Earticle

Nous présentons ici une étude exploratoire et empirique de certains aspects du processus oral et écrit de la traduction à l’intérieur de ce qui construit la compétence traductrice.L’objectif est double : la recherche d'essais d'habilités spécifiques de la compétence traductrice dans des tâches de traduction et la comparaison des données obtenues en traduction à vue et en traduction écrite afin de vérifier empiriquement si la traduction à vue peut être considérée en tant que modalité d'interprétation. L’expérience s’est appuyée sur un échantillon de 22 étudiants de Traduction et Interprétation inscrits en dernière année universitaire, une fois passés leurs examens finaux.Les tâches ont consisté en une traduction à vue et une autre écrite du même texte et, ensuite, cet échantillon d’étudiants a répondu à un questionnaire rétrospectif sur les causes des problèmes de traduction et les stratégies employées pour résoudre ceux-ci. Pour la révision des traductions, trois critères ont été établis: non-sens, faux-sens et contre-sens. Les résultats démontrent des différences qualitatives et quantitatives selon les problèmes, les causes et les stratégies employées. La quantité d' stratégies détectée était supérieure dans la traduction à vue. Les résultats qualitatifs ont été fructueusement incorporés au modèle théorique de la compétence traductrice du groupe PACTE. Il a également été observé que les composantes psychophysiologiques présentent plus d'incidence en traduction à vue, tel qu'il arrive dans les modalités de l'interprétariat. La compétence du transfert ayant trait à la re-expression est plus relevante dans la traduction écrite.

Artificial intelligence translation programs have achieved a high level of reliability with respect to the accuracy of their translation output. In this regard, the goal of this research is to investigate to what level the AI translation program, especially the Google translation program, has advanced so far, and how soon the translation program will take us to the next level. To test the accuracy and reliability of the Google translator’s output, numberous Korean data were used for input including those which were carefully made up or collected to elicit lexical and syntactic ambiguities. The tests conducted in this study are aimed at identifying the types of errors that the Google translation program produces. The findings of the tests show that the Google translator makes various types of errors including lexical, syntactical and idiomatic ones. Lexical and syntactical errors are found to be made in the translation of ambiguous sentences in which contexts human translators are unlikely to make any translations errors, which is attributable to the Google translator's inability to understand non-linguistic cues or information. These results indicate there still remain a number of areas in translation that require human intuition and intervention.

Translation quality is a key to discussing translation, and it is essential to have the criteria that help determine a “good” target text in order to discuss translation quality. Despite endless debates on the issue, the academic community has yet to agree on explicit standards that a “good” target text needs to satisfy. Therefore, translations have been assessed according to subjectively-set criteria. This paper is a preliminary report on designation of different weights by translation error category. To devise a clear and consistent translation evaluation model, it is vital to first set adequate categories for evaluation according to directionality i.e. Japanese into Korean and vice versa. Therefore, the study attempts to establish evaluation criteria for Korean-Japanese technical translations by using Translation Evaluation System(TES), which a Japanese company, Hitachi Ltd. utilizes to evaluate Japanese-English technical translations. Once the criteria were determined, the target texts produced by the paper's subjects (students) were examined to sort identified errors by the categories. The errors were then sent to native speakers of Japanese for a five-step assessment based on naturalness of expressions and readability. By designating different weighted value to each error category, this study aims to contribute to a more systematic and quantitative evaluation of translated texts.

Stanislavsky's Complete Works of 8 Volumes are 'actors’ bible' in which Stanislavski's acting method is fully explained. However, the acting book series widely known in Korea as Stanislavski's collections are based on double translation of the English, Japanese, and French translations of the 2nd, 3rd, and 4th volumes of the Russian collection. The problem is that these double translations have a lot of mistranslation, distortion, and deletion compared to the Russian original, which is an obstacle to properly understanding the Stanislavsky acting system. Therefore, this study aims to compare and analyze the Russian original, English translation, and Korean double translations of the English translations, focusing on the 2nd volume, to determine which translation errors from double translation are included in the Korean version and to examine what kind of misunderstanding these translation errors can lead to in understanding the Stanislavsky system.

