A learner corpus based study on second language lexicology of Turkish students of English

The emergence of machine-readable corpora in the linguistics field in the 1960s shifted the direction of a considerable body of linguistic research from syntax and phonology, which was by then the focus of linguistic research to a number of domains that were mostly neglected under the umbrella of traditional approaches. And lexicology, which is the target of this research, was a major beneficiary of that shift. By utilising a computer learner corpus-based approach, this study addresses multidimensional lexical aspects of a machine-readable corpus of the writing of Turkish students of English as a foreign language (TICLE). Lexical investigation of this corpus, required a similar sized authentic corpus, which was compiled from Louvain Corpus of Native English Essays (LOCNESS). Employing the computerized contrastive and analytical methods, this dissertation aims at exploring: (1) learners' lexical complexity and richness, (2) how far the learner corpus is deviant from the reference corpus in terms of the features and percentages of the top most 200 frequent tokens? (3) what are the most salient and stereotype features of the learner corpus? And how far the learner corpus influenced by the learners? L1? Findings show that: (1) the learner corpus is much less complex in terms of lexical diversity and density than the reference corpus. (2-3) Learners? top 200 tokens are markedly characterized by vague lexica, underuse and overuse of some lexica, resulting from the influence of the linguistic and rhetorical features of learners? L1. Key words: Corpus linguistics, interlanguage, lexical analysis, learner corpus, underuse/overuse