CONFRONTING EUROPEAN DIVERSITY: DELIBERATION IN A TRANSNATIONAL AND PLURI-LINGUAL SETTING Abstract In this article, we confront some commonly held assumptions and objections with regard to the feasibility of deliberation in a transnational and pluri-lingual setting. To illustrate our argument, we rely on an analysis of group discussions from EuroPolis, a transnational deliberative experiment that took place one week ahead of the 2009 European Parliamentary elections. The European deliberative poll is an ideal case for testing the viability of deliberative democracy across political cultures because it introduces variation in terms of constituency and group plurality under the controlled conditions of quasi-experimental scientific setting. For measuring group dynamics and interactions we apply a modified version of the Discourse Quality Index (DQI) that is combined with a qualitative content analysis of selected sequences of discussions. Findings show that participants of transnational deliberative polling 1) generally recognise the EU polity as a reference point for exercising communicative power and impact on decision-making, and 2) are in fact able to interact and debate across languages and cultures, developing a self-awareness of citizens of a shared polity and thereby turning a heterogeneous group of randomly selected citizens into a constituency of democracy. IRENA FIKET ESPEN D. H. OLSEN HANS-JÖRG TRENZ Irena Fiket is Post-Doc Fellow at the Center for the Study of Political Change, University of Siena; e-mail: fiket@unisi.it. Espen D. H. Olsen is Senior Researcher at Arena -Centre for European Studies, University of Oslo; e-mail: e.d.h.olsen@arena.uio.no. Hans-Jörg Trenz is Deputy Director of the Centre for Modern European Studies (CEMES), University of Copenhagen and Research Professor at ARENA - Centre for European Studies, University of Oslo; e-mail: trenz@hum.ku.dk. IV IV LO d cp rN o o rN rN 15 > A 3 a 0 z * I (I) 0 c « IV Ln Introduction The debate on the democratic deficit of the EU has increasingly involved calls for a more profound engagement of "ordinary" citizens in European politics. The low level of political participation in EU politics has become all the more acute in recent years as the so-called "permissive consensus" has been cast aside by growing discontent among citizens with the integration project (Hooghe and Marks 2009); most clearly visible in a string of popular rejections of the Constitutional and Lisbon Treaties. Moreover, the new diversity of the enlarged Europe is potentially a further obstacle to mutual understanding among the peoples of Europe and the activation of European citizenship. European Parliamentary elections thus far rather amplified the problem addressing the citizens mainly as uninformed voters, displaying decreasing turnouts and increasing support for populist parties and Eurosceptic opposition. After the "period of reflection" that followed the 2005 rejection of the EU Constitutional Treaty, EU institutions took a deliberative turn manifested in the implementation of numerous deliberative experiences based on the involvement of lay citizens in public debate across Europe. Even though the institutional and procedural design of the expanding participatory regime of EU governance has been widely analysed, only few empirical studies have thus far analysed the conditions and capacities for citizen deliberation within the EU (Abels 2009; Huller 2010; Friedrich 2011). EU analysts have however repeatedly emphasised that the generation of democratic legitimacy in a setting of enhanced socio-economic, political and cultural diversity is constrained by the lack of a common public sphere that guarantees a certain degree of uniformity of public opinion and will formation (Grimm 1995; Schlesinger and Fossum 2007). In this article we analyse whether enhanced diversity of a group of deliberating citizens and the fuzziness of the polity to which they respond impacts the constitution of a mini-public of democratic self-government. We argue that the potential of citizens' deliberation to generate democratic legitimacy is dependent on participants' disposition to recognise, first, the polity as a legitimate entity to exercise political authority, and second, to identify as a constituency, i.e. as members of a political community that is (self)empowered to authorise and control government. The EU poses a challenge to both the polity and the constituency dimension. In formal terms, the EU exercises political authority, but is it recognised by the citizens as a legitimate entity for delegating collectively binding decision-making (the polity dimension)? Secondly, the EU has established a complex citizenship regime, but do citizens identify as a constituency of rights holders and democratic agents of public authorisation and control (the constituency dimension)? 3 a these two constitutive dimensions of democratic legitimacy which are commonly treated as independent variables in deliberative experiments within national political settings. Our argument is that both dimensions, the recognition of political authority and the identification of the citizenry, are not simply to be considered as constitutive elements of democracy. Both dimensions are rather to be seen as possible outcomes of the process of building democratic legitimacy through reflexive y^ reason-giving (see Eriksen 2005). In the following we therefore explore the possibility of a positive correlation between participation in deliberative polling and the formation of a democratic CT> constituency. The guiding hypotheses are that, as an effect of deliberation with ^ other European citizens, participants of deliberative polling start to a) recognise the EU as a legitimate authority for collective problem solving (polity hypothesis), and b) identify as a constituency of democratic politics (constituency hypothesis). Deliberation can thus trigger reflexive processes conducive to the establishment of both legitimate government and the democratic agents of authorisation and control. In the context of citizen deliberation, this polity and constituency generating power of deliberation (see Cooke 2000) is overlapping and mutually dependent. We conceptualise the reflexivity of the deliberative setting in procedural terms through the generation of knowledge and shared normative perspectives among the participants that help to qualify (or validate) the substantial policy issues at stake and to establish mutual understanding and agreement (Eriksen 2005, 17). In the Deweyan sense, then, we observe whether and how individuals identify as democratic agents (as a public) through critical practices, which authorise political power and, at the same time and through the same practices, constitute the community of citizens that is in charge of the control of this power (Dewey 1927). Confronting European Diversity: The Polity and the Constituency Contested Deliberative polling has thus far been applied mainly within national and monolingual