Does the European Union Need to Become a Community?

Many theorists contend that for the European Union to become a viable democratic polity its citizens must develop an overarching communal identity. I take issue with this claim, arguing that the norms, motivations and perceptions that make supranational democracy possible can also emerge through processes that do not presuppose shared communal identifications. These include the gradual externalization of domestic democratic norms and practices to the EU level, the incorporation of the resulting supranational democratic attachments back into existing national identifications and the build‐up of transnational political trust propelled by the practice of supranational democracy itself. Such an outcome is not inevitable, but it is conceivable in that it is theoretically coherent and has limited empirical analogies and precedents. The range of options for the EU's further democratic development is therefore broader and the chances of its success greater than many analysts assume.


Introduction
Analysts of European integration often contend that the European Union must build not only institutions but also an identity. Assuming that the Union cannot get by on (at any rate uncertain) popular perceptions that it is economically beneficial, they argue that only a shared sense of belonging to an overarching European communal unit could help Europeans develop the trust and commitments a democratic polity needs. Past and present difficulties in the integration process, accordingly, are often interpreted as a symptom of Europe's alleged 'identity deficit' (Fahey, 2009, p. 35) and 'community deficit' (Etzioni, 2007, p. 23), ranging from the defeat of EU treaty referendums and the rise of Eurosceptic political parties in several Member States to the current crisis in the European monetary union (EMU). As Gideon Rachman (2011, p. 13) argues in a widely cited commentary, the reluctance by many Member States to stabilize the currency project through deeper fiscal integration and the issuing of Eurobonds reflects 'a lack of a common identity' that is 'the single currency's true fatal flaw'. Weak overarching identifications impede attachments to the EU while ' [t]he fact that national loyalties are much stronger than any common European loyalty means leaders are constrained in the solutions they can feasibly consider' (Rachman, 2011, p. 13; see also Auer, 2011).
Even so, the 'identity deficit' hypothesis divides theorists of European integration. For its proponents, it often serves as a basis for contemplating the themes and symbols around which a putative European identity could (and should) revolve (Habermas, 1991;Howe, 1995) and the potential mechanisms of its emergence (Etzioni, 2007;Eriksen and Fossum, 2004). Others, by contrast, question not only the desirability and/or viability of a potential European identity-building project as many of its advocates define it (for critiques, see Nicolaïdis and Pélabay, 2009;Castiglione, 2009) but also, frequently, the belief that such a project is indispensable to secure popular support for the EU.
Such objections draw in part on a broader claim regarding the possible sources of political loyalties. While some social and political projects that demand strong commitments from their members typically do attract communal identifications (the modern nation-state often serves as the archetypical example), few theorists would argue that this applies to all associations, organizations and collaborative projects that engage people in their everyday lives. In Andrew Mason's terminology, some of these may instead be social formations people interact in and 'belong to' without necessarily identifying themselves as 'belonging together' as members of a single overarching communal unit (Mason, 2000, pp. 127ff.;Honohan, 2001). By some accounts, even the EU itself has long displayed such qualities. On the one hand, popular identifications with the Union and citizens' self-identifications as European or EU-European have remained relatively weak and heavily concentrated on sections of the better off and better educated (Fligstein, 2008, chapter 5;Eurobarometer, 2008, p. 67; also see further below). On the other hand, however, average popular support for the EU, for one's country's membership in it, and even for an expansion of EU competences into areas that promise few direct material benefits to individual citizens (for example, foreign policy) has consistently been much stronger by comparison. Such support prevails even in several of those Member States that have experienced regular flare-ups of popular opposition to specific EU projects such as the common currency and the ill-fated Constitution Treaty (Díez Medrano, 2009). If one accepts this, it follows that at least some of the support the Union currently enjoys derives from sources other than a sense of membership in an overarching community of Europeans.
Applied to the EU's future development this proves and disproves nothing, but it cautions against premature theoretical closure. It suggests that efforts by some theorists to contemplate models of supranational polity-building that do not depend on an overarching bounded European identity (see Lacroix, 2009;Nicolaïdis and Pélabay, 2009) cannot be dismissed out of hand simply based on an assumption that the absence of such an identity makes sustained political commitments ipso facto impossible. Rather, the conceptual foundations of the 'identity deficit' hypothesis itself demand closer investigation.
