Language use and investment among children and adolescents of Somali heritage in Sweden

ABSTRACT This article explores language use and investment among Somali-speaking children and adolescents in Sweden, through group interviews and survey data. Our findings indicate that there are incentives to invest in Somali language learning considering the reported language use patterns and the expressed positive attitudes towards Somali mother tongue instruction. The Somali language was perceived to be ‘naturally’ linked to Somali identity and to being able to claim ‘Somaliness’, not only by the adolescents but also by the surroundings. Thus, advanced Somali language proficiency was perceived as necessary for being able to pass as ‘culturally authentic’ (Jaffe, A. [2012]. “Multilingual Citizenship and Minority Languages.” In The Routledge Handbook of Multilingualism, edited by M. Martin-Jones, A. Blackledge, and A. Creese, 83–99. London: Routledge). Furthermore, being perceived as unproficient in Somali or unable to transmit the language to future generations was experienced as guilt-provoking. Nevertheless, the adolescents articulated a compliance with the dominant linguistic order in Sweden, and their school’s assimilatory language rules (‘Swedish-only’). This compliance was associated with good manners and moral behaviour, thus reflecting the potentially harmful and pervasive nature of assimilatory language ideology and policy for individual students. The findings exemplify in many ways the struggles it entails to maintain and develop a minoritised language in a majority language context and the complex ‘ideological enterprise’ of language learning with its educational and ethical dilemmas.


Introduction
Many studies have shown the struggles it can entail to maintain and develop a minoritised language (cf. heritage language) in a majority language context (for an overview, see e.g. García 2009). Generally, these studies have shown that the degree, frequency, and quality of use of a minoritised language are good predictors of the kind of proficiency young speakers will develop in the language (e.g. Kim and Pyun 2014;Lü and Koda 2011). Young people's use of the language will also depend on the wider society's attitudes and reactions to the use of the language and to multilingual practices in general. Furthermore, research has shown that opportunities provided for developing literacy competencies in the language as well as possibilities for engaging in diverse in-and-out-of-school literacy practices are also crucial for language maintenance (e.g. Baker 2011;Freebody and Freiberg 2001;Haneda 2006).
In this article, we explore language use patterns and the incentives to invest in different languages among young individuals of Somali-speaking heritage in Sweden. This includes an exploration into the language ideological enterprise that language learning and use encompasses for them (cf. Harris and Rampton 2003), the relative worth they ascribe to different languages, and their experiences of, and beliefs about the role of so-called Somali 'mother tongue instruction' (henceforth MTI), i.e. the subject teaching of Somali as offered within the Swedish national curriculum (see further below).

The Somali diaspora in Sweden
As a result of increased global migration, Sweden, as many other countries in Europe, has become increasingly heterogeneous and multilingual in the last few decades. Many children in Sweden grow up learning and speaking various languages in addition to Swedish, both inside and outside of their homes. The immigration to Sweden from the Horn of Africa (Somalia, Ethiopia, Kenya, and Somaliland, Djibouti) began in the late 1980s, as a consequence of the collapse of the Somali state and the following years of severe civil wars. The migration has since continued, with a notable peak in the early 2000s (Open society foundations 2014). As a consequence, there is a relatively large Somali diaspora in Sweden today, and Somali is estimated to be one of the 10 largest languages spoken in Sweden (Sweden Statistics 2015). This can be seen in school statistics in the number of children who are entitled to Somali MTI in compulsory school. According to recent statistics, 22.9% of all students in the Swedish compulsory school are entitled to MTI, and 7% of these students are specifically entitled to Somali MTI (Sweden Statistics 2015). In comparison to most other languages offered through MTI, the enrolment rate is comparatively high for Somali, that is, a relatively high percentage of Somali-speaking parents choose to enrol their children in Somali MTI.

MTH in Sweden
As determined by the Swedish Education Act (Utbildningsdepartementet, SFS 2010:800 2010, all students in Sweden with a parent or a legal guardian who speaks a language other than Swedish as their 'mother tongue' are entitled to MTI on the condition that (a) the language is used on a daily basis in their home and (b) the student also has a basic knowledge in the language in question (these two restrictions do not apply to speakers of Sweden's five national minority languages, who are entitled to MTI regardless). MTI, as other subjects within the national curriculum, has its own subject syllabus and its own pedagogical goals and subject requirements.