settings: Respondents were chosen from one legally demarcated, and socially and culturally recognised constituency. In addition they respond mainly to one (either local, regional or national) level of government. Deliberative theorists, including the designers of deliberative polling, have however emphasised that diversity of opinions needs to be considered as a necessary procedural condition for deliberation to facilitate opinion change and learning (Fishkin and Luskin 2005; Sunstein 2009; Thompson 2008). This is based on the argument that deliberating citizens from diverse socio-cultural backgrounds can learn to acknowledge and respect the plurality of values and views that exist within a polity and contribute to the construction of the public good (Benhabib 1994; Gutmann and Thompson 1996). We argue that experimenting with citizen deliberation in the transnational setting of Europolis creates two additional challenges. Firstly, the applicability of the experiment is put into question by the fact that the group of randomly selected participants is situated within a "non-finished" polity. Political authority of the EU is neither legally consolidated nor socially accepted. The EU polity as the reference point for the sample is not the familiar environment of national or regional government but a complex multi-level governance arrangement. This introduces uncertainty with regard to the question of which type of administration, legislative procedures and formal government deliberation should exert influence on. Is communicative power expressed through transnational deliberative bodies renationalised in the sense that participants target mainly domestic institutions and decision-making processes or do such "mini-publics" also pay tribute to the complexity of multi-level governance in the sense of empowering European institutions and supranational authority? Secondly, the constituency from which the representative sample is chosen for deliberative polling lacks concrete political recognition in the EU setting. Participants of Europolis were not members of a pre-established demos or a fully recognised political community. Participants rather represented lay citizens from 27 member states and spoke 23 different languages. The constituency of democratic politics in the EU is arguably neither fully legally recognised nor does it recognise itself as a politically bounded and culturally distinct community. EU constituents are unbounded, multi-dimensional and contested (Abromeit and Schmidt 1998; Fos-sum and Trenz 2006). Statistical indicators for drawing a representative sample of European citizens can therefore not rely solely on the background assumption of a relatively homogeneous and monolingual population but must take into account the existence of pluri-ethnic and pluri-lingual fragmented groups as well as shifting minorities and majorities. In confronting these two challenges, the feasibility of deliberative democracy in the EU has been discussed widely. From the one side meaningful and equal deliberation in a complex and culturally pluralistic EU is seen as an impossible project. The democratic deficit of the EU is seen here as structurally rooted in the absence of a European demos. The people(s) of Europe do not simply have a common identity as citizens of the same polity. They also lack the socio-cultural prerequisites to become united, e.g. through a common language, a shared cultural background or through participation in an encompassing public and media sphere (see e.g. Grimm 1995; Offe and Preuss 2007; Kraus 2008). In this line of reasoning, deliberation works best, if political culture is contextualised, pluralism of opinions is contained, participants speak the same language, share a common ethical understanding and pay each other respect as co-citizens (Habermas 1998, Wessler 2008). The upshot of this is, then, that democratisation in terms of engaging citizens and fostering a vibrant public sphere on the transnational level is an impossible task. If cultures demarcate different discursive universes, discourses between cultures must be seen as principally problematic (Leigh 2004). From the other side, different solutions have been discussed for reconciling political equality with deep diversity in multicultural societies (Fossum 2003, Fray ser and Honneth 2004). There is a tendency in homogeneous deliberative settings that familiarity and closeness leads to unjustified extremism (see Sunstein 2009, 3). Like-minded people tend to amplify their pre-existing views and reduce their internal diversity (ibid., 8). High degrees of solidarity and pre-existing affective ties even increase these effects of group polarisation (ibid.: 42-44). Deliberative polling contributes to avoiding such polarising traits of groups by creating a setting where people do not start out with strong convictions. Participants are confronted instead with plural views and new information that breaks familiar settings. In this line of reasoning, we would expect that group heterogeneity is a favourable condition in transnational settings. Group polarisation effects should be unlikely and deliberation should be on average more balanced and less "extreme" than in national settings. The upshot is: deliberation works better if it includes diverse people: "Cognitive diversity is crucial to the success of deliberative democracy" (ibid., 142-43). Another manifestation of deep diversity in the EU setting is its pluri-lingual c > traits (Kraus 2008). The question of whether and how groups can interact and seek understanding across languages is therefore highly relevant. Findings from social ^ movement research show that pluri-linguism at the European level does not nec- essarily impair the inclusivity and epistemic quality of deliberative settings such as the European Social Forum as compared to the exchange among movement activists at the national level (Doerr 2008, 2009). Translations can potentially help out in exceptional circumstances, for instance, in the case of professional groups or among experts. But can the epistemic condition of democracy be met in a pluri-lin-gual random sample of citizens? The critical issue in confronting European diversity is how transformative the deliberative poll can potentially be in a transnational setting. The transformative power of deliberation can alter individual preferences towards the identification of a common good (Mansbridge 2010). But deliberation can be also transformative in the sense of social learning and group identification. In Europolis the participants were at the same time empowered as potential voters in European Parliament elections. Against their diverse socio-cultural backgrounds they were confronted with problems of common relevance and through their communicative exchanges around these issues, they learned to articulate their shared concerns, (e.g. in the question rounds with experts and political representatives), fleshed them out with reasons and justifications and critically reflected