Language shift among adolescent ethnic German immigrants: Predictors of increasing use of German over time

Abstract

This study examined language shift from speaking Russian to German longitudinally in a sample of 229 adolescent immigrants (M_age = 16.14, M_{length of residence} = 6.26, 67% female). Our aims were to test whether language shift can be found in adolescent first generation immigrants and to test whether variables indicative of (a) linguistic adaptability, i.e., the efficiency to learn, speak and use a second language, (b) the amount of contact with native Germans, and (c) the motivation to use German, predicted differences in adolescents’ levels and rates of change in German language use. Results showed an overall decelerated increase in German language use over time, which resembled a “learning curve” leveling off at an average “frequent” use of German. Differences between adolescents in language shift were mainly associated with variables indicative of interethnic contact and motivation, but not to linguistic adaptability. In more detail, speaking German increasingly in daily life was related to an increasing share of native peers, a decreasing self-identification as Russian and an increasing orientation towards natives. Language shift thus seems to result from an increasing sense of belonging to the receiving society.

Introduction

Language shift, which is defined as “the change from the habitual use of one language to that of another” (Weinreich, 1953, p. 68), is often regarded as a prerequisite for the integration of immigrants and as an indicator of adaptation (Remennick, 2004). The extent to which immigrants use their second language in daily life instead of their heritage language was found to be a predictor of immigrants’ adaptational outcomes such as self-esteem, earnings, or substance consumption and deviant behavior (Chiswick, 1998, Epstein et al., 2003, McQueen et al., 2003, Schnittker, 2002). However, surprisingly little is known about longitudinal change in second language use and its precursors (Clément, Baker, & MacIntyre, 2003). This secondary data analysis of a sample of adolescent ethnic German immigrants investigated change in second language use, or language shift, over four annual measurement occasions. We were especially interested in whether there is change in language use at all, which pattern language shift follows over time, and on which level this change takes place. Furthermore, we wanted to know which variables indicative of linguistic adaptability, contact, or motivation predict differences in language shift between adolescents.

We focused on the time period of adolescence as this period is not only formative for adult years, but also seems a very sensitive time for change in second language use (Tseng & Fuligni, 2000). Especially for adolescent immigrants language is not only a means of communication, but also of expressing group affiliations. Language use thus can be assumed to be related to achieving developmental tasks in the new context, such as finding new friends and defining one's identity (Caldas & Caron-Caldas, 2002). Further, our focus was on ethnic German immigrants (also called ‘Aussiedler’ or repatriates). With over 2.5 million immigrants since the breakup of the former Soviet Union (FSU) in 1989, ethnic Germans are the largest group of recent immigrants in Germany. Ethnic Germans are diaspora migrants to Germany from Eastern Europe and Russia, whose ancestors were invited centuries ago by authorities to settle in areas not yet reclaimed to help cultivate land or secure border lands (Dietz, 2006). Due to this history, they are regarded as Germans by the German law and receive full citizenship upon arrival as well as economic support. However, having lived for generations in territories of the FSU, and being separated from their German ties especially in times of the Cold War, they became more and more included in mainstream Russian culture. For example, they were prevented from speaking German in public in the 1950s and 1960s due to Soviet ideology. Thus, many lost fluency in German. Especially the young generation grew up speaking only Russian, because marriages between ethnic Germans and native Russians became common, which made it difficult to keep German as a family language (Dietz, 2006). For these reasons, although being of German ancestry, most ethnic German immigrants did not speak German when they entered Germany. Apart from their privileged status, ethnic German migrants face similar challenges as other migrant groups, such as problems with the transferability of their, often quite high, Russian educational qualifications (Konietzka & Kreyenfeld, 2001), their less favorable economic situation, and they often have only contact to other ethnic German immigrants from the FSU (Dietz, 1999a).

Language shift is usually assumed to be a process spanning decades or involving different generations. Veltman's (1983) classical concept of language shift as a process covering three generations suggests that first generation immigrants predominantly maintain their mother tongue, the second generation speaks both languages, and the third generation almost exclusively uses the second language. Although Veltman's (1983) concept assumes that change in second language use takes place between generations, there are theoretical and empirical arguments to assume that change also happens in a single migrant generation (Jia and Aaronson, 2003, Pease-Alvarez, 2002).

