What then should a teacher do? There are two possible courses of action. One is to abandon grammar teaching. This is what Krashen (1982) recommends. He suggests that teaching grammar results in "learned" knowledge, which is only available for monitoring utterances that learners produce using their "acquired" knowledge, and, as such, is of very limited value. Krashen recommends instead that teachers concentrate on providing lots of comprehensible input so that learners can "acquire" a second language naturally like children acquiring their mother tongue.

This is an attractive proposal -- particularly for teachers who don't like grammar! But it has several problems. One is that students are often convinced that "learning" grammar is of value to them and, therefore, expect the teacher to teach grammar. Another more serious problem is that learners do not seem to master the grammar of a second language even when they get plenty of comprehensible input. Studies of learners in immersion classrooms (e.g. Swain 1985) have shown even after ample exposure to the target language learners continue to make a lot of grammatical errors. In other words, Krashen's claim that learners "acquire" grammar naturally is not entirely correct.

This suggests, therefore, that the second course of action might be better -- trying to find a way of teaching grammar that is compatible with how learners learn grammar. Teachers may not be able to make learners speak and write grammatically, as I found to my cost in that classroom in Zambia many years ago, but they may be able to help learners become grammatical. It is this idea that has motivated much of my own research over the past 10 years (see, for example, Ellis 1993 and 1995).

To my mind, then, the key question is "How can we teach grammar in a way that is compatible with how learners acquire grammar?" Second language acquisition research suggests that grammar teaching should take into account three key principles:

• Learners need to attend to both meaning and form when learning a second language.
• New grammatical features are more likely to be acquired when learners notice and comprehend them in input than when they engage in extensive production practice.
• Learners' awareness of grammatical forms helps them to acquire grammatical features slowly and gradually.

These three principles have guided my own approach to teaching grammar.

Attention to form and meaning
Current second language acquisition theories view grammar learning as best accomplished when learners are primarily focused on meaning rather than form, as Krashen has argued. However, contrary to Krashen's position, these theories also claim that some attention to form is necessary for learning to take place. The problem is that learners are limited language processors who find it difficult to attend to both form and meaning at the same time. Thus, when they are focused on meaning they are unable to attend simultaneously to form and, conversely, when they are focused on form, their ability to understand or make themselves understood suffers.

For this reason, they need meaning-based tasks that also allow them the opportunity to process language as form. In the materials I have been developing, students are first required to process a text for meaning and then, afterward, to attend to how a particular grammatical form is used in the text.

Learning grammar through input
Grammar has traditionally been taught via production practice. That is, students have been required to try to use a grammatical structure in controlled and free exercises. However, current theories of second language acquisition see production as the result of acquisition rather than the cause. It follows that grammar can be taught more effectively through input that through manipulating output.

An interesting study by Tanaka (1996) provides evidence to support such a claim. Tanaka compared two ways of teaching Japanese high school students relative clauses. One way involved the use of input practice, and the other traditional production practice. Tanaka found that input practice led to better comprehension of the target structure and, in the long term, to production that was just as accurate. In other words, the input practice helped learners to process relative clauses in both input and output, but the production practice only helped output.

What does input practice involve? It involves "structured input tasks." These are tasks that require students to (1) read or listen to input that has been specially designed to include plentiful examples of the target structure and (2) consciously attend to the target structure and understand its meaning. In one kind of structured input task, a text is gapped by removing words containing the target structure and asking students to fill in the missing words.

In the grammar teaching materials I have been working on, the structured input tasks are all oral rather than written -- learners have to listen to the texts rather than read them. This is because oral texts require students to process grammatical structures in real time, which is exactly what is needed to help students acquire them. Furthermore, oral texts also serve to practise the important skill of listening.

The role of awareness
Learners can acquire a new grammatical structure only very gradually and slowly. It can, in fact, take several months for them to master a single grammatical structure. For this reason, grammar instruction, no matter how well designed, is unlikely to achieve immediate success. This suggests that grammar teaching needs to emphasize awareness of how grammatical features work rather than mastery. Learners who are aware of a grammatical structure are more likely to notice it when they subsequently encounter it. Thus, awareness can facilitate and trigger learning; it is a crutch that helps learners walk until they can do so by themselves.

How can teachers develop awareness of a grammatical structure? One way, of course, is simply to tell the students how it works. This is the traditional way. Japanese students have plenty of experience of listening to teachers lecture about grammar! An alternative way, which I think is more promising, is to use consciousness-raising tasks. These are tasks that provide students with "data" about how a particular grammatical structure works and help them to work out the rule for themselves. In this approach, students discover how grammar works on their own. Such tasks make the students much less dependent on the teacher.

Fotos (1984) carried out a study to see how well consciousness-raising grammar tasks worked with Japanese college students. She found that the students' awareness of the grammatical structures she targeted was just as accurate when they worked out the rules for themselves as when they were told them. Moreover, in Fotos' study, the students had to work in groups to discover the rules and talked in English together as they did so. Thus, the consciousness-raising tasks doubled up as communicative tasks!

In an attempt to incorporate these principles into materials for teaching grammar, I have developed the following sequence of tasks:

• Listening task (i.e. students listen to a text that they process for meaning).
• "Noticing" task (i.e. students listen to the same text, which is now gapped, and fill in the missing words).
• Consciousness-raising task (i.e. students are helped to discover how the target grammar structure works by analyzing the "data" provided by the listening text).
• Checking task (i.e. students complete an activity to check if they have understood how the target structure works).
• Production task (i.e. students are given the opportunity to try out the target structure in their own sentences). The aim of the production task is to encourage students to experiment with the target structure, not its mastery.

The aim of such materials is not so much to "teach grammar," as this is often not possible, but rather to help students to "become grammatical." This is a lesser goal but it is a worthwhile one. Furthermore, it is a goal that is more compatible with the current emphases on communication and student autonomy.

References
Ellis, R. 1993. Second language acquisition and the structural syllabus. TESOL Quarterly 27: 91-113.
Ellis, R. 1995. Interpretation tasks for grammar teaching. TESOL Quarterly 29: 87-105
Krashen, S. 1982. Principles and Practice in Second Language Acquisition. Oxford: Pergamon.
Swain, M. 1985. Communicative competence: some roles of comprehensible input and comprehensible output in its development. In S. Gass and C. Madden Eds. Input and Second Language Acquisition, pp. 235-252.
Tanaka, Y. 1996. The comprehension and acquisition of relative clauses by Japanese High School students through formal instruction. Unpublished Ed.D dissertation, Temple University Japan, Tokyo.