Every person. . . has an accent. Your accent carries the story of who you are—who first held you and talked to you when you were a child, where you have lived, your age, the schools you attended, the languages you know, your ethnicity, whom you admire, your loyalties, your profession, your class position: traces of your life and identity are woven into your pronunciation, your phrasing, your choice of words. Your self is inseparable from your accent. Someone who tells you they don't like the way you speak is quite likely telling you that they don't like you. (Matsuda, Reference Matsuda1991, p. 1329)
INTRODUCTION
We begin our review by quoting Mari Matsuda, an acclaimed American lawyer, scholar, and activist who has spoken and written extensively on issues of feminism, critical race theory, and civil rights. The above quote begins Matsuda's eloquent essay condemning the practice of discrimination based on accent or societal and legal prejudices stemming from people's judgment about other speakers’ speech. Although accent discrimination falls outside the scope of this review, which is to discuss the link between ethnic identity and second (and subsequent) language learning, Matsuda's poignant description of accent seems remarkably relevant. It highlights the inextricable bond between our speech patterns and the myriad of variables that shape perceptions of who we are, including our sense of ethnic identity. The goal of this review is to unravel at least some aspects of this bond, by showing evidence from social-psychological and applied linguistic research. We start by defining and contextualizing ethnic identity and its possible relationships to language. We then review recent empirical evidence for the link between ethnic identity and second language (L2) measures and provide theoretical framing for these findings. We conclude by describing several avenues for advancing this area of research.
DEFINITION
Ethnic identity can be broadly defined as a subjective experience of being a part of an ethnic group (Ashmore, Deaux, & McLaughlin-Volpe, Reference Ashmore, Deaux and McLaughlin-Volpe2004), and in the case of L2 learning and use, the ethnic groups in question typically are learners’ own (ancestral) ethnic group and one or more target language (L2) communities. The construct of ethnic identity embraces the feelings, experiences, and behaviors that in their totality amount to individuals’ positioning with respect to their membership in a single or multiple ethnic groups. Although many components of ethnic identity exist (Cameron, Reference Cameron2004; Leach et al., Reference Leach, van Zomeren, Zebel, Vliek, Ouwerkerk and Spears2008), it appears to include at least three elements: centrality of the group to one's self (perceived importance of the group membership), positive affect associated with the group (feelings of joy and pride of being a group member), and in-group ties (perceived strength of affiliation to one's group). Ethnic identity forms part of a large field of identity studies, which spans the domains of clinical, social, and cross-cultural psychology; sociolinguistics; sociology; political science (especially in the realm of immigration and language policy); and forensics and law. Of particular interest to researchers interested in L2 development, however, is the question of how individuals’ sense of ethnic identity relates to their linguistic performance and, vice versa, how language practices might shape one's ethnic identity.
HISTORICAL PERSPECTIVE AND OBJECTIVE
The pioneering work of Wallace Lambert and his colleagues paved the way for understanding how individuals’ conception of ethnic identity might relate to language learning (for another perspective, see Le Page & Tabouret-Keller, Reference Le Page and Tabouret-Keller1985). Central to Lambert's research were the ideas that language learning involves contact between individuals or groups, which increases the chance that matters of personal and group identity will become salient. Consequently, becoming bilingual or multilingual involves not only the learning of language as a linguistic system but also the interpretation of its cultural aspects, including personal and ethnic identity (Lambert, Reference Lambert1967, Reference Lambert, Giles, Robinson and Smith1980). Lambert described several patterns of how individuals’ conceptions of ethnic identity might interact during language learning. One pattern illustrates subtractive bilingualism or assimilation, whereby individuals (usually members of a minority group) acquire the language of a majority group and often lose their own language and culture. Another pattern reflects additive bilingualism or integration, with speakers embracing a new language and culture despite a strong sense of ethnic identity, thus adding a new language and culture without losing their own. Although particular interaction patterns depend on the given social and political context (Sachdev, Giles, & Pauwels, Reference Sachdev, Giles, Pauwels, Ritchie and Bhatia2012), Lambert emphasized that adding a new language and culture should not lead to a loss of identity.
Clearly, describing ethnic identity in terms of discrete interaction patterns may prove simplistic, which is a view supported by many identity scholars (e.g., Block, Reference Block2007; Norton & Toohey, Reference Norton and Toohey2011). Nevertheless, the categories of assimilation and integration outlined by Lambert, as well as related labels, such as adaptation, convergence, and divergence (Giles, Reference Giles2012), are useful in framing L2 ethnic identity research. This is because they provide anchor points for integrating and interpreting psychological research on ethnic identity, which focuses on social aspects of cross-cultural communication, migration, and settlement (Nguyen & Benet-Martínez, Reference Nguyen and Benet-Martínez2013), and research in applied linguistics, whose objective is to understand L2 development. Therefore, the aim of this article is to pull together evidence from social psychology and applied linguistics by focusing on the identity–language link from the perspective most relevant to L2 development, namely, by considering how ethnic identity might be implicated in L2 learning.
