Who speaks what language to whom and when – rethinking language use in the context of European Schools

2.1 The European School

The European School was founded in 1957 as a result of a parental initiative in Luxembourg (Housen 2002). Today, the European School is primarily intended to cater to the schooling of the children of EU employees. Currently, there are 14 European Schools in 7 different countries, comprising approximately 25,000 students (Euresc 2015) of which a little over 600 are Swedish (Ottosson 2017). All European Schools follow the same curriculum, leading to a European Baccalaureate. In accordance with the language policy of the Council of Europe (Ecml 2015), the European School aims at maintaining the students’ L1 and strive towards academic literacy in at least two languages (Baetens Beardsmore 1993; De Mejía 2002). As part of The European School’s multilingual education two or more languages function as languages of instruction (Baetens Beardsmore 1995); English, French and German are offered as languages of instruction together with the students’ L1. The schools are divided into language sections (Euresc 2016), and depending on the school, Swedish is taught as a subject or used as a language of instruction. In the latter case, mathematics, natural sciences and philosophy are taught in Swedish. In total, roughly 16 lessons per week are taught in Swedish.

2.2 TCK and elite bilinguals

The students in European Schools, like students in international education in general, are often characterized as globally mobile children (e.g. Baetens Beardsmore 1979; Hayden 2012). During the last decades, the term Third Culture Kids (TCK) has become a common and widely used term for the group of children and teenagers growing up in the international community. The term itself (TCK) refers to the belief that TCK grow up neither in their parents’ home culture (first culture), nor in the culture of the host country (second culture) but in the international society’s conglomerate of cultures, the “third culture” (Pollock and Van Reken 2009).

In addition, TCK do not only grow up in multicultural international environments, but also in the presence of several languages. It is not unusual that TCK in general, and especially TCK with a non-English-speaking background, often are required to relate to three languages in their everyday lives. An example of such a circumstance would be a student who is part of international education with English as language of instruction in a country in southern Europe and whose parents are native speakers of Swedish and Japanese.

As mentioned earlier, it is not unusual that students in international education, i.e. TCK, serve as an illustration of elite bilinguals. The term elite bilinguals often applies to voluntary migrants who have become multilingual by their own free choice (Skutnabb-Kangas 1981; Romaine 1995; Boyd 1998) and who are well-educated and proficient in at least one of the languages spoken by the international community (De Mejía 2002). Elite bilingualism (Paulston 1978) tends to be seen rather as a status-driven investment (Butler and Hakuta 2004) towards future access to the global market of work and higher studies, rather than a life sustaining necessity. As de Meija puts it, “elite bilingualism may be seen as a valuable personal possession which is consciously chosen and worked for” (2002: 41). Baker and Prys-Jones go as far as to argue that elite bilinguals “have continuously accepted that bilingualism causes no problems in thinking, academic achievement or cultural acceptance” (1998: 16). However, contradictory to Baker and Prys-Jones’ claim, my previous research shows that “the alleged free choice and privileged social class of TCK, and elite bilinguals, do not result in a multilingualism more superfluous and less complicated than that of other multilingual teenage groups” (Rydenvald 2015).

3.1 Multilingualism and related terminological issues

In the wake of the increased mobility of globalization, the understanding of multilingualism has changed. Aronin and Singleton claim that “[b]ecause multilingualism and globalization are so inextricably intertwined, all the major attributes of the globalization phenomenon characterize multilingualism as well” (Aronin and Singleton 2012: 56–57). One of the keywords of globalization is mobility, which also is found in contemporary sociolinguistics together with keywords such as dynamics, interplay, complexity and unpredictability (Blommaert 2012: 12). In line with the changing perspective, multilingualism has gone from being perceived as a parallel, dualistic (cf. Matarese 2013: 294) system of separate codes which have additive or subtractive effects on each other (Lambert 1977), to being perceived as a dynamic process. With the term dynamic multilingualism, García (2009) aims at describing how multilingualism is developed through an interplay between the individual’s linguistic resources and competences as well as the social and linguistic contexts she/he is a part of.

