Maarja Siiner a a Institute of Journalism and Communication, University of Tartu, Tartu, Estonia Published online: 26 Feb 2014.

Transcription

1 This article was downloaded by: [ ] On: 28 February 2014, At: 06:48 Publisher: Routledge Informa Ltd Registered in England and Wales Registered Number: Registered office: Mortimer House, Mortimer Street, London W1T 3JH, UK Journal of Multilingual and Multicultural Development Publication details, including instructions for authors and subscription information: Decentralisation and language policy: local municipalities' role in language education policies. Insights from Denmark and Estonia Maarja Siiner a a Institute of Journalism and Communication, University of Tartu, Tartu, Estonia Published online: 26 Feb To cite this article: Maarja Siiner (2014): Decentralisation and language policy: local municipalities' role in language education policies. Insights from Denmark and Estonia, Journal of Multilingual and Multicultural Development, DOI: / To link to this article: PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the Content ) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms &

2 Conditions of access and use can be found at

3 Journal of Multilingual and Multicultural Development, Decentralisation and language policy: local municipalities role in language education policies. Insights from Denmark and Estonia Maarja Siiner* Institute of Journalism and Communication, University of Tartu, Tartu, Estonia (Received 30 September 2013; accepted 25 January 2014) The present article contributes to attempts to re-conceptualise the top-down perspective on language policy, by analysing the role of local and city governments agency in language education policy making. Only few studies analyse the role of lower administrative levels in language policy, other than in implementation of governmental policies, why their policy appropriations are seen as policy deviations. Language policy researchers have however recognised that local governments can, due to the regional or local character of some language education problems, also be given a more active role in policy making. My claim is that in order to do so, the analysis has to be based on a conceptual framework that sees language policy developments as part of the general democratisation and decentralisation processes in society. The article therefore also problematises attempts to analyse and make sense of language policy developments separately from political and economic transitions of society. Based on Giddens structuration theory and language governance studies, I will analyse how two different language policy models in Estonia and Denmark in terms of allocation of resources and authority frame local municipalities opportunity for agency in language education policy matters. Keywords: language education policy; local government; decentralisation; Denmark; Estonia; LGR Introduction As the first point of contact between government and the general public, the municipal level, the lowest sub-national government level, is the place where linguistic problems have to be dealt with on a day-to-day basis. Most European Union (EU) countries have assigned to the sub-national levels at least some responsibility for preschool, primary and secondary education, as part of general decentralisation and democratisation trends, based on such arguments as efficiency and transparency in the delivery of public goods (Escolano et al. 2012). In his book on language policy (LP), Johnson discusses how city governments and other smaller administrative bodies can design language education policy (LEP), and by doing so, both in practice and in ideology, give a boost to, slow down, or modify national language plans (Johnson 2013, 237). Backhaus furthermore analyses LPs on lower administrative levels, based on an ethno/linguistic model, distinguishing between societies according to the number of linguistic groups present (Backhaus 2007). These actions can also be studied as part of the general democratic decentralisation process by focusing on how governments at the local level can become * maarja.siiner@ut.ee 2014 Taylor & Francis

4 2 M. Siiner more responsive to citizens desires and more effective in service delivery (Linnas 2011; Crook and Manor 1998). Local municipalities more active involvement in solving local problems is, furthermore, a tool for the political education of the population, contributing to the society s democratisation (Jaanson 1999, 19). The present article will provide examples from Denmark, with its long tradition of decentralisation and democratisation processes, and from the still transitional society of Estonia, where the roles and responsibilities of local municipalities are yet undecided. By analysing the agency of local governments, I hope to shed some light on language policy and planning (LPP) challenges in the broader infrastructural context of the democratisation and decentralisation of power and public services in a state, and to contribute in a technically more substantive way to LP analysis. My aim is also to contribute to the discussion of whether the sustainability of language policy (in terms of flexibility in the infrastructure needed to meet the changing needs of language users in the globalising world) can be achieved separately from the sustainability of the whole society in terms of improving the welfare of the people and the coherence of social structures and institutions (Siiner and Vihalemm 2013; SE , 31). In the next section, I will introduce the conceptual basis for my paper and the analytical paradigm applied in my analysis of LG involvement in LEP, based on ethnolinguistics, Giddens structuration theory (Giddens 1984) and the theory of language governance (Walsh 2012; Drechsler 2004). Then I will briefly explain why an analysis of LGs involvement in LP issues is important and why I have chosen to analyse these two cases to exemplify my points. In the fourth section, I will outline the main differences in LP and public administration models in Denmark and Estonia. In the fifth section, I will apply the analytical paradigm introduced in my analysis of the Tallinn and Copenhagen city councils involvement in LEP. I will finish with a discussion of the consequences of weak vs. strong involvement of local municipalities and the pros and cons of decentralisation in the last section of the paper. The interdisciplinary nature of LP as a field of study The role of LG in LP research has received only limited attention, mainly due to the lack of a conceptual basis and methodological framework, although in many cases LP researchers have recognised the central role of LGs in public education (Spolsky 2004, 53). In addition, other prominent LP researchers recognise that language policy is no longer only about managing linguistic diversity, but also about cohesion, civil rights, democratisation and decentralisation in a society (Jernudd and Nekvapil 2012). LP as a field thus extends beyond the margin of mono-disciplinary (linguistic or political) approaches, calling for a combination of tools and approaches from several disciplines (Jernudd and Nekvapil 2012, 17). Seeking an extended understanding of LP from an interdisciplinary angle, the present paper is embedded in a theoretical framework that goes beyond socio- and ethnolinguistics, reaching out into social science, where the focus on structure and agency is one of the main concerns. The conceptual framework is partly based on an understanding of LP as a multilayered and multi-agent process, as introduced in the ethnographic approach (McCarty 2011). Implementation (or appropriation ) is not just what happens after policy is made, but it is a link in the chain of the policy process in which different actors or agents potentially have input. The levels at or through which LP moves and develops are, in the ethnographic approach, roughly divided into the macro (state), meso (LG/ institutions) and micro (individual practitioners) levels (Hornberger and Johnson 2011, 275). The

