Understanding Second Language Acquisition (2nd ed)

by Rod Ellis

  By viewing learning itself as primarily a social activity—albeit one that involves cognitive processes. This view sees learning as embedded in the processes by which social contexts are constructed through interaction. In other words, learning is no longer seen as essentially an individual, cognitive affair but as a participatory and social affair.

  It is this latter sense of the ‘social turn’ that critics of cognitive SLA have promoted and which is the focus of this chapter. However, my starting point will be SLA research based on a structural view of social context and its role in L2 acquisition. This research predated the later research and thus—from a historical perspective—is an appropriate starting point. I will then move on to examine in some detail the objections to ‘cognitive SLA’ and consider the alternative social (and social-cognitive) theories that this critique gave rise to.

  Social factors and L2 achievement

  The social context in which learners live and work has an effect on how successful they are in learning an L2. This is especially true of learners in a second language context—i.e. a context where the majority or official language is the target language as in the case of L2 learners of English who are living in the United States or United Kingdom. However, it is also true of learners in a foreign language context (i.e. a context where the target language is not the official or majority language as, for example, with Chinese learners of English in China). In both cases, social factors determine the opportunities that learners have to engage with the L2 and their motivation to do so.

  The two theories I will now consider focus respectively on the role of social factors in second language settings and foreign language classrooms.

  Schumann’s Acculturation Model

  Schumann’s Acculturation Model aimed to account for the variable success of immigrants in second language settings. The key concept was ‘acculturation’. As Schumann (1978a) put it:

  … second language acquisition is just one aspect of acculturation and the degree to which a learner acculturates to the target-language group will control the degree to which he acquires the second language (1978a: 34).

  Schumann focused on one of the six learners (Alberto) studied by Cancino, Rosansky, and Schumann (1978). Whereas the other five manifested considerable development over the ten-month period of the study, Alberto did not advance in most of the structural areas Schumann investigated. Table 9.1 provides a summary of Schumann’s ten-month study of Alberto.

  Learner’s background Alberto was a 33-year-old lower middle class Costa Rican working in a factory in the USA with other non-native speakers of English. He lived within a small Spanish-speaking minority in a Portuguese area.

  Data collection Three types of audio recorded data were collected:

  spontaneous speech recordings of conversations with the researcher

  experimental elicitations (e.g. transforming a positive into a negative utterance)

  pre-planned sociolinguistic interactions involving taking Alberto to parties, restaurants, and sports events in order to collect speech in varied natural situations.

  Analysis The analysis focused on specific grammatical structures (e.g. negatives, interrogatives, and auxiliary verbs). It involved frequency analysis (see Chapter 4).

  Main findings Alberto showed very little linguistic development during the period of study—for example, his negatives were predominantly of the ‘no’ + verb type (‘I no use television’), a large proportion of his interrogatives were not inverted (‘Where you get that?’), and only one auxiliary (copula is) was mastered. In short, Alberto’s English was ‘pidginized’ (i.e. manifested the same features as pidgin languages).

  Discussion Schumann considered a number of factors that could explain Alberto’s lack of development. He dismissed ability and age as factors and instead examined the factors that could result in pidginization. Schumann also suggested that Alberto experienced substantial psychological distance (for example, displayed negative attitudes and little motivation to learn English).

  Table 9.1 Schumann’s (1978b) study of Alberto

  The extent to which learners acculturate determines the level of their social distance from the target language community. Schumann identified a number of social factors that influence the degree of social distance:

  Social dominance—i.e. whether the L2 group is politically, culturally, technically, or economically superior (dominant), inferior (subordinate), or equal to the target language group.

  Integration pattern—i.e. whether the L2 group assimilates (gives up its own lifestyle and values in favour of those of TL group), seeks to preserve its lifestyle and values, or acculturates (adopts lifestyle and values of TL group while maintaining its own for intra-group use).

  Enclosure—i.e. the extent to which the L2 group shares the same social facilities (low enclosure) or has different social facilities (high enclosure).

  Cohesiveness—i.e. the extent to which L2 group is characterized by intra-group contacts (cohesive) or inter-group contacts (non-cohesive).

  Size—i.e. whether the L2 group constitutes a numerically large or small group.

  Cultural congruence—i.e. whether the culture of the L2 group is similar or different from that of the TL group.

  Attitude—i.e. whether the L2 group and TL group may hold positive or negative attitudes towards each other.

  Intended length of residence—i.e. whether the L2 group intends to stay for a long time or a short time.

  A learning situation can be ‘bad’ or ‘good’ (Schumann 1978a) depending on social distance. Overall, Alberto’s situation was a bad one: he belonged to an L2 group that was socially inferior to the TL group; neither his group nor the target language group sought assimilation; his L2 group was characterized by high enclosure; the L2 group was relatively cohesive and large; the level of cultural congruence was not high; negative attitudes predominated in both groups. As a result, Alberto did not acculturate and had very little need to communicate with members of the TL community.