This article scrutinizes two propositions on which the postulated link between communal identity and European polity-building is often seen to hinge: the proposition (1) that communal identifications inspire affectively charged, non-instrumental commitments to an overarching democratic polity and to its defining values and institutions and (2) that a communal identity engenders transnational trust in that these commitments are indeed shared. I argue that these claims are plausible to the extent that any democratic polity needs non-instrumental commitments as well as trust, and that communal identifications can encourage both. Even so, in several deductive steps I try to show that this does not preclude the possibility of such perceptions and motivations emerging through interrelated processes that do not presuppose a sense of overarching European communal belonging. These include the gradual externalization of domestic democratic norms and practices to the EU level, the incorporation of the resulting supranational democratic attachments back into existing national identifications and the accumulation of transnational political trust driven by the practice of supranational democracy itself. Though far from inevitable, such an outcome is conceivable in that it is theoretically coherent and has limited empirical analogies and precedents. Among them is the EU itself as a gradual absorption of EU-related norms and attachments into national self-definitions and a strengthening of trust between its member populations have far outpaced the emergence of overarching European identifications. I conclude by discussing the conditions likely to benefit the continuation of such processes as well as the factors liable to impede them, and in that context I briefly return to the eurozone crisis. The EMU situation should not be read as a sign that the Union's continued democratization as much as many other forms of integration beyond a technocratically administered common market are intrinsically unviable in the absence of strong overarching communal identifications. It does, however, bear out the danger of policies whose scope, rigidity and near-irreversibility upset the delicate balance between supranational integration and national adaptation while consuming more trust and commitments than the fragile European polity can generate at this relatively early stage of its formation.

I. Proposition 1: 'Communal Identity Engenders Non-instrumental Commitments'
The initial claim underpinning this proposition is fairly uncontested. As a partially democratic supranational polity itself composed of democratically governed Member States the EU depends on a measure of popular support. Such support must express itself in national electorates voting for governments that favour continued EU membership. Moreover, the quality of a pan-European democratic process ultimately depends on citizens being willing and able to follow and engage in democratic debate and deliberation, to vote and, most crucially, to accept the possibility of being outvoted by a majority from other Member States without rejecting the legitimacy of the supranational democratic process itself (Weiler, 1996).
Often drawing on the founding work of Émile Durkheim (1933Durkheim ( [1892), a long tradition in analytical social theory holds that social and political systems cannot generate such commitments by relying on rational-strategic calculations among their members alone. For while instrumental calculations by definition imply an end they seek to attain, they 'cannot, at a societal level explain the antecedently given ends that [strategic] action presupposes as a motive in the first place' (Abizadeh, 2002, p. 495;March and Olsen, 1989). For example, while instrumental considerations might help determine the means people employ when trying to maximize the welfare of their families or nations, they fail to account for why individuals come to value their families or nations as ends worthy of having their welfare maximized. Some theorists refer to the latter type of commitments as 'social commitments', defined as 'non-instrumental and infused with emotion or affect' (Lawler et al., 2009, p. 3). Their non-instrumental character partially insulates social commitments from the often more volatile logic of instrumental cost-benefit calculations and confers upon their beneficiaries a perceived 'self-referential claim to survival' (Theiler, 2003, p. 254) -not simply as means to an end, but partially as ends in themselves.
In some interpretations this might still allow the EU to get by on instrumental loyalties alone provided its citizens deemed it conducive to their own welfare or that of social units to which they had developed non-instrumental attachments (for example, their respective Member States). Yet even some otherwise strongly 'utilitarian' approaches to popular attitudes in the EU 1 deem this unlikely, arguing that while instrumental loyalties are beneficial they probably are not sufficient to secure long-term popular backing for a supranational European polity. Underpinning this is in part a widely shared conception of democracy as entailing the deliberative pursuit of shared social ends as opposed to mere bargaining or negotiation among strategically interacting, self-interest-maximizing individuals (Honohan, 2001). That aside, integration becomes harder to legitimize on exclusively instrumental grounds in areas where citizens perceive it as delivering fewer direct tangible benefits (for example, foreign policy). Finally, instrumental support for one cause risks being overridden by more affectively charged support for another when the two causes are perceived to clash (Mercer, 2010;Lawler et al., 2009), leaving an EU that enjoyed purely instrumental backing vulnerable to challenges from inside the Member States provided these retained their affective, non-instrumental dimension.
None of this logically implies that the EU must become the strongest level of social commitment for its citizens overall. But for many theorists it suggests that the Union requires at minimum what Etzioni (2007, p. 32, emphasis added) calls a 'narrow trumping loyalty'. In this conception, the EU must 'trump' lower-level attachments in 'selected matters' (Etzioni, 2007, p. 32) -that is, presumably where it has some powers over its Member States and where loyalties are thus most likely to conflict. Following the same logic, the extent of 'trumping' required would increase as the EU obtained additional powers since this would generate more loyalty clashes the Union needed to 'win'. It would also grow as the EU became subjected to greater democratic control since this would entangle citizens in a more extensive supranational democratic process for which they needed to acquire the necessary affective dispositions while simultaneously increasing the ability of citizens who lacked such dispositions to damage the Union at the ballot box.
