Identity, language learning, and social change

Abstract

In this review article on identity, language learning, and social change, we argue that contemporary poststructuralist theories of language, identity, and power offer new perspectives on language learning and teaching, and have been of considerable interest in our field. We first review poststructuralist theories of language, subjectivity, and positioning and explain sociocultural theories of language learning. We then discuss constructs of investment and imagined communities/imagined identities (Norton Peirce 1995; Norton 1997, 2000, 2001), showing how these have been used by diverse identity researchers. Illustrative examples of studies that investigate how identity categories like race, gender, and sexuality interact with language learning are discussed. Common qualitative research methods used in studies of identity and language learning are presented, and we review the research on identity and language teaching in different regions of the world. We examine how digital technologies may be affecting language learners' identities, and how learner resistance impacts language learning. Recent critiques of research on identity and language learning are explored, and we consider directions for research in an era of increasing globalization. We anticipate that the identities and investments of language learners, as well as their teachers, will continue to generate exciting and innovative research in the future.

1. Introduction

In restaurant was working a lot of children but the children always thought that I am – I don't know – maybe some broom or something. They always said ‘Go and clean the living room’, and I was washing the dishes and they didn't do nothing. They talked to each other and they thought that I had to do everything. And I said ‘No’. The girl is only 12 years old. She is younger than my son. I said ‘No, you are doing nothing. You can go and clean the tables or something’.

(Interview with Martina, Norton 2000: 99)

Martina was an English language learner from eastern Europe who had immigrated to Canada for a better life for her three children. Partly because she was not a proficient speaker of English, she struggled to find work in her profession as a quantity surveyor, and was employed in a fast food restaurant in the greater Toronto area. Her co-workers, as well as the manager's children (who frequently visited the restaurant), were all born in Canada, and spoke English fluently. What Martina communicates in this extract is that engaging in social interaction with her co-workers was a struggle, primarily because she was positioned as a dehumanized and inanimate ‘broom’. To resist these marginalizing practices, Martina reframed her relationship with her co-workers as domestic rather than professional, and from the identity position ‘mother’, rather than ‘immigrant’ or ‘broom’, she claimed the right to speak.

While this data has been discussed more fully in other publications (Norton Peirce 1995; Norton 2000) the vignette is a sobering reminder of the powerful relationship between identity and language learning, which is of central concern to many scholars in the field of language education. Indeed, over the past 15 years, there has been an explosion of interest in identity and language learning, and ‘identity’ now features in most encyclopedias and handbooks of language learning and teaching (Norton & Toohey 2002; Ricento 2005; McKinney & Norton 2008; Norton 2010; Morgan & Clarke 2011). In the broader field of applied linguistics, interest in identity has also gained considerable momentum. There is work, for example, on identity and pragmatics (Lo & Reyes 2004; Spencer-Oatey & Franklin 2009), identity and sociolinguistics (Joseph 2004; Omoniyi & White 2007; Edwards 2009); and identity and discourse (Benwell & Stokoe 2006; Wodak et al. 2009; Young 2009). An extended state-of-the-art article on identity and language learning is timely. We have defined particular areas of interest from a rapidly expanding literature in order to provide readers with both depth and breadth in our review of this topic.

Given that the article represents a contemporary review, we focus on the literature published since Norton's often cited article ‘Social identity, investment, and language learning’, published in TESOL Quarterly in 1995 (Norton Peirce 1995). This article, along with succeeding early publications (Norton 1997, 2000, 2001), drew on poststructuralist theories of language and identity to offer new perspectives on language learning and teaching, and introduced Norton's construct of investment to the field. In reviews of the literature, many scholars cite Norton's work as pivotal in reframing debates on identity (Menard-Warwick 2005; Ricento 2005; Block 2007a, 2007b; Swain & Deters 2007; De Costa 2010a; Morgan & Clarke 2011) and it was in the context of such work, as Zuengler & Miller note (2006: 43), that identity was established as a research area ‘in its own right’. Further, as Block (2007a: 864) notes, a poststructuralist approach to identity ‘has become the approach of choice among those who seek to explore links between identity and L2 learning’. We therefore focus on poststructuralist theories of identity and language learning in framing this review of the literature.

Further, while much research on identity focuses on second language (L2) learning, poststructuralist theory is also of great relevance to foreign language learning (see, for example, Kanno 2003, 2008; Pavlenko 2003; Kinginger 2004; Kramsch 2009) and will be incorporated in this article. Readers interested in work on group identity and intergroup relations, common in the 1970s and 1980s, would find McNamara's review (1997) particularly helpful. In his article, McNamara highlights the important work of Tajfel (1981), who was centrally concerned with identity relationships between in-groups and out-groups in social life.

Central arguments in our review are discussed in Section 2, with explications of relevant theoretical frameworks examined in Section 3. The relationship between identity and second language acquisition (SLA), discussed in detail in Section 4, has been of much interest to the field, particularly in view of Firth & Wagner's (1997: 285) call for an ‘enhanced awareness of the contextual and interactional dimensions of language use’. Further, while much of this research explores the multiple and intersecting dimensions of language learners' identities, there is also a growing body of research (see Section 5) that seeks to investigate the ways in which particular relations of race, gender, and sexual orientation may impact the process of language learning. Also of interest are the research methods associated with research on identity (Section 6), as well as implications for teaching (Section 7). We then move to critiques of the field (Section 8), and conclude, in Section 9, with directions for the future.

2. Central arguments

In 1998, sociolinguist Susan Gass made the important argument that the theoretical relevance of identity categories to L2 learning needed to be established. The wide range of research studies we discuss here shows that new theories of identity and language learning permit a conceptual shift in research about L2 learning, and offer important insights about the language learning process. The points below summarize the claims made by identity and language learning researchers, with illustrative reference to Martina's vignette above, and are further developed in subsequent sections:

1. (i) Contemporary identity theories offer ways to see the individual language learner situated in a larger social world. While some previous SLA research defined learners in binary terms (such as motivated or unmotivated, introverted or extroverted, inhibited or uninhibited), identity theorists see these affective descriptors as constructed in frequently inequitable social contexts, as variable over time and space, and sometimes co-existing in contradictory ways within a single individual. As illustrated in the data from Martina, identity is theorized as multiple, changing, and a site of struggle.

2. (ii) Identity theorists highlight the diverse positions from which language learners are able to participate in social life, and demonstrate how learners can, but sometimes cannot, appropriate more desirable identities with respect to the target language community. As Martina found, while some identity positions may limit and constrain opportunities for learners to listen, speak, read, or write, other identity positions may offer enhanced sets of possibilities for social interaction and human agency.

3. (iii) Language learning theory and research needs to address how power in the social world affects learners' access to the target language community, and thus to opportunities to practice listening, speaking, reading, and writing, widely acknowledged as central to the SLA process. Identity theorists are therefore concerned about the ways in which power is distributed in both formal and informal sites of language learning, such as Martina's workplace, and how it affects learners' opportunities to negotiate relationships with target language speakers.

4. (iv) Identity, practices, and resources are inextricably linked and mutually constituted. The variable practices and resources of specific settings, and an individual's access to them (as Martina found), relate powerfully to the ways in which identities of individuals are constructed. From a poststructuralist perspective, practices, resources, and identities are both produced and inherited. Examination of these in relation to language learning offers promise for improving and enhancing learning contexts.

5. (v) L2 learning is not entirely determined by structural conditions and social contexts, partly because these conditions and contexts are themselves in states of production. In addition, language learners who struggle to speak from one identity position, as Martina did, may be able to reframe their relationship with their interlocutors, thereby changing their access to practices and resources, and claim alternative identities from which to speak, listen, read, or write. If learners are successful in their bids for more powerful identities, their language acquisition may be enhanced.

6. (vi) The sociological construct of investment complements the psychological construct of motivation in SLA. Norton, who first introduced this construct (Norton Peirce 1995; Norton 2000; Norton in press) was concerned that most psychological theories of language learning motivation did not do justice to the complex identities of language learners, and the often inequitable relations of power they negotiated in different sites. The construct of investment seeks to make a meaningful connection between a learner's desire and commitment to learn a language, and the language practices of the classroom or community. Although Martina was a highly motivated language learner, she was not invested in the language practices of her workplace, where she experienced discriminatory practices. Such theorizing has helped to shift contemporary debates on motivation in the field of SLA (see Dörnyei & Ushioda 2009).

