Taking up again the idea of grammar as a process, or of doing it whilst interacting, grammar is something that develops, or grows organically. Scott Thornbury calls it a complex system and makes 4 points about its composition.
1. It is dynamic and non-linear, it doesn't develop step by step but moves as a whole. Likening it to the way a shoal of fish moves.
2. It is adaptive and sensitive to feedback. This time likened to an ant hive adapting their movements due to their environment.
3. It is self-organising. It wants to see and make rules. Similar to the way a child's world and language knowledge develop.
4. Finally, it is emergent. It has the power to make connections from component parts. E.g. it can make the noun phrase 'not all of the three famous Andrews sisters' from individual meetings with e.g. 'both of the men', 'two dogs', 'the Blues Brothers', 'not half' .
(Thornbury 2001, 48-52)

Our goal is therefore to facilitate this emergent, complex system to grow and to do this in the most effective way.
A learner will acquire a structure when he or she is ready for it, by a process that could well involve steps backwards as well as forwards. E.g. the u-shaped curve of acquiring an irregular past tense - eat, ate (the learner uses it as a lexical item), eated (they apply the rule for regular), ated (maybe) and finally, ate (again). (Thornbury 2001,46)
This process is following a pattern of noticing a feature, formulating and applying a rule, experimenting, re-noticing and again experimenting.
Batstone proposes the cycle of noticing, structuring, re-noticing and re-structuring as the pattern that emerging grammar takes (Batstone 1994,42) and suggests that helping learners to do this will facilitate learning.
However, more is needed to make the language fully useable. The learner must practice to be able to use the language they have been (re)noticing and (re)structuring.
As language is stored and used in pre-composed chunks (Lewis 1993,19) it is suggested that this is what fluency is based upon.

It is our ability to use lexical phrases that helps us speak with fluency. This prefabricated speech has both the advantages of more efficient retrieval and of permitting speakers (and hearers) to direct their attention to the larger structure of the discourse, rather than keeping it narrowly focussed on individual words as they are produced.
(Nattinger and DeCarrico, p32, quoted from The Lexical Approach, 1993,19)

This observation has significant implications for the learning of grammar, in that it is not until the language has been proceduralised, i.e. stored as something ready to be instantly produced, that acquisition can be said to have fully taken place. (Batstone 1994,45) Another relevant point which emerges from the above quote is that of using language whilst paying attention to the discourse and other factors of production and comprehension such as articulation and perception together in real-time. This synthesisation is necessary so a learner can get on with the business of interaction without having to worry about processing language at the same time.

To sum up:
(re)notice - (re)structure - proceduralise
(Batstone 1994,45)

Some points to keep in mind when teaching grammar.

I would like to briefly refer back to 2 previous points. The first is that of by encouraging learners to use strategies and through the nature of conversation, we are only really providing fluency practice, which is needed for making language fluid and automatic, but through its nature could actually be detrimental to the learners developing interlanguage. In most spoken activities the pressure on the students is high, encouraging learners to rely on lexis instead of grammar to realise interaction.
The second point is related to that of distance or a gap in shared knowledge.
Scott Thornbury uses these 2 principles to create his 2 rules of thumb when designing grammar activities.

1. The greater the distance, the more grammar. (That is to say, where there is a knowledge gap or a social gap, words alone will be insufficient to bridge that gap.)
2. The fewer the processing demands, the more grammar. (That is to say, where there is reduced pressure and more planning time, there is a greater likelihood of 'grammaring' as opposed to 'wording'.)
(Thornbury 2001,21)

Thornbury goes on to point out that combined with these 2 points, learners will need to know how precise they have been and that feedback must be explicit and immediate, giving 4 points to bear in mind in grammaring activities:

· Low context dependence.
· High incentive for precision.
· Low pressure.
· High feedback.
(Thornbury 2001, 21)

How should we go about teaching grammar?

What should we do to encourage the process of (re)noticing, (re)structuring and proceduralisation?
Firstly, the language needs to be made available, a lot of input must be provided and awareness raised to linguistic features to turn it into intake, following the principle that once something has been noticed, it is likely to be noticed again and repeated noticing leads to acquisition.
Thornbury suggests consciousness-raising or grammar interpretation.
Among the characteristics of consciousness-raising (C.R.) there is:
· The attempt to isolate a specific linguistic feature for focused attention.
· The provision of data which illustrates the target feature, as far as possible from texts which have already been processed for meaning.
· The requirement that learners utilise intellectual effort to understand the targeted feature. Learners are encouraged to hypothesise and test these hypotheses.
(Willis and Willis 1996,64)

A feature of C.R. exemplified by Thornbury is that form should matter to meaning in solving the task, for example, using answer phone messages from his friends who are either still on holiday or already home students are forced to pay attention to the use of either past simple or present perfect to decide which. For this and more of this type of exercise see appendix 1.
C.R. activities of this type 'are designed to raise learners' awareness regarding specific grammatical items in order to promote the 'restructuring' of their mental grammar'
(Thornbury 2001,100)

This type of activity is valid at all levels. At upper intermediate level it can be particularly useful for raising awareness to the finer points of grammar and lexical grammar that are quite likely to go unnoticed as the students are happy that they understand without taking the chance to develop further by paying attention to the language.
Three types of activity which work well for this purpose are retranslation, reformulation and dictogloss or grammar dictation where high feedback is provided in the comparison of the students' version and the 'ideal' version. For examples of this type, see appendix 2.

Finally, students need to produce the language to help them proceduralise it. When designing tasks for this purpose we need to refer back to the ideas raised about processing demands and distance.

Some ways to increase distance are:
· reduce the students shared knowledge, e.g. giving them different pictures to tell a story together without showing each other the pictures.
· Increase formality, e.g. getting them to write a formal report of the task they have done.
· Bend their schema, e.g. writing something in the style of a newspaper report, where the chronological order and written order do not coincide, or just change things away from the usual way of their happening.
For examples of this type see appendix 3.

Some ways to reduce processing demands we are:

· Repeat a task. Martin Bygate demonstrates some excellent improvements in accuracy, range, complexity, repertoire and lexical selection and collocation when learners do this. (Bygate, in Challenge and Change 1996, 142)

· Give preparation time before a task. Pauline Foster demonstrates surprising improvements in fluency, syntactic and lexical variety, complexity and accuracy. (Foster, in Challenge and Change 1996,134)
· Change the medium from spoken to written, or mirror an internet chat room in a slow motion conversation. (Thornbury 2001,25)