Linguistic terrioriality under stress

This article revisits a well-known dichotomy (the ‘territorial’ and ‘personal’ principles) and develops a four-element classification of state approaches (from the most generous to the most menacing, from the perspective of speakers of minority languages). The article examines the implications for language policy of geographically dispersed or spatially concentrated patterns of distribution of speakers of particular languages. We begin by exploring the general literature on language policy, focusing in particular on the territorial and personal principles, the use of ‘threshold rules’ at municipal and other subnational levels, and the hybrid language regimes that are often a consequence of sociolinguistic complexity. We consider the extent to which responses to linguistic diversity across Europe may be understood by reference to these principles and categories. We explain why we have selected particular case studies (the Baltic republics, Transylvania, Switzer-land, Belgium and Ireland) for further exploration. We conclude that, notwithstanding the value of the typologies we consider, real-life cases are almost invariably more complex, with states implementing policies that defy categorisation, that may change over time, and that may treat different language minorities by reference to different principles.


Introduction
The relationship between language and territory, and, more particularly, the implications for public policy of geographically dispersed or spatially concentrated patterns of distribution of speakers of particular languages, is our starting point in this special issue. Our object is to confront the major contours of the substantial literature to which this demographic challenge has given rise by narrow-ing the focus to a small number of case studies and seeking to draw out from these the major lessons that have emerged across a range of disciplines.
In this introductory contribution we examine the context of the articles that follow in this issue. We begin by exploring, in the first section, the general and comparative literature. We consider the origins of the distinction between the 'territoriality' and 'personality' principles in public international law, and the manner in which they were carried over into such disciplines or cross-disciplinary fields as sociolinguistics, political science and cultural geography. The next section describes the variety of responses to linguistic diversity across Europe, relying in particular on the well-known territoriality-personality dichotomy -the notion that state responses to problems of societal bi-or multilingualism may build on provisions for the specific territory or region concerned (the principle of territoriality), or be constructed around 'the language chosen by or attributed to the individuals being served' (the principle of personality; McRae, 1978, p. 332). In the last main section, we indicate our reasons for selecting particular cases for further analysis, and we outline the tentative conclusions to which these articles point.

The debate
The principle that the population should use the language of its rulers -cujus regio, ejus lingua -describes the linguistic territoriality principle advocated by Philippe Van Parijs (2011a, 2011b. In the normative literature on language regimes, the territorial approach is usually opposed to the personal approach to describe two alternative principles of language policy and planning. At the root of the territorial approach lies an assumption that the most basic language rights are secured whenever it is possible to use the language in all contexts of life, and the concentration of a language community within the boundaries of a territory enables the establishment of a sociolinguistic 'context of choice' (Kymlicka, 1989) -an approach based on the questionable hypothesis of homogeneous linguistic territories with clear-cut boundaries (De Schutter, 2008).
In the domain of policy discussion, the territoriality-personality divide gained particular momentum during the debate on Canada's constitutional options in the last three decades of the 20th century and later (see Cardinal & Léger, 2018, for an overview). It re-emerged in the wake of the enlargement of the European Union and discussion of the drafting of a European constitution in the early years of the new millennium (Van Parijs, 2004;Gazzola, 2006;later discussed in Aparicio Fenoll & Kuehn, 2016;Burckhardt, 2018;Burckhardt & Gazzola, 2018). These two waves were followed by scholarship focusing on the link between language regimes -intended as the principles and practices guiding state policies on language -and the institutional traditions underpinning them (Cardinal & Sonntag, 2015a). A comparative analysis based on empirical and retrospective research facilitated a sidelining of the traditional divide between personal and territorial approaches. Language regimes can be rarely -if ever -summed up as the implementation of prescriptive normative principles, as illustrated by the Swiss case (Burckhardt, 2021). They can be best understood as the outcome of historically rooted approaches to statehood (Cardinal & Sonntag, 2015b), leading to a shift in focus towards the combination of territorial and personal elements to describe complex language regimes, including the concepts of multilingual citizenship and institutional completeness (Cardinal & Léger, 2017). In fact, a closer inspection of the clear-cut cases of territoriality and personality brings to the surface a hybrid system which is a combination of aspects of both principles. In this special issue, we argue that these hybrid systems play a crucial role in practice, as the articles that follow will demonstrate.