7. (vii) The theoretical constructs imagined communities and imagined identities contribute usefully to understanding SLA, because a learner's hopes for the future (or their children's future) are integral to language learner identity. For many learners, the target language community is not only a reconstruction of past communities and historically constituted relationships, but also a community of the imagination, a desired community that offers possibilities for an enhanced range of identity options in the future. An imagined community assumes an imagined identity, and a learner's investment in the target language can be understood within this context.

3. Theoretical frameworks

Poststructural theories of not only language, but also subjectivity and positioning, inform recent work on identity and language learning. Sociocultural theory also offers perspectives on learning that are often drawn upon in recent work on this topic. We address each of these areas in turn, illustrating the ways in which they have been taken up in research on language learning and teaching.

3.1 Poststructuralist theories of language

Poststructuralist theories of language have become increasingly attractive to identity and language learning researchers (Norton & Morgan in press). Structuralist theories of language, often cited as originating with the work of Swiss linguist Ferdinand de Saussure (1966), emphasized the study of the linguistic knowledge (competence) that allowed idealized speakers/hearers to use and understand language's stable patterns and structures. From this perspective, actual instances of language usage (performance), which could be affected by memory lapses, fatigue, slips, errors, and so on, were not seen as revealing of idealized patterns, and thus were of little interest in the scientific study of language. However, poststructuralist theories of language, proposed by many, but particularly by Russian literary theorist Mikhail Bakhtin (1981, 1984, 1986) saw language not as a set of idealized forms independent of their speakers or their speaking, but rather as situated utterances in which speakers, in dialogue with others, struggle to create meanings.

For Bakhtin, language had no independent existence outside of its use, and that usage was of course social. He used the metaphor of speech communication as a chain, an ongoing conversation that new speakers (for example, children or newcomers to speech communities) strive to join. While structural theories might see language learning as a gradual individual process of internalizing the set of rules, structures, and vocabulary of a standard language, Bakhtin saw language learning as a process of struggling to use language in order to participate in specific speech communities. Using language meant using a tool others had used before, and Bakhtin saw speakers as constrained by those past usages. However, he also saw speakers as able to use language to express their own meanings (with both custom and innovation characterizing language use). Further, Bakhtin pointed out how social positions outside language might affect any individual's speaking privilege.

More recently, Hall, Cheng & Carlson (2006) discussed a usage-based view of language knowledge that saw people learning through joint engagement with others in activities, using cultural tools. Recognizing the increasing body of theory and research that points out how using even one language relies on complex sets of understandings of context, they argued that speakers of multiple languages are able to engage in interactions in those languages as a result of their access to participation in the activities in which those languages are used. Unlike those who believe that language competence precedes language performance, Hall et al. understood language competence as proceeding from participation in performance in activities using particular language tools. They also recognized that individuals differ in their access to participation, according to their social and cultural positioning.

French sociologist Pierre Bourdieu's work directly addresses the poststructuralist study of the politics of language (Bourdieu 1977; Bourdieu & Passeron 1977; Bourdieu 1991; Albright & Luke 2008). While poststructuralists are, of course, not the only theorists interested in language and power, Bourdieu explicitly drew attention to the importance of power in structuring discourse, with interlocutors seldom sharing equal speaking ‘rights’. For Bourdieu, ‘legitimate’ and ‘illegitimate’ speakers were distinguished by their differential ‘rights to speech’ or their ‘power to impose reception’ (1977: 648). For Bourdieu, using language was a social and political practice in which an utterance's value and meaning was determined in part by the value and meaning ascribed to the person who speaks. Recognizing that the ascribed value of a person or group can vary, depending on circumstances or contexts (in Bourdieu's terms, ‘fields’), he saw linguistic discourse as ‘a symbolic asset which can receive different values depending on the market on which it is offered’ (1977: 651). He further noted that dominant usage is associated with the dominant class. Heller (2008: 50) explicitly paralleled access to language with access to other resources that are also ‘produced, attributed value, and circulated in a regulated way, which allows for competition over access, and typically, unequal distribution’. From this perspective, not only individuals’ but also groups’ ascribed identities structure access to and opportunities for language use and learning.

3.2 Poststructuralist theories of subjectivity

Christine Weedon (1987/1997), one of the best-known scholars working in the feminist poststructuralist tradition, understood, like Bakhtin and Bourdieu, the importance of ascribed individual and group identity positions in structuring the extent to which language practices are valued. However, she also argued that it is in language that the individual constructs her ‘subjectivity’, which she saw as ‘the conscious and unconscious thoughts and emotions of the individual, her sense of herself, and her ways of understanding her relation to the world’ (1997: 28). Her use of the term ‘subjectivity’ reminds us that an individual can be simultaneously the subject of a set of relationships (e.g. in a position of power) or subject to a set of relationships (e.g. in a position of reduced power). Thus, for Weedon, social relationships are crucial in how individuals are constructed and construct themselves.

Weedon used the terms subject and subjectivity to signal a break with dominant Western humanist views of the individual. While Western humanist philosophy stressed the essential, unique, fixed, and coherent core of an individual, Weedon's view, like that of other poststructuralists, was that the individual (i.e. the subject) was diverse, contradictory, dynamic, and changing over historical time and social space. Like Foucault (1980), Weedon argued that subjectivity is discursively constructed, and is always socially and historically embedded. Holland & Lave (2001) discussed the apparent paradox of identity being experienced as unitary and durable, while being, at the same time, variable and situated in dynamic practice. Like many other poststructuralist theorists, they emphasized that ‘both the continuity and the transformation of social life are ongoing, uncertain projects’ (2001: 4) and that individuals maintain ‘histories in their persons’. These theories of identity are central in Norton's early work, and have been taken up by many identity theorists, including Kramsch (2009), whose compelling book, The multilingual subject, focuses on the subjectivity of the foreign language learner.

Language educators have found poststructural observations about subjectivity helpful in theorizing how education can lead to individual and social change. A conceptualization of subjectivity as multiple, non-unitary, and dynamic leaves room for the view that individuals need not be locked forever in particular positions. Rather, from this perspective, although some contexts and practices may limit or constrain opportunities for learners to listen, speak, read, or write, other contexts and practices may offer enhanced sets of possibilities for social interaction and human agency. Thus, pedagogical practices have the potential to be transformative in offering language learners more powerful positions than those they may occupy either inside or outside the classroom. In poststructuralist theory, subjectivity and language are seen as mutually constitutive, and are thus centrally important in how a language learner negotiates a sense of self within and across a range of sites at different points in time. It is through language that a learner gains access to, or is denied access to, powerful social networks that give learners the opportunity to speak.

Post-colonial theorists such as Stuart Hall (1992, 1997) and Homi Bhabha (1994) used poststructuralist identity theory to analyze how categories such as race and gender have been essentialized. In theorizing cultural identity, Hall focused on identity as in process, ‘becoming’, and stresses that identity is ‘not an essence, but a positioning’ (1997: 226) in particular historical and cultural environments. This means of theorizing difference has not been entirely satisfactory to those who would assert their identities as homogenous and unitary, foregrounding a particular aspect of their experience such as gender, race, or religious affiliation (see Section 5 below). Current worldwide expressions of nationalism and religious fundamentalism testify to this. Such unitary assertions of identity are often explained as strategic essentialism in service of political goals (Yon 1999). The terms identity politics or the politics of difference reference this particular coalescence of identity and power relations.

3.3 Poststructuralist theories of positioning

Bakhtin (1981) was particularly interested in how position or status was signaled in language in works of fiction, and in conversation in general. Many identity and language learning researchers and theorists also stress the importance of considering how contexts shape positioning among particular interlocutors. While positioning has been discussed by many poststructuralist theorists (Foucault 1980; Henriques et al. 1984; Weedon 1987/1997; Hall 1997), it was Davies & Harré (1990: 7) who explicitly used position as ‘the central organising concept for analysing how it is that people do being a person’. They and other poststructuralist theorists have reminded us that identities are contingent, shifting and context-dependent, and that while identities or positions are often given by social structures or ascribed by others, they can also be negotiated by agents who wish to position themselves. As Davies & Harré put it: ‘discursive practices constitute the speakers and hearers in certain ways and yet at the same time are a resource through which speakers and hearers can negotiate new positions’ (1990: 7).

Recognition of the apparent paradox of positioning, reflecting the socially-given and the individually-struggled for, has been important in many studies of language learning. Menard-Warwick (2007), for example, identified particular positioning speech acts of both a vocational English language class teacher and her Latina students, such that learners were enabled or constrained to claim voice in the classroom. Noting that while vocational teachers often aim at empowering their students, Menard-Warwick observed that customary classroom materials and activities, as well as powerful societal discourses, often constrain students’ possibilities for claiming desirable identities. She pointed out that teachers should be alert to how students position themselves in classroom discourse and approach language instruction from a critical perspective to enable learners to name, and perhaps struggle against, some of the disempowering tendencies of the linguistic practices of their new cultures.