Culture learning and culture shedding theories (Berry, 2005, Masgoret and Ward, 2006) may help to explain change in language use within one generation, as they focus on explaining individuals’ behaviors over time in the new country. According to these models, changes in skills and behaviors follow a “learning curve” pattern, i.e., increase over time and level off as new culture-specific skills and behaviors are acquired (Ward, 1996, Ward, 2001). Older skills and behaviors that are not needed anymore are shed in favor of new skills and behaviors which allow individuals to better fit with the society of settlement (Berry, 2005). Although such a pattern is typically expected for skills, in the acculturation context it also seems to apply to behavioral changes which aim at increasing the fit between individual and society. For example, Titzmann and Silbereisen (2009) found that increases in the share of natives in adolescent immigrants’ social networks are more pronounced early after immigration than later on. Most likely, this is due to the fact that newcomer immigrants need to reach a level of sociocultural adaptation that allows them to participate both in the home and the host society to a degree they deem sufficient and after which only smaller changes can be observed (Masgoret & Ward, 2006).

We base our hypotheses on both these theoretical underpinnings. According to culture learning and shedding theories we expect that second language use increases over time but that changes are more pronounced earlier than later (“learning curve” or “decelerated increase”). Furthermore, we do not assume a full adaptation over time for the first generation adolescent immigrants examined in this study. As first generation immigrants, our participants should on average predominantly maintain their mother tongue, as is expected by the three generation model (Veltman, 1983). However, as young immigrants with average lengths of stay of several years, we expect language shift to level off at a frequent use of German.

Hypothesis 1

Increase in second language use as a behavior is expected to show decelerated increase, meaning that change is more pronounced earlier, eventually flattening out over time.

As there is no specific theory which predicts first generation immigrants’ language shift, we base our expectations on theories on second language acquisition but combine these with acculturation theories on immigrants’ sociocultural adaptation. There are several theories on second language acquisition, reaching from a micro-perspective of acquiring the grammar, vocabulary, pronunciation, and other functional aspects of a second language, typically covered by psycholinguists (cf. Mitchell & Myles, 2004), to a ethnolinguistic macro-perspective of group differences in language use (e.g., Landry & Allard, 1994). There are also several social psychological approaches to second language acquisition, however, they were only seldom applied to immigrants (Masgoret & Gardner, 1999). Probably the best known of these models is Gardner's (1985) socio-educational model, according to which “it is hypothesized that individual differences in ability (both intelligence and language aptitude) and motivation account for differences in how successful students are in acquiring a second language” (Gardner, 2006, p. 244). Although this model was designed for examining second language achievement in the classroom context, some aspects are also relevant for the more informal language acquisition of immigrants, especially the aspect that a second language learner's level of motivation is based on different reasons, among which the desire to learn a second language in order to meet and communicate with members of the other group is particularly important (Masgoret & Gardner, 1999).

MacIntyre and Charos (1996) examined second language use outside the classroom in a bilingual Canadian community, combining Gardner's socio-educational model with the pyramid model of willingness to communicate (MacIntyre, 2007, MacIntyre et al., 1998), a model that directly aims at second language use. According to this heuristic model, willingness to communicate, which is defined as the intention to speak given the opportunity, is influenced by various factors ranging from more distant ones, such as intergroup climate and the speaker's personality, to situated antecedents, such as the desire to communicate with a specific person in a particular situation. In their final path model, MacIntyre and Charos (1996) found that frequency of communication in the second language was not only predicted by motivation, as was expected from the socio-educational model, and willingness to communicate, but that there also were direct paths from perceived competence, and exposure to second language speakers. This parallels economic and sociological research on immigrants’ second language acquisition in terms of proficiency and use, which assumes that opportunities for contact with the new language and its speakers are as relevant as the efficiency to learn and speak a new language and the incentives to use it (e.g., Chiswick, 1998, van Tubergen and Kalmijn, 2005). Also acculturation theories, which directly aim to predict changes in skills and behaviors in immigrants, name three groups of predictors, which are, cultural knowledge, degree of contact, and positive inter-group attitudes (Berry, 2006, p. 53).

The aim of this study was not to test any of the aforementioned theories, but to use them to guide our research with regard to potential predictors of language shift among adolescent immigrants. Although in these different areas of research predictors received various labels and were conceptualized slightly differently, we noticed that as an organizational framework three groups of predictors can be derived which we will refer to as (a) linguistic adaptability, (b) contact, and (c) motivation.