This study compares how Papago, Google Translate, and ChatGPT perform when translating police dialogues, using BLEU scores, manual assessment, and error analysis. The analysis results are as follows. All three tools showed low BLEU scores compared to reference translations. In manual evaluation, ChatGPT significantly outperformed others, scoring 8.6 out of 10, compared to Google Translate's 6.5 and Papago's 5.6. Error analysis confirmed ChatGPT's superiority, with only 41 errors, while Google Translate and Papago produced 123 and 152 errors respectively. Across all tools, substitution was the most common accuracy error, followed by omission and addition. These results suggest that ChatGPT is the most reliable tool for police communication with foreign nationals.

This paper aims to categorize the errors of the Korean-Russian machine translation and establish pre-editing rules to improve the quality of machine translation. Result shows that types of errors include vocabulary and context error, typo, others, verbal error, (vocabulary/phrase) omission, (vocabulary/phrase) structural error and an error on a proper noun and number/time/date error in following order. Based on these, 10 pre-editing rules have been established and suggested. 1) The formal name must be used instead of an abbreviation for a particular organization. 2) numerical writing methods which use Korean units must be avoided. 3) address must be divided for each unit by using a comma. 4) in the case of date, correct information must be provided to identify the context such as year, month, date and etc. 5) the usage of polysemy or homonym must be avoided. 6) the meaning must be clarified. 7) the usage of dependent nouns which can cause confusion in identifying the context must be avoided 8) addition of unnecessary information(symbol) must be avoided 9) abbreviated subject must be clarified. 10) the sentence structure must be simplified. As a result of using these rules, it has been confirmed that the errors of machine translation have been significantly improved.

The purpose of this study is to investigate punctuation translation strategy of the Korean-Russian machine translation and categorize the types of its errors. This study is also aims to seek practical implications for improving product quality of machine translation by establishment of pre-editing rules and its application. For this goals, corpus of articles in KBS World Radio were established, and quantitative and qualitative analyses were carried out. The analysis results showed that both the punctuation translation strategy of machine and of human were almost same. However, the punctuation error rate by machine translation was 24.19% while the error rate by human translation was only 2.4%. To determine the cause of this, only the errors about punctuation by machine translation were classified, and the translation strategy was addition > shift > literal translation > omission. The type of punctuation marks causing errors in the source texts by machine translation was revealed. Then, the errors by machine were categorized by punctuation marks, and Pre-editing rules were created based on this. In addition, to verify the rules’ utility, the source texts and the texts translated by machine, the pre-edited source texts and the newly-produced texts by machine were comparatively analyzed. The analysis result showed that only the pre-editing rules applied to punctuation marks considerably improved sentence structure, grammar, omission, mistranslation and non-sentence.

The paper aims to demonstrate the usefulness of incorporating corpus search techniques into Korean-English translation classes where English is the students’ B language. To this end, the paper frist gives a brief overview of the literature advocating the utilization of corpora in translation trainee training. After that, it focuses on COCA (the Corpus of Contemporary American English) as the most useful large-scale public English corpus for the present study’s purposes, offers a summary account of its key search functions and showcases a variety of search strategies as they are applied to resolving three types of collocational errors, using examples drawn from the author’s translation classes.

While translation is basically converting one written language into another, it can be described as rewriting. Moreover, the competence of an English-into-Korean translator can be measured by his/her ability to avoid reviving some aspects of English grammar in their translation output. Resorting to the strategy of ‘form-based’ or literal translation is generally unfeasible. This research deals directly with the subject of translation errors. In this analysis, translation errors beyond-thesentence have been more closely examined and applied to real cases of communication breakdown as well as to successful efforts. The aim of this research focused on the output of translator trainees aimed to discover whether errors found in the English-into-Korean translations form patterns. The results of this investigation clearly reflect the fact that there are error patterns: among these, sense consistency with the original message, lexical decision and logical cohesion are considered the most important parameters in translation quality assessment. This study culminated in a final exam that included translation errors made by seventeen Korean undergraduate students of translation studies when they performed an English-into-Korean non-literary translation task involving a New York Times article entitled “Poll Finds Faith in Obama, Mixed With Patience”, published in January 18, 2009.