The importance of territory for activities in the public domain is well established: government programs typically require a defined geographical space over which they are carried out. But might governments relate to citizens outside the framework of territorial politics? We may identify at least two relevant contexts in which the territorial principle is challenged. In the first place, leaving aside for the moment the question of language policy, emerging concepts of citizenship began to make a distinction between the principles of geography and descent at an early stage (Slama, 2017). By the early twentieth century there were significant differences between states that followed the principles of jus soli, acquisition of citizenship by virtue of birth within a particular territory, and jus sanguinis, acquisition of citizenship on the basis of blood relationships; in general, the former principle was predominant in the Americas, and the latter in Europe (Scott, 1930, pp. 28-29). This appears to have reflected contrasting state-building trajectories, with citizenship law being marked in varying degrees by a legacy of organic conceptions of nationality that stressed ties of community and kinship (Safran, 1997, pp. 313-314). Of course, too much should not be made of this distinction: one comparison of citizenship conceptions in France and Germany -traditionally seen as prototypes of the 'territorial' and 'descent' approaches respectivelyconcluded that the picture is a complex one, with criteria sometimes overlapping (Brubaker, 1992, pp. 10-11).
Our central concern here, however, is with a second context: the implications of territoriality for language policy rather than for citizenship or political autonomy. One of the earliest sustained policy debates addressed problems of linguistic co-existence in Canada, as mentioned above. This was spurred by the emergence of the language issue in Canadian politics in the 1960s, and found expression in the Royal Commission on Bilingualism and Biculturalism established by the government in 1963. The remarkable five-volume report of the commission, published in 1967-70, was based on painstaking research. It addressed the issue of territoriality directly in its first volume, Chapter 4 of which was entitled 'the territorial principle and the personality principle' (Royal Commission, 1967, pp. 75-88). The chapter noted that 'only a few countries have accorded full equality to their linguistic minorities' , and singled out four as having something to teach Canada: Finland, Belgium, Switzerland, and South Africa. While this referred to South Africa under the apartheid regime, when only the white population enjoyed full civil and political rights, the commission found other reasons for classifying this country as an outlier: it was characterised by a high level of personal bilingualism (between English and Afrikaans); its two main languages, like those of Canada, were not indigenous to the area but were a result of settlement; and there was a relatively high degree of spatial mixing, unlike the other countries, where there were many unilingual areas. As a consequence, the commission concluded, the personality principle, not the territoriality one, was dominant in South Africa. In its view, the principle of territoriality was 'a theoretical concept'; the Finnish, Belgian, and Swiss systems did not represent pure examples of this, since they also admitted 'some considerations of personality' (Royal Commission, 1967, pp. 81-82).
The Canadian commission was influential not just in shaping public policy but also in sparking widespread academic interest. One of those extensively involved in the commission, the late Kenneth McRae, developed its findings in respect of the tension between the 'territoriality' and 'personality' factors in a path-breaking article that was to shape his own research profile for decades and to have a major influence on the broad field of language policy and politics (McRae, 1975(McRae, , 1978. Building on an older tradition of legal scholarship, he applied the distinction between jus soli and jus sanguinis to the domain of language policy. He made a major contribution to empirical research by building further on the work of the commission with seminal, book-length studies of language policy and its sociolinguistic context in Switzerland (McRae, 1983), Belgium (McRae, 1986) and Finland (McRae, 1997); for an elegant overview, see McRae (2007). He did not include South Africa, presumably in part because of the distinctiveness that the royal commission had observed but also because of the increasingly obvious unsustainability of the apartheid regime, and the exclusion of most of the population from rights even more fundamental than language ones.
More than two decades into the twenty-first century, we have accumulated a large amount of experience in analysing language policy as regards its anticipated and actual outcomes. The dilemma remains unchanged: the coexistence of languages in bi-or multilingual societies becomes increasingly problematic as sub-state public bodies and agencies grow in political and administrative significance. Indeed, the propensity of language contact to lead to linguistic conflict -often taking the form of a metaphorical life-or-death struggle between languageshighlights the extent to which problems arising from the coexistence of two or more languages on a single territory urgently require resolution (Laponce, 1987, pp. 53-93;1993). The tension between 'official' languages of state and other languages which do not enjoy this status is a relevant problem for this special issue, which also considers patterns of competition between geographically dominant languages, whether or not they enjoy official status, and spatially dispersed ones, which typically are socially, politically and legally marginalised.
