A Corpus-Based Analysis of Lexical Bundles in Non-Native Post Graduate Academic Writing and a Potential L1 Influence

Abstract

This study was conducted to examine the lexical bundles used by non-native speakers of English and explore any potential L1 influence on L2 lexical bundle use. Following a corpus-based approach, the frequency and types of English four-word lexical bundles in the postgraduate academic writing of Turkish and American students were analyzed, and the bundles unique to the Turkish students were compared with Turkish lexical bundles produced by Turkish post-graduate students. For this purpose, three sub-corpora were compiled: English MA/PhD theses by Turkish, English MA/PhD theses by American, and Turkish MA/PhD theses by Turkish students, all from the area of language teaching. Data analysis showed that the Turkish students used twice as many types of four-word lexical bundles in their English theses (N = 125) as the American students (N = 69). Moreover, 62 lexical bundles were significantly overused by Turkish students, and 37 of these lexical bundles never occurred in the theses of American students. With respect to cross-linguistic influence, the findings showed that Turkish postgraduate students were likely to transfer 24.8% of lexical bundles from their native language, Turkish, to a foreign language, English. Moreover, four-word lexical bundles that were very frequent in Turkish theses were also found to be very frequent in English theses of Turkish students. These findings are discussed in light of previous studies, and pedagogical implications are offered.