Broadly accepting such reasoning, many analysts from otherwise very different theoretical traditions go further. Insisting that the Union needs non-instrumental commitments, they also maintain that it could attract such commitments only if its citizens developed a shared overarching communal identity -not necessarily replacing national and sub-national identifications, but subsuming them into an outermost circle of communal belonging shared by all Europeans. More specifically, the argument is not that such a communal identity would be sufficient to sustain supranational democracy (even people who feel intensely European may reject the EU), but that it would be a necessary catalyst for the non-instrumental dispositions without which supranational democracy cannot flourish. Roland Axtmann, for example, encapsulates this as follows: [N]otwithstanding as to how the delineation of the 'demos' has been achieved, unless the citizens form a solidaristic body there cannot be a democracy. And this is for the simple reason that in a democracy the willingness to listen and to be swayed by an argument put forward by others is influenced by a perception of the interlopers' legitimate membership of the conversational community, by their shared status as fellow citizens. There has to be a sense of belonging together, a sense of loyalty, a sense of identity, even. (Axtmann, 2006, p. 95; see also Howe, 1995;Etzioni, 2007) Such an interpretation thus reflects more than just the well-established finding that shared communal identifications can benefit mutual social commitments -a claim that reverberates throughout social psychology (Hogg and Abrams, 1988), historical sociology (Calhoun, 1997;Posen, 1993) and many related fields. It rather implies that the link between communal identifications and democratic commitments is sufficiently compelling to make the former indispensable for the latter, with an outermost layer of communal identification constituting the 'logical space for democratically united citizens' in the EU (Habermas, 2001, p. 107, emphasis added). The proposition, in short, is that democratic decision-making demands non-instrumental commitments and that these require shared communal identifications however the overarching communal space might be demarcated and symbolized.

Assessing Proposition 1
The logic just outlined is vulnerable to the following objection: even conceding, first, that the EU requires non-instrumental commitments and, second, that communally anchored commitments are typically stronger than their non-communally rooted counterparts, the claim that supranational democracy necessarily presupposes the EU partially 'trumping' national attachments is not compelling. Nor, therefore, is the assertion that the Union requires a communal quality to make 'trumping' viable. To develop this argument I turn to the literature on consociational democracy in some socially and culturally deeply fragmented multi-level domestic systems where identity 'trumping' is deemed unattainable or undesirable and where loyalties to different political levels develop through rather than in competition to one another. These examples are potentially relevant for the EU, regardless of whether one accepts the applicability of the wider consociational blueprint to European integration.
Sketched very broadly, the consociational logic of identity accommodation implies that the need for identity 'trumping' diminishes as attachments to different-level polities develop through each other and thereby become mutually implicit. 2 This presupposes two concurrent processes. First, different constituent polities must be in a position to define their membership in the overarching political system in mutually divergent ways. Members of Switzerland's different constituent groups, for example, have long defined their 'Swissness' differently from one another, shaped by their particular local cultural and historical circumstances. The political centre encourages this heterogeneity of overarching identity definitions by allowing the various cultural groups, cantons and localities to use Swiss-wide historical myths and symbols selectively and to interpret them idiosyncratically. 3 Strong sub-unit autonomy in socially and culturally sensitive areas such as cultural policy and education help sustain this variability. Second and conversely, in such systems the sub-polities must interpret and practise their respective norms and cultural particularities in such a way as to make them 'fit' into the overarching polity. The need for such adaptation is least in areas where sub-unit autonomy is strongest. It becomes more important where the overarching political centre has greater powers, where interaction between the centre and the sub-units is thus more intense and where norm incompatibilities would therefore run a greater risk of producing destabilizing friction and loyalty clashes.
Where it materializes this two-way rapprochement leads to a kind of mutual incorporation: of the sub-polities into the overarching polity and, vice versa, of overarching polity-defining norms and identifications into the sub-polities. As the two become normatively and symbolically implicit in each other, they become partially symbiotic: identifications with the sub-polities strengthen commitments to the overarching polity. Staying with the Swiss example, such shared loyalties to the federal state differ from more conventional nationalisms in Europe and elsewhere in 'content', intensity and symbolic manifestation. But they nonetheless sustain a remarkably extensive system of Swiss-wide democratic norms and practices as it has evolved over the past 150 years.
In this interpretation, deeply fragmented consociational systems such as Switzerland thus illustrate more than just what many analysts of European integration habitually single them out for -namely that political unity and cultural diversity can coexist and that citizens can hold multiple political loyalties. More interestingly, they also suggest that this does not presuppose hierarchically layered multiple identities with the outermost layer 'trumping' the others, and it is in this regard that they are potentially most instructive for the EU.