Malgré l’intérêt qui ne cesse de croître en Corée pour la traductologie, l’une de ses sous-disciplines, la critique de la traduction, en est encore à ses balbutiements. Pour garantir la qualité de la traduction dans une société donnée, il est essentiel que cette société puisse faire appel à une solide tradition de critique de la traduction. Cet article propose donc un modèle à la fois complet et systématique pour assurer un bon niveau d’évaluation des traductions et, par extension, un modèle de critique de la traduction puisant dans les riches perspectives offertes par la traductologie. La première partie consiste en un résumé de l’état actuel des recherches dans le domaine de la critique de la traduction en Corée, suivi d’une analyse des écueils qui guettent chaque filière adoptée en critique de la traduction. Ensuite, trois textes sont analysés afin de montrer la nécessité d’incorporer les perspectives offertes par la traductologie dans la critique de la traduction. En conclusion, on appelle les chercheurs en traductologie à participer de façon plus active aussi bien au développement de modèles qu’à la pratique dans le domaine de la critique de la traduction.

Due to globalization and the development of multimedia technology, the use of Korean-English subtitled movies and dramas has dramatically increased around the world, and especially through the Internet. However, the importance of Korean- English subtitling has been under appreciated in the field of translation. It is not difficult to find grammatically misused forms in Korean-English subtitled materials. The purpose of this paper is to highlight the importance of Korean-English subtitling and the necessity of an academic use to subtitling and translating tasks. The subtitling errors in a selected Korean movie and drama currently available on Internet websites in the U.S. were collected and analyzed. Errors in he data were then divided into two groups: Grammatical Errors and Lexical Errors. This paper focuses on the grammatical errors, which were subcategorized into six grammatical groups: Articles, Pluralization, Capitalization, Prepositions, Verbs, and Pronouns. This study hopes to help reduce occurrence of the most common errors made by Korean translators in Korean-English subtitling.

The paper discusses the relation between error tendency and translation quality in additional translation with the Chinese-Japanese translation corpus for patent name as a subject of the study. The paper divides errors of additional translation into three classes: Additional translation [missed]: there is no additional translation if the additional translation is necessary; Additional translation [excessive]: there is an additional translation if the additional translation is unnecessary; Additional translation [wrong]: there is an additional translation if the additional translation is necessary, but the additional translation is wrong. The corpus adopted for the study is Chinese-Japanese translation corpus with 676 translated patent documents provided by a translation company in Dalian, including the Chinese original version, Japanese version translated by translators with Chinese nationality and Japanese version proofread by proofreaders with Japanese nationality. The proofreaders graded the translated documents in the corpus into 5 grades based on the translation quality, i.e. translation versions with grade 1, 2 and 3 have low translation quality and with grade 4 and 5 have high translation quality. The corpus consists of 55 translated patent documents with grade 5 (8%), 473 translated patent documents with grade 4 (70%), 130 translated patent documents with grade 3 (19%) and 18 translated patent documents with grade 2 (3%). The study conclusion is as follows: 1. Among the translation errors, errors of additional translation account for more than 50%, take the second place among errors in translation versions with low quality and take the first place among errors in translation versions with high quality. 2. Among the errors of additional translation, missed translation accounts for more than 70% and is positively related to the translation quality; excessive translation accounts for 10%~30% and isn’t related to the translation quality; wrong translation accounts for less than 10% and isn’t related to the translation quality. 3. Among the errors of additional translation, functional word errors account for more than 90% and content word errors account for less than 10%. A Japanese auxiliary word, “no”, takes the first place and isn’t related to the translation quality.

