The Future of Education

There have been many reports on advances in the development of learner corpora that have made it possible to effectively use these collections of texts for the benefit of the learning process. This paper lists all possible applications in English courses taught to Bachelor students of a middle-size learner corpus REALEC, which comprises student written works supplied with expert annotation of mistakes, browsing and search options, and some optional automated tagging system. Annotation in the corpus is given by either experts (mostly, EFL instructors), or by learners themselves under the supervision of their EFL instructors. As the first point, the paper argues that when EFL methodology requires that students apply the error classification in the process of annotating their peers’ essays and gradually their own essays as well, their understanding of subtle areas of grammar, vocabulary and discourse improves, and correspondingly, the number of errors in their written works decreases. The second argument concerns the tool for the development of placement and progress tests, which makes use of sentences with mistakes made by other learners – contributors to the corpus. In the suggested design of the tests sentences are automatically extracted from the same corpus, manually divided into three echelons according to the complexity of the change required in the correction of the mistake, and then administered to learners as a way of automated measurement of their proficiency in English. The submitted test is scored automatically within minutes. The third possibility considered in the research is the possibility to supplement the corpus with the platform of trainers automatically or semi-automatically set up on the basis of frequently marked errors in the corpus. In conclusion we point out the ease and usefulness of the proposed applications both for EFL instructors and English learners.

Key words: EFL/ESL methodology, learner corpus, error annotation, automated trainers; corpus-generated tests