CONTEXTUALIZATION
However, before addressing the main issue, which is the identity–language link, we first briefly contextualize ethnic identity from conceptual and theoretical perspectives. Such framing is important because ethnic identity, as any other identity construct, is a complex phenomenon. First, ethnic identity is believed to be a highly subjective dimension, in that it does not reflect externally derived social or political labels but instead corresponds to the individual's subjective positioning towards a group. Ethnic identity is also only one dimension among many social identities individuals may possess, such as gender, age, race, sexuality, or social class identities. In addition to ethnic identity, one's personal histories and experiences, which make an individual's personal identity, might contribute to the successes and failures of intercultural contact. Furthermore, people can possess multiple ethnic identities and can integrate them in a multitude of ways depending on the micro and macro contexts of interaction. Last but not least, ethnic identities are likely multifaceted and dynamic, with the consequences that they can be “negotiated” not only across different types of intercultural contact but also within a single intercultural transaction (e.g., Block, Reference Block2007; Edwards, Reference Edwards2009; Norton & Toohey, Reference Norton and Toohey2011).
The multidimensional and contextualized nature of the ethnic identity construct is reflected in the variety of theoretical frameworks proposed to explain it. With respect to applied linguistics, identity theories can be roughly categorized into social-psychological theories (e.g., Noels & Giles, Reference Noels, Giles, Ritchie and Bhatia2009) and sociocultural theories (e.g., Lantolf & Thorne, Reference Lantolf and Thorne2006). Common to social-psychological perspectives is the emphasis on intergroup behaviors, such as communication, accommodation, acculturation, social change, or discrimination, with respect to individuals’ and groups’ identity profiles. In turn, sociocultural constructionist views maintain that identity, including ethnic identity, emerges through patterns of interaction during which multiple identities are negotiated dynamically. These views also hold that negotiation of identities takes place in socially constructed contexts or communities of practice and involves the exchange of cultural knowledge for positioning one kind of identity over another (Norton & Toohey, Reference Norton and Toohey2011). Although at first glance these views might seem to conflict because of their orientations towards either psychological or sociocultural research traditions, they nevertheless appear highly complementary in that they highlight the complex nature of ethnic identity as a social, cognitive, and cultural phenomenon (Hulstijn et al., Reference Hulstijn, Young, Ortega, Bigelow, DeKeyser, Ellis and Talmy2014).
ETHNIC IDENTITY AND L2 LEARNING: ESTABLISHING A LINK
It comes as no surprise that language and ethnic identity are inextricably linked. Barring a handful of exceptions (e.g., Northover & Donnelly, Reference Northover and Donnelly1996), ethnic groups typically consider language as a salient identity symbol (e.g., Edwards, Reference Edwards2009; Fought, Reference Fought2006). As shown by Labov's (Reference Labov, Gumperz and Hymes1972) classic study, speakers can manipulate their speech patterns (e.g., in terms of vowel quality) to project a certain identity, distinguishing themselves from members of another group. In this sense, the speech patterns of individual speakers or groups, which are commonly described as accent and refer to dimensions associated with linguistic attributes of spoken language (e.g., prosody, segmental accuracy), appear to be some of the most salient markers of speakers’ ethnolinguistic belonging (Edwards, Reference Edwards2009; Lippi-Green, Reference Lippi-Green2012; Sachdev et al., Reference Sachdev, Giles, Pauwels, Ritchie and Bhatia2012). For instance, language users, including bilinguals and multilinguals, exploit variations in speech patterns to project various degrees of convergence or divergence towards their interlocutors (e.g., Boberg, Reference Boberg2004; Gordon, Reference Gordon2000; Schilling-Estes, Reference Schilling-Estes2004).
The special status of accent as a salient marker of ethnic identity is unsurprising for several reasons. Preference for “familiar” accents emerges in the first year of life (Kinzler, Dupoux, & Spelke, Reference Kinzler, Dupoux and Spelke2007). A nonnative accent is a powerful cue guiding children's and adults’ perceptions and actions, such as making an evaluative judgment or choosing friends (Girard, Floccia, & Goslin, Reference Girard, Floccia and Goslin2008; Kinzler, Shutts, DeJesus, & Spelke, Reference Kinzler, Shutts, DeJesus and Spelke2009; Rakić, Steffens, & Mummendey, Reference Rakić, Steffens and Mummendey2011). And accent is linked to various types of stigma and profiling (Baugh, Reference Baugh1999; Gluszek & Dovidio, Reference Gluszek and Dovidio2010a). Besides accent, other linguistic markers of identity include the use of vocabulary, code-mixing, and code-switching, as well as the choice of one language or dialect variety over another, all linked to speakers’ expression of identity through language (e.g., Bailey, Reference Bailey2000; Lo, Reference Lo1999; Rajadurai, Reference Rajadurai2007; Rampton, Reference Rampton2005; Reyes, Reference Reyes2005).
