Despite extensive research on metadiscourse, methodology descriptions in research articles provide limited guidance on how to identify, classify, and, particularly in cases of clustered items, quantify metadiscourse markers. This article discusses the methodological challenges of analysing metadiscoursal adjectives, using the example of novice academic writers’ use of adjectival interactional metadiscourse markers. Our exploratory analysis of the corpus (654,925 tokens) revealed that, in a non-negligible number of cases (13.5%), metadiscoursal adjectives co-occurred with other linguistic items that were performing different metadiscoursal functions, thus putting a different interpretation on the initial observation. The phenomenon whereby both the adjective and its co-occurring item exercise a prominent metadiscoursal function—which we labelled superimposition —has been observed in previous studies but has not been adequately explored and has led to divergent, and often incomparable approaches, to metadiscourse quantification. We argue that metadiscoursal superimposition as a methodological approach can help bridge the gap between the individual marker analysis and their use in academic writing discourse, thus providing a structured, quantifiable, and functionally richer framework. We discuss the benefits, possible pitfalls, and implications of our proposal that superimposition be included and quantified as a supplementary step in metadiscourse quantification analyses.
Takač et al. (Fri,) studied this question.