In the earlier dualistic perspective on bilingualism and multilingualism there was a certain focus on proficiency in defining the concepts. However, in contemporary research on multilingualism there is a tendency to highlight language use as one of the defining criteria, as the following quote clearly illustrates:

‘Multilingualism’ is here taken to be the use by individuals, groups, organizations or countries in more than one language in everyday life; this includes second, third or multiple languages. ‘Bilingualism’ is a subcategory thereof and is taken to be any real-life use of more than one language rather than the equally proficient use of more than one language.

(Radein Initiative Aalberse et al. 2011).

In this article, I use the term multilingualism as an umbrella term for bilingualism and multilingualism. Secondly, I define multilingualism in accordance with the usage based perspective presented in the quote above and define multilingual teenagers as teenagers who in their daily lives encounter and need to use two or more languages regardless of their proficiency in them (cf. Axelsson et al. 2005: 8). In this article, I will touch upon proficiency on a couple of occasions, but otherwise I will focus on the usage of language.

Generally, multilingual individuals have at least one L1 and one L2. Both terms, together with the related term native speaker are difficult, and have been criticized by many scholars (e.g. Coulmas 1981; Rampton 1990; Firth and Wagner 1997; Fraurud and Boyd 2011). Rampton (1990) for instance, suggests a replacement of the terms native speaker and mother tongue, in order to make transparent the connotations of various belongings, e.g. ethnic, national and linguistic, attached to the terms. He instead proposes the terms language expertise, language affiliation and language inheritance, where “language affiliation refers to the attachment or identification they [ESL students] feel for a language whether or not they nominally belong to the social group customarily associated with it” (Leung et al. 1997: 555).

The term L1 is used in this article to denote the first language/languages learned by the participants from their parents. The term mother tongue is not used, in order to avoid the ideological issues attached to the term. Neither is the term first language used, since it within the scientific field of international education also is used to denote the language the individual has the highest proficiency in (De Mejía 2002). It is not unusual for multilingual individuals who have done their schooling in another language than their L1 to become more proficient in the language of instruction than in their L1.

The term L2 is often used to denote the languages the individual has learned after the L1/L1s has/have been established. Moreover, the term L2 often implies that the language in question is a majority language in the society where the individual lives. The individual learns the language because she/he needs it in daily life, as opposed to, e.g. foreign languages learned in school context (Hyltenstam 2004: 37). In today’s globalized world and especially with the increased use of English it is not entirely unproblematic that L2 is used as a cover term for L3, L4 etc. (cf. Cenoz and Jessner 2000), neither are the borders between L2 and foreign languages clear cut and uncomplicated (Jessner 2008: 271), which is clearly illustrated by referring to English as a lingua franca in a general sense (cf. House 2003). The terminology simply does not seem to be able to fairly describe the life situation of multilinguals in today’s globalized world. Hammarberg (2010: 102) writes that the terms become “misnomers, suggesting a too narrow conception of the notions they stand for”. The participant group could be used as an illustration of Hammarberg’s claim, since in many cases their languages do not fit these categories.

3.2 Domain theory

Within the vast research fields of language use and language variation early theories tended to aim at “ordering the diversity” (Hymes 1972: 71) of human interaction by explaining its nature and delineating underlying premises, e.g. Hymes (1972) concept of speech events, summarized in the acronym taxonomy SPEAKING, speech accommodation theory (SAT) (Giles et al. 1987: 14) and the audience design theory (Bell 1984). An early theory that describes bilingual language use through a model which organizes, categorizes and operationalizes human interactional behavior is Fishman’s domain theory (Fishman 1965; Fishman 1971; Fishman 1972a; Fishman 1972b). One contribution put forward by the domain theory is the view of bilingualism as a stable condition rather than a temporary transit period between monolingualism in the native language in the one end, and the target language at the other end (Fishman 1965). In his theoretic model, Fishman refers to the five essential domains of family, friendship, religion, education, and work. These domains are “commonly associated with a particular variety or language” (Fishman 1972a: 44). He defines them in terms of place, subject and the role-relations of the participants (Fishman 1972a: 37). Domains are often referred to as either formal, e.g. religion and work, or informal, e.g. family and friendship. In multilingual settings the majority language is often seen to be related to the language use in more formal domains and the minority language to the more informal ones (Hyltenstam and Stroud 1991: 47). Fishman also argued that in a multilingual context one language often becomes dominant in a domain, where he equaled dominant language with most frequently used language (García et al. 2006: 12).