The use of language, including the L2, to project a certain kind of ethnic identity is not at all a simple matter, both in terms of the linguistic repertoires employed and the identities construed. For instance, the Italian Canadians studied by Giampapa (Reference Giampapa2001) showed versatility in their language choice, their code-mixing of dialects, and their use of specific lexical expressions to project different identities in various settings. Studying Polish immigrants in the United Kingdom, Temple (Reference Temple2010) showed how these speakers constructed multiple, complementary identities in various life domains (home, work, education) and projected different ethnic selves through their use of Polish and English with various interlocutors (family members, friends, coworkers). In fact, complex, hybrid identities appear to be the hallmark of bilingual and multilingual language use, with such identities often constructed as a way of coping with adverse circumstances, such as discrimination experiences or immigration and settlement struggles (e.g., Oropesa, Landale, & Greif, Reference Oropesa, Landale and Greif2008).
Ethnic identities revealed through language appear to be mutually constructed by all interactants and embedded in the particular contexts of language use. For example, in Miller's (Reference Miller1999) study of three migrant children learning English in Australia, the children constructed complex ethnic and personal selves that were shaped by their immediate environment. These identity expressions ranged from espousing a common “we don't speak English” identity in a migrant reception center to being marginalized as nonnative speakers relegated to talking only to other speakers of the same ethnic tongue in initial years of high school. The identity–language link is also often characterized by human agency, in that projecting certain identities through language may gain speakers access to—or instead prevent them from using—the needed resources to advance their goals. For example, Cervatiuc (Reference Cervatiuc2009) studied 20 academically and professionally successful immigrants to Canada from a variety of ethnic backgrounds. She identified several strategies pursued by her participants, such as seeking and gaining access to social interaction with members of the host community and other successful immigrants. Most notably, successful L2 users created a hybrid identity that embraced both their ethnic and target cultures, suggesting that successful learning experiences might be tied to the adoption of multilingual, multicultural identities (see Norton & McKinney, Reference Norton, McKinney and Atkinson2011).
ETHNIC IDENTITY AND L2 LEARNING: EXPLORING THE RELATIONSHIP
Compared to a rich body of literature documenting identity–language links, there is relatively little research documenting how speakers’ identification with their own ethnic group and with the target community might relate to L2 development, with most research relying on learners’ self-ratings of their L2 ability collected at a single time. At least hypothetically, in line with Lambert's additive bilingualism or integration pattern, language users might embrace L2 learning to become bilingual and bicultural. In contrast, illustrating Lambert's subtractive bilingualism or assimilation pattern, language users may acquire the language of a majority group and lose their own language and culture. Or language users may refrain (overtly or covertly) from acquiring an L2, especially if they experience threat to the survival of their ethnic group. Alternatively, language users’ sense of ethnic belonging to a particular group might have little bearing on the rate and ultimate success of their L2 learning.
Positive learning outcomes
One common finding across several learning contexts is that L2 users can successfully maintain the ethnic identity of their primary (home) group while acquiring an L2, suggesting that they can adopt a new cultural identity as part of L2 learning. For example, Henning-Lindblom and Liebkind (Reference Henning-Lindblom and Liebkind2007) surveyed 291 Swedish-speaking youths in Finland and found that better Finnish (L2) skills and more extensive Finnish interaction networks were linked to the speakers’ increased identification with the target L2 group. Most importantly, speaking Finnish as the L2 did not subtract from the speakers’ Swedish identity; in fact, the youths appeared to hold multiple simultaneous identities of bilinguals, Swedish speakers, and Finnish nationals without major adverse effects of these identities on each other (see also Noels & Clément, Reference Noels, Pon and Clément1996; Noels, Pon, & Clément, Reference Noels, Pon and Clément1996). In Israel, Ellinger (Reference Ellinger2000) showed a positive association between Russian and Hebrew learners’ strength of ethnolinguistic identification and L2 (English) achievement levels. In essence, a stronger sense of identification with the home group was associated with higher L2 proficiency scores. A similar positive association was reported by Coupland, Bishop, Williams, Evans, and Garrett (Reference Coupland, Bishop, Williams, Evans and Garrett2005) for Welsh high school students in Wales.
One possibility is that L2 learning is most efficient when speakers espouse a double-positive orientation, that is, a favorable view of their own ethnic group and of the L2 community. For example, Gatbonton and Trofimovich (Reference Gatbonton and Trofimovich2008) showed that those native French speakers of L2 English in Quebec who expressed willingness to be identified as both Canadian and French Canadian were the speakers with the highest self-rated L2 ability scores. In a recent study of Kurdish learners of Turkish in Turkey, using listener-rated measures of L2 accent, Polat and Schallert (Reference Polat and Schallert2013) reported that the most successful learners included either those who demonstrated strong identification with their home and L2 groups or those who strongly affiliated themselves with the L2 group. The disadvantaged group included the speakers who were singly oriented towards their home ethnic group.
