한국언어과학회

This study was to analyze errors of machine translations(MTs) of the texts for middle and high school students, and to offer guidelines on how to use them. To this end, total 4760 machine-translated sentences from two test-preparing books and one novel for middle and high school students were analyzed to 5 types of errors. 5 types errors used were ‘incorrect meaning errors’, ‘context errors’, ‘omission error’, ‘grammar errors’, and ‘typing error.’ The results of the analyzation showed that Naver and Google Translations had 3.8% error and 96.2% proper sentences. The most frequent error was ‘incorrect meaning errors’, followed by ‘omission error’, ‘context errors’, ‘grammar errors’, and ‘typing error’ in high frequent order. Naver and Google machine translations used in the study offered 96.2% proper sentences, so middle and high school students can compare their own English sentences with machined translated sentences. Moreover with the knowledge of errors of MTs, students can correct errors of MTs and use them properly.