A means to control how writers mark their presence, negotiate knowledge claims, and engage with their audience, metadiscourse is one of the most prominent approaches to analysing academic writing. The present systematic review attempts to take stock of the existing literature by investigating how metadiscourse has been researched in ...
A means to control how writers mark their presence, negotiate knowledge claims, and engage with their audience, metadiscourse is one of the most prominent approaches to analysing academic writing. The present systematic review attempts to take stock of the existing literature by investigating how metadiscourse has been researched in academic writing by analysing a sample of 370 high-quality empirical studies published between 1990 and 2021. Studies were coded for their conceptual frameworks, research designs, data sources, study contexts, writers, texts, corpora, and reporting practices. It was found that over 80% of research involved cross-sectional descriptive corpus-based analysis, drawing on intercultural rhetoric. Owing to its impact, ease of application, and study comparability, most research adhered to the ‘broad’ tradition in metadiscourse. Representative of this approach, Hyland's interpersonal framework and models of stance and engagement were prevalent, although difficulties in undertaking a ‘thick’ analysis of such a wide variety of features coupled with publishing constraints meant that many authors narrowed their focus to a few select features (especially hedges, boosters, and self-mentions). Approximately 37% of corpus-based research followed the ‘thin’ tradition, with an emphasis on marker frequency counts over contextually-bound interpretations. Corpora of English texts, notably, research articles, were prominently studied, with little research taking place outside of university contexts or recruiting human participants as informants. We discuss avenues to advance research in metadiscourse, through identifying possible future inquiries and improving study quality.