3.4 Sociocultural theories of language learning

Sociocultural theories draw on L. S. Vygotsky's (1978, 1987) insights into the social nature of learning but also on the work of more contemporary theorists who have extended and modified his ideas (e.g. Wertsch 1998; Rogoff 2003). Vygotsky (1978) argued that humans act on the world with tools (both physical and symbolic), and emphasized the symbolic tool of language, proposing that children gain ‘increasing control over the mediational means made available by their culture, including language for interpersonal (social interaction) and intrapersonal (thinking) purposes’ (Lantolf 2000: 8). From this perspective, learning is a social process in which culturally and historically situated participants engage in culturally-valued activities, using cultural tools. They thus develop the sorts of behaviors required for participation, and in so doing, change the activities and the tools. This foregrounding of dynamic social activity and the tool mediators of that activity are special features of sociocultural theory. Educational studies grounded in this perspective pay careful attention to the activities provided for learners in their diverse environments and to the qualities of the physical and symbolic tools, including written language, that learners use. In addition, they stress the importance of learners' access to cultural resources, and how learners might change those resources over time. As Rogoff (2003) noted, development (or learning) is ‘changing participation in the sociocultural activities of a community, which also change’ (2003: 368).

Sociocultural theories have been frequently drawn upon in studies of SLA. Such studies are sometimes more psychological in orientation (examples are Lantolf 2000; Lantolf & Thorne 2006; Swain et al. 2010); sometimes they are more anthropologically- and sociologically-focused (Kostogriz & Tsolidis 2008; Toohey 1998, 2000; Toohey & Norton 2010). The theories represent a shift from seeing learners as individually internalizing stable systems of language knowledge, to seeing them as differentially-positioned members of social and historical collectivities, using (and thus learning) language as a dynamic tool. This moves observers toward examining the conditions for learning, and the issues of access of learners for appropriation of practices, in any particular community.

Lave & Wenger's (1991) work has been used by many language researchers allied with sociocultural theory, especially with respect to their construct, legitimate peripheral participation, which represented their view that communities are composed of participants who differentially engage with the practices of their communities, and that this engagement or participation in practice is ‘learning’. For them, the ‘oldtimer’ as well as the ‘newcomer’ are simultaneously learning through practice. Stressing the importance of local analysis of communities, they pointed out that conditions vary with regard to ease of access to expertise, to opportunities for practice, to consequences for error in practice, and so on. Lave & Wenger discussed the importance of not sequestering newcomers away from participation in community activities, if they are to learn. They noted that, ideally, learners must ‘see’, or be in the presence of, mature practice. This theme is taken up by SLA researchers who examine, in particular, learners' access to L2 communities.

4. Identity and SLA

A great deal of language learning research in the 1970s and 1980s conceptualized the ‘identities’ of language learners as their fixed personalities, learning styles, and motivations. In contrast, more recent work on language learner identities adopts poststructural understandings of identities as fluid, context-dependent, and context-producing, in particular historical and cultural circumstances. From this perspective, personalities, learning styles, motivations, and so on are not fixed, unitary, or decontextualized, and while context ‘pushes back’ on individuals’ claims to identity, individuals also struggle to assume identities that they wish to claim. Constructs of investment and imagined communities/imagined identities have been particularly important in these debates.

4.1 Investment, motivation, and SLA

When Norton conducted research with immigrant women in Canada (Norton 2000), she observed inconsistencies in the predictions made by studies of motivation in SLA, on the one hand, and what she found from careful ethnographic observation of language learners, on the other. Most motivation studies at that time framed motivation as a fixed characteristic of individual language learners, and hypothesized that learners who failed to learn the target language did not, for various reasons, have sufficient (or appropriate) desire to learn the language. Seeing language learning as mainly an individual accomplishment, these studies of SLA motivation were not, by and large, concerned with power relations between language learners and target language speakers. What Norton found was that high levels of motivation did not necessarily result in ‘good’ language learning, and that unequal relations of power between language learners and target language speakers were often salient in her learners' accounts. Accordingly, she found it necessary to develop the construct of investment to complement constructs of motivation in the field of SLA (Norton Peirce 1995; Norton 2000, 2010, in press).

The construct of investment in language learning draws on economic metaphors, associated particularly with the work of Bourdieu. Bourdieu & Passeron (1977) used the term cultural capital to refer to the knowledge, credentials, and modes of thought that characterize different classes and groups. They argued that cultural capital has differential exchange value (or ‘currency’) in different social fields. Drawing on these metaphors, Norton observed that learners ‘invest’ in the target language at particular times and in particular settings, because they believe they will acquire a wider range of symbolic and material resources, which will, in turn, increase the value of their cultural capital. As the value of learners' cultural capital increases, so learners reassess their sense of themselves and their desires for the future. Norton argued that investment and identity together signal the socially and historically constructed relationship of learners to the target language and their sometimes ambivalent desire to learn and practice it.

The notion of investment recognizes that learners often have variable desires to engage in the range of social interactions and community practices in which they are situated. Previous work on motivation frequently conceived of individuals as having unitary, fixed, internalized and ahistorical ‘personalities’. Investment, on the other hand, sees language learners as having complex identities, which change across time and space, and which are constructed on the basis of the socially given, and the individually struggled-for. Thus, while motivation can be seen as a primarily psychological construct (Dörnyei 2001), investment is a sociological construct, and seeks to make meaningful connections between a learner's desire and commitment to learn a language and their changing identities.

Seeing engagement in language learning as investment calls forth different sets of questions about a learner's commitment to learning the target language. In addition to asking ‘To what extent is the learner motivated to learn the target language?’ the teacher or researcher can also ask ‘What is the learner's investment in the language practices of this classroom?’ A language learner may be highly motivated, but may nevertheless have little investment in the language practices of a given classroom or community, which may, for example, be racist, sexist, elitist, anti-immigrant, or homophobic. Alternatively, the language learner's conception of good language teaching may not be consistent with that of the teacher, compromising the learner's investment in the language practices of the classroom. Thus, the language learner, despite being highly motivated, may not be invested in the language practices of a given classroom. The learner could then be excluded from those practices, or choose not to participate in classroom activities. In time, the learner could be positioned as a ‘poor’ or unmotivated language learner by others (Norton & Toohey 2001).

A classroom-based study conducted by Duff (2002) in a multilingual secondary school in Canada serves to illustrate these theoretical issues. Duff found that the teacher's attempts to provide speaking rights for non-local students in the classroom had ambiguous results. It became apparent that many of the English language learners in the class were afraid of being criticized or laughed at by native English-speaking peers because of their limited command of English, and thus avoided oral interaction. Duff (2002: 312) noted that ‘Silence protected them from humiliation’. Learners' silence, however, was represented by local English speakers as ‘a lack of initiative, agency, or desire to improve one's English or to offer interesting material for the sake of the class’ (2002: 312). In addition, the teacher's efforts to provide non-local students with opportunities to speak sometimes positioned students in ways Duff called ‘awkward’. While some students resisted their subordinate positioning verbally, others were seemingly content to remain silent, investing more heavily in the written activities of the classroom. It could be argued that rather than being unmotivated, many of the silent English language learners in the class were not invested in the language practices of their classroom, actively resisting practices in which they occupied unequal relations of power vis-à-vis local English speakers.

The fields of applied linguistics and language education have seen lively interest in the construct of investment (Pittaway 2004), including a special issue on investment that appeared in the Journal of Asian Pacific Communication (Arkoudis & Davison 2008; Norton & Gao 2008). McKay & Wong (1996), for example, investigated the English language development of four Mandarin-speaking students in Grade 7 and 8 in a California school, noting that these students’ investment in the school's language was related to their needs, desires, and negotiations. Skilton-Sylvester (2002) argued that the psychological construct of motivation did not sufficiently describe the complex lives of four Cambodian women in adult ESL (English as a Second Language) classes in the USA with whom she worked, and that understanding the domestic and professional identities of these women was necessary to explain their investment in particular adult ESL programs.