From an international perspective, the legal support for MTI appears quite impressive, but one can also argue that the implementation of the MTI policy has never been on par with its robust legal foundation (e.g. Ganuza and Hedman 2015;Hyltenstam and Milani 2012;Lainio 2013). Unlike other subjects in the curriculum, MTI is a non-mandatory subject, which means that parents need to actively request tuition for their children. A provision in the Ordinance for compulsory school (Utbildningsdepartementet, SFS 2011:185 2011 also prescribes that municipalities are only obliged to arrange for MTI if there are at least five students requesting instruction in the same language, and if they are able to find a suitable teacher in the language. This means that not all students who are entitled to or request MTI in fact receive it. When offered, MTI entails approximately 40 minutes of instruction per week. Many MTI teachers can attest that this time is insufficient in order to fulfil the MTI subject requirements (e.g. Ganuza and Hedman 2015).
On behalf of the Language Council of Sweden, Spetz (2014) conducted a written survey on the organisation of MTI, which was sent to 290 municipalities in Sweden. The results showed that only about a quarter of the municipalities participating in the survey were able to provide MTI to all students who had requested it. In addition, surveys were also sent to Finnish-, Persian-, and Somali-speaking families with the intention of investigating parents' attitudes to MTI, and their reasons for having decided to enrol, or not enrol, their children in MTI. The majority of the parents considered MTI to be advantageous for their children's development of language competencies in the 'mother tongue', and also for strengthening their children's 'cultural identity'. The majority of the Somali-speaking parents also considered MTI to be favourable for their children's Swedish development and for their school results in general.
Despite all of these above-mentioned limitations, studies have shown that participation in MTI has a positive impact on students' language and literacy competencies in the minoritised language (e.g. Bylund and Diaz 2012;Ganuza and Hedman 2017a), and on their school achievements in general (e.g. Ganuza and Hedman, forthcoming;Skolverket 2008).

Aims and theoretical considerations
The overarching aim of this article is to gain a deeper understanding of language use and investments among young individuals of Somali-speaking heritage in Sweden. We do this by studying their articulations of subjective positions regarding language use and the relative value they ascribe to different languages, mainly within educational contexts but also in regard to domestic realms. In the article, we analyse and discuss how students' positioning index and relate to both large-scale societal discourses and to local discourses of language, education, and the relative worth of languages.
In our analyses, language ideologies are understood as 'beliefs, or feelings, about languages as used in their social worlds' (Kroskrity 2006, 498) and 'dominant' ideologies are here seen as those that historically have been imposed by society and subsequently come to be embodied, internalised, and viewed as natural by most people (Kroskrity 2006).
Studies have shown how institutions, in particular, educational institutions, often tend to be leading in the production and reproduction of language hegemonic structures (cf. Heller and Martin-Jones 2001), thereby contributing to the power of hierarchy between different languages in the society at large. In Sweden, Swedish is prescribed as the official and principal language in the Language Act (Kulturdepartementet, SFS 2009: 600 2009. It is also the main language of schooling, and the perceived 'natural' language, following the logic of historical language hegemony in Sweden (Lindberg 2009;Salö 2016). Hult (2012), who succeeded Josephson (2004), visually represents Sweden's current linguistic order in the form of a hierarchic pyramid, where Swedish, closely followed by English, is on the top of the pyramid, other major European languages as well as the other Scandinavian languages are placed in the middle, and the recognised national minority languages (Samí, Meankiäli, Finnish, Jiddish, and Romani chib) and the so-called newer immigrant languages at the bottom-end of the pyramid.
Closely related to language ideology, is Bonny Norton Peirce's (1995) idea about investment (Peirce 1995), which brings to the fore the social aspects of language learning. With reference to the economic metaphors of Bourdieu, an investment in a language is perceived to be worthwhile when the language-learning endeavour can be converted into a cultural and a social capital. Thus, learners' investment in a language is dependent on their understanding of how language learning will 'increase the value of their cultural capital and social power' (Darvin and Norton 2015, 37). How this enterprise relates to ideology is illustrated in Darvin and Norton (2015).