5 Journal of Multilingual and Multicultural Development 3 dynamics of LP can be detected in language legislation, where regulations can open up free spaces (Hornberger and Johnson 2007) for necessary local initiatives, interpretations and adjustments, by either stating their role explicitly in policy texts or by leaving some areas unregulated. I will return to the importance of the free spaces later on. Since the ethnographic approach has dealt very little with the role of LGs in the policy-making process, I will include insights from other approaches. Some studies have already dealt with interactions between language regulations, practices and social structural surroundings, asking How do legal and institutional settings frame language (political) activities? (Siiner and Vihalemm 2013; Vihalemm and Hogan-Brun 2013; Öst 2012; Meiorg 2012). The present article is an attempt to shed more light on the role of LGs as part of social structures in LEP, combining theoretical insights from ethnolinguistics with insights from language governance (Loughlin and Williams 2007; Walsh 2012) and Giddens structuration theory (Giddens 1984). What makes a combination of these perspectives useful as a set of thinking tools (Bourdieu, cited in Thomas 2007, 83) in LP analysis is that they, on the one hand, offer a plausible technical explanation of the importance of involving agents on other than the state level in LPP, creating flexibility in LP design (Siiner 2012, 13). The structuration theory helps to explain how laws passed by national governments explicitly frame the possibility for agency in language political matters. Combined with insights from language governance studies and the ethnographic approach, it helps to understand how LP process is influenced by the interaction of local, regional, national and international actors, each seeking to achieve its own form of governance. This combined approach also makes it possible to examine the conflicts between lower and higher administrative units of a society participating in LP making (Backhaus 2012; Walsh 2012). Language governance studies borrow the term governance from political studies of public management and decentralisation processes; the involvement of sub-state administrative levels is here an important parameter in measuring a state s administrative capacity in solving language problems and can be analysed by how authority and resources are allocated on state and sub-state levels (Drechsler 2004, 388). LP in a state is, in language governance, seen as part of the broader context of public administration and thus of the democratisation and decentralisation processes (Walsh 2012). The division of LP responsibilities between local governments (LGs) and central government (CG) is also a question of the balance between trust by the state and the community in the LGs ability and commitment to solve ad hoc local problems, and control by the state in order to guarantee democracy (Hanberger 2009, 19), in which decentralisation issues are also issues of democratisation and accountability (Blair 2000). The main thrust of the argument of governance theorists (Blair 2000) is that as society becomes more complex and differentiated, the traditional method of governing from above becomes more difficult. This leads to governance understood as steering rather than directing, which, it is claimed, supplements or at times even replaces traditional government. In terms of language, governance is a complex, multifaceted concept. It occurs in the context of globalisation, operates at several levels (local, regional, national and supranational) and involves a myriad of organisations (national, regional and LG, state agencies, transnational corporations, non-governmental organisations and international bodies of governance). However, despite globalisation, national government retains a key role as a strategic site to unite the various infrastructures of governance (Loughlin and Williams 2007, ). Giddens analytical paradigm of allocative and authoritatively allocated resources is particularly useful in the analysis of language governance and the role of LGs, since it

