TRANS-EUROPEANISING PUBLIC SPACES IN EUROPE HAKAN G. SICAKKAN Abstract Are there any trans-border interactions and networking patterns, any common systems of competing political discourses, and/or any common channels, platforms, or arenas of communication or action that can be regarded as the beginnings of a European public sphere? If so, how is this embryonic European public sphere being structured? Based on a comparative analysis of discursive configurations and networking patterns of more than 240 civil society organisations in sixteen European countries and eight European civil society networks, this article finds discursive gaps between the views of member state-level and European-level civil society organisations on diversity, the future of the EU polity, and who they see as their legitimate addressees. Networking patterns indicate this gap is not only in discourses but also in interactions. Considering the current segmentation along national lines, this may imply the beginnings of a development toward the emergence of a horizontally and vertically segmented European public sphere. Hakan G. Sicakkan is Associate Professor at the Department of Comparative Politics, University of Bergen; e-mail: Hakan.Sicakkan@isp.uib.no. rN ro O Q. a. rN i—I o rN o~> £ £ 3 a 0 £ 1 * (R 0 C ! m o Introduction The aim of this paper is to assess the extent to which the participants of public debates in Europe are interconnected through transnational networks, collaboration, and discourses. A focus on transnational interconnectedness is important for the European public sphere (EPS) research for several reasons. Normatively, from a democracy point of view, a transnational public sphere with a transnational public, which is conscious of its role of overseeing the actions of the supranational policymakers, is desirable in Europe due to the increasing powers of the European Union (EU). Theorists of democracy on the neo-functionalist and cosmopolitan flanks call for a transnational European public which can assume the task of holding the supranational power-holders accountable (cf. Eriksen 2005). On the other hand, the intergovernmentalist and communitarian wings do not entirely recognise the need for a transnational public sphere in Europe as their proponents view supranational policymaking to be primarily a result of collective decision-making by democratically elected, legitimate representatives of the citizens of the EU-member state. Theoretically, identity (Risse 2010), universal values like democracy and human rights, economic interdependency and common market, common interests in international politics, and common law and political institutions, among other things, have been highlighted as factors that can energise the growth of a transnational public and a European public sphere in Europe. In this debate, the intergovern-mentalist and neo-functionalist camps have focused on, respectively, what divides and what brings Europeans together. Empirically, in the current decade, research has gone beyond the question of whether a European public sphere exists. Empirical focus has been on Europe-anisation of national media due to the assumption that, with its public outreach, accessibility, and openness, the media sphere is the best empirical equivalent of the concept of public sphere (Habermas 1974). Media research that offers a structural y approach has used (1) media's attention to "European themes" (e.g. Gerhards 2000; Trenz 2003), (2) the degree of reporting the same events at the same time (e.g. Eder and Kantner 2000), (3) whether news are reported with a "European framing" or "similar framing" (Peters et al. 2005), (4) visibility and resonance1 beyond national borders (Eder and Kantner 2000; Eder and Trenz 2003; Koopmans 2004; Olesen 2005), (5) legitimacy of foreign speakers in national public spheres (Risse and Van de Steeg 2003). This line of research has documented that media's attention to Europe-related themes is gradually increasing. Media research that deploys "common/similar discourses" or "common/similar meaning frames" as ^ an indicator of the European public sphere reports at best contradictory findings because the degree of transnational similarity in discourses and meaning frames varies with respect to the "policy fields one studies" (Koopmans and Erbe 2004, 114) or a "halting" process of Europeanisation (Peters et al. 2005). Every step forward in the conceptualisation of the EPS increased our knowledge of the commonalities and differences among the national media in Europe. How- G C > ever, considering the media is not a channel that only mirrors reality, but also forms it in different ways, there is no guarantee that the commonalities found in media q research is the artwork of a European public. Except few outstanding examples (e.g. Koopmans 2004, 2007; Splichal 2011), the media research on the EPS has not given us a solid idea about whether a European public exists and how it is structured LO and interconnected. This is because we have hitherto tried to understand the public 2 sphere by looking at it "directly" through its appearance mirrored by the media and, at the same time ignored its preordained component: the public. Conceptually, the EPS and the European public are imagined in various ways by scholars. On the one hand, those who sought in Europe the classical Habermasian model of a public sphere as a single space shared by a unified, critical European public, were quick to recognise that they were looking in vain. On the other hand, those who view the EPS and the European public as overlapping public spheres and multiple publics (cf. Schlesinger 1999) are still working to map out the areas of overlap. A European public is difficult to imagine in isolation from national publics just as national public spheres cannot be imagined without the subaltern and sub-national public spaces (Sicakkan 2006) that constitute them. Nor can the EPS be imagined in isolation from the territorially and level-wise polycentric and hierarchical European power structures - recently, some researchers have discovered that the EPS might be following the multi-level governance (MLG) system of the EU (cf. Koopmans and Erbe 2004). Indeed, the EPS should be conceptualised as a sphere that consists of several different types of public spaces at different levels, where the transnational European (trans-European) public sphere is only one of the constituent public spaces that co-exist. The same holds true also for the conceptualisation of a European public: a trans-European public is only one of the multiple types of publics that constitute the European public. Note that I am not deploying "European," "transnational," and "trans-European" interchangeably. In order of their appearance in European history, the major types of public spaces that currently co-exit are (1) essentialising ethnic, religious, or national spaces, (2) nationalising public spaces of the modern nation states, (3) trans-Europeanising public spaces, and (4) globalising/transnationalising public spaces, which correspond, respectively to, ethnic and religious publics, national publics, trans-European publics, and transnational/global publics. Through European integration, each of these public space types has found its expression and representation at different levels. The EPS has come into being with the emergence of a trans-Europeanising public space and a trans-European public that stretches over different levels of the EU political and social systems and co-exists and interacts with the other current public space types. The important empirical question at this juncture is "how do the different types of public space types and public form the EPS in interaction with each other?" In the following, I primarily focus on the impacts of the collective actors operating in trans-European and national arenas with an empirical focus on the patterns of their discourses and networks. Trans-Europeanising Public Spaces in Europe The reason for labelling the new public space a "trans-Europeanising public space" is two-fold: First, by using this term, I emphasise that trans-Europeanisation is an ongoing process. Second, the term can also be understood as a function of certain common arenas, networks, and