These findings thus imply that a double-positive orientation towards the home and the target groups is a helpful but not a required condition for learners to show strengths at least in some aspects of the L2. For example, in the context of learning a heritage language (Korean) by Korean Americans in the United States, a stronger sense of orientation towards the Korean culture was associated with stronger Korean competence (Kang & Kim, Reference Kang and Kim2012). And for ethnically Russian adolescents in Germany, shifting language use patterns from Russian to L2 German was tied to holding positive feelings towards the L2 community, while identifying less strongly with the home culture (Michel, Titzmann, & Silbereisen, Reference Michel, Titzmann and Silbereisen2012). Therefore, what seems to matter for L2 learning and use, at least in some contexts, is a positive outlook towards the target L2 group.
Negative learning outcomes
In some cases, issues of ethnic identity seem to contribute negatively to L2 learning, implying that language users may not achieve the level of L2 performance that might otherwise be possible. One example comes from the study by Paladino et al. (Reference Paladino, Poddesu, Rauzi, Vaes, Cadinu and Forer2009), who explored the effect of stereotype threat, which refers to a negative judgment about a social group, on Italian-dominant speakers in an Italian-German bilingual community in Italy. Italian speakers underperformed in written and oral tests of German (their L2) after a negative stereotypical judgment about the German language abilities of their ethnic group was voiced explicitly and even when such a judgment was only implied. This suggests that perceived negative attitudes about one's social group, especially when they concern language, may negatively impact L2 performance. This result is also in line with an early finding that, for native French speakers of L2 English in Quebec, a higher degree of perceived threat to the survival of their home (French) group was associated with lower L2 (English) proficiency (Taylor, Meynard, & Rhéault, Reference Taylor, Meynard, Rhéault and Giles1977). In essence, lower-than-expected L2 performance might not only be linked to perceived threat to a group's vitality but also be tied to perceived negative attitudes.
Further evidence that perceived negative attitudes about L2 speakers’ ethnic groups might negatively impact L2 learning comes from a series of recent studies by Gluszek and colleagues. In a U.S. survey of native and nonnative speakers, Gluszek and Dovidio (Reference Gluszek and Dovidio2010b) showed that, compared to speakers of regional and standard dialects, foreign-accented speakers demonstrated greater problems in communication and greater perception of stigmatization from the L2 community. Most importantly, compared to other groups, foreign-accented speakers felt less affiliated with the United States, with stronger accents linked to the feeling of not belonging to the United States. In a follow-up study, Gluszek, Newheiser, and Dovidio (Reference Gluszek, Newheiser and Dovidio2011) employed causal modeling to show that, for nonnative speakers of English, identification with the American culture was tied to perceived L2 accent strength, such that weaker identification with the L2 group was associated with more accented L2 speech (see also Moyer, Reference Moyer2007). Perceived negative attitudes or stigmatization from the L2 community might thus contribute to language users’ increased sense of being an outsider as well as to their less-than-ideal L2 performance.
One aspect of the ethnic identity construct that appears most amenable to negative associations with L2 learning is speakers’ sociopolitical views concerning their identity. For example, for native French speakers of English in Quebec, where French and English are respectively majority and minority languages, speakers with the strongest sovereigntist views (e.g., Quebec becoming an independent nation) were judged as being most accented as well as least comprehensible, fluent, and proficient overall in their L2 (Gatbonton & Trofimovich, Reference Gatbonton and Trofimovich2008). Speakers with the strongest sovereigntist views were also those who performed poorly in terms of overall pronunciation accuracy in producing a particular segment (English /ð/) and demonstrated the lowest level of phonological development attained with this target (Gatbonton, Trofimovich, & Segalowitz, Reference Gatbonton, Trofimovich and Segalowitz2011). A similar negative relationship emerged for ethnic Latvians speaking L2 Russian in Latvia, where Latvian is the majority and Russian is a minority language. Ethnic Latvian speakers with a stronger sense of ethnic identity and stronger political views towards the Latvian language were also those with lower self-rated L2 (Russian) ability (Trofimovich, Turuševa, & Gatbonton, Reference Trofimovich, Turuševa and Gatbonton2013). As argued previously, at least in part, these negative associations might be attributed to perceived threat to the vitality of the speakers’ home group. Additionally, French speakers of L2 English in Quebec appear to link L2 proficiency with loss of loyalty to the home group (Gatbonton, Trofimovich, & Magid, Reference Gatbonton, Trofimovich and Magid2005), so fear of being labeled a traitor to the ethnic group might deter some speakers from attaining the level of L2 skill that would otherwise be possible.
