초록 열기/닫기 버튼
이 글은 한국어 구어 담화 연구에 적용하기 위한 대화 행위의 주석 체계를 논의하는 데 그 목적을 둔다. 형태와 통사 주석은 마련된 것이 있지만 담화 차원에서의 주석 체계에 대해서는 아직까지 명확한 기준을 제시한 연구가 부재하기에 본 연구에서는 한국어 학습자 말뭉치 구축에 있어서 담화 차원의 대화 행위 주석 체계를 실험적으로 모색해 보고자 한다. 아울러 대화 행위 주석 체계 정립의 가능성을 타진하고 주석 작업의 신뢰도를 높이기 위한 단계적 절차 수행 방안에 대해 논의한다. 그리고 대화 행위 주석이 실제로 어떠한 연구로 활용될 수 있는지 보이기 위해 주석이 부착된 자료를 활용하여 일상대화에서의 담화 전개 양상을 분석해 보이고 그것이 한국어 담화 교육에서 어떻게 적용될 수 있는지 논의한다. 합리적 차원의 대화 행위 주석 체제를 마련하는 것은 다차원적이고도 역동적인 담화 의미를 규명하기 위한 출발점이 된다. 또한 학습자 구어 말뭉치의 구축 과정에 이를 적용함으로써 교수·학습과 교재 및 교육 자료 개발에 기여함은 물론 국제 학계에서 한국어 담화 분석과 한국어교육학의 효율적 소통을 기대할 수 있을 것이다.
This paper focuses on discussing the dialogue act annotation system that can be applied when studying about Korean oral discourse. Since 1990’s, in the field of discourse analysis, an objective research on investigating the characteristics of Korean oral discourse, started to take place. And since the start of the 21st century, Korean language teaching academia began to actively do analytical research on the oral discourse of Korean language learners. More recently, At a national level, a task of establishing and utilizing Korean language learners’ corpus is in progress. In order for a large national corpus research to be applied continuously in future researches, developing a annotation system that can effectively handle spontaneous information is necessary. Even though there are forms and syntax annotation system that are already developed, we still lack researches that clearly defined the standard for a annotation system at a discourse level. Therefore, this research seek to develop a dialogue act annotation system at a discourse level through experiments. The foundational theories that helped develop a reasonable Korean dialogue act annotation system were the various researches done by different corpus language departments, but the main reference that was used was the results of research and development by ISO. In first experiment, we provided 14 tags and requested 2 inexperienced taggers to do a trial conversation. Using the problems and implications as base, we added an additional tag and performed the second trial using 15 tags. To increase the credibility on the second experiment, we broke down each stage of the trial. First, the taggers performed their own test. Second, they cross examined each others results and finally gave chance to discuss and revise the results together. If the taggers cannot come to an agreement, one of the skilled researchers will listen to the sound file and judge the final tag. Through the experimental process, we were able to further examine the possibility of establishing a dialogue act annotation system, and discuss ways to increase the credibility by separating each part of the experiment stages. And to find out how dialogue act annotation can be utilized in other experiments, we applied the materials that includes the tags to daily conversation and analyzed the developing aspect of conversation. Using the result of the analysis, we further examined how it can be applied to teaching Korean discourse. Establishing a rational dialogue act annotation system is a starting point to deeply understanding the multi-dimensional and yet more dynamic meaning of the corpus. Also, by utilizing the information when establishing learner’s oral corpus, it will not only contribute to developing teacher and student learning materials, but can also expect more efficient communication on analytical Korean discourse and Korean linguistic education at international academia.
키워드열기/닫기 버튼
oral language corpus, learner’s oral language corpus, dialogue act, dialogue act annotation, Korean language teaching and learning, teaching of Korean discourse.