번역가가 영어의 정관사를 잘못 사용함으로 인해 오해가 발생하는 경우는 거의 없다는 것이 일반적인 견해이나, 실제로는 관사를 올바로 사용하지 않음으로써 의미가 변화하거나, 의사소통에 문제가 생기는 경우들도 있다 (Berry 1991:252-253). 본 연구는 번역을 공부하는 한국 학생들의 영어 관사 오류를 분석함으로써 오류의 원인을 알아보고자 했다. 선행연구(Hawkins et al 2006;Kang 2005; IKW 2004; IKW 2003) 에서는 한국어에 [특정성 SR]을 나타내는 문법 요소가 존재하는 지에 대한 가정이 연구자 마다 다를지라도 한국 학생들이 [특정성]을 구분할 수 있다고 전재하고 있다. 그리고, 학생들이 영어에서 관사 선택의 기준이 되는 [청자의 (지시대상에 대한 선행)지식 HK]이 아닌 지시대상의 [특정성] 여부에 따라 관사를 선택한다고 주장한다. 그러나 본 연구에서는 학생들이 [청자의 지식]에 대한 명확한 개념을 습득하지 못했을 뿐 아니라, [특정성]에 대한 개념 또한 영어를 모국어로 사용하는 사람들과 다름을 발견했다. 학생들은 ‘화자’가 지시대상의 범위를 축소 또는 지적하도록 하는 암시적 또는 명시적 신호가 있는 경우 [특정성]을 잘못 판단하게 된다. 본 연구에서는 관사 사용의 오류가 학생들이 정관사 ‘the’를 지시 대상의 범위를 축소하거나 지적하는 표시로 이해하고 있고, 이러한 개념이 청자의 지식에 대한 개념을 포함하고 있지 않다는 것을 인식하지 못함으로 인해 발생한다는 것을 발견했다.

This paper aims to shed light on the distinction of translation competence between professional translators and translation major learners by going over the translation error at lexical level that translation major learners often make in literary translation and comparing them with professional translators’ translation of the same work. In PACTE’s study, translation competence is made up of five sub-competencies as well as psycho-physiological components. The five sub-competence is bilingual sub-competence, extra-linguistic sub-competence, knowledge about translation sub-competence, instrumental sub-competence and strategic sub-competence(2003: 58). Translation error can be used as a standard for measuring translation competence. In this paper, error patterns of translation major learners were classified into five categories: Errors in vocabulary related to culture, Errors caused by interference of Sino-Korean word, Errors due to conceptual meaning, Errors in correspondence between vocabularies, Error in selecting advanced vocabulary. By way of conclusion, this paper had revealed the translation error’s meaning in the distinction of translation competence between two groups.

Terminology translation plays a key role in technical machine translation. This paper conducts quantitative and qualitative evaluation and presents a comprehensive error analysis on terminology translation from Korean to English. For this study, 637 terms are selected from a patent domain corpus. Among them, 425 terms are translated correctly while other 212 terms have one or more errors. Those identified errors are classified into seven error categories which include addition, omission, lexical errors, order errors, morphological errors, partial errors, and miscellanies. Among them, partial errors and morphological errors are the two top error categories, and lexical errors and omission cases are also high. The study also presents n-gram wise error rates. The calculation of n-gram wise terminology translation error rates shows that the curve for the Google Translation results goes upwards with the increasing size of n. This shows that neural machine translation can be vulnerable in translating a higher-order of n-gram terms. This study has implications for Korean-English machine translation research, in that it classifies error cases spotted in machine translation of terminology which is critical in securing accuracy of translations.

오류 판정은 텍스트의 유형에 따라, 평가자가 누구인지에 따라 달라질 수도 있으며, 시대에 따라서도 달라질 수 있다. 그럼에도 불구하고 번 역 결과물의 오류에 관한 지속적인 분석 연구가 필요하고, 외국인 유 학생들에게 다양한 오류 분석의 틀을 보여 각각의 장점과 한계점에 대 해 지도할 필요가 있다. 이에 본고는 인간 번역에서 기계번역에 이르 기까지 최근의 번역오류 연구 동향을 살펴 외국인 유학생을 위한 번역 교육의 기초 자료를 마련하는 데 목적을 둔다. 혼재되어 있는 오류 유 형화의 틀을 범주 별로 분석해 보고 문제점을 분석하여, 한국어 번역 오류 연구의 방향성을 고찰해 보고자 함이다. 이를 위해 먼저 서두에 서 문제를 제기하고, 이어 번역 품질평가 측면에서 오류가 어떻게 유 형화되는지를 보았다. 다음으로 번역 전략적 관점에서 오류 분석이 어 떻게 이루어지고 있는지를 비문학, 그리고 문학 영역으로 나누어 살펴 보았다. 다음으로 최근 주목 받고 있는 기계번역의 오류 분석에 대해 살펴보았다. 마지막으로 논의를 통해 한국어 번역 오류 유형을 체계화 하여 제시하였다. 최근 널리 인용되고 있는 Vilar 외(2006)의 분류를 실마리로, 형태, 통사적 접근, 의미, 화용적 접근, 텍스트 차원의 접근 등 세 개의 대분류 아래 한국어의 언어학적 특성에 맞는 세부 분류를 두었다.

