Contesting European regions

ABSTRACT Contesting European regions. Regional Studies. A regional or ‘meso’-level of regulation and policy-making has emerged in Europe. This cannot adequately be explained by functional imperatives or drivers. A constructivist perspective sees the region as the outcome of political contestation over the definition and meaning of territory. Six competing conceptual frames for regionalism are proposed: integrative; competitive; welfare; identity; government; and the region as a refraction of social and economic interests. Any given case will reflect a balance among these conceptions. Such an understanding permits a combination of comparative analysis with an understanding of individual cases and avoids both dismissal of territory and territorial determinism.


CONCEPTUALIZING REGIONS
This article is a reflection on the constitution of the region in the 50 years since Regional Studies was founded. The term 'region' is used in multiple ways in the social sciences. In international relations, it refers to the supranational level. In other disciplines, the focus is on the sub-state level, at a number of different scales. Regions can also be conceptualized as inter-state spaces, crossing state boundaries. Of interest in this article is the rise of the intermediate or 'meso' (Sharpe, 1992) region, between the state and the local level.
The rise of the region has often been explained by reference to functional imperatives or drivers. Some emphasize economic globalization, which is eroding the capacity of the nation-state and enhancing the efficacy of smaller units in facing the challenges of competition and meeting citizen preferences (Ohmae, 1995). Others put forward arguments about efficiency and the optimal level for service delivery and regulation (Hooghe & Marks, 2009). Some point to the tendency of government to respond to underlying patterns of social identity (Erk, 2007;Kymlicka, 2007). Functional arguments, however, are teleological, explaining change by reference to its effects. At best, functional arguments present reasons for changing the spatial scale of government, but reasons are not the same as causes. Reasons are, moreover, normatively contested as they rely on some transcendent objective. For example, arguments about large or small units of government have shifted over recent decades, from consolidation in the interests of planning and economies of scale to fragmentation (under the influence of public choice theory) in the interest of competition and variation. Regions are too different in their geographical scale and institutional realization to be explained by a single set of drivers. Determinist theories also underplay the role of agents and political competition.
Instead, a constructivist approach sees regions as the outcome of contestation among social and political actors in specific conditions. It does not take the ontology of regions for granted but sees them as potential spaces to be filled in with social and economic content, and often contested. This is consistent with modern theories of territory which have moved away from a rigid definition of boundaries and emphasized flexibility and multiple meanings. Territory is not just a topological concept but a sociological one, which is socially constructed (Lefebvre, 1974) and constituted by its social and economic content and its utility in explanation of social processes and outcomes. Territory is constructed in two senses: its definition and meaning are a matter for interpretation by social, political and economic actors and by citizens; and actors themselves seek to construct systems at particular spatial scales and give them particular meanings. Such an approach also meets the demand for conceptually consistent but contextualized understanding (Rutten & Boekma, 2012;Storper, 2011a); but this makes it difficult to do comparative analysis. Finally, territory is contested in that its definition, significance and uses have important outcomes for the distribution of power and resources. There is no 'objective' or purely technical definition of the region or the 'right' spatial level at which to conduct particular policies or regulate economic, social and environmental systems. Instead, different conceptualizations of the region have developed across time and in different places, and have competed with each other.
Spatial rescaling is currently transforming the nationstate as economic, social and political systems that previously were largely contained within its boundaries migrate to new levels (Brenner, 2009;Brenner, Jessop, Jones, & MacLeod, 2003;Keating, 2013). The outcome is not a single territorial grid but a multiplicity of possible spaces and constructions (Goodwin, Jones, & Jones, 2012;Paasi, 2009;Pike, 2007;Varró & Lagendijk, 2013). None of this means that the region should be dismissed as a category of analysis or as not 'real', but rather that it is a conceptual rather than a reified phenomenon. So this produces multiple regionalisms, which may or may not coincide. This article does not seek to create taxonomy of regionalisms in which each case belongs in a given category since that would simplify matters and not aid understanding. Instead, six conceptual or interpretive frames are proposed which explore diverse ways of constructing the region as an economic, social and political project. The frames refer to the key dimensions in the social construction of territory and are: integrative regionalism; competitive regionalism; welfare regionalism; identity regionalism; regions as government; and regionalism as a refraction of social and economic interests. 1 Any given case will probably be open to more than one interpretation and it is the contest among these that produces the outcome.