on their experiences. Deliberation is in this sense embedded in social learning processes (Trenz and Eder 2004). It spurs not only reflection over the policy issues at stake but also over the process itself as a collective experience. This community generating potential of deliberation is at the core of our empirical analysis. The critical yardstick for concluding on the community generating effects of deliberation in Europolis is, then, the extent to which a randomly selected group of individuals from diverse national and linguistic backgrounds turned into a critical public, which recognised EU authority and developed a notion of identity. Method and Data Taking place one week ahead of the 2009 European Parliamentary election, Eu-roPolis was set up to conduct a transnational deliberative experiment that engaged 348 citizens from all EU Member States. The Europolis research design followed standard Deliberative Poll design.1 The event specifically addressed climate change and immigration control, two high-profile issues of recent political debates in Europe. The participants were assigned into 25 small groups consisting of two or three languages. Discussions were led by moderators who had the task to raise certain pre-determined issues for debate as well as to manage the workings of the group, but still with a minimalistic approach to moderation. In addition, there was a host of translators involved with each group due to their pluri-lingual character, thus allowing verbal exchange in the participants' mother tongue. Europolis produced two sets of data. The first were collected using questionnaires and those allow us to measure both, pre- and post-deliberation opinions and knowledge level and the perceptions of the participants at the end of experiment. The second were audio recordings of the debates in all small group discussions of the event. The small group debates have been coded by using a modified version of the Discourse Quality Index (DQI). The DQI is a measurement instrument2 that relies on qualitative coding of debates based on a carefully constructed coding scheme3 and represents a quantitative measure of deliberation based on Habermas' concept of discourse ethics. The main goal of the DQI is to tap a continuum that ranges from the complete violation of Habermas' discourse ethics to the ideal speech act. The initial DQI was constructed to analyse elite deliberation in parliamentary assemblies (Steiner et al. 2004). Europolis was, however, about citizen deliberation. The DQI was subsequently modified to include a new set of categories.4 We coded and analysed small group discussions on immigration control in 4 groups of the following language composition: the first was composed of Italian and English speakers; the second of English speakers and Bulgarians; the third of English, French and Portuguese; and a fourth group of Italians, Spaniards, and Swedes. Beyond Attitudinal Data: Opening the Black Box of Deliberation Deliberative polling has, thus far, relied principally on attitudinal data to measure the transformative effects of deliberation. In our study we take attitudinal changes of participants of Europolis, in both the "polity" and the "constituency" dimension, as a starting point of our discussion. With regard to the first dimension, the questionnaire measures participants' attitudes on decision-making levels before they started the discussions but after they read the briefing materials and after deliberation with regard to the two debated issues (immigration and climate change) and two "control" issues (unemployment and crime).5 The results show that participants of the groups that were analysed became more favourable of shifting decision-making powers to the supranational level on the "control" issues6 (unemployment and crime) but less favourable on immigration and climate change. With regard to the second dimension, the Europolis questionnaire results indicate that deliberation in a transnational setting shows a clear potential to spur identity change among the participants.7 More concretely, the share of participants that perceived themselves as national citizens only decreased significantly after participation in the deliberative poll. Participants turned from identifying in exclusive nationalist terms to becoming "inclusive nationalists," i.e. they also identified as members of a community of Europeans. In order to understand better why deliberation led participants to become more nationalist in their polity preferences while at the same time socialising in a transnational group setting it is necessary to analyse micro processes of debate and group formation in deliberative polling. Relying principally on aggregated questionnaire data, deliberative polls have so far not systematically covered "real-life" experiences of deliberating citizens. Our in-depth analysis of deliberative group discussions is an attempt to enter the black box of deliberation. For that purpose, we rely on behavioural data from audio recordings of the debates in 4 out of 25 small groups that discussed the issue of immigration. Qualitative content analysis of the transcribed discussions was applied to select speech acts, in which the polity and constituency dimensions were raised by the participants. Relevant text was tagged during the coding process to expose arguments and story lines used by participants. We do not claim that the quotes are "representative" of overall discussions, but rather use the quotes to firstly illustrate and elaborate on findings from the quantitative data regarding the polity dimension, and secondly to highlight specific sequences of critical reflexivity among citizens regarding the constituency dimension. This approach made it possible to ^ tease out substantive issues raised by participants in actual deliberative moments. c > The selection of quotes was as such done in a sequential manner. Based on initial descriptive analysis we then went back to the substantive debates to recover quotes ^O that underlined the results. Our qualitative analysis further needs to be considered as exploratory in the sense that we cannot rely on any comparative baseline as well as no comparable data from other deliberative polls or mini-publics. Indeed, the experimental character of deliberative polling raises the question of its relevance to political analysis. We argue that the counterfactual nature of deliberative polls in itself can be used as a starting point for answering the questions of polity recognition and constituency formation in an EU setting. In this sense, Europolis as an orchestrated multicultural and pluri-lingual event is meant to put key notions of deliberative theory to test. Against this background of a counterfactual setting (the deliberative poll as the opinions of citizens if given opportunity to deliberate) and the uniformity of deliberative theory (a modicum of linguistic and cultural understanding needed for deliberation) we can explore the potential of citizen deliberation in a