Nordic language policies for higher education and their multi-layered motivations

Language policies have been drafted in Nordic higher education with the obvious, but unproblematised and unchallenged motivation caused by internationalisation. In this article, we analyse the various motivations for drafting language policies in Nordic higher education and the ideological implications of those motivations. We do this by approaching the question from multiple (macro, meso and micro) viewpoints, in order to make visible some of the undercurrents in higher education language policy. We are particularly interested in the explicit motivations for language policy change, and the explicit and implicit actors and action represented in our data. We will first discuss the background for internationalisation in Nordic higher education and then move on to our analysis of policy documents, survey data on the motivations for language policy drafting in Nordic higher education institutions. Our results indicate that internationalisation turns into a national question in the motivations. It also appears that the institutions are reactive (rather than active) in responding to perceived needs to draft a language policy.

One of the major responses of non-Anglophone countries has been to offer English language study programs, in order to deal with what Hughes (2008) has termed Anglophone asymmetry, i.e. the domination of English-speaking countries in the international student market. These new stand-alone Master's programmes created a new entry point to European higher education systems for students from third countries (Brenn-White and van Rest 2012). They also strengthened the base of the international cooperation networks, which originated in a post-WW II aim of supporting peaceful coexistence between nations by increasing intercultural understanding. In recent decades, as Hultgren et al. (2014) point out, internationalisation has generally been linked with marketisation of education, which in turn has been construed as forming something of a threat to the traditional welfare ideals of the Nordic countries.
Thus, internationalisation and processes related to it challenge Nordic higher education not only from the point of view of the language aspect (national languages versus English) of higher education institutions, but, from the Nordic perspective, also from the fundamental perspective of universities and other higher education institutions as national institutions, providing a public service in a globalising world (see Marginson 2006 for a discussion of the national-global interface in the Australasian context). Increasing international cooperation further increases pressures towards language of tuition, as English increases its share in a situation previously dominated by the local Nordic languages. Here, as recent studies show (see, for instance, articles in Hultgren et al. 2014 or the thematic issue 216, 2012 of International Journal for the Sociology of Language on internationalisation of higher education), the Nordic countries have solved the issues of language in higher education in different ways, resulting in different kinds of dynamics.
While this article and the whole special issue at hand (see also Airey 2012;Fabricius et al. 2016;Airey et al. 2015;Hellekjaer and Fairway 2015;Ljosland 2015; and the introduction by Saarinen) concentrate on the Nordic countries, the approach of focusing on implications of language usage in higher education policies is by no means limited to that geographical context: similar trends are seen not only in globalising Western countries (see Doiz et al. 2013) but also in the expanding Asian and South American contexts (Marginson 2006;Lim and Low 2009).
In this article, we focus on language policies and strategies of higher education institutions in the Nordic countries. Language policy has been understood in different ways that reflect the position of language policy both as an approach to understanding the functions of language in society, as well as a research field. For instance, Kaplan and Baldauf (1997) see language policy as a collection of ideas, regulations and practices that steer the change of language conditions in societies. While Kaplan and Baldauf recognise the existence of formal and informal inputs to policy, they primarily see language policy as enacted by a government or ''other authoritative body or person'', i.e. as authoritative (Kaplan and Baldauf 1997, xi). Spolsky (2004:5), in turn, sees language policy of a speech community as composed of three components: the language practices, language beliefs (ideologies) and language management of the said community. McCarty (2011) expands the understanding of policy by including the elements of ''human interaction, negotiation, and production mediated by relations of power'' (McCarty 2011, 8), thus taking a more dialogical and networked approach to the question, and influenced by ethnographic approaches. Johnson (2013, 9) synthesises language policy in a way that introduces a layered nature to the process of language policy, in defining language policy as official regulations; unofficial and implicit mechanisms; processes of multi-layered policy creation, interpretation, appropriation and instantiation; and policy as discursive and ideological (Johnson 2013, 9).
