A Corpus-based Study Of High-frequency Multi-word Constructions In Native And Chinese Academic Writing

An increasing number of studies have shown that language is made up of a largeproportion of formulaic sequences that are of special importance to second or foreignlanguage learners who want to achieve native-like fluency and accuracy. Corpuslinguistics has made one special type of formulaic sequence a research hotspot, i.e.multi-word construction (MWC), which refers to uninterrupted sequences of two-or morewords that commonly go together in natural discourse. However, most domestic empiricalstudies on MWC focus on undergraduates‘composition writing, and studies focusing onMWC use in academic writing are still few. The present study aims to investigate the useof high-frequency MWCs in native speaker and nonnative speakers‘(L1Chinese)academic writing, which will contribute to the teaching of academic writing in China.Six Corpora in three pairs, amounting to7million words, are built by collectingnative and nonnative speakers‘research articles, PhD dissertations and MA theses inLinguistics. In the present study, a list-based and corpus-driven approach is adopted toidentify the TOP100most frequent MWCs in six corpora. The MWC identification iscompleted by two steps. Firstly, Liu‘s (2012) list of most frequent MWCs in generalacademic writing is used as the database/guide for our MWC identification in six corpora.Secondly, in consideration of discipline specificity of this study, some high-frequencyMWCs in Linguistics may not be identified based on Liu‘s List. Therefore, the empiricallycorpus-driven approach is employed to identify these discipline-specific MWCs in sixcorpora in order to complement Liu‘s list. After these steps, the TOP100most frequentMWCs in six corpora are identified and then categorized functionally according to Liu‘ssemantic/functional classification. We make a further analysis of the characteristics in thefunctional distribution of these TOP100MWCs in six corpora. In addition, by takingresearch article (RA) corpora as an example, a detailed contrastive analysis is conducted to investigate the similarities and differences in native and nonnative speakers‘use of MWCsin research article.Research findings indicate that there are both similarities and differences in nativeand nonnative speakers‘use of high-frequency MWCs in RA, PhD dissertation, MAthesis writing. The results from six corpora indicate that there exists a high degree ofsimilarity in native and nonnative speakers‘use of most frequent MWCs in RA, PhDdissertation, MA thesis writing. In terms of MWC types, native and nonnative speakersshare a large proportion of TOP100MWCs (76%,73%,69%respectively). In terms ofMWC function, both native and nonnative speakers use the largest proportion ofreferential MWCs and the smallest proportion of stance MWCs. Referential MWCs arethe main conveyor of information, characterizing a high density of information inacademic texts. On the other hand, the results from RA corpora indicate that there stillexists significant difference in native and nonnative speakers‘use of most frequentMWCs in RA writing. Chinese professionals tend to underuse, overuse and misusecertain MWCs, which exerts a negative influence upon the idiomaticity of their languageproduction. For instance, they underuse some MWCs typical of Written English, butoveruse those typical of spoken English, and misuse some MWCs with grammatical,semantical and collocational errors. The investigation indicates that Chineseprofessionals‘overuse, underuse and misuse of certain MWCs is probably related to theirregister awareness, linguistic competence, L2input, and cultural background.This study reveals the characteristics of native and nonnative speakers‘use ofhigh-frequency MWCs and enhances our understanding of MWC use in academic writing.The findings will enrich studies of nonnative speakers‘MWC use, and can provideguidance for MWC identification method and ESP/EAP teaching．...