6 4 M. Siiner makes it possible to distinguish between different types of actors in terms of their capability or potential for action and success (Giddens 1984, 258). In his structuration theory, Giddens illuminates the constitutive, although amenable to change, character of social structures. Giddens structure refers to a set of rules, which individual or collective social agents draw on to enact or change social practices. One of the central ideas in the structuration theory is social activity or agency, which is based on different social agents access to necessary resources in the production and reproduction of social practices (Giddens 1984). Giddens resources are of two types. Authoritative resources are the non-material resources drawn on in controlling and influencing the circumstances of other agents actions or in coordinating the activity of others, and allocative resources are connected with the harnessing of material resources, in terms of goods, technology and the environment. Laws, as an attempt to institutionalise some rules, make these resources more durable, an important tool in distributing resources, in granting some agents rather than others the power of getting things done, i.e. access to sufficient allocative and authoritative resources (Giddens 1984, 283). According to Giddens, laws that frame the possibility of agency, either by leaving the possibility of (legislative) intervention to other agents at lower levels of society or by leaving certain areas unregulated (Siiner 2012, 67), highlight the central role of the state in the distribution of resources and as a framing agency. The distribution of resources can also be implicit: where existing rules (i.e. laws passed at the state level) or the lack of rules leave the possibility of agency to other agents who might, or might not, have the necessary means to fulfil the task. Embedded in this interdisciplinary theoretical framework that goes beyond the ethnography of language in reaching out into social and political science, I will, with a focus on structure and agency and a concern for the distribution of resources, apply an analytical paradigm of allocatively and authoritatively allocated resources in my analysis of LGs role in LP matters. In order to do so, I have gone through relevant language political texts (including laws, regulations, development plans, policy reports and monitorings of policy developments) regulating the area of LEP in the basic and lower secondary schools in Denmark and Estonia. By applying critical textual analysis, my aim was to uncover the explicit or implicit distribution of allocative and authoritative resources to LGs in these texts, i.e. which agents are included and which excluded from assuming agency in solving LP problems. Furthermore, as an example of the dynamics of LP making, I analyse two attempts at agency by LGs, one successful and the other one unsuccessful, by the city council of Copenhagen and the city council of Tallinn. These attempts will be explained in relation to the LP models adopted in these countries and the social structures surrounding and affecting them. But I will begin with a brief explanation of why an analysis of the LG role in LEP is important. LGs involvement in LEP LGs, as the government tier closest to citizens, usually play a crucial role in decentralisation processes and increased accountability in a country. While most states tend only to become involved in language matters when they are intertwined with bigger and more important political issues, such as the economy, preservation of power and opposition to state enemies, where unitary solutions are needed (Ricento and Hornberger 1996, 414), LGs are often more capable of finding locally sensitive solutions. In addition, the involvement of LGs in supplying basic public (social) services to local citizens contributes to democratic governance and sustainable development (Echebarría 2001). Previous studies reveal that an LEP succeeds better if it incorporates the needs and wishes of a community, while a strictly

7 Journal of Multilingual and Multicultural Development 5 top-down policy could be met with resistance (Freeman 2004). Attempts by governments to intervene in local social situations without the consent of the affected parties are always problematic (Ricento and Hornberger 1996, 416). Offering an analysis of three municipal LPs in different parts of the world, Backhaus (2012, 240) reveals the potential for legal and ideological friction between municipalities (local policy makers) and higher administrative levels (central government), as local municipalities have to take much greater care than national governments to communicate with their residents and can, furthermore, be placed in the unclear position of being either policy implementers or political decision makers (Backhaus 2012, 226, 242). According to Rees famous principal-agent problem (Rees 1985), LGs can be hard to motivate to act in the best interests of the state and tend rather to act in their own and local voters interests (Cullis and Jones 1992, 322). Highlighting this ideological tension between CG and LGs as an attempt at local appropriation (Ricento and Hornberger 1996) can thus shed some light on the dynamics of LP making, while also contributing to policy analysis and development in a technically substantive way, as it highlights how the role of local municipalities in offering efficient public service and representing the community s needs is crucial not only to the democratisation and decentralisation processes, but also to defining and solving local language problems. My main focus here is on the consequences of a weak vs. strong involvement of LGs in LEP. In order to illustrate my points, I will analyse how LGs in two countries with different administrative systems and LP designs have taken steps to solve language educational problems. I have chosen Tallinn and Copenhagen for my case study, since they are both capitals of small member states of the EU (Estonia and Denmark, respectively), and thus face the same challenges of globalisation and multilingualism in education. Also, an interest in and attempts at comparison of Estonian and Scandinavian designs of public administration systems in terms of the division of tasks and responsibilities dates back to the beginning of the transition period in Estonia, at the beginning of the 90s (Jaanson 1999, 44). My aim, with the help of these contrastive cases, is to demonstrate how language policy development and success is dependent on the infrastructural constraints of societies. There are, however, considerable differences in historical, demographic, geographic and political factors between Denmark and Estonia, differences that have also shaped public administration, including the role, power and responsibilities of the local municipalities in the countries. These will be outlined in the next section, with an explanation of how they have influenced development of the LP models adopted. LP models in Denmark and Estonia Denmark Denmark represents the type of LP in which the government has not taken the symbolic act of ratifying the status of the official language as the sole principal language of the state. The language ideological debates (Blommaert 1999) over the need to protect Danish against the influence of English in the areas of science and research have taken place for decades, but have been complicated by the general positive attitude towards English, the fact that Danish both as a written and an oral language is not seen as being in a threatened position, and the prevailing language attitude that sanctioning and controlling the use of language is discriminatory (Sprogudvalget 2008). LPP activities in Denmark are however not so much non-existent as they are hidden, as many of the decisions concerning the use, acquisition and status of languages are influenced by

