This study investigates how Chinese scholars in Applied Linguistics construct different authorial stances in their English and Chinese research articles (RAs) by using interactional metadiscourse comprising boosters, hedges, and self-mentions. A specially designed corpus of 22 Chinese and 22 English RAs written by the same group of Chinese scholars was compiled and examined for metadiscourse forms and functions. We found that (a) while the Chinese scholars employed similar frequencies of boosters in both their Chinese and English RAs, they used significantly more hedges in their English RAs than in their Chinese RAs; (b) while they used more boosters than hedges in their Chinese RAs, a reverse pattern was found in their English RAs; (c) they used significantly more self-mentions, particularly first-person pronouns in their English RAs, than in their Chinese RAs. These findings indicate that the same Chinese scholars have displayed different epistemic stances and authorial identities in their Chinese-medium and English-medium publications.