Paper

SECOND LANGUAGE ACQUISITION:

Interlanguage and The ‘Natural’ Route of Development

by : Group A Class E
Alim ( 05.62. )
Priawan R ( 04.62.0066 )
Titin I ( 04.62.0079 )
Erna S ( 04.62.0036 )
Novita ( 03.62.0000 )


Introduction

Language acquisition, the process of learning a native or a second language. Second language learners (L2) acquire a knowledge of a language of a second language in a fixed order as a result of a predisposition to process language data in highly specific ways. These claims stand in stark contrast to behaviourist accounts of second language acquisition (SLA), which emphasized the importance of environmental factors and first language (L1) interference. Where language acquisition was concerned, the key concept in the revised thinking about the process of learning was that of interlanguage. This was used to refer to the systematic knowledge of language which is independent of both the learners L1 and the L2 system he is trying to learn. Interlanguage was the theoretical construct which underlay the attempts of SLA is to identify the stages of development through which L2 learners pass on their way to L2 proficiency.
Language acquisition followed a ‘universal’ route that was largely uninfluenced by such factors as the age of the learner, the context in which learning took place, or the learner’s L1 background. According to this view of language acquisition, the controlling factor was the faculty for language that all human beings possess and which was also responsible for L1 acquisition. Inevitably the question arose as to what extent the order of development in language acquisition paralleled that in L1 acquisition. The validity of the L2 similar to L1 hypothesis has been a recurrent issue in language acquisition. However, although learner internal factors are powerful determinants of language acquisition, the conviction that they are capable of accounting for the entire process, which in some circles at least has been suggested, is not warranted.
To begin with, this paper will briefly consider the background theory in first language acquisition. It will then examine the notion of interlanguage, before reviewing in some detail the research upon which claims about a natural route of development have rested. Finally, a number of caveats regarding the centrality of learner internal processes in accounts of language acquisition will considered.

Mentalist Accounts of First language acquisition

Chomsky’s ( 1959 ) attack on Skinner’s theory of language learning led to reassertion of mentalist views of first language acquisition in place of the empiricist approach of behaviourists. He stressed the active contribution of the child and minimized the importance of imitation and reinforcement. He claim that the child’s knowledge of his mother tongue was derived from a Universal Grammar which specified the essential form that any natural language could take. As McNeill ( 1970: 2 ) put it:
The facts of language acquisition could not be as they are unless the concept of a sentence is available to children at the start of their learning. The concept of a sentence is the main guiding principle in a child’s attempt to organize and interpret the linguistic evidence that fluent speakers make available to him.


Lenneberg (1967) emphasized the biological prerequisites of language. Only homo sapiens was capable of learning language. Thus, whereas even severely retarded human beings were able to develop the rudiments of language, even the most socially and intellectually advanced of the primates, chimpanzees, were incapable of mastering the creativity of language. Lenneberg argued that the child’s brain was specially adapted to the process of language acquisition, but that this innate propensity was lost as maturation took place. His work provided empirical and theoretical support for the concept of a built in capacity for language as part of every human being’s biological endowment.
One further feature of mentalist accounts of language acquisition mentioning, the child built up his knowledge of his mother tongue by means of hypothesis testing.

In summary, therefore, mentalist views of L1 acquisition posited the following:
Language is a human specific faculty.
Language exists as an independent faculty in the human mind. For instance although it is part of the learner’s total cognitive apparatus, it is separate from the general cognitive mechanisms responsible for intellectual development.
The primary determinant of L1 acquisition is the child’s acquisition device, which is genetically endowed and provides the child with a set of principles about grammar.
The ‘acquisition device’ atrophies with age.
The process of acquisition consist of hypothesis-testing, by which means the grammar of the learner’s mother tongue is related to the principles of the ‘universal grammar’.

