Introduction

State-mandated education policy plays a crucial part in regulating which languages are to be taught in schools and, by extension, which are to be used in a given society. Education thus has an impact on the relative worth and legitimacy of different languages—including competences in them and the speakers of them (Bourdieu 1993; Heller and Martin-Jones 2001). In this light, mother tongue instruction (henceforth MTI) provides scholars of language in society with a useful lens by which to view national responses to multilingualism, as well as diversity more generally. Taking MTI as an object of inquiry, we posit in this article, allows for an understanding of the ‘field of struggle’ that grants legitimate power over various species of capital in a society, including linguistic resources (Bourdieu 2014). Ultimately, this perspective has merits for producing a better understanding of what language policy is and what education does.

Indeed, as notions and classifications, ‘mother tongue’ and, accordingly, ‘MTI’ are language ideologically vested in ways that make their use and reuse shot through with complications (see, e.g., Kroon 2003). As used here, MTI refers to the teaching of minority languages within national systems of education. Thus, ‘minority languages’ are those other than the state-backed national languages, and ‘linguistic minority children’ are speakers of such languages or otherwise associated with them. While the provision of MTI may differ between states, it is implemented worldwide, sometimes under a different label, and sometimes primarily on a private or community basis rather than within the national curriculum or state-sponsored educational structure. However, as we account for in this article, the institutional motifs for and intended objectives of MTI may vary greatly even within what seem to be very similar structures. Illustrating this point, this article investigates MTI in Sweden and Denmark. These two neighboring Scandinavian states are interesting to compare because, while they resemble one another in central ways, they also exhibit significant differences that impact on the organization and valorization of MTI. Descriptively, both Sweden and Denmark may be regarded as multicultural states, in the sense that their populations are ethnically, linguistically, and culturally diverse. However, in terms of multiculturalism as policy, the two countries diverge to the degree that they are often depicted as each other’s opposite (e.g., Hedetoft et al. 2006; see also Lӕgaard 2013). In particular, post-2001, the oft-evoked image is that of ‘the liberal, multicultural Sweden; the restrictive, nationally oriented Denmark’ (Brochmann and Hagelund 2012: 236). These differences are discernible in MTI policy, as we outline presently.

MTI in Sweden and Denmark

In Sweden, MTI has relatively strong legal and institutional support, integrated as it is as an autonomous subject in the national school curriculum. Despite persistent implementation issues (e.g., Ganuza and Hedman 2015), the preconditions for MTI are favorable compared to those in many other countries. Since 2009, Sweden’s Language Act assigns special status to Swedish and the so-called national minority languages (Finnish, Meänkieli, Romani, Sámi, and Yiddish) but states additionally that persons with other first languages ‘are to be given the opportunity to develop and use their mother tongue’ (Swedish Ministry of Culture 2009, Section 14, official translation). Furthermore, as regulated under the Education Act, the state demands that municipalities ensure MTI is provided to children, independent of their linguistic background or country of origin, on two conditions: (1) the language is the child’s everyday language of interaction at home, and (2) the child has basic competence in the language in question (Swedish Ministry of Education 2010). Swedish legislation allows for the arrangement of bilingual education, with a fifty–fifty allocation between Swedish and other languages as the medium of instruction (Swedish Ministry of Education 2011). However, in practice, MTI has come to signify (only) a 40- to 60-min language lesson per week.

Compared to Sweden, MTI in Denmark is more loosely anchored in legal and institutional frameworks, and thus it occupies ‘a vague and unstable position in the national curriculum as well as in a local context’ (Daugaard 2015: ix). The organization of MTI falls upon the responsibility of individual municipalities. What is more, existing national regulations provide municipalities with the opportunity to offer MTI to some groups and not others. In accordance with EU regulations, MTI is mandatorily provided only to children whose parents are citizens of states belonging to the EU or the European Economic Area (EEA), along with Greenland and the Faroe Islands. Municipalities can also optionally organize MTI for children from other areas, but only at the expense of the municipality or (fully or partially) funded by their parents. For groups eligible for MTI, provisions are contingent upon there being a sufficient number of students and qualified teachers. Publically organized MTI covers a maximum of 150 h and a minimum of 60 h per year, depending on the number of students, the grade range in the classes, or whether MTI is organized ex gratia on the part of municipalities. Actual practices thus differ in ways that make MTI provisions miscellaneous and difficult to oversee.

According to the Swedish national curriculum, the objective of MTI is to provide students with the opportunity to develop knowledge in and about their mother tongues. At the same time, it stipulates that in gaining knowledge about their mother tongue, students are to ‘become conscious of its importance for their own learning in different school subjects’ (Swedish National Agency for Education 2011: 83, official translation). This phrasing subtly indicates that skills in the mother tongue can be transferred to other forms of knowledge but leaves it to the students, or, rather, their teachers, to work out how. The national curriculum also makes reference to the link between MTI and students’ cultural identity; it states that ‘the instruction ‘should give pupils the opportunities to develop their cultural identity and become multilingual’ (ibid.). In a similar vein, according to Danish policy (Danish Ministry of Education 2009, our translation), the aim of MTI is for students to ‘acquire competences and skills in order to become capable of understanding the spoken and written language and of expressing themselves in speaking and writing.’ However, the policy makes no reference to the objective of strengthening a non-Danish ethnic or cultural identity. It does, however, stipulate that ‘The instruction will give the students insight into the culture and society of the country of origin, among other things, in order to ease the students’ possible return to this country’ (our translation). What is more, in Danish educational policy and debate, MTI is, for the most part, discussed as an instrument for enhancing the acquisition of Danish or other skills taught in the ordinary educational system, whereas other learning objectives are largely overlooked. It is particularly striking that knowledge in the languages labeled as ‘mother tongues’ of minorities has little currency per se in the linguistic market of Denmark; skills in the designated mother tongues per se are accepted chiefly insofar as they are exchangeable into the capital of Danish language skills (Daugaard 2015; Kristjánsdóttir and Timm 2007).

