2019 Volume 173 Pages 16-30
This study investigates Japanese learners' language use in speaking and writing tasks on a particular theme, using the International Corpus of Japanese as a Second Language (I-JAS). The purpose of this analysis is to quantitatively compare language production between the two tasks by the same participants through text-mining. The results show that: 1) a comparison of the word token counts reveals that fewer words are elicited in the writing task than in the speaking task; 2) a comparison of word type counts reveals that the higher a participant's proficiency level, the greater the variety of words used in the writing task, and 3) a comparison of characteristic words reveals that the higher a participant's proficiency level, the more similar are the language forms observed between the writing and speaking tasks. This study focuses on Japanese learners' language form use and quantitatively explores its linguistic characteristics in more than 600 participants' written and spoken language data; this differs from the qualitative research methods used for such studies in the past.