We approach policy in a manner close to Johnson's (2013) definition, by asking our questions from multiple viewpoints, in order to make visible some of the ideological undercurrents in higher education language policy. We thus acknowledge the multi-sited (Halonen et al. 2015) nature of policy making, where policy needs to be conceptualised in different, historically and contemporarily emerging layers, in order to understand the nature of policy change. In this article, we analyse the various motivations for drafting language policies in Nordic higher education and the ideological implications of those motivations. We do this by analysing the explicitly stated motivations for language policy change, and the actors represented in our data. We will first discuss the background for internationalisation in Nordic higher education and then move on to our analysis of legislation, national and institutional policy documents, and survey data on the motivations for language policy drafting in Nordic higher education institutions, in order to form a multilevel understanding of the language policy situation and the language ideological implications of this. The research questions and data are presented more explicitly in the chapter on Purposes and data.

Language policy making in Nordic higher education
Recent literature links language policy development in Nordic higher education inexplicably to internationalisation (Gregersen 2012;Lauridsen 2013;articles in Hultgren et al. 2014). This is easily understood: increases in large scale and systematic internationalisation since the 1980s (Nokkala 2007), in the form of more systematic exchange programmes or international research cooperation, have made contacts between the national language(s) and language(s) of internationalisation (the latter often unproblematically presented as English; Saarinen 2012) visible in a new way.
Internationalisation of higher education can be conceptualised as a macro (changing global environments), meso (integrating an international dimension to higher education activities) and micro (international publications, staff and student mobility) activity (Frølich 2006). The individual micro-aspects of internationalisation have occurred over centuries with differing intensities, with changes in international policy having had different effects on the mobility of individuals, as witnessed by changes in mobility volumes after the First World War and the directions of student mobility flows after the Second World War. It seems, however, that the position of language(s) in internationalisation has not been problematised explicitly at any level until recently (see, for instance, Saarinen 2012; Doiz et al. 2013). A recent survey on the role of English in the International Master's Programmes in Europe (Lam and Maiworm 2014) shows an increased awareness of the need to focus on language and not on the core subject curriculum only. Airey (2012) links the language question into the debate on academic literacies and emphasises that the disciplinary knowledge structures and the epistemological foundations determine to an extent the language(s) the students need in order to understand the disciplinary content on any deeper level (see also Kuteeva and Airey 2014). Also, a growing understanding of language as not a monolithic entity but a localised and situational element in the learning process requires acknowledging language(s) in the explicit learning outcomes (see Pennycook 2010).
These perspectives combined with the international students' varying educational and linguistic backgrounds inevitably highlight the need to focus on the role of language(s) in higher education policy in a new way, as language cannot be seen merely as a communications tool. Language is not monolithic, and consequently monolithic language policies do High Educ (2017) 73:597-612 599 not work in global higher education. ''One-size-policy'' clearly does not fit all students, programmes or contexts (Kuteeva and Airey 2014; see also Airey et al. 2015).

Trends in internationalisation of Nordic higher education and the position of language
Nordic higher education institutions have initiated internationalisation programmes since the 1980s, in response to European policies for increasing staff and student exchanges. While the first internationalisation policies were an extension of the micro-level mobility schemes to a policy level, by the 1990s and particularly the 2000s, internationalisation had become a strategic, macro-level issue (also) for the Nordic nation state in global economy (Haapakorpi and Saarinen 2014). Particularly in the small language areas such as the Nordic countries, the number of English language study programmes has increased as an attempt to attract international students. (Salö 2010;Garam 2009;Hilmarsson-Dunn and Kristinsson 2010;Schwach 2009;Wächter and Maiworm 2008). According to Wächter and Maiworm's (2008) survey, Northern European countries (with Southern European Cyprus and Central European Slovakia breaking the pattern) seemed to dominate the charts of proportions of programmes provided in English. Finland and Denmark were on top of the list together with the Netherlands, followed by Slovakia, Norway and Sweden. In a more recent survey by Brenn-White and Faethe (2013), based on data from StudyPortals.eu, Nordic countries are among the top ten of European countries organising English-speaking Master's programmes, both in terms of absolute numbers of programmes and in terms of rate of growth during the last couple of years (Brenn-White and Faethe 2013).