Apart from linguistic consequences for L2 performance, negative associations of ethnic identity with L2 learning might lead to certain behavioral choices. For instance, in a study of Iranian immigrants in the United States, language preference, which correlates both with acculturation strength (Noels et al., Reference Noels, Pon and Clément1996) and language proficiency (Brenneman, Moris, & Israelian, Reference Brenneman, Moris and Israelian2007), was found to be associated with immigrants’ desire to either stay or leave the United States. In essence, immigrants who preferred to read study materials in their ethnic language (Persian) were more likely to express a desire to repatriate to their country of birth, compared to those who chose the English version of the materials (Hojat, Foroughi, Mahmoudi, & Holakouee, Reference Hojat, Foroughi, Mahmoudi and Holakouee2010), suggesting that the identity–language link might be tied to behaviors with rather dramatic consequences for language learning and use. Yet ethnic beliefs need not always be in line with behavioral choices. For instance, even though Chinese speakers of L2 English in Quebec ascribed less loyalty to peers speaking better English, compared to those who spoke English more poorly, these speakers would still choose nonaccented peers as leaders for events in mono- and bi-ethnic situations (Gatbonton et al., Reference Gatbonton, Trofimovich and Magid2005). Thus, speakers seem to balance the costs of maintaining loyalty to their group with the rewards associated with nativelike L2 ability (see also Goldberg & Noels, Reference Goldberg and Noels2006).
No association
It might come as no surprise that ethnic identity might bear little relationship with L2 performance, at least in certain contexts and for some L2 speakers. For example, Diehl and Schnell (Reference Diehl and Schnell2006) analyzed extensive self-report data from first- and second-generation immigrants to Germany over a 17-year period. Across time, most immigrants showed an increase in their self-identification with the German culture (feeling totally German), along with a marked decrease in their identification with the home group (feeling totally like a member of the country of origin), especially in the second generation. However, immigrants’ L2 skills appeared to “stagnate” over time, meaning that first- and second-generation speakers estimated their German ability (speaking German well) as high in the 1980s as they did in early 2000s, although again second-generation speakers rated themselves higher than their parents’ generation. Although these data are not longitudinal, they nevertheless imply that positive self-identification with the L2 group need not go hand-in-hand with improved L2 skills.
Besides, not all elements of the ethnic identity construct should matter to L2 development, or should matter to the same degree. For example, the same French speakers of English who showed negative links between the strength of their sovereigntist views and various measures of L2 ability demonstrated no associations between measures of L2 ability and other aspects of their ethnic identity, such as pride in knowing the group's history, in displaying its symbols and learning its language, and in willingness to defend the group (Gatbonton & Trofimovich, Reference Gatbonton and Trofimovich2008; Gatbonton et al., Reference Gatbonton, Trofimovich and Segalowitz2011). This suggests that speakers might weigh certain aspects of ethnic identity more highly than others. In the context of Latvia, unlike ethnic Latvians, Russian speakers showed no negative associations between L2 Latvian ability and their ethnic identity as Russian speakers, with the consequence that a stronger sense of ethnic identification through language was unrelated to their L2 (Latvian) ability (Trofimovich et al., Reference Trofimovich, Turuševa and Gatbonton2013). As a linguistic and political minority in Latvia, ethnic Russians may have preferred not to associate their strong ethnic beliefs with the ability to speak L2 Latvian, perhaps both to maintain a strong ethnic identity and to gain access to the socioeconomic benefits associated with speaking the country's official language. These results highlight the situation-specific nature of ethnic identity, implying that different contexts might be characterized by a separate set of ethnic identity dimensions and that even the same context might not engage all possible dimensions of ethnic identity for every speaker or group in every communicative transaction.
In concluding this section, it needs to be emphasized that associations between ethnic identity and L2 learning are likely much more complex than portrayed here, as identity–language links appear to depend on generational differences (e.g., Noels, Leavitt, & Clément, Reference Noels, Leavitt and Clément2010), speakers’ gender (e.g., Polat & Mahalingappa, Reference Polat and Mahalingappa2010), and other situational and contextual factors, such as speakers’ country of origin, language distance, political climate in the home and host societies, and prior language learning histories (e.g., McCrone & Bechhofer, Reference McCrone and Bechhofer2008; van Tubergen & Kalmijn, Reference van Tubergen and Kalmijn2005; Vedder & Virta, Reference Vedder and Virta2005). These associations might also vary as a function of research instruments used to assess them, for example, depending on the language of the interview or the questionnaire employed (Harzing, Reference Harzing2006), or the type of instrument used, such as explicit tests or implicit measures (Pantos & Perkins, Reference Pantos and Perkins2012). In sum, the precise contribution of various contextual factors to the identity–language link needs to be clarified in future research.
IDENTITY–LANGUAGE LINK: RECIPROCAL AND SELF-REINFORCING?
The discussion of the identity–language links in L2 learning thus far has shown that beliefs about ethnic identity might be linked to either positive or negative aspects of L2 performance, or might reveal no such relationships altogether. However, underlying these relationships was the assumption that they operate in a particular direction, namely, from identity expression to L2 learning and use. Because most social-psychological research on identity is based on correlations, with ethnographic or observational identity research similarly implying associations (e.g., Yip, Reference Yip2005), it is only reasonable to hypothesize that the identity–language links might function in the opposite direction as well, from L2 learning and use to one's identity expression.
