The relationship between the use of fillers and Japanese proficiency: Focusing on OPI level assessment

This study employed a Cumulative Link Mixed Model (CLMM) to examine the relationship between Japanese learners' filler use and OPI (Oral Proficiency Interview) level ratings. The results showed no significant association with proficiency for the i(い)-group, u(う)-group, ano(あの)-group, and eeto(ええと)-group fillers. In contrast, the a(あ)-group, e(え)-group, o(お)-group, and n(ん)-group fillers were significantly negatively associated with higher proficiency, while the use of maa(まあ)-group and sono(その)-group fillers was significantly positively associated with higher proficiency. Furthermore, an analysis of maa(まあ)-group and sono(その)-group fillers based on their functions and discourse roles showed that turn-internal maa(まあ) and inferential/re-presenting sono(その) require at least two clauses to be produced. This indicates a connection with paragraph-level language production, which is one of the criteria for 'Text Type' in OPI assessment. It was also found that the use of these fillers with such functions is significantly positively associated with higher proficiency. On the other hand, for fillers such as ano(あの) and eeto(ええと), as well as maa(まあ) used in adjacency pairs, no significant increase with proficiency was observed, confirming that their use is not associated with paragraph-level production. The use of fillers linked to paragraph-level production can be considered one of the key linguistic factors that can help explain the relationship between filler use and the development of oral language proficiency.