Objectives, material, and procedure

The states of Sweden and Denmark, then, have put into place divergent language policies with regard to MTI. How can these differences in policy be explained? What ends do these policies serve, and how have they gained legitimacy? In addressing these questions empirically, drawing on Brubaker (1990), we adopt a historical, comparative perspective to unravel some of the inter-state divergences in MTI policy and discourse. It is vital to recognize that nationally system-specific features are historically manufactured. Understanding contemporary specificities of MTI policies, therefore, entails understanding the processes by which such specificities are products. The comparative component fulfills the objective of accounting for the specificities of each polity by contrasting it with the state of affairs in the other as a source of comparison—examined, thus, ‘in light of observable alternatives’ (Kocka and Haupt 2009, 4). In this way, we seek to understand the conditions that have affected MTI in Denmark in relation to those at play in Sweden, and vice versa. The scope of this endeavor is limited to the conditions surrounding MTI in Sweden and Denmark during the formative period of MTI-related policy-making, viz. the 1960s, ’70 s, and ’80 s, although some mention is also made of newer developments. Theorizing these conditions, we adopt an analytical grid constructed within the framework of Pierre Bourdieu, which has many virtues for exploring the intersection of (state) policy, education, and language within a coherent framework (see below).

We study MTI, on the one hand, as policy, broadly perceived as the vision for MTI as expressed in central regulatory documents at the national level. On the other hand, we investigate MTI in discourse produced in the political and academic field of each country. As a result, the empirical material analyzed consists of a rather motley collection of text documents, produced at different points in time and relevant to MTI in different ways. Firstly, this set consists of legal texts inscribed in national acts on language and education, as well as national curriculum documents and equivalent policy documents of different times (Danish Ministry of Education 1976, 1984, 1993, 2009; Swedish Ministry of Culture 2009; Swedish Ministry of Education 2010, 2011; Swedish National Agency for Education 2011). Secondly, the material consists of academic and political discourse, often canalized through print press and, hence, via the journalistic field of each country (e.g., Andersson et al. 1991; Vibjerg 2012). Where relevant and applicable, we also make reference to reports and regulations produced by national and international institutions (Council of Europe 1977; Swedish Ministry of Employment 1974; Swedish National Audit Office 1990; UNESCO 1953). Analyzing this material, we employ the line of sociohistorical analysis developed in Bourdieu’s work, which is attuned to emphasizing struggles and effects within and across fields (see Gorski 2013). Methodologically, this entails interpreting the place of MTI in light of the history of multiculturalism in the two countries, including developments in immigration policy (e.g., Brochmann and Hagelund 2012). It follows from this that we rely additionally on meta-analysis of empirically validated findings, including a body of earlier studies relevant to multiculturalism and minority education. This material is subsequently interpreted through the lens of Bourdieu’s notion of fields in order to add a sociological understanding to divergences between MTI policies in the two countries.

State policy, education, and language: a Bourdieusian perspective

According to Bourdieu, societies are differentiated into fields and subfields—that is, separate social universes imbued with their own logic and functioning forces, each populated by human agents as well as institutions that struggle over something recognized as being of mutual value and interest (e.g., Bourdieu 1993: 72–77). In his later writings, Bourdieu (2014) came to theorize the state as a ‘meta-field’ within such societies, one that is able to exercise power over other fields because of the concentration of capital it has come to accumulate (e.g., Bourdieu 1994, 2014; Wacquant 1996). Correspondingly, it possesses a form of ‘meta-capital’, granting power over other kinds of capital (Bourdieu 2014: 345), and acts as ‘the central bank of symbolic credit [which] guarantees and consecrates a certain state of affairs’ (Bourdieu 1996: 376; see Karlander 2017). Of particular interest here is that the state is in the position to exercise power over the exchange rates between different species of capital and, thereby, over the relative worth of their holders (Bourdieu 1994: 4). Impacting on such visions is in itself a central stake in struggles between holders of capital—politicians, scholars, journalists, etc.—who struggle for power over the state within the social space referred to as the field of power (Bourdieu 1994: 5).

In this sense, the state is in and of itself a field of struggle over legitimate power. Here, various fields have stakes and interests attached to affecting the exchange rate between different kinds of capital (Bourdieu 2014: 198). As archetypes of fields located within the field of power, Bourdieu (2005a) foregrounds the political field, the academic field, and the journalistic field. Operating from within their own fields, politicians, academic scholars, and journalists produce accounts of the social world, and struggle to make them known and recognized. Noteworthy, however, in Bourdieu’s framework is that these agents are not conceptualized primarily as individuals, but as agents positioned within particular social fields, and as holders of particular forms of capital effective within those fields. Ultimately, then, it is the fields that interact with one another, confront one another, and, in so doing, struggle to ‘impose the legitimate vision of the social world’ (Bourdieu 2005a: 40).