The incremental nature of L1 acquisition is evident in two ways. First, the length of children’s utterances gradually increases. Initially the utterances consist of one word. Later two words, then three and four word utterances follow. Second, knowledge of the grammatical system is built up in steps. Tense or the auxiliary ‘do’ are not acquired at the same time, but in sequence. Similarly, complex grammatical system such as negatives or interrogatives are learnt slowly in piecemeal fashion and involve rules quite unlike those in the target language.
According to mentalist accounts of L1 acquisition, language acquisition is a universal process. The term process, which is common in acquisitional studies, is used with two related meanings. It refers both to the sequence of development and to the factors that determine how acquisition takes place. The mentalist claim that the processes are internal and operate largely independently of environmental influences is no longer entirely defensible, as we shall see. First, however, it is necessary to consider the impact that mentalist theories and empirical research had on accounts of language acquisition.

Interlanguage

The concept of interlanguage to its background in mentalist views on language acquisition and then to show how early interlanguage theory provided an impetus for empirical research into both the nature of L2 errors and the sequence of development in language acquisition. The term interlanguage was first used by Selinker (1972). Various alternative terms have been used by different researchers to refer to the same phenomenon; Nemser (1971) refers to approximative systems, and Corder (1971) to idiosyncratic dialects and transitional competence.
The assumptions underlying interlanguage theory were stated clearly by Nemser (1971). They were: (1) at any given time the approximative system is distinct from the L1 and L2; (2) the approximative system form an envolving series; and (3) that in a given contact situation, the approximative systems of learners at the same stage of proficiency roughly coincide. Meanwhile, Corder (1967) made this comparison explicit by proposing that at least some of the strategies used by the L2 learner were the same as those by which L1 acquisition takes place. In particular, he suggested that both L1 and L2 learners make errors in order to test out certain hypothesis about the nature of the language they are learning. He saw the making of errors as a strategy, evidence of learner internal processing. This view of the language acquisition presented in the contrastive analysis hypothesis. Hypothesis testing was a mentalist notion and had no place in behaviourist accounts of learning.
Selinker (1972) suggested that five principal processes operated in interlanguage these were:
(1).language transfer (this was listed first, perhaps in deference to the contemporary importance attached to L1 interface);(2)overgeneralization of target language rules;(3)transfer of training(i.e. a rule enters the learner’s system as a result of instruction ); (4) strategis of L2 learning (I.e.’an identifiable approach by the learner to the material be learned(1972:37),and (5)strategies of L2 communication (i.e.’an identifiable approached by the learner to communicanition with native speakers’(1972:37)
The five processes together constitute the ways in which the learner tries to reduce the learning burden to manageable proportions and, as such, it has been suggested by Widdowson (1975b) that they can be subsumed under the general process of ‘simplification’. Learners have limited processing space and, therefore, cannot cope with the total complexity of a language system, so they limit the number of hypotheses they test at any one point in time.
Selinker also noted that many L2 learners (perhaps as many as 95 percent) fail to reach target language competence. That is, they do not reach the end of interlanguage continuum. They stop learning when their interlanguage contains at least some rules different from those of the target language system. He referred to this as fossilization. It occurs in most language learners and cannot be remedied by further instruction. Fossilized structures can be realized as errors or as correct target language forms. If, when fossilization occurs, the learner has reached a stage of development in which feature x in his interlanguage has assumed the same form will occur. If, however, the learner has reached a stage in which feature y still does not have the same form as the target language, the fossilization will manifest itself as error. On occasions, the learner may succeed in producing the correct target language form, but when the learner is focused on meaning especially if the subject matter is difficult he will ‘backslide’ towards his true interlanguage norm. Selinker and Lamendella (1978a) argue that the causes of fossilization are both internal and external. It can occur both because the learner believes that he does not need to develop his interlanguage any further in order to communicate effectively whatever he wants to, or it can occur because changes in the neural structure of his brain as a result of age restrict the operation of the hypothesis testing mechanisms. In one respect, however, mentalist theorizing cannot be easily carried over into language acquisition research. According to Chomsky and others, the true determinant of L1 acquisition was the child’s ‘acquisition device’, but this changed with age such that automatic, genetically-endowed language acquisition was not possible after puberty.
Selinker’s 1972 paper was seminal. It provided the theoretical framework for interpreting language acquisition as a mentalistic process and for the empirical investigation of language learner language. Subsequent discussions of interlanguage focused on its three principal features, all of which were raised by Selinker in one way or another. The following are the essential characteristics of language learner;
Language-learner language is permeable
In the sense that rules that constitute the learner’s knowledge at any one stage are not fixed, but are open to amendment.