8 6 M. Siiner political decisions made on other topics, such as primary education and integration, dictated by changing governments hidden political agendas (Milani 2008). The changes in power constellations in 2001 led to major changes in the political agenda, leading to a general negative attitude towards immigrants and bilingualism (seen as obstacles rather than assets) in the official discourse and media (Holmen and Jørgensen 2010). Amendments were made to citizenship legislation entailing a tightening of restrictions for family reunions, including a demand of knowledge of the Danish language as a prerequisite for the granting of Danish citizenship. Amendments made to Folkeskoleloven (the Basic Schools Act) by the new government decreed that it was no longer obligatory for municipalities to offer free classes in the mother tongue to bilingual children, while the evaluation of Danish language competences for all children and offering necessary language stimulation in day care became obligatory (Statuary Notice of the Act on Pre- Primary Education 2011). Screening language competences was at the core of greater control of school achievement, which had not been obligatory previously in Danish primary schools or daycare institutions. The Danish education system has, at the preschool and basic school levels, traditionally been characterised by local involvement: the Minister of Education sets the target subjects and the guidelines for curricula, and organises leaving examinations (the examinations were optional until 2007), but municipalities and schools themselves decide how to reach the targets. However, due to generally strong decentralisation tendencies, the local municipalities have maintained their role as the main agents responsible for the language screening of bilingual children and for supplying extra classes in Danish as a second language if needed. The administrative reform in Denmark was carried out in 2007 as part of decentralisation trends, diminishing the number of municipalities by two-thirds, to 98, with between 20,000 and 500,000 inhabitants (Danmarks Statistik). Danish municipalities have also extended their responsibilities in social welfare, primary education and employment, including integration of immigrants (Aalbu, Böhme, and Uhlin 2008). The Danish administrative reform was also marketed as a local democracy reform, as more responsibilities were transferred to organisations closer to the people (Aalbu, Böhme, and Uhlin 2008, 40). As a result of the reform, whose primary focus was on the smooth delivery of services, the municipalities extended their responsibilities in such areas as social welfare, primary education (while secondary and higher education remained the responsibilities of the state), spatial planning, the environment and nature, culture, roads and industrial development and employment (Aalbu, Böhme, and Uhlin 2008, 21), which resulted in teaching Danish as the L2 to immigrants (Siiner 2012, 54 55). Denmark can, in the EU context, be described as highly decentralised, in which the sub-national government expenditure share as a proportion of general government expenditure, describing the spending power of subnational governments is almost two-thirds, one of the highest in the EU (Escolano et al. 2012, 6). The economic capacities of sub-national governments (consisting of local and regional governments, with the latter solely responsible for health care) must therefore be sufficient to meet their responsibilities. The critiques of the high decentralisation in Denmark point to the danger of setting up an elaborate administrative system that can lead to inefficiencies in both local and national governance and high administrative costs (Schmidt 2008). We can, however, conclude that in Denmark, local municipalities have the necessary authoritative and substantive allocative resources to solve challenges in (language) education policy.

9 Journal of Multilingual and Multicultural Development 7 Estonia Estonia has, on the other hand, developed a more centralised type of language management, with the state being the main agent in initiating language political interventions, stating goals and exercising control. In order to normalise the situation after decades of one-sided bilingualism in the Soviet area, the post-communist Estonia developed a control-oriented, top-down and thick type of LP, where language planning explicitly appeared in repeated laws and regulations, with agencies designated to carry out regulating and control functions (Siiner 2006; Rannut 2004). The first Language Act, passed in 1989, was primarily a symbolic act, but was also expected to work as one of the main tools of change. Although the latest version of the Integration Programme (Eesti Lõimumiskava ) places significantly more stress on civic participation and the economic dimensions of integration, linguistic integration is still the main goal of the integration policy of Estonia, and is sometimes equated with it (Vihalemm and Siiner 2011, 128). As part of the Soviet legacy, Estonia inherited a parallel school system, with Russianmedium and Estonian-medium schools. The future of the Russian schools cannot be determined separately from the question of regional development, since the schools mainly lie in the Russian enclaves, in the north-eastern corner of Estonia and in the capital, and are therefore part of the distinctive demographic and structural features of post-communist Estonia. A large survey carried out in Estonian Russian-medium schools revealed that the future of the Russian-medium schools is connected to political struggles, not only in language skills and education, but also in regional development, the growing exodus, social cohesion and democracy, in terms of trust in local municipalities and governments final reform goals (Kello, Masso, and Jakobson 2011). A study of risk communication revealed that less-integrated Russian-speakers in Ida-Virumaa, especially inhabitants of Narva, did not trust that the official information from either CG or LGs was in their best interests (Harro-Loit, Vihalemm, and Ugur 2013, 235). To integrate Russian-language schools into Estonian education, amendments to the law on basic and upper secondary schools were introduced in 1997, which required Russian-language gymnasiums to begin the transition to the Estonian language in 2007, with the final aim of including at least 60% of the entire curriculum by 2011 (Heidmets et al. 2011, 97). The education reform was part of the general normalisation processes in the transitional state (Smith 2003). However, the reform did not deal with the language of instruction in the basic and lower secondary Russian-medium schools, which was to remain Russian. Estonian-medium teaching is not compulsory in basic schools, but many Russian-medium schools offer one or more subjects taught in Estonian. In addition, the state has supported the implementation of language immersion programmes, which follow the example of the Canadian French immersion model. Approximately one-fifth of non-titular pupils currently either take part in an immersion programme or learn Estonian in depth at the basic school level, and the number is increasing (Kello, Masso, and Jakobson 2011), as Russian-speaking parents clearly prefer basic schools with Estonian language immersion programmes or Estonian-medium schools (Metslang et al. 2013, 9). Parents choices also reflect the infrastructural deficit in Estonian LEP. Contrary to the solution applied in Denmark, the state exercises control over children s language competence once they are adults (as Russian-speakers have to pass Estonian language exams in order to gain Estonian citizenship), but does not distribute sufficient funds to support successful language socialisation and acquisition starting at the preprimary and primary school levels. Improving the linguistic skills of both Russian-medium schools teachers and pupils has been mainly delegated by the state to the administrators of