Domain theory applies a macro-perspective on language use, also pointed out by Fishman: “Domains are particularly useful constructs for the macro-level (i.e., community-wide) functional description of societally patterned variation in ‘talk’ within large and complex diglossic speech communities” (Fishman 1972b: 44). The macro-level perspective implies a top-down function which describes norms of language behavior, rather than allowing a bottom-up perspective which examines how multilinguals actually use their languages. Seen through the lens of the macro-perspective the domains appear as rather static, i.e. one single language is palpably dominating each domain and the boundaries between the domains are clear and firm.

However, the language use of multilinguals has proved to be more dynamic than the domain theory suggests. Not only do multilingual individuals use different languages in domains, but they also use different languages in the same domain. The premises for the participation in and belonging to different domains are likely to change during a lifetime (e.g. García 2009; Grosjean 2001; Skutnabb-Kangas 1981). The borders between the domains are thus blurred and the domains overlap. For instance, the school domain does not only relate to education but also social life between peers, resulting in the domains of school and friendship domains to be, to a large extent, interwoven. Furthermore, in the super-diversity (Vertovec 2007) that characterizes the globalized society the interpretation of a domain is liable to vary, e.g. the perception and the scope of the family domain. Another aspect that concerns contemporary society are the multitudes of networks an individual participates in, and the fact that networks can stretch over domains, e.g. religion could be practiced at home with friends. The IT-revolution over the last couple of decades in general and, in particular, the explosion of social media have also blurred the boundaries between official and intimate domains.

Despite the fact that domain theory appears to be out of tune with a society characterized by mobility as well as contemporary research on multilingualism, “[d]omains continue to be used as an organizing principle” (Bell 2014: 134). As an instrument for categorization on the macro-level domains may serve a function. However, domain theory is not a sufficient model for describing the complexity of the multilingualism and language use of the participants in this study, and, I argue, nor multilingualism in general in today’s globalized world.

4.1 Nexus analysis

Nexus analysis engages exploration of “social action at the intersection of three main elements” (Scollon and Scollon 2004: 19). The three elements consist of discourses in place, interaction order and the historical body. Central to nexus analysis is that nothing happens in a social vacuum, and therefore everything is linked to earlier actions and will result in new actions. Social actionScollon and Scollon (2004: 12) see as “any action taken by an individual with reference to a social network”. The social action in this study is the language choice taken by the participants in the networks linked to school and home.

“All places in the world are complex aggregates (or nexus) of many discourses which circulate through them” (Scollon and Scollon 2004: 14). In nexus analysis the element discourses in place refers to all discourses that are relevant for the social action, and not only spoken discourses but those in the form of e.g. architectural styles or different material (Scollon and Scollon 2004: 14) as well. In this study I will use the element of discourses in place to denote the participants’ thoughts about the reasons for their language choice (cf. Palviainen and Boyd 2013) as they describe them in the interviews.

The element interaction order, originating from Goffman (1983), is in nexus analysis (Scollon and Scollon 2004) used:

to talk about any of the many possible social arrangements by which we form relationships in social interaction. Our interest in the interaction order grows out of the fact that people behave differently depending in part whether they are alone when they act or if they are acting together in consort with other people as they might in having a conversation between friends, taking a university class, or consulting with a lawyer or medical doctor.

(Scollon and Scollon 2004: 13)

Metaphorically, interaction order could be described as a genre where certain interactional behavior is expected in the social activities we engage in. In this study I am interested in the language behavior of the multilingual participants, i.e. what languages they use in the certain social arrangements. Interaction order refers here to the social activities between peers in school as well as on their way to and from school, and social activities taking place outside school involving family or friends. In this study interaction order refers to the patterns of language in the self-recordings and in the questionnaire.