interaction patterns although the objectives associated with them may not be Europeanisation. An example is the nationalist, intergovernmentalist, and anti-EU organisations' cooperation throughout Europe. Although these organisations are against any change that would reduce the sovereignty of the member states, the organisations' trans-border interactions contribute to forming a trans-Europeanising political space. In operational terms, a trans-Europeanising political space is defined as a system of multiple competing discourses advocated and voiced by different types of collective actors at national and European levels and/or a set of trans-border networks/structured interactions between collective actors located in different countries. That is, when either the criterion of transnationally shared discourses, or the criterion of transnational interactions, or both, is satisfied, one can start talking about trans-Europeanising political spaces. Table 1: A Conceptual Framework for Trans-Europeanising Political Spaces Is the Discourse Europeanising? YES NO Does the Organisation Have Trans-European Ties/Networks? YES I Trans-European organisations (e.g., Social Platform) II Non-Europeanising organisations in trans-European arenas (e.g., UEN) NO III Europeanising organisations in non-trans-European arenas IV Non-trans-European organisations u S Table 1 gives a schematic overview of the categories that constitute trans-Europeanising political spaces. In this framework, a nationalising discourse, for instance, can be observed in trans-European and national arenas, and similarly a Europeanising discourse can be observed in national and trans-European arenas. An organisation may be disseminating Europeanising discourses and simultaneously getting involved in trans-European networks (model I). An organisation may also be engaging in trans-European networks while disseminating primarily nationalising discourses (model II). Further, an organisation may be disseminating Europeanising discourses in its own member-state context without participating in trans-European networks at all (model III). Finally, an organisation may be deploying nationalising Q discourses only in a member state without engaging in trans-European networks (model IV). The organisations (actors) that fall under models I, II, and III, their trans-European affiliations (networks), and their views (discourses) on selected policy issues altogether constitute the trans-Europeanising public spaces. Model IV in Table 1, however, refers to the public spaces that are not trans-European as these organisations operate with typically non-Europeanising discourses only in national or local arenas. The different elements of this conceptual framework are further elaborated in the following sections and used as a heuristic tool to depict ^ the current structuring of trans-European political spaces. Discourses. For this research, I measure and assess the discourses focusing on organisations' statements about (1) which groups to include in the organisations' vision of a diverse society and whether an ethno-nationally diverse society II is acceptable/desirable/inescapable in their mindset, (2) the role they envision for the EU central political institutions and member states in the EU, and (3) which institutions/organisations/networks the groups want to receive their political mes- 2 sages. These three themes lie at the core of the tension between the gatekeepers and trespassers of borders and boundaries of many kinds in Europe as well as different levels of government within the EU political system. I simply distinguish between Europeanising and non-Europeanising discourses. Europeanising discourses tend to contain inclusive attitudes favouring (1) diversity of all kinds and (2) central EU institutions' participation in policymaking at different levels along with the existing national and local political authorities, and (3) defining different European intergovernmental and supranational institutions as receivers of the political messages - along with the existing national authorities. Non-Europeanising discourses, on the other hand, are characterised by disfavouring and excluding attitudes toward (1) diversity caused by non-native groups of people and (2) intergovernmental and supranational authorities' involvement in policy matters, as well as (3) regarding non-national (intergovernmental and supranational) political institutions as irrelevant addressees for the political messages. Networks. Analytically, the network dimension of trans-Europeanising public spaces can be approached in two ways. The first approach focuses on "horizontal" (Koopmans and Erbe 2004) networks where social and political actors seek and get involved in transnational collaboration and communication without attempting to build a higher hierarchical level that structures their interactions. The second approach emphasises "vertical" (Koopmans and Erbe 2004) networks that seek to articulate more structured, and often institutionalised, channels of collaboration and communication, at the European level. The second approach can be further elaborated in terms of bottom-up and top-down networks. Bottom-up networks emerge through social and political actors' own initiatives to build trans-European networks seeking to structure and/or institutionalise their collaboration at the European level. Top-down networks emerge through elite-led European-level initiatives that attempt to bring different social and political actors together under their umbrella. Each process and mechanism for forming a trans-European network implies a specific preference for a particular model of a EPS. Different preferences concerning involvement in horizontal and vertical trans-European structures, on the one hand, and in bottom-up and top-down structures, on the other hand, imply different approaches to diversity, as well as different attributions of ontological priority to the individual, the collectivity (of different types), the sub-national, the national, and the European. In other words, I expect some actors to deliberately rule out participating in vertical structures because the actors do not want to contribute to a hierarchical EPS structure. Therefore, in trans-European constellations of nationallevel organisations, I expect to find not only pro-European orientations but also diverging ideas and strategies concerning how the EPS should be structured (or not be structured at all) - e.g., a strictly segmented EPS along the lines of a Europe of nations, or an EPS as an arena that facilitates only limited trans-national collaboration on certain issues that cannot be dealt with only at the national level, or an EPS of overlapping European publics that follows the multi-level governance structure of the EU, or an ideally integrated single EPS, etc. In this article, the network dimension of trans-European political spaces is measured through the following indicators: (1) the operative level of the networks (regional, national, trans-European interactions), (2) the scope of collaborative interaction (collaborative projects/actions, joint projects/actions, attempts to formulate common objectives, efforts to formulate common actions to address common concerns, synchronising existing projects/action plans, mutual information sharing), and (3) membership status in networks (active membership, passive membership, observer status). Research Design, Sample and Data The data about the collective actors is measured at two levels: institutional level data about organisations, gathered from organisations' printed and online official documents, and individual level data, obtained from in-depth interviews with persons that are in leading positions in the organisations (elite interviews). Organisations and the Institutional Data Sample. The research design focuses specifically on those organisations and elites that have high visibility in public debates - representing the most visible mainstream and alternative discourses and networks. In each of the 16 European countries, I focus on three political parties (the party leading the government, the main opposition party, and the most visible Maverick party in each context), three non-governmental