Toward written error correction with a Japanese-English parallel corpus: Data-driven learning in the japanese EFL classroom

Abstract

The concept of data-driven learning (DDL) – direct student use of corpora – has been gaining attention among researchers and teachers. DDL gives learners the chance to take an inductive approach to learning by recognizing patterns in corpus data rather than following abstract rules. However, very few studies focus on presenting clear practices that other teachers can readily use in their classes, particularly related to written error correction. This paper begins by reviewing research pertaining to DDL in student writing as well as using Japanese-English parallel corpora. It then outlines a recent study by Jenifer Larson-Hall (2015), notes weaknesses in the study, and outlines a current project (McGuire & Larson-Hall, under preparation).