In assessing the manner in which the political and administrative implications of the twin principles of 'territoriality' and 'personality' have been spelt out in European states, we need to consider these two principles more carefully. The 'personality' principle implies that citizens should enjoy the same set of (official) language rights no matter where they are located spatially within the state; the 'territorial' principle posits that language rights may vary from region to region, according to local conditions (Kymlicka & Patten, 2003, p. 29). The latter approach thus applies a geographical principle in protecting language rights in a particular area (May, 2012, pp. 187-188). The 'personality' principle (corresponding to jus sanguinis) implies that citizens should enjoy the same set of (official) language rights no matter where they are located spatially within the state; the 'territorial' principle (like jus solis) posits that language rights may vary from region to region, according to local conditions, notwithstanding its limitations, such as the arithmetical criterion to be adopted -the cut-off point at which there are deemed to be a sufficient number of speakers of a particular language to warrant active language protection and the use of such languages in the public domain (May, 2012, p. 190).
The case for recognizing and guaranteeing the 'equal dignity' of languages by acknowledging their official status in a given territory is indeed a strong one. In a multilingual area in which people learn languages and interact linguistically with one another, a coercive regime can be either territorially or functionally differentiated: its provisions can vary according to either the location of interaction or according to the category to which the people involved belong. In the former case, linguistic borders are crucial; in the latter, linguistic community membership is (Van Parijs, 2011a, p. 136). Van Parijs defends the territorial regime on the basis that, in the EU case, for example, it permits each of the co-existing languages to be 'king' within the limits of its territory, and for the identity associated with that language to be protected through its role as the official language of the population as a political community (Van Parijs, 2011a, pp. 146-149). The Van Parijs model sees personal bilingualism as comprising two overlapping territorial layers: in the EU case, a common layer, or ground floor, encompassing the whole of the EU, with English as the common language; and multiple territorially circumscribed layers, or upper floors, in which ancestral languages 'reign' , though such a twolevel model is not our primary concern here.

The options
In principle, states may respond in at least four quite different ways to the challenge of linguistic diversity. First, in the 'preservationist' approach the state's policy on minority languages has a conservationist dimension, seeking to cultivate them and ensure their survival as cultural forms, even if their speakers are few in number and show little interest in the survival of the language or in being identified as a distinct community. State policies are rooted in a top-down approach which may go well beyond the demands of linguistic minority activists (if these exist; in some cases, there are none). Second, in the 'protectionist' approach states acquiesce in the demands of minority activists, arriving at a mutually acceptable accommodation that typically makes provision for a bilingual or multilingual regime. The shape and extent of this regime will depend on geolinguistic and sociolinguistic considerations, as well as on the balance between minority demands and state priorities. Third, in the 'proscriptive' approach states make no (or only insignificant) concessions to linguistic minority activists. The state is defined as unilingual, and the official state language enjoys a monopoly in the public sector, including the education system; use of minority languages is proscribed. Fourth, the state may adopt a 'permissive' approach, at least formally, defining itself as language-neutral and leaving substantial room for manoeuvre to communities to organise themselves around non-state actors such as religious organisations or secular charities, which may receive public funding to implement state policies.
Not surprisingly, the tension between these approaches has resulted in a sizeable literature. Early scholars, such as Laponce (1987), opted for a 'protectionist' regime, relying on the territoriality principle to provide a language with the best chance to maintain its integrity and its specific features. The principle of equality between languages was advocated also by Kymlicka & Patten (2003), and relabelled as 'equal dignity' or 'parity of esteem' , as in the work of Van Parijs (2011a). Here, again, the strongest protective principle, at least as far as Van Parijs is concerned, is territorial status for the languages involved. There may indeed be more pragmatic reasons for adopting this approach. As Csata et al. (2021) show, 'protectionist' language policy may promote efficiency in economic exchange and the construction of social trust between speakers of the majority and minority languages, contributing to enhanced socioeconomic development. This continues to be an area of interdisciplinary debate. One extensive strand in the literature stresses the negative effects of territorial diversity on growth in developing countries (Easterly & Levine, 1997;Alesina et al., 2003) and US counties (Alesina, Baqir & Easterly, 1999). But different findings may also be cited: Arcand and Grin (2013) conclude that linguistic diversity does not have negative effects on macroeconomic variables such as growth or per-capita GDP, while Florida (2005) highlights the positive outcomes of territorial diversity on innovation in wealthy metropolitan areas.