INTEGRATIVE REGIONALISM
Mainstream social science was long wedded to a vision of modernization that saw it as a process of territorial integration and functional differentiation (Finer, 1997;Parsons, 1971); Emile Durkheim (Durkheim, 1964, p. 187) declared that 'we can almost say that a people is as much advanced as territorial divisions are more superficial'. This perspective was influenced by the advance of industrial production with its distinct rationality, the breakdown of traditional and peripheral societies in the late 19th century, the creation of unified markets, cultural integration, and the institutional incorporation of territories. Territorial integration occurred largely within the rigid territorial parameters of the nation-state but, within these, students of state-and nation-building emphasized deterritorialization (Deutsch, 1972;Giddens, 1985;Lipset & Rokkan, 1967).
By the 1970s there was more appreciation that national integration was not always complete and that significant regional economic, social and political cleavages remained within nation-states (Rokkan, 1980(Rokkan, , 1999. Then came a recognition that regional differentiation was not just the legacy of older, pre-industrial society, but was reproduced under conditions of modernity (Tarrow, 1978). Focus then moved towards territorial management, the strategies that states use to integrate peripheral territories into national political and economic systems (Keating, 1988;Rokkan & Urwin, 1983).
From the 1960s, the region became a key unit for policy design and modernization. The context was the 'Keynesian welfare state' resting on the twin pillars of national economic regulation and national welfare standards. The territorial counterpart was 'spatial Keynesianism' (Brenner, 2004) in the form of regional policy designed to overcome market imperfections and integrate lagging regions into national economic space. Regions, and local governments, were used to extend public services and welfare provision.
Post-Second World War regional policies tended to be depoliticized and integrative, focusing on technical logic.
Regional policy was essentially positive-sum, offering declining regions help to develop, boosting the national economy by mobilizing idle resources, and helping wealthy regions by relieving congestion and providing markets for their goods. Preferred institutional mechanisms were development agencies and centralized allocation of resources.
Starting in the 1970s and 1980s, the European Union (as it now is) pursued essentially similar strategies. The European Regional Development Fund (later Structural and Cohesion Policy) sought a spatial dimension to the single market by overcoming structural obstacles and facilitating an optimal spatial division of labour. Once again, the reasoning was integrative and technical and the European Commission has consistently presented an economic, rather than a redistributive, justification for the policy (Begg, 2010;Keating, Hooghe, & Tatham, 2015). Its own spatial map of NUTS (Nomenclature des unités territoriales statistiques) reflected a purely technical logic.
In practice, depoliticized and integrative regionalism proved difficult to sustain. Both states and the European Commission have had to institutionalize the policy and provide for delivery mechanisms. They have sought to engage local economic, social and political actors in development policy, sometimes seeking to marginalize existing local elites in the name of modernization and sometimes co-opting them. There were regional development commissions, public-private partnerships and mechanisms for citizen input. The European Union shied away from interfering with structures of territorial government, which are the prerogative of the member states. It did, however, insist on the principles of subsidiarity, taking issues to the local level, and of partnership, including local civil society. The whole process inevitably raised questions about development priorities and whether state visions of the spatial division of labour corresponded to local priorities. Development based on a technical, economic logic was challenged by opposition groups and social movements concerned with the social implications of change (Keating, 1988). The result was a politicization of regional development and a contest for the definition of the region, its economic and social meaning, and its institutionalization by states, the European Union and regional actors.