Language-learner language is dynamic

The L2 learner’s interlanguage is constantly changing. However, he does not jump from one stage to the next, but rather slowly revises the interim system to accommodate new hypotheses about the target language system. This takes place by the introduction of a new rule, first in one context and then in another, and so on. A new rule spreads in the sense that its coverage gradually extends over a range of linguistic contexts.


Despite the variability of interlanguage, it is possible to detect the rule based nature of the learner’s use of the L2. For the L2 learner, however, the true norms are contained in the interlanguage system he has constructed. Interlanguage theory was based on ‘behavioural events’. Early interlanguage theory, then, was closely associated with Error Analysis. As this served as one of the main devices for examining the processes of interlanguage, the principles and methodology of Error Analysis will be considered briefly.

Language-learner language is systematic

Interlanguage was based on ‘behavioural events’. As Selinker acknowledged, the behavioural events that have aroused the greatest interest in discussion of SLA have been ‘errors’. Early interlanguage theory, then, was closely associated with error analysis. As this served as one of the main devices for examining the processes of interlanguage, the principle and methodology of error analysis will be considered briefly.

Error Analysis

Error Analysis provides two kinds of information about interlanguage. The first which is relevant to the first of the two questions posed above concerns the linguistic type of errors produced by L2 learners. Error Analysis provides a synchronic description of learner errors, but this can be misleading. The second type of information which is relevant to the question about the strategies used in interlanguage concerns the psycholinguistic type of errors produced by L2 learners.
The most significant contribution of Error Analysis, apart from the role it played in the reassessment of the Contrastive Analysis Hypothesis, lies in its success in elevating the status of errors from undesirability to that of a guide to the inner workings of the language learning process. As a result of interlanguage theory and the evidence accumulated from Error Analysis, errors were no longer seen as ‘unwanted forms’ (George,1972), but as evidence of the learner’s active contribution to second language acquisition.

Origins of Interlanguage

The starting point of interlanguage becomes a major issue when second language acquisition is seen as a recreation continuum rather than a restructuring continuum. If the learner builds up his interlanguage by gradually increasing the complexity of the system, what is his starting point? Corder (1981) considers two possibilities. One is that the learner starts from scratch in the same way as the infant acquiring his mother tongue. But Corder considers this possibility unlikely, as it is highly implausible that entire processes of language acquisition will be replicated. The second possibility is that the learner starts from ‘some basic simple grammar’ (Corder,1981:150). Corder suggests that language learners regress to an earlier stage in their own linguistic development before starting the process of elaboration. In other words, the starting points is the learner’s knowledge of how to get a message across without the help of grammar.
This part of any language user’s capacity and can be called upon whenever there is a need to create ‘simple’ messages. The same process is evident in the speech that mothers address to children, and that native speakers address to L2 learners to facilitate communication. In general, however, the question of the origins of interlanguage is still an open one, subject to speculation. There have been very few empirical studies of the initial stage of the interlanguage continuum.