The steady rise in the number of international students (see Nordic Council of Ministers 2013) has gradually lead to an interest in regulating language policies for higher education institutions. An indication of this is a joint Declaration on a Nordic Language Policy (Nordiska ministerrådet 2007), which, while not legally binding, steers Nordic higher education institutions towards developing language strategies in the spirit of the apparently practical but problematic concept of parallel language use (see also Airey et al. 2015), originally introduced by Davidsen-Nielsen (2008) in 2002. According to Thøgersen (2010), the term was initially welcomed by universities and policy makers as a seemingly functional solution when balancing the use of the national language(s) and English in higher education institutions. The idea was that the national language and English would be treated as two equal languages, used pragmatically in the most convenient ways in different situations. However, Hultgren (2014a) found in Denmark that applying the ''parallel language policy'' tended to emphasise the use of Danish on the state level and English on the institutional level. Further, as Thøgersen points out, the local language and the language of internationalisation do not have an equal status simply because not all people have access to both or all of the languages. In addition, concerns have been voiced about the equal legislative rights of employees and students regarding access to study places and job recruitment. All these points link language issues to questions of equal participation in university processes.

Purpose and data of article
The purpose of this article is to analyse the motivations for and actors presented in language policy construction in Nordic higher education. As Hornberger (2006) has noted, formal language policies reflect different (historical and contemporary) political and ideological tensions and relationships, which have wider connotations than those strictly within the domain of language policy. In the case of the Nordic countries, the discussion of higher education language policies inevitably also includes their (common and separate) histories, as well as their educational traditions. In order to grasp these political and ideological underpinnings, we have asked the following questions: 1. How is the need for a language policy motivated in legislation, institutional documents and by local actors? What motivations are explicitly mentioned in the documents? What motivations are textually implied? 2. What language policy actors are presented as having a role in language policy construction? What potential actors are absent?
With the first question, we wanted to know whether either an internal or an external event or chain of events creating a need for a policy was construed, either in the documents or in the survey. In other words, we were interested in finding out how implementing a policy was motivated not only in legislation and macro-level documents, but also in the institutions. The second question outlines and examines what actors were presented as active and which ones were missing. From the social point of view, hiding some actor or group of actors can be ideologically at least as significant as mentioning another. Van Leeuwen's (1996) categorisation of social actors allows for a wide range of actors, including both human and non-human, to be acknowledged. Thus, political (non-human) concepts such as ''quality'' or ''globalisation'' can be presented as active forces in a policy process alongside more traditional (human) actors such as administrators and politicians. Thus, based on research on internationalisation and language policies as reviewed earlier in this article, we expected to see at least university administrators, national (political and administrative) bodies, (language) legislation, economy, internationalisation and institutional-level strategies presented as actors.
To answer the questions, we have used data from different (macro, meso and micro) levels of higher education, as we believe that a multi-layered approach (Halonen et al. 2015) is needed to answer our questions. The data consist of language and higher education legislation from the Nordic countries, as well as of national and institutional policy documentation on internationalisation and language policy. A survey was conducted in October and November of 2011 with an electronic questionnaire that was sent to all universities and other institutions of higher education in the Nordic countries (N = 151). The Webropol questionnaire was made available in Finnish, Swedish and English. It was sent to those responsible for the international activities of the universities, as our initial observation, based on a survey of institutional websites, was that language policies were mainly initiated because of demands of internationalisation. However, in order to have the best possible expert answer the survey, the recipients were encouraged to send the link to a person more suitable for answering if they felt that the survey was out of their expertise. After one round of reminders, we received altogether 53 answers, which means a response rate of approximately 35 per cent.
To analyse the replies, we conducted a content analysis of the open-ended answers and the language policy documents, focusing on arguments for internationalisation and for implementing language policies and the actors related to language policy work. Some quantitative analysis of the survey was conducted to identify reply categories for our own purposes, but in the current article, the analysis is mainly qualitative, based on coding and classifying the open-ended replies.