Thus, for Bourdieu, agents of various fields commonly seek to anchor their agendas within the state, or, more precisely, within what we shall here term the bureaucratic field (Bourdieu 1994). Comprehended as the display of the dominant vision of a given society, state policy can be seen as an indicator of the success of various fields. From this view, a policy is, firstly, a product of struggles within fields and, secondly, the product of cross-field effects (Lingard et al. 2005). This position allows for a historical understanding of language policy—one that entails recognizing and accounting for the active labor of field agents involved in shaping a given policy and the instruments used to enforce it. For Bourdieu, the state imposes its visions through its institutions and the agents who act in its name. On that note, Bourdieu places the lion’s share of the weight on the role of education, which he sees as a pivotal instrument for the effective imposition and inculcation of the dominant vision of societies: ‘Not the army, the asylum, the hospital, and the jail, but the school is the state’s most potent conduit and servant’ (Wacquant 1996: 162). Not least, schooling is a crucial site upon which the state’s unifying power is manifested, and it is accordingly used to designate legitimacy to languages to dominate the linguistic market (Bourdieu 1994: 7–8). The result of this process is readily visible in the linguistic markets of Sweden and Denmark, each dominated by a state-mandated language that sets the benchmark against which the value of all other languages is defined (e.g., Bourdieu 1993).

MTI in Sweden and Denmark: historical struggles and cross-field effects

In both Sweden and Denmark, MTI gained a foothold in the school system throughout the 1970s. Nonetheless, the motifs and driving forces differed in each context. Historically, Swedish and Danish polities have had different relationships to their ‘old’ minorities, and this fact has had a steering impact on their relationships to ‘new’ minorities, as well as on their approaches to multiculturalism more generally (e.g., Hedetoft et al. 2006; Hegelund 2002; Wickström 2015).

State politics, MTI, and academic knowledge

Denmark has a history as an empire that includes Greenland and the Faroe Islands, but, apart from a German faction, there are no recognized autochthonous linguistic minorities within the current territory of the Danish state (see, e.g., Moldenhawer and Øland 2013; Jørgensen and Holmen 2005). Sweden, by contrast, has had a strenuous relationship with its autochthonous minority groups, in particular, the Roma, Sámi, and the Tornedalen Finns. The era of nation-state building was marked by overt discrimination of these minorities, which included the hardline imposition of Swedish in the school system. Disinclination towards supporting cultural diversity was also pertinent to the immigrants who settled in Sweden after World War II, either as war refugees or labor immigrants. However, in international relations and foreign policy Sweden had, at this point, made a name for itself as a strong advocate of peace and social justice, which, in the eyes of agents from the political field, did not square with Sweden’s strained relationship with minorities on its home ground (e.g., Borevi 2012; Brochmann and Hagelund 2012).

The Swedish position came to be informed by a debate in the academic field on so-called ‘semilingualism’ (see Stroud 2004), as drawn into focus by the Swedish linguist Nils Erik Hansegård. In the 1960s, Hansegård coined the term ‘semilingualism’ to describe Finnish-speaking youth in the northern parts of Sweden who, in his view, mastered neither Swedish nor Finnish ‘entirely’ as a result of not having received formal instruction in their mother tongue. In chorus with other renowned linguists, he considered this to be sociolinguistically unfair, morally unsustainable, and unworthy of a country such as Sweden (Hansegård 1968; see Malmberg 1964: 156f. and note 1). Hansegård’s argument was widely disseminated via his 1968 book ‘Tvåspråkighet eller halvspråkighet?’ [Bilingualism or Semilingualism?], where several mentions were made of the UNESCO recommendation of 1953, which stated: ‘It is axiomatic that the best medium for teaching a child is his mother tongue’ (UNESCO 1953: 11). Coterminous with Sweden beginning to receive increasing flows of labor immigration throughout the 1960s, demands for MTI in non-Swedish languages were raised by representatives of the Finnish, Estonian, and Jewish groups (e.g., Wickström 2015). Hansegård’s argument was soon drawn into the Swedish political debate on MTI (see Borevi 2002; Hansen 2001). This is a cross-field effect that exemplifies a case of import of scientific thought—entextualized in the concept of semilingualism. Politicians and advocates of MTI alike could argue that the inequality bred by semilingualism in the north threatened an entire generation of immigrant children, and that this situation was incompatible with the Swedish notion of equality hailed at the time (Wickström 2015). By and large, the late 1960s and early 1970s in Sweden were marked by a radical move away from the previous assimilationist position, as the country instead opted for a multicultural stance that encompassed not just recognition but promotion of difference (e.g., Runblom 1994). The thought of implementing MTI in Sweden in the 1970s aligned well with this vision.