Variabilty in Interlanguage

The natural route does not manifest itself in a series of clearly delineated stages. Rather each stage overlaps with the one that precedes and follows it. Each new rule is slowly extended over a range of linguistic contexts. Therefore, at any given stage of development, the learner’s interlanguage system will contain a number of competing rules, with one rule guiding performance on one occasion and another rule on a different occasion. In addition, each interlanguage system contains linguistic forms that are in free variation; that is, forms that are not guided by rules and whose use is not systematic at all. These types of variability pose both practical and theoretical problems for second language acquisition research. The practical problems concerns concern how to collect data in order to study language learner language. If the learner’s performance varies from one task to another, how can the researcher evaluate the particular data he has collected? The theoretical problems concern how to reconcile the inherent variability of the learner’s performance with claims that interlanguage is systematic, and how to account for non systematic variability. Not all non-sytematic variability, however, involves free variation. Psycholinguistic factors to do with the learner’s emotional or physical condition can lead to slips, hesitations, and repetitions. This type of variability can be called performance variability.
Variability is, of course, not only a characteristic of language learner language. It occurs in language use. Before we turn to variability in interlanguage, therefore, the general nature of variability in interlanguage, therefore, the general nature of variability in language use will be considered.
It is possible to identify two basic types of variability: systematic and non-systematic variability. Each of these will be described separately.

Systematic Variabilty

The status of the variability in the language use of native speakers has been viewed very differently by linguist, depending on whether they approach it in terms of a homogeneous competence model or a heterogeneous competence model. In a homogeneous competence model all variability is classified as non-systematic variability. That is, the language user is credited with a system of rules (his ‘competence’) which is homogeneous. In this competence model, linguistic knowledge is separated from non linguistic knowledge, even though this may be called upon when linguistic knowledge is put to use. According to the heterogeneous competence model, therefore, the user’s knowledge of language rules interlocked with his knowledge of when, where, and with whom to use them. According to this view of language, the user’s ability to perform needs to be understood in terms of communicative rather than linguistic competence. In a heterogeneous competence model, therefore, stylistic variability is viewed as an integral part of competence, not performance.
Labov (1970) operates within a heterogeneous competence model. He lists five axioms about how to study language use. Because these are of central importance for understanding the nature of user variability, they are summarized below.
All speakers possess several ‘styles’. That is, they adapt their speech to make it fit the social context.
‘Styles can be ranged along a single dimension, measured by the amount of attention paid to speech.’ A language user varies in the degree to which he is able to monitor his speech in different situations.
The vernacular is the style in which minimum attention is given to monitoring speech. It is the style associated with informal, everyday speech. It provides ‘the most systematic data’ for linguistic study.
It is possible to tap the vernacular style of the user by systematic observation of how he performs in a formal context (such as an experiment).
The only way to obtain good data on the speech of a language user is through systematic observation.
Labov also deals with variability determined by linguistic factors. he examined the use of the copula (‘be’) in utterances produced by speakers of Black English Vernacular in New York and found that the presence or absence of the copula was largely systematic. Its use depended on the specific linguistic context provided by an utterance. For instance, the copula was more likely to be used when the preceding syntactic environment constituted a noun phrase rather than when it was a pronoun. Similarity, it was more likely to occur when the following grammatical structure was ‘gonna’ than it was a noun phrase. It is important to note, however, that the systactic environment did not determine the absolute occurrence of the copula, only its relative frequency. Thus ‘be’ was variable in all contexts, but certain contexts predisposed the speaker to use it more frequently than other contexts.
This approach involves viewing language as an intersecting set of idiolects. Any one speaker will have access to a number of such idiolects, but is unlikely to be capable of the full range. Furthermore, these features can be seen to be hierarchically ordered, such that the presence of one feature in an idiolect entails the co-occurrence of one or more other features. The important points are that variation is systematic, governed by both situational and linguistic factors, and that the user’s linguistic form presupposes the presence of one or more other forms.

Non-Systematic Variability

Native speaker language use is also characterized by non-systematic variability. As discussed earlier, this is of two types. Performance variability, however, is not part of the user’s competence. It occurs when the language user is unable to perform his competence. It is the second type of non-systematic variability – free variability – which is of greater interest for understanding second language acquisition and which will, therefore, be discussed here.
In general, non systematic variability has not received much attention from linguists, who have preferred to explore systematic variability. Bickerton (1975), however, has disputed the prevailing ‘contextual theory’ of linguistic variation. He observes as follows:
While with the help of a little hindsight, a plausible contextual explanation can be given for many stylistic shifts, there are many more that operate in quite unpredictable ways. (Bickerton,1975:183).


