The learning of a second language
4.2. Main models or theories of SLA
As we said at the beginning of this subject, explaining the process of SLA is not an easy task and, in fact, none of the existing theories has been able to provide a full and exhaustive account of all the variables intervening in the language learning process. It will be impossible to deal in detail with each and every single one of the large number of the SLA models proposed. However, with the purpose of systematising this broad area, it is possible to identify four main SLA model areas. These four areas should not be regarded as completely distinctive since there is some degree of overlap between them. We will be looking briefly at each of the models listed below later in this chapter, and in doing so we will again mainly follow Ellis (1985, 1994, 2009).
- Sociolinguistic models
These give primary importance to the role of environment and to the attitudes of the learners to the target language community. Schumann's Acculturation theory together with Andersen's Nativisation, and Giles and Byrne's Accommodation models can be classified under this heading.
- Linguistic models
These focus specifically on the language being learned. Here we can include Hatch's Discourse Theory, Ellis' Variable Competence Model and Chomsky's theory of Universal Grammar.
- Cognitive Models
These maintain that learners construct internal representations of the language being learned. One may think of these internal representations as 'mental pictures' of the target language. Krashen's Monitor Theory, Lamedendella's Neurofunctional system, Pienenann's Multidimensional System, McLaughlin's Information Processing Scheme and Rumellhart's and McClelland's Parallel Distributed Processing Model (1986) can be considered as cognitive in nature.
4.2.1. Sociolinguistic models
4.2.1.1. The acculturation model (Schumann 1978)
This model was devised with the sole aim of explaining language acquisition by L2 immigrants in second language environments. It specifically excluded those settings where learners received formal instruction.
It was Schumann (1978) who first coined the term "acculturation", by which he meant the process of adapting to a new culture. In his view, "second language acquisition is just one aspect of acculturation" and there exists a perfect match between the degree to which a learner acculturates to the target language group and the degree to which s/he acquires the second language.
Acculturation implies social and psychological distance between the learner and the target language culture. Social variables govern whether the learning situation is "good"
or "bad". On the other hand, psychological factors are affective in nature and they include:
a) language shock, b) culture shock, c) motivation, and d) ego boundaries.
Schumann also describes the kind of learning which takes place. According to him, the first stages of SLA are characterized by the same features and processes which are responsible for the formation of pidgin languages, that is, languages with very basic grammatical structures. Schumann refers to this as the pidginisation process. When pidginisation persists, learning fossilises, that is, learners do not make any progress in their learning.
4.2.1.2. The nativisation model (Andersen 1979)
Andersen puts together some of Schumann's principles as well as his own emphasis on the learner's internal processing mechanisms. According to this model, SLA is the result of two general factors: nativisation and denativisation. Nativisation consists of assimilation; learners make input conform to their own internalised view of what constitutes the L2 system. They simplify the learning task by testing hypotheses based on their knowledge of the world and of the L1. Nativization is apparent in pidginisation and the early stages of both first and second language acquisition.
Denativisation involves accommodation; learners adjust their internalised system to suit input by the use of a series of inferencing strategies which enable them to remodel their interlanguage system in accordance with the "external norm".
It is important to point out that both nativisation and acculturation models try to account for the fact that second language learners do not reach native-like competence in the L2 in the same way as L1 learners. Social and psychological distance are held to be responsible.
4.2.1.3. Accommodation theory (Giles and Byrne 1982)
This theory has much in common with Schumann's model. Giles and Byrne are concerned with successful language acquisition in multilingual environments.
They believe that it is the way in which the learner's social group (the
"ingroup") defines itself in relation to the target language community (termed the "outgroup") that is important for SLA. For Giles and Byrne the intergroup relationships are dynamic and fluctuate according to the shifting views of identity held by each group in relation to the other.
According to Giles and Byrne, as for Gardner and Lambert (1972), motivation is the primary determinant of L2 proficiency. They consider motivation to be a reflex of how individual learners define themselves in ethnic terms. This is governed by a series of key variables:
KEY VARIABLES
1. Identification with in-group. Weak identification Strong identification. the learner sees his/her group as holding a low or high status and as sharing or excluded from institutional power).
Low perception. High perception.
4. Perception of in-group boundaries (whether the learner sees his/her in-group as culturally &
linguistically separate from the out-group or as culturally and linguistically related).
Soft and open. Hard and closed.
5. Identification with other social categories (whether the learner identifies with few or several other in-group social categories (religious, gender) & as a consequence whether s/he holds an adequate or inadequate status within their in-group).
As with Schumann's theory, this model of SLA only applies to second language environments and does not account for the developmental sequence. Of particular interest is Giles and Byrne's attention to linguistic output which attempts to explain the differences found between learners' use of 'ingroup speech markers'. If learners are positively motivated towards the target community, ingroup speech markers are scarcely noticeable, while the opposite is true if learners are negatively motivated towards the outgroup.
4.2.2. Linguistic models
4.2.2.1. Discourse theory (Hatch 1978)
According to this view, language development should be considered in terms of how the learner discovers the meaning potential of language by participating in communication.
This is similar to Halliday's view on FLA.
The main principles of this model can be described as follows:
1. SLA follows a "natural" route in syntactic development.
2. Native speakers adjust their speech in order to negotiate meaning with non-native speakers. This means that negotiation of input is a key factor in SLA.