10 8 M. Siiner the schools, without offering them any training in change management (Mehisto 2009). Teachers often need not only linguistic training but also methodological consultation and emotional support. Due to the low birth rate in the early 1990s, the overall number of schoolchildren in grades 1 12 has declined. An especially steep decline in the number of pupils occurred in schools with Russian as the language of instruction (Heidmets et al. 2011, 97), affecting the quality and content of instruction in Estonian in these schools (Metslang et al. 2013, 5). This tendency continues and, combined with the general decrease in population in Estonia (Statistics Estonia 2011), the closing of a number of municipal schools is unavoidable, contributing to insecurity and stress among teachers. The first monitoring of the preliminary results of the present Development Plan for the Estonian Language (an important strategic planning document whose principles are taken into account in drawing up draft laws and planning the budget of the language domain) clearly reveals that especially school headmasters and teachers in Ida-Virumaa struggle with insufficient language competences, while inhabitants of this region demonstrate the most sceptical attitudes towards the value of Estonian language in their everyday life (Development Plan, 28). A closer analysis of the Basic Schools and Secondary Schools Act and the Development Plan of the Estonian Language reveals that municipalities are not assigned any specific role in LP implementation, in spite of the fact that many schools in Estonia are municipal. Municipalities as owners are in charge of general management of a school and have to apply for education licence to run a school from the Minister of Education and Research (Basic Schools, 63). In that application they have to document existence of teachers meeting the qualification requirements. However, an analysis of existing practices, also reflected in schools development plans, reveals that considerable responsibility is given to schools headmasters, who are in charge of guaranteeing a smooth transition to Estonian-medium education (Kello, Masso, and Jakobson 2011). The school budgets for the transition are determined directly by the state, on the basis of the schools development plans. An analysis of these plans reveals that schools are responsible for finding teachers with necessary competences in Estonian, which has long been a serious problem in Ida-Virumaa. This problem can not be solved by the schools alone, as it demands reconsideration of teachers training for Russian medium schools in areas with minimum presence of Estonian language outside of school classes and of how to attract teacher to jobs in Ida-Virumaa. The regional differences in Estonia in social and economic development and in the ethnic composition of the population have been described in detail in several studies and reports (Uus and Kaldur 2013; Lauristin et al. 2011), clearly indicating the need for regional solutions, which have been hampered not only by the above-mentioned distrust between the CG and LGs but also by unevenly distributed resources. Estonian municipalities are rather small, having between 500 and 18,000 inhabitants, with the exception of the five largest towns, which have 40, ,000 inhabitants (Statistics Estonia 2011). There is at present a trend towards an increase in regional imbalance in Estonia and towards the extinction of peripheral regions (SE , 24). The considerable differences in the allocation of financial resources among local municipalities (depending on the size and income of their population) led the Congress of Local and Regional Authorities to recommend that the Estonian state allocates financial resources to local authorities commensurate with the increasing responsibilities assigned to them (Council of Europe 2010).