It is as yet unclear how L2 learning might shape the strength or the breadth of one's ethnic identity beliefs. For example, Ersanilli and Koopmans (Reference Ersanilli and Koopmans2010) compared Turkish immigrants to France, Germany, and the Netherlands using several cultural integration and linguistic measures collected through a telephone survey. They found that Turkish immigrants in France and Germany, where several linguistic and cultural preconditions to naturalization exist, showed greater identification with the target group and demonstrated higher L2 proficiency and more extensive L2 use, compared to Turkish immigrants in the Netherlands, where (until recently) no preconditions for naturalization existed. In essence, these data might support the idea that higher L2 proficiency (which is required for obtaining citizenship in France and Germany) might be linked to increased identification with the L2 group, such that improving L2 skills also enhances one's sense of belonging to the L2 community (see also Ersanilli & Koopmans, Reference Ersanilli and Koopmans2011; Henning-Lindblom & Liebkind, Reference Henning-Lindblom and Liebkind2007).
In a recent analysis of large-scale survey data targeting first-generation immigrants in Germany, Hochman and Davidov (Reference Hochman and Davidov2014) tested both hypotheses simultaneously, namely, that immigrants’ linguistic integration influences the strength of their identification with the German group and that the strength of ethnic integration has a backward effect on immigrants’ German proficiency. These researchers obtained only one significant association, with German language proficiency influencing the strength of immigrants’ identification with the German group. The reverse effect, whereby ethnic identification influences L2 proficiency, was weak mainly because immigrants’ L2 skills improved little over time. It appears, then, that a certain threshold of L2 skills might be needed for speakers to feel and subsequently enrich a sense of belonging to the target group. Yet a greater sense of belonging might not always translate into higher L2 proficiency, likely because of difficulties of the immigrant experience, in which extensive language learning often takes a back seat to more pressing obligations, such as securing employment and caring for children (Temple, Reference Temple2010).
At least one way to conceptualize the identity–language relationship might be to assume that they could be described in terms of reciprocal causation, meaning that potential links are running in both directions. In psychological research, reciprocal causation could be described in terms of Matthew effects, after a Biblical quote from Matthew: “For everyone who has will be given more, and he will have an abundance. Whoever does not have, even what he has will be taken from him” (Matthew 25:29). The Matthew effect is defined as “the amplification of any initial advantage (e.g., economic resources, health status, cognitive ability) that leads to cumulative differences that widen preexisting gaps” (Ceci & Papierno, Reference Ceci and Papierno2005, p. 149). Put differently, an early success with a particular skill gives an individual an advantage later, compared to individuals who do not have an advantageous experience with this skill early on.
It is possible that the relationship between language users’ ethnic identity and their L2 development (at least ideally) also involves self-reinforcing bidirectional processes described by Matthew effects. Those speakers who reach a certain threshold of L2 skills are often those who readily avail themselves of different opportunities to use the L2 because understanding and producing language cause them little difficulty. Opportunities for L2 use, in turn, provide these speakers with ample experiences for cross-cultural contact, leading to appreciation of members of the target ethnic group and its culture and to the sense of inclusion in L2 cultural networks, which fosters language development further through opportunities of authentic language use (see Clément Reference Clément, Giles, Robinson and Smith1980; Clément, Baker, & MacIntyre, Reference Clément, Baker and MacIntyre2003). At least some aspects of this positive learning–identity loop have been supported by Rubenfeld, Clément, Lussier, Lebrun, and Auger (Reference Rubenfeld, Clément, Lussier, Lebrun and Auger2006) in the context of French-English bilingualism in Canada. In contrast, other speakers for whom L2 skills pose difficulty will avoid opportunities to use their L2 precisely because listening and speaking prove to be a laborious and grueling task. These speakers might overtly or covertly project negative emotions onto the L2 community, particularly when their home group's identity is threatened, ascribing negative feelings to the L2 group's members, which in turn would deter them from further contact with the group and from further opportunities for language development.
Whether or not relationships between ethnic identity and L2 learning fit with Matthew effects, one factor that seems to underpin these associations is language use, such that the link between identity and L2 learning might be instantiated through experience with language. There is indeed evidence that various aspects of the language use factor might mediate L2 learning–identity links. For example, for native French speakers of L2 English in Quebec, the associations between ethnic identity factors and various measures of L2 performance vanished after the speakers’ self-reported amount of L2 use was partialled out, suggesting that the identity–language link was mediated through L2 use (Gatbonton & Trofimovich, Reference Gatbonton and Trofimovich2008; Gatbonton et al., Reference Gatbonton, Trofimovich and Segalowitz2011). Moreover, several factors appear to contribute to speakers’ sense of belonging to the L2 community, such as frequency and amount of contact with the L2 group (Clément et al., Reference Clément, Baker and MacIntyre2003), breadth of L2 social networks (Michel et al., Reference Michel, Titzmann and Silbereisen2012), amount of access to L2 written and audiovisual media (Clément, Baker, Josephson, & Noels, Reference Clément, Baker, Josephson and Noels2005), and duration of residence in the L2 community (Walters, Phythian, & Anisef, Reference Walters, Phythian and Anisef2007). Perceived communication difficulties, which arguably represent one aspects of interactive language use, seem to mediate the link between speakers’ L2 accent and strength of their identification with the L2 society (Gluszek et al., Reference Gluszek, Newheiser and Dovidio2011).