COMPETITIVE REGIONALISM
Since the 1980s, there has been an important reconceptualization of regional economic development. The region is seen less as an object of state and European policy and more as an economic unit in its own right. One factor has been the decline of national regional policies, only partly compensated by European regional policy. The opening of national economic borders through global and European free trade and capital mobility has hampered inter-regional redistribution and diversionary policies. The old logic, by which transfers to poor regions came back to the wealthy regions as orders for their goods, no longer holds as consumers can spend their money on imports. Regional policy is expensive and less effective and is restricted by European competition policies.

Contesting European regions 11
Both national and European regional policies have been affected by new thinking about spatial economic development captured by the term 'new regionalism' (Keating, 1998), which stresses the importance of the local and regional scale to the understanding and steering of economic development (Cooke & Morgan, 1998;Scott, 1998;European Commission, 2013; European Commission Directorate-General for Regional and Urban Policy, 2014). New Economic Geography (Krugman, 2011) stays close to classical economics in emphasizing how the proximity of producers can lower transaction costs and exploit economies of agglomeration in supply chains or 'clusters' (Porter, 2001). Other accounts (Gertler, 2010;North, 2005) focus on the role of institutions in promoting that balance of competition and cooperation in which markets thrive. Attention has moved from the traded dependencies of transaction costs models to 'untraded interdependencies' in the form of tacit knowledge and face-to-face exchange, an idea dating back to Alfred Marshall (Marshall, 1920). Institutional accounts fade into sociological accounts, focusing on the characteristics of local societies, including social structures and mechanisms for sharing knowledge and fostering collaboration. These, in turn, fade into cultural explanations, focusing on the characteristics of the population, including social capital and trust (Malecki, 2012;Putnam, 1993).
Regions and localities are portrayed as more than mere locations of productions but rather as production systems, with their own internal logic and interdependencies but linked into global trading systems and supply chains (Crouch, Le Galès, Trigilia, & Voelzkow, 2001). In this way, regions are constructed on the basis of economic factors, but using economic sociology rather than neo-classical principles.
A further move has been to portray these local and regional production systems as being in competition as Ricardian comparative advantage (in which each territory has an optimal role in the spatial division of labour) gives way to absolute, or competitive, advantage in which some territories can gain an unassailable lead (Kitson, Martin, & Tyler, 2004;Scott, 1998). This goes against a principle of neo-classical economics that only firms compete. The concept of regional competition and competitiveness has been strongly criticized as incoherent (Bristow, 2005(Bristow, , 2010. It reifies regions as the appropriate unit of analysis; even if territory is relevant, it does not necessarily correspond to regions in other senses. Yet this is precisely the way in which the theme is used to construct the region. Ohmae (2001) is impatient with political concepts, insisting that regions emerge from market order; Alesina and Spolaore's (2003) theories about the 'size of nations' (by which they mean states and regions) do the same. Regional politicians can use the theme of competitive growth to construct a shared regional interest, expanding their home electoral and social support base. States and the European Union can use it to step back from diversionary regional policy and equalization, putting the responsibility back to regions. Competition also rests upon a dominant form of justification in the modern world by evoking the market.
The notion of the competitive region potentially narrows the agenda for regional policies to a concern with productivity and market advantage (Bristow, 2005). It is true that policy-makers have sought to rebut criticisms that this amounts to entrenched neo-liberalism by extending the definition beyond the narrow criterion of production costs. The European regional competitiveness index (Annoni & Dijkstra, 2013) combines disparate factors and refers to matters that are assumed to enhance productivity. Others open up beyond considerations of firm productivity to include social and environmental factors (Aiginger, Bärenthaler, & Sieber, 2013;Aiginger & Firgo, 2015; Organisation for Economic Co-operation and Development (OECD), 2014). These might be relevant to a discussion about regional productivity and welfare, but they do not address the central issue about competiveness: that it is inherently relational and a zerosum game. This puts competitive regionalism at odds with other concepts of the region, such as welfare.