organisations (NGOs) or social movement organisations (SMOs) - civil society organisations that are the most visible in their contexts and represent the mainstream or alternative discourses, three think tanks (a policy research organisation, an academic think tank, and an advocacy think tank in each context), three print media actors (two main-player newspapers and one smaller newspaper that exhibits anti-establishment views in each context), and two broadcast media actors (one public and one commercial TV-channel that are main players in each context). The research design includes collective actors operating at different levels of governance. Therefore, I also planned to include three European political party federations (The Party of European Socialists "PES," European People's Party "EPP," and Union for Europe of Nations "UEN"), three trans-European networks of NGOs/ £ SMOs (Social Platform of European NGOs "Social Platform," European Network against Racism "ENAR," and European Women's Lobby "EWL"), and two trans-European networks of think tanks (European Policy Institutes Network "EPIN" and Trans European Policy Studies Association "TEPSA"). However, there are no Europe-wide media actors that are followed by a large European population: Euronews, which comes closest to what may be called a trans-European media channel, is not amongst ^ the significant news sources utilised by European citizens although it broadcasts in several languages. Facing this fact, the research design had to omit the "trans-European media." Due to concern for representing the actors that are the most visible in the public debates, the final sample includes a larger number of organisations: 242 organisations at member-state level (56 political parties, 67 social movement organisations, 46 think tanks, 44 newspapers, and 29 TV-channels, which are spread throughout sixteen European countries) and 8 European umbrella organisations that are the trans-European counterparts of these. In terms of both discourse and networking, q these exhibit varying degrees of affiliation with or dissociation from trans-Euro- peanising political spaces. Some are contained in national arenas in terms of both discourse and networks; some operate with Europeanising discourses in trans- C^ European arenas.2 2 Elites and the Interview Data Sample. From each organisation, a number of persons in leading positions have been interviewed. Understanding the internal diversity within the organisations that are active in public debates is very important with respect to the theoretical points of departure of Eurosphere. One of the project's aims is to identify the organisations and the persons in organisations that are pushing for more trans-Europeanisation or nationalisation. Thus, in each organisation, either the leader, or the vice leader, or someone in the steering board known to be endorsing the leader's views, was selected. In addition, for each organisation, a person known to be the opinion leader but not holding an official leadership position was selected. In cases where the official leader and the opinion leader were identified as the same person, an interview with an additional opinion leader was not conducted. Further, at least one leading person who had official responsibility for, or was known to be interested in the policy areas that Eurosphere is researching on, was included in the sample. Further, for those organisations with internal groups like women's groups, minority groups, youth groups etc., we included those persons who led the group that was the most visible in public debates. Thus, the size of the qualitative sample in each country is determined by four factors: (1) the number of the organisation types (which is four - political party, NGO/SMO, think tank, print media), (2) the number of the organisations' positions in the public debates (which is three - mainstream, main opposition, Maverick/alternative/anti-anti-establishment), (3) the number of the elite types (which is four - formal leader, opinion leader, internal opposition leader, sub-group leader), and (4) the saturation point for representing internal diversity in each organisation. The research design stipulates that including 48 elites from each country (representing 4 organisation types, 4 elite types, and 3 positions: 4x4x3=48) will provide the optimum coverage of important collective actors that participate in public debates. This makes a total of 768 interviews required to conduct the project in 16 countries. However, 54 interviews were planned for each country in order to avoid ending up with too few interviews, making a total of 864 interviews with organisations at the member state level: seven persons from each political party, four from each NGO/SMO, three from each think tank, and three from each print media. The number of interviewees from political parties is larger because they accommodate almost all types of elites and internal groups. In addition, 24 interviews were planned with the leaders of eight trans-European networks. These are the central operative units of eight European networks, the majority of which are located in Brussels. By operative units, I refer to leaders, boards, and secretariats of European umbrella organisations that bring together national level organisations under their transnational networks. The final interview data set contains 764 interviews because, in some organisations, the saturation point was reached below the maximum number of planned interviews - indicating a low level of internal diversity in the respective organisations. That is, interviewing more persons would not result in new information about the respective organisation. The second factor is inaccessibility of print media elites in the UK. Methods of Analysis. I use organisations, networks of organisations, and people who are in leading positions in these organisations (elites) as units of observations in different analysis stages. Discourses about diversity, the European polity, and the European public sphere are mapped through interviews with elites. The information about networking and collaboration patterns comes from institutional-level data collected from the organisations' official printed documents and other online publications as well as secondary literature on these organisations. Each of the three dimensions - views on diversity, European polity, and the European public sphere - and the networking and collaboration patterns are mapped by using multiple variables. In order to create concise indicators, the number of the variables is reduced with principal components analysis (PCA). To create the new scores, I use regression factor scores since they consider the importance (loadings) of the variables constituting the respective dimensions. All the PCA-results tables in this paper report rotated component matrixes based on varimax rotation with Keiser normalisation. For the question of whether a system of competing Europe-wide discourses and trans-European interaction patterns exists, I adopt an exploratory approach. By using a series of discriminant analyses (DA), I identify the member state-level organisations that display discourse and networking patterns similar to those of trans-European networks, and vice versa. The grouping variable in each DA is simply a dummy variable indicating whether an organisation is a national-level organisation of a trans-European network. The final classifications of the cases are cross-validated. Elite Discourse Patterns in the European Public Sphere Interviews with leaders of national and trans-European-level organisations show there are clear differences in the organisations' approaches to diversity, EU polity, and the public sphere. Although the whole spectrum of views is represented at both levels, the set of views that dominate at each level differs. Differences between National and Trans-European Elites' Views on Diversity. The interviewees were asked to mention persons and groups that they see as relevant to their own idea of a diverse society. After the interviewees talked about their £ own preferences, they were asked to consider whether they would like to include other categories. The answers were then registered in a common database. Table 2 presents