Central to the analogy is the observation that the EU, too, has experienced a gradual mutual accommodation and incorporation of different-level representations and attachments in a way broadly reminiscent of the domestic examples just discussed. On the one hand, prevailing conceptions of what the Union 'is' have traditionally differed from one Member State to the next -a political and geographical extension of Revolutionary values in France and of Christian (and Social) Democracy in Germany, a vehicle for bypassing post-colonial dependencies in Ireland, an institutional embodiment of western liberal norms in many central European countries, an association of free-trading liberal democracies in the United Kingdom and so forth (Marcussen et al., 1999). Many of these interpretive frames have deep roots in the Member States' respective national histories and shape not only popular preferences towards particular EU policies and institutional setups, but also the very meanings, emotions and cognitions the Union invokes in different national contexts (Díez Medrano, 2003). On the other hand, despite strong differences in national perceptions of the EU, a growing body of research charts a parallel process of normative socialization that has gradually redefined many patterns of national identification and has made them more amenable to EU membership. Examples include the 're-imagination' of Ireland as a country at the economic and political core of EU-Europe in Irish mainstream political discourse since the 1970s (Hayward, 2009) and the ascendancy of EU-compatible 'liberal nationalisms' in large parts of post-1989 central Europe (Auer, 2004), among many others. In combination, these developments suggest a two-way process: mutually divergent national interpretations of the EU reproduce themselves but against the backdrop of a gradual 'Europeanization' of national identity definitions. While perceptions of Europe remain heavily 'nationalized' national identifications have become partially 'Europeanized'.
These mutual adaptation and incorporation processes have been uneven and their continuation is not inevitable. Even so, they suggest that in principle the Union is amenable to a pattern of identity accommodation reminiscent of those internally fragmented multi-level democracies where the outermost level does not 'trump' the others. Such a 'marble cake' constellation (Risse, 2004, p. 251) differs fundamentally from conventional visions of the EU as a political formation whose social legitimacy invariably depends on citizens developing concentric circles of communal identification: rather than presupposing an outermost level of communal belonging invariant in content, meaning and affective significance for all EU citizens, it has national and European norms and representations flowing into, shaping and adapting to one another as part of a continuous dialectical interchange. Commitments to the overarching European polity develop through rather than parallel to national identity definitions -not 'European and Irish' so much as 'European as Irish' in subtle but important difference to conventional multi-level identity and 'selective trumping' accounts.
Such a conceptualization resonates with the work of several theorists who have contemplated potential futures for the EU as something other than a multi-level polity comprising hierarchically layered levels of communal attachment. This includes older attempts to directly apply consociational blueprints gleaned from domestic politics to the EU (Taylor, 1991). It also encompasses more recent calls for the EU to resist the lure of a 'European identity with a common substantive basis' in favour of a shared discursive space in which mutually recognizing and deliberating national demoi retain their 'narrative diversity' while becoming part of a 'transnational overlap of national overlapping consensuses' (Nicolaïdis and Pélabay, 2009, pp. 182, 185;Nicolaïdis, 2004;Besson 2006). These proposals differ from one another in many respects but they share the same underlying assumption this section has tried to corroborate: the assumption that democratic polity-building in the EU does not inexorably depend on its citizens developing a shared European communal identity furnished with invariant content and 'trumping' power, and that the EU could evolve so that 'different citizens may effectively identify with their common political institutions for different reasons' (Lacroix, 2009, p. 151;Castiglione, 2009). Provided we accept this in principle, it helps us move beyond the question of whether the EU could generate communal identifications strong enough to partially 'trump' their national counterparts to contemplating what kinds of institutional arrangements could best prevent the EU and the Member States from making competing loyalty claims. I return to this below. First, I turn to a second rationale for the belief that the EU requires communal identifications, which centres on the need for trust.

II. Proposition 2: 'Communal Identity Engenders Trust'
If domestic and supranational democracy alike relies on citizens' non-instrumental commitments, many analysts insist that democracy also requires trust. More than that, based broadly on the notion that non-instrumental dispositions can benefit a democratic polity only if its members trust one another that they are in fact disposed in this way, prominent strands in political sociology and related disciplines have long depicted trust as an elementary building block and glue of democratic life on which social commitments and deliberative capacity alongside a raft of other social virtues in turn depend (for example, Putnam, 2000). It is at this point that in the study of European integration the communal identity assumption frequently enters the argument once more, in the form of a postulated link between communal identifications and trust. Many analysts see this link as close to self-evident and use the two concepts almost interchangeably, interpreting the strength of one to also indicate the presence of the other (for example, Niedermayer, 1995;Eriksen and Fossum, 2004).