transnational setting. To investigate the transformative potential of deliberation in relation to this set of data, we can operationalise our guiding polity and constituency hypotheses as follows: Polity Dimension. As an indicator of recognition of the EU polity, we expect that participants justify their arguments less in terms of particular group interests or references to their country of origin but by referring to the benefits of EU/Europe, or to common good principles. We further expect that European or common good-oriented justifications increase as an effect of group discussion while particularistic (nationalist) justifications decrease over time. The DQI has a variable called "content of justification." This variable allows for the measurement of justification of arguments. Originally, this was set up to capture whether arguments were made in terms of narrow group interests, in terms of the common good, or in terms of both (Steiner et al. 2004, 58). To capture the specifica of citizens' deliberation8 in Europolis on European issues, we added the category Europe justified speech acts. Through content analysis of the transcribed group discussions we expect to find specific instances of "polity contestation" that involve participants spontaneously in debates about the delegation of political authority and the preferred institutional/ constitutional design of the EU. Constituency Dimension. As an indicator of the formation of a political community, we expect that participants participate equally in group discussions and that no linguistic group dominates over others.9 Furthermore, we look at the role of the facilitator, in order to determine to what degree equal participation has been encouraged and if that was the case, which participants needed encouragement.10 We further measure degrees of interactivity between participants across languages and whether these interactions include positive, neutral or negative reference to other participants' arguments.11 Through content analysis of the transcribed group discussions we expect to find specific instances of "constituency contestation" that involve participants spontaneously in debates about the confinement of the political community and degrees of common identification in Europe. Analysis: Deliberative Citizens in Action The Polity Dimension £ 3 a 0 £ 1 * tf) 0 c (B KD Coding the group discussions by the use of DQI helps us to understand how participants in a transnational deliberative setting exchange arguments in political debates on the EU and how they justify the appropriateness of collective choices and levels of decision-making. The underlying assumption is that citizens by expressing preferences with regard to specific policy solutions (that were measured trough use of questionnaires) also raise validity claims relating to the common good of the issue under debate: Within what particular institutional arrangements are arguments held to be valid and who should be the main beneficiary of a given policy solution and to whom shall collective decisions apply? In a given setting, should the collective choice respond to the needs of the participants' own country, the EU, or global community? To analyse whether European or global-oriented justifications increase as an effect of group discussions we adopted a sequencing approach. Different sequences were separated following the agenda of the group discussion. The first phase of discussion evolved mainly around the identification of the problems at stake. The second phase was more strongly influenced by the moderators, who coordinated the more formalised task to formulate questions for the plenary session. These questions were based on a selection of previously justified contributions to the debate, which made a further engagement of the participants in justificatory discourse during this phase redundant. The third phase was again more open and allowed participants to synthesise debates on the issue, express their opinions on the plenaries with experts and reflect their experiences. Deliberation in this last phase was found to be frequently less issue focused and more geared towards reflection of the common understanding of the citizens and their attitudes towards the EU. Leaving aside the more formalised setting of the second phase, we would expect to observe that European common good orientation among the participants increase from the first to the third phase. Table 1 largely confirms this hypothesis. As an effect of knowledge increase, learning and socialisation during the experiment participants became more "European" in their justifications and recognised the EU polity as a legitimate entity of problem-solving. At the end of the experiment one third of the justifications delivered contained a European common good reference, while references to national interest clearly diminished. Table 1: Justification All groups (7, 8, 11, 12) Discussion Formulation of questions Discussion after plenary N speech acts (excluded moderator) 202 74 73 N of speech acts justified among which: 109 40 25 group interests, own country 25.9% 25.5% 12.5% Europe 31.5% 19.3% 45.8% global or common good references 42.5% 55.2% 41.7% Overall, Table 1 indicates that European and common good orientations with regard to the issue of immigration control prevail over national interests. Citizens LO demonstrate a clear tendency to look beyond the national context to validate their ^ claims and to raise competing polity preferences. Citizens do in fact also engage with the issue of the EU polity and express strong opinions on European integration in relation to alternative local, national or global polity settings. Selected speech acts identified through content analysis confirm that citizens exhibited preferences for a more powerful EU that develops capacities of immigration control. The lack of coordinated action in this area is seen as a problem and responsibilities should be shared instead of blaming single member states: But what I understood today at the plenary meeting was that everybody blames individual member states. Italy is sending back immigrants. Spain does not want to do anything. Greece the same. Still, Italy, Spain and Greece are receiving these people, trying to select them in the best possible way. I would ask the 'lords' of the European Union what they are doing for these countries (Italian male, SG [small group] 12). In addressing and recognising the EU as a polity, where borders and the insider/ outsider logic matters, these debates also suggest that "polity recognition" correlates with an evolving sense of community. As we will see from the following example, polity and constituency dimensions frequently overlap: The only thing I want to say is that we can't really allow to sit on the fence. If we truly want to build a true Europe, we should talk about external borders only, and the EU member states should relinquish some of their sovereignty. I think that there's no other way to go about it (French male, SG 11). So, I think