Clearly, the nature of a language regime -whether or not territorially baseddepends on political and contextual conditions (McRae, 2007;May, 2012;Williams, 2013;Dembinska et al., 2014;Vizi, 2016). Quite often, minority and migrant languages are granted specific rights in the context of a regime that prioritises an established official language. The nature of the rights that are applied is determined by a number of factors: domestic ones such as size, history and tradition, and international ones, including geopolitical weight, relations with kinstates, international treaties and other outcomes of globalization and of the impact of European agreements and norms. The importance of the broader European dimension is exemplified by two conventions of the Council of Europe which help to protect the position of traditional language minorities in affiliated states: the European Charter for Regional or Minority Languages (ECRML) adopted on 5 November 1992, and the Framework Convention for the Protection of National Minorities (FCPNM) signed on 1 February 1995 (Grin, 2003).
Closely related to the notion of territorial and non-territorial criteria underpinning language policy is that of territorial and non-territorial autonomy. The territorial approach (which finds expression in autonomy, devolution, federalism and related instruments, as illustrated by such cases as Switzerland) is well known, and is the subject of a very large literature (for a recent overview, see Loughlin et al., 2013). The second, non-territorial autonomy, is less familiar, and has only recently been subjected to comparative analysis (see, for example, Nimni, 2005;Smith & Cordell, 2008;Smith & Hiden, 2012;Nimni, Osipov & Smith, 2013;Chouinard, 2014;Coakley, 2017). Historically, non-territorial autonomy was common before the creation of the modern state. Tribal peoples, often nomadic, related to their kings or rulers by bonds which 'resided in the blood-tie, not in the territory' , so that 'the law under which a man lived was personal, always attaching to him, in virtue of his origin, and irrespective of his domicile' (Macartney, 1934, p. 60). Thus, in medieval times the Germans of Bohemia lived under their own corpus of laws, jus Teutonicorum, and were so recognised by the King of Bohemia; at a later period the Székely of Transylvania were recognised as an autonomous, non-territorial 'nation'; and the non-Muslim millets ('nations') of the Ottoman empire enjoyed a similar status (Macartney, 1934, pp. 60-66;Csata & Marácz, 2016). Autonomy of this kind fits uneasily in the constitutional structure of the modern state, but there have been some examples of notable experiments with non-territorial autonomy, usually restricted to cultural and linguistic rights, as in Moravia in the last years of Habsburg rule, interwar Estonia, post-1970 Belgium and several Central and East European states after the fall of communism (Coakley, 2016a, pp. 13-17).
Resistance to territorial autonomy and absence of consensus on solutions gave way to parallel proposals -mostly in Central and Eastern Europe -for cultural and linguistic autonomy based on the personality principle. This model acquired popularity at the end of 19th century in the Austro-Hungarian Empire, in particular in the work of the social democratic activist Karl Renner (for an account, see Nimni, 2007). With a view to mitigating risks commonly associated with territorial autonomy (such as secession, the creation of new minorities and ethnic discrimination), ethno-linguistic minorities were to be granted collective constitutional rights and non-territorial cultural self-rule. Cultural autonomy recognises the status of collectivities whatever their geographical location; collective rights are thus dissociated from territory and can be attributed to dispersed groups. The model potentially gives control to communities over their cultural and linguistic affairs. Local minorities are not subject to the cultural practices of the local majority but enjoy 'sovereign areas of competence' (Nimni, 2007). In the Austro-Hungarian Empire the 'nationalities' , i.e. the autochthonous ethnolinguistic groups, were granted individual and collective linguistic rights, recognised in offices, schools, courts and in county and communal assemblies (Marácz, 2020, pp. 63-4). This law was not always respected in practice in the Hungarian part of the empire, where speakers of minority languages continued to decline (for example, the number of Slovak speakers fell by 8% between 1880 and 1910; Marácz, 2020, pp. 68-69). In the contemporary world, where the traditional nation-state model gives way to new governance structures such as the European Union, it seems appropriate to consider cultural autonomy as an alternative to territorially conceived solutions to protracted ethnic conflicts. But we should not expect too much if we stray outside the domain of language policy: comparative analysis of regimes of at least nominal non-territorial autonomy has concluded that 'there may be less -perhaps, much less -to non-territorial autonomy than meets the eye' (Coakley, 2016b, p. 178).
