A corpus-driven investigation into lexical bundles across research articles in Food Science and Technology

Abstract

The purpose of this study was twofold: (a) to identify the most frequent 4-word lexical bundles and (b) to analyse the functions these lexical bundles may serve. To those ends, a corpus of 4,652,444 in Food Science and Technology (hereafter FST Corpus) was developed, using 1,421 research articles (RAs) across 38 Food Science and Technology (FST) journals. Setting frequency and range as two criteria, we used AntConc to identify the most frequent lexical bundles. We also used Hyland's (2008b) functional taxonomy to analyse the functions of the lexical bundles. The results of frequency and range showed 153 lexical bundles in FST Corpus. Functional analysis of the lexical bundles revealed 86 text-oriented, 63 research-oriented, and four participant-oriented lexical bundles, suggesting the central role text-oriented functions may play in FST. Implications for the explicit instruction of lexical bundles, for graduate students in FST, and for EAP curriculum developers and materials producers are discussed.