Hedging and rhetorical moves in research article discussion sections across disciplines: a corpus-based analysis

Although research on Research Article (RA) Discussion sections has extensively examined move structures, it has seldom integrated analyses of hedging across disciplines. As a result, understanding of how rhetorical functions are realised through linguistic choices remains limited. This study investigates rhetorical moves, hedging distribution, and their interaction in RA Discussion sections across hard disciplines (Electrical Engineering, Chemistry, and Physics) and soft disciplines (Linguistics, Sociology, and Humanities). Using a mixed-methods approach, the study combines qualitative move analysis with corpus-based hedging analysis supported by computational tools and statistical testing. The findings show that Move 4 (Commenting on Results) is the most frequent across all disciplines and serves as the primary locus of hedging. Move 1 (Background Information) and Move 2 (Reporting Results) also occur frequently, while Move 7 (Deductions from Research) typically functions as closure. Disciplinary variation is evident, with soft disciplines exhibiting higher density and variability of hedging compared with hard disciplines. Statistical analysis further confirms that both move type and disciplinary context significantly influence hedging frequency. These findings provide a more integrated account of how rhetorical structure and linguistic resources interact in academic writing, contributing to cross-disciplinary genre analysis and informing academic writing pedagogy.