As with the previous proposition, the trust hypothesis' two component assumptionsthat democracy needs trust and that communal identifications can strengthen trust -enjoy solid theoretical and empirical backing. In addition to the intuitively plausible claim that trust is more conducive to democracy than its absence or even its opposite (that is, distrust), the first assumption accords with the observation that democratic systems whose members display high levels of mutual trust tend to be more stable and cohesive than their lower-scoring counterparts on dimensions ranging from voter participation and civic engagement to the wider social legitimacy of democratic institutions. This relationship seems broadly solid despite some lingering problems of definition and measurement and the inevitable problem of separating between cause and effect (Jamal and Nooruddin, 2010;Rothstein and Teorell, 2008). For its part, the second component assumption is reflected in a large social-psychological literature which suggests that communal identifications can strengthen not only mutual social commitments among the identifiers, but also notions of mutual compatibility, reliability and predictability which in turn help foster trust (Mercer, 2010;Hogg and Abrams, 1988, chapters 3 and 4). Historically oriented accounts of political integration in state-and nation-building contexts often posit the same basic relationship. They show how in such settings the growth of communal sentiments facilitated trust-dependent co-operation and collective action among people most of whom remained mutual strangers (Anderson, 1991;Calhoun, 1997). Even if one accepts both claims, however, this again does not settle whether an overarching communal identity is an indispensable (as opposed to merely a beneficial) condition for democracysustaining trust in the EU -a question to which I now turn.

Assessing Proposition 2
Trust among interpersonal strangers is -by necessity -category-based and thus stereotypical. We associate people with a social category we perceive most relevant to their evaluation in a particular context and seek to infer from this whether they and us share compatible social and 'symbolic universes' and similar norms and motivations (Berger andLuckmann, 1991 [1966], pp. 104ff.; Hogg and Abrams, 1988, chapters 3 and 4). While, as suggested, a shared communal identity can amplify a belief in shared norms, people nonetheless can and often do assume that others with whom they share no communal affiliation resemble themselves in certain respects and share some of their norms. Such perceptions can form by way of trust generalization -that is, the projection of trusting personal encounters with individual group members to the group at large (Putnam, 2000;Delhey and Newton, 2003). They may also result from exposure to indirect, socially mediated representations of the out-group -for example, through school curricula and the mass media.
A good example of inter-societal trust deepening over time is western Europe itself, as attitude surveys have consistently documented a dramatic increase in trust among west European democracies after 1945. Strikingly, for example, for French respondents Germany has long become the second most trusted foreign country on earth (Inglehart, 1991, p. 164) with similarly impressive trust levels characterizing relations even between many of those west European societies whose pre-1945 relationship was frequently hostile or even violent and that retain some negative mutual stereotypes (Gerritsen and Lubbers, 2010;Niedermayer, 1995;more broadly, Delhey, 2007). Existing surveys do not indicate over which areas precisely such trust extends. However, it is bound to encompass at least a belief in a commitment to mutual non-violence on all sides -the most elementary form of trust in any inter-group relationship and a necessary (but not sufficient) condition for its extension into other areas (Deutsch et al., 1957). The long-lasting, total and entirely uncontested absence of mutual violence scenarios in political debates in all west European democracies highlights the intersubjective entrenchment of such trust: not only are most west Europeans unable to imagine another war with, for example, Germany (and vice versa), but they cannot imagine Germans imagining such a war (and again vice versa). All sides have internalized norms of mutual non-violence as well as one another's internalization of these norms.
Yet while west Europeans unconditionally trust in one another's peaceful mutual dispositions they have not experienced a corresponding rise in overarching communal identifications (that is, as European or EU-European). For example, 51 per cent profess no attachment to the EU at all even when given the option of declaring multiple identifications and of rating those with the EU weakest. A similar proportion cannot identify with the European flag, one of the main overarching symbols the EU has promoted for decades. By contrast, 91 per cent of EU citizens feel attached to their country and 87 per cent to their city, town or village (Eurobarometer, 2008, p. 67;Eurobarometer, 2007, p. 80).
These figures mask persistent variations between different Member States and socioeconomic classes, with EU identifications typically stronger among those with higher incomes and education levels (Fligstein, 2008, chapter 5). They may also somewhat overestimate actual communal attachments in some respects while underestimating them in others. 4 Even so, read in conjunction with one another they leave little doubt that communal identifications in the EU lag behind the dramatic and well-documented rise in transnational trust over the past half-century. This makes it necessary to explain such trust without conceptualizing an overarching communal identity as the principal causal factor in its emergence.
The international relations literature offers a good starting point for such explanations as many scholars attribute western Europe's trust-laden post-war pacification to two overlapping processes. In the following paragraphs I briefly outline these, arguing that they suggest parallels with how trust in shared political commitments, too, could strengthen.
A first explanation takes a broadly constructivist perspective by charting a dialectical process that involves transnational perceptions, expectations and behaviour. Initial co-operative moves under United States tutelage after 1945 accumulated into a growing legacy of co-operation and peaceful coexistence, thereby steadily improving transnational perceptions and stimulating trust. Still more intense co-operation thus ensued, while embryonic overarching institutions helped codify co-operative norms and socialize national political elites into these norms (Checkel, 2002;Wendt, 1999). The outcome, in this rendering, has been a virtuous cycle as '[c]ooperative behaviour leads to a feeling of trust, and the feeling of trust is evidence that one should cooperate' (Mercer, 2010, p. 5).