Variables related to linguistic adaptability. We use the term linguistic adaptability to refer to variables related to the efficiency to learn and speak a new language, perceived competence, and cultural/linguistic knowledge. We examined perceived hassles with speaking the second language as one indicator of linguistic adaptability. Based on the willingness to communicate model, we assume that “the speaker's self-perception of competence is considered more relevant than objective measures of linguistic skills” (Clément, Noels, & MacIntyre, 2007, p. 61). The language problems a second language learner experiences should thus reflect his self-perception of competence. A second, rather indirect indicator for linguistic adaptability used in this study was parental education. The educational level is not only a common proxy for socioeconomic status but also a reflection of the capability to speak a second language (e.g., Hakuta & D’Andrea, 1992). Higher educational qualifications of parents relate to higher cognitive abilities (Neiss & Rowe, 2000), and are associated with a more stimulating home environment, and more learning activities for the child (Myrberg & Rosén, 2008). Thus, higher parental educational levels can accelerate the process of second language acquisition (Chiswick, 1998, Hakuta et al., 2000).

Variables related to contact. Contact refers to the amount and intensity of time spent interacting with speakers of the home and host society, which also influences the opportunities for speaking the second language. In this study, contact was examined on two different levels. First, by the ethnic composition of the neighborhood, which affects the possibilities of maintaining the first language (e.g., Chiswick, 1998, Stevens, 1992). The more the speakers of a first language live in close proximity to one another, the greater the likelihood of their not using the second language and creating an environment conducive to maintaining the first language (MacIntyre and Charos, 1996, Mesch, 2003). Another measure of contact concerned direct interactions between immigrants and natives. As natives usually do not speak the immigrants’ first language, immigrants need to use the second language in those encounters. In particular, using the second language with native friends should ultimately transfer to other situations as the use of the second language becomes habitual (Caldas & Caron-Caldas, 2002) especially in adolescence, when peers exert a major influence on behavior (Vandell, 2000).

Variables related to motivation. According to Gardner (2006), motivation is “a very complex phenomenon with many facets” (p. 242). In this study, we cannot assess all the cognitive, affective and conative characteristics that are associated with being highly motivated. Instead, we investigated three reasons that might lie behind a strong motivation for speaking the second language, especially focusing on integrativeness and instrumentality (Gardner, 2006). The educational aspirations of the adolescents are considered to be an instrumental motivation (Grenier, 1984, Mesch, 2003), as they provide an incentive to use the second language in order to gain social or economic rewards, such as attaining a higher social status and higher earnings. In contrast, integrative motivational variables are characterized by positive attitudes towards the second language and its speakers and a desire to integrate into the receiving society (Gardner & Lambert, 1972). A higher self-identification with the culture of origin is expected to be related to less second language use, as adolescents who identify strongly with their culture of origin have to struggle more with inconsistencies in behavior and identity when using the second language in daily life, because people generally strive to keep identity and behavior congruent (Berzonsky, 1997). Another indicator for integrativeness in this study was the cultural orientation towards the receiving society, because immigrants who have positive attitudes towards the new country and its inhabitants can be expected to adopt behaviors of the new culture more readily, such as using its language (Schumann, 1986).

Although earlier research already provided some evidence for the association of some of these predictors and language use, variables of linguistic adaptability, motivation and contact were thus far hardly studied in combination. Furthermore, longitudinal studies on this subject are still scarce. Especially longitudinal analyses can challenge models with competing predictors, as not only a temporary state is assessed, where associations of variables might occur for all kinds of reasons, but change within persons in predictors as well as in second language use (Davies, 1994). Applying a developmental perspective and analyzing longitudinal data of quite a heterogeneous sample of adolescent immigrants, who were in Germany for a few months to over ten years at the first measurement occasion, we were able to test whether these groups of variables relate to both the level at the first measurement occasion and subsequent rate of change within individuals over time. It might be that different predictors can become more or less relevant at different phases in the process of acculturation (Beiser et al., 2002, Titzmann et al., 2008). If a predictor variable was closely related to language shift over time, associations of both levels and rates of change should be found. If, however, a predictor variable was only related to language shift soon after migration, only associations of the levels at the first measurement occasion of our sample should be found. Accordingly, associations of only the rates of change would be indicative of the onset of a co-developmental process starting some years after migration. Furthermore, we wanted to examine whether each group of variables uniquely contributes to second language use. Knowing which groups of variables and, in more detail, which particular predictor variables, are connected to change in second language use will help to create parsimonious and tailored models for language shift.

Hypothesis 2

We expected that predictors of each of three groups of variables were related to immigrants’ second language use at the first measurement occasion (Hypothesis 2a). In more detail, immigrants were expected to use their second language more in daily life, if they have a higher linguistic adaptability, have more contact to natives, and have a higher motivation for second language use. Furthermore, change in these predictors was also expected to be related to change in second language use over time, more specifically, we expected that each group of predictors was related to change in language shift over time (Hypothesis 2b).