The valuing of their capital is an affirmation of their identity, a legitimation of their rightful place in different learning contexts. At the same time, because of the pull of ideology, the capital they possess may not be accorded symbolic value by structures of power, or the capital they desire becomes difficult to attain because of systemic patterns of control. (Darvin and Norton 2015, 46) Apart from the social aspects of language learning, the language learner's agency is also a key component in this model. Hence, in our analyses of data, we take into account the notion of investment as well as the prominent role of human agency.
Furthermore, in addition to Norton's model of investment and identity, we also account for Michel Foucault's notion of ethics (cf. Foucault 1983;Foucault and Blasius 1993), as ethical dimensions may also be related to language useas part of an individual's 'self-formation' or identity.
According to Foucault, the construction of the individual subject includes 'the means by which we change ourselves in order to become ethical subjects' (Foucault 1983, 239) or 'the kind of being to which we aspire when we behave in a moral way' (239). Thus, Foucault's ethics 'determines how the individual is supposed to constitute himself as a moral subject of his own actions' (238). These ethical dimensions of self-forming activities relate to an imagined identity, which is different from that introduced by Peirce (1995). According to Foucault, ethical self-formation requires work on the self by the self in various explicit and implicit ways, for example through bodily expressions as well as thoughts and beliefs.

Participants and data
In the article, we draw mainly on data from three group interviews with adolescents of Somalispeaking background, which were conducted by the first author of this article as a part of a research essay that she wrote in connection with the teacher training programme (Palm 2016). As a point of comparison we also draw on background data on the language and literacy practices of 120 six-to twelve-year-old students of Somali-speaking background, which was collected as a part of a research project conducted by the second and the third author (e.g. Ganuza and Hedman 2017a, Forthcoming). About a third of the younger participants went to the same school as the adolescents in the group interviews. One common trait for the participants from both of these data sets is that they have grown up in Somali-speaking families in Sweden. The majority of them were born in Sweden, or came to Sweden at an early age. Most of them live in areas where relatively many people of Somali-speaking heritage have settled, and to where there is a continued flux of migration from the Horn of Africa.
The group interviews were conducted with a total of 13 students in the 9th grade. The only criterion for recruitment for the interviews was that the students had experience of MTI. All but two of the students interviewed reported that they currently attended Somali MTI, while the other two reported that they were enrolled in Arabic MTI. The group interviews were conducted during school hours, in groups of 3-4 students. Each interview lasted for 40-50 minutes. All participants were informed of the purpose of the interviews and they gave their oral and written consent to participation in the study. The group interviews followed a semi-structured format and focused in particular on the participants' language use, their experiences of MTI and their experiences of and beliefs about language and multilingualism. All interviews were audiorecorded and later transcribed.
The data from the background survey were collected as part of a research project on the role of MTI for students' literacy proficiency in the so-called mother tongue (e.g. Ganuza and Hedman 2017a). The project focused in particular on Somali language acquisition and Somali MTI. The background survey was conducted orally and individually, based on a written questionnaire, which included questions about the participants' language and literacy biography and use. The survey was conducted prior to a larger set of tests of vocabulary and reading in Somali and Swedish. In this paper, we only report results from the background survey.
As mentioned previously, about a third of the younger participants who responded to the survey went to the same school as the adolescents who participated in the group interviews. All of these participants attended a private school, with a Muslim profile, which is located in a larger Swedish city. The school receives many students of Somali-speaking background. The school follows the national Swedish curriculum, but students also have some lessons dedicated to Islam. Almost all of the students in this school attended MTI, and MTI attendance is encouraged by the school's principal as well as the teachers. Nevertheless, as will be discussed below, the school also has a language use policy that urges students, and teachers, to always use Swedish in the school, except during MTI, English and modern language classes.