Importantly, in the post-war era, immigrants to Sweden were conceptualized as long-term residents rather than guest workers (Hyltenstam and Tuomela 1996: 23). This was not so in Denmark. Here, conversely, the impetus for implementing MTI differed since multiculturalism was never embraced as official state policy (e.g., Lӕgaard 2013; Schmidt 2013). Accordingly, the question of organizing MTI surfaced not as a way of meeting the demands of ethnic activists or autochthonous minorities, but rather, as a facet of labor politics (Hegelund 2002). As in many other Western European countries, the industrial boom of the 1960s led to a labor shortage, and, as part of a guest worker scheme, the Danish government invited labor immigrants to temporarily reside and work in Denmark. The expectation was that they would return to their countries of origin when their services were no longer needed (Padovan-Özdemir and Moldenhawer 2016). Denmark joined the European Cooperation in 1973, and early policy developments concerning MTI can be seen as an effect of Denmark’s aligning with common European frameworks (e.g., Council of Europe 1977). Kristjánsdóttir (2006: 146ff.) notes that in the debates on how to deal with the linguistic needs of the children of immigrant laborers, references were made to Sweden’s developments in this area. As in Sweden, it was generally held that the acquisition of the majority language would be facilitated by developing the mother tongue. MTI was provided sporadically in Copenhagen in 1972, and linguistic minority children were included in the Primary School Act in 1975 (Danish Ministry of Education 1976).

Political decision-making and cross-field effects

Notably, academic knowledge production in Sweden (e.g. Hansegård 1968) had a steering impact on Swedish MTI policy developments in ways that do not apply to Danish developments. Importantly, the Danish Act of 1975 stated that it was up to the minister of education to decide upon the rules and regulations concerning the education of linguistic minority children. This phrasing was in agreement with a common structuring feature in Danish political culture, namely, the practice of allocating significant power to the responsible minister of a given political realm (e.g., Wenander 2016). As a consequence, while state-sponsored MTI was offered to children irrespective of their cultural and linguistic backgrounds, already from the start it failed to gain traction in the higher order of shared views and values legitimized by the field of power. Not only did it lack solid anchorage in the educational system, but legitimation in the state field as well. In Denmark, the chief motivation for MTI was to preserve the children’s language skills in order to facilitate a future repatriation to their countries of origin.

As opposed to Denmark, the general cohesiveness surrounding MTI in the Swedish political field allowed the academic field to occupy a prime position in the Swedish field of power. As Hammar (1997: 167, our translation) puts it, the politicians ‘handed over to experts within and outside of the parties to come up with technical solutions.’ In fact, the Swedish decision-making system is set up in a way that greatly facilitates this form of cross-field effect by virtue of the long-upheld and widely used practice of installing ad hoc commissions prior to making important policy decisions. Following Bourdieu (2005b: 104), we may think about the commission as a ‘bureaucratic form of consultation’ that makes possible exchanges between fields, such as the legislating field and the knowledge-producing academic field. Subsequently, commissions provide ‘a cat-flap’ between science and politics (Wisselgren 2008)—that is, a window through which agents from the academic field can impose their visions on state policy. Their work materializes in state-commissioned reports, which are performative discourses imbued with quasi-official authority (Bourdieu 2014: 26).Footnote 1

Anton (1969) singled out the commission system as one of the most striking features of Swedish political culture and described it as being key to minimizing political disagreement and conflict. In Sweden, typically, representatives from the state and political fields cooperate with experts from academic fields in conducting thorough investigations, so as to produce fact-based foundations on which well-informed political decisions can be made. Apart from fulfilling an important consensus-building function in Swedish politics, the Swedish commission practice is said to have contributed to a notable respect for expert knowledge (Johansson 1992). In Sweden, resulting from commission work conducted between 1968 and 1974, a new immigration policy was presented in the highly influential government report ‘Invandrarutredningen’ (‘The Immigrant Commission’) (Swedish Ministry of Employment 1974). On the one hand, the report advocated ‘equal conditions’ between immigrants and ‘the majority population’, for example, in preserving and developing ‘their own language’ (p. 95). On the other hand, it advocated freedom of choice: immigrants were to decide for themselves the degree to which they wished to adapt the majority values vis-à-vis retaining, among other things, ‘their linguistic identity’ (ibid.). It also stressed the need to not just recognize the children’s linguistic backgrounds but to actively promote the use of their so-called home languages as the medium of instruction (Swedish Ministry of Employment 1974: 256–257). Jacobsson (1984: 78) claims that academics and experts alike played a crucial role in convincing the committee members that MTI was worth supporting. As Borevi (2012) and others show, Sweden’s new multicultural position was swiftly becoming official policy. In fact, following a law passed in 1976 (Swedish Ministry of Justice 1976), the Swedish Constitution (cp.1 Section 2, our translation) was supplemented by language stating that ‘The opportunities for ethnic, linguistic and religious minorities to preserve and develop a cultural and social life of their own should be promoted.’ In 1977, this vision moreover materialized in the Home Language Reform, which has since constituted the cornerstone for the organization of MTI in Sweden (e.g., Hyltenstam and Tuomela 1996; Lainio 1997).Footnote 2 In the political debates that preceded this reform, as well as in its preparatory work, agents pointed to the fear of semilingualism, and all political parties were in agreement about the importance of immigrant children’s mastery of their ‘home languages’ (Borevi 2002: 209–210). In view of this, the reform made MTI into a general right offered to all children, irrespective of linguistic background. This provision would in due time come to signify a divide between the implementations of MTI in Sweden and Denmark.

