Ellis, R. (2006). Current issues in the teaching of grammar: An SLA perspective. Tesol Quarterly, 40(1), 83-107.
The field of applied linguistics seems to be undergoing a stage of, to use Thomas Kuhn’s terminology, “paradigm shift”, during which “anything goes”. Before the completion of this paradigm shift, people in the field of language teaching are constantly trying new methods, proposing new theories, and inventing or redefining terms. It is truly an era of excitement to theorists, but might be quite frustrating to practitioners.
Ellis raised several seemly easy questions, such as should we teach grammar, what grammar should we teach, and how should we teach it etc.. It turns out that people have different answers, which are all, to some extent, supported by empirical evidence. I think the problem is not that people cannot agree to one set of answers, but the questions themselves. Maybe, Ellis did not ask the right questions. The field of language teaching concerns teaching methods, which should be rooted in human cognition. To answer the many questions raised in this paper, one has to know how human cognition works. Without addressing the cognitive process of language learning, any speculation about grammar teaching is futile. Yes, some of the speculations can be validated by empirical studies focusing on one or two grammatical structures in a controlled environment (e.g. intensive or extensive). However, as Ellis pointed out these studies cannot paint the full picture of language learning. Since it is practically impossible to test all grammar structures at one time, there’s no way to know whether any type of method (e.g. explicit v.s. implicit) works for all the grammatical structures.
Ellis thoroughly examined the several arguments on grammar teaching, without mentioning what he meant by “grammar”. He used the term “grammar” freely, as if “grammar” is a homogenous entity, while in fact it is not. Ellis himself addressed this issue in the “what grammar should we teach” and the “explicit v.s. implicit” sections. He concludes on p.102 that “there is ample evidence to demonstrate that teaching grammar works”. I think this conclusion is too broad to be of any pedagogical significance. According to Ellis’ own account, empirical results differ depending on what the researchers meant by “teaching” (methods), “grammar”, and “works” (measurement).
Ellis’ 10 ”beliefs” as stated at the end of this paper are thought-provoking. Yet, I am more interested in understanding why he believes so. For example, how does he know “corrective feedback” is important? In the framework of statistical learning, corrective feedback can easily be overridden by strong prior biases.