Crafting Scholarly Discourse: A Corpus-Based Descriptive Study of Lexical Bundles in Malaysian Postgraduate Theses

Academic writing is considered one of the most demanding skills for postgraduate students, particularly in English as a second language (ESL) contexts. Lexical bundles, which are recurrent multiword expressions, are central to academic fluency because they enhance cohesion, support argumentation, and reflect disciplinary conventions. However, research on the use of lexical bundles in Malaysian postgraduate theses remains limited. This study investigates the most frequent four-word lexical bundles in the Findings and Discussion sections of 130 theses from Universiti Putra Malaysia (UPM). The bundles were categorised using Hyland’s (2008) functional taxonomy. A one-million-word corpus, the UPM Findings and Discussion Corpus (UPMFD), was compiled and analysed with AntConc, applying frequency and dispersion thresholds of at least 40 occurrences per million words and presence in at least five texts. The analysis identified 85 four-word bundles. The most frequent bundles, including on the other hand and as shown in figure/table, were predominantly text-oriented. Research-oriented bundles, especially those indicating quantification and location, were also common, while participant-oriented bundles appeared less frequently. The findings suggest that postgraduate writers emphasise textual organisation and statistical reporting but show limited use of stance and engagement. This study highlights the importance of explicit instruction in functional lexical bundles to improve coherence, authorial positioning, and academic competence in postgraduate writing.