we should also strengthen the borders of Europe, because if we make all these people legal, we will have a massive arrival of migrants, and we do not have the capacity to welcome all these people. And it will only be to the detriment of the migrants themselves (Luxembourg/Portuguese male, SG 11). The Constituency Dimension Any viable polity depends on a modicum of identification from its citizens. As frequently highlighted, the EU lacks the typical identity signifiers that are held to be constitutive of nation-states (Giesen 2003; Delanty 2005; Castiglione 2009). Since a strong political identity that would replace the existing identities of the nation state seems unattainable and for many also undesirable, the question is whether the European setting is based on a zero-sum relationship between existing national identities or conducive to a positive sum relationship of nested identities (Gora et al. 2011). From the first perspective, participants in transnational mini-publics would be expected to defend primarily national views and interests. The group discussions would lead to a nationalist clash among the participants who would become more introverted in defending the integrity of the national community and mapping their attitudes onto a cultural cleavage towards their follow participants from other member states. From the second perspective, participation in group discussions would stimulate citizens to engage with others' views and interests. This would lead, in turn, not only to attitude change but also trigger off micro processes of identity change and socialisation of participants as citizens of Europe. Europolis provided an ample "laboratory" for gauging the degree to which group heterogeneity and language differences affected the deliberative mini-public as a democratic constituency. Did participants in the multinational and pluri-lingual setting of Europolis have equal opportunity to participate in the debate and to contribute to deliberative exchange and opinion formation? In the following analysis, we analyse possible effects of ethno-cultural heterogeneity and language pluralism with special emphasis on socialisation, group reflexivity and identity formation. Group Solidarity. Qualitative data from transcripts give numerous examples of the development of what we call reflexivity of participants which turned group deliberations into critical voice of the citizens. Critical reflexivity was partly encouraged by the specific task the group had to perform in formulating expert questions and addressing policy makers. The confrontation with experts and other groups in the plenaries created shared expectations that were exchanged among the participants especially in the last round of the debate. The development of critical and reflexive attitudes as part of group deliberation can be considered as an important identity marker. We can distinguish between different layers of deliberations which can encompass a critical reflection on the role of participants as citizens, on the purpose of the scientific experiment and their role therein, and finally, a meta-discourse on Europe and its complex identity questions. For obvious reasons, critical reflexivity as part of the group discussions is unequally developed; in some instances, it is given only sporadic expression and restricted to single statements, in other instances, it unfolds in longer sequences through dialogue among the participants. First, group solidarity is enhanced by the processes of becoming reflexive as citizens of Europe and expressing critique towards the experts and politicians. In the following statement, an Italian participant confronts the unitary visions of the citizens (the participants of the panel) with the still divided positions of the political representatives (the experts of the plenary). We (the citizens) can make proposals and provide solutions for problems, which we feel are ours. We can, in principle, convert from nationals to Europeans. But they (the politicians) are not able to give y substance to a European identity. They do not know how to use the opportunities (like a citizen forum) for us but only for them. They do not take up our ideas but only follow their opportunistic interests: [we should] ... make a question to our political representatives of Europe: Wheth-^ er (and when) Europe will give substance to a European identity. (...) We could give the proposals and solutions to our problems (but) we must feel them as our problems, we should feel as Europeans... the fact that we still (mainly) have a national identity is limiting strongly our ability and participation. In fact, the participation in European elections' was shameful. The sense of belonging (to Europe) is lacking (Italian male, SG 7). More specifically, the experts and politicians are criticised for their unwillingness to provide concrete answers or their incapacity to make themselves understandable to the citizens. This lack of responsiveness is then generalised as a European ^ experience that marks the citizen-elite divide of the EU and justifies the democratic response of indignant citizens against the elites in Brussels: ^^ What I experienced as a person, I felt that these young women (experts and pol- (^q iticians who participate at the plenary sessions) ... even if they wanted to give 0 £ tf) G KD us the answers, those were not the answers in my opinion. ... And I am sorry for this but if I could decide who should occupy those places at the European l^v Union I would suggest placing us there. Why so many people are moved from their homes (to work for the EU) if they can't give us concrete answers? (Italian female, SG 12). This indignation about the incomprehensible experts and elites is also shared by other participants. In the following statement, the upcoming elections are seen as an opportunity to mark a difference. Again, a "we"-feeling is created by distinguishing participants of the experiment as the forerunners of a European citizenry who should guarantee that only the "really qualified" are elected. ... now we have the European elections, and we should all do the 'advertising' in order to select the qualified people. So they will not come there only to be 'chair warmer' (Italian male, SG 12). Secondly, group solidarity is enhanced by the processes of becoming "reflexive" as being part of a European experiment. Reflections on the purpose of the experiment are a recurrent topic of group discussions. Participants see themselves confronted with the expectation that they should develop a common understanding and we-feeling as Europeans. In general, this possibility is not rejected but taken up as an opportunity for further reflection: ... the purpose of this research is to understand how the discussions may change our views ... and I think that the possibility to communicate with each other could help us to understand each other better and could lead us to feel more European ... because we get to know other people and we discuss with them ... so