  While the Acculturation Model affords a convincing explanation of Alberto’s lack of progress in learning English, a true test of the theory would require proof that learners who vary in the extent of their social distance also vary in the extent to which they acquire the L2. A number of studies (see Schumann 1986b) attempted to show this, but with mixed results. One reason for this lies in the difficulty of measuring the social variables that are claimed to influence acculturation. Apart from the problem of obtaining reliable measures of each social factor, there is no principled way of weighting the different variables. Another problem with the theory is that it has nothing to say about how social factors influence the quality of contact that learners experience. The model is premised on the assumption that social distance determines the quantity of contact. Finally, the theory is clearly deterministic; that is, social factors determine the extent to which learners participate (or do not participate). It takes no account of learner agency—i.e. the ability of learners to confront and overcome potentially negative social conditions through their own actions.

  The Socio-educational Model

  The Socio-educational Model was introduced briefly in Chapter 3. In contrast to the Acculturation Model, it was developed to explain L2 learning in classroom settings, in particular the foreign language classroom. The model is socio-psychological in nature—it sought to relate social and psychological factors to explain differences in L2 achievement—and was based on the importance that Gardner (1985) attached to ‘integrativeness’ (i.e. the individual’s willingness to identify with members of the target language group). The model plots the connections between four aspects of L2 learning: (1) the social and cultural milieu; (2) individual learner differences; (3) the setting, and (4) learning outcomes. It proposes that L2 learning—even in a classroom setting—is not just a matter of learning new information but of ‘acquiring symbolic elements of a different ethnolinguistic community’ (Gardner 1979: 193).

  Our primary interest here lies in th
e role the model attributes to the social/cultural milieu. It predicts that the relationship between the social/cultural milieu and L2 achievement is an indirect one. Factors in the social and cultural milieu are seen as causally related to ‘integrativeness’, which—in turn—is causally related to motivation and—via this—to achievement. Gardner’s research (for example, Gardner 1985) lends support to the basic claim of the model, namely that there are identifiable causal paths between the different components of the model. Social factors impact on the nature and level of the individual learner’s motivation, which in turn affects achievement.

  Missing from the model, however, is any account of how particular social factors operative in a social/cultural milieu influence attitudes. In this respect, the socio-educational model is much more limited than the Acculturation Model. Like the Acculturation Model, however, the Socio-educational Model is deterministic. There is no allowance for the possibility that learners can themselves influence their social/cultural milieu.

  SLA—a cognitive or a social enterprise?

  Both the Acculturation Model and the Socio-educational Model are de facto input-output models. That is they assume that L2 learning is a mental phenomenon involving individual learners. Learning is conceptualized in terms of ‘success’ and ‘achievement’ with social factors having an indirect effect through learners’ access to input and their motivation to make use of it for learning. This is a position that has been challenged by post-modernist accounts of L2 acquisition which dispute the existence of a simple linear relationship between social context and learning.

  Critiques of cognitive SLA

  One of the first critiques of cognitivist SLA was Firth and Wagner (1997). They claimed that there was an imbalance in SLA research in favour of cognitive-oriented theories and methodologies and argued that there was a need to reconceptualize SLA as a more balanced enterprise by acknowledging the social dimension of L2 acquisition. They suggested that this required three major changes:

  (a) a significantly enhanced awareness of the contextual and interactional dimensions of language use, (b) an increased emic (i.e. participant-relevant) sensitivity towards fundamental concepts, and (c) the broadening of the traditional SLA data base.

  (Firth and Wagner 1997: 286)

  In relation to (a), they emphasized the importance of recognizing that meaning ‘is not an individual phenomenon consisting of private thoughts executed and transferred from brain to brain, but a social and negotiable product of interaction, transcending individual intentions and behaviours’ (p. 290). They objected, for example, to the input modification studies we considered in Chapter 7 on the grounds that there is no such thing as a baseline way of talking and interacting. They also challenged the assumption that talk involving L2 learners is inherently problematic, arguing that what cognitivist researchers view as deviant and problematic forms may simply reflect learners’ resourceful and strategic use of their available linguistic resources. They emphasized the need to consider the ‘local agenda’ that figures in specific interactions rather than applying a pre-determined universal model for the negotiation of meaning.

  In relation to (b), they challenged the status of some fundamental SLA concepts—in particular those of ‘native speaker’ and ‘non-native-speaker’. They argued that these terms ignore the participants’ own sense of social reality. When learners speak, they do not do so as ‘non-native speakers’ but as a friend, a guest, an expert or a novice, a stranger, etc. Similarly, when so-called native speakers talk to learners they do not do so as native speakers, but from whatever social identity is relevant to the social context.

  ‘An approach that takes the participants’ own perspective into account—i.e. an emic approach—requires acknowledging that they have particular social identities which shape how specific encounters unfold. They also rejected the view of the learner as a ‘defective communicator’ that is implicit in cognitive-interactionist SLA.

