학습자 말뭉치를 이용한 한국어 학습자 오류 분석 연구

Abstract

Kim, Yu-Mi. 2002. A study of Error Analysis of Korean learners by using ‘Learner Corpus’. Teaching Korean as a Foreign Language, 27, 141～168. Developing Korean learners' corpus of Korean as a foreign language and analyzing its learners' errors enable to figure out the properties of their interlanguage. However the previous studies are lack of researched without many relevant materials or resources, so it was hard to draw some generalization. In addition, they limit the themselves in reviewing or analyzing the restricted linguistic facts by disregarding various data, but focused on small numbers of linguistic materials. In order to overcome this difficulty, a corpus can be used ; it does not only retain a large amounts of materials, but also studies diverse linguistic capabilities in Korean studying. To develop Korean learners corpus, it needs to collect materials such as writing tests and papers from language education centers of three domestic universities and develop a raw corpus by entering those materials in the database system. Furthermore, it is required to mark up the personal information for the raw corpus, and tag the grammatical annotation for Korean education. For this, it is necessary to create a collection or dictionary of Korean errors in usage, attach error tags and classify them. First, researching the occurrence rate of errors and the usage frequency by part of speech should be surveyed. Then the errors by the spearkers of different mother tongues such as English and Japanese can be analyzed. The particles among lexical categories are investigated for the analysis of errors. Case particles, auxiliary particles, and conjunctive articles are included. Case particles such as nominative, objective, possessive, and adverbial are divided by omission, addition, and replacement to analyse error patterns. The main factors of errors are interference of mother tongue, simplification, and the overgeneralization affected by the compound of target language structures. Besides, avoidance is regarded as a psychological factor for the errors. To determine how the error patterns of particles are related with learners' level of the study in Korean. We analyzed the data of Korean learners based on the Japanese language culture. The errors rate decreases as the level goes up to the advanced. In addition, the repeated study is essential sauce the some error pattern shown a the level of the beginners still appears in the advanced level. To prevent the particle errors, it is desirable to provide the beginners with the basic grammar explanations and the example of actual errors to the beginners. For the intermediate and the advanced, the corpus concordance, the information of sentence patterns including particles related to specific verbs can be given in term of the corpus concordance. (KyungHee University)