WELFARE REGIONALISM
If regions are in competition for absolute advantage, it follows that there will be an increase in inter-regional disparities. The economic marginalization of regions in the global division of labour and the territorial impact of plant closures in turn promotes the idea of the region as a defensive space and a 'revolt of the poor'. There is also a 'revolt of the rich' as politicians in prosperous regions cite the need to compete in European and international markets to complain about the burden of transfers to poorer compatriots. Such transfers, whether explicitly through fiscal equalization and regional policy or implicitly through national welfare programmes, have become increasingly salient as the region has become a recognizable unit and regional accounts are available. So demands to limit transfers have become a major political issue in Germany, Italy, Spain, Belgium and the UK. On the other hand, systems of fiscal transfer have proved quite resilient. They are institutionally embedded in national political systems and survey evidence suggests that public support for inter-regional transfers remains rather high. Inter-regional solidarity is weaker, however, in wealthy regions and those with a strong sense of distinct identity (Henderson, Jeffery, & Wincott, 2013). This raises normative issues about inter-regional equity and the idea of territorial justice (Storper, 2011b).
Rescaling also presents questions about intra-regional equity. Competitive development is presented as in the interests of all within the region. Yet any given development strategy will produce winners and losers, whether these be defined by class, gender, age or location. Another possible consequence of inter-regional competition is that of a 'race to the bottom' as regions, needing to attract investment and wealthy taxpayers, cut public expenditures, especially on welfare, along with taxes. It is for this reason that theories of fiscal federalism have traditionally advocated that redistributive policies should be located at the higher, federal level, where externalities can be taken into account (Oates, 1999). It is also why for most of the 20th century the social democratic left tended to favour centralization while those on the market right favoured decentralization.
This can no longer be taken for granted. In some cases, regions have been constructed as sites of opposition to austerity policy and locations of social solidarity. This is clearly visible in Scotland, where support for devolution and independence are linked to themes of welfare, but is also visible in cities and regions elsewhere. Offsetting the race to the bottom, there are also signs of a 'race to the top' in the provision of public services as regional parties and leaders compete to innovate and impress their electorates (Costa-Font & Rico, 2006;Gallego & Subirats, 2011) or to the middle as regions converge (Dupuy, 2012). In regions with a strong identity, the affective solidarity that underlies welfare states may play at that level, so creating alternative communities for sharing (Béland & Lecours, 2008;McEwen & Moreno, 2005).
The welfare state has been both restructuring and rescaling spatially (Ferrera, 2005;Hemerijck, 2013;Kazepov & Barberis, 2008). There is a move away from passive social assistance towards active labour market policy, aimed at getting people into work. The older pattern of need based upon low incomes or periodic unemployment has been supplemented by new social risks linked to precarious employment, disability and other forms of disadvantage. Activation policies are increasing linked to local and regional labour markets and to other spatial policy instruments, notably economic development. In many European countries, they are managed by partnerships of business and labour, often with a European input via the social funds. This has strengthened the region as a space for social compromise and negotiation and for policy delivery.
Governments and the European Commission seek to reconcile competitive and welfare regionalism with successive formulas like socially inclusive growth, but the dilemmas remain. The 'place-based' approach (Barca, 2009) yet again rehearsed the need to promote both growth and social inclusion, but also recognized that they are distinct. When policies are put into practice the conflicts between the economic and the social often re-emerge. Economic and social issues are thus increasingly presented in a territorial framework, but this does not mean that there is really as shared regional interest, that regionalism is caused by economic or social grievance. In order to express such grievances territorially, there has to be a territorial framework, identity or repertoire of symbols to map onto. These are provided by territorial identities, government, and the territorialization of economic and social interests themselves.

IDENTITY REGIONALISM
Regions are sites of social identity formation, which can be integrative or autonomist. Integrative regionalism goes back to the late 19th century, seeking to sustain historic identities and loyalties, but within a national political modernizing project (Applegate, 1990;Núñez Seixas, 2001). It competed with autonomist regionalism, which re-emerged in the 1970s in the form of the demand for recognition of national diversity. It is tempting to make a distinction between regionalisms and minority nationalisms, the latter being committed to separatism, but this is too simple. Since the late 19th century, the term 'nation' has carried with it the connotation of a right to self-determination, so movements asserting the existence of a distinct people with the right to set their own constitutional future have adopted the term, putting them on the same normative plane as states themselves. In the modern world, however, the nature of statehood has been changed by transformation of the state and the emergence of new conceptions of sovereignty. The degree of self-government desired is an empirical question to be examined in context, not a defining feature.
