THE USE OF SPOKEN LEARNER CORPORA TO DETECT PROBLEMS WITH LEXICAL ACCURACY
THE USE OF SPOKEN LEARNER CORPORA TO DETECT PROBLEMS WITH LEXICAL ACCURACY

Summary/Abstract: The purpose of this study is to offer a solution for the easier and more accurate detection of speaking problems, especially with the use of lexical items, through the use of a corpus methodology. Although the compilation of spoken corpora is a difficult and time consuming process, it has a great potential to shed light on the use or misuse of English by foreign language learners. For the purposes of this study, a spoken corpus of learner English was compiled. The participants are students at the intermediate to upper-intermediate proficiency level of English. The corpus used in the study, namely ‘Corpus of Learner Monologues’ (CLM) consisted of transcriptions of 35 spoken accounts by participating foreign language learners on two different topics. The topics assigned for the compilation of the corpus were chosen from among IELTS (International English Language Testing System) exam speaking section topics. The recordings were transcribed and hand coded for error categories. After the coding of errors, the corpus was analyzed by using AntConc 3.2.4w, a special software for corpus analysis. The results revealed that verbs are the most problematic lexical group in students’ speech. Other common errors included adverb, noun, and adjective word groups. Results also highlighted that “wrong word choice” was the most common reason for students’ lexical errors. The study carries implications for the teaching of speaking skills.

• Details
• Contents