The idea of interlanguage is founded upon the assumption that an L2 learner, at any particular moment in his learning sequence, is using a language system which is neither the L1, nor the L2. It is a third language, with its own grammar, its own lexicon and so on. The rules used by the learner are to be found in neither his own mother tongue, nor in the Target Language. Thus, Nemser cites Serbo-Croat learners of English who will produce “What does Pat doing now?”, although this construction belongs neither in English, nor in Serbo-Croat. The lesson to be learned, suggest applied linguists such as Nemser, Pitt Corder and Selinker, is that we need to understand the learner’s language as a system in its own right. This is both possible, and interesting because learners tend to go through a series of interlanguages in systematic and predictable ways.

How does the learner create her interlanguage? According to Selinker, there are a number of basic processes – but, particularly in his later work, he insists upon learning strategies – that is, activities that the learner adopts in order to help her acquire the language.

Language transfer – the learner uses her own L1 as a resource. This used to be looked upon as a mistake, but it is now recognised that all learners fall back on their mother tongues, particularly in the early stages of language acquisition, and that this is a necessary process

Overgeneralization – the learner uses an L2 rule in situations in which a native speaker would not. This can occur at a number of levels

• thus at the phonetic level, for example, learners of English, after having learnt to master the English ‘r’, may take to placing it at the end of words, whereas in RP it is not pronounced.
• at the grammatical level, a learner in the early stages may use nothing but the present tense. Later, there may be extensive, non-native use of ‘be – ing’ forms of the verb.
• at the lexical level – learners tend to use base terms and to stretch them – thus a ‘goose’ might be referred to as a ‘chicken’, or a teaspoon may be a ‘little spoon’.
• at the level of discourse, lexical items and expressions may be used in inappropriate social contexts. Someone learning French as an L2, and who has been staying with a friendly family with teenagers may find themselves using the ‘tu’ form to strangers, members of the CRS and so on.
• Simplification – both syntactic and semantic – the learner uses speech that resembles that of very young children or of pidgins. This may be either because they cannot, in fact, as yet produce the target forms, or because they do not feel sure of them.
• Let us look more closely at transfer. It can have several different effects

• Negative transfer

• Until the morpheme studies of Dulay and Burt, it was often assumed that most errors were derived from transfer of the L1 to the L2 – this was referred to as interference. It is now no longer clear where errors derive from. As we have seen, Dulay and Burt believe that the majority of errors are not based on transfer. However, it is not always a simple matter to decide whether an error is L1 based or not.
  • For example, when French speakers use ‘have -en’ forms in inappropriate settings, is it because of overgeneralization, a developmental error, or an interference error based on the Passé Composé?
• Indeed, it is not always easy to decide whether an error has occurred at all. Take again the case of the ‘have -en’ forms. A French speaker learning English may use the form in the correct setting, but actually derive it from the French Passé Composé – he has done the right thing, but for the wrong reasons. Has an error actually occurred? How would we know?
• Consider this dialogue, derived from :
  • A : I (look for) Bob. You (see?) him.
  • B : Yes, I (see) him half an hour ago
• A French learner might produce
  • A : I’m looking for Bob. You have seen him?
  • B : Yes. I have seen him half an hour ago.