With reference to foreign language learning, Kinginger (2004) documents the experiences of a young American woman called Alice, who, over a four-year period, negotiated many facets of her identity in her struggle to learn French, both in the USA and in a study abroad experience in France. Kinginger addresses the identity changes that Alice underwent as she sought to reconcile an ‘imagined’ France with her mixed language learning experiences, concluding that Alice's efforts towards French language competence were clearly an investment in social identity. In a different context, Haneda (2005) examined the engagement of two university students in an advanced Japanese literacy course, concluding that their membership in multiple and differing communities may have shaped the way they invested in writing in Japanese. Potowski (2007) explored students’ use of Spanish in a dual Spanish/English immersion program in the USA, noting that even if a language program is well run, learners' investments in the target language must be consistent with the goals of the program if language learning is to meet expectations. Arguing that investment has emerged as a ‘significant explanatory construct’ (2006: 59) in the L2 learning literature, Cummins (2006) drew on the construct of investment to develop a definition of an identity text, and he and Early provide powerful examples of such texts in their forthcoming book (Cummins & Early 2011).

4.2 Imagined communities and imagined identities

In modern daily life, people interact directly with members of many communities: they may be involved in neighborhood, workplace, educational, medical, and religious communities. As Wenger (1998) suggested, however, these are not the only communities with which people are affiliated; they also affiliate with communities of the imagination. Anderson (1991), who originally coined the term ‘imagined communities’, observed that nations are imagined communities, ‘because the members of even the smallest nation will never know most of their fellow-members, meet them, or even hear of them, yet in the minds of each lives the image of their communion’ (1991: 6). Thus, in imagining ourselves allied with others across time and space, we can feel a sense of community with people we have not yet met and with whom we may never have any direct dealings.

The term ‘imagined community’ was applied to SLA theory by Norton (2001), who was particularly interested in the relationship between imagined communities and imagined identities. These concepts were further developed in Kanno & Norton (2003) and Pavlenko & Norton (2007), and have proved productive in diverse research sites (see, for example, Silberstein 2003; Murphey, Jin & Li-Chi 2005; Carroll, Motha & Price 2008; Dagenais et al. 2008; Kendrick & Jones 2008). There is a focus on the future when learners imagine who they might be, and who their communities might be, when they learn a language (see also Jenkins 2005, 2007 with regard to English as a lingua franca). Such communities include affiliations, such as nationhood or even transnational communities, which extend beyond local sets of relationships. Such imagined communities may well have a reality as strong as those in which learners have current daily engagement, and might even have a stronger impact on their investment in language learning. Norton (2001) argued that a lack of awareness of learners' imagined communities and imagined identities could hinder a teacher's ability to construct learning activities in which learners can invest.

Further discussion of these issues is available in a co-edited special issue of the Journal of Language, Identity, and Education on ‘Imagined Communities and Educational Possibilities’ (Kanno & Norton 2003). In this publication, international scholars explored how the notion of imagined communities figured in language learning in the specific situations with which they were familiar. More recent work by Kanno (2008) in Japan examined the relationship between school programs and access to bilingualism in five schools that promoted bilingual education. While additive bilingualism was promoted for upper-middle-class students, schools serving immigrant and refugee children were more likely to promote subtractive bilingualism. Kanno concluded that educators' various visions of children's imagined communities drew forth different forms of bilingual education, and maintained existing inequities among upper-middle-class and immigrant and refugee children.

Imagined community also served as a helpful theoretical construct in the work of Dagenais et al. (2008) in Canada. These researchers investigated the linguistic landscapes around two elementary schools in Vancouver and Montreal that enrolled L2 learners. The project drew on innovative resources such as digital photography to show graphically how the children imagined the language(s) of their neighborhoods, and how they constructed their identities in relation to them. Children in both cities were also encouraged to exchange letters, posters, photographs, and videos. Dagenais et al. argued that this documentation of the actual and imagined neighborhoods, as seen by children, provided rich information on the children's understanding of their community. Understanding children's actual and imagined communities had important implications for language teaching and learning in these sites.

Kendrick & Jones (2008) conducted research in Uganda to analyze drawings and photographs produced by primary and secondary girls. Like Dagenais et al. (2008), this study used multimodal methodologies to investigate the girls’ perceptions of their participation in local literacy practices, and to promote dialogue on literacy, women, and development. The girls’ visual images were found to provide insight into their imagined communities, communities in which command of English and access to education was available. They concluded (2008: 397):

Providing opportunities for girls to explore and consider their worlds through alternative modes of communication and representation has immense potential as a pedagogical approach to cultivate dialogue about the nature of gender inequities, and serve as a catalyst for the positing of imagined communities where those inequities might not exist.

In investigating racial discourses found in educational documents in the UK, Blackledge (2003) also used the concept of imagined communities. He found that educational decision-makers’ imagination of a monocultural and monolingual community, seen as normative and natural, marginalized the cultural practices of Asian minorities making regular visits to their heritage countries. From the perspective of this normative homogeneous imagined community, the Asian practices were positioned ‘as aberrant, Other, and damaging to the educational prospects of minority children’ (2003: 332). For him, the discourses of the dominant group, which assumed an imagined community of homogeneity, racialized and stigmatized the cultural practices of Asian families, and accomplished this with apparently ‘common-sense’ arguments.

5. Identity categories and language learning

At the same time that recent research on language learning emphasizes the multiplicity of learners' identities, a growing group of researchers is interested in exploring how such relations or identifications as race, gender, and sexual orientation may impact the language learning process. These researchers do not regard such identity categories as ‘variables’ but rather as socially and historically constructed processes within particular relations of power. As Gal (1991: 176) argued with respect to gender, such relationships are ‘system[s] of culturally constructed relations of power, produced and reproduced in interaction’. Language educators have increasingly examined how identity categories and language learning might be intertwined. Special issues of the TESOL Quarterly on ‘Gender and Language Education’ (Davis & Skilton-Sylvester 2004) and ‘Race and TESOL’ (Kubota & Lin 2006) examined such issues, and books by Fought (2006), Rampton (2006), Heller (2007), Lin (2008), and May (2008) explored issues such as ethnicity and privilege in relation to language learning. We here examine some of the research on language learning with respect to identity categories of race, gender, and sexual orientation.

Race and ethnicity have long been recognized as connected to identity, and several scholars have been interested in the relationship between race and language learning (Ibrahim 1999; Lin et al. 2004; Curtis & Romney 2006; McKinney 2007; Kubota & Lin 2009). Ibrahim (1999) conducted research with a group of French-speaking continental African students in a Franco-Ontarian high school in Canada to ascertain the impact on language learning of ‘becoming black’. Recognizing that there are particular linguistic styles associated with ‘being black’ in North America, he argued that the African students’ use of Black American Stylized English was a way for these students to construct identities familiar to (and often esteemed by) their peers, and to connect to North American imagined constructions of ‘blackness’.

Research in South African multilingual contexts has provided a different set of insights into issues of race with respect to language learning, and the learning of English in particular (e.g. Makubalo 2007; McKinney 2007; Nongogo 2007). McKinney (2007) conducted research on the language practices of black South African students attending high schools that had previously enrolled white students. She showed that these black youth had sophisticated understandings of themselves and others in relation to different ‘brands’ of English as well as to the use of local African languages. Although South Africa has eleven official languages, it is English that is the language of power. One learner referred to the prestige variety of English as ‘Louis Vuitton English’, representing English as a commodity (McKinney 2007: 14). Despite the criticism endured by black students who are acquiring a prestige variety of English (being seen as ‘becoming white’), black students resisted this derogatory positioning and showed their awareness of the different kinds of cultural capital carried by varieties of English and local languages. They showed that they were mindfully acquiring English for their own uses rather than identifying with white first language speakers of English in their language acquisition processes.

In Kubota & Lin's 2006 special issue of TESOL Quarterly on ‘Race and TESOL’ (Kubota & Lin 2006), and their 2009 book (Kubota & Lin 2009), the authors presented a wide variety of research that investigated the relationship between race and language learning. In their introduction to the special issue, Kubota & Lin argued that while many other academic disciplines have for some time concerned themselves with the problematic social category of race, the field of TESOL has been remiss in this regard, but that it ‘could initiate unique and vibrant inquiries to build on these topics and investigate how they influence identity formation, instructional practices, program development, policy making, research, and beyond’ (2006: 473). Noting that ‘race parallels the nation as imagined community’ (2006: 474), Kubota & Lin presented several studies that problematized the concept of race, investigating the ways in which race has been historically constructed and how the discourse of race has affected language teaching and learning.

