Repetitive bundles in Malaysian learner writing

Abstract

One way to investigate learner writing is by analyzing the most frequently recurring sequences of words, that is, lexical bundles. This paper presents results for lexical bundles analyses of a Malaysian corpus (MCSAW) against its reference language variety, LOCNESS (Louvain Corpus of Native English Essays). Key 4-word lexical bundles are firstly investigated in terms of their frequencies as well as distribution in both corpora. Following this, key lexical bundles are further categorized and analyzed according to their functions, including qualitative analysis of the most recurrent bundles by examination of concordance lines. Results show that learners use simple types of lexical bundles repeatedly compared to their native speaker counterparts. Evidence of tautology can also be found in learner writing. The findings highlight that using lexical bundles appropriately is important to achieve native-like fluency, while the absence of more varied lexical bundles in learners’ discourse may result in unidiomatic-sounding writing style.