A second, complementary explanation focuses on the entrenchment of liberal democracy throughout most of western Europe after 1945 and with it of norms pertaining to the non-violent management of political conflict. It credits this with a process of norm externalization as political elites in liberal democracies project their own internalized rejection of political violence to similarly constituted polities, thereby collectively redefining their relationship as a mirror of their domestic environment (Cederman, 2001). As Risse-Kappen (1995) argues, such a process ultimately hinges on recognition and perceptions, too: a belief in their own peaceful dispositions leads democratic societies to attribute the same proclivities to other democratic societies and to act accordingly, thereby creating increasingly self-fulfilling expectations.
These two trust-building trajectories can become mutually reinforcing. Each has dynamic and expansive qualities, which can turn transnational trust into an emergent social property. In line with this, the 'older' west European democracies are on average more trusting of and more trusted by other Europeans than their counterparts in central and eastern Europe in respect to which such trust-generating processes have had less time to gain momentum (Gerritsen and Lubbers, 2010).
Crucially for the present argument, neither process has an overarching communal identity among the trust-forming parties as an indispensable logical prerequisite, which squares with the empirical evidence from western Europe discussed earlier in this section. Both instead depict the norms, motivations and perceptions that sustain peaceful co-operation as emerging through an inter-societal process of norm-and perceptionshaping interaction and thus endogenous to the social practice of peaceful co-operation itself (Wendt, 1999).
What does all this mean for the prospects of political trust strengthening in Europe? The analogy is imperfect of course -not least in that supranational democracy goes beyond mere inter-state pacification and indeed presupposes it. Nonetheless, assuming that both pertain to norm-governed, intersubjectively rooted practices and that like norms linked to the non-violent management of political conflict, norms of democratic representation, deliberation and decision-making are already embedded in the Member States' domestic political cultures, potential parallels between the two come into view. In principle, Europeans could develop trust in a shared commitment to a supranational democratic polity through processes broadly analogous to those that once produced trust in a shared commitment to mutual non-violence. Central to this would once more be a reciprocal relationship involving behaviour and trust and the externalization of domestic norms to the relationship level, except that this time the norms concerned would encompass a commitment to supranational democracy rather than just to the peaceful resolution of political conflict. Social learning would not pertain to democratic norms per se (as these are already embedded domestically) so much as to their projection and application to the supranational sphere in interaction with other democratically constituted societies.
In addition to the institutional conditions discussed in the next section such a trustbuilding dynamic would benefit from two broad factors. First, just as awareness of one another's domestic pacification helped post-war west Europeans develop trust in a mutual commitment to peaceful co-operation, a greater awareness inside the Member States of how other member populations have struggled for and practise democracy in its many national idiosyncrasies and permutations could help Europeans build up trust in one another's democratic dispositions and expectations. The relative mutual insulation of national mass media and other forums for political socialization and deliberation still sometimes makes such recognition difficult (Theiler, 2006). Most importantly, Europeans cannot develop trust in one another's supranational (as opposed to just domestic) democratic commitments outside the actual practice of supranational democracy itself. So it was with western Europe's trust-laden post-war pacification which hinged not only on professed commitments to peaceful co-operation on all sides, but on progressively intensifying co-operative behaviour. Europeans could thus 'experience' these commitments and gradually internalize them and the behaviour-trust-behaviour dialectic described could gain momentum.
The trust on which supranational democracy depends could thus strengthen gradually through mutual observation, interpretation and experience among European publics, in step with the expanding practice of supranational democracy itself. It would therefore also be intertwined with a strengthening of shared democratic commitments as described in the previous section, and it would benefit from the fact that democratic norms are already firmly established inside the Member States rather than having to evolve ex nihilo. Among many other social practices in many other contexts, western Europe's trust-laden post-war pacification illustrates that trust and the social practices that depend on it can strengthen one another even if the trust-forming parties share no or only weak overarching communal affiliations. Trust requires, at minimum, trusting and trust-confirming behaviour, and in turn makes such behaviour possible.

III. Empirical Implications
The argument I have tried to develop over the previous sections has two main applications. First, it points to the institutional conditions under which supranational democracy could best deepen and consolidate even assuming the continued weakness of overarching communal identifications in the EU. Second, it helps us to better appreciate the factors contributing to some past and present difficulties in the integration process. In this section I briefly discuss the first and then turn to the eurozone crisis as an example of the second.
