Turkish Online Journal of Educational Technology, vol.2016, no.1, pp.1594-1601, 2016 (Peer-Reviewed Journal)
It is an old consensus by now that a computerized corpus-based analysis of semantic prosodic profiles in English language provides various explanations for how EFL learners exploit the target language during their learning process in terms of lexical selection. The prosodic analysis of lexical selections of EFL learners is likely to give us much better means of understanding the acquisition and learning processes of EFL learners as well as the factors influencing these processes (Granger, 2008). This study investigated the semantic prosodic awareness of Turkish EFL learners in terms of using pragmatic markers such as " I mean, I think, so, well, you know, so on, or so ". The analysis was done by comparing two argumentative academic corpora, KTUCALE (Karadeniz Technical University Corpus of Academic Learner English) and BAWE (British Academic Written English). In order to measure possible overuses and underuses as well as statistical significance, frequency counts and log-likelihood ratios were used. Corpus linguistics software AntConc 3.2.2 and SketchEngine online corpus software interface were used in order to collect data. The data was analysed quantitatively by comparing frequencies and calculating log-likelihood ratios. The results of this study validated the commonly held view that a limited number of pragmatic features are overly used in the academic argumentative papers of tertiary level Turkish EFL students. Other findings were that BASE, spoken native corpus, corpus included the highest number of pragmatic markers of all the three corpora and BAWE, written native corpus, contained restricted number of markers such as " so " with ample use.