results from a PCA of the categories mentioned by the respondents. The first dimension indicates global and transnational understanding in the sample. All the variables loading on this dimension concern categories that are unrelated to the notion of a homogenous nation state - but other phenomena, other groups, and belongings that contest it. I labelled this dimension "Global and Transnational Orientation to Diversity." It measures the respondents' tendency to include all types of diversity, not only group-based diversity but also individual diversity. This includes diversity generated by internal mobility within the EU. The second dimension measures the extent to which a respondent is willing to include gender, disability and sexuality groups, different generations, and social classes in his or her definition of a diverse society. I labelled this dimension "Bodily and Individualist Orientation to Diversity." These variables are associated with social class as the majority of the respondents were concerned that such belonging might affect people's social class/status. c > Table 2: Principal Components Analysis of Groups Seen as Relevant for Definition of the Diverse Society V1.1: Which groups are relevant today for defining a diverse society? (Valid N= 741) Component 1 2 3 Transnational belonging (groups that are identifying with more than one country) .874 .292 .137 Shifting belongings (people whose belongings are under a process of change) .848 .273 .136 European belonging (groups identifying with the EU) .842 .281 .173 Global belonging groups (identification with humanity) .835 .308 .181 Multiple/mixed belongings (people identifying with more than one group) .826 .255 .149 Life-style groups (people identifying with different sorts of life-styles) .695 .262 .214 Territorial belonging (groups identifying with a specific region in a country) .690 .255 .111 Ideological groups (people identifying with a specific ideology) .601 .239 .390 Migrant groups (people coming from non-European countries) .531 .172 .078 Gender groups (men/women) .191 .782 .227 Disability groups (people with physical and mental disadvantages) .390 .709 .062 Sexuality groups (e.g., gays, lesbians, transsexuals, homosexuals, etc) .200 .649 .390 Generation (e.g., youth/elderly) .393 .643 .143 Social Class (e.g. workers, employers, farmers, rich, poor, etc) .370 .519 .179 Ethnic groups (people identifying with a specific ethnic group) .023 .261 .734 Religious groups (people identifying with a specific religion) .189 .268 .704 National belonging (people identifying with a specific nation) .459 -.019 .580 Contribution to explained variance (%) 49.90 8.65 5.50 The third dimension clusters the indicators measuring whether the respondents include national, religious, and ethnic groups in their definitions of a diverse society. I labelled this dimension "Traditional Orientation to Diversity." In this dimension, we measure how inclusive respondents are to group-based diversity created by the nation-state itself. DA of the three scales with the grouping variable "national vs. trans-European organisation" gave the results shown in Table 3, 21.8 percent of the interviewees from national organisations and 52.9 percent from trans-European organisations agree on a globally/transnationally-oriented definition of a diverse society. Inversely, 78.2 percent of national and 47 percent of trans-European elites agree on a national orientation to a diverse society. These results show nationalising and Europeanising discourses are disseminated at national-level and trans-European-level organisations, but the national orientation is stronger at the national level whereas the transnational/global orientation is stronger at the trans-European level. Table 3: Classification Results from Discriminant Analysis of Groups Relevant for the Definition of Diversity V6 National or Transnational Organisation? Predicted Group Membership Total National Trans-European Cross-validated Count % National 566 158 724 Trans-European 8 9 17 National 78.2 21.8 100.0 Trans-European 47.1 52.9 100.0 My second indicator concerning diversity views relates to the normative, onto-logical, or instrumental status each interviewee gives to ethno-national diversity. The respondents were asked what they thought about ethno-nationally diverse societies. The responses were classified according to whether the respondents regard ethno-national diversity as a normatively desirable goal in itself, or an inescapable fact, or a matter that defines the meaningful existence of persons, or a means to achieve other goals. Respondents' answers were coded into multiple categories when the answers fit more than one category. Table 4: Principal Components Analysis of the Status Given to Ethno-national Diversity V2.1 What do you think about ethno-nationally diverse societies? (Valid N=720) Component 1 2 3 The respondent sees an ethno-nationally diverse society as a desirable goal to achieve .869 -.301 -.214 The respondent does not attribute any normative or ontological status but sees ethno-national diversity as an inescapable fact of a social life -.835 -.376 -.214 The respondent sees an ethno-nationally diverse society as an ontological matter without which society's and/or an individual's existence would not be possible -.001 .969 -.054 The respondent sees an ethno-national diversity as means for achieving some other goals and not as a goal in itself -.014 -.044 .986 Contribution to explained variance (%) 36.54 29.65 26.11 Results from a PCA of these four categories are presented in Table 4. The first dimension is labelled "Normative vs. Realist Approach," and it measures respondents' tendency to view an ethno-nationally diverse society as a goal in itself or as an inescapable fact. Large positive values indicate perception of ethno-national y diversity as a goal in itself. Negative scores with larger absolute values indicate perceptions of ethno-national diversity as an inescapable fact whether or not one sees it as desirable or not. The second dimension is labelled "Ontological-Existential Approach." The higher scores with positive values on this scale indicate the respective respondents do not necessarily favour or not favour ethno-national diversity, but they accept it since they regard ethnicity and nationality as the foundation of people's social existence. Higher scores with negative values mean that the respective respondents do not perceive ethno-national diversity as an existential matter, but acceptable for other reasons. The third dimension is labelled "Instrumental Approach." Specific statements - e.g. ethno-national diversity "is enriching our culture," "stimulates economic development and innovation," "is a god way of fighting an aging society," "should be tolerated if we want to share our wealth with poor people," "is acceptable since it leads to a more just society/world," "is a necessary tool for protecting human c > rights," "needed if we want to have a more colourful society etc - are coded into this category. Higher positive values on this scale thus indicate instrumentalist approaches to ethno-national diversity. r\i Table 5: Classification Results from Discriminant Analysis of Views on Ethno- national Diversity PO V6 National or Transnational Organisation? Predicted Group Membership Total National Trans-European Cross-validated Count % National 452 294 746 Trans-European 6 12 18 National 60.6 39.4 100.0 Trans-European 33.3 66.7 100.0 The distribution of these views between levels is given in Table 5, 39.4 percent of the interviewees from national-level organisations and 66.7 percent from trans-European organisations share a normative view of diversity as a goal to achieve. However, 60.6 percent of the national and 33.3 percent of the trans-European interviewees share an instrumentalist and realist approach to diversity. That is, among the national-level elites, ethno-national diversity is acceptable because it is unavoidable, a necessity for meaningful social existence, and needed to achieve other goals. Views that do not see ethno-national diversity as a goal in itself dominate among the national-level elites. Inversely, views that regard ethno-national diversity as a goal in itself dominate among elites