INVESTIGATING LEXICAL BUNDLES IN THE CORPORA OF ENGLISH AND INDONESIAN RESEARCH ARTICLES WITH THE SKETCH ENGINE

Abstract

The low publication rate of Indonesian researchers in reputable international journals, particularly in arts and humanities,
is caused, among others, by difficulties they faced in producing precise expository texts in English, which are different
from texts in Indonesian. The present study examines lexical bundles in the corpora of English and Indonesian research
articles (RA) on literature and linguistics to describe the similarities and differences of conventionalized phraseology in
the scientific genre of English and Indonesian by using corpus software, namely Sketch Engine. The study focuses on
the frequency, structural and functional characteristics of lexical bundles using a mixed-method research design. The
English corpus comprises 1,351,048 words derived from 124 RA, while the Indonesian corpus consists of 637,910 words
collected from 124 RA. We found that three-word lexical bundles are more prevalent than four-word lexical bundles in
both corpora. Based on the structural forms, prepositional-based bundles are the most frequent form in English RA, while
noun-based bundles are the most common form in Indonesian RA. There were no participant-oriented bundles found
in the Indonesian RA corpus in terms of functional classification, whereas the English RA corpus involved more varied
functional categories of lexical bundles. The findings provide an understanding of phraseological combinations in English
and Indonesian scientific writing, characterizing disciplinary discourse as well as native and non-native English speakers’
rhetorical style, and have pedagogical implications for EAP practitioners.