Countries vary both in the extent of the suite of language measures conceded to minorities and the manner in which this changes over time, often in the context of prolonged struggle. Thus, for example, in Poland, minority representatives asked for an 8% threshold, which would permit 79 municipalities to be defined as bilingual (with established minority languages such as German, Ukrainian or Belarusian coexisting with Polish). This was extensively discussed in the Polish parliament between 2002 and 2005 and was finally rejected by the main Polish parties, some of which refused to consider any bilingual measures at all, while others were prepared to accept a 50% threshold, which would only allow five municipalities to get bilingual topographic signs and to use minority languages in public administration. After months of parliamentary debate, the initial proposal of 50% was downscaled to 20% on the model of Hungary and Slovakia; this would affect 51 municipalities according to 2002 census data. Threshold solutions thus vary greatly in their application of the territorial principle, with elements both of individual rights (to use the minority language) and territorial arrangements (such as administrative and topographic bilingualism). They are also subject to varying interpretations at local level. There may, for example, be a demanding requirement for minority registration, as in Poland; in Slovakia, even one speaker of Slovak who does not speak the minority language can block its use at municipal council meetings (Marácz, 2011).
In responding to European framework documents such as the ECRML and the FCPNM, the Central and Eastern European states arrived at certain creative but still contested solutions. One of these, as we have seen, is the 'threshold rule' , a type of hybrid system with both territorial and personality aspects. Most Central and East European countries accorded minorities a set of linguistic and educational rights in regions where a 'substantial' number of minority members resided. Formal definitions of 'substantial' vary, sometimes leading to controversy. The 'threshold rule' is arbitrary and politically driven, with great variation from country to country. Thus, for example, at municipal level in Finland an 8% threshold is applied if a minority is to be conceded certain language rights; in Kosovo the corresponding figure is 10%; in Romania, Hungary, and Slovakia the threshold is of 20%; and at regional level in Estonia a much more demanding 50% limit is specified (Dembinska et al., 2014, p. 364).
Even a generous threshold is no guarantee of the survival of bilingual territorial units. Following each census, bilingual municipalities may be threatened with potential withdrawal of their minority regimes. Thus, Hungarian speaking inhabitants of Transylvania's capital, Cluj-Napoca, may no longer use Hungarian in communicating with the municipal administration as their share of the population had dropped to 16.4% by 2011, below the 20% threshold. This still represents a substantial absolute number, 49,565 -much more than in smaller Transylvanian towns with a Hungarian majority, where Hungarian can be used in communication with administrative officials. 1 As Wickström (2019, 2020, p. 10), has argued, threshold rules restricting minority language rights may violate constitutional equality and impede the implementation of minority language rights. Such threshold provisions do not help dispersed minorities, such as Ukrainians in Poland, who were forcibly displaced from their historical territory in the 1940s. The arrangement continues to be subject to intra-state politics and to changing relations between the host-state and the kin-state (Brubaker, 1996;Batory, 2010). In Slovakia, similarly, the 1995 minority law was modified in 2009 for nationalist reasons, restricting the use of Hungarian to the private sphere and penalizing violations of this (Marácz, 2011, p. 167). Elsewhere, arrangements may be more benign. Minorities may, for instance, be protected by a 'grandfather clause' that allows declining minority languages to retain official status even when they drop below the formal threshold (for example, in Finland, provided they remain above 6%). There are similar provisions that discriminate positively in the case of Romansh speakers in Switzerland (Grin, 2010;Stojanović, 2010) and, de facto, in defining Irish-speaking districts in Ireland (Coakley, 2021). We need also to note a caveat: states may adjust the manner in which the threshold criterion is applied by switching from the resident population to the citizen population, or vice versa; minorities normally find it easier to reach the threshold if the resident population is the base. Efforts to achieve optimum balance between equity and efficiency may lead to granting the same rights to all speakers of minority languages, with full implementation of the personality principle for majority and minority alike (Wickström, 2019(Wickström, , 2020.

The case studies
In seeking to explore further the range of approaches to linguistic diversity reflected in language regimes, we need to specify more precisely the manner in which such regimes respond to the sociolinguistic realities that they are designed to address. We may identify at least three domains within which the provisions of language law are relevant. In each, the authorities may interact with the population through a single language, through the medium of two or more languages of equal status, or, in an intermediate position, mainly through a single language. These three sectors may be described as follows: -the central state -including federated entities with full legislative competences -and its agencies: the public representation system and political institutions, the civil service, and public sector bodies such as the education system, health and welfare organisations and economic development agencies -substate public bodies: local authorities at various levels (region, district, municipality, etc.) and any related agencies; and devolved agencies whose jurisdiction is non-territorial, such as cultural councils or institutions connected to sub-state language communities -the private sector, including not just commercial enterprises but also such organisations as churches, sporting bodies and social institutions, which may be subject to specific regulations on language use to promote a 'context of choice' .
