Developing an inventory of core lexical bundles in English research articles: a cross-disciplinary corpus-based study

Abstract

Abstract Corpus linguistics has revealed that formulaic language forms an important part of written academic discourse. Recently, much research has focused on lexical bundles as a new category of formulaic language. The primary goal of the present study was to identify core lexical bundles in English research articles across different disciplines and to explore the functions they fulfil. The secondary goal was to determine the coverage of core bundles over lexical bundles from a corpus of research articles in applied linguistics. Using AntConc programme, the analysis started with the creation of three separate lists of lexical bundles from three corpora. Then, the common lexical bundles from these three lists were identified. In total, 661 core lexical bundles common across the three disciplines were identified. Functional analysis revealed the predominance of text-oriented bundles. In addition, core bundles covered 30.19% of the list of bundle types in the research articles. The functional utility of core bundles and their coverage in a domain-specific discipline reinforces the idea that lexical bundles can be transformed into useful learning and testing materials.