The stakes of immigration: union and disunion in the political fields

In Denmark by the end of the 1970s, it was clear that guest workers were not necessarily going to return to their countries of origin, and immigration became an increasingly heated issue in the Danish political field. More specifically, from this point onwards, the country saw the electoral success of the right-wing Progress Party, and, owing much to the discourse introduced and maintained by this party, immigration in Denmark developed into a salient stake in the political field. This fact created division within the political field, since the left-wing parties were still in favor of a pluralistic approach (Jønsson and Petersen 2012: 106–107). In 1982, a liberal-conservative coalition came into office and remained for a decade. This administration framed immigration as a social and economic burden to the model of the Danish welfare society (see Moldenhawer and Øland 2013). In 1984, a new statutory order launched new regulations for MTI (Danish Ministry of Education 1984), which stressed that its provision should be grounded in the Danish context, thus breaking with the previous position of linking MTI to the children’s country of origin (see Hegelund 2002: 98f. for a commentary). Throughout the 1980s, the Danish debate was characterized by public discourses on the ability of immigrant children to adapt to Danish culture. This proclaimed ‘cultural gap’ motivated endeavors to improve their Danish language acquisition by providing MTI (Padovan-Özdemir and Moldenhawer 2016: 6). Still, the learning objectives stated that MTI should maintain and develop language knowledge in the mother tongues to ensure sustained bonds to their origins and, it is worthy to note, to facilitate the possible return to their countries of origin (Hegelund 2002: 99, 104–105; Padovan-Özdemir and Moldenhawer 2016: 7).

Throughout the 1980s, Sweden was marked by a retreat from the multiculturalist stance of the 1970s. Consequently, few ventures were made to boost or develop MTI further—in fact, while government subventions continued to support MTI, the number of participating students dropped. Likewise, the Swedish administration now felt the need to inspect the educational model launched in the previous decade and the large-scale enterprise of bilingual education it had sustained. The Swedish National Agency for Education (then: Skolöverstyrelsen) commissioned the internationally renowned expert Christina Bratt Paulston to produce a report on the Swedish debate on bilingualism in education. Among other things, her report (Bratt Paulston 1983) pointed out that the debate had been heavily informed by the concept of semilingualism and the phenomenon it invoked, which was deemed non-scientific and non-existing, and that agents active in the debate had recurrently overstated the importance of the mother tongue in identity formation. However, while Bratt Paulston’s position deviated from the mainstream stance in the Swedish academic field, she did not question that MTI was worth promoting. The rationale for organizing MTI did not merely hinge on the idea of developing skills in the mother tongues but stood on several grounds. Thenceforth, secondary values were strategically invoked, notably the idea that MTI was beneficial to the students’ learning of Swedish.

MTI and impact from the academic fields

In Denmark, coterminous with anti-immigration discourses, MTI arose as a stake in the academic field in the 1980s, where scholars increasingly began to engage in the debate. Among the most active commentators were Birgitte Rahbeck and Tove Skutnabb-Kangas, both of whom were overtly pro-MTI. The latter scholar had previously taken up positions also in the Swedish debate on MTI but now came to engage more actively in the Danish academic discussion. Rahbeck and Skutnabb-Kangas made several arguments for supporting multilingual education, and they framed multicultural education as being linked to linguistic (in)equality and democracy. However, as Kristjánsdóttir (2006) argues, while these scholars stood out as authorities in the debate, their viewpoints and rhetoric were perceived as being too far-reaching in relation to Danish praxis, which was defended by other agents of the academic field, encompassing agents from a wide range of disciplinary subfields. For example, Søndergaard (1982: 308) argued that a qualified debate on pedagogic practice in Denmark was hindered ‘when, for instance, the prominent bilingualism scholar Tove Skutnabb-Kangas refers to the Danish model as physically and content-ideologically integrative and assimilatory’ (quoted in Kristjánsdóttir 2006: 184, our translation). In this vein, opposing positions were taken in the academic debate on MTI, and a professed disparity in scholarly standpoints provided a sense of incredibility to research on MTI in Denmark in the early 1980s. Policy-makers were able to make tensions and divergent positions within the scholarly field into a resource for arguing that while some academic experts hold skills in the mother tongue to be important, others express doubt about such effects. It also meant that pro-MTI arguments could be dismissed as belonging to the tropes of perceived left-wing radicals (Kristjánsdóttir 2006). An important consequence of these struggles was that agents from the academic field were no longer listened to and that the value of scholarly expert knowledge in MTI decreased. In the early 1990s, the Danish government could point to disagreement among linguists about the role and importance of mother tongue in educational contexts. Thus, in 1993, when a new educational act was launched (Danish Ministry of Education 1993), it differed little from the previous rules of 1975: the content and structure of MTI was still the decision of the responsible minister, and it was the obligation of municipalities to organize classes.

