There is a classic example of a child's developing grammar in the literature on first language acquisition where the efforts of a parent to correct his two-year-old daughter is shown to be less than effective:

Child: Want other one spoon, Daddy.

Daddy: You mean, you want "the other spoon".

Child: Yes, I want the other one spoon, please, Daddy.

Daddy: Can you say "the other spoon"?

Child: Other one spoon.

Daddy: Say "other".

Child: Other.

Daddy: Spoon.

Child: Spoon.

Daddy: Other spoon.

Child: Other spoon. Now give me other one spoon? (1)

At this point, apparently, Mummy intervened and told everyone to get on with their breakfast.

What this extract is normally used to exemplify is that correction is ineffective during child language acquisition, because the child at each phase in his or her language development which changes from day to day has a notional grammar of the language they're learning and no amount of persuasion on the part of parents or other adults will change it. For the child at this phase, the phrase 'another one spoon' is grammatically correct according to her grammar at the time and what Daddy's on about is anybody's guess.

In fact, children (whom we always underestimate) know in some sense that this is going on. Another example shows (another) Daddy imitating the grammatical forms of the child and receiving the complaint, 'NO! Daddy, I say it that way, not you.' (2)

Interlanguage

The word used to describe the developing stages of a grammar in second language acquisition is usually 'interlanguage', of course. The question is, is this interlanguage susceptible to outside intervention by direct methods (overt correction and modelling) or is it the case that the learner simply needs to hear enough language around him or her to build up a grammar by incremental means so that it approximates the developed grammar of the target language (TL)?

The interlanguage of a learner seems to be based on a static model of language development, or at least on a deficiency model. The interlanguage is seen as either a degenerating form of the L1 grammar as it maps L1 structures on to the L2, or it is seen an incomplete version of the target language grammar, deficient in detail and control. In either case, the evidence of the interlanguage what the learner is actually doing with the language at specific moments is usually analysed as failing in some respect or other and hence requiring remediation or treatment in the form of what we loosely call teaching.

An alternative proposal is to see the developing interlanguage as representing a series of temporary but functioning grammars in their own right. That is, the learner has judged that such and such a way of structuring the target language seems to fit with the notions he or she has about how the TL works, so this is how we speak. This has sometimes been called hypothesis testing, as if the learner were conducting ongoing linguistic experiments and paying close attention to the resulting effects, weighing them up and estimating the probability or not of having achieved an acceptable version of the TL.

One wonders if language learners really have the time or the inclination to function as hypothesis testers. Such a notion seems to imply that any feedback that is available is coherent and detailed enough to allow for a clear identification of a problem when it occurs, at the same time demanding sufficient analytical expertise on the behalf of the learner to be able to locate the precise nature of the problem. It might instead be the case that with exposure to the language and the gradual accumulation of examples and counter-examples, the 'this-site-under-construction' nature of the internal grammar steadily and inexorably builds an ever more sophisticated TL grammar until such point as this becomes indistinguishable from that of a mother tongue user of the language.

Phonology

If we look at how interlanguage phonology develops over time (3), we can see that the process is not random. In a study which looked at how Japanese learners changed their realization of the / l / phoneme in English (that notorious obstacle), it could be seen that the processes of variation in production were clearly related to the sounds that came before and after the <l>s in the test words. That is to say, the underlying language behaviour was rule-based. As the realization of the phoneme became less Japanese-like and more English-like, the learners demonstrated the application of a series of variable rules which reflected the patterned (as opposed to random) variation of their production.

Rule-based linguistic behaviour is grammatical. As the reseacher says, 'sound system learning proceeds by the gradual and systematic modification of rules in a newly developed grammar'. The rule-based behaviour gives rise to variable performance that is predictable and that proceeds in stages. The study of these variant forms, then, is emphatically not a study in error analysis, but is a variablility analysis.

Definitions of interlanguage as a concept as described above seem not to include the possibility of variable forms, but deal with degenerate or incomplete attempts at achieving an end-state normative standard. If we re-think the definition of interlanguage so that it can include the dynamic processes that characterise language learning, we may be able to break away from the deficiency paradigm for intermediate states of grammatical development.

The L2 Learner

Cook's (1999) article on centering the language learning attention on the actual language learner shows that this implies paying more attention to the linguistic production in terms of what the learner is trying to do, given the linguistic background represented by the mother tongue, rather than how he or she isn't doing what native speakers do. Instead of the learner being somehow marooned along the continuum from L1 to L2, struggling to deal with new grammatical concepts, or testing ill-formed hypotheses about the TL, we might conceive of him or her as proceeding more or less deliberately along a series of intermediate steps, each of which is coherent on its own terms, and each of which represents a fully functional grammar at that moment and for that learner.

If linguistic behaviour is rule-based along this pathway, learners with a L1 in common should travel along it in recognisable if not exactly parallel ways. Therefore, language teaching which is based on a detailed knowledge of how speakers of language A have previously dealt with the question of learning language B will be potentially more flexible and responsive to individual learner needs at each successive phase in solving this question. This responsiveness in turn implies a body of knowledge, a data base or corpus, of the typical, rule-based variant forms that are specific to each L1/L2 encounter.

Error Correction

What does this mean in terms of teaching practices related to error correction? If we see variant forms as just that variant forms, not incorrect forms that are derived from rules that are specific not to one language or another but rather to the way those two languages interact when you are trying to learn one of them as an L2, then our pedagogic interventions can be that much more subtle and well-defined. We can judge our learners' linguistic production not in terms of what he or she is failing to do in terms of the native speaker conception of 'correctness', but instead in terms of how well those typical variability rules are manifesting themselves at each stage in the process. Intervention would then be aimed at refining the variant forms in ways that are recognisably rule-based for that L1 - L2 moment to keep the learner on the track towards less L1-like and more L2-like linguistic behaviour.

References

Cook, V. 'Going beyond the native speaker in language teaching'. TESOL Quarterly, 33, n 2, 1999.
Dickerson, W. 'The psycholinguistic unity of language learning and language change', in Ioup, G & Weinberger, S. Interlanguage Phonology: the acquisition of a second language sound system, Cambridge, Mass.: Newbury House, 1987.
Pinker, S. Word and Rules: the ingredients of language. Basic Books, 1999.

Notes

See Fromkin, V and R. Rodman An Introduction to Language. New York: HBJ, 1993, p 404.
See Pinker, S. Word and Rules: the ingredients of language. Basic Books, 1999, p. 199.
Dickerson, 'The psycholinguistic unity of language learning and language change', in Ioup, G & Weinberger, S. Interlanguage Phonology: the acquisition of a second language sound system, Cambridge, Mass.: Newbury House, 1987.