한국어 의미. 화용적 특성에 의한 영한기계번역 오류 연구

Abstract

Abstract Thanks to the tremendous improvement in the quality of simple results of machine translation, which is much faster than humans, the trend to use it has become more and more common, having been established to some extent. However, due to the enormous amount of equivalent variables in translation, it is still difficult to guarantee the safety of its quality with lots of missing parts. It also means that the language expressing human thoughts is complex, and the best alternative and most widely used solution to overcome this problem is Machine Translation Post-Editing (MTPE). Therefore, post-editing can be said to be a collaboration between humans and machines. Nevertheless, it is a disprove that the role of human translators is still of consequence. Then, irrespective of the personal ability of the translator, it is important for the translator to apply certain linguistic characteristics according to the characteristics of Korean native language in the post-editing of machine translation. This study focuses on the linguistic characteristics of English, which can be considered the most common language among numerous languages, and Korean among them, and examines what errors occur during machine translation, limited to the case of translating English into Korean (target language) by type. After examining it, we decided to focus on what unique characteristics of Korean should be differentiated from English and applied during post-editing. To this end, this study was conducted as follows: From Chapter 2, we investigated semantic and pragmatic characteristics of Korean as a preceding study prior to machine translation. First, through the history and research direction of Korean semantics, various categories of semantics are considered, and lexical forms, individual words, complex vocabulary expressions and meanings, compound words, idiomatic expressions, collocations, redundant expressions, and vocabulary borrowing as well as grammatical forms have been examined. The grammatical categories, postpositions and endings according to the meaning of the phase, the meaning of the corresponding expression, besides, the meaning of the sentence were identified. Next, the definition of Korean pragmatics, research fields and major issues were grasped, and furthermore, an in-depth analysis was conducted by linking pragmatics and Korean language education with lexical and grammatical theories. Based on this, the lexical and morphological characteristics, syntactic characteristics, and semantic pragmatic characteristics of Korean seen in machine translation were analyzed. From Chapter 3, we first looked at the development process of machine translation, divided its error analysis into foreign and domestic preceding studies, and then analyzed actual errors according to error types. In addition, in this chapter, after reviewing the concept and role of post-editing first, the preceding research and guidelines were determined through the transword selection method, and post-editing work was performed through error analysis of actual machine translation. In this study, machine translation, post-editing, and semantic and pragmatic characteristics of Korean were combined to roughly apply real cases. However, in the future, the difference between English-Korean and Korean-English, concrete cases of post-editing for actual foreign Korean learners, structural differences between Korean and English, and grammatical differences due to linguistic characteristics, including cognitive and pragmatic differences, follow-up studies such as verification work through detailed comparative analysis of translation and post-editing will be necessary.