Reported language use and language use expectations
The majority of the young respondents to the language and literacy background survey reported that they speak 'mostly' or 'only' Somali at home with both of their parents (see also Ganuza and Hedman 2017a). A smaller number of respondents reported that they used 'Somali and Swedish equally much', and relatively few that they used Swedish more than Somali at home with parents. Only one respondent out of the 120 reported that s/he used 'only Swedish' with parents. Reported language use with siblings and friends showed the opposite pattern, that is, the large majority reported that they speak 'mostly' or 'only' Swedish with their siblings and friends, some that they use Swedish and Somali equally much, and only a few that they speak 'mostly' or 'only' Somali with their siblings and friends. These reported language use patterns are similar to those reported in many other studies, that is, comparatively more use of the minoritised language with parents, and comparatively less use of the minoritised language with siblings and friends, with whom the language of the wider society tends to dominate (e.g. Aitsiselmi 2004;Boyd 1985;Gregory 2001;Obied 2009;Yamamoto 2001). The majority of the young respondents also reported that they currently attended Somali MTI and had done so for more than one year (96 of the 120 respondents).
The overall language use patterns reported by the adolescents in the group interviews were practically the same, although the adolescents gave a more comprehensive and detailed description of their language use patterns. In general, the adolescents explained that they rarely spoke 'only Somali' in their everyday encounters. Instead, their out-of-school language practices tended to be characterised by an interchangeable use of Somali and Swedish: 'it's like Swenglish but Swedish and Somali instead', as one of the interviewees put it. The adolescents claimed that they rarely thought of their languages as separate, as exemplified in excerpt 1.
Excerpt 1: Faduma: It feels like you're speaking one language […] we just mix it together, it all comes naturally. Amira: It's like one. Faduma: It just comes naturally.
The excerpt exemplifies how the adolescents experienced their everyday multilingual use as something natural and seamless, at least among peers in certain local discourses. However, this multilingual use did not mean that they were not aware of their own and/or other's language use. On the contrary, they recounted how they navigated their linguistic use in accordance with the surrounding, in relation to who might or might not understand and in relation to what language use was expected. Meanwhile, the participants expressed that they sometimes found it uncomfortable if their multilingual practices excluded friends who do not know Somali well enough to partake in the conversation. Thus, the adolescents' language use was to some extent also governed by courtesy (excerpt 2). Excerpt 2: Shadya: Sometimes it can be difficult, well, as you are used to talk with your friend like that because most of them are Somalis here, but some don't understand and that can be hard, well it is not something you mean or, like, to backbite someone or so.
Excerpt 2 also reveals a fear of how 'wrong' language use may entail a risk of being viewed as someone who backbites or even cheats in class: 'If you have a class with a teacher who does not understand what you say, the teacher might think that you talk about them or maybe cheat', as another participant pointed out. The (un)official language policy in the adolescents' school was to use Swedish only when in school (outside of MTI, English, and modern languages). Students who did not comply with these language regulations tended to be reprimanded by the school staff, and sometimes also by their peers (see also Hedman 2015, 2017b). Some of the articulated reasons for this language regulation rule were to avoid exclusion and abusive language. Most of the students also defended the school's language rule and agreed that one should always use Swedish when in school in order to avoid excluding anybody. In this way, the role and use of Swedish become entangled with ideologies of inclusion and good and moral behaviour. What these different examples show is how the adolescents adapt their language use in fine-tuned ways in order to 'fit in' and meet the varying expectations in different interactional contexts. Here, it is important to note the role of language beyond 'conveying meanings', that is, how it acts as a means to position oneself and perform or represent identities (cf. Rampton 1995).
Both the young respondents to the language and literacy background survey and the adolescents in the group interviews tended to rate their proficiency in Swedish higher than their proficiency in Somali, especially for reading and writing. One of the participants even associated lack of minority language proficiency to a sense of collective weakness (in the context of a diaspora). However, the adolescents explained that they had the impression that people around them, especially teachers in the school, often expect them to be more proficient in Somali than in Swedish, as if 'being Somali' automatically means that they are able to speak Somali at a high level of proficiency. This was, for example, visible in a statement by one of the participants who claimed that 'it's like it's already a fact that we should know Somali so well just because we're Somalis' and, likewise, by another who claimed that 'they look at us like we've just arrived from Somalia, and at the same time I don't even know Somali, I don't know it, like I've never even been there'.