who work in trans-European organisations. Differences between National and Trans-European Elites' views on EU Polity. Application of PCA on the five items listed in Table 6 resulted in three dimensions. The first dimension measures the extent to which the respondents want a development where policymaking/decision competences between the member-state and EU levels are differentiated and divided between levels according to different policy areas. Based on an inspection of the answers about different policy areas in qualitative interviews, I have interpreted this dimension as measuring the preference for a system of multi-level governance (MLG). In addition, an inspection of the respondents' preferences concerning decision levels in different policy areas in the quantitative data set supports this interpretation. Large positive values mean a preference for multi-level governance whereas large negative views mean the absence of this preference. Table 6: Principle Components Analysis of the Views on EU Polity Development V3.1 In which direction should EU polity develop in the future? (Valid N=663) Component 1 2 3 More centralisation, but in certain policy fields .804 .003 -.158 More autonomy for member states, but in certain policy fields .782 -.037 .007 More federalisation at large -.293 .802 -.270 More autonomy for member states -.339 -.722 -.380 More centralisation -.156 -.024 .919 Contribution to explained variance (%) 29.83 23.53 21.36 The second dimension can be interpreted as measuring the preference for a multi-level federal polity (MLP) versus more autonomy for member states in all areas. "Autonomy for member states" and "federalisation at large" load on the same dimension with opposite signs, making this dimension meaningfully bipolar. Large positive values imply a pro-federalisation attitude, and large negative values imply pro-member state autonomy attitudes. The third dimension measures the extent to which a respondent is for more EU centralisation regardless of policy areas - that is, a preference for building a centralised EU polity (EUP). Large positive values indicate pro-centralisation attitudes, and large negative preferences mean the absence of this preference in a respondent. Cases with very low values on all three dimensions display a general anti-EU preference, and even a preference for dissolving the EU. Table 7: Classification Results from Discriminant Analysis of the Views on EU Polity Development V6 National or Transnational Organisation? Predicted Group Membership Total National Trans-European Cross-validated Count % National 545 160 705 Trans-European 9 8 17 National 77.3 22.7 100.0 Trans-European 52.9 47.1 100.0 As indicated in Table 7, 22.7 percent of the interviewees from national organisations and 47.1 percent of the interviewees from trans-European organisations agree on establishing a MLG or (to less extent) a MLP. However, 77.3 percent of national-level interviewees and 52.9 percent of trans-European interviewees agree on more decentralisation and more autonomy for member states. Differences in Elites' Preferred Addressees in the European Public Sphere. y This section is based on a set of variables measuring the extent to which, and Q whom, actors want to target as the addressees of their messages or claims in their £ communications and interactions. The first column in Table 8 lists the different authorities and organisations the respondents mentioned as their addressees. A PCA resulted in two dimensions. The first dimension encompasses the different European and EU political and judicial authorities - that is, the addressee is an institution at the European level, 0 and the communication is upward. The second dimension measures the extent to which an actor's targeted addressees are other organisations, networks, groups, etc., including the European Commission, the European Parliament, and European parties/party families. Unlike the first dimension, communication and collaboration here do not necessarily imply a vertical or hierarchical but rather a horizontal structure of communication. Table 9 shows 2.3 percent of the interviewees from national-level actors and 31.3 percent of the interviewees from trans-national actors want to be involved c > in vertical communication structures. However, 97.7 percent of the national actor interviewees and 68.8 percent of the transnational actor interviewees want to be primarily involved in horizontal communication structures. Table 8: Principal Components Analysis of the Actors' Addressees in the Public Sphere LO V5.10 Which actors on all levels (international, supranational, national, sub-national, i.e., regional and/or local) do you want to address with your activities? (Valid N=544) Component 1 2 European Court of Auditors .844 .079 European Ombudsman .841 -.021 European Economic and Social Committee .774 .310 Presidency of the Council .757 .321 European Committee of the Regions. Agencies .745 .232 Council of the European Union .724 .269 Council of Europe .713 .234 European Council .677 .325 European Court of Human Rights .652 .224 European Court of Justice .643 .193 European Commission .441 .375 Gender organisations/networks .174 .709 Ethnic minority organisations/networks .189 .672 Religious organisations/networks .181 .665 Political parties and/or party families .058 .634 Lobbies .229 .622 Citizens in general .128 .454 European Parliament .374 .443 Contribution to explained variance (%) 41.53 10.08 Elites at the national and trans-European levels clearly prefer horizontal trans-European interactions. This trend is much more pronounced within the national-level organisations. A closer examination of the in-depth interviews also shows many of those who favour involvement in horizontal networks and who simultaneously want to involve EU political institutions as little as possible in their trans-European affairs do so because they are sceptical about the EU's democratic qualities, and they do not want to be part of the legitimisation mechanisms the EU has devised. Some political elites stated they already had good communication and collaboration channels with their sister parties in other countries, through party federations and one-to-one contacts between the party elites. Further, the national-level SMO/NGO leaders who prefer horizontal Europeanisation say this process started before the European Union existed and should continue especially now in the new political Table 9: Classification Results from Discriminant Analysis of the Actors' Addressees V6 National or Transnational Organisation? Predicted Group Membership Total National Trans-European Cross-validated Count % National 516 12 528 Trans-European 11 5 16 National 97.7 2.3 100.0 Trans-European 68.8 31.3 100.0 context of Europe, which is characterised by pooling of sovereignties so that the new concentrated power can be effectively criticised and controlled by citizens. The interviewees also think for issues on which some national governments are not responsive enough (e.g. women's rights, minority rights, environmental protection), European-level institutions can be a good tool for making national governments change their courses of action. Since the interviewees' own aim is to make sure that the interests they voice be protected, horizontal Europeanisation uninfluenced by EU premises is, for them, a better alternative. If necessary, European political institutions can be addressed for this purpose, but the European level should not, in their eyes, be taken for granted as a legitimate authority in all matters. This trend is clear concerning organisations operating at the national level. In addition to those who favour horizontal trans-Europeanisation, we find national-level elites who seek to address only national governments and authorities in their activities. Here, the concern is the survival of the nation state rather than the democratic legitimacy of EU political institutions. Trans-European elites, on the other hand, perceive their role as mediators between European Union institutions and the national-level organisations. Trans-European elites are aware they cannot claim to be representing anybody, but what they do is important and needed, because the new power structures in Europe require