Bickerton points to Labov’s famous example of phonological change in the speech of Martha’s Vineyarders. They attempted to maintain their own group identity in the face of an influx of outsiders by their choice of a specific sound. However, the sound they used was not specially introduced, but rather a sound that already existed in their speech and which they intensified. In other words, a sound which began its existence in free variation with other sounds came to be used systematically to convey social meaning. This process of exploiting a form in free variation is a key feature of interlanguage development.

The same types of variability are evident in interlanguage as in natural languages. However, although there is no difference in kind, there is difference in degree; variability is extensive in interlanguage. The following account will first consider contextual variability and then free variability.

Contextual Variability

It has been pointed out that contextual variability is of two kinds: that which is determined by situational context and that which is determined by linguistic context.
Variability that is the result of situational context is analogous to the stylistic variability observed in native-speaker usage and described in previous section with reference to Labov’s work. Dickerson (1975) examined the occurrence of /z/ in the speech of ten Japanese speakers studying English at University level. She collected data on three separate occasions in a nine month period, using a three part test involving (1) free speaking, (2) reading dialogues aloud, and (3) reading word lists aloud. Dickerson found that the correct target language variant was used most frequently in (3) and least frequently in (1), with the frequency in (2) in between. This order was maintained over the three occasions. The same order was observed for other variants that were closest in form to the target language sound. In other words, Dickerson found that L2 learners employed multiple variants (one of which might be the correct target language form, but need not be). They used the target language variants or those variants linguistically closest to it in situations where they were able to audio-monitor their speech, and those variants linguistically distant from the correct target language form in situations where audio-monitoring was not possible. Development over time involved an increase in the proportion of the target-like sounds.
Other studies of phonological development have come up with observations similar to those of Dickerson. Schmidt (1977), for instance, noted that Arabic-speaking students of English became more accurate in their use of English ‘th’ sounds in a formal task than in an informal task. Also Schmidt noted that these learners did exactly the same in L2. English as they did in L1 Arabic, where they also style-shifted from relative low to high frequency in the use of ‘th’ sounds, depending on whether they were speaking colloquial Arabic (associated with formal situations). In other words, the patterns of style-shifting were the same in L1 and L2. Bebee (1980) also found evidence of the direct transfer of a formal feature of the L1 into the L2 when the situation demanded a formal style. The studies by Schmidt and Bebee, while reinforcing the overall picture provided by Dickerson’s study, make it clear that when a learner is able to attend closely to his speech (i.e. in a careful style), he may produce not only a higher incidence of target language forms, but also a higher incidence of L1 forms, if these are associated with formal use in the L1. Language transfer, therefore, is also a variable phenomenon.
Learners from a variety of language backgrounds always included the second verb in such sentences in free oral production, but increasingly omitted it in proportion to the degree of monitoring permitted by different tasks (i.e. elicited imitation, written sentence-combining, and grammatical judgment). Lococo (1976) found that twenty-eight university students enrolled in an elementary Spanish course produced systematically fewer errors in adjectives, determiners, and verbs when the task was translation than when it was free composition, with the scores for the picture description task intermediary. Thus it is seen once again that when the learner is able to monitor his performance, he produces target language forms with greater regularity.
At one end of the continuum is the vernacular style, which is called upon when the learner is not attending to his speech. This is the style that is both most natural and most systematic. At the other end of the continuum is the careful style, which is most clearly evident in tasks that require the learner to make a grammatical judgment (e.g. to say whether a sentence is correct or incorrect). The careful style is called upon when the learner is attending closely to his speech. Thus the stylistic continuum is the product of differing degrees of attention reflected in a variety of performance tasks. Variability as a result of the linguistic contexts occurs when two different linguistic contexts induce different forms even though in the target language they require the same form. For example, the learner might produce correct exemplars of the third person singular ‘-s’ when the linguistic context consist of a single clause utterance as in:

Mr Smith lives in Gloucester.

but fail to do so when the linguistic context consists of a subordinate clause as in:
Mr Smith who live in Gloucester married my sister.
This variability may not involve a correct target language form at all. it may consist of the use of two (or more0 deviant forms.