Some observers have used the terms 'ethnoregionalist' (Newman, 1996) or 'ethnonationalist' (Connor, 1994) to denote the new movements, but that is to introduce the difficult, contested and normative language of ethnicity. Substate nationalist movements in Catalonia and Scotland have taken pains to disavow ethnic particularism in favour of a 'civic nationalism' that is inclusive of the whole population, including incomers (Keating, 2001a(Keating, , 2001b. Basque nationalists have abandoned the ethnic exclusivism of Sabino Arana, who founded the Basque Nationalist Party at the end of the 19th century. Some movements in the Balkans have sought to distance themselves from the prevailing ethnic nationalism by stressing a consciously multi-ethnic regionalism (Stjepanovic, 2012). The term 'regional nationalist' (Keating, 1988) is less prejudicial although not uncontested.
Regional nationalism in the 1970s was widely attributed to retarded modernization or a 'revolt against modernity' (Lipset, 1975) in line with contemporary modernization theories, but has since been subject to a theoretical reappraisal as part of the latest phase of modernity (Olsson, 2009). While all nationalisms look both back and forward, the new regional nationalisms are characterized by their relationship to the new opportunities created by rescaling.
A key element in this is the use of Europe as a new framework for the discursive projection of the political community and for new forms of autonomy (Duerr, 2015). For some, like the Scottish National Party (SNP), the European Union lowers the cost of independence, since it guarantees market access and limits the power of large states by pooling authority. Others have gone further, seizing on the pooling of sovereignty in Europe to make a 'post-sovereign' argument for autonomy within a multilevel system of Europe, states and regions (Keating, 2004). This was long the position of mainstream Catalan nationalism and is now the dominant strand in the Basque Nationalist Party. Europe thus provides a discursive space in which to articulate demands for recognition as something more than a region defined by the state constitution. There were also hopes that Europe could provide more concrete opportunities for regions to act in transnational space and exercise real power, focused on the movement for a Europe of the Regions. The Maastricht Treaty established the Committee of the Regions (CoR) (Piattoni & Schonlau, 2015) and gave regions access to the Council of the European Union, Contesting European regions 13 but as part of national delegations. The CoR never gained substantial power, one reason being that regions are defined very differently in different member states and the European Union has not cultivated regions as rivals to the states but rather uses them as objects of policy and territorial integration. Its key term is not autonomy but territorial cohesion. Justifications by those claiming self-determination have moved away from primordial conceptions of the nation towards democratic principles, based on the right of people to decide their own fate. Yet, as Jennings (1956) argued, the principle of self-determination of peoples makes little sense unless someone can first define the people. Contemporary regional nationalists construct their political community in a variety of ways. One is by historical interpretation and revision, to establish the ontological reality of the political community as something constructed through time, if not primordial. This provides a counternarrative to state-based histories founded upon a teleology of integration and unity. It is linked to arguments about the foundations of political authority and sovereignty (Keating, 2001b). Basque nationalists insist on the primacy of their historic rights (fueros) as pre-constitutional, while Spanish nationalists claim that all political rights derive from the 1978 Constitution (Herrero de Miñon, 1998). Scottish nationalists reject the absolute sovereignty of the Westminster Parliament and argue that Scottish conceptions of sovereignty are distinct and pluralist (MacCormick, 1999). Catalan nationalists take as a reference point Catalonia before 1714 as an autonomous trading nation within a federal Crown of Aragon, itself nested in a Spanish confederation and with links across Europe and the Mediterranean. Flemish nationalists point to the glorious era of Flanders as a cultural centre and trading nation in the early modern period. Of course, these accounts have varying degrees of plausibility both in the units to which they refer and in the historical experience. Flanders as a historical reference point is rather different from the modern region, both in geographical scope and social and political meaning. Padania (northern Italy) is a modern invention in spite of efforts by Northern League propagandists to present it as 'the oldest community in Europe' (Oneto, 1997). Yet such historical narratives are not mere nostalgia but, as often happens with history, project contemporary debates into the past, while giving their own nation a modernist, progressive and democratic teleology to match that of the states themselves. Normative assessments, however, have varied from seeing them as forms of emancipatory democracy (Gagnon, 2014;Guibernau, 2013) to condemning them as unfit to join the community of European nations (Weiler, 2014).