In addition to the questionnaire survey, website data (institutional language policies and internationalisation policies) have been collected from a purposefully selected combination of 1-2 universities and 1-2 other institutions of higher education in each country (N = 11). The documents were selected purposefully from institutions that reported having language policy documents and linked their documents in the electronic survey. The selection is not, in other words, in any way representative of all Nordic institutions of higher education, but rather collected to illuminate the various developments in the field (''Appendix'').
This holistic approach is challenging in the way that it is potentially superficial in attempting to cover as wide a range of different levels of data. Heller (2011, 192) has characterised this kind of analysis as ''ethnography of discursive shift'', describing the analysis as discursive work ''in spaces that often need to be constructed out of leftover interstices, taking form over here while everyone else is looking over there''. The elusive policies that we have analysed call for a multi-sited and multi-layered (see Halonen et al. 2015) analysis, i.e. the acknowledgement that policy discourses often take place on a number of potentially interconnected situations simultaneously, which, in turn, creates a potential for clashing interpretations of the political reality. The article offers a previously unexplored view into internationalisation of higher education from the point of view of language.

Multi-layered perspectives on language policies for Nordic Higher Education
Legislative language policy steering in Nordic countries and higher education institutions The Nordic countries share a history of particular kinds of political values that promote an existence of a welfare state and egalitarian ideals (see Hultgren et al. 2014); on the other hand, Nordic language policies have taken shape based on very different kinds of historical developments. This chapter analyses the legislative and other formal steering documents for language policy making in the Nordic countries in general and higher education in particular.
While the Nordic countries all seem to have reacted to the changes in the global higher education market by increasing English language educational offerings, their legislative backgrounds differ quite significantly. Denmark does not have an explicit language legislation, which, according to Siiner (2012) has led to a situation where national legislative steering is based on covert input from other sectors, such as integration and primary education sectors. Sweden, in turn, got her first language act in 2009, stating explicitly for the first time that Swedish is the official language in Sweden, as well as recognising five official minority languages. Finland is constitutionally a bilingual (Finnish and Swedish) country, which had its first Language Act in 1922, with a renewal in 2004. Iceland introduced a new Act on the status of the Icelandic language and Icelandic sign language in 2011. Bilingual (Nynorsk and Bokmål) Norway has no specific language law, but the status of languages is defined in separate legislation with respect to, for instance, usage of languages in civil service, or the status of the Sámi languages.
In short, the legislative basis in the Nordic countries varies on a continuum from no language legislation (Denmark) to historically formed, constitutionally defined and legislatively binding state bilingualism (Finland).
However, regardless of legislation, all Nordic countries have introduced national language policies and strategies in some form, which reflect on the situation of the national language(s) under pressures of globalisation. These include documents such as Sprog på Spilet udspil til en dansk sprogpolitik 2003 (Denmark); Bästa språketen samlad svensk språkpolitik 2005 (Sweden); Mål og meining 2008 (Norway); and Kansalliskielistrategian väliraportti (2015) from Finland. It appears that while the legislative base differs, the countries have in common a need to define, systematise, and ultimately secure the position of the national language(s).
At the university legislation level, the situation is equally heterogeneous. Denmark currently has no legislative regulation at the university legislation level; closest to national regulations are recommendations drawn by the Danish Rectors' Conference (Rektorkollegiet 2004). Swedish higher education legislation does not regulate on language(s) in higher education, and the new national language legislation does not explicitly define languages of tuition for higher education institutions. However, an Ombudsman's decision from 2009 (Justitieombudsmannen 1811/2008) declares that the Language Act only applies to the core functions of universities (decisions, meeting protocols and other official documentation), which should take place in Swedish; thus, the Act does not extend to the language of tuition and publication at universities (Salö 2010). In Finnish higher education legislation, the development since the turn of the Millennium has gone in the opposite direction from the rest of the Nordic countries, increasing the higher education institutions' own powers and possibilities for using other languages in addition to Finnish and Swedish. Recent decisions by the Chancellor of Justice's office (see Saarinen 2014) have, however, stated that universities would need to be more explicit about their foreign language tuition: while universities can offer tuition and degrees in other languages than the official Finnish and Swedish, the students would have to be aware of their constitutional right to take examinations and essays in the national languages. The Icelandic higher education legislation from 2006 does not regulate languages explicitly (although it mentions the task of the universities in strengthening the infrastructure of Icelandic society), but as described above, the new 2011 language legislation covers all levels of education (Kristinsson and Bernharðsson 2013). In Norway, a mention of languages was added to university legislation in 2009 in the form of universities and higher education institutions having responsibility to maintain and develop Norwegian terminology.