the time we spent here is good for us (UK female, SG 07). Another Italian participant reflects about the privileged experience to participate in the scientific experience, which for him is also a "human experience." He is however fully aware of the isolated character of the experience and deplores the lost opportunity for the EU to not making a more systematic use of the ideas and proposals that are produced by the citizens: I wonder why this opportunity is not used by the EU. It could have been an opportunity for the EU and for the people to bring up new ideas. In fact, it could have given the space for (our) new ideas that could have become active and not only passive proposals (like now). Instead, it's only good for us as an experience, but in the end it only remains a 'discourse' that we carry with us, but this benefit does not sufficiently justify that the EU is losing this opportunity (Italian male, SG 07). Participants thus combine their critical reflection about the experiment with the expression of critique of the EU and the state of European democracy. As a case of second-order reflexivity, this transformation of becoming a European citizen can again become an element of reflexive group deliberation. It is then recognised that the experiment was not helpful in an instrumental sense to arrive at better policies and solutions but rather in a symbolic sense to make participants aware of the dimensions of European citizenship: ... this meeting, at least in my opinion, did not help us to solve or to clarify the problems of immigration. But, it increased the awareness of European citizen- ship. Not because they made me feel more European, but because they made me be more careful towards the people we are selecting to represent us in Europe (Italian female, SG 12). While acknowledging the potential for transnational identity formation, another participant also underlined its possible pitfalls and limits. A contrast to this is found in the interplay between different modes of identity that might change over time: For me, Europe and the world are a village (...) A Frenchman for example, coming from the south of France to the north of France, is like in a foreign country. And with the years, he will become used to the people of Northern France and the people of Northern France will become used to him. (...). And I think that those who welcome the migrants should create situations where people can better integrate themselves. I think if you do that, you're not going to lose your identity, you're not going to lose your origins (...) (Luxembourg/Portuguese male, SG 11). These findings on critical reflexivity of group discussions strongly back an understanding of reflexive public deliberation as an effective means to overcome cultural incommensurability (Bohman 2003). Socialisation factors of taking part in an assembly like a deliberative poll matter to explain the transformative force of deliberation in intercultural settings. The challenges of cultural pluralism are thus minimised by the effects of group reflexivity. Participants from diverse eth-no-political groups are committed to shared practices for providing evidence and discussing solutions to common problems. What is more, participants from diverse socio-cultural background are critically engaged in contesting political authority and defining their role as European citizens. These findings on group reflexivity and socialisation as a counter-effect to cultural fragmentation are also strongly backed by the post-deliberation questionnaire poll.12 The views and perceptions of participants on the behaviour of other participants, provides answers to the degree of cohesion and "group-ness" in the transnational mini-public of Europolis. Overall, the participants evaluated their experience of participating in the deliberative event as highly positive. Only 28 percent of participants felt that their group fellows mainly cared about their own country and not about European Union. 88 percent agreed that participation J was equal in small group discussions. The experience of meeting and talking with other people from all across the continent and with different cultural background also had an impact: 81percent of the participants thought that they had learnt a lot about people different from themselves, "about who they are and how they live." 84 percent felt that their fellow participants respected what they had to say, even if they did not necessarily agree. On average, the participants thought the event extremely balanced and considered the quality of the group discussions they took part in to be high. Most importantly, participants from other member states were not seen as hostile players who defended diverging interests but as equals who expressed strong views and provided accessible justifications. On this score, then, we can conclude that, overall, the results of analyses of EuroPolis groups show that contrary to the communitarian assumptions, ethno-cul- 0 £ i _________„.....,______________________,...___________________________,..................... tural plurality has no significant impact on deliberative quality and the possibility for citizens from different member states to debate and find agreement on issues ^ of common concern. en KD Discussion Formulation of questions Discussion-after plenary SG 7 (n participants 11) Language of moderator - Italian Total speech acts (excluded moderator) 46 21 21 Moderator intervene to engage individual participants 15.2 % 4.5 % 5.0 % Moderator intervene to engage linguistic group (engl) 0.0 % 0.0 % 10.0 % Neutral or positive reference to other arguments 50.0 % 90.9 % 85.0 % UK and IRL 56.9 % 50 % 68.1 % Italian n. of speech acts 43.1 % 50 % 31.9 % Total 100% 100% 100% SG 8 (n participants 10) Language of moderator -English Total speech acts (excluded moderator) 64 35 20 Moderator intervene to engage individual participants 11.00% 8.6% 30.00% Moderator intervene to engage linguistic group 0.0 % 0.0 % 0.0 % Neutral or positive reference to other arguments 39.1% 34.3% 50.0 % UK 67.2% 44.2% 50 % Bulgarian n. of speech acts 32.8 % 55.8% 50 % Total 100% 100% 100% SG 11 (n participants 13) Language of moderator -English Total speech acts (excluded moderator) 46 9 16 Moderator intervene to engage individual participants 20.7% 22.2 9.1 Moderator intervene to engage linguistic group 0.0% 0.0% 0.0% Neutral or positive reference to other arguments 50.0% 42.2% 66.7 FR LUX % of speech acts 30.8% 15.6% 25.8% UK IRL % of speech acts 69.2% 84.4% 74.2% Total 100% 100% 100% SG 12 (n. participants 15) Language of moderator -Italian Total speech acts (excluded moderator) 76 9 16 Moderator intervene to engage individual participants 2.6 % 0.0 % 0.0 % Moderator intervene to engage linguistic group (sw) 3.9 % 0.0 % 0.0 % Neutral or positive reference to other arguments 35.5 % 11.1 % 43.8 % Swedish n. of speech acts 23.3 % 23.5 % 37.5 % Italian n. of speech acts 65.1 % 70.6 % 62.5 % Spanish n. of speech acts 11.6 % 5.9% 0.0 % Total 100 % 100 % 