외국어대학교 논문은 저작권에 의해 보호받습니다.
지도교수: 홍종명
참고문헌: p. 107-117
I804:11059-200001022270
0
상세조회0
다운로드다국어 초록 (Multilingual Abstract)
This study aims to analyze the error patterns in Vietnamese–Korean machine translation (MT) outputs and the MTPE performance of Vietnamese learners of Korean in order to provide foundational data for Korean translation education tailored to Viet...
This study aims to analyze the error patterns in Vietnamese–Korean machine translation (MT) outputs and the MTPE performance of Vietnamese learners of Korean in order to provide foundational data for Korean translation education tailored to Vietnamese learners.
To achieve this objective, a two-stage study was conducted. In the first stage, a total of 323 sentences were selected from three text genres commonly encountered by learners—airline announcements, bank product guides, and news articles from the Nhan Dan newspaper—and translated using Google Translate and Naver Papago. The resulting translation errors were analyzed using the error classification framework proposed by Han and Kang (2022). In the second stage, a survey instrument was developed based on the semantic-level errors that occurred most frequently in the first stage. The survey was administered to 70 Vietnamese learners of Korean to examine their MTPE performance across three stages: error detection, cause identification, and error correction.
The results showed that a total of 862 errors were identified in the Vietnamese–Korean MT outputs, with semantic-level errors accounting for the largest proportion. Among the subcategories, functionality errors, fluency errors, and clarity errors were particularly prominent, indicating that the limitations of MT are more evident in functional appropriateness and natural meaning transfer than in accuracy at the form level. Furthermore, the frequency of MT errors did not correspond directly to the level of difficulty learners experienced during MTPE. While clarity errors were relatively easy to detect and correct, fluency errors proved to be the most difficult to identify regardless of learners' Korean proficiency levels. Moreover, error detection, cause identification, and error correction functioned as distinct cognitive processes. Even when learners successfully detected an error, they often struggled to accurately identify its cause or produce appropriate corrections, a tendency that was particularly evident in cohesion errors and functionality errors. Finally, a comparison of MTPE performance across Korean proficiency levels revealed that improvements in Korean language proficiency did not necessarily translate into improved MTPE performance. Although advanced learners demonstrated stronger error detection abilities, they also showed a greater tendency to introduce new errors during the correction process.
Based on these findings, this study proposes several pedagogical directions for MTPE instruction for Vietnamese learners of Korean, including collocation-based fluency instruction, step-by-step instruction that distinguishes the stages of error detection, cause identification, and error correction, explicit instruction on genre-specific terminology and error types, and differentiated instructional approaches that reflect learners' proficiency levels.
By going beyond the analysis of machine translation errors and empirically demonstrating how Vietnamese learners of Korean detect, understand, and revise semantic-level errors, this study contributes to Korean language education by providing foundational data for establishing the content and instructional priorities of MTPE instruction.
목차 (Table of Contents)