In sum, on the national level, the ethos of ''safeguarding'' the national language(s) in higher education is relatively strong, regardless of the legislative status of languages.
On the institutional level, Nordic universities and other higher education institutions have increasingly started to draft and implement explicit language policies and strategies since the 2000s. One reason for this appears to be the increasing mobility since the degree reforms spurred by the Bologna process. Formal language policies may also be a reaction to the Nordic language policy declaration (Nordiska Ministerrådet 2007).
Danish universities and colleges tend to have language policies in some form, which in practice recommend parallel use of Danish and English (with or without committing to that controversial term also promoted in the Nordic language policy declaration; see Hultgren 2014b). In Sweden, the National Agency for Higher Education (Högskoleverket) gave a recommendation in 2008 that higher education institutions should draft their language policies, but according to Salö and Josephson (2014), only approximately 15 universities or higher education institutions out of a total of 26 institutions have done so. In Finland, in January 2014, 12 out of 14 universities have explicit language strategies, as required by the Council of State Decision in 2007. Six out of seven Icelandic universities (Kristinsson and Bernharðsson 2013) have explicit language policies, mainly starting with the premise of Icelandic as the principal language of the universities. Instruction in English should mostly be confined to the graduate level, and the importance of Icelandic terminology is emphasised as well as exquisite usage of the Icelandic language. Norwegian higher education institutions have had a similar recommendation by the The Norwegian Association of Higher Education Institutions, but it seems that about one in five institutions have a strategy at the moment, and additionally some have covered language policy issues in their other strategies such as internationalisation policies (Kristoffersen et al. 2014). Table 1 summarises the situation of legislative and strategic language steering at Nordic higher education institutions.
Why institutional language policies? Findings from the survey Nordic universities and HEIs increasingly seem to produce language policies or language strategies either as separate, explicit strategy documents and policies, or as part of another (strategy) document (see also Salö 2010). About half of the institutions that answered the questionnaire had a language policy or were in the process of drafting one.
In this chapter, we analyse the survey results from the point of view of motivations for language policies.
One of the questions in the survey was on motivation for creating a language policy or guidelines for language use. The arguments fell roughly into three categories: (1) legislation or ministerial incentive, i.e. governmental actor seen as significant; (2) regulation of the relationship between the national language(s) and English; or (3) increasing demands for more structured internationalisation (international staff, international students, international study programmes, international labour market for graduating students), especially by higher education administration actors at national and institutional level.
The first category can be characterised by statements like it is a self-evident thing, the ministry requires a local university language policy to complement the national language strategy (Swedish university) where the local needs are embedded into a document that has been required by national-or institutional-level actors (i.e. ministries or university administration) for external reasons.
The second category has a more emotional tone of safeguarding the national language on the one hand, and presenting English as a language with an undisputed position on the other: Kravet på [university] att leva upp till rollen som ett nationellt och internationellt universitet, där svenska och engelska används parallellt… [The requirement for the university to live up to its role as a national and international university, where Swedish and English are used in a parallel fashion.] Category three, in turn, is a more practical category where internationalisation is presented as an actor that prompts a need for clearer language policy guidelines, rules and regulations for action. A simple statement from one of the respondents rätt språk i rätt sammanhang [right language in the right situation; Swedish university] resonates well with this idea, which turns the ideological question of language choice in higher education into a practical one. This observation is supported by Björkman's (2014) finding of the choice of English being promoted on ''practical'' grounds.