trans-European organisations that can articulate the common interests of European civil societies. However, trans-European organisations strive on both fronts. Access to EU decision-making mechanisms is difficult although some of the organisations have been defined by the European Commission as official consultation partners in the matters they specialise in. They think it is also difficult to gain the full trust of national-level member organisations because they are sometimes regarded as too close to the EU. This view was confirmed by interviews with national-level political party and SMO/NGO elites. In addition to the perception that trans-European elites may be ideologically closer to the EU than to the grassroots, national-level elites are also y concerned about the EU terminology adopted by trans-European elites. In the eyes of national-level elites, the difficulty of this terminology makes communication between national and trans-European-level elites at times ineffective, and this challenge also makes it difficult for national-level elites to actively participate in trans-European-level activities. However, trans-European elites tend to see EU terminology as a practical necessity that makes it possible to communicate with and disseminate contention toward EU policymakers. The majority of the trans-European elites state that it is important that the national-level civil society and political organisations understand the necessity of acting together on issues that require European-level solutions, but it is not always easy to persuade their member organisations to be more active. Further, the elite interviews and our institutional data document that trans-European organisations usually operate with a very small number of full-time staff members, which makes it difficult to prioritise integration activities for nationallevel organisations. The most ambitious trans-European organisation in creating a 0 c > high level of integration, by creating a common understanding of common problems, is the EWL. This organisation uses considerable staff resources and voluntary resources to integrate women's organisations from Central and Eastern European countries. In addition, ENAR appears to be concerned about linking with member state-level anti-racist organisations. On the other side of the coin, 2.3 percent of the national-level and 33.3 percent ^ of the trans-European-level elites say they want to address the intergovernmental and supranational bodies in Europe with their activities. The trend within the trans-European organisations is not negligible. Among the trans-European organisations, the Social Platform appears to be the most oriented toward using the European Union institutions, and specifically the European Commission, as one of the primary addressees of their activities. Discursive Misalignments between National and Trans-European-level Elites? These findings point to misalignments between the values of national and trans-European elites. If trans-European organisations are supposed to represent/aggregate the interests of European civil society regarding the EU, this can be perceived as a legitimacy problem on the part of the trans-European organisations. Even when we assume a somewhat less ambitious mission for them, such as articulating interests, it is not possible to ignore this mismatch. Certainly, diversity of views and political polarisation in the public sphere are necessary and desirable from a democracy point of view. However, what we observe here is not only a horizontal polarisation but also a vertical, hierarchical polarisation between the member-state and trans-European-level organisational elites. Some of the trans-European elites interviewed work in organisations officially involved in EU-level policy processes as regular consultation partners - this is especially true for the Social Platform, ENAR, and the EWL. Although an overwhelming majority of the interviewed trans-European NGO/SMO elites are aware they cannot claim to represent the European civil society, they claim to represent social and political norms for the good of all - thus investing in output legitimacy rather than input legitimacy. The three party federations we interviewed are supposed to represent their member parties, and they have representatives in the European Parliament. Low electoral turnout, combined with mismatches between national-level and trans-European-level elite views, also points to a hierarchical structuring of the trans-European political spaces. Although the think tank networks - EPIN and TEPSA - and their member organisations we interviewed are not expected to represent anybody other than themselves and their expertise, they provide policy assessments, evaluations, and advice to the European Union. The European Commission and other EU political institutions take these trans-European organisations as the most relevant conversation partners in certain policy issues, and have privileged them and institutionalised their participation in consultation processes in different ways. However, the views these institutions disseminate about diversity, ethno-national diversity, and legitimate addressees in the European public sphere are fundamentally different from the views expressed by elites working in national-level organisations. In addition, the European Union's consultation system provides opportunities for other organisations and individual citizens to express their views on policy issues. Organisations' Networking Patterns in the European Public Sphere In the following set of PCAs and DAs, the unit of observation and analysis is organisation. Data about the organisations networking and interactive patterns were gathered from their printed and online documents (annual reports, activity reports, leaflets, brochures, descriptions of ongoing projects and project partners, and secondary literature where available). The following principal components and discriminant analyses of organisations' networking patterns include sub-national, national, and trans-European interactions. Organisations' Collaboration Patterns. Table 10 shows the results from a PCA of the operative levels of networks the interviewed organisations are actually involved in. The 46 media actors in the data set are excluded from this analysis as the networking they do is not comparable with the networking of the three other types of organisations. Table 10: Principal Components Analysis of the Organisations' Networks Organisations/networks the organisation collaborates with N=158 Component 1 2 Regional organisations/networks .921 -.063 National organisations/networks .631 .543 Trans-European organisations/networks -.012 .938 Contribution to explained variance (%) 49.64 31.22 u S i 0 £ * i * (R 0 C « 00 The first component measures the extent to which an organisation is involved in sub-European (regional and national) networks, and the second measures an organisation's involvement in trans-European networks and national networks. The variable "national organisations/networks" loads on both dimensions. This indicates the majority of the organisations in our data material have national networks. However, those with large positive scores in the first dimension are also involved in sub-national networks, and those with large positive scores in the second dimension, in addition to their national networks, are involved in trans-European networks. This implies the presence of and a distinction between national multilevel and European multi-level networking structures in Europe, strengthening my expectation that national boundaries and European multi-level governance structures would lead to this type of networking structure. Table 11: Classification Results from Discriminant Analysis of the Organisations' Networks National or transnational? Predicted Group Membership Total National Transnational Cross-validated Count % National 142 3 145 Transnational 5 2 7 National 97.9 2.1 100.0 Transnational 71.4 28.6 100.0 Table 11 presents the distribution of these two networking patterns between trans-European and national-level organisations. We observe that 98 percent of C^ member state-level organisations collaborate primarily with organisations' national ^ and sub-national