One of the studies in the Kubota & Lin special issue is that of Motha (2006), who observed that ‘school and classroom practices provide the terrain in which meanings of racialized identities are dynamically and continuously constructed and negotiated’ (2006: 497). Motha's research with four American teachers attempting to create anti-racist pedagogies illustrated the complexities such a commitment involved. For example, the Korean American teacher (the only teacher of color among the research subjects) believed that her professional legitimacy was judged inadequate by colleagues, and that this judgment contributed to her feelings of professional inferiority. Motha's analysis of the conversations around race of the four research subjects implied what has become a common theme: teachers must critically examine, with their students, how language practices can oppress or liberate, and how the equating of whiteness with ‘standard English’ should be challenged.

Shuck (2006), in the Journal of Language, Identity, and Education, explored the ways American public discourse links language with race as a way of positioning groups. White undergraduates who spoke English as a first language at a southwestern US university told Shuck in interviews that non-native speakers with non-European origins were incomprehensible, intellectually inferior or lesser, and responsible for their ‘non-integration’ in American society. These students placed responsibility for creating comprehensibility with the non-native speaker, not the native English-speaking interlocutor.

A variety of scholars have made important contributions with regard to another identity category, gender, and its impact on language learning (Pavlenko et al. 2001; Norton & Pavlenko 2004; Sunderland 2004; Cameron 2006; Menard-Warwick 2009; and Higgins 2010). These scholars, like many others, conceptualize gender not only in terms of male/female divides, but as a system of social relationships and discursive practices that may lead to systemic inequality among particular groups of language learners, including women, the poor, minorities, the elderly, and the disabled. What these scholars found, amongst other things, was that different forms of hybridity were common in studies of gender and language learning. One such example was Taylor's (2004) ethnographic research conducted in an anti-discrimination camp in Toronto, which uncovered a combination of discourses she called ‘racialized gender’. With reference to interviews with two young women, Hue from Vietnam and Khatra from Somalia, Taylor demonstrated convincingly how discourses of racialized gender were prominent in the lives of these secondary school students.

Hybridity is also a key theme in Kamada's study with young women in Japan, comprehensively addressed in her compelling book, Hybrid identities and adolescent girls: Being ‘half’ in Japan (Kamada 2010). Drawing on a longitudinal study with six adolescent girls of mixed background, Kamada found that the girls were all engaged in daily battles for respect and recognition both inside and outside the classroom. While the girls learnt from a very young age that ‘the nail that sticks up gets hammered down’ (2010: 101), they also learnt that practices of marginalization were marked by gender. While boys were more likely to be boisterous and overt in their negative behavior, girls were more likely to use more covert, and sometimes supportive, practices. As one of the participants, Rina, noted ‘if you're girls they just let you be, but if they're boys they might bully you’ (2010: 102). Significantly, it was only over time that the girls began to appreciate that there was cultural capital associated with being ‘hybrid’.

The extent to which gender impacts the language learning of adult women is explored with great insight by Menard-Warwick (2009), who examined how the gendered life histories of immigrant women in California shaped their participation not only in the English classroom, but also in the education of their children. Of particular significance in the study was the larger sociocultural context of Latin American immigration to the USA, and struggles over poverty, unstable migrant status, childcare, and schooling. The study highlights the fact that profound diversity exists both within and among groups of women, notwithstanding commonalities of experience.

Still another identity category, sexual orientation, has been interrogated in studies of language learning (King 2008; Moffatt & Norton 2008; Nelson 2009). Many of these scholars are interested in the ways teachers can create welcoming classrooms for gay, lesbian, transgendered, or questioning learners. Nelson (2009) pointed out the differences between a pedagogy of inclusion, which aims to introduce images as well as experience of gays and lesbians into curriculum materials, and a pedagogy of inquiry, in which students and teachers alike investigate how linguistic and cultural practices naturalize only some sexual identities, most notably heterosexuality. Such inquiry requires teachers and students to be alert to discourse and other means used to make non-heterosexual relationships ‘other’, and it also aims at provoking and supporting action on the basis of inquiry. Nelson's distinction, and her description of a pedagogy of inquiry, is helpful and applicable in other sites where aspects of learners' identities are ignored or stigmatized.

6. Methods of studying identity and SLA

The methods required for investigating the intersection between identity positions and language learning are complex, given a poststructural approach to identity and sociocultural theories of learning (Pavlenko & Blackledge 2004; Norton & McKinney 2011). For example, methods that rely on static, inherent, and measurable learner ‘variables’ are not consistent with some of the major understandings of these approaches. Other research methods are required to deal with such complexities, and the focus on issues of equity and power that we see in much work in this tradition calls for qualitative research designs that are informed by critical research. Methods that scholars use in identity approaches to language learning therefore often draw on critical ethnography, feminist poststructuralist theory, sociolinguistics, and linguistic anthropology. Identity and language learning scholars rely frequently on three methodological understandings (see also Norton & McKinney 2011).

First, most reject the view that any research can claim to be objective or unbiased. Poststructural researchers must be reflexive about their own experiences, recognizing that their perspective on that which they are observing or analyzing is not the only one, and that their conclusions will inevitably be ‘situated’ and partial. This is not to say that qualitative research is lacking in rigor; but it is to recognize that all research is situated, and that the researchers and their tools are integral to the progress and ethical conduct of a research project (Kramsch & Whiteside 2007; De Costa 2010b; Norton & Early in press). Ramanathan (2005) noted, for example, ‘Questions and issues of what are “present” and “absent” clearly underlie what are “visible” and “invisible” in literacy events and practices and are determined, to a large extent, by the researcher's lens’ (2005: 15).

A second common understanding with respect to method is the ‘structure and agency’ issue: that is, identity researchers must account for not only how structural conditions and social practices place individuals, but also how individuals struggle to situate themselves in the contexts in which they find themselves. For example, identity and language learning researchers must examine identity categories like race, class, gender, sexual orientation, and other structural issues that might be associated with inequalities in access to language learning. However, they also need to ensure that they leave conceptual room for the actions and investments of human agents. Holland et al. (1998) put this elegantly: ‘Human agency may be frail, especially among those with little power, but it happens daily and mundanely, and it deserves our attention’ (1998: 5). Methods for examining L2 learning and identity thus need to pay close attention to how individuals are placed by common societal practices, but also how they place themselves by engaging in societal practices in innovative ways.

A third and related understanding is that the methods that identity researchers use must seek a better understanding of how political and economic issues interact with language learning, constraining or enabling human action (Cummins 2000; Fairclough 2001; Pennycook 2007; Janks 2010). Such researchers often draw on Foucault's (1980) insights about the relationship between knowledge and power, and the subtle and complex ways in which power circulates in society. Foucault's conceptualization of power as discursively produced and reproduced is of special interest to language educators as they investigate particular learning environments and how they privilege or stigmatize learners. Pennycook (2007: 39) noted that

Foucault brings a constant scepticism towards cherished concepts and modes of thought. Taken-for-granted categories such as man, woman, class, race, ethnicity, nation, identity, awareness, emancipation, language or power must be understood as contingent, shifting and produced in the particular, rather than having some prior ontological status.

This approach encourages language education researchers to reject ‘grand theories’ and methods, and to come to understand the particularity of the persons, environments, and processes they wish to examine.

In research on identity and language learning, there has thus been a strong methodological focus on learner and teacher narratives, collected either through fieldwork (Goldstein 2003; Miller 2003; Stroud & Wee 2007; Barkhuizen 2008; Botha 2009) or from autobiographical and biographical accounts (Pavlenko 2001a, 2001b; Johnson & Golombek 2002; Benson & Nunan 2005; Todeva & Cenoz 2009, Wildsmith-Cromarty 2009; Nunan & Choi 2010). Pavlenko (2001b: 167) argued for the particular contribution that narrative can make:

L2 learning stories. . . are unique and rich sources of information about the relationship between language and identity in L2 learning and socialization. It is possible that only personal narratives provide a glimpse into areas so private, personal and intimate that they are rarely – if ever – breached in the study of SLA, and at the same time are at the heart and soul of the L2 socialization process.

Narrative inquiry has become an increasingly important research method in many of the social sciences, and particularly in education, and is the topic of a forthcoming special issue of TESOL Quarterly (Barkhuizen in press). Connolly & Clandinin (1990) pointed out that it is a method that has potential for ‘the improvement of practice and of how researchers and practitioners might productively relate to one another’ (1990: 12). Specifically, they note that as researchers recognize that their own lives and experiences have effects on their research efforts and products, narrative inquiry might become a way for researchers and research participants to produce ‘collaborative stories’ (1990: 12) (see also Clandinin & Connolly 2000).