Categories two and three follow the idea of parallel languages, a need to create an environment where English and the national language exist side by side. This approach seemed to be part of the language policy discourse, especially in Danish and Swedish institutions. In sum, the respondents mostly seem to construe language policies as a reaction to perceived current needs.
The replies on the question on institutional strategic steering are largely focused on language skills of the international staff, on offering of courses in other languages (in other words in English) and on student recruitment. An example from a survey answer from a Danish university exemplifies this: Yes, it is necessary for our foreign researchers to learn Danish in order to become well-integrated in the society. Furthermore, it is necessary for the Danish speaking employees to have a certain level of English knowledge as a lot of the communication with the researcher and students are in English (Danish university).
The strategic steering comes in the form of the political interest in attracting more international students to the Nordic higher education institutes. Replies such as Yes, a greater demand and an emphasis on recruiting more international students will lead to this (a Norwegian higher education institution) illustrate this view. The negative replies reflect mostly a satisfaction with the status quo: No specific need, international study programs and study units are enough for active student exchange and double degree agreements. (A Finnish university of applied sciences). Some respondents felt that the current amount is enough: Yes, probably. But personally I feel the current steering is enough (A Swedish university). These replies imply that the current mobility programmes and funding structures for both teacher and researcher mobility have become an established component of internationalisation activities.
In our survey, the respondents brought up future expectations only when explicitly asked about them. Approaches to the future are somewhat neutral, and the replies do not appear to reflect a vision of an active steering of it. Increasing internationalisation and the requirements it puts to languages are often represented as self-evident: internationalisation and globalisation are presented as actors pushing to an increased number of international degree programmes. Some caution could, however, be detected in the replies, as the following excerpt shows: Study programme development: more master programmes, more PhD programmes entirely, or partly, taught in English is a challenge for the staff, as the institutional policy is to increase of programmes taught in English. (a Norwegian higher education institution).
In some answers, future (and current) changes in higher education policy were stated as overriding language policies, as in the following example: The current transformation of the universities of applied sciences also affects the language policies so this point in time is High Educ (2017) 73:597-612 605 not the best for any changes or new thinking (a Finnish university of applied sciences). In Finnish higher education, structural developments and the legislative position of Swedish were stated as such overriding actors. Interestingly, such ideas on the predominance of large-scale policy changes were presented also when such policies did not, in fact, exist. An example of this was a Swedish response, stating that language policies were drafted because of a top-down legislative demand, a requirement that does not exist as such.
Responses to motivation and strategic steering of language policies appear reactive, i.e. construed as reactions to perceived needs. Consequently, the policies may turn out to be collections of individual language policy requirements rather than systematic and consistent policies. Ideologically, this leads to a policy that is not explicitly drafted in a democratic process.

Higher education institutions as language policy actors: the institutional policies
Next, we will look at some selected case institutions and their language policies. The selection of cases was explained in the previous chapter.
A content analysis of a selected number of language policies (1-4 from each country, both from universities and other HEIs, 11 in all; see ''Appendix'') gave five main categories with which language policies were argued for.
The relationship between the national language and the international language (with few exceptions English) was a major concern in all the language policies that we analysed. In cases from Denmark, Sweden and Finland, this relationship was either explicitly (Copenhagen University in Denmark, KTH Royal Institute of Technology in Sweden, Aalto University in Finland) or implicitly (University of Jyväskylä, Finland) described in terms of parallel languages: the right language for the right context. There were, however, differences in the presentation of this principle. At the University of Copenhagen, parallel language policy seemed to be drafted to secure the availability of English rather than Danish at the University, for the purposes of internationalisation. (Copenhagen University; see also Hultgren 2014a, b). KTH Royal Institute of Technology, on the other hand, presents the principle of right language in the right context in rather neutral terms. University of Jyväskylä, while not explicitly discussing the notion of parallel languages, apparently balances between the positions of Finnish and English in the everyday activities of the university.
A specific case in construing the relationship between the national and the international is the discourse of protecting and safeguarding the national language and culture. While, for instance, University of Jyväskylä highlights its monolingual history and tradition in its motivations, Aalto University similarly comments on its bilingual traditions, when stating needs of internationalisation and multiculturalism as motivations for a language policy. This ideological balancing, as well as the reference to internationalisation without the weakening of national languages (Aalto University), may reflect the need to base language policy arguments of bilingualism on the Constitution .
