by Rod Ellis
If learning is ‘participation’, then doing learning is learning. This is the position taken by sociocultural theorists and also by some conversation analysts. Such a position would be disputed by cognitive theorists who require evidence that ‘doing learning’ results in some change in the learner’s linguistic system. Some conversation analysts also acknowledge that ‘learning is understood as changes in doing, rather than as doing in itself’ (Sahlström 2011: 45). We do not know for sure that J will remember the names of the English mealtimes following T’s input. As Kasper (2006) noted ‘CA does not provide a ready-made framework to examine learning, whether language learning or the learning of other things’ (p. 91). How then does CA-SLA endeavour to show that learning has taken place?
Short-term development
Given CA-SLA’s commitment to relying exclusively on the evidence provided in the naturally-occurring interactions that learners participate in—i.e. eliciting additional data by means of tasks and tests is not allowed—it is necessary to look for evidence within the same or in successive interactions involving the same learner to demonstrate that change has occurred. ‘Short-term development’ is demonstrated if change occurs within a single interaction or within subsequent interactions that occur shortly afterwards (for example, within the same lesson).
Long-term development
To investigate how learning takes place over time, Markee (2008) advocated a ‘learning tracking methodology’. This consisted of (1) learning object tracking and (2) learning process tracking. The former involved identifying when a specific learning object (for example, a word or a grammatical structure) occurred in the interactions involving the same learner(s) over a period of time. The latter involved showing ‘how and when participants orient to, and potentially incorporate, particular learning objects that occur in different speech events in their interactional repertoires’ (p. 409).
An example of this kind of approach can be found in Firth and Wagner (2007). They focused on a lexical item—‘blowing’—as this occurred in telephone conversations between a Danish cheese exporter (H) and an Egyptian importer (A). These extracts are reproduced below. On the first occasion, A refers to the problem with the cheese he has received from (A), noting that it is ‘blowing’ (a non-standard way of referring to the fact that the cheese has started to ferment). A self-initiated/other-repair sequence ensues where H indicates he does not understand ‘blowing’ and A responds by providing synonyms (i.e. ‘bad’ and ‘fermenting’). Finally—in l.6—H demonstrates he has understood by providing his own synonym (‘gone off’). In this extract, however, H only produces ‘blowing’ when repeating the word following A’s turn. In the second extract (Extract 3 below)—which occurred in a telephone conversation between the same two people two weeks later—H is able to produce ‘blowing’ without any assistance from A. Firth and Wagner argued that this second encounter shows how learning ‘is carried over in time and space’ (p. 808) and that ‘H has now learned this lexical item in a way that extends beyond the concrete local context where he first became acquainted with it’ (p. 809).
Extract (2)
1 A We don’t want the order after the cheese is uhh blowing.
2 H See, yes.
3 A So I don’t know what I can we can do uh with the order now. (.)
What do you think we should do with this is all blowing Mister Hansen (0.7)
4 H I am not uh (0.7) blowing uh what uh, what is this uh too big or what?
5 A No the cheese is bad Mister Hansen (0.4) it is like (.) fermenting in the Customs cool rooms.
6 H Ah it’s gone off.
7 A Yes, it’s gone off.
Extract (3)
1 A Yes (.) Mister Hansen.
2 H Hello Mister Akkad (.) hh we haf some informations for you about the cheese (.) with the blowing.
3 A Yes Mister Hansen (Firth and Wagner 2007: 808)
Firth and Wagner aim to show how ‘learning in action’ takes place. In CA-SLA this involves ‘being contingently adaptable as the unfolding context requires’ (2007: 808). The analyses they provide also illustrate another important claim of the conversation analytical approach, namely that change occurs as learners discover how to become members of a particular community of practice—a cheese exporter/importer in the case of Firth and Wagner’s study. It is in this sense that language learning is a social practice.
Conversation analysts are also concerned with how learners acquire the discursive practices associated with particular social events. Young and Miller (2004), for example, investigated an adult Vietnamese learner’s participation in the revision talk that occurred in weekly writing conferences. They documented how this changed over time from peripheral to fuller participation. They described a participation framework for this particular social event consisting of eight speech acts. Initially, the instructor performed seven of these. Four weeks later, however, the student performed many of the acts himself (for example, he now identified problems, explained the need for revision, and suggested possible ways of carrying out revision) without any direction from the instructor. In other words, the student demonstrated that he had now acquired ‘interactional competence in the practice of revision talk’ (p. 533).