The third element is the historical body, which refers to the habits, experiences, beliefs and attitudes accumulated by an individual (Palviainen and Boyd 2013; Scollon and Scollon 2004) during the course of her/his life time. The concept of historical body concerns the issue of how the “participants all come to be placed at this moment and in this way to enable or to carry out this action” (Scollon and Scollon 2004: 160). Thus, the historical body does not simply describe the individual’s background, but also incorporates the experiences mounted and embodied by it. The reported results from the survey (Rydenvald 2015, 2016) and the interviews performed in this article describe the participants’ historical bodies.

4.2 Data and participants

Another advantage granted by nexus analysis is that both reported quantitative data and qualitative data from face-to-face interaction can be incorporated in the same analysis. Within MDA, triangulation is preferably accomplished with data from the following four sources: (1) participants’ generalization, e.g. interviews or surveys where the participants makes statement about their actions; (2) “neutral” observations, e.g. recordings of actions; (3) individual participants’ experiences, e.g. life histories from interviews; and (4) observer’s interaction with participants, e.g. discussions with participants about results from earlier data collection (Scollon 2001; Blåsjö 2013). As will become evident from the discussion below, data from these four sources constitute the material in the present study. (1) In the survey the participants have answered questions about language use, attitudes and preference in regard to different domains. In the interviews the participants also discuss their actions in the recordings. (2) The participants have recorded their own language use in different social settings. (3) In the interviews the participants describe, for instance, how their multilingualism has changed over their lives. The background questions in the survey also constitute a form a life history concerning language use, migration and schooling. (4) During the years of the data collection, I have had an ongoing dialogical interaction with the participants, presenting and discussion results.

The data in this study thus consists of self-recordings, interviews and a questionnaire. The data has been gathered over a period of 3½ years. The participants in the study consist of 56 secondary school students in a Swedish section in a European School in Belgium. They were between 15–19 years old at the time the data was gathered. All but one have at least one parent who has Swedish as L1, and the majority (N=32) come from families where both parents have Swedish as L1

The purpose of the questionnaire was to provide an overview of the participants’ perception of their multilingualism and language use. It was answered by 49 participants in October 2012. As a complement to this reported data, self-recordings of face-to-face interaction were done by the participants. The second data source thus consists of self-recordings done by a subset of 15 students from the population of the survey in December 2014. Of these, 7 had not answered the questionnaire at the previous occasion but did so by this data collection session. The 15 students recorded conversations in school, at home, on their way to and from school, and in their spare time activities. In school, recordings were done in corridors, recreation areas, canteen, and school yard. Only one recording took place in class, as the school administration was not receptive to recordings done during lessons. The self-recordings cannot be regarded as representations of the participants’ entire repertoire. Rather, they represent a sample of their repertoire. The participants did 72 recordings that comprise a total of 20 hours 46 minutes and 51 seconds. The recordings were analyzed with regard to where the conversations had been held, whether the interlocutors present in the conversation came from networks of family or friends, and what languages were used. Simple transcripts were made of some passages.

The third data source consists of semi-structured individual and group interviews conducted with a subset of students that participated in the self-recordings. The interviews take their starting point in the language use from the self-recordings, but they also generate a general discussion about multilingualism and language use. The interviews were made in March 2016 with 9 students and then transcribed. These 9 students were a subset of the population of the 15 students in the self-recordings. The individual interviews were made with the students, whose self-recordings revealed a multilingual situation in the home domain, comprising a total of 6 interviews with 4 students. Group interviews were done on two occasions with all 9 students. Together the individual and group interviews comprise a total of 2 hours, 30 minutes and 26 seconds, divided into individual interviews (1 hour, 16 minutes, 20 seconds), and group interviews (1 hour, 14 minutes, 6 seconds). The interviews have been transcribed. In connection to each data gathering session the informants have given their permission to participate and for the data to be used. The informants and other participants in the recordings as well as the interviews have been anonymized. They have been given new names that, like their own given names, function in a Swedish, English, and French context. The excerpts used in this article have been translated into English.

As mentioned, throughout the 3½ years of data gathering, I have had an ongoing discourse with the students, their Swedish teacher as well as teachers in other subjects, the school principal and responsible authorities in Sweden, e.g. the Swedish National Board of Education. At the time for the third data collection visit the study based on the survey was finished and had resulted in a publication (Rydenvald 2015), which was presented to and discussed with the students and the principal. The students participating in the self-recordings and their teachers have also read and commented on this article.