One important inference from the argument so far is that the factors liable to strengthen overarching trust and commitments require effective counterbalances lest they acquire the opposite effect. For Europeans to 'experience' supranational democracy as they 'experienced' peaceful transnational co-operation after 1945 and through this to develop stronger overarching democratic commitments and trust, the supranational democratic process must be visible, engaging and consequential in their everyday lives. Yet as was equally argued, if citizens are to develop commitments to the EU through their existing national identifications, overarching institutions and processes must limit their visibility and intrusiveness especially in socially and culturally sensitive areas and sometimes retain a measure of 'constructive ambiguity' to facilitate their incorporation into the different national identity constellations. Similarly, while provisions for the open pursuit of political conflict are a necessary and defining feature of every democratic system, in highly fragmented polities conflict more easily risks degenerating into damaging degrees of intense polarization. Such systems must thus structurally facilitate the clash of competing political interests while at the same time mitigating the danger posed by its most acrimonious and extreme potential manifestations (Dahrendorf, 1990).
In illustrating how specific kinds of institutional arrangements can help maintain such balances and equilibria the Swiss example is once more, up to a point, instructive: direct democracy at the federal level through popular initiatives and referendums renders the overarching polity salient and 'experienceable', partially offsetting the general inertia, indirect nature, limited visibility and conciliatory quality that define the Swiss-wide political process in many other respects. The majoritarian and polarizing slant of direct democracy is then for its part counterbalanced by the highly restricted scope of federal competences and by long consultative cycles and tolerance for regional discrepancies and ambiguities when implementing the results of popular votations (Linder, 2010). In the EU, the introduction of Europe-wide popular referendums in narrowly delineated areas of supranational competence accompanied by a stronger ring-fencing of national prerogatives especially in identity-sensitive areas such as cultural policy and education could help achieve a similar balance. In combination, such policies would make supranational democracy more visible and salient while helping to safeguard the integrity and reproductive capacity of the different national polities as the primary units of communal identification and building blocks of Europe's overarching democratic system (Theiler, 2006).
To the same end, the Union could go further still. As argued, perceptions of and preferences towards the EU have traditionally differed from one national polity to the next, leading several Member States to participate in some aspects of the integration process while abstaining from others (for example, Schengen and the single currency). If such 'variable geometry' at the institutional level were to deepen alongside moves towards the Union's greater democratization, the logical corollary would be a corresponding variability at the level of democratic participation. Supranational European electorates of various sizes and compositions would deliberate and vote on different issues, with a given national polity's participation contingent on its Member State having transferred relevant competences to the EU. Rather than a single invariant European demos, supranational democracy in such a scenario would engage overlapping configurations of national polities that committed, trusted and democratically interacted in different areas to different degrees. 5 Such a constellation would mark a profound break with prevailing models of democracy evolved inside the modern state (including its consociational variants) and it would, of course, face many difficulties. Even so, the argument I tried to make suggests not only that a variable European polity lacking communal closure is conceivable, but also that it would be easier to reconcile with the present and likely future variability in overarching political commitments and communal identifications in the EU. Bereft of a communal anchoring and to some extent politically decentred, such a polity could nonetheless progressively deepen and consolidate. In the ways shown, democratic interaction can strengthen the very trust and commitments on which democracy depends.
At the same time, the reciprocal processes charted throughout this article can veer in either direction, not only building up trusting political relationships but also eroding them. To illustrate this logical flipside of my argument I briefly return to the eurozone crisis.
Two aspects of the EMU situation are central to such an account. First, since its inception the currency project has made every participating member's economic welfare more directly dependent on the actions of every other member -above all on a shared willingness and ability to abide by EMU's various fiscal stability commitments. Interdependence spells mutual vulnerability which can generate fear and rejection. Since the beginning of the crisis this has manifested itself in growing opposition to the currency among many member populations (see below), in increased hostility towards those Member States frequently blamed for destabilizing the euro and, vice versa, in resentment against economically dominant countries (especially Germany) accused of imposing austerity and structural adjustment policies on weaker eurozone members even if their populations perceive such policies as riding roughshod over their needs and democratically derived preferences. The recent resurgence of derogatory national stereotypes even among some mainstream political elites seems to embody such grievances -from Chancellor Merkel's reliance on dubious factual evidence to berate southern Europeans' supposedly deficient work ethic (see Böll and Böcking, 2011) to increasingly virulent anti-German rhetoric in some Member States, most visibly in Greece.