Final comments
The conversation-analytic approach views language learning as social activity. Perhaps, more than any of the other social approaches examined in this chapter, it is representative of the ‘social turn’ in SLA. It avoids consideration of exogeneous factors as explanations of learningNOTE 4. It is broad in scope – it addresses both how linguistic items are acquired and how interactional competence is developed – and it views the former as embedded in the latter. It asks questions such as ‘How do learners learn to do disagreement, repair, etc. as members of specific communities?’ and ‘In so doing how do they acquire L2 linguistic resources in the process?’ It views identity as multiple and negotiated rather than fixed and situationally determined. Thus identity is not something that is ‘accessed’ by learners when they interact but rather is ‘interactionally produced, locally occasioned, and relationally constituted’ (Kasper and Wagner 2011: 122). At times it is the learner’s identity as learner that is made prominent but at other times other social identities emerge.
The greatest challenge facing conversation analytical approaches is being able to provide evidence of long-term learning. Naturally-occurring data do not readily afford instances of reoccurring ‘learning objects’ and it is difficult to know what constitute comparable interactional events. The rejection of tests to establish long-term learning can perhaps be justified on the grounds that tests do not show what learners can do in naturally occurring interactions but this creates problemsNOTE 5. CA-SLA is also limited because it declines to consider how psychological factors (for example, working memory or motivation) affect how specific interactions unfold.
The social identity approach
Sociocultural theory, sociocognitive theory, and conversation analysis all acknowledge the importance of social identity. I turn now to a theory that foregrounds the role of social identity. The social identity approach constitutes an ‘unequivocally social approach to SLA’ (Norton and McKinney 2011: 87).
Social identity as a postmodern phenomenon
From a postmodern viewpoint, social identity is understood as a relational rather than an individual phenomenon: our social identities do not constitute fixed traits, but rather exist in a state of flux as they are discursively constructed through participation in interactions with different people in different contexts. Nevertheless, we do tend to think of our own and other people’s identities as ‘inscriptions’. We view our identities in terms of gender (male or female); race (African or European); marital status (married or single); occupation (university professor or chef), etc. Language learners may similarly see themselves as native or non-native speakers—despite Firth and Wagner’s rejection of such inscriptions—although they do not necessarily always act out these identities. These two ways of conceptu
alizing social identity, however, are not necessarily contradictory. As Miller and Kubota (2013) pointed out our identities are constructed and reconstructed repeatedly so that ‘our ongoing social performances become sedimented into recognizable durable identities that persist over time’ (p. 233). By and large, however,—as Miller and Kubota acknowledged—SLA research that has investigated the role of social identity has drawn on these sedimented identities.
Postmodern accounts of social identity emphasize the following:
Social identity is multiple and non-unitary. The language learner is not just a ‘language learner’, but can draw on a number of different social identities when interacting in the L2. This is most obviously the case for learners functioning in multilingual naturalistic settings, but is also true for classroom learners.
Social identity is complex and inherently subjective. For example, learners’ sense of what it means to be ‘male’ or ‘female’ intersects with their other identities (race, ethnicity, social class, etc.) and thus varies from one learner to another. Thus, it is not possible to generalize how gender affects learning opportunities across all male or all female learners. McKay and Wong (1996), for example, found little commonality in how four male Chinese students in a Californian secondary school enacted gender in their social contact with other students.
Because social identity is multiple, conflicts can arise when the behaviour associated with one identity becomes incompatible with that of another. Skilton-Sylvester (2002) reported that Ming—a Cambodian learner of English—was unable to balance her identity as a student with that as a wife when her husband demanded she discontinue her language classes because he perceived her expanding social networks as a threat to their marriage. She felt obliged to give up her literacy studies.
Power relationships are inextricably linked to the social identities that are evoked when learners engage in interactions with other speakers of the language and thus impact on the learning opportunities available to learners. Inequitable power relationships can impede learners’ rights to speak and to be heard and thus restrict their access to the L2. Learners, however, have ‘agency’. That is, they can resist how they have been positioned by drawing on an identity option that affords them more equal status with their interlocutors and, in this way, they can achieve ways of participating in the kinds of social exchanges that foster opportunities for learning. Social identities, then, are a site of struggle.
The social identities of learners need to be understood not just in terms of those identities that they have already constructed but also in terms of imagined communities and identities they envisage for their future lives. Learners conceive of groups of people who are not yet accessible, but who may become so in the future. ‘Imagined communities are no less real that the ones in which learners have daily engagement and might have a stronger impact on their identities and investments’ (Norton and McKinney 2011: 76).
Language learning requires ‘investment’ on the part of the learner. This notion—first introduced by Norton Peirce (1995)—differs from that of instrumental motivation (see Chapter 3). It acknowledges that when learners speak, they are not just concerned with the material benefits of learning a language, but also with organizing and reorganizing who they are. Learners’ decisions whether or not to invest in language learning depends on how they see themselves in relation to their existing or imagined social communitiesNOTE 6.