networks. However, 71.4 percent of the trans-European organisations also primarily collaborate with national and sub-national-level organisations, whereas 28.6 percent of trans-European organisations cooperate with national organisations and other trans-European networks. As the percentage of national-level organisations collaborating with other national organisations and simultaneously with trans-European networks is low (2.1 percent), and the percentage of transnational organisations that collaborate with national-level organisations is high (71.4 percent), trans-European organisations collaborate with only a small selection of national-level organisations. This is certainly true in the case of the trans-European think tank networks, which prefer to include only one think tank from each EU member country. The same argument goes for party federations, which collaborate with a limited number (preferably only one) political party in each member country. As to the SMOs and NGOs, ENAR and the EWL also have limited the number of organisations from each country, often to only one, in their membership lists. However, the Social Platform is a network of networks, and individual organisations cannot be members in the Social Platform. Even without considering the results presented in Table 11, the membership structure of trans-European organisations demonstrates the number of nationallevel organisations involved in trans-European networks is quite low. The results I obtained from the analysis of the interviews (Table 9) are almost identical with the results from this analysis of the institutional data. Combining these results, I conclude organisational elites are quite consistent in their intensions and actions: To a large degree, they do not want to have intergovernmental and supranational authorities as addressees of their activities; in practice, they do not collaborate with trans-European organisations that have these authorities as the main addressees of their activities. Scope of Organisations' Collaboration with Networks and Other Organisations. A PCA of six variables indicating how organisations collaborate in their national, sub-national, and trans-European networks resulted in one component (Table 12). The variables in the first column measure different types of collaboration forms. The variables "attempts at mutual information sharing," "efforts to synchronise separate projects/action plans," "collaborative projects/actions," "joint/projects/actions," "attempts to formulate common objectives to address common concerns," and "attempts to formulate common objectives" represent ordinal-ranked categories of the variable collaboration scope. However, the PCA did not distinguish between variables measuring project-/action-based collaboration and more strategic collaboration to achieve long-term objectives; I will stick to interpreting this scale as an indicator of the organisations' collaboration scope. Thus, the extracted single component can be interpreted as a measure of the size of the collaboration repertoire of organisations. The higher an organisation's score, the more collaborative activity types in which the organisation participates. Smaller scores indicate less collaboration activity with networks and other organisations. However, the largest scores with a positive sign are also forms of collaboration Table 12: Principal Components Analysis of the Organisations' Actions in Trans-European Networks N=158 Component 1 Efforts to synchronise separate projects/action-plans .786 Attempts at mutual information-sharing .763 Attempts to formulate common objectives .721 Joint projects/actions .719 Collaborative projects/actions .702 Efforts to formulate common objectives to address common concerns .622 Contribution to the explained variance (%) 51.93 aiming to achieve longer-term common objectives. Whereas the indicators I constructed in the previous section measure the extent to which organisations network with organisations operating at different levels, this indicator tells us what they do when they collaborate. Table 13 shows 60 percent of trans-European-level organisations have a larger collaboration scope or repertoire, and 76.5 percent of the national-level organisations have smaller collaboration repertoires. This is certainly not surprising since the survival of trans-European networks largely relies on collaboration with member organisations and other networks. Table 13: Classification Results from Discriminant Analysis of the Organisations' Actions in Networks u S i 0 £ * I * (R 0 C « o CN National or transnational? Predicted Group Membership Total National Transnational Cross-validated Count % National 117 36 153 Transnational 2 3 5 National 76.5 23.5 100.0 Transnational 40.0 60.0 100.0 What do these numbers tell us about national and trans-European-level organisations? First, a much smaller percentage of national-level organisations than trans-European organisations get involved in collaboration that requires agreement on common objectives. Second, a considerable portion (40 percent) of the trans-European organisations has this collaboration repertoire. Still, 23.5 percent of national-level organisations and 60 percent of trans-European-level organisations do get involved in collaboration that either may lead to or has led to formulation of common objectives. Indeed, this is a lot and implies individual organisations are coming together to stand on the different poles of whatever kind of political spaces they are operating in. The results cover collaboration at all levels (local, national, or European). Organisations' Membership Status in Networks. Our institutional data also covers information about the organisations' membership status in trans-European networks. The PCA presented in Table 14 is based on three variables indicating whether organisations have active or passive membership status or observer status in their networks. The analysis gave two components that distinguish between organisations that are members and organisations that have only observer status in their networks. Table 14: Principal Components Analysis of the Organisations' Membership Status in Networks Status of the organisation in selected networks N=160 Component 1 2 Passive membership status (only voting rights) .820 -.147 Active membership status (with voting and representation rights) .688 .267 Observer status .039 .961 Contribution to explained variance (%) 40.03 32.1 The first component measures whether an organisation has active membership status in the network with voting and representation rights (large positive values). The higher scores indicate membership with voting and representation rights, and the smaller values indicate only passive membership status without representation rights. The second component measures whether a non-member organisation has observer status in an organisational network. Larger values indicate observer status, and smaller values indicate the absence of observer status. Organisations that score low on both dimensions are those that do not have membership or observer status in any organisational networks; however, this does not mean the organisations do not collaborate with networks. Table 15: Classification Results from Discriminant Analysis of Membership Status in Networks National or transnational? Predicted Group Membership Total National Transnational Cross-validated Count % National 145 10 155 Transnational 4 1 5 National 93.5 6.5 100.0 Transnational 80.0 20.0 100.0 Table 15 shows 6.5 percent of national-level organisations and 20 percent of trans-European organisations have strong membership statuses in organisational networks. The one trans-national organisation with strong membership status in a network is ENAR - which is a member of the Social Platform. The Structuring of Trans-Europeanising Public Spaces The conceptual framework of this paper defines "the articulation of trans-Eu-ropeanising public spaces" in terms of two features: (1) generating trans-European discourses and (2) creating trans-European networks. Fulfilling either of these criteria means contributing to creating trans-Europeanising public spaces. Although Europeanising and non-Europeanising discourses exist in national and