Regional nationalist movements have adopted the themes of the new regionalism to construct both their political communities and their policy prospectuses. So in some wealthy regions, the region/nation is discursively constructed as a space of economic dynamism, endowed with competitive capacity. Tradition is linked to modernity, as in the Bavarian slogan laptop und lederhosen and, far from being an obstacle to modernization, is seen as a source of social capital as the region is credited with just those virtues that underpin new regionalism (Keating, Loughlin, & Deschouwer, 2003). Social solidarity is also pressed into service, so that the region/nation is seen as essentially cohesive, albeit in different ways. Scotland and Wales have been constructed as bastions of social democracy, while Flanders is presented as more individualistic and less welfaredependent than Wallonia; northern Italy is portrayed as being less dependent than the south.
It is very difficult to measure how far popular identities have shifted in the direction of sub-state regions since identity is a multidimensional concept and national surveys use different meanings and questions. On the rare occasions when the same question has been used cross-nationally, its meaning is different in different places. The Linz/Moreno Question (Guinjoan & Rodon, 2016), asking people on a five-point scale whether they identify with the state or the sub-state nation, has been asked regularly in Spain, the UK and Belgium (but in Belgium Europe and the local level are also included). The main finding is that even in places with strong regional nationalist movements, most citizens have developed dual identities rather than identifying exclusively with one level or another. Identity is thus not a given but another field by which regions can be constructed and given meaning.

REGIONS AS GOVERNMENT
These multiple meanings of region do not produce a definitive political outcome or institutional response. Some conceptions point to regions as arms of the central state; others to regions as self-regulating economic systems; others again raise the issue of political autonomy and citizen representation. The combination of these conceptions has, however, encouraged moves towards elected regional government. States have sought mechanisms to recapture functional systems that have escaped their purview through rescaling, and also to incorporate territorial social and political actors. They, and the European Union, seek legitimacy for their new modes of intervention, territorial administration and regulation. Opposition movements, challenging economically driven or technocratic regionalism in the name of social considerations, seek to broaden the agenda of regional policy and expand the range of stakeholders. Regional nationalists seek autonomy as a first-order objective. These competing visions have led to the politicization of the regional question. The institutional response has been the establishment of representative and accountable government as the only way to broker and compromise the divergent meanings of regionalism and to provide legitimate outcomes. There has been a rise in regional government (Hooghe, Marks, & Schakel, 2010), but the response is far from uniform, and some states have been reluctant to establish territorial government that might correspond with political identities. This was long the case in France and remains so in Central and Eastern Europe (Yoder, 2012).
Much has been written about a move from government to governance (Bellamy & Palumbo, 2010). The latter term is notoriously elastic but usually refers to a mode of regulation based on networks rather than hierarchy. Multilevel governance in turn refers to networks spanning spatial levels and the public and private sectors (Bache & Flinders, 2004;Hooghe & Marks, 2001;Piattoni, 2010). As a general comment on the state of the world this might be unexceptionable, as any social system can be analyzed by territory and function. The claim that the world is moving from government to multilevel governance, however, suggests that there is a shift from the multipurpose, centralized, hierarchical state to a looser order of networks. The term 'governance' (and its multilevel variant) is applied with particular frequency to the European and regional levels, probably because of the lack of strongly integrated governmental institutions. Yet, in so far as governance refers to networks and loose constellations of institutions, the trend is away from this, to elected, multipurpose regional government. Governance may be no more than a transitional stage towards government as new scales are institutionalized (Goetz, 2008).