ETHNIC IDENTITY FROM A SOCIOCOGNITIVE PERSPECTIVE
The preceding discussion—and especially our brief contextualization of the ethnic identity construct—makes it abundantly clear that identity is a complex socially constructed phenomenon. Therefore, one challenge for L2 researchers is to conceptualize ethnic identity within testable theoretical frameworks that ideally could intersect social and cognitive dimensions of L2 learning, linking both the learner (as a social and a cognitive agent) and the social context of learning to L2 learning outcomes. In this section, we describe several potential frameworks that might fit this requirement, while also acknowledging that the identity construct might require far greater contextualization than what is possible through a single theoretical lens.
One proposal that seems promising in this regard is the sociocognitive approach (Atkinson, Reference Atkinson and Atkinson2011). This approach assumes that language development is driven by a dynamic interaction between the mind, body, and world. In essence, people's cognitive states, such as person-specific individual variables and mental representations (i.e., the mind), are instantiated in overt behaviors, such as bodily actions, orientations, or emotions (i.e., the body), which are in turn fully embedded in particular social contexts (i.e., the world). Language is seen as an instrument of social action, as a flexible and adaptable tool of making a change in a given social environment (e.g., purchasing tickets or brokering ceasefire). Language development is conceptualized as a gradual, interactive adaptivity or alignment of the learner with a social and cognitive learning environment. For example, a speaker might align with an interlocutor within a given cross-cultural transaction in terms of the complexity of utterances, body gestures, voice volume, and rate of speech (Atkinson, Churchill, Nishino, & Okada, Reference Atkinson, Churchill, Nishino and Okada2007; Churchill, Okada, Nishino, & Atkinson, Reference Churchill, Okada, Nishino and Atkinson2010). This view of learning as social and cognitive alignment, which is compatible with both cognitive research on interactive alignment (Garrod & Pickering, Reference Garrod and Pickering2009) and social-psychological research on social accommodation (Giles, Reference Giles2012), appears to be promising for conceptualizing identity–language links. For instance, it may be possible to study how L2 speakers align with their interlocutors (or fail to do so) in how they position their ethnic identity in interaction and how such alignment (or lack thereof) might impact L2 learning and use (for examples of alignment research, see Trofimovich, Reference Trofimovich2013).
Another possible framework for conceptualizing identity–language links includes dynamic systems theory (de Bot, Lowie, & Verspoor, Reference de Bot, Lowie and Verspoor2007), which views language learning as an iterative (repetitive) process characterized by variability both within and across individuals. This process occurs on many time scales (e.g., within an interaction, across lessons, during semesters of coursework, throughout years of language experience) and features a number of developmental stages (called attractor states). Van Geert, Steenbeek, and van Dijk (Reference van Geert, Steenbeek, van Dijk, Schmid and Lowie2011) applied this theory to account for socially mediated L2 learning, thus encompassing both cognitive and social aspects of L2 development. In this model, language development occurs through the interaction between novices (typically, learners) and experts (e.g., teachers), with learning determined by the interplay between the situation-specific goals of the learner (e.g., the need to acquire certain knowledge, to exert less effort in learning, and to preserve aspects of own ethnic identity) and the goals of the teacher (e.g., the need to complete certain learning tasks, to motivate learners, and to satisfy requirements from the employer or the curriculum). Learning is thus conceptualized within this model as a continuous, dynamic adaptation of teacher and learner behaviors, which can both influence and be influenced by ethnic identity beliefs, with teachers adapting their actions to the perceived needs and sociocognitive states of learners.
Other theoretical views for conceptualizing L2 development and ethnic identity include the willingness-to-communicate framework, which embraces several cognitive, social, and experiential factors to explain learner desire to engage in L2 communication (MacIntyre, Clément, Dörnyei, & Noels, Reference MacIntyre, Clément, Dörnyei and Noels1998), as well as the social context model (Clément Reference Clément, Giles, Robinson and Smith1980), which focuses on intergroup contact targeting such factors as L2 confidence, competence, and identity. In a version of this model, frequency and quality of L2 contact are linked to L2 confidence, which is in turn related to both willingness to communicate and identity, which are both linked to frequency of L2 use (Clément et al., Reference Clément, Baker and MacIntyre2003). Segalowitz, Gatbonton, and Trofimovich (Reference Segalowitz, Gatbonton, Trofimovich, Dörnyei and Ushioda2009) proposed a conceptualization of L2 pronunciation learning that includes cognitive and social influences. In this view, ethnic identity is part of a larger motivation system that determines whether learners engage in L2 use (for details on motivation–identity links, see Csizér & Magid, Reference Csizér and Magid2014). Language use is important because it provides learners with opportunities to tune their perceptual and cognitive systems for the processing of L2 input. This cognitive and perceptual tuning is driven by psycholinguistic variables, which include frequency (i.e., how often a given pronunciation target occurs in an L2) and cross-language similarity (i.e., perceptual differences between learners’ two languages that determine the ease or difficulty of certain aspects of L2 pronunciation). In essence, ethnic identity and motivational variables shape particular patterns of L2 use, and use in turn impacts language learning outcomes by allowing learners to practice their cognitive processing skills through L2 input and/or output.