The 1990s resulted in a number of setbacks for MTI in Sweden that came as a result of discourse in the economic field, which focused on the societal costs of its implementation. Borevi (2002) asserts that in the period from 1988 to 1998, there was an emphasis on state finances in relation to immigration-related policy. Due to a financial crisis, MTI was subject to cutbacks, after having been targeted by a set of economic arguments produced by what Bourdieu terms ‘the right hand of the state,’ thus, as a cross-field effect imposed by agents in the economic field. For example, a report published by the Swedish National Audit Office (1990) criticized the implementation of MTI and argued that the money spent on the reform did not apparently yield the profit that was intended (Hyltenstam and Tuomela 1996: 23). Implementation issues, however, were already known in the academic field, and the report met with heavy criticism. In 1991, an article appeared in Dagens Nyheter, one of the largest dailies in Sweden, with the message that ‘immigrant children are being sacrificed’ as a result of cutbacks to MTI (Andersson et al. 1991). The article was signed by some 50 researchers from the academic field, distributed across the universities of Gothenburg, Lund, Stockholm, and Uppsala. This illustrates the broad support for MTI from academics, as well as their capacity to mobilize and gain an audience within the political as well as the bureaucratic field. In this vein, MTI was again able to endure, despite the retreat from the previous multiculturalist stance. Proponents of MTI increasingly pointed to additional benefits linked to MTI, such as enhanced prospects for developing skills in Swedish and other core school topics, that is, indirect forms of capital exchange (cf. Heller 2010), holding currency in the academic field as well as in the political field.

The persistent impetus of immigration in field struggles

In Sweden, immigration had surfaced as a ‘problem’ in the late 1980s, but in comparison to Denmark the issue was not conflict-ridden (e.g., Green-Pedersen and Krogstrup 2008). Indeed, in Sweden local political parties and xenophobic movements emerged around this time, including the right-wing populist party New Democracy, which made it into parliament in the elections of 1991. While this party was overtly against MTI, its existence proved to be short-lived, and its discourse on MTI had little or no impact on the political field. Accordingly, MTI remained in the Swedish educational system despite that the rules and entitlements to MTI were somewhat restricted (Cabau 2014).

In Denmark, the year of 1995 saw the establishment of the Danish People’s Party, a far-right party with roots in the populist movement of the 1970s, whose criticism of immigrants and immigration brought further disunion to the political field. In particular, it left the Social Democrats—governing in coalition from 1993—deeply divided, while liberals and conservatives latched on to the critical discourse that was now in circulation (Hedetoft 2010; Jønsson and Petersen 2012: 121–131). In 2001, a few months after the events of 9/11 in the U.S., a conservative and liberal government entered office with the support of the People’s Party. This new coalition ‘skillfully played the immigration card’ (Lӕgaard 2013: 176), which included the successful leveling of criticism against the past social-liberal government for being too soft in matters of immigration, integration, etc. MTI, as well as other symbols of ‘non-Danishness’, were caught up in an increasingly heated debate on immigration as a social problem and an emphasis on the importance of imposing ‘Danish values’ on non-Danish subjects residing within the boundaries of the polity (Lӕgaard 2013: 188–189; Schmidt 2013: 205ff.). Within this debate, the educational system was emphasized as a key instrument for reproducing perceived Danish core values and norms (Moldenhawer and Øland 2013: 406). In the process leading up to the elections of 2001, the Liberal’s leader, Anders Fogh Rasmussen, promised to bring MTI to a halt should he win. Accordingly, after winning the elections, a mission statement detailed that ‘The undertaking of improving foreigners’ knowledge in Danish is strengthened. Obligatory mother tongue education is abolished’ (Danish Prime Minister’s Office 2001, our translation). However, due to EU regulations a less drastic policy was eventually implemented. This policy stated that municipalities no longer needed to provide MTI to children who did not come from the member states of the European Union; from countries outside of the EU that are part of the EEA area; and for children from the Faroe Islands or Greenland. Municipalities would also remain free to voluntarily organize MTI for ‘non-EU’ children, but it was no longer obligatory. As a result, the number of linguistic minority children to whom municipalities were obliged to offer MTI was reduced dramatically: in 2008, there was a decrease from 70,000 to 7500. The remaining 62,500 children would be able to receive publicly funded MTI only by virtue of benevolent initiatives of individual municipalities (Timm and Kristjánsdóttir 2011: 76–77). This situation has given rise to criticism in the academic field, but the criticism has remained relatively aloof from or ineffectual in the political field. As the Danish politician Inger Støjberg, currently the minister of immigration and integration, put it in 2012: ‘In Denmark we speak Danish. You have the right to learn all the languages you want, but it has to take place in your spare time’ (Vibjerg 2012: 11).

Comparative discussion

In this article, we have sought to account for differences in MTI in Sweden and Denmark by focusing on a half-century of relations between the academic field and the political field, and the relation of these two fields to the bureaucratic field, where policies are produced (Bourdieu 1996: 386). This approach has disclosed a number of points in which cross-country comparisons are relevant to our understanding of MTI and the language policies it entrenches. In both Sweden and Denmark, MTI is a language policy enacted in the educational field, and, over time, it has grown into becoming closely linked to immigration. Relatedly, Hyltenstam and Tuomela (1996: 11–12) have argued that MTI in Sweden has often come to stand proxy for discussions about immigrants and immigrant policy more generally. While concurring with this claim, we would argue that it holds even truer in respect to state-mandated MTI policy in Denmark, where it seems particularly evident that ‘Linguistic struggles may not have obvious linguistic bases’ (Bourdieu 1993: 80). Rather, as Kristjánsdóttir and Timm (2007: 17) regretfully note, while policies on linguistic minorities in Sweden are grounded in theories on language and education, the equivalent policy area has been served up as integration politics in Denmark. Consistently, the rationale for organizing MTI in the Swedish educational system is based on a battery of ostensibly positive effects for the individual as well as for society at large. In Denmark, conversely, MTI is viewed predominantly as an instrument for enhancing the acquisition of Danish. By the same token, whereas MTI in Sweden has been seen as a right within a multiculturalist scheme, MTI in Denmark has been seen as a reluctant restraint, or as an arrangement for repatriation without affording migrants an expectation of staying.