100% *The percentages of each linguistic group speech acts are calculated by weight of each language within the composition of small group. Impact of Language Pluralism. For the purpose of this analysis, language is used as the second analytically distinct though not independent variable in constituting a critical public. In the post-deliberation questionnaire poll language was seen by only 12 percent of the participants as a barrier to follow the debate and "understand" their fellow European citizens. When analysing language group participation and interactions in different sequences of deliberation in the four groups analysed the results are more mixed (Table 2). Table 2: Equality, Language Groups and the Role of Moderators* In our analysis of group discussions we approached the criterion of equal participation by weighting each linguistic group's share in deliberation. Table 2 indicates that all linguistic groups participated in group discussions. Furthermore, moderators rarely intervened to engage specific linguistic groups in discussion but barely encouraged individual participants to get on board in the debates. There are, however, some patterns of language dominance in two groups (11 and 12) that correlate with the language spoken by the moderator while in the other two groups (8 and 7) moderating effects on language dominance did not become salient. Our data set is too small to further enquire this question of language dominance. Possible intervening variables that explain the variation on the share of group participation are the design of the group setting, delays in waiting for translations and individual styles of moderation. Another possible explanation for the minimisation of language as an impact on deliberative quality is that pluri-lingual settings are in fact especially conducive towards certain "habits of listening" (Doerr 2008; 2009). Transnational groups might turn out to be more attentive listeners and overcome habits of hearing in familiar national settings. In a discussion among co-nationals we know intuitively whom to listen to and whom to ignore. In a transnational setting, this familiarity is not given. In EuroPolis this was amplified by the technical equipment (simultaneous translations, headphones and microphones) which helped focus the attention of the participants. Participants were routinely asked to speak slowly and keep their speech intelligible in order to facilitate translation and thus mutual understanding. The higher listening requirements of the pluri-lingual setting might thus have worked positively for the deliberative quality. In all, then, our results with regard to the equality of participation and status of language groups confirm the overall trend of the Europolis experiment that plurality is not a principled barrier to deliberation. Participants did not isolate themselves but engaged in debate with citizens from other language groups. Conclusion In this article, we have explored whether deliberation is feasible if participants respond to a polity of dispersed authority and interact in a transnational and pluri-lingual setting. The analysis of group discussions from Europolis has J furnished two main findings. Firstly, the EU polity is recognised and taken as a reference point by citizens for exercising communicative power and impact on decision-making. In this sense, Europolis generated a counterfactual and microcosmic European "public," where citizens from highly diverse backgrounds and despite language pluralism have debated and contested each other on issues of principle and policy related to European integration. Problems of understanding related to the use of several languages in heterogeneous group settings can thus be partly ^ overcome, though there remain restrictions in how the principle of political equality can be approached and how the overall representativity of the experiment can be defended (Olsen and Trenz 2013). Secondly, the constituency created in Europolis was mainly one of critical reflexiv-ity toward experts and political elites. As such, our qualitative data highlights deliberation's community-generating and transformative role against the communitarian view that certain pre-political requisites must always be in place for deliberative democracy to function effectively. The participants did not all become wholehearted 0 £ i C (6 o Europeans or came to share a specific political identity. We show, however, that they in certain sequences of the deliberation developed a critical and collective problem-solving capacity on issues of shared relevance (Dewey 1927; Dryzek 2009). In other words, Europolis instilled in its participants a deliberative reflexivity which went beyond mere preference exchange: we have provided examples of how some of them developed what Dryzek (2009) has highlighted as a capacity to identify systemic shortcomings after confrontation with experts and politicians. Our analysis shows, then, that the EU polity received "recognition through criticism." By giving citizens the opportunity to discuss and voice opinion, deliberative polling raises awareness of the complexities of political decision-making and democratic legitimacy. We therefore conclude that communicative barriers as deriving from dispersed authority and group heterogeneity in the post-national constellation are for the most part practical and not substantial. They can be overcome by careful design of the deliberative setting which facilitates encounters among the participants and generates habits of respect, listening and learning. Notes: 1. For an overview of how deliberative polls are organised, see Fung 2003. See also Fishkin 2009. 2. The unit of analysis of the DQI is a speech act delivered by a participant. The entire discussion is broken down into smaller speech units and each speech act is coded separately. Every speech act is coded for all the variables included in the coding scheme. 3. See Steiner et al. (2004) for the coding scheme. 4. New categories of interest for this study will be listed and explained later in the text. 5. The question used to measure those attitudes read: And on a scale from 0 to 10, where "0" means "entirely at the EU level," "10" means "entirely by the individual Member States," and "5" is "exactly in the middle," at what level do you think decisions should be made in each of the following areas? Immigration; Climate Change; Fighting unemployment; Fighting crime. 6. "Control" issues are those that were not discussed during Europolis, namely: unemployment and crime. 7. The questions that allowed the measurement of identity change read as follows: On a scale from 0 to 10, where "0" is "not at all," "10" is "completely," and "5" is "exactly in the middle," how much would you say you think of yourself as being European?; And on the same 0 to 10 scale, how much would you say you think of yourself as just being from your [country]? Only country=10; And if you had to choose just one of the following alternatives, what would you say you see yourself as...? 1-nationality only/ 4-European. 