During the visits to the school, the authors also experienced how some of the staff referred to this imagined abundant use of Somali, and how they used this as an argument for justifying why the school should be an exclusive space for learning and using Swedish only, because 'where else will they hear Swedish'.
Although some adolescents questioned the assumption that 'being Somali' and having Somali as your 'mother tongue' automatically means knowing how to speak Somali, almost all of them still stressed the importance of knowing Somalioften referred to as 'my' or 'our' languagein order to be able to claim 'Somali identity' or 'Somaliness' (cf. Bigelow 2010). As one participant put it: 'Knowing your language strengthens the sense of fellowship. If you know the language you feel like you belong to the people who speak that language'. Accordingly 'losing the language' would, for the same reason, entail losing a culture. As a consequence of this assumed link between language and identity, many of the adolescents feared what the result might be should they not be able to learn and use Somali well enough, or should they be unable to pass the language on to future generations (see excerpt 3; cf. Bigelow and Tarone 2004).
Excerpt 3: Amira: Like our next, our children, like our parents' grandchildren, what are the odds that they will know their language? Faduma: Exactly. That's why it's so important for us to learn Somali so that like the next generation. The odds are not that high that they will know Somali very well, or even understand perhaps. But at the same time you can only try your best Beyla: But like the language is all we have with us from our country. You know it is an important part of us. It's important, I think.
In fact, a large part of the group interviews centred round issues related to the participants' proficiency in Somali and their fear of not being proficient 'enough', something which was perceived as a shortcoming and as a potential source for embarrassment (see excerpt 4). What these different examples reveal is how the Somali language is perceived as an important and constitutive factor of the participants' individual, as well as collective identities (cf. May 2012). Here, advanced Somali language use even becomes part of a normative or 'moral behavior', pivotal to the 'right' or 'good' way of 'Somaliness' (cf. Foucault 1983, 239), given the feelings of embarrassment and shame that not knowing Somali seems to render. The participants in excerpt 5 expressed embarrassment not only for the individual who does not 'know' Somali but also for the listeners. These examples show how the adolescents tend to presume the existence of an essentialist connection between linguistic competence and ethnic identity (e.g. Ag and Jørgensen 2013;Stroud 2004), possibly forming a notion of an imagined ideal 'self' (cf. the notion of 'ethical self', Foucault 1983). Moreover, the participants' statements show how they tend to see themselves and be seen by others as 'culturally inauthentic' (Jaffe 2012) when they only reach limited levels of minority language competence.

The role and valorisation of Somali MTI
All of the adolescents in the group interviews reported that they currently attended MTI, as did most of the other students at their school. As mentioned previously, many students of Somali-speaking background attend this particular school. One could argue that their school has been successful in promoting MTI and in achieving a high rate of MTI attendance, for example, by integrating MTI during the ordinary school hours (and not in the late afternoons as is done in most other schools), and by locally employing several MTI teachers, who are present and visible in the school every day. However, despite these favourable circumstances the students do not receive more hours of MTI than in other schools, and MTI is also not particularly well integrated with other school subjects (cf. Ganuza and Hedman 2015). During the group interviews, MTI was regarded by the adolescents as a school subject with unquestioned legitimacy within the premises of the local school. All participants agreed that MTI is important and that one should enrol if entitled to it, despite it not being a mandatory subject. They also stressed their parents' appreciation of MTI and considered MTI to be an important arena for maintaining and developing their proficiency in Somali, particularly their literacy skills (see excerpt 5).
Excerpt 5 Nadifa: Well, we don't only learn Somali, we develop it. Jamilla: Well look, we knew Somali and so on when we were little kids but we learned how to spell Somali when we received instruction. Faduma: Spelling, well, it is a big difference between speaking Somali and writing Somali […].