While we refer to all three of these sectors in the special issue, our particular interest is in the second one (substate public bodies). We are interested in the way in which such bodies are designed; but in particular we want to explore the manner in which these institutions accommodate linguistic diversity. What kind of language regime prevails -is it based on territorial self-administration (language rules of local authorities) or on the personality principle (voluntary adhesion to cultural councils and linguistic autonomy)? In the latter case, how are these entities formed, and how significant are their functions? In the former case, what kind of 'threshold' rule determines linguistic status (provisions for minority languages)?
In selecting cases for inclusion in this special issue, we are conscious that it would be impossible to cover the extraordinary diversity of Europe's language regimes, no two of which are identical. Instead, we have selected five representative cases which illustrate the challenges facing the modern European state. These cases represent the three main types that we have already discussed (we do not illustrate further the fourth, 'permissive' type, where there is no formal language regime). The three others are of course ideal types, and it is not possible to shoe-horn functioning language regimes comfortably into such categories as the 'preservationist' , 'protectionist' and 'proscriptive' ones discussed above. We need also to take into account the geopolitical context within which the language regime operates, notably relationships with adjacent polities, and the state tradition that marks its bureaucratic culture (Cardinal and Sonntag, 2015a, pp. 119-120), with varying degrees of emphasis on the individual and the group as basic actors (Coakley, 2018).
First, while most European countries can proceed with developing a language regime relatively free from external constraints (notwithstanding the implications of international frameworks such as those of the Council of Europe and related bodies), some feel under particular pressure because they live in the shadow of a powerful state which is vigilant in protecting the interests of its own expatriate kinfolk. Thus, the Baltic republics of Estonia, Latvia and Lithuania face a particular challenge, as discussed by Ádám Németh: their independence from the Soviet Union is still recent -barely three decades old -and two of them, Estonia and Latvia, are home to large Russian minorities (very large in the case of Latvia). The clash between different stances on language policy can be observed in Németh's article. In this case the policy inconsistency is diachronic and not synchronic. Although Russian is currently the language of a demographic minority in the Baltic states, these territories were part of the Russian Empire until 1918 and of the Soviet Union between 1940 and 1990. In the former period, it was German rather than Russian that was the language of the dominant group in Estonia and in the greater part of Latvia; in Lithuania, Polish was the traditional language of the elites. In the Soviet period, Russian took over a hegemonic role in relation to Estonian, Latvian, and Lithuanian. After independence, the three republics moved to a language regime in which the whole territory of the respective republic was now considered as officially unilingual.
Second, although most countries have a relatively uncomplicated history of incremental state building, some are heirs to a complex heritage of geopolitical evolution. An example is Romania, created in its present form in 1918 by merging the former 'old kingdom' (formally part of the Ottoman empire until 1878) with the former Hungarian region of Transylvania and certain other territories. While the 'old' core of the state was overwhelmingly Romanian speaking, Transylvania was deeply divided linguistically between Romanians, Hungarians and Germans, and Hungary for long laid claim to it. This makes the case of Romania, and in particular its Transylvanian region, particularly interesting as a case study in the evolution of language policy. Zsombor Csata, Roman Hlatky, Amy Liu and Ariel Pitre Young take this up in their analysis of linguistic polarisation in Transylvania. Their article goes further than merely describing the linguistic regime at municipal level; it also explores the link between linguistic polarisation and human development. Indeed, they build a bridge between the streams of literature dealing with language policy and development economics by studying the impact on local development of a language regime involving minority language protection. Their quasi-experimental analysis of the position in the municipalities of Transylvania leads to the conclusion that while in the areas where Hungarian is not recognised ethnolinguistic polarisation has a negative effect on economic development, this effect is absent in the municipalities in which Hungarian enjoys official recognition. Their conclusion suggests a lowering of the threshold for recognition, in order to promote local development through minority language rights.
Third, almost all European countries are unitary states; when they are not, the formulation of language policy is obviously more challenging. This is illustrated by the fascinating case of Switzerland, a classic federal state where the evolution of language policy is analysed by Till Burckhardt. Noting the complexities associated with the use of demographic thresholds and quotas, he explores in depth the legislative process which eventually led to the recognition of territorial and personal language policy concepts in the Swiss federal constitution.