Another example of stressing the national over the international is provided by the policy of the University of Iceland for 2011-2016, which also particularly stresses the University's commitment to Icelandic society, culture and language. On the other hand, while Icelandic is the principal language of the university, English is explicated, in the context of human resources and internationalisation, as a de facto second language in administration. The policy also encourages learning of Icelandic for international students and staff, and an English test for incoming Master's and doctoral students. The language policy of the university from 2004 is available only in Icelandic, which is in itself emblematic of the strong position of the Icelandic language in it.
An interesting side note is that the university language policies in Norway do not differentiate between the two national languages nynorsk and bokmål but instead only mention Norwegian. It is possible that this conceptualisation is, in itself, a symptom of Anglicisation, as internal differences get diluted. 1 Another prominent category in the language policies was the discussion of the position of multilingualism in general or languages other than English or the national language(s) in particular. This was often done in a very general manner, either as explicit recommendations to promote languages other than English, or as more vague statements of encouraging multilingualism, where the context implied that multilingualism included not only English but other languages as well. An example of this is the language policy of the University of Jyväskylä, which uses the concept of dynamic multilingualism when developing multilingual activities in all operations of the university, in order to increase the language and culture sensitivity of students and staff.
An interesting exception to the Anglocentric view of multilingualism is provided by the University College Sjaelland (UCSJ, Denmark), which has an explicit policy of promoting German alongside English as the foreign language. This may be linked to the regional role of the UCSJ, as stated in its strategy. Additionally, English is presented as a possibility rather than a self-evident requirement. This together with the focus on competences in Danish as the first goal thus seems to suggest that from the point of view of language policies, the UCSJ has a local and regional profile.
Some documents expanded the discussion of languages as individual languages to cover different functions of language, ultimately leading into a discussion of our third category, view of language. The University of Jyväskylä language policy is based on an idea of language as communicative and interactive competence (rather than from the definition of in which context different languages are to be used, which is often the case with language policies). While the University as a working environment takes as its implicit starting point the internationalisation of recruitment, the University as a study environment section, in turn, starts with the functions of language in knowledge building and continues with emphasis on the Finnish language and culture. Particularly the aspect of knowledge production is rare in university language policies (see also Kuteeva and Airey 2014).
The official policy document of Dalarna University (Sweden) clearly states that staff and students alike are equally responsible for the language use at the Högskolan. In particular, the written academic texts need to be of a very high quality when it comes to clarity, reader friendliness and unambiguousness. The university also maintains that it will increase awareness of the meaning of language and of language competence development that the students undergo during their degree studies. The documents seems to seek a balance between what is done inside the institution in relation to language and with how the reputation of the institution is nurtured and protected.
The goal of the language policy guideline of the Norwegian University of Life Sciences (UMB) is presented in clear and simple terms: increased linguistic and cultural awareness. This is accomplished by raising awareness regarding language use at UMB.
The fourth category that emerged from the HEIs was the discussion of the position of English in the language policy. For instance, the Seinäjoki University of Applied Sciences (Finland) general strategy mentions languages in four contexts: twice in the context of English language degree programmes and twice as internationalisation goals are defined (in the latter case, also, the goals are linked with increasing English language tuition). As the general strategy links language with internationalisation, we also looked into the international strategy of the institution. In it, the position of foreign language (implied as English, see Saarinen 2012) tuition is even more pronounced than in the general strategy; of the twenty mentions of language, seventeen refer to foreign language teaching, two to teaching of Finnish to international students, and one generally to the English language. Internationalisation is, in other words, mainly operationalised as English language degree programmes and other English language tuition, making English a practical rather than a political choice.