3. The conversational strategies used to negotiate meaning, and the resulting adjusted input influence the rate and route of SLA in a number of ways:
- the learner learns the grammar of the L2 in the same order as the frequency order of the various features in the input;
- the learner acquires commonly occurring formulas and then later analyses these into their component parts;
- the learner is helped to construct sentences vertically; vertical structures are the learner utterances which are constructed by borrowing chunks from the preceding discourse and then adding elements of the learner's own.
- The natural route to SLA is the result of learning how to hold conversations.
This theory has been strongly criticised by those theorists who believe that successful language learning can actually be achieved without negotiation of input, as for example by self and independent study. As with the previous models, learner's cognitive processes are not considered.
4.2.2.2. The variable competence model (Ellis 1984; Tarone 1983; Widdowson 1979; Bialystok 1978)
This model is based on two main distinctions:
a) The first refers to the process of language use, and b) The second is related to the product.
This theory claims that the way a language is learned is really a reflection of the way it is used. The product of language is said to comprise a continuum of discourse types ranging from entirely unplanned (spontaneous, lacks preparation) to entirely planned (carefully thought out).
The process of language is understood in terms of the distinction between the linguistic knowledge of rules (theoretical knowledge) and the ability to make use of this knowledge or procedures (procedural knowledge). Language users make the knowledge of linguistic rules work according to both the situational and the linguistic context. They actualise their abstract knowledge of sentences to create utterances in discourse.
From this we gather that the product of language use is the result of either or both of the following:
- a variable competence, i.e. a heterogeneous rule system available to the learner.
- variable application of procedures for actualising knowledge in discourse.
Broadly speaking, the main features of this model can be summarised as follows:
- There is a single knowledge store containing variable interlanguage rules according to how automatic and how analysed the rules are.
- The learner possesses a capacity for language use which consists of primary and secondary discourse and cognitive processes.
- L2 performance is variable as a result of whether primary processes employing analysed L2 rules are utilised in unplanned discourse, or secondary processes employing analysed L2 rules are used in planned discourse.
- Development, that is, progress in language learning occurs as a result of
- acquisition of new L2 rules through participation in various types of discourse, that is, new rules arise in the application of procedural knowledge;
- activation of L2 rules which initially exist in either a non-automatic analysed form or in an automatic analysed form so that they can be used in unplanned discourse.
4.2.2.3. The universal hypothesis (Chomsky 1976, 1986)
The fundamental hypothesis in this model is that all languages have certain basic structures in common - these are called universals. The study of linguistic universals has contributed to explanations of SLA in two ways. First, it has been proposed that the linguistic properties of the Target Language (TL), that is, the language to be learned, may vary according to the degree of effort needed to acquire them; furthermore, the complexity of these language features is in its turn determined by their condition as universal, that is, common to all languages, or as language-specific. This means that those properties of the TL which are common to many or all languages are easy to learn in comparison to those properties that are found in few languages, or only in the target language.
Second, linguistic universals can be used to predict which differences lead to difficulty and which ones do not. Thus, the study of linguistic universals has led to a reconsideration of transfer theory. Chomsky maintained that by studying a single language in depth, it would be possible to discover the highly abstract principles of grammar that constrain the form of any specific grammar. These principles are known as Universal Grammar (UG).
Greenberg (1966) and Comrie (1984), following a similar direction to that of Chomsky's, tried to identify universals by examining a wide range of different languages with the purpose of discovering the features in common. They refer to these as typological universals.
UG is said to innate. It can be defined as the set of principles which apply to all languages and is composed of two types of universals:
- substantive, which include distinctive phonetic features of which sounds are made, and syntactic categories such as noun, verb, subject, object.
- formal, which are statements about what grammatical rules are possible. These universals are more abstract than the previous ones.
Universals constitute constraints on the kind of grammar that a child can develop. Thus it has been claimed that it is impossible for a child acquiring the L1 to produce utterances which contravene UG - these 'impossible' utterances are known as "wild grammars" (Goodluck 1986). However, the mere presence of UG alone is not enough to explain how children achieve the complex task of FLA. As well as having the innate capacity of UG- a special 'language faculty'- children need input data in order to set the parameters of UG as it relates to their specific L1. Furthermore, the grammar of a language contains two parts: core and periphery. Core grammar is equivalent to the rules that the child discovers with the help of UG. Periphery grammar corresponds to the elements which are not constrained by UG. Core rules are unmarked since they accord with the general tendencies of language; in contrast, periphery rules are marked since they are exceptional in some way.
4.2.2.4. Swain's Output thesis
Along similar lines to Hatch's ideas, Merrill Swain's work brought about one of the great unresolved controversies in the field of Applied Linguistics, largely due to its apparent rejection of Krashen's main thesis. The Comprehensible Output thesis, developed by Swain, states that learning takes place when encountering a 'gap' in the linguistic knowledge of the L2. By noticing this gap the learner becomes aware of it and might be able to modify his output so that he learns something new about the language.
Although Swain does not claim that comprehensible output is solely responsible for all or even most language acquisition, she does claim that under some conditions, CO facilitates second language learning in ways that differ from and enhance input due to the mental processes connected with the production of language. This hypothesis is closely related to the 'Noticing Hypothesis' (Schmidt 1990). Swain defines three functions of output:
1. The Noticing function: Learners encounter gaps between what they want to