11 Journal of Multilingual and Multicultural Development 9 In spite of good advice, local government reform (LGR) attempts in Estonia have been rather fragmented, non-holistic, unfinished or often not even developed beyond the planning stage (Linnas 2011, 181). The main obstacles in the development of a capable state and administration have been not only economic constraints but also the neo-liberal and technocratic world-view of the political elite, leaving the future of the public administration reform unresolved (Drechsler 2004, 394). Linnas (2011, 181) points out that the lack of clarity in LGR goals hinders their implementation, including consensus building and embodying perspectives of all major stakeholders. As early as 1999, the National Association of Local Authorities in Denmark (NALAD) recommended forming sub-commissions; a sub-commission on education might consist of members having specialised knowledge in the field, e.g. school headmasters. It was recommended that these sub-commissions should be involved in the LGR process (NALAD 1999, 30), which has not happened. In addition, co-operation between LGs ( 12 of the Local Self- Government Organization Act), through county associations, to create opportunities for improved performance, such as joint outsourcing or co-sourcing of different functions, is also still limited in Estonia, hampering also the possibility of founding schools jointly by several smaller municipalities (cf. Basic Schools, 61). The role of LGs in the LEP, as in integration issues, has been thus rather vague and incidental (Eesti Lõimumiskava ). There is, however, one opportunity for policy adjustment or agency being assigned to local governments: they can apply for permission from the national government to have a different language than Estonian as the language of instruction at one of their municipal schools (Basic Schools, 21, 2 3). I will come back to that in the next section, offering two case studies of LEP agency taken by LGs. Comparative analysis of decentralisation as redistribution of authoritative and allocative resources Copenhagen The share of immigrants of the population of Copenhagen was 22.2% in 2012, with Arabic, Turkish, Somali and Urdu speakers as the largest language groups. In LEP, local municipalities in Denmark possess considerable authoritative and allocative resources, being the main agent responsible for making sure that all children are prepared to start school, which includes the task of offering the screening of Danish language competence to all three-year-old children in and outside of day care, with the language screening of bilingual children being obligatory at the start of day care (at the age of three), at the beginning of school, at the age of five or six and during basic school, usually in the fourth and seventh grades, in order to ensure that children have sufficient Danish language skills to keep up with their studies. If needed, extra classes in Danish as a second language are offered. Since there was an especially clear positive correlation between ethnic minority students poor reading skills and their parents socioeconomic status, a further step was taken regarding legislation, which also had an impact on the parents situation. Parents are not only involved in language screening and instructed in how they can stimulate their children s language development, but the LG has the responsibility of informing parents whose children do not attend day care about the opportunity to place their children in day care (Primary and Lower Secondary Schools Act 2009, 5). The LGs role as the central agents in solving the problem of linguistic integration includes the responsibility of offering support (allocative resources, i.e. know-how and feedback to parents) and control, revealing the state s distrust of bilingual children s parents capability of supporting their children s integration sufficiently.

12 10 M. Siiner Furthermore, the parents employment situation is included in the evaluation of a child s situation and stimulation needs. Parents are also obliged to let their children participate in language screening and language stimulation classes. If parents themselves want to attend language stimulation classes, they are obliged to inform the LG in written form, including, as minimum information, who will attend language stimulation, which children will participate in language stimulation and where it will take place. The municipal council is in charge of controlling and inspecting the language stimulation classes attended by parents (Statuary Notice of Act on Pre-Primary Education 2011, 11). Since the PISA Ethnic 2009 demonstrated that school settings also have an impact on students reading skills (both native and ethnic minority students attending schools with 40% or more bilingual students achieve significantly lower scores on the PISA reading test compared with native students who attend schools with fewer than 10% bilingual students), yet another amendment was made to the basic and lower secondary school act in In Copenhagen, the municipal authority launched the Copenhagen Model (københavnermodellen) in order to reverse the disturbing trend of a growing concentration of bilingual children in a number of municipal primary and lower secondary schools in certain areas. The local municipality offered bilingual children the opportunity to attend basic schools in different catchment area schools. The municipality, at the same time, offered schools resources to renovate and modernise schools, thus enabling them to attract monolingual children from socioeconomically advantaged Danish-speaking families. The Copenhagen municipality made a plea to these families to assist bilingual families by sending their children to the public schools, instead of sending them to private schools with fewer bilingual pupils, as was the tendency at the time. Here the town council was not a passive implementer, but an active language political agent. In the case of a covert LP (Denmark), where the state has not taken an official position regarding the symbolic status of the state language, the prevailing understanding of language (problems) is instrumental and dynamic in the sense that problems have to be solved contextually and locally, and local municipalities, who also are responsible for primary education, are seen as the main agents in solving language education problems, having both the authoritative and allocative resources to take initiative in finding locally workable solutions. Tallinn In Tallinn, the share of Russian speakers in 2012 was 36.2% and Estonians 55.5% (Danmarks Statistik 2013; Statisticas Estonia). Russian speakers are not a homogeneous group, as they consist of several ethnic groups. According to the Estonian Department of Statistics, there are more than 120 ethnic groups living in Estonia (but only 15 of them exceed 1000 members), but less than 40% of them have retained their ethnic languages, with most of them having shifted to Russian, many already during the Soviet period (see more in Siiner 2006). In Lamberts terms, Estonia and Tallinn can thus be described as dyadic rather than mosaic in terms of presence of different linguistic groups (Russian and Estonian) (Lambert 1999). Tallinn is the biggest municipality, with almost 0.4 million inhabitants, constituting about one-third of the total population (Statistics Estonia 2011). The municipal government of Tallinn has been dominated by the Center Party, which is also the largest opposition party and has been the most popular party among the vast Russian-speaking minority since The past decade has witnessed the escalation of the conflict between the state and local municipalities, fuelled by the growing political power of the populist