Although Switzerland is usually associated with the linguistic territoriality principle, this approach is the consequence of bottom-up linguistic self-determination at the level of the federated entities, and, in the case of multilingual cantons, the municipalities (though there is considerable variation in levels of municipal autonomy across cantons). The initial attempt to recognise linguistic territoriality as a principle was later replaced by a formula in which it is recognised as a tool to reach the objective of linguistic peace, or 'harmony between language communities' . Minority language rights are territorially restricted to the areas of their 'ancestral presence' . This prevents potentially controversial language issues, such as the portability of language rights for internal migrants who are speakers of national languages of foreign descent (such as Italian immigrants in Frenchand German-speaking areas and their descendants). The institutionalisation of the federal language regime departs from this traditional territorial approach by defining language rights on the basis of the personality principle and by introducing a system of representative bureaucracy based on language community quotas. In the case of the Italian-speaking community in particular, but also in relation to the French-and German-speaking communities, this leads to an asymmetry between extensive rights to representation in the federal administration and more restrictive rights to education in the minority language outside its area of 'ancestral presence' .
Fourth, most European countries are marked by the overwhelming dominance of one language group over others. Belgium is an outstanding exception: the demographic and socio-economic strength of the Dutch-and Frenchspeaking communities is fairly evenly matched. Here, the linguistic territoriality principle was implemented as a language policy tool in the last half of the 20th century through a top-down approach, by defining four linguistic territories: Dutch-speaking Flanders, the French-and German-speaking territories of Wallonia, and the bilingual Brussels-Capital Region. The hybrid nature of the resulting system is clear. As Helder De Schutter shows, Belgium defies categorisation in respect of either the territoriality or the personality principles. For complex historical reasons, the territoriality principle enjoys primacy in the Belgian state; but the system becomes hybrid due to the existence of 27 municipalities that offer language facilities for the three official languages, displaying a continuum between the territoriality and personality regimes. The capital region, Brussels, contributes to this hybrid system as an almost perfect case of the personality regime located at the end of the continuum.
Fifth, language policy is normally a response to linguistic diversity, but it may also sometimes be found in a society with a single language of wider communication. This is the case in Ireland, where the ancestral language, Irish, like other marginalised languages such as Breton and Scots Gaelic, has receded over the cen-turies to a few isolated pockets. State policy in Ireland has had a dual focus: on preserving the spoken language in the Irish-speaking districts, and on disseminating knowledge of the language, as the 'national' one, over the rest of the country and in state structures. The tension between these two approaches to language policy is discussed by John Coakley, who examines the apparent failure of language policy either to halt the decline of Irish as an everyday language or to bring the language into more widespread use outside the Irish-speaking districts.

Conclusion
This article and those which follow are intended to explore a central issue in language policy: the manner in which states respond to the demography and geography of linguistic diversity. To conclude this overview, we present the broad findings that may be drawn from these articles. Four stand out: the value of the typology of language regimes that we presented, the porousness of the border between territorial and non-territorial approaches, the significance of the contrasting state traditions of Western Europe on the one hand and Central and Eastern Europe on the other, and the importance of what might be described as the 'powerful neighbour effect' in shaping language policy.
We presented above a simple, four-part typology of language regimes. Three of the categories constitute a type of scale, from the most generous (preservationist, where the state proactively tries to keep marginalised languages alive), through an intermediate category (protectionist, where a language rights regime compatible with minority demands is upheld), to the most restrictive (proscriptive, where speakers of minority languages are prevented from using these in interacting with the state, which recognises only one official state language). The evidence of the case studies suggests that, while this typology may offer useful insights into aspects of language policy, its categories are not ones in which specific language regimes may usually be placed. Not only are there many intermediate points along the continuum that they define; governments may react in quite different ways to language minorities within the state structure. In Scotland, for example, the government is preservationist in relation to Gaelic, but proscriptive in relation to Scots; in Switzerland, it is protective in relation to the three main languages, but also preservationist as regards Romansh. Furthermore, our fourth category, permissive, is not represented among the case studies, but was to be found in certain traditional states, such as the Ottoman empire (Barkey and Gavrilis, 2016). This may, however, in reality be a variant of the proscriptive approach: absence of a formal language policy may simply imply de facto endorsement of a single language as unquestionably the official one, as in the Jacobin model so widely followed in twentieth-century Europe.
