Native English scholarly writers use interactional metadiscourse markers in their research articles to indicate their stance and negotiate their claims with readers. Hedges are a type of interactional metadiscourse marker often used in scientific research articles to soften writers’ claims and protect themselves from criticism (Hyland, 2005a). When Japanese researchers write research articles in English, they tend to use fewer English hedges than native English writers. Although hedges are used in research articles written in Japanese, their usage appears to differ from that of English hedges.
This study analyses the use of hedges in research articles by Japanese writers in both English and Japanese in comparison with English hedges employed by native English writers in order to reveal the differences in the use of hedges between the two languages. The analysis focuses not only on the discrepancies in usage between native English and Japanese writers, but also on the characteristics of hedges used in academic articles written in English and Japanese. Furthermore, this study investigates the impact of Japanese writers’ first language on their use of English hedges in articles written in English. A total of 30 published empirical research articles in soft science disciplines in English and Japanese were used for the quantitative and qualitative analyses of the use of hedges, revealing both the writing and linguistic differences between the two languages. Thus, this study aims to offer pedagogical suggestions for Japanese learners of English to use hedges more effectively in their research papers written in English.

The fluctuation hypothesis (Ionin et al., 2004) holds that learners of English as a second or foreign language (ESL/EFL) will continue to make errors in their use of English articles until they switch their criteria for English article selection from specificity to definiteness. However, L1 language acquisition research indicates that young children who are L1 speakers use English articles accurately on the basis of whether the referent is specific or nonspecific before they acquire the definite–indefinite distinction. It is possible that Japanese EFL learners, like young L1 learners, can accurately use English articles employing the specific–non-specific distinction even before they master the definite–indefinite distinction. This study examined this hypothesis using Bickerton’s (1981) semantic wheel-based taxonomy and Díez- Bedmar and Papp’s (2008) tag-coding system. An analysis of 38 essays from the Nagoya Interlanguage Corpus of English Reborn showed that the specific– non-specific distinction can contribute to a highly accurate use of English articles for non-specific referents and relatively accurate article use for specific referents by Japanese EFL learners who have difficulty applying the definite– indefinite distinction. The pedagogical implications of the results were discussed.