Second, in trying to ease such tensions policy-makers cannot resort to the same strategies that helped defuse many previous crises in the integration process, from the 'empty chair' stalemate in the 1960s to the EU's foreign policy failure during the American-led invasion of Iraq. Corresponding broadly to the logic outlined earlier, this included pausing, the informal adjustment of a disputed policy's original remit and a (temporary) increase in 'constructive ambiguity' to better accommodate different national preferences while formally retaining a previously agreed policy and leaving the acquis intact (Ross, 1995;Marquand, 2011). Leaving aside the momentous option of breaking up the eurozone, EMU's economic logic, by contrast, is such that attempts to stabilize it cannot but stimulate measures of precisely the opposite kind to those just invoked, continuously raising rather than (temporarily) lowering the stakes. Actual or planned stabilization measures range from ever more extensive and (at least in creditor-country public perceptions) more redistributive bail-out mechanisms to unprecedented intrusion by the EU (possibly reinforced by some notionally semi-distinct structure to be established under the proposed intergovernmental Fiscal Stability Treaty) in domestic policy processes even in areas as sensitive as national wage-setting (Erne, 2012). They also include radically beefed-up penalties to boost Member State compliance. The overall result is a further increase in EU interference in domestic democratic processes, a decline in tolerance for divergent national implementation speeds and a reduction in ambiguity.
In some respects this increase in EU powers recalls the impetus from crisis to 'spillover' postulated by neofunctionalist theories of integration, but it carries significant dangers. Attempts to stabilize the EMU project by making its members still more mutually dependent and the Union still more intrusive and redistributive risk continuously upping the ante with public support for such measures demanding the very commitments to and trust in the currency's long-term viability the crisis has already helped undermine. A majority or near-majority of citizens in several key Member States (representing various financial predicaments) now believes that the euro damages their economies while simultaneously opposing a greater role for the EU in enforcing national fiscal discipline so as to stabilize the currency project (German Marshall Fund, 2011, pp. 19-20). Alienation is especially strong in Germany -the inevitable financial pivot of all present and future stabilization schemes. By late 2011 only 17 per cent of Germans still fully trusted the single currency (Institut für Demoskopie Allensbach, 2011, table A4), while 75 per cent opposed German participation in the European Financial Stability Facility despite warnings by their government and mainstream opposition parties that the fund is essential to avert the currency's demise (Peel, 2011).
How all this will affect long-term support for the EU itself is unclear. The Union can draw some comfort from the earlier observation that commitments and trust are multidimensional and, up to a point, divisible: distrust in another Member State's economic reliability does not necessarily erode trust in, say, its commitment to liberal democratic values. Moreover, opposition to a particular EU policy does not necessarily signify a wholesale rejection of the EU. Whatever its long-term consequences, interpreting the eurozone crisis in light of the earlier argument brings out what I think are important tensions at the heart of the single currency project: EMU's near-irreversibility and economic imperatives of discipline, intrusiveness, rigidity and uniform enforcement clash with the political need for balance, gradualism, (partial) reversibility and tolerance for national idiosyncrasies and varying implementation speeds central to the growth of a nationally 'incorporated' European polity through the various interrelated processes described. Brought in relatively early in the European polity-building process, the eurozone project upset the delicate balance between supranational integration and national adaptation and consumed more transnational political commitments and trust than Europeans are yet able to generate.
Those commentators cited at the outset who contend that if Europeans had developed stronger overarching identifications the eurozone crisis would be easier to alleviate might therefore well be right. As suggested, shared feelings of communal belonging can be a potent catalyst for trust and commitments sufficiently strong and uniform to make certain political projects possible. Yet even if the EMU experiment should ultimately stumble over the weakness of such identifications, it would be wrong to infer from this that a shared communal identity is an inescapable prerequisite for all forms of further integration -the continued development of EU-wide democratic norms and institutions included. For while EMU is in many respects antithetical to the needs of a gradually evolving, variable and nationally 'incorporated' European polity at least during these relatively early stages of its formation, a potential expansion of supranational democracy through the various interrelated processes outlined throughout this article is compatible with these needs. Much will therefore depend on whether the EU heeds the lessons of the single currency experience and proceeds -for the time being at least -with policies that allow for pauses, partial reversals and the often messy accommodation of mutually divergent democratically derived national preferences and political cultures. The more it succeeds in this, the greater are its chances of nurturing a polity of mutually trusting and democratically interacting citizens regardless of whether these identify themselves as belonging to an overarching community of Europeans.

Conclusions
A democratic European polity needs non-instrumental commitments and it needs trust among citizens that these commitments are indeed shared. A stronger overarching European communal identity, if it evolved, could probably reinforce both. Yet this does not exclude the possibility of such commitments and trust emerging independently of communal identifications -a possibility that is theoretically coherent and has limited empirical analogies and precedents. Such an outcome is not inevitable and the eurozone crisis is an important reminder that the reciprocal processes I have charted can also operate in reverse. Even so, deducing that under certain conditions supranational democracy is possible even without an overarching communal identity leads to a cautiously optimistic conclusion for the EU. It implies that the options for the Union's democratic development are broader and the chances of its success greater than many theorists suggest. Even if their European 'identity deficit' refuses to narrow, Europeans could choose to build a democratic polity around it.