Corpus linguistics has vastly developed and been addressed in order to mirror the frequencies of naturally occurring lexical items not only in English, but also in many other languages. Learner corpora represent the written interlanguage performance of L2 or foreign language users coming from different mother tongue backgrounds. International Corpus of Learner English (ICLE) is the first computer learner corpus comprised of the argumentative essays written by advanced learners of English representing 16 different mother tongue backgrounds. In this study, the Turkish subcorpus of ICLE (TICLE) which represents the written performances of Turkish users of English as L2 or foreign language has been analyzed and the ten most frequent words listed. TICLE is preferred as it can be be compared with a comparable reference corpus; namely, Louvain Corpus of NativeEnglish Essays (LOCNESS). In addition, LOCNESS has been preferred to reveal the top ten high-frequency words in native usage of English by American English L1 speakers. The results of this data-driven study have been discussed on Contrastive Interlanguage Analysis (CIA); firstly, top ten high-frequency words have been illustrated; secondly, the top ten high-frequency words in L2 usage have been compared with that of the native performance and then the overuse, underuse and statistical significant difference tests have been conducted to reveal any properties of interlanguage and native use. Finally, the linguistic properties of those top ten words have been discussed in detail with implications to ELT in Turkey.
Gülten Koşar, Social Sciences University of Ankara, Turkey
Yunus Emre Akbana, Kahramanmaraş Sütçü İmam Iniversity, Turkey
This paper is part of the ECLL2015 Conference Proceedings (View)
View / Download the full paper in a new tab/window