13 Journal of Multilingual and Multicultural Development 11 Center Party in several municipalities in Ida-Virumaa and in Tallinn. The Center Party, making up the majority in the Tallinn City Council, claims to worry about the future of Russian-medium schools and about Russian-speaking children s insufficient Estonian language skills, and thus about their poor future prospects. Although Estonian language proficiency among teachers in Russian-medium schools has been consistently improving, the problems in several secondary schools in Tallinn and Narva are so severe that schools have been forced to close down their upper secondary classes (while the lower grades [1 9] continue). In 2013, the municipal government of Tallinn petitioned the Estonian government to allow a number of upper secondary schools to continue instruction in Russian, citing the Basic Schools and Upper Secondary Schools Act, 21 (3), which leaves it to the board of trustees of a municipal gymnasium based on the school s development plan to ask the LG or city council to apply for permission from the Estonian government to use a different language than Estonian for instruction. Their petition was rejected by the Ministry of Education and Research with reference to the Language Act and the Constitution, which states that everybody should have the right to receive education in the Estonian language (Tallinna Linnamäe 2013). As a result, the Tallinn city government advised the Tallinn Linnamäe Vene Lütseum (hereafter TLVL), where 40% of teachers had, by 2013, not passed the required B2 level in Estonian, to become a private school, as according to the Private Schools Act 1998, 15 (1) the board of trustees of a private school is free to select the language of instruction. The national government responded by making amendments to the Private Schools Act, so the choice of the language of instruction would in future be decided on the basis of the Basic School and Upper Secondary School Act when a shareholder, founder or member of a private school was the state or the LG. The aim was to eliminate the possibility that the Private Schools Act could be used to avoid a school being shut down due to problems with transition to Estonian-medium instruction. The distrust between the Tallinn city council and the national government can also be explained as a principal-agency problem, known from political science (Rees 1985). Here the CG as the principal lacks trust in a powerful LG s (leaded by an opposition party) ability and willingness as the agent to act in the best interest of the state, rather that in its own interests. The problem arises when the two parties have different interests and asymmetrical information (the agent having some information that the principal does not). The situation is also complicated by the fact that in Estonia there are both state and municipal schools (owned and run by the municipality). The owner of the school also establishes the statutes of the school, including the rules for composing and the functions of the board of trustees (whose goals are to observe and create better opportunities for teaching and education), and establishes the rules of its procedure (the Basic Schools and Upper Secondary Schools Act, 66). There have been complaints from representatives of the opposition parties in the Tallinn city council that the Center Party appoints only its own party members to the schools boards of trustees, making it more difficult for politicians from other parties in the Tallinn city government to deal with Tallinn municipal schools (Pilvre, personal communication). The Center Party has a special interest in Lasnamäe, the area of Tallinn with the biggest concentration of Russian-speakers and where the TLVL has been established. The TLVL development plan for clearly states that, due to the specific situation in Lasnamäe, schools should recognise their roles not only as educational institutions, but as contributors to the development of the whole area (Arengukava 2010, 3), while municipal schools in areas with fewer Center Party voters have experienced problems with getting the Tallinn city government to attend to their needs (Pilvre, private communication).

14 12 M. Siiner What at first glance seems to be a question of deviations in the implementation of LEP (the language of instruction) is at a deeper level a question of incomplete decentralisation and democratisation processes, hampered by the political conflict between the national government and the Tallinn City Council, dominated by the Center Party. Both parties have used the means available to them to hamper democratisation processes and transparency, by obscuring information about the future of Russian medium schools in general (by the Government) and some specific Russian medium schools at Lasnamäe (by the Tallinn city council). On the other hand, it has become clear that locally and regionally sensitive policy appropriations are needed, demanding cooperation between schools and municipalities, since unitary (state) solutions do not work (Lindemann 2013; Haveri 2003). Leaving the responsibility for solving the problem of transition to Estonian-language instruction to a large extent to the schools and headmasters (Masso and Kello 2011, 130) ignores the fact that these problems cannot be solved separately from the political, economic and socio-cultural transitions of the society. Discussion The present paper is an attempt to contribute to the analysis of LP developments as an integral part of the development of the whole society and its social structures. Based on an analytical paradigm of allocation of resources and agency, my aim has been to shed more light on the role and importance of lower administrative levels in solving language problems, as language problems may have structural and regional features and be related to social and infrastructural problems, and thus also to the democratisation processes in a society. I have analysed two attempts at agency taken by two city councils in two countries with two different language policy models. According to the ethnographic approach, language legislation and regulations can open up free spaces (Hornberger and Johnson 2007) for necessary local initiatives, interpretations and adjustments, by either stating them explicitly in policy texts or by leaving agency to actors at lower levels, as was the case with the Copenhagen municipality, who, by applying the Copenhagen model, tried to solve language socialisation constraints. However, a lack of flexibility, in which LP matters are perceived as solely the business of the state, can create a vacuum, where the unitary solutions offered by the state are not sufficient to solve local language problems and the schools, responsible for implementation of the reform (transition to Estonian medium instruction) do not have access to appropriate resources to solve the infrastructural problems (lack of qualified teachers in some areas). My analysis demonstrated that the role of LGs in LP could not be understood and analysed separately from the decentralisation and democratisation processes in the state, raising the issues of accountability and local involvement. Increasing regional differences and uneven allocation of resources creates some powerful municipal agents, which can, by undemocratic means serve their own (their voters ) interests rather than the interests of the whole municipality or the interest of (a better cohesion in) the state. The described ideological tension present between a (too) powerful LG and the CG can create a principal-agent problem, where the LG can be hard to motivate to act in the best interest of the whole society, contributing thus to the CG conviction that LGs can not be trusted. Acknowledgement The article was prepared in the framework of research grant Collective Identities in Estonia in the Changed Political and Economic Context financed by Estonian Science Foundation [Grant No. 8347].