EXPANDING CURRENT RESEARCH AND MOVING FORWARD
Theoretical frameworks aside, research on ethnic identity has been and continues to be a productive area of applied linguistic research. With a view towards the future, it might be worthwhile to identify some research strands that seem particularly promising, especially in terms of their reach beyond language learning. One such strand involves the study of how ethnic and professional identities of language users intersect in various professional domains. For instance, Harrison (Reference Harrison2013) studied nonnative social workers in Australia, documenting how their ethnolinguistic identity was often perceived as a detriment to their credibility and depreciated their knowledge and skills. Harrison also problematized the incompatibility of the official “equality” and “diversity” discourses espoused by governments and businesses and individual language users’ sense of exclusion and marginalization, which further highlights that the identity–language links may have far-reaching, and often detrimental, social and economic consequences for language users (e.g., being passed over for promotion). In another study, Kokaliari, Catanzarite, and Berzoff (Reference Kokaliari, Catanzarite and Berzoff2013) examined psychotherapists’ perception of the role of language in treatment of bilingual patients, showing that patterns of language choice in therapy sessions were closely tied with expression of identity, both ethnic and personal, and were linked to strong affect expressions, death, and trauma. In the medical domain, a cancer patient's ethnicity (being originally from outside the English-speaking world) might even add to this patient's clinical distress during treatment (Thomas, Carlson, & Bultz, Reference Thomas, Carlson and Bultz2009), suggesting that ethnolinguistic discourse surrounding physical or psychological trauma related to illness, natural disasters, or military conflicts might be critical to people's well-being (see Higgins & Norton, Reference Higgins and Norton2010; Kaldor, Reference Kaldor2013).
Among other promising agendas for expanding ethnic identity research are investigations of identity–language links in cases of language revitalization or loss, targeting such socially and economically vulnerable populations as, for example, aboriginal or Native American speakers in North America (e.g., Peltier, Reference Peltier2010). Also seemingly underresearched are ethnic identities of classroom language learners and teachers, with the consequence that identity issues need to be considered alongside other social, cognitive, pedagogical, and linguistic factors in classroom L2 teaching (e.g., Ajayi, Reference Ajayi2011; Choi & Yi, Reference Choi and Yi2012). Additionally, sociolinguists might continue elaborating how ethnic identity shapes variability in linguistic patterns across various domains of language (e.g., Hall-Lew & Yaeger-Dror, Reference Hall-Lew and Yaeger-Dror2014). Researchers working within lingua-franca perspectives might examine how ethnic identity interacts with speakers’ construction of pan-linguistic and pan-cultural identities in lingua-franca contexts (e.g., MacKenzie, Reference MacKenzie2009), and scholars working within sociocultural frameworks might continue problematizing the links between ethnic identity and such notions as power, capital, and desire, as they relate to language learning and use (e.g., Chowdhury & Phan, Reference Chowdhury and Phan2014; Motha & Lin, Reference Motha and Lin2014).
Last but not least, a most promising (and perhaps a very rewarding) area of research appears to be related to the relationship between the expression of ethnic identity, language use, and discrimination and stigma (e.g., linguistic profiling). The social power of this research lies not only in its potential to help raise awareness of discrimination and stigma in various research, business, and professional communities (e.g., Collins & Clément, Reference Collins and Clément2012; Gluszek & Dovidio, Reference Gluszek and Dovidio2010a), but also in its value for suggesting possible ways of eradicating or minimizing discriminatory practices faced by bilingual and multilingual populations. Researchers have documented several interventions, such as perspective taking, which might reduce discriminatory judgment (Hansen, Rakić, & Steffens, Reference Hansen, Rakić and Steffens2014; Kang, Rubin, & Lindemann, Reference Kang, Rubin and Lindemann2014; Weyant, Reference Weyant2007).
We conclude this article targeting the relationship between ethnic identity and language by referring back to Mari Matsuda's (Reference Matsuda1991) impassioned plea against intolerance based on ethnic prejudices. Matsuda argued that “accent tolerance suggests a radically pluralistic re-visioning of . . . identity” (p. 1403). This view deeply resonates with many (if not all) of those who are involved in language research and practice communities because it once again emphasizes that matters of language learning, teaching, and use are defined through multiplicity of voices, identities, languages, and beliefs.
