From lexical bundles to lexical frames: uncovering the extent of phraseological variation in academic writing

The contextual knowledge of a word is closely related to the knowledge of phraseological sequences as words are often used in the phraseological forms, either continuous or discontinuous. Much has been done to examine the continuous phraseological sequences for various purposes. However, studies on phraseology often overlook the potentially useful discontinuous phraseological sequences that allow for more flexible and productive use of language forms. To bridge the gap in phraseology studies, this study therefore employed a corpus-driven approach to analyse the characteristics of a form of discontinuous phraseological sequence, namely lexical frames in a one-million-word corpus of research articles in International Business Management (IBM). The characteristics of lexical frames were observed in four aspects: the degrees of variability and predictability of lexical frames, the structures as well as the variable slot fillers of lexical frames. The corpus tool, Collocate 1.0 was used to extract three- and four-word lexical bundles while kfNgram was used to extract three- and fourword lexical frames from the lexical bundles. The results revealed that three-word lexical frames are more prevalent in IBM. The degree of variability analysis indicated that there are more fixed lexical frames in the category of three-word lexical frames compared to the four-word category. In terms of the degree of predictability, the category of four-word lexical frames contains more predictable lexical frames than the threeword category. Also, most lexical frames are function word frames and the lexical frames are mostly filled up by content words rather than function words. This study contributes to the understanding of phraseological variation in academic writing