This study compares metadiscourse markers in result and discussion sections of literature and engineering research papers. To this end, 40 research articles (20 literature and 20 engineering) are selected from two major international journals. Based on Hyland’s (2005) model of metadiscourse, the articles are codified in terms of frequency, percentage, and density of interactive and interactional markers found. The two corpora are compared to see to what extent the genre of lectures are different. Data are analyzed through Chi-square and the results of the quantitative analysis reveal significant cross-linguistic differences for overall frequency of metadiscourse except for frame markers and boosters.
