The comparative effect of direct written corrective feedback and metalinguistic explanation on learners’ explicit and implicit knowledge of the English indefinite article

• Direct corrective feedback does not lead to greater accuracy in an Error Correction Test or in new pieces or writing.
• In contrast metalinguistic explanation is effective.
• The effects of the metalinguistic explanation disappears over time.
• Explicit knowledge is valuable in helping learners monitor for accuracy but is not durable.

The study extends current work on written error feedback in writing in two ways. First, it examines whether it has an effect on adult ESL learners’ L2 implicit and explicit knowledge. Second, the study compares the effect of one common type of feedback – direct corrective feedback (DCF) – with an alternative type of error feedback – the provision of metalinguistic explanation (ME). The effect of these two types of error feedback was measured by an Error Correction Test (ECT) and by examining the accuracy of use of the target feature (the English indefinite article) in both a revised text and in new pieces of writing by 49 low-intermediate ESL students in an intensive language programme in the United States. In addition, eye-tracking data and self-reports elicited from the learners provided information about the use that they made of the DCF and ME. It was found that the DCF had no effect on accurate use of the target feature suggesting that it benefited neither implicit nor explicit knowledge. In contrast, the ME led to gains in accuracy in the ECT and in a new piece of writing completed immediately after the treatment but not in a second new text completed two weeks later. These results are interpreted as indicating that the ME helped to develop learners’ L2 explicit knowledge but that the effect was not durable and thus probably had no effect on their implicit knowledge. Learners’ self-reports indicate that the learners receiving the DCF did not develop awareness of the rule whereas those receiving the ME did and were able to use it when revising their original text. These findings are discussed from the perspective of both SLA theory and language pedagogy and suggestions for further research are put forward.