Social identities are dynamic: they evolve over time and new identities are constantly being formed. This is clearly so in the case of the ‘language learner’. In the opinion of some researchers this results in a ‘third space’—a hybrid identity that evolves when learners come to see themselves as multilingual and transnational individuals, rather than as language learners. Kramsch (2009) proposed that the third space allows learners to take over the language and ‘give it meanings other than the native speaker would’ (p. 238).
Norton’s theory of social identity in L2 learning
The importance of social identity in L2 learning came to prominence in a series of publications by Norton in the 1990s (for example, Norton Peirce 1995; Norton 1997). Norton sought to ‘rethink language acquisition in its social, cultural, and political contexts, taking into account gender, race, and other relations of power as well as the notion of the subject as multiple and formed within different discourses’ (Norton Peirce 1995: 20).
Norton (1997) defined social identity as ‘the relationship between the individual and the larger social world, as mediated through institutions such as families, schools, workplaces, social services, and law courts (p. 420). She argued that it was related to but distinct from both ‘cultural identity’ (‘the relationship between individuals and members of a group who share a common history, a common language, and similar ways of understanding the world’) and ‘ethnic identity’ (‘the relationship between the individual and members of the race to which the learner belongs’). Norton elected to investigate social identity because she believed that it best captured the heterogeneous and dynamic nature of identity in the learners she investigated.
Implicit in Norton’s view of social identity, L2 use, and L2 learning is the importance of ‘ownership’ of the language being learned. Norton rejected the view that native speakers own the language and asserted that learners need to see themselves as legitimate speakers of it and challenge being positioned as illegitimate. Thus—like Firth and Wagner—she rejected the traditional notions of ‘native speaker’ and ‘non-native speaker’.
Norton (2000) reported on a number of case studies of adult immigrant women in Canada. She documented how these women often had social identities imposed on them which denied them opportunities to speak and be heard and which they were uncomfortable with. In some cases, this resulted in them withdrawing from contact with native speakers. In other cases, however, they fought to establish a preferred social identity in their interactions with native speakers which afforded more equal discourse rights and—in Norton’s view—greater opportunities for learning English. Mai is an example of the first. She became the subject of derision in her workplace because of the preferential treatment she received from the management on account of her single status. As a result, she was not able to establish an identity that created the conditions where she could comfortably interact with other speakers. In contrast, Martina—a Czech woman who came to Canada with her husband and children at the age of 37—was able to make use of her identity as the primary caregiver in her family to engage in ‘counter-discourses’. She refused to be silenced despite experiencing frequent feelings of shame. Instead, she challenged the ‘rules’ governing interactions between immigrants and Anglophones—for example, by entering into an argument with her landlord about the terms of her lease. However, Norton provided no evidence that the learning opportunities afforded to Martina led to learning.
Transnational identity
Where Norton was concerned with under-privileged immigrant learners, other researchers studied L2 learners who were not so socially disadvantaged. Block (2006), for example, investigated the multilingual identities of five Japanese female graduate students in London. These learners were all middle-class, highly educated, and economically secure and thus possessed ‘the requisite social and cultural capital to make the move from Japan to London’ (p. 104). While there were some marked differences in these learners’ identities, they were ‘internationalist women’—that is, women who had elected to live their lives, to some degree at least, outside the confines of traditional Japanese society. They viewed English as ‘a language of liberation in which they can develop new femininities’ (p. 97).
An increasing number of studies have focused on the transnational identity of multilingual learners. Rampton (1995)—for example—documented the insider identities that arise in the multiracial communities found in urban settings. He described the ‘language crossing’ (i.e. ‘the use of speech varieties which are not normally thought to belong to the speaker’) that occurred in White and As
ian adolescents’ use of West Indian Creole. Rampton suggested that these speakers’ access to minority languages enabled them to play with different identities when interacting with members of their peer group. Li and Hua (2013) investigated the ‘translanguaging’ of a group of Chinese university students in London. Originating from different countries and social backgrounds, they developed a transnational and multilingual network amongst themselves. Li and Hua illustrated how these students developed fluid identities—not as Chinese from specific countries, nor as Chinese in Britain—but as Chinese students at universities in London.
Final comments
A limitation of these—and other—studies is that, although they address how ‘learning opportunities’ are created through the social identities learners assume, they do not demonstrate to what extent these opportunities actually result in learning. There is a tendency to uncritically equate ‘learning opportunities’ with ‘learning’. In part, this limitation is a result of the methodology researchers have adopted. By and large, the research has consisted of short-term case studies. In such studies it is not possible to document how social identity and learning interrelate over time. Also, as Miller and Kubota (2013) pointed out, there has been a reliance on learner narratives collected through interviews. Researchers have tended to take what learners say about their learning experiences at face value without questioning the veracity of self-report data.