8. The DQI category "content of justification" allows to assess whether justification of the statements or speech acts have been backed referring to benefits and costs of all. The DQI distinguishes three types of justification: Explicit statement concerning constituency or group interests (own country); Explicit statement in terms of a conception of Europe in utilitarian or collective terms; Explicit statement in terms of the common good or difference principle (solidarity, quality of life, justice, etc.). 9. We counted the number of speech acts delivered by participant/linguistic group. 10. The DQI category that captures the nature of moderator intervention is also a new category for the purposes of Europolis. In our study we specifically probe whether moderators of selected groups intervene to engage specific language groups in the debate. 11. The DQI categories that allows the measurement of degrees of interactivity is coded "Respect toward other arguments" and distinguishes between: No reference to other participants' arguments; Negative reference to other participants' arguments; Neutral reference to other participants' arguments: and Positive reference to other participants' arguments. 12. The questionnaire data are available on http://cdd.stanford.edu/polls/eu/ under the heading "Results." References: Abels, Gabriele. 2009. Citizens' Deliberation and the EU Democratic Deficit: Is There a Model for Participatory Democracy? Tübinger Arbeitspapiere zur Intergrationsforschung 1. Abromeit, Heidrun and Thomas Schmidt. 1998. Grenzprobleme der Demokratie: konzeptionelle Überlegungen. In B. Kohler-Koch (ed.), Regieren in entgrentzen Räumen, 293-310. Opladen: Westdeutscher Verlag. Benhabib, Seyla. 1994. Deliberative Rationality and Models of Democratic Legitimacy. Constellations 1, 1, 26-52. Bohman, James. 2003. Reflexive Public Deliberation. Democracy and the Limits of Pluralism. Philosophy & Social Criticism 29, 1, 85-105. Castiglione, Dario. 2009. Political Identity in a Community of Strangers. In J. T. Checkel and P. J. Katzenstein (eds.), European Identity, 29-51. Cambridge: Cambridge University Press. Cooke, Maeve. 2000. Five Arguments for Deliberative Democracy. Political Studies 48, 5, 947-969. Delanty, Gerard. 2005. The Quest for European Identity. In E. O. Eriksen (ed.), Making the European Polity: Reflexive Integration in the EU, 127-142. London: Routledge. Dewey, John. 1927. The Public and its Problems. Chicago: Gateway Books. Doerr, Nicole. 2008. Deliberative Discussion, Language, and Efficiency in the World Social Forum Process. Mobilization 13, 4, 395-410. Doerr, Nicole. 2009. Language and Democracy 'In Movement': Multilingualism and the Case of the European Social Forum Process. Social Movement Studies 8, 2, 149-165. Dryzek, John S. 2009. Democratization as Deliberative Capacity Building. Comparative Political Studies 42, 11, 1379-1402. Eriksen, Erik O. 2005. Reflexive Integration in Europe. In E. O. Eriksen (ed.), Making the European Polity - Reflexive Integration in the EU, 9-29. London: Routledge. Fishkin, James S. 2009. When the People Speak. Oxford: Oxford University Press. Fishkin, James S. and Robert C. Luskin. 2005. Experimenting with a Democratic Ideal: Deliberative Polling and Public Opinion. Acta Politica 40, 3, 284-298. Fossum, John E. 2003. The European Union In Search of an Identity. European Journal of Political Theory 2, 3, 319-340. Fossum, John E. and Hans-Jörg Trenz. 2006. The EU's Fledgling Society: From Deafening Silence to Critical Voice in European Constitution Making. Journal of Civil Society 2, 1, 57-77. Fraser, Nancy and Axel Honneth. 2004. Redistribution or Recognition. A Political-Philosophical Exchange. London: Verso. Friedrich, Dawid. 2011. European Governance and the Deliberative Challenge. Paper presented at ECPR General Conferences. Reykjavik, August 2011. Fung, Archon. 2003. Survey Article: Recipes for Public Spheres: Eight Institutional Design Choices J and Their Consequences. Journal of Political Philosophy 11, 3, 338-367. Giesen, Bernhard. 2003. Collective Identity of Europe: Constitutional Practice or Community of Memory. In A. Triandafyllidou and W. Spohn (eds.), Europeanisation. National Identities and Q Migration. Changes in Boundary Constructions between Western and Eastern Europe, 21-35. London: Routledge. Gora, Magdalena, Zdzislaw Mach, and Hans-Jörg Trenz. 2011. Reconstituting Democracy in Europe and Constituting the European Demos? In E. O. Eriksen and J. E. Fossum (eds.), Reconstituting Democracy in Europe. Theory and Practice, 159-178. London: Routledge. Grimm, Dieter. 1995. Does Europe Need a Constitution? European Law Journal 1, 282-302. GGutmann, Amy and Dennis F. Thompson. 1996. Democracy and Disagreement. Cambridge. MA: Harvard University Press. Habermas, Jürgen. 1998. The Inclusion of the Other: Studies in Political Theory. Cambridge. MA: MIT Press. Hooghe. Liesbeth and Gary Marks. 2009. A Postfunctionalist Theory of European Integration: From Permissive Consensus to Constraining Dissensus. British Journal of Political Science 39, 1, 1-23. ^ Hüller, Torsten. 2010. Playground or Democratisation? New Participatory Procedures at the European Commission. Swiss Political Science Review 16, 1, 77-107. £ * i * tf) 0 C (6 Kraus, Peter. 2008. A Union of Diversity, Language, Identity and Polity Building in the EU. Cambridge: Cambridge University Press. p-j Leigh, Jennifer. 2004. Reflections of Babylon: Intercultural Communication and Globalization in the New World Order. Globalization 4, 1. Mansbridge, Jane. 2010. Deliberative Polling as the Gold Standard. The Good Society 19, 1, 55-62. Offe, Claus and Ulrich K. Preuss. 2007. The Problem of Legitimacy in the European Polity: Is Democratization the Answer? In C. Crouch and W. Streeck (eds.), The Diversity of Democracy: Corporatism, Social Order and Political Conflict, 175-201. Cheltenham: Edward Elgar. Olsen, Espen D. H. and Hans-Jörg Trenz. 2014. From Citizens' Deliberation to Popular Will Formation? Generating Democratic Legitimacy Through Transnational Deliberative Polling. Political Studies 62, S1, 117-133. Schlesinger, Philip and John E. Fossum. 2007. The European Union and the Public Sphere. A Communicative Space in the Making? London: Routledge. Steiner, Jörg, André Bächtiger, Markus Spörndli, and Marco R. Steenbergen. 2004. Deliberative Politics in Action. Analysing Parlimentary Discourse. Cambridge: Cambridge University Press. Sunstein, Cass R. 2009. Going Extremes. How Like Minds Unite and Divide. Oxford: Oxford University Press. Trenz, Hans-Jörg and Klaus Eder. 2004. The Democratizing Dynamics of a European Public Sphere. Towards a Theory of Democratic Functionalism. European Journal of Social Theory 7, 1, 5-25. Thompson, Dennis F. 2008. Deliberative Democratic Theory and Empirical Political Science. Annual Review of Political Science 11, 497-520. Wessler, Hartmut. 2008. Investigating Deliberativeness Comparatively. Political Communication 25, 1, 1-22.