abstract in written mode of language, metadiscourse markers are used commonly to help writers in general and academic writers in particular to produce coherent and professional texts. the purpose of the present study was to compare introduction sections of applied linguistics and physics articles regarding their use of interactive and interactional metadiscourse markers based on the model proposed by hyland (2005). in order to carry out this study, 50 articles from each of the disciplines were selected whose writers were native farsi speakers. both qualitative and quantitative approaches to data analysis were employed in the present study. it was found that applied linguistics used significantly more stances of metadiscourse markers in general and interactive metadiscourse in particular. in the case of interactive metadiscourse subcategories, applied linguistics articles used more code glosses, physics articles employed more evidentials and they were approximately the same in using frame markers. neither of the disciplines differed significantly from one another in the case of interactional metadiscourse. no significant difference was found for interactional metadiscourse in general, but it was found that applied linguistics articles used significantly more boosters compared to the physics articles, physics articles employed more hedges and both of the groups were approximately the same in using self-mentions. the findings of the present study are useful to syllabus designers in emphasizing the critical importance of allocating some sections to metadiscourse markers in their textbooks.