Introduction

Views on the role of explicit instruction in the acquisition of implicit knowledge, often taken to be the goal and product of L2 teaching and learning, continue to be a contentious issue in L2 methodology (cf. Pawlak 2006). As implicit knowledge of L2 correlates with fluent language proficiency, approaches advocating rich input and “learning by doing” have been favoured as the basis for developing L2 competence. However, as the critics of the structural and audiolingual methods noticed already in the 1960s, such methods teach “speech, not language” (N. Ellis 1994: 37). Generally speaking, approaches that depart from the teaching of metalinguistic rules, which often overemphasise the quantity of output at the expense of its quality, produce fluent speakers whose language may be riddled with lexical and grammatical mistakes. As Ammar and Spada (2006: 544) have observed:

The downplaying of the role of explicit instruction (and explicit learning) brings in its wake a downplaying of grammatical competence in general and accuracy of expression in particular. This is especially worrying from the point of view of long-term L2 development: gaps in language competence of relatively advanced language learners are particularly difficult to eliminate as they tend to stabilise.