trans-European-level organisations, non-Europeanising discourses domi- nate in national-level organisations, and Europeanising discourses dominate in trans-European organisations. Concerning discourses on diversity and the future development of an EU polity, the gaps between the views of national- and trans-European-level elites are more or less similar across the different topics analysed here. The largest gap between trans-European-level and national-level elites' views is between their acceptance of the EU political institutions as legitimate addressees in the public sphere. Organisations' networking patterns at the institutional level also indicates that, while the majority of member state-level organisations and all of the trans-European-level organisations are involved in horizontal trans-European relations with organisations in other European countries, very few national-level organisations are involved both horizontally and vertically in trans-European relations. This finding is consistent with the findings in the analysis of their discourses concerning legitimate addresses in the European public sphere. These findings show the most active and influential social and political actors at the member-state level prefer and are working to achieve a horizontal trans-Europeanisation in Europe - by leaving out from their communication paths and collaborative work EU political institutions and trans-European networks that draw on EU institutions as their addressees. These results point to the existence of trans-Europeanising political spaces, with Europeanising discourses and/or trans-European ties between organisations at the national and European levels. Earlier research - on especially the media public sphere - convincingly shows the current European public sphere is horizontally segmented along national lines in Europe. While this study shows the same tendency exists in the discourses and networking patterns of the central organisations participating in public debates, it also finds that there is a notable discursive rapprochement between member-state and trans-European-level elites. More importantly, trans-Europeanising political spaces, i.e. the component of the European public sphere, which is expected to contribute to the weakening of the national boundaries, may also potentially divide the European public sphere vertically. There are some discursive gaps between the views of national and European-level elites. Further, networking patterns also show this gap is not only in discourses but also in interactions. This implies a significant lack of interconnectedness between national and trans-European publics. In the future, this currently weak vertical division may contribute to the emergence of a horizontally and vertically segmented £ European public sphere. However, if Stein Rokkan's conclusions (Rokkan 1975, Rokkan et al. 1987) pertaining to European national state-building processes hold ^ true for the building of the European Polity, such vertical segmentation may also create common transnational reactions from the grassroots, resulting in integration of the European peripheries against the multiple political centres of the EU. Acknowledgement More than 130 Eurosphere researchers in 16 European countries collected and registered the data that are used in this article. They also wrote country reports and comparative reports from which I benefited immensely when trying to understand the data in depth. I would like to express my sincere gratitude to them. A Notes: m 1. According to Olesen (2005), "Visibility refers to the degree to which frames are heard and seen in the public sphere," and "Resonance refers to the degree to which frames elicit a response from t—I interested parties; for example likeminded activists and social movements, media, politicians and the targets of claims (for example states and institutions)." 2. For detailed information about rules and procedures for selecting organisations and interviewees, see Eurosphere Research Notes no. 9 and 13 at http://eurospheres.org/publications/research-notes/. References: Eriksen, Erik Oddvar. 2005. An Emerging European Public Sphere. European Journal of Social Theory 8, 3, 341-363. Eder, Klaus and Hans-Joerg Trenz. 2003. The Making of a European Public Space: An Inquiry into Relating Supranational Politics and Transnational Spaces of Communication. The Case of Justice and Home Affairs. In B. Kohler-Koch (ed.), Linking EU and National Governance, 111-134. Oxford: Oxford University Press. Eder, Klaus and Cathleen Kantner. 2000. Transnationale Resonanzstrukturen in Europa. Eine Kritik der Rede vom Öffentlichkeitsdefizit In M. Bach (ed.), Die Europäisierung nationaler Gesellschaften. Sonderheft 40 der Kölner Zeitschrift für Soziologie und Sozialpsychologie, 306-331. Wiesbaden: Westdeutscher Verlag. Gerhards, Jürgen. 2000. Europäisierung von Ökonomie und Politik und die Trägheit der Entstehung einer eu-ropäischen Öffentlichkeit. In M. Bach (ed.), Die Europäisierung nationaler Gesellschaften. Sonderheft 40 der Kölner Zeitschrift für Soziologie und Sozialpsychologie, 277-305. Wiesbaden: Westdeutscher Verlag. Habermas, Jürgen, Sara Lennox and Frank Lennox. 1974. The Public Sphere: An Encyclopedia Article (1964). New German Critique 3, 3, 49-55. Koopmans, Ruud. 2004. Movements and Media: Selection Processes and Evolutionary Dynamics in the Public Sphere. Theory and Society 33, 3-4, 367-391. Koopmans, Ruud 2007. Who Inhabits the European Public Sphere? Winners and Losers, Supporters and Opponents in Europeanised Political Debate. European Journal of Political Research 46, 2, 183-210. Koopmans, Ruud and Barbara Pfetsch. 2006. Obstacles or Motors of Europeanization? German Media and the Transnationalization of Public Debate. Communications 31, 2, 115-138. Koopmans, Ruud and Jessica Erbe. 2004. Towards a European Public Sphere? Innovation: The European Journal of Social Science Research 17, 2, 97-118. Olesen, Thomas. 2005. Transnational Publics: New Spaces of Social Movement Activism and the Problem of Global Long-Sightedness. Current Sociology 53, 3, 419-440. Peters, Bernhard, Stefanie Sifft, Andreas Wimmel, Michael Brüggemann and Katarina Kleinen vom Königslöv. 2005. National and Transnational Public Spheres: The Case of the EU. European Review 3, 1, 139-160. Risse, Thomas and Marianne Van de Steeg. 2003. An Emerging European Public Sphere? Empirical Evidence and Theoretical Clarifications. Paper presented to the conference Europeanization of Public Spheres, Political Mobilization, Public Communication and the European Union. Science Center Berlin, June 20-22, 2003. Risse, Thomas. 2010. A Community of Europeans? Transnational Identities and Public Spheres. Ithaca and London: Cornell University Press. Rokkan, Stein. 1975. Dimensions of State Formation and Nation-Building: A Possible Paradigm for Research on Variations within Europe. In C. Tilly (ed.), The Formation of National States in Western Europe, 562-600. Princeton: Princeton University Press. Rokkan, Stein, Derek W. Urwin, Frank H. Aarebrot, Pamela Malaba and Terje Sande. 1987. Center- Periphery Structures in Europe. New York: Campus Verlag. Schlesinger, Philip. 1999. Changing Spaces of Political Communication: The Case of the European Union. Political Communication 16, 3, 263-279. Sicakkan, Hakan G. 2006. Glocal Spaces as Prototypes of a Future Diverse Society: An Exploratory Study in Six European Countries. In H. Sicakkan and Y. G. Lithman (eds.), What Happens When a Society is Diverse? Exploring Multidimensional Identities, Studies in Political Science 30, 105-148. New York: The Edwin Mellen Press. Splichal, Slavko. 2011. Transnationalization of the Public Sphere and the Fate of the Public. New York: Hampton Press. Trenz, Hans-Jörg. 2000. Korruption und politischer Skandal in der EU. Auf dem Weg zu einer europäischen politischen Öffentlichkeit. In M. Bach (ed.), Die Europäisierung nationaler Gesellschaften. Sonderheft 40 der Kölner Zeitschrift für Soziologie und Sozialpsychologie, 332-359. Wiesbaden: Westdeutscher Verlag. Trenz, Hans-Jörg. 2003. Media Coverage on European Governance. Testing the Performance of National Newspapers. EUSA Biennial International Conference. Nashville, TN: March 27-29. s I 0 £ * 1 * (R 0 C 117