Elected regional government has the effect of fixing the territorial scale and boundaries of regional systems, much as the nation-state 'caged' social and economic systems (Mann, 1993). Regions, however, are less tightly bounded than states, given the competing territorial imaginations and the ease of functional systems in economy and society to escape their borders. These borders in themselves have little that is natural about them but reflect historical patterns and the balance of political forces. Having been established, however, they show great resilience as parties and leaders use them to establish power bases and institutional resources. Even the most apparently arbitrary boundaries then become entrenched, as in the German Länder. The French regional reform of 2015, merging regions into larger units, is an exception, and these were among the least socially embedded regions in Europe.
The rise of the meso has created an effective system of multilevel federalism, even if the term 'federalism' is not always used. Students of federal government make a distinction between coordinate and cooperative federalism. In the coordinate ideal type, each level of government has its own competences, which it exercises independently. In the cooperative mode, competences are shared and the emphasis is on joint policy-making, within vertical policy communities spanning the two levels. During the Keynesian welfare era, the emphasis was on cooperation. There is now a growing interest in a third variety, competitive federalism (Dente, 1997). Regions compete for investment, technology and markets, as in the competitive regionalism discussed above. They also compete over policy innovation and service provision. Governments in stronger regions have called for more autonomy and disentanglement of central and regional competences. This is visible in Germany, Belgium, Italy and Spain, where regions have complained about the centre using framework laws to restrict regional autonomy. There has also been pressure from richer regions to devolve tax power; poorer regions have been much less keen on the idea. Poorer German Länder have slowed down federal reforms and Wallonia is less enthusiastic about further federalism than Flanders. So the tension between integrative regionalism and regionalism as differentiation has increased.

REGIONS AND THE REFRACTION OF INTERESTS
Modernization and integration theory tended to assume that the growth of class and sectoral alignments would efface territorial differences within the nation-state. Even after the emergence of regional government, Pastori (1980) could write of 'regions without regionalism ' and Trigilia (1991) and Le Galès (1997) of the 'paradox of the regions'. The idea was that regional governments had been set up, but the articulation of social and economic interests continued to be on a national scale. Regions were, in other words, not constructed around substantive interests, which left regional arenas rather hollow. Experience in recent years, however, shows a more complex picture. Class and sector do not displace territory but are refracted by it in distinct ways at different territorial levels under the twin influences of rescaling of functional systems in economy and society, and the emergence of regional government, which encourages an institutional isomorphism in which interest articulation reflects government structures. The differential territorialization of sectoral and class interests thus becomes a key factor in the construction of the region. This process has not been fully explored in the literature, apart from the work of Keating and Wilson (2014), which explores the organization and territorial orientation of representative groups and the horizontal and vertical relations among them and with government.
Business groups are aware of the importance of territory to economic development and how context affects the success of investment. Large firms take great care in choosing investment locations and assuring the supply of infrastructure, labour and technological innovation. Large business, however, is wary of capture by territorial political systems, especially where left-wing or environmentalist forces are strong. They tend, therefore, to favour functional regionalism, depoliticization of development policy and agencies dominated by business interests. Small businesses are less mobile and more dependent on local public goods and support, on local markets and, in some places, protectionist networks to defend local traders. They are often closer to local political concerns as their owners and managers are local citizens and may share local identities. This makes them more supportive of regional structures and government, sometimes combining free-market rhetoric with practical protectionism and dependence on territorial government.