Students with experiences from other schools claimed that their possibilities to draw on their Somali language repertoire were greater in this particular school, due to the presence of many other Somalispeaking students. Many of them also expressed concerns that there would be fewer opportunities for Somali language use after their transition to upper secondary school. Still, within MTI, some of them felt that they were unable to negotiate how their use of Somali in Sweden may differ from uses of Somali as practiced in the 'traditional homelands' (cf. Ganuza and Hedman 2015;Canagarajah 2013). As exemplified in the authors' earlier studies (e.g. Ganuza and Hedman 2017b), many MTI teachers tended to embody and practice purist ideologies of the minoritised 'mother tongue' during the MTI lessons, something which the adolescents claimed to have experienced. However, the participants also told about their positive experiences of an earlier Somali MTI teacher who had been familiar with and permissive of ways of using Somali typical of children and adolescents in Sweden, as exemplified in excerpt 6. This teacher's understanding and shown valorisation of the students' linguistic and cultural resources was perceived by the students as an important motivational factor for using Somali. According to Norton and Darvin's model of investment (2015), learners' investment in a language is dependent on to what extent their linguistic capital is recognised and to what extent and how it 'can serve as affordances to their learning' (46). The teacher's affirmation of their linguistic capital was, hence, perceived as an affirmation of their identity and 'a legitimation of their rightful place' (46) in the MTI learning context.

Embodiment of the established linguistic order
Despite the valorisation of Somali (see above), Swedish was seen as the language of highest prestige and as the unquestionable lingua franca of the school. Swedish was discursively considered to be the 'natural' and most prioritised language, as exemplified in excerpt 7.
Excerpt 7: Amira: Well one prioritizes Swedish. Faduma: Yeah. Amira: Well one prioritizes the language depending on the country. If you're in England then you prioritize English because that what you'll have the most use for. Amira: You know, we are in Sweden, right. Jamilla: You know, if we were in Somalia then Somali would be most important, but we are here now.
Most of the students expressed that they value multilingualism, that 'the more languages you know the better', and also stressed that multilingualism would be a valorised asset by future employers, 'that is what the employers want, that one should know many languages'. Nonetheless, when it came to Somali, the adolescents expressed that they did not think that they would be able to use it neither in future studies nor in their future professional life, unless they got a job as an MTI teacher or as an interpreter. When the first author asked the students whether they wanted to be able to use Somali more frequently in school, for example, for learning subjects other than MTI, they declined and claimed that the MTI lessons were enough for their needs (see excerpt 8). The adolescents' valorisation of different languages is in analogy with ideologies of so-called immigrant bilingualism, that is, the students argue that the main reasons for Somali language maintenance and use is for personal gains and for the development of a desired Somali cultural identity, but less so for instrumental reasons such as social, career and economic relevance (cf. Arthur 2004). However, there were a few rare occasions during the interviews when some students questioned MTI's and Somali's marginalised role in education. For example, one participant reasoned that MTI should at least be assigned as much teaching time as modern languages in the curriculum, 'if we have Spanish twice, then I think we should be able to have Somali twice a week too, at least'.

Discussion
The results of this study show that there are incentives for children and adolescents of Somali-speaking background in Sweden to invest in Somali language learning and identity. The majority of the younger children reported that they spoke 'mostly' or 'only' Somali at home with both of their parents, and the adolescents stressed the importance of knowing Somali in their daily lives and took it for granted that one should attend Somali MTI in order to develop the language. In fact, all of the adolescents, as well as a majority of the young participants in the survey, currently attended Somali MTI. The adolescents' positive attitudes towards Somali and MTI were likely, at least in part, a result of their local school's successful encouragement of MTI attendance. A prominent trait in the interviews was the perception of Somali language to be 'naturally' linked to Somali identity, as Somali language proficiency was considered necessary for being able to claim 'Somaliness' (cf. Bigelow 2010) and for being able to pass as 'culturally authentic' (Jaffe 2012). As a consequence of this perceived essentialist linkage, feelings of not being proficient 'enough' in Somali, for example, in order to be able to transmit the language to future generations, was experienced as guilt-provoking. One participant expressed embarrassment caused by someone's lack of Somali proficiency in MTI class, not only for the speaker but also for the listeners. Thus, high language proficiency seems to usurp a central role for 'real' or 'good' 'Somaliness'.