Indeed, a key difference between the two countries is that immigration has had a more significant stake in the political field in Denmark than in Sweden, and for a longer period of time. In Sweden, all political parties, including the political right, have been in general support of a liberal, multicultural policy since the formative phases of Swedish immigration policy. Partly, therefore, in the Swedish political field, overt anti-multiculturalist positions were generally absent until the 1990s. Meeting the linguistic needs of non-majority speakers through policies such as MTI has therefore been widely supported as the politically and, in addition, scientifically correct thing to do—aligned with national narratives of solidarity, egalitarianism, and anti-racism. As a result, MTI has a relatively solid foundation in Swedish national legislation and is offered to linguistic minority children and adolescents irrespective of their linguistic and ethnic backgrounds, notwithstanding persistent implementation issues (e.g., Ganuza and Hedman 2015). Since 2010, however, the Sweden Democrats on the far right of the Swedish political spectrum have had tremendous success in the elections of recent years. Importantly, though, they have been largely secluded from the political establishment insofar as no other parties—as of yet—have been willing to formally cooperate with them. In fact, Swedish politicians often point to Denmark as a warning of the possible outcome of legitimizing right-wing populist parties and allowing them an agenda-setting role in immigration politics.

Compared to Sweden, the Danish political field displays more conflict and ideological disagreement, where agents belonging to the right wing recurrently ‘demand a more assimilatory policy’ (Brochmann and Hagelund 2012: 240). Here, correspondingly, the right to state-sponsored MTI has been abolished for children who are deemed to be of non-European descent. In Denmark, the emergence of far-right populist parties had transpired previously in the 1970s and ’80s. This fact contributed to creating a divided political field where immigration and immigration-related policies in general have become objects in political struggles. With its populist roots, the Danish People’s Party has not been stigmatized in the same way as the Sweden Democrats (Rydgren 2010). On the contrary, the Danish People’s Party has had a profound impact on political discourse in the immigration area. Correspondingly, policies that aim to preserve immigrants’ cultural identities are politically controversial in Denmark. Accordingly, MTI in Denmark is opposed by agents with a critical stance on immigration policy more generally; in Denmark, such agents are distributed across the entire political spectrum. As a consequence, as Hedetoft (2010: 110) asserts, ‘leading Danish politicians from all agenda-setting parties […] have repeatedly stressed that Denmark is not and does not intend to be a multicultural society.’

Because migration in Sweden was not overtly controversial in the early formative years of MTI policy, it was not high on the agenda in the political field. This fact made possible considerable cross-field effects produced by academics in the new and relatively insignificant policy area of immigration. In respect to MTI more specifically, agents acting from within the Swedish academic field have mainly been language scholars, which might explain the ostensible unity characterizing this field. At any rate, they have remained united by a common position-taking that MTI is to be supported. In Swedish scholarly discourse, MTI has been deemed beneficial for the development of students’ cultural identities, as well as for their enhanced prospects for developing skills in Swedish and other core school topics. Arguably, this may have rendered MTI less vulnerable in struggles where the future of MTI has been at stake.

Adding to these dynamics, the Swedish decision-making system is set up in a way that provides well-positioned agents from the academic field with a window into the process of crafting a given policy, often through the long-established Swedish commission system. It may be argued, firstly, that the commission practice partly explains Swedish political consensus on MTI: political disagreement can to some extent be avoided when scholarly produced knowledge claims are on the table. Secondly, and more generally, it may be argued that this situation has resulted in a form of respect for expert knowledge that academics do not enjoy elsewhere, and particularly so on matters of immigration in Denmark (see Jørgensen 2011). While this is not to suggest that Swedish academics ‘get to decide’ carte blanche how language policy should be shaped and implemented, it does suggest that Swedish academics have been better positioned within the field of power compared to their counterparts elsewhere. At any rate, it is clear that the Danish academic field has been less influential in informing MTI policy with authority. There are several reasons for this. Firstly, the Danish decision-making system is set up in a way that attributes less authority to experts. While ad hoc commissions exist in Denmark, the Danish political system is organized in ways that instead bequeath the authoritative power to act to the appropriate minister (e.g., Wenander 2016). The opportunities for academics to impose their vision of the world have been particularly rare in Danish immigration policy, which has been largely dominated by agents from other fields (Jørgensen 2011). Secondly, and partly related to the first point, while there are counter-positions present in the Danish MTI debate, such as those of academics, these views on multiculturalism and MTI are seldom taken into account. From the start, MTI was anchored foremost in the administration of the state, and thus, unlike in Sweden, it was potentially subject to the ebb and flow of party-political interests. Arguably, this circumstance is related to the active involvement of Danish academics from a broad range of disciplinary subfields in the early era of MTI policy-making, which may have yielded counter-productive effects. Scholarly debaters in the 1980s appeared to be in profound disagreement over the preferred organization of MTI, both between each other and in relation to the position of the Danish state (Kristjánsdóttir 2006). Since then, Danish scholars such as Kristjánsdóttir and Timm (2007), Horst and Gitz-Johansen (2010), and others have pointed to the fact that the knowledge they produce fails to inform policy, because the Danish state field effectively shields off multicultural positions.