Trade unions are in principle based on class and have historically expanded across national territories, overcoming local particularisms. They favour national welfare standards and in some states have been incorporated into statelevel forms of social partnership and social compromise. Increasingly, however, they have been drawn into local and regional alliances in defence of threatened sectors, which have taken on a territorial as well as a sectoral focus. They have adopted new regionalist themes of economic development, emphasizing human capital, labour market policies and activation. They seek at regional level to recover positions they lost with the demise of nationallevel corporatist structures and collective bargaining. Consequently, they have moved away from their previous emphasis on the central level to embrace regional arenas and decentralization. As mass organizations (albeit declining in numbers) they are also open to popular forms of identity and must take account of new regionalist and minority nationalist politics in places like Spain, Belgium or the UK. So they are cross-pressured, which generates internal tensions.
Environmentalist groups often have local origins and the impact of environmental change is first experienced locally; but regulation is at a wider level. Environmental groups remain locally rooted and are often in alliance with groups defending local traditions. At the same time, they are connected to the European level, which is more accessible and capable of making rules binding all the way down the spatial scales.
Social and economic interests are thus refracted by territory, with the emergence of a territorial-level interest articulation. This may take the form of the construction of a common territorial interest, underpinned by the rhetoric of territorial competition. Alternatively, the region may become an arena for interest intermediation and compromise. Regional policy communities may emerge within or across sectors. There may be new social alliances and oppositions; at the regional level there is often a productivist alliance of business and trade unions against environmentalists.
The rescaling of policy communities thus serves also to construct the region as a space of social compromise, facilitated by regional governments, which seek to incorporate interests as a means for strengthening regional capacity and by the preference of the European Union for the incorporation of civil society in the delivery of its own development programmes.

CONCLUSIONS: CONTESTED REGIONS
Regions have been studied from a variety of disciplinary perspectives: economic, legal, geographical, sociological, historical, cultural and political. None is determinant but an interdisciplinary approach enables a richer understanding of the phenomenon. The constructivist approach allows one to see territory as a field in flux, with multiple influences at work and no definitive outcome. As the state's monopoly on the definition of territory is weakened, the field is more open and contested. In some places the very foundations of sovereignty are challenged, while in others there are territorialized struggles over resources. There is competition between states and the European Union over the definition of regions for spatial policy interventions. The drawing of regional boundaries influences patterns of inclusion and exclusion and the balance of political and social forces. The same is true of the internal constitution of regions. Elected regional government has often been in tension with corporatist forms of representation. Regions as vehicles for state policy are in tension with regions as a form of territorial autonomy. Regions are arenas for playing out some of the most important political issues such as the balance between economic competition and social solidarity.
Yet regionalism is not a one-way project leading to a 'regional world' or Europe of the Regions. States remain important actors in shaping space and, at a time of austerity, have sought to reinforce their control. European requirements on debts and deficits have led the Eurozone states to introduce constitutional limits applicable to all levels. In Spain, austerity has had asymmetrical effects, with Catalonia pushing towards independence while other regions, unable to take the burden, have asked the state to take competences back. France has imposed mergers of regions in the name of economy, while Italy has rolled back promises of federal reform and extensive devolution. The European order is characterized by both centrifugal and centripetal tendencies.
The six conceptualizations proposed here provide frames of analysis for understanding the complexities of rescaling and the rise of the region. This is not an exhaustive taxonomy, nor is it an empirical finding based on hard facts, but a means of identifying key aspects of the social reality and interpreting cases. It is offered not as a conclusive way of understanding regions but as a means of capturing the fluidity of relationships while not destroying the basic object of study. In this way, it contributes to understanding how social, political and economic relationships are refracted by territory, while avoiding territorial determinism.

DISCLOSURE STATEMENT
No potential conflict of interest was reported by the author.

NOTE
1. It is of the essence of the constructivist approach that this schema is not the product of inductive reasoning or empirical generalization. It is, rather, a scheme of interpretation whose test is not a comparison with a hard empirical reality but rather its utility in explanation and understanding (Hacking, 1992;Kratochwil, 2008). It is to be judged not by whether it is right or wrong but according to whether it is useful. By the same token, it is not claimed that other conceptualizations are wrong but merely that they should be judged in the same way.