If speakers of different mother tongues do, in fact, make different mistakes, and if these mistakes do appear to be related to structures in the mother tongue, then it would seem reasonable to speak of ‘interference errors
At the level of phonology, this certainly appears to be the case
– there are typical accents, and it is comparatively easy to distinguish between the English pronunciation of, say, a German L1 speaker, a French L1 speaker or a Japanese.
However, even here, there appear to be rules that are target language specific – progress through to full acquisition of the ‘th’ appears to follow a fairly regular pattern, which is similar to that of an English child learning her L1.
b) Positive transfer
Not all effects of language transfer are negative – indeed, we may consider that without some language transfer, there would be no second language learning. We have seen that, in the cases of Genie and Chelsea, it is very difficult to master a language after the age of 11 or 12 years of age, unless one already has a mother-tongue to fall back on. It may be that younger children are able to pick up an L2 without reference to their L1, but for adolescents and adults, the mother tongue is a major resource for language learning.
Where languages are historically and linguistically related to each other, the positive effects of transfer may be obvious. French-speaking learners of English and English speaking learners of French quickly come to realise that they share an enormous amount of vocabulary, for example – there are far more ‘Vrais Amis’ than there are ‘Faux amis’, and it makes sense to take advantage of this.
For Japanese speakers learning Chinese, there is a great advantage when it comes to studying the written language in the fact that the Japanese ideographs are based upon the Chinese. This saves considerable time.
However, the Chomskian perspective has lead specialists in SLA to believe that there are deeper levels at which the L1 may aid in language learning. If all languages are fundamentally the same, then it makes a lot of sense to use the rules of the mother-tongue as initial hypotheses about the rules of the L2. We will come back to this point in a later lecture, when considering implicational hierarchies.
We must conclude that – The teacher who tries to forbid his students from having recourse to their L1 may be doing them a disservice, for L1 can, in fact be extremely helpful.
c) Avoidance
Where certain structures are very different from L1, students may simply avoid using them. Schachter (1974) found that Chinese and Japanese learners of L2 English made less errors in the use of relative clauses than did Persian or Arabic learners – but this was because they tried to use them less often. This is because Persian and Arabic relative clauses are structured in a similar way to English ones, while the two Oriental languages treat them in a very different way.
It is difficult to know when a student is using avoidance as a strategy – he must show some evidence that he knows of the structure that he is avoiding, and it must also be so that a normal speaker of the target language would have used the structure in that situation.
Kellerman distinguishes 3 types of avoidance :
1. Learner can anticipate that there is a problem, and has some idea of what the correct form is like.
2. Learner knows the target form well, but believes that it would be too difficult to use in the circumstances in which he finds himself – free-flowing conversation, for example.
3. Learner knows how to use the target form, but will not do so because it breaks a personal rule of behaviour – ready use of ‘tu’ form by person coming from a culture where formality is highly valued.
d) Overuse
This may be a concomitant of avoidance. Students will use the forms that they know rather than try out the ones that they are not sure of. It may also reflect cultural differences – thus Olshtain (1983) found that American college students, learning Hebrew in Israel, were much more likely to use direct expressions of apology than were native speakers. This also seems to be true of English speakers of French.
How do teachers actually treat errors? In fact, there is considerable variation from one teacher to another, and also the treatment of error by any one teacher may vary from one moment to the next.
Studies of what teachers do have shown that very often they are inconsistent. Also, some errors are more likely to be treated than others – discourse, content and lexical errors receive more attention than phonological or grammatical errors – and here there is variation between native and non-native-speaker teachers. Many errors are not treated at all. Further, the more a particular kind of error is made, the less likely the teacher is to treat it. Moreover, teachers sometimes correct errors that have not taken place.
Another question is ‘Who does the repairing?’. In natural settings, there is a preference for self-initiated and self-completed repair. However, in the classroom, it is the teacher who initiates repair – at least during the language-centred phase – while he expects the student or one of his peers to produce the correct form.
Error treatment seems to have little immediate effect upon student production – thus the teacher may correct an error made by student A to have student B make exactly the same error five minutes later – and hear student A do it again before the end of the lesson!
Some experts – Krashen among them – have deduced that this suggests that correction is a pointless exercise. However, we should be aware that there are no studies as yet of the long-term effects of error correction.
What about students’ attitudes to error correction? In the main they say that they want to be corrected, both in the classroom, and in conversation with native speakers. However, when they are taken at their word, they feel uncomfortable with the resulting style of discourse.
Our recommendations for action can only be very tentative, and lack empirical backing. However, it would appear that the following rules are accepted by most members of the profession now – which does not mean that they are right!
1. Teachers should respect student errors – they are a part of the learning process. Respecting does not mean taking no notice of them, but it does mean that they are not to be treated as necessarily being evidence of stupidity, idleness or evil intent on the part of the learner.
2. Only treat those errors that students are capable of correcting, according to the state of their interlanguage at the time of the error. Written scripts should not be returned with simply everything underlined in red ink.
3. Self-repair is preferable to other-repair, as the student feels better about it. Being corrected by the teacher, or by other students, may be humiliating.
4. Teachers need to develop strategies for overcoming avoidance. The student needs to be put in a situation where he or she is forced to use the unassimilated structure and to think about the problems that this poses. However, this needs to be treated as a process of discovery rather than as a minefield.
Most important, remember that the students errors are a precious resource for the teacher, which inform her about the state of her pupils’ interlanguage. This is why it so important to avoid negative marking, where the student simply learns that if he makes an error he will lose points.