It should be noted that the emphasis of this article has been on the educational system as conceptualized mainly in national terms. However, in both Sweden and Denmark, but particularly in the latter, the educational system is decentralized in ways that position municipalities as the executive agent of schooling. While their autonomy is never total vis-à-vis the state, it gives them extensive room for enacting policies that do not necessarily adhere to the position of the state. As a case in point, densely populated municipalities, such as Aarhus and Copenhagen in Denmark, have offered MTI to children with linguistic backgrounds beyond the EU area, although without financial support from the state (see, Daugaard 2015). By the same token, the national gaze opted for here does make Swedish MTI appear remarkably well organized, whereas, in practice, Sweden’s lofty ambitions have been criticized for not having been accompanied by the political will and resources necessary to realize its intentions (e.g., Lainio 1997). Be that as it may, national policy is not the only level relevant to empirical scrutiny, and it is possible that a national comparison focusing on MTI in practice would disclose less striking differences between the two states studied here. Studies comparing the practical implementation of MTI policies in Sweden and Denmark would therefore be valuable.

Conclusion

Controlling the educational system is vital to the labor of controlling linguistic markets, including the valorization and distribution of linguistic resources. Because policymakers have a profound influence over the curriculum, they are in the position to decide what and whose resources should be recognized and capitalized through state-funded education, and to what purposes. As illustrated by the case of MTI, this labor of ‘managing’ and ‘handling’ multilingualism is extensively achieved through state-backed language policy enacted in the educational field, where linguistic assets gain salience and value (Heller 2010; Heller and Martin-Jones 2001). However, as we have aimed to highlight in this article, state policy is the product of historical struggles in and between fields, and it may be at variance between national systems of education. Through the lens of Bourdieu’s (e.g., 1994) sociology of the state, what we prima facie may think of as state-mandated MTI policy is in fact the product of cross-field effects (Lingard et al. 2005), forged as a result of a historical struggle in which a diverse range of agents intervene: politicians, of course, but also university academics.

Viewed through this lens, we posit that Swedish–Danish differences in MTI policy reflect the differing ways in which the political field vis-à-vis the academic field of each country has managed to impose its visions on the bureaucratic field and the policies it produces. Ultimately, the outcome of such struggles accounts for the fact that while the Swedish discussion on MTI may be characterized as having been markedly structured by interventions from the academic field, the Danish discussion has remained a matter of ‘purely’ political consideration. Ergo, the existing divergences concerning MTI in Sweden and Denmark reflect differences relating to the order of the field of power of each society, where distinct forms of field-specific capital have dominated in the two societies at particular points in time. In Sweden, a relative lack of party-political opposition in immigration-related policy has allowed agents operating from within the academic field to continuously provide impetus to MTI, by creating a cross-field effect on the policies engendered by and consecrated within the field of power, often through the long-established commission practice and, thus, with the consent of political agents. These conditions have been profitable for the maintenance of a strong ‘left hand of the state’—that is, a social sector that ‘safeguards the interests of the dominated’ (Bourdieu 2003: 34–35; see also Lingard et al. 2005). By the same logic, it seems clear that the relative lack of successful pro-multicultural academic interventions has weakened the left hand of the Danish state. In Denmark, immigration was a stake in the political field early on, a fact that has led to a differentiated political field where MTI has not been broadly supported. As a result, its anchorage in legal frameworks has remained comparably weak, and in contrast to their Swedish colleagues, Danish academics have not been well-positioned within the group of agents who have ‘had a say in this series of decisions’ (Bourdieu 2014: 17) leading up to the organization of MTI as currently observable. Multicultural position-takings of the academic field have been curtained from the state field where organizational decisions on MTI are made, which is to say that, relationally speaking, academic forms of capital have been granted a low degree of significance relative to MTI policy within the Danish field of power.

What we can learn from these two comparable, yet distinct, cases is that the legitimacy of policies, as well as their scope and content, is the result of historical struggles over symbolic power waged between agents of various fields. Thus, following Bourdieu (2014), we argue that any language policy points to past battles and treaties in the field of power, and therefore reflects the historical outcome of cross-field effects within the field of struggle where state legitimacy is granted and the exchange rate between different kinds of capital is set. Seen through this lens, MTI is a useful object for scholarly approaches to language policy, not least as an indicator of the ways in which national systems of education respond to multilingualism and accrue symbolic value to diversity more generally. It duly illustrates that one of the things that education does is adjust the exchange rates that regulate the relative value of linguistic resources on the market, so as to determine the relative hierarchy of legitimization of different languages and, in effect, the social worth of different speakers.