Fieldwork-based studies of identity and language learning often combine diverse methods of data collection, including ethnographic observation, interviews (including life history interviews), diary studies, journal writing, and written responses (narrative or other) to researcher questions. Ethnography's traditional reliance on longitudinal observation has characterized some such studies, and sociocultural theory ‘seeks to understand mental development and learning by considering not only the contextual specifics but also the process over time, rather than focusing only on a particular moment of spoken or written production’ (Swain et al. 2010: xii).

In the ethnographic classroom research conducted by scholars such as Toohey (2000, 2001), Hawkins (2005), and De Costa (2010b), several ethnographic data collection methods were used, including regular classroom observation documented in field notes, audio recordings and video recordings; interviews and ongoing informal discussions with the children's teachers; and home visits, where parents were interviewed. The combination of these methods provided rich data for the analysis of learners' classroom language learning as socially, historically, and politically constructed, and of the classroom as a site where teachers, children, and the researcher were engaged in identity negotiation.

Despite its strengths, qualitative methodology in research on language and identity has important challenges, as recognized by other social scientists like Hammersley (1992) and Gao (2007). From their research on task-based language learning in urban settings in the UK, Leung, Harris & Rampton (2004) pointed out that qualitative researchers often have difficulty presenting their data completely, in all their ambiguity and inelegance. In a study they conducted, the method was to collect naturally-occurring data with the use of video and audio recordings, supplemented by field notes. These data did not unequivocally support the conceptual framework in which the researchers were working (in their case, task-based language learning), with some data being inexplicable. Another problem they encountered was how to decide which data were relevant, and how many data should be analyzed. They suggested that rather than ‘smoothing’ this messiness, qualitative researchers should expressly pay attention to data that do not fit their chosen conceptual framework. In so doing, they suggested, researchers can acknowledge the complexities of research participants’ experiences and can also advance theory-building in accounting for more of what they observe.

In an explicit attempt to acknowledge the complexities of conducting research on identity and language learning in classrooms, Denos et al. (2009) describe a research collaboration between teachers and university-based researchers in Vancouver, Canada, with the mutual goal of investigating what practices in classrooms would make a difference to the learning opportunities of minority-language children. While collaboration was an explicit goal of the project, the collaborators found the activities and intellectual currency of academia and of schools to be highly divergent, and in some cases, collaboration proved problematic. For example, while teachers greatly appreciated the opportunity to discuss and critique educational practices with university colleagues, they expressed ambivalence towards writing at length about practice in publishable academic papers (an important activity of academia), noting that they felt little ownership of the academic language characteristic of many published journals. Recognizing the problem of incommensurate discourses between researchers and teachers, Sharkey & Johnson (2003) initiated a productive and engaging written dialogue between them, as a means of demystifying research and theory that addresses themes of identity, power, and educational change.

Future research focused on identity and power in SLA will likely expand our methodological tool-kits and theoretical frameworks. For example, Wagner (2004) indicated an interest in the method of conversation analysis of L2 conversations to support an empirical ‘understanding of learning as empowerment of social participation’ (2004: 614). Block (2007a) also recently commented on how analysis of naturally occurring interactions would enhance our understanding of identity and language learning. While there are several identity-focused analyses of L2 classroom talk (e.g. Harklau 2000; Duff 2002; Pomerantz 2008; Talmy 2008), analyses of talk outside the classroom are less common, and should provide very helpful information to enhance our insight.

Innovative research on the written text will also increase our understanding of identity and language learning (Ivanič 1998; Starfield 2002; Le Ha & Baurain 2011). How texts mediate identities is a theme of recent scholarship in actor network theory, which advocates tracing how technologies (e.g. texts), once invested with power or responsibility, act on behalf of humans, have agency, and influence practices (including identity practices) in local settings (Clarke 2002; Hamilton 2009). Hamilton, for example, showed how a particular assessment document positioned and conferred identities on adult literacy students, and led to a reductive view of literacy.

7. Identity and language teaching

The relevance of identity research for classroom teaching has been investigated by a number of language scholars. Drawing on anthropologists Holland et al.'s notion (1998: vi) that ‘Identities – if they are alive, if they are being lived – are unfinished and in process’, such researchers encourage teachers to regard students’ identities as potential, and to experiment with activities that do not lock students into ‘finalized identities’. Ascertaining what is both desirable and possible requires ongoing negotiations among teachers, administrators, and policy-makers, in the context of broader material conditions that can serve to constrain or enable the range of identity positions available to students (Mohan, Leung & Davison 2002; Tsui & Tollefson 2007; McKinney & Norton 2008; Lo Bianco, Orton & Gao 2009; Blackledge & Creese 2010). If language educators recognize that diverse classroom practices offer learners a range of positions from which to speak, listen, read, or write, it is important for educators to explore with students which identity positions offer the greatest opportunity for social engagement and interaction. And, of course, if there are identity positions that silence students, teachers need to investigate and address practices that marginalize students. We address this topic in more detail, with reference to a range of perspectives from the international community, digital innovations, and classroom resistance.

7.1 International pedagogical perspectives

Apart from the studies discussed so far, there have been many other research studies conducted in diverse global settings that investigate how particular pedagogical practices can either constrain or enable the range of identity positions available to language learners in classroom settings. Lee's (2008) research in a Canadian post-secondary institution suggested the complexities involved when teachers want to enhance the range of identities available to their students, but when customary classroom practices re-assert fixed positions. Despite teachers’ best intentions, Lee argued, pedagogical decisions can reinforce subordinate student identities, and limit students’ access not only to language learning opportunities, but also to their imagination of more desirable identities.

Lee's findings echo those of Ramanathan (2005), who found that teachers’ language instructional practices in secondary schools in India promoted and strengthened existing inequities among students learning English in two different kinds of schools. She examined the adjustment of students who had been educated in either Gujarati- or English-medium schools through grades K-12, when they moved to English-medium tertiary level institutions. She found, perhaps not surprisingly, that students who received English-medium instruction through high school were more successful in English-medium colleges than those schooled in the vernacular. In explanation, she further noted that in the English-medium secondary schools, students were required to engage in creative and high level analysis of English literature, while in the vernacular-medium secondary schools, students, who were mostly lower-caste, were engaged in grammar and translation exercises. Ramanathan suggested that language instructional practices (mandated by curricular statements that might not be under the control of local teachers) like grammar and translation, which do not call on higher-level thinking skills, may limit language learners' progress and thus their access to more powerful identities. This study showed how larger institutional arrangements may limit teachers’ efforts to enable students to appropriate more powerful identities.

Another example of institutional practices that constrain possibility is found in the Mexican context, where López-Gopar (2007) discussed the identity assignation of indigenous Mexicans as ‘illiterates’. He quoted Molina Cruz (2000: 405), who described this situation:

The stigma of illiteracy hurt our parents and grandparents when, because of not knowing how to read and write or how to speak Spanish, they were considered ignorant, people without culture, and even worse, in our indigenous mother tongue they were called ‘bene tont’, a term only used for the inept, the retarded, the crazy [people]; ‘bene tont’ equals illiterate. ‘Bene’ means person or man and ‘tont’ means fool, a word we borrowed from Spanish.

López-Gopar argued that this alienating identity comes from narrow views of literacy that preclude recognition of the many multimodal literacies of indigenous peoples. He further argued that building upon the multiliteracies that indigenous children brought to Mexican schools would ameliorate the schooling difficulties of many such children (see also Martin-Jones & Jones 2000; Hornberger 2003; Prinsloo & Baynham 2008).

More promising pedagogical projects have been described by researchers in another Mexican study, in diasporic Chinese communities, and in South Africa, Uganda, the USA, and the UK. These classrooms, and many other transformative classrooms that have been discussed in the literature (see, for example, the many projects discussed in Norton & Toohey 2004), are led by language teachers who have broad conceptions of ‘language’ and thus of ‘language teaching’.

In Oaxaca, Mexico, Clemente & Higgins (2008), for example, described their research with non-native English-speaking student teachers of English, showing how these pre-service teachers questioned the dominant role that English plays in the globalized political economy, and how they sought to appropriate and perform English without sacrificing their local Mexican identities. Describing their research site as a contact zone, Clemente & Higgins demonstrated that the student teachers in their study satirized and, in effect, ‘de-throned’ English through various forms of language play in both English and Spanish, and argued that the student teacher groups were safe havens in which participants could play with both languages. Such performances allowed them to explore various identity positions, as a counter-discourse to dominant discourses of the native English teacher. As one student teacher said (2008: 123):

I have a Mexican accent. English is mine from the very moment I put it into practice and I am able to establish communication. But when I say that the English language is mine, I do not mean to say that I want to take the culture that comes with it.