Our second conclusion relates to interplay between the territorial and personality principles and the difficulty of differentiating clearly between these in practice. This challenge is picked up by De Schutter (2021), who redesignates these two principles as poles of a continuum, with an indefinite number of intermediate, mixed cases, rather than as clearly defined categories. Indeed, a hybrid system incorporating aspects of both the territoriality and personality principles appears more frequently than has been picked up in the literature so far. On closer inspection, this is also true of the several case studies in this special issue: the Baltic states, Transylvania, Switzerland, Belgium and Ireland. Although in these cases distinctive language regimes operate, these regimes combine in varying degrees aspects of the territoriality and personality principles. Clear-cut instances of these models do not occur in the studies investigated here, notwithstanding McRae's efforts to draw an unambiguous distinction between them.
The third important consideration has to do with the fault-line between West European language regimes and those in Central and Eastern Europe. In the latter cases the federal approach is now rare, and minority languages are given official status on the basis of a mixed system in which the personality principle is applied alongside the territorial one. These hybrid systems rest on a threshold rule, on cultural and linguistic autonomy, and on other individually driven group rights for minority language speakers. This is illustrated in the status of the Hungarian, German and Roma minority languages in Transylvania, in Romania. In some cases the personality principle is not even applied formally, as in the Baltic cases, where official rights for Russian speakers are substantially withheld. In West European polities, language regimes are usually rooted in well-established republican state traditions in which the concepts of 'national minority' and 'minority language' are not defined precisely, and census questions on language have in general been introduced rather reluctantly and belatedly. If a question on language is asked at all, it is more likely to refer to knowledge of languages rather than, as in Central and Eastern Europe, to mother tongue, habitual language or domestic language (Coakley, 2018, pp. 260-265). In the federal system of Switzerland, as discussed in this special issue, and of Belgium, the territoriality principle is implemented to promote different monolingual 'contexts of choice' , although in both cases partial systems based on the personality principle are present.
Our fourth noteworthy observation has to do with the 'powerful neighbour effect' . Many language communities (especially smaller one) operate under the shadow of a neighbouring state whose dominant geopolitical weight endows its language and culture with formidable advantages over smaller rivals. Thus, the English language has continued to push Ireland's indigenous language aside, a process that continued even after Irish independence. The Baltic republics -and especially Latvia -experienced similar pressure from Russian. The protectionisttype language regime of the interwar period in Latvia was succeeded after annexation to the Soviet Union in 1940 by new citizenship and language regimes, with Latvian nationality dissolved into Soviet citizenship and the Latvian language becoming de facto a territorial minority language alongside Russian (by 1989, Latvian speakers constituted just half of the population). After independence in 1990, the citizenship law was nullified (Bowring, 1994), and Latvian nationality was reestablished on the jus sanguinis principle: it was confined to those who were Latvian nationals in 1940 and their descendants. Others -mostly ethnic Russiansbecame 'non-citizens' , and were thus, as presumed 'migrants' , excluded from the provisions of the ECRML (to illustrate the impact of the new definition of citizenship, ethnic Latvians, who constituted 52.5% of the resident population in 1989, now became 78.5% of the citizen population; Coakley, 2012, pp. 226-227). A high threshold requirement for the recognition of minority languages at municipal level has created further difficulties for Russian speakers, though they still fare reasonably well by comparison with Bosnians in Austria, Algerians in France and French speakers in Leuven or Bern.
Notwithstanding the simplicity of our four-category typology of language regimes, or the revealing character of the territoriality-personality dichotomy, attempts to fit actual state approaches to the management of linguistic diversity into these categories are likely to be only partly successful. This is the case even if we restrict ourselves to traditional regional languages, to the exclusion of new immigrant languages, which raise further difficult normative questions (Grin, 1994). Moreover, the notion of territory is evolving in a new geopolitical order, shaped by strengthened non-state actors, global supply chains, new migration flows, and the reality and potential of digital communication in transcending territorial boundaries (Badie, 2014;Edquist, 1997;Scholte, 2005). As ideal types, the categories we have used are revealing; but most (if not all) language regimes are hybrid ones. The status and rights of minority languages will be determined in the best case scenario by a mixture of the territoriality and personality principles which cannot guarantee full equality of the languages involved. Equality and democracy can only be satisfied in such cases if there is an evolution from a mixed system to one where the territoriality principle is applied to all languages with a significant presence, as advocated by Jean Laponce and Philippe Van Parijs; but reaching this position in the real world is a formidable challenge.