Lexical bundles in EAP: corpus-based studies of writing and speaking - Multi-word units and metaphor in ESP

Vocabulary and English for Specific Purposes Research - Averil Coxhead 2018

Lexical bundles in EAP: corpus-based studies of writing and speaking
Multi-word units and metaphor in ESP

Lexical bundles in professional academic writing were the focus of research by Byrd and Coxhead (2010), using Coxhead’s corpus (3,500,000 words), which was the corpus for Coxhead’s AWL study (2000). The texts in this corpus included textbooks, journal articles, laboratory manuals, book chapters and technical reports. Byrd and Coxhead (2010) investigated four-word lexical bundles in four sub-corpora: Arts, Commerce, Law and Science. These four disciplines shared 73 bundles, which represented 1.1% of the total corpus (each bundle occurring at least 20 times per million words). Byrd and Coxhead (2010) then took the range of occurrence into account, and selected only the lexical bundles that met a minimum of 10% in each of the four disciplines in the corpus. On the other hand, for example, had a frequency of 353 times in the whole corpus, and its range was 23% (Arts), 27% (Commerce), 35% (Law) and 15% (Science). Like Hyland (2008) and his finding of lower levels of lexical bundles in Applied Linguistics, Byrd and Coxhead (2010) study found that Arts contained the lowest percentage of lexical bundles (1.44%). The Sciences also had fairly low amounts of lexical bundles at 1.46%. In contrast, bundles in Law accounted for 5.44%, and in Commerce, they accounted for 2.65%. This study shared 21 bundles with those found by Biber et al. (2004) and Hyland (2008) (for more on lexical bundles in professional arenas, such as Ha (2015) on lexical bundles in Finance, Crawford Camiciottoli (2007) and Nelson (2000) in Business Studies and Verdaguer, Laso and Salazar (1996) in Biomedicine, see also Chapter 8).

Nesi and Basturkmen (2006) investigated the amount and discourse functions of lexical bundles in an academic spoken corpus, made up of 160 lectures from Nesi’s British Academic Spoken Corpus (BASE) (see Thompson & Nesi, 2001) and 40 lectures from the MICASE (Simpson, Briggs, Ovens & Swales, 2002). In total, the corpus contained 1,270,798 words in four disciplines: Arts and Humanities, Social Sciences, Life Sciences and Physical Sciences. Nesi and Basturkmen (2006) focused in particular on bundles which occurred 10 or more times in each discipline and 50 times in the whole corpus for closer analysis. Seventeen of the most frequent 20 bundles in this study were also in Biber et al. (2004) top-20 bundles in classroom discourse (for example, the end of the, at the same time, and if you want to). Nesi and Basturkmen (2006) then worked carefully on concordances of bundles to examine their role in the cohesive discourse of the lectures. They reported in some detail on referential bundles and discourse organisers from the lecture corpus, based on the categories from Biber et al. (2004), and concluded that second language learners in academic contexts need to be aware of these bundles and the roles they play in the discourse of these very common academic listening events.