Paradoxically, while the adolescents expressed their fear of not being proficient enough in Somali, they simultaneously claimed that others often expect the opposite; given an essentialist view that 'being and looking Somali' automatically implies Somali language proficiency. This was, however, contested by the adolescents as some of them claimed that 'I don't even know Somali, I don't know it, like I've never been there.' The adolescents also thought that many of their teachers believed that they spoke only, or mostly, Somali outside of school, despite this not being the case. These statements reveal a perceived mismatch between others' expectations for their language use and identity and the adolescents' own perception, that is, a mismatch in the dynamics between the identities chosen by the self and the identities chosen for the individual by others (Blackledge and Pavlenko 2001;Zienkowski 2017).
Although the adolescents expressed generally positive attitudes towards multilingualism, Somali, and Somali MTI, they simultaneously articulated beliefs that reflected a compliance with the linguistic order advocated both by the school and by the Swedish society at large (cf. Hult 2012). This compliance was, for example, visible in their overt understanding of the school's assimilatory language regulation rules (i.e. 'Swedish-only'). By using languages that teachers do not understand, the adolescents risked being thought of in negative terms, as the use of Swedish was seen as a precaution against exclusion, abusive language, and cheating. One consequence, albeit unintentional, is that Swedish language use then becomes associated with good manners and moral behaviour (cf. Jonsson 2015), disregarding the fact that the unwanted behaviour could possibly also happen in Swedish. Such a consequence reflects how language usage norms and regulations may become entangled with ethic dimensions in intrinsic ways. In fact, regulations of language, apart from language/bodily expressions, also include thoughts and beliefs and as in this case; moral grounds. Hence, these regulations may even target which language use isand is notto be considered to be part of an imagined ideal or 'better' self (cf. Foucault's notion of ethical self-formation), which points to the pervasive and potentially harmful nature of assimilatory language ideologies and policies for the individual.
As documented in previous research on minoritised language use, the interviewed adolescents tended to associate Somali language use mainly with sentimental values, the home and the personal sphere, and less with instrumental values pertaining to upward social mobility and empowerment (cf. Cabau 2014). Thus, in spite of the fact that the students aspired for identity claims such as 'Somaliness', they never challenged the dominant linguistic order in the group interviews.
In the participants' accounts, investments in Somali language learning and use can be converted into cultural and social capital in certain respects. For example, in alignment with Darvin and Norton (2015), one could argue that the adolescents' 'desire to be part of an imagined community' (Darvin and Norton 2015, 47) and to be able to claim 'Somaliness' renders them important incentives to invest in Somali language learning and use. The school's encouragement and valorisation of Somali MTI also plays an important role in providing incentives to invest in Somali. However, many of the adolescents expressed uncertainty about being able to continue to use and invest as much in Somali after transitioning to upper secondary school. The value and prestige ascribed to Somali in the local school context could be viewed as a form of resistance to the wider institutional and social order upheld in Sweden (Heller and Mc Laughlin 2016). However, this local resistance will unlikely have any effect on the perceived value and use of Somali in the wider society. That is, the relative worth ascribed to Somali will most likely remain local, which reveals a consolidated power structure.
In sum, this study highlights the complex 'ideological enterprise' of language learning and its educational and ethical dilemmas. The findings also exemplify the struggles it entails to maintain and develop a minoritised language in a majority language context and/or in a context of diaspora. The strength of the prevailing linguistic structure as well as linguistic and ethnic essentialism cannot be overlooked as those dimensions are deeply intertwined in the students' narratives. However, nor may any examples of student agency that counteract linguistic hegemony, essentialism, and purism be overlooked in order to more closely account forand understandthe relationship between institutional processes and legitimating ideologies of language (Heller and Mc Laughlin 2016). The findings reflect both counteracts and compliance with linguistic hegemony, which possibly constitute everyday life of all the multilingual participants in the study. Although we may not easily find conditions that manage to break the 'glass roof' of the prevailing linguistic order, we still need to continue studying the value that language users attribute to languages in a wider context (cf. Blommaert 2005).