An error analysis of L2 English discourse reference through learner corpora analysis

Abstract

This research investigates errors of referring expressions used in L2 narrative discourse through two learner corpora, namely the Cambridge Learner Corpus (CLC), and a subsequent controlled corpus created in CLAN. By adopting the cross-sectional approach to learner language used in the English Profile Programme, the research identifies the frequency and type of errors of reference made by Mandarin and Korean L2 English groups from A1 (beginner) to C2 (advanced) levels of the CEFR (Council of Europe, 2001), and asks if and when L2 learners cease making errors of reference, and whether L1 background is a factor in the frequency and type of errors made. The results suggest that L2 learners produce little to no syntactic marking of reference at lower proficiencies, gradually incorporating the appropriate markings of the L2 target at higher proficiencies. It was also found that Korean L2 English learners produce more errors compared to the Mandarin L2 English group at each CEFR level. The difference in the type and frequency of errors between the L2 groups is suggested to lie in the potential grammaticization of numerals and demonstratives in L1 Mandarin to sharing the same functions as the English indefinite and definite articles. (Li and Thompson, 1976, 1989, Hedberg, 1996, Chen, 2004), giving the Mandarin group an advantage in mapping syntactic form (articles) to pragmatic function (introducing and maintaining reference) in the L2.