The purpose of this research was to examine the typology of translation errors in translation studies published to date. An increasing number of international students are studying Korean translation errors, and this study was an endeavor to develop a framework to assist scholars with their analyses of translation errors. This study first determined how errors were classified in terms of translation quality evaluation, and then examined how error analyses were performed from a translation strategy the perspective. The data were divided into non-literary and literary domains. The error analysis of a machine translation was subsequently scrutinized, comprising an error censorship of the machine translation system itself, as well as an analysis of human errors. Researchers’ views and discussions on the kind of analysis framework that would serve the purpose best, had to be incorporated. Determining and classifying errors is not an easy task, and continual analytical studies are consequently needed on errors in translation results. To do so necessitates that the typification of errors should first be done, and this study makes a significant contribution in that respect. I believe that this study will be helpful to foreign students studying Korean translation errors.

The aim of present research is to examine translation errors that occur with Korean to English translations on Korean statues in a public place located in Seoul. The authors collected data of the English information boards on the statues three times during the period of 2013-2019, which included August 2013, March 2015, and February 2019. The translation error patterns of these three different versions on English information boards were compared. Through careful investigations, the format errors, grammatical errors, and contextual errors were detected in the revised versions as well as in the original version. Furthermore, it was also found that there was little improvement with the revised versions of English information boards. As a result, this paper proposes that thorough preparations and planning should be implemented when information boards on statues are translated from Korean into English.

Jinnamgwan in Yeosu was designated as the 304th Korean national treasure in 2001. The Korean name of the national treasure can be translated into English as Jinnamgwan in Yeosu. Jinnamgwan is the name of the building and Yeosu is the place where the building is located. This paper examines the information board of Jinnamgwan. The translator of the information board seems to have had difficulty in comprehending the context, since many difficult words are used in the Korean information board and the Korean sentences are very difficult to understand. Furthermore, there are many words which are usually used only by construction specialists. These seem to be the main reason for the errors made in translation. Some measures must be taken to prevent future errors in the translations. Possible suggestions include presenting the translator's name or contact information of the managing office. Next step will be to recruit and train translation specialists and have them translate the material.

This paper analyses the errors due to Korean interference when writing/translating into Spanish. We claim that, when learners once input into memory a Spanish form “α” that most preferentially corresponds to a Korean form “A,” they tend to equate this correspondence as ‘A = α’ and apply it to cases where such correspondence should not be used. To find out how many learners make this type of error we conducted an experiment with 42 students. In the experiment, we divided the sentences into two groups. The sentences that can produce an appropriate result were placed in group A if we use the correspondence ‘A → α’, whereas those that can produce an inappropriate result were placed in group B if we do the same. Our analysis showed that the most common error in group B is to reuse the correspondence used in group A. Therefore, learners need to avoid an “excessive equation” to reduce errors. Finally, for each sentence, we suggested ways to reduce or avoid errors.

Este trabajo analiza los errores debidos a la interferencia coreana al escribir/traducir al español. Sostenemos que, cuando los alumnos ingresan una vez en la memoria una forma española α que corresponde preferentemente a una forma coreana A, tienden a equiparar esta correspondencia como ‘A = α’ y la aplican a los casos en los que dicha correspondencia no debería usarse. Para averiguar cuántos alumnos cometen este tipo de error, realizamos un experimento con 42 alumnos. En el experimento, dividimos las oraciones en dos grupos. Las oraciones que pueden producir un resultado apropiado se colocaron en el grupo A si usamos la correspondencia ‘A → α’, mientras que las que pueden producir un resultado inapropiado se ubicaron en el grupo B si lo hacemos. Nuestro análisis mostró que el error más común del grupo B es reutilizar la correspondencia utilizada en el grupo A. Por lo tanto, los estudiantes necesitan evitar una “ecuación excesiva” para reducir errores. Finalmente, para cada oración, sugerimos formas de reducir o evitar errores.