Dickerson’s (1975) study also provides evidence of contextual variability according to linguistic context. She noticed that her subjects used a number of variants for English /z/ according to the linguistic environment. The phonetic quality of the sound they produced depended on what consonants and vowels were adjacent to /z/. Thus when /z/ was followed by a vowel, the learners used the correct target language form every time, even on the first occasion; but when /z/ was followed by silence, they used three variants, only one of which was /z/. Progress from one occasion to the next consisted of the increased use of /z/ in environments where initially it was little used, and also of increased use of variants that were phonetically closer to /z/.
The effects of the linguistic and situational context interact to influence jointly the learner’s use of interlanguage forms. If linguistic contexts are seen as continuum, then the use of any particular interlanguage form can be plotted on the intersection of these two continua. For example, it can be predicted that, to begin with, use of the third person singular will be most frequent when the linguistic context is simple and the style careful.
Conclusion
Second language acquisition was seen as a series of evolving system which comprised the interlanguage continuum. Each system was considered to be internally consistent, in the sense that it was rule-governed. It was, however, also permeable to new rules, and, therefore, dynamic. The continuum was initially viewed primarily as a restructuring continuum stretching from the learner’s L1 to the target language. Later it was viewed as a recreation continuum in which the learner gradually added to the complexity of interim systems. Most learners never reached the final stage of the continuum; their interlanguage fossilized some way short of target language competence.
Both interlanguage theory and the empirical studies that supported it have had a major impact on our thinking about the nature of SLA. The switch from a behaviourist to a mentalist framework proved a source a great insight into both L1 acquisition and SLA. It has become generally accepted that the human language faculty is a potent force in language in language acquisition. But internal processing and innate mechanisms are not the full story. It is also necessary to consider how the linguistic input contributes to the process of acquisition.
The process of acquisition is related to this variability. (1) The learner slowly extends the contextual range of the forms he has acquired, by mastering their use in additional stylistic and linguistic contexts. (2) The learner slowly resolves the free variability that exists in his interlanguage by developing clear form-function relationships. Both of the processes will be examined in detail. First, however, the different ways in which variability in language learner language is handled in SLA research will be considered.


References
Beebe, L. 1980. ‘Sociolinguistic variation and style shifting in second language acquisition’ Language Learning 30/1: 433-47
Bickerton, D.1975. Dynamics of a Creole System. Cambridge: Cambridge University Press.
Chomsky, N.1959. Review of Verbal Behaviour by B.F Skinner. Language 35 : 26-58
Corder, S. 1981a. Error Analysis and Interlanguge. Oxford: Oxford University Press.
Dickerson, L. 1975. ‘Interlanguage as a system of variable rules. ‘TESOL Quarterly 9: 401-7.
George, H. 1972. Common Errors in Language Learning: Insights from English. Rowley, Mass,: Newbury House.
Labov, W. 1970. ‘The Study of Language in its Social Context.’ Studium Generale 23 : 30-87.
Lenneberg, E. 1967. Biological Foundations of Language. New York: Wiley and Sons.
Lococo, V. 1976. ‘A comparison of three methods for the collection of L2 data: free composition, translation and picture description.’ Working Papers on Bilingualism 8:59-86.
McNeill, D. 1970. The Acquisition of Language. New York: Harper Row.
Schmidt, R. 1977. Sociolinguistic Variation and Language Transfer in Phonology.’ Working Papers on Bilingualism 12: 79-95
Selinker, L. 1972. ‘Interlanguage.’ International Review of Applied Linguistics X: 209-30.
Selinker, L. and J. Lamendella. 1978a. ‘Fossilization in Interlanguage’ in C. Blatchford and J. Schachter (eds.). On TESOL ’78: EFL Policies, Programs, Practices. Washington D.C.:TESOL.