15 Journal of Multilingual and Multicultural Development 13 References Aalbu, H., K. Böhme, and Å. Uhlin Administrative Reform Arguments and Values. Nordic Research Programme Report 6. Stockholm: Nordregio. Arengukava Tallinna Linnamäe Vene Lütseumi Arengukava [Development Plan for Tallinn Linnamäe Russian Gymnasium]. Tallinn. Backhaus, P Linguistic Landscapes: A Comparative Study of Urban Multilingualism in Tokyo. Clevedon: Multilingual Matters. Backhaus, P Language Policy at the Municipal Level. In The Cambridge Handbook of Language Policy, edited by B. Spolsky, Cambridge: Cambridge University Press. Basic Schools and Upper Secondary Schools Act Passed by Riigikogu on 9 June 2010, published in Riigi Teataja 2010/41/240. Blair, H Participation and Accountability at the Periphery: Democratic Local Governance in Six Countries. World Development 28 (1): doi: /s x(99) Blommaert, J The Debate is Open. In Language Ideological Debates, edited by J. Blommaert, Berlin: Mouton de Gruyter. Council of Europe Local Democracy in Estonia. Recommendation 294. Accessed 13 January = &Site = Congress. Crook, R., and J. Manor Democracy and Decentralisation in South Asia and West Africa: Participation, Accountability and Performance. Cambridge: Cambridge University Press. Cullis, J., and P. Jones Public Finance and Public Choice. London: McGraw-Hill. Danmarks Statistik Indvandrere i Danmark [Immigrants in Denmark]. Copenhagen: Danmarks Statistik. Development Plan of the Estonian Language Tallinn: Estonian Language Foundation. Accessed 13 January keele+ arengukava+ inglise.indd_.pdf. Drechsler, W Governance, Good Governance, and Government: The Case for Estonian Administrative Capacity. TRAMES 8 (4): Echebarría, K Government Modernization and Civil Service Reform: Democratic Strengthening, Consolidation of the Rule of Law, and Public Policy Effectiveness. Regional Policy Dialogue, Public Policy Management and Transparency Network. Paper. Washington, DC: Inter-American Development Bank. Eesti Lõimumiskava [Estonian Integration Strategy ]. Accessed 13 January Escolano, J., L. Eyraud, M. M. Badia, J. Sarnes, and A Tuladhar Fiscal Performance, Institutional Design and Decentralisation in European Union Countries. IMF Working Papers 12 (45). Freeman, R Building on Community Bilingualism. Philadelphia, PA: Caslon. Giddens, A The Constitution of Society: Outline of the Theory of Structuration. Cambridge: Polity Press. Hanberger, A Democratic Accountability in Decentralised Governance. Scandinavian Political Studies 32: 1. doi: /j x. Harro-Loit, H., T. Vihalemm, and K. Ugur Information Channels and Response Patterns in a Situation of Risk. In Developing a Crisis Communication Scorecard, edited by M. Vos, R. Lund, Z. Reich, and H. Harro-Loit, Jyväskylä: University of Jyväskylä Bookstore. Haveri, A Intermunicipal Cooperation as a Part of Local Governance. Finnish Local Government Studies 31: Heidmets, M., A. Kangro, V. Ruus, A. V. Matulionis, K. Loogma, and V. Zilinskaite Education. Historical Perspective: 20 Years of Reforms. Estonian Human Development Report 2010/2011, Tallinn: Eesti Koostöö Kogu. Holmen, A., and J. N. Jørgensen Skærpede holdninger til sproglig mangfoldighed i Danmark [More Strict Attitudes towards Linguistic Diversity in Denmark]. In Sprogs status i Danmark 2021, vol. 58., edited by J. N. Jørgensen and A. Holmen, Copenhagen: Københavnerstudier i tosprogethed. Hornberger, N. H., and D. C. Johnson Slicing the Onion Ethnographically: Layers and Spaces in Multilingual Language Education Policy and Practice. TESOL Quarterly 41 (3): Hornberger, N., and D. C. Johnson The Ethnography of Language Policy. In Ethnography and Language Policy, edited by T. McCarty, New York: Routledge.