The situatedness and hybridity of language learners' identities is also emphasized in the language and teaching of diasporic Chinese communities around the world, where researchers are investigating the heterogeneity of learners of Chinese as a Heritage Language (CHL), and their diverse needs and desires for learning (He & Xiao 2008). In this regard, the observation by Jason, as quoted in He (2010: 66) is particularly poignant:

My home language is Chinese. My parents are from China. They praised me, scolded me, all in Chinese . . . My Chinese is really bad. I can't read and I can only write my name. But when I think of Chinese, I think of my mom, dad, and home. It is the language of my home, and my heart.

In research on CHL, which extends to research on the learning of Chinese as a foreign language, many scholars are pointing out that identity (ethnic, cultural, and linguistic) is a major factor in learners' investment in learning Chinese (Li & Duff 2008; Duff et al. forthcoming).

In a different educational context, Stein (2008) described transformations in South African schools in which local forms of representation had been marginalized and devalued by the apartheid system. Teachers changed these under-resourced township schools by re-appropriating and re-sourcing local textual, cultural, and linguistic forms. Teachers provided opportunities for their students, English language learners, to make use of multimodal resources, including linguistic, bodily, and sensory modes, in order to engage in representing meaning. Learners were enthusiastic participants in production of multimodal counter-texts that subverted the usual ‘proper’ school topics and readings, and drew on topics considered by some to be taboo.

Recent research in Uganda investigated how instructional activities involving drawing, photography, and drama could be incorporated more systematically into English language classrooms (Kendrick et al. 2006; Kendrick & Jones 2008). Working in two Ugandan sites, Kendrick and her colleagues found, as did Stein, that multimodal activities offered teachers new and expanded ways to validate students’ literacies, cultures, and identities, and in addition, such activities were engaging and supportive of English language learning in the classroom. In a photography project, for example, students began to use English for communication and expression of students’ own meanings, a development from their former perception of English as a somewhat restrictive and artificial medium of instruction.

In the UK, Wallace (2003) described her work with adult language learners in which she presented a variety of popular texts, including newspaper and magazine articles and advertisements. Her work with students in critically analyzing such texts made clear to students how power and meaning structure texts, and how students’ engagement in critical reading could help them question and reshape the powerful discourses that surround and sometimes subordinate them. Wallace pointed out the contrast between dominant ESL methodologies (such as communicative language teaching and task-based learning) and a critical approach, arguing that dominant approaches often ‘domesticate’ learners, focusing primarily on activities designed to adapt students to dominant practices rather than engaging them in the important work of critiquing disempowering discourses.

The teachers described in the above projects conceived of language not primarily as a static linguistic system, but rather as a social practice in which experiences are organized and identities negotiated. These teachers recognize that if learners are not invested in the language practices of the classroom, learning outcomes are limited, and educational inequities perpetuated. Further, such teachers take great care to offer learners multiple identity positions from which to engage in the language practices of the classroom, the school, and the community. In diverse regions of the world, innovative language teachers are seeking to provide learners with a range of opportunities to take ownership over meaning-making, and to re-imagine an expanded range of identities for the future.

7.2 Digital technology, identity, and language learning

In the modern world, print literacy is not enough. People need to be literate in a great variety of different semiotic domains. (Gee 2004: 19)

Multimodal technologies impact twenty-first century citizens and have been investigated by a number of scholars interested in language learning (see, for example, Kramsch & Thorne 2002; Warschauer 2003; Lewis & Fabos 2005; Lam 2006; Mutonyi & Norton 2007; Snyder & Prinsloo 2007; White 2007; Pahl & Rowsell 2010). These scholars recognize the increasing multiplicity of communications channels and media, in which meanings are communicated not only through text but also through music, sound, images, and a variety of digital media, thus ‘redefining what it means to be literate in the twentyfirst century’ (Coiro et al. 2008: 10). Indeed, Kramsch & Whiteside (2008) have argued that what they call symbolic competence is becoming increasingly important in contemporary society. Similarly, Poyntz (2009) reminded us ‘Beginning in infancy, young people now grow up learning the language of mass media through a constant diet of screen images, audio messages, and text-based communication’ (2009: 369).

The way that technology provides language learners with the means to construct imagined lives has been investigated by Lam (2000, 2006), for example, who examined the computer-mediated transnational identities that immigrant youth in the USA were fashioning for themselves as multilingual, multicompetent actors. She found that these identities afforded them possibilities to provide language learning opportunities for themselves that were broader than those available to them in school, where they were stigmatized as immigrants and incompetent language users. Similarly, in Australia, the distance language-teaching programs examined by White (2007) were designed to provide learners with a wider range of choices for foreign language learning in school. She concluded that as innovations in distance learning and teaching expand, it is imperative that the field finds ways of addressing the philosophical, pedagogical, and professional issues that arise, and that issues of identity, for both teachers and learners, are significant factors in each of these domains.

Lewis & Fabos (2005) examined the use of Instant Messaging (IM) by seven young people in the USA for evidence of how these youths' social identities shaped and were shaped by their engagement in this form of digital meaning-making. Not surprisingly, they found that the youth engaged in IM to enhance their social relationships and statuses across contexts, sometimes assuming multiple identities online. IM permitted them to engage in literacy practices in ways they were not able to do in school, and they, like many others, argued that schools would do well to try to link students’ expertise in outside-school literacies with their in-school literacy instruction.

Similarly, Pahl & Rowsell (2010: 76) describe projects in the UK and the USA that use ‘artifacts as a means to reconfigure students who feel marginalized by school literacy to find a place in the English classroom’. Finding that students ‘valued the artifacts in their lives over many of the texts studied in English class, which bear little resemblance to their own worlds’ (2010: 79), teachers asked children to use their knowledge of digital social networking and combine this with investigations of traditional school-identified texts, thus providing a school space for hitherto marginalized students.

Most of the studies that investigate how digital technologies affect identity and language learning have been celebratory, with most observers convinced that ‘Networked electronic communications have given rise to new social spaces, linguistic and semiotic practices, and ways of fashioning the self’ (Lam 2006: 171). However, like analysts of the use of digital technologies with youth generally, Lam also cautioned that these technologies may not ‘necessarily provide the analytical tools that may empower youths to critique and change existing social structures in positive directions’ (2006: 186). A cautionary note was also sounded by Kramsch & Thorne (2002), who indicated, in a study of the synchronous and asynchronous digital communications between American learners of French in the USA and French learners of English in France, that these communications were not always identity-enhancing. They found that students had little understanding of the larger cultural framework within which each party was operating, leading to problematic digital exchanges.

Hull, Jury & Zacher (2007) and Poyntz (2009) argue that teachers play an important role in instructing children and youth to use multimodality for critical purposes. The interest of many scholars in how digital technologies might contribute to enhanced language learning and learner identities will no doubt add to our increased understanding of this relationship in the future. Further, as scholars such as Warschauer (2003), Mutonyi & Norton (2007), Snyder & Prinsloo (2007), and Andema (2009) have noted, digital research on language learning in wealthier regions of the world may be of limited application in other contexts, so understanding how digital literacies are taken up in poorly-resourced communities is essential in future research.

7.3 Identity and resistance

Resistance in language learning was not studied a great deal in the past, but the relationships among identity, language learning, and classroom resistance have become of major interest to several scholars in language education. Such research typically examines how structural constraints and customary classroom practices might position learners in undesirable ways, but that such constraints and practices are sometimes resisted by learners so as to create innovative and unexpected identity relationships.

Canagarajah (2004a), for example, investigated how language learners negotiate learning a new language or dialect (sometimes associated with colonial relationships) but nevertheless maintain membership in vernacular communities and cultures. In his research, he noted that language learners of English in the USA and Sri Lanka were sometimes ambivalent about the learning of an L2 or dialect, recognizing social and economic benefits to such learning, but also serious social losses. He noted that this ambivalence sometimes led to clandestine literacy practices on the part of students to create pedagogical safe houses in the language classroom. In both contexts, the clandestine literacy activities of the students are seen to be forms of resistance to unfavorable identities that learners may be assigned because of their participation in the L2 community. Canagarajah argued that these safe houses serve as sites of identity construction, allowing students to negotiate the often contradictory tensions they encounter as members of diverse communities.

McKinney & van Pletzen (2004) provide a second example of research on identity and resistance in language learning. They introduced critical reading into their first year English studies course using two curriculum units on South African literature. Their students, relatively privileged, and studying at a historically white and Afrikaans university in South Africa, felt uncomfortably positioned by the curriculum materials (concerned with South Africa's apartheid past), and therefore resisted these materials. In response, the authors described their attempts to use the course so that both they and their students could explore the many private and political processes through which identities are constructed. Students’ resistance became framed as providing impetus for consideration of these important and complex issues.