  For Firth and Wagner (1997), the traditional SLA data base is overly narrow as it consisted mainly of interactions from a laboratory or classroom setting. They argued that the L2 is a means of participating in a whole range of different social contexts, with typical encounters involving multilingual participants communicating with each other for social purposes. Thus, it is necessary to investigate how learners communicate in real-world situations. This criticism of cognitivist SLA, however, is perhaps less valid for—as we have seen—there has been plenty of research that has drawn on a rich, ‘naturalistic’ data base, including Schumann’s study of Alberto; the European Science Foundation work on the basic variety (see Chapter 4); and—from a structural perspective—the work on variability in learner language (see Chapter 5).

  Like Firth and Wagner, Block (2003) sought to broaden the scope of SLA by adopting a more interdisciplinary and socially-informed approach. He critiqued the way that the terms ‘second’, ‘language’, and ‘acquisition’ are conceptualized in the input-output model of L2 acquisition. He argued that the term ‘second’ fails to capture the experiences of multilingual speakers and proposed that a better term would be ‘other’ or ‘additional’NOTE 1. Block pointed out that in mainstream SLA ‘language’ has been interpreted as referring to linguistic competence and that it has failed to examine communicative competence. He noted that ‘acquisition’ has been understood in terms of information-processing models (as we saw in Chapter 7) and that such models typically fail to recognize that the processes involved are social and external as much as they are mental and internal. Above all, Block argued for rich, contextualized descriptions of learners and the interactions they participate in.

  It is important to note, however, that both Firth and Wagner and Block did not reject the importance of cognitive aspects of L2 acquisition. Firth and Wagner concluded their article by emphasizing that language is both a social and a cognitive phenomenon and that what was needed was to investigate ‘how language is used as it is being acquired through interaction’ (p. 296, italics in original). Block considered it possible to integrate information processing and sociocultural approaches in a single theory. However, by and large the ‘social turn’ has prioritized the social over the cognitive.

  Responses to the critique

  Firth and Wagner’s (1997) article led to a number of responses published one year later in The Modern Language Journal. Some of these responses (for example, Rampton 1998) were clearly sympathetic to Firth and Wagner’s position whilst others were dismissive. The central point of contention was the site of acquisition. Is acquisition a matter of taking in and thereby possessing knowledge of the L2? The term ‘acquisition’ implies this—i.e. that we can ‘acquire’ a language in the same way we acquire a house or car. Alternatively, is acquisition best conceived of as ‘participation’—the taking part in social activity, which becomes the site of learning? It was this latter view of acquisition-as-participation that was the focus of the objections. Long (1998), for example, argued that changes in the social setting have not been shown to have any effect on the way in which learners acquire an L2 (for example, no differences in error types or developmental sequences are evident). In a similar fashion, Gass (1998) argued the approach that Firth and Wagner advocated ‘is not actually part of SLA, but part of the broader field of L2 studies’ (p. 84). She emphasized the need to distinguish ‘participation’ and ‘acquisition’. In effect, Long and Gass simply reasserted the legitimacy of the input-output model and SLA as a branch of cognitive science.

  Summarizing the differences between cognitive and social SLA

  This debate is on-going. Table 9.2 provides a summary of the principal differences between cognitive and social SLA. As Larsen-Freeman (2007) noted these paradigms are both internally consistent and ‘exist in parallel worlds’ (p. 781).

  Dimensions Cognitive SLA Social SLA

  Scope Focus on the universal aspects of L2 acquisition. Priority is given to linguistic competence. Focus on the discursive characteristics of interactions involving L2 learn
ers and how learning of both the micro and macro aspects of language is embedded in these interactions.

  Social context The structural view is dominant—the social context determines the L2 data made available to the learner and the learner’s attitudes to learning. The interactional view is dominant—the social context is seen as jointly constructed by the participants through interaction.

  Learner identity The learner is viewed as a ‘non-native speaker’. Learner identity is static. The learner is viewed as having multiple identities that afford different opportunities for language learning. Learner identity is dynamic.

  Learner’s linguistic background The learner has full linguistic competence in his/her L1. Learners may be multilingual and may display varying degrees of proficiency in their various languages.

  Input Input is viewed as linguistic ‘data’ that triggers acquisition. Input is viewed as contextually constructed; it is both linguistic and non-linguistic.

  Interaction Interaction is viewed as a source of input and an opportunity for output. Interaction is viewed as a socially negotiated event and a means by which learners are socialized into the L2 culture.

  Research methodology Methodology is quantitative and confirmatory—inquiry is seen as ‘scientific’ and value-free, aimed at testing specific hypotheses with a view to making generalizations. Methodology is qualitative and interpretative—emphasis is placed on uncovering the ‘local agenda’ through detailed analysis of naturally-occurring interactions. Learning is demonstrated by tracking specific learning objects over time to demonstrate that change has occurred.