5.1 Historical body

The historical body aims at explaining why the participants actually are in the site of engagement executing the social action. The European School is primarily intended for the children of the civil servants in the European Union, and thus the parents’ work constitutes an important common feature among the participants. In addition to the parents’ professional connection with the EU, the participants show similarities in socioeconomic family background; 92% live with both their parents, 88% of the parents work and 89% of the parents have a university education (Rydenvald 2015). They also show similarities in their migrational background as the majority of the participants was either born in the country of residence (N=11/56) or immigrated there before the age of six (N=33/56).

Moreover, the participants share a common linguistic family background. The majority (N=32/56) of the participants have Swedish as their only L1 or together with another L1. In the families (N=23/56) where the parents have different L1’s, the majority language of the country where they live is only present in six families. The exception being one participant who comes from a family in which neither of the parents have Swedish as L1. This participant was born in Sweden and was raised there until eight years of age when the family emigrated from Sweden to the country of residence. This participant reports Swedish as L1 together with the parents’ L1. All participants report a close contact with Sweden which could be seen as a linguistic extension of the family. The majority reports several visits to Sweden each year, the majority have summer residences in Sweden where they also spend a period of at least three weeks during the summer.

Finally, the participants share an educational background. The majority (N=36/56) has had their whole schooling in the European School. Another four started in Year 2 at primary level and yet another two have been away as exchange students for one year. In sum, 75% of the participants have attended the European School in which data has been collected since their second school year. The school is situated in a European country where French is a national majority language. However, in school the participants belong to the Swedish section; one of the nine language section in the school. The participants have Swedish as their L1 language of instruction, and the majority of them has English as their L2 language of instruction. Thus, linguistic and national belongings interact in the participants’ historical background and the interviews show that they are aware of this interplay, which excerpt (1) exemplifies:

The majority of the participants have been raised outside Sweden, in one or several countries. Many students participating in international education, i.e. TCK, lead mobile lives (Pollock and Van Reken 2009), which also applies to the participants in this study. The results show that 43% of the students have lived in one or more countries in addition to the country of birth and the country of residence. The pattern of migration for these families follows two paths: either they have lived in a series of countries before moving to Belgium, or have Belgium as a residential base which they leave for another country for a period of years before returning. In both cases, the majority of the countries concerned have been hosting major EU institutions.

To sum up, both the quantitative results from the survey and qualitative results from the interviews show that the participants share many features in their historical bodies, e.g. family background, relating to parents’ professional careers as well as linguistic and migrational aspects, and educational aspects. The majority of the participants have done the entire or most of their schooling in the European School, with Swedish and English as languages of instruction.

5.2 Interaction order

The interaction order could be described as an interactional behavior within a given social activity. The interactional behavior differs depending on which sort of social activity the interlocutors engage in. Furthermore, the interactional behavior changes depending on where the social action takes place and the relation to other interlocutors and the numbers of interlocutors participating in the conversation. Finally, our historical body plays a part in our interactional behavior, in that our previous knowledge, experience, and beliefs about the interaction order have an impact on our behavior. In this study the interaction order is concerned with the different languages the participants use in different social activities. The social activities referred to in the nexus analysis are those that are present in the self-recordings done by the participants. Comparisons are also made with the reported results from the questionnaire.

The majority of the self-recordings are done in participants’ homes, in school or on their way to or from school. Some recordings are done in the homes of friends or relatives, in stores or when the participants are out walking with friends or dogs. In the recordings done at home the participants talk with parents, siblings and pets, but also with friends, girlfriends and boyfriends. In some cases they talk to friends of their parents. The recordings done in school mostly consist of conversations among friends and peers. The majority of the conversations took place in either groups, i.e. four or more interlocutors, or in pairs.

5.3 Discourses in place

The third element of the nexus analysis that influences the social action is called discourses in place, referring to all discourses that are relevant for the social action, which in this case is the language choice of the participants. In this study the discourses appear mainly through the interviews with the participants based on the language use in the self-recordings. From the analysis, two principles emerge that dominate the discourses in place, namely the principles of inclusion and, what I metaphorically define as, “the least common denominator”.