A third example is provided by Talmy's (2008) investigation in Hawai'i of the multiple ways in which secondary school English language learners resisted being positioned as ‘ESL students’. The school expected that ESL students would bring required materials to class, read assigned fiction, do bookwork, meet assigned dates, follow instructions, and work for the full class session. However, resistant ESL students frustrated these expectations, leaving materials ‘at home’, talking with friends, and playing cards. Talmy pointed out that the students’ resistance led to a particularly undesirable end, transforming the ESL program into precisely what the students disliked most, ‘an easy, academically inconsequential program that did little to meet their L2 learning or educational needs’ (2008: 639).

8. Critiques of the field

While many language scholars have been excited about new directions in identity research and the conceptual tools provided by poststructuralist theory, they have also engaged in vibrant critique of the field (Norton & Morgan in press). As mentioned in Section 6 above, one area of debate concerns the relationship between societal structure and human agency. Luke (2009: 293) noted, for example, that resistance to marginalizing discourses may be compromised by students’ ‘phenotypical features, their gender or sexuality, their language and accent [which] are not chosen, not wholly malleable through discourse’. He argued that, notwithstanding the multiplicity of identity celebrated by poststructuralist theory, some identifications so strongly determine social relations that resistance is difficult. He also noted that strategic essentialism (2009: 292), ‘the claim of shared historical origin and unity for political and cultural purposes’, might be important for groups who wish to reclaim cultural and linguistic solidarity.

A second area of debate discussed by Norton & Morgan (in press) concerns the challenges that language teachers, learners, and researchers might experience in contexts in which ‘truth’ remains a relative term. As they note, ‘when truth, reality, and meaning become pluralized and destabilized – as the work of Derrida and Foucault would indicate – we can become politically paralyzed’. While the destabilization of knowledge and meaning can be liberating, the challenge is to determine a principled basis for action. Here the issue of ethics and values, of great interest to an increasing number of language education scholars, becomes central (Johnston 2003; Clarke 2009; Crookes 2009).

In the area of methodology, Sharkey (2004) raises important concerns about the limitations of autobiography in language teacher education. Drawing on a self-study while still a graduate student, Sharkey reflects on her use of autobiography in relation to what is ‘untold’ in autobiographies. She discusses, in particular, two examples of self-censorship with respect to issues of Jewish identity in the story of Amy, a methods student of Sharkey, and sexual identity, Sharkey's own story. Sharkey notes as follows, ‘Our acts of self-censorship raised numerous questions regarding autobiographies produced in teacher-education classrooms and has led me to question the kinds of pedagogical spaces I help create and sustain’ (2004: 507). To address these concerns, Sharkey suggests that language teachers should carefully attend to what students share outside of written texts, and that silence should not necessarily be seen as a deficit, but as a political act of resistance.

There also exists in the literature on identity and language learning a lack that seems curious, given the political affiliations of many such researchers, and this concerns the analytical construct of ‘class’ with respect to issues of economic privilege. While there are some researchers such as Kinginger (2004), Luke (2004), Ramanathan (2005), Rampton (2006), and Kanno (2008), who address economic privilege (or the lack of it) in some depth, the category ‘class’ has not received the kind of attention accorded other identity categories such as gender, ethnicity, and race. Block (2007b) speculated that the relative absence of class identification in language learning research may be the result of contemporary uneasiness about the perception of determinism in Marxist rhetoric about class, and/or because economic structures have been so radically changed by globalization that ‘classes’ are difficult to distinguish. Whatever the reasons for limited research on class, differential privilege is widely acknowledged in language education and identity literature, and examinations of class and its effects are relatively common in the wider educational literature. In any case, sustained attention to privilege (in perhaps the nuanced way that Bourdieu writes about different kinds of capital) is still to come.

9. Future directions

This review of foundational and contemporary research on identity, language learning, and social change shows the vibrancy and energy that are stimulating many researchers and much debate. We have noted that current language education scholars see language and language learning as situated in particular contexts that are complex and dynamic. No longer are static views of language as system and language learning as internalization of that system seen as adequate in a world in which boundary-crossing, multilingualism, and human agency are recognized. As we have argued in this review, language learning researchers are increasingly turning to literature in diverse fields such as anthropology, sociology, post-colonial and cultural studies, and education to better understand the language learning contexts in which they work. Researchers who investigate identity and language education in the future will need to be comfortable with this interdisciplinarity (see Warriner 2007). The review also suggests that researchers will require increasingly sophisticated understandings of difference, and of cosmopolitan identities (Luke, Luke & Graham 2007). Morgan & Ramanathan (2005) underlined the importance of understanding that learners live in globalized sociocultural worlds. Such an understanding will require, even more than before, consideration of global social structures that impinge on the identity of language learners and their learning.

Accordingly, we are convinced that an understanding of identity and SLA processes must be enriched by research conducted in postcolonial and indigenous sites, where multilingualism is ubiquitous and language acquisition processes can be quite different from language learning experiences in the West (for example Ndebele 1987; Block & Cameron 2002; Lin & Martin 2005; Morgan & Ramanathan 2005; García, Skutnabb-Kangas & Torres-Guzmán 2006; Canagarajah 2007; Pennycook 2007, 2010; Rassool 2007; Blommaert 2008; Tembe & Norton 2008; Alim et al. 2009; Higgins 2009). Canagarajah (2007) challenged the monolingualist assumptions underlying much of SLA theory, arguing that ‘insights from non-Western communities should inform the current efforts for alternate theory building in our field’ (2007: 935). In such multilingual contexts, it is likely that the term SLA itself is inappropriate: Block (2003: 5) noted that the term ‘second’ doesn't capture the ‘experiences of multilinguals who have had contact with three or more languages in their lifetimes’. In applied linguistics more generally, two recent issues of the AILA Review of the International Association of Applied Linguistics on ‘Africa and Applied Linguistics’ (Makoni & Meinhof 2003) and ‘World Applied Linguistics’ (Gass & Makoni 2004) have begun to broaden the scope of the field.

With respect to the learning and teaching of English, which remains a key interest in applied linguistics research, Morgan & Ramanathan (2005) have argued that language educators need to discover ways to decolonize English language teaching, and thus to provide a wider range of identities for English language learners. Recognizing that Western interests have an almost exclusive hold on the language teaching industry, they argued that much more attention needs to be put to the issues, problems, and triumphs in language teaching in other than Western sites, and to how English language learners in those sites are positioned. Supporting agency and professionalism in poorly resourced communities is seen as important (Canagarajah 2004b; Higgins 2009), as is the need to better understand local vernacular modes of learning and teaching (Bhattacharya et al. 2007; Le Ha 2008). Indications of progress on these aims are growing. Examples include special issues of the TESOL Quarterly on ‘Language in Development’ (Markee 2002), ‘Language Policies and TESOL’ (Ramanathan & Morgan 2007) and ‘Migration and Adult Language Learning’ (Burns & Roberts 2010).

Another focus for research gaining momentum in the investigation of identity and language learning is that of the language teacher and the language teacher educator (Morgan 2004; Hawkins 2004; Pennycook 2004; Creese 2005; Varghese et al. 2005; Richards 2006; Tsui 2007; Clarke 2008; Hawkins & Norton 2009; Nunan & Choi 2010; Norton & Early in press). Pennycook (2004), for example, reflected on his experience of observing a teacher in a TESOL practicum in Sydney, Australia. Noting that a great deal of language teaching happens in community programs, places of worship, and immigrant centers where resources are limited, he considered the ways in which teacher educators must understand the differences between these contexts and those of the privileged language education institutes in which they work. He also argued that there are critical moments in language teacher education that can provide opportunities for student teachers to situate their own and their students’ lives and concerns in recognition of power and authority in the wider society, and to understand how their teaching might be directed toward social and educational change.

In conclusion, we have drawn on a rapidly increasing body of research to argue that language learners' identities are always multiple and in process, and that learners often have different investments in the language practices of their classrooms and communities. There are widespread representations of complex and embodied language learners living in socially stratified worlds that constrain as well as enable the exercise of human agency. Future research on identity and language learning should further the goal of coming to understand and contribute to more equitable and agentive language teaching and learning practices and environments. We have also made the case that the imagined identities and imagined communities of learners are central in the struggle for legitimacy. As language learners in every region of the world claim the right to speak and be heard, their identities and investments will continue to generate exciting and innovative research.