A fifth category in the language policies referred to a phenomenon known as internationalisation at home (Nilsson 2000). The principle refers to the institutional purpose of internationalisation as offering local students a chance to get international experiences at the home institution. The language policy of Kungliga tekniska högskolan proposes an approach to English language tuition as a progressive process: the undergraduate-level students are expected to function mostly through Swedish and have only passive knowledge of scientific English. The master's level studies include an idea of parallel language where the students show an increased competence of English. On doctorate level, the students use both Swedish and English fluently. Similarly, at UMB the first-degree-level teaching is mainly carried through Norwegian and the master and doctorate levels both English and Norwegian should be used as language of instruction. This continuum implies the idea that the students know Swedish well enough to begin studying in it, i.e. that they are local. The document does not state how this progress is supported in the curriculum; it may well be that it is supposed to be a natural and expected progression in the scientific disciplines.

Conclusions and discussion
Internationalisation emerges as the primary motivation in drafting and implementing language policies. The outcome may be a relaxing of language policy steering (as is the case in Finland) or a tightening protection of the national languages (as, for instance, in Sweden and Iceland). As internationalisation is operationalised in the strategies and our survey as (mostly incoming, in fewer cases outgoing) mobility of students and staff, language policies become operationalised as regulations on usage of English at universities and the relationship of English to the national language(s). This makes visible the ideological tensions between the national and the international roles of higher education institutions.
Is there, then, a Nordic language policy, and how could it be characterised? Legislatively, the backgrounds of each Nordic country vary considerably. Finland (with a strong historical tradition of constitutional bilingualism and the ensuing language legislation) and Denmark (no national language legislation) would provide the ends of the continuum, with the remaining Nordic countries having different historical and contemporary solutions to the language question. Also in terms of university language legislation, there are different ideological solutions, with Finland having moved towards more relaxed language steering, and countries like Norway and Iceland becoming more protectionist towards the national language. At the university level, there seems to be a similar need to draw language policies and other guidelines for language use. In these, the outcome is often a pragmatic solution of how to best organise the parallel use of the national languages and English, although without explicitly using the term parallel languages.
When language is mentioned in the context of medium of instruction, it is done in an instrumental context of enabling mobility rather than in a knowledge-intensive context of problematising knowledge production (see also Kuteeva and Airey 2014). From the point of view of language, this development is problematic. Kuteeva and Airey (2014) noted in their analysis of disciplinary differences in the use of English in higher education that providing generalised pragmatic policy steering on language use tends to overlook disciplinary differences in language and literacy practices. We suggest this may have unintended consequences for disciplinary knowledge production practices and may also impact research paradigms, as different language publication is likely to be directed at different audiences; typically national languages may be used for professional audiences and international languages for academic readers. Presenting particular language policies that have their explicit ideological impetus in internationalisation may have unintended effects elsewhere (see also Airey et al. 2015).
The role of language policies, then, seems to link to the organisation of international mobility rather than questions of teaching and learning. An interesting feature in the survey and the policy documents is that global and national economies do not appear as motivation in internationalisation, and consequently, language policies. Frølich (2006) noted a similar feature in her study of internationalisation of research: even if economic arguments prevail in higher education policy (see also Saarinen 2008), in internationalisation policies they appear to take place slowly.
It appears that language policies in (Nordic) higher education are construed in complex interaction between different actors and often competing ideologies, stressing the need for a multi-sited and multi-layered approach. While the state-level legislation and other data in the Nordic countries appear to reflect national needs, the institutional policies promote an ideology of internationalisation in English. Making English the ''practical'' choice (see Björkman 2014) hides the ideological and political aspects of language policy. Internationalisation, in fact, becomes a national question, when it is operationalised into how international students and staff are to be integrated (either by use of English or the national language) into the national higher education system, or how the national language is to be safeguarded against the flow of English.
In Finland, the institutions reported various national policy issues (language policy issues, higher education policy reforms, etc.) having a bearing on drafting a language policy, even if internationalisation would be stated as the prime motivation. In Denmark, and to some extent Sweden, there appeared to be more mentions of the needs of national language instruction in higher education, again referring to the tensions between motivations to internationalise and to the requirements of the national setting. In all cases, language policy drafting seems reactive rather than active: language policies are construed as a reaction to perceived external rather than internal needs.