MAPPING EUROPEAN IDEOSCAPES: Examining newspaper debates on the EU Constitution in seven European countries^* Open Access

The public debate over the EU Constitutional Treaty (EUCT) – culminating in the French ‘non’ and the Dutch ‘nee’ in May 2005 – revealed a clash of markedly different understandings of the Treaty and its significance for the future of the European Union. While some greeted the failure of the Treaty as an indication of the long-awaited revenge of nation-states against the all-mighty Brussels, others saw it as a regrettable victory of national interests over common European interests. Yet others evaluated the outcome of the two referenda on the basis of a completely different set of criteria, interpreting it as a sign of hope for the European social model, or as a disappointing triumph of protectionism over the market forces.

This paper charts the distribution of these different interpretations in the newspaper coverage in seven European states: the Czech Republic, Germany, France, the Slovak Republic, Slovenia, Switzerland, and the UK. In doing so, however, it departs from the existing discussions and empirical research on European mass communication, that examines European communication predominantly, if not exclusively, in terms of a Europeanization of national communicative spaces, spheres or publics (van de Steeg 2002; Brüggeman et al. 2005; Machill et al. 2006). We argue that existing theorizing and research on the European public sphere, while declaratively abandoning the nation-state-centred model, actually fails to move beyond methodological nationalism, and therefore unintentionally perpetuates an erroneous understanding of the ‘national’ public sphere. Despite embracing the rhetoric of globalization and transnationalization, most scholars continue to assume at a methodological and practical, if not theoretical level, the existence of fairly homogenous national communicative spheres, discourses, and publics, largely coterminous with state-borders. In contrast, this paper develops a number of conceptual and methodological solutions that allow us to scrutinize how European mass communication is segmented by the mutually entwined imbrications of class and nation, on sub-national and sub-state as well as inter-state and transnational levels.¹

The past decade has witnessed a sweeping change of social scientific theoretical imagery. In trying to come to grips with developments in the latter part of the twentieth century, social theorists have devised a series of concepts and promoted a number of all-encompassing theoretical shifts. We were told that in order to make sense of a world in which the walls of nation-states came crumbling down, we should move from the examination of the ‘space of places’ to the exploration of the ‘space of flows’ and ‘networked connections’ (Castells 1996). If we wanted to understand the contemporary world, we needed to abandon the ‘sociology of societies’, and develop instead a ‘sociology of mobilities’ (Urry 2000). When studying culture, we ought to avoid looking solely at national environments, and examine instead the global cultural flows in terms of various de-territorialized ‘scapes’ – ‘ethnoscapes’, ‘mediaspaces’, ‘technoscapes’, ‘finanscapes’ and ‘ideoscapes’ (Appadurai 1996).

The rise of the new theoretical imagery, characterized by the metaphors of ‘flows’, ‘networks’ and ‘scapes’, is discernible also in the writings on the EU and Europeanization. In a recent revision of research and theorising on the topic, Gerard Delanty and Chris Rumford (2005: 1) make an attempt to ‘outline a conception of Europeanization in terms of a theory of society beyond national societies’. According to them, ‘the emerging social order cannot be fully understood by reference to traditional comparative methods of analysis which all presuppose national societies and their convergence’ (ibid: 4). The EU is increasingly approached as a multilayered polity, operating through a network of multiple power centres engaged in governance at different levels (Grande 1996; Hooghe 1996; Hooghe and Marks 2001), and adapted to functioning in an increasingly globalized environment (Held et al. 1999; Held and McGrew 2002). Castells (1997), for example, argues that the EU is best understood as a ‘network state’, a novel form of state characterized by multi-level governance and fully embedded into globalizing trends.

Just as there is a growing consensus that the EU should not be seen as just another nation-state, there also seems to be a growing agreement that the European public sphere should not be seen as a variation on the theme of the national public sphere. In other words, the European public sphere is no longer conceived as a pan-European communicative sphere that is structurally similar to but independent from individual national public spheres (van de Steeg 2002; Schlesinger and Fossum 2005; Machill et al. 2006). Instead, this communicative sphere is now increasingly being described in terms of a network (Schlesinger 2003: 10–3). The EPS is thus seen as ‘a sphere of publics’ (Schlesinger and Kevin 2000), ‘a highly complex network [that] branches out into a multitude of overlapping international, national, regional, local and subcultural arenas’ (Habermas 1997: 373–4), ‘an aggregate’ of media spheres (van de Steeg 2002: 508), and ‘an increasingly interconnected grouping of overlapping communicative communities with the potential to become a loosely integrated communicative space’ (Schlesinger 2003: 4).

Despite this obvious shift away from the nation-state-centred approach, however, there remains a stubborn insistence on national public sphere being the fundamental, if not the only, building block of such a network-like European communicative space. In one way or another, national public spheres are seen as the starting point for the emergence of a European public sphere. As a rule, research into the so-called Europeanization of the mass media – usually limited to the investigation of quality newspapers (Eder and Kantner 2000; Trenz 2004, 2005; Wimmel 2004; Downey and Koenig 2006) – focuses on the extent of Europeanization of national public spheres that approached as homogenous. Europeanization is thus seen as ‘a process that enlarges the scope of public discourse beyond the territorial state’ (Brüggeman et al. 2005), and the European public sphere is expected to emerge on the basis of a ‘Europeanization of national publics’ (Machill et al. 2006: 63).

While the continuing importance of the national in European communication cannot be denied, paying attention exclusively to the relationship between the national and the transnational (including the European) can easily obscure important differences arising along other lines, and lead to a form of methodological nationalism. Together with Slavko Splichal (2006: 707) we could ask: ‘Why should a nation have the privilege of generating the public? Why can't other types of collectivities also have their own publics?’ An almost direct consequence of taking the national as the only important analytical distinction is the tendency to overstate the internal homogeneity of national units. This tendency is a long-standing feature of comparative research, not limited only to media and communication studies, and it has often been argued that nations are not appropriate units of comparison (Livingstone 2003). Some authors dealing with issues of European communication do acknowledge that the internal homogeneity and distinctiveness of national public sphere is often exaggerated. Marianne van de Steeg (2002: 501–2), for example, points out that ‘the assumption in most of the literature on a European public sphere is that each member state has its own national public sphere’, and that these national spheres are ‘clearly identifiable and self-contained’, and ‘internally highly homogenous’, The account of the national public sphere is therefore highly idealized: ‘the heterogeneity, variety and difference within what is deemed to be the national public sphere are missed’ (506). In a similar vein, Hans Jörg Trenz (2004: 292–3) questions the assumption of a unitary national public sphere, points out that ‘different media do not always travel along the same roads’, and admits that assessing the (non)existence of the European public sphere on the basis of quality newspapers alone introduces a particular bias. Yet neither of the authors makes the further step of suggesting an alternative approach, one that would allow us assess the diversity of content in European mass media in a more complex manner.

We have witnessed a new wave of attacks on methodological nationalism in recent years. Unlike the early criticisms arising in the 1970s (Chernilo 2006: 6–9), this new wave arose in response to the perceived fundamental transformation of modern societies in the latter part of the twentieth century, and has been taken up not only by methodologically-minded scholars, but also by some of the mainstream social theorists. The equation between society and the nation state has been identified as one of the main impediments to understanding the dynamics of the increasingly interconnected and globalized society of the new millennium. Due to that, we need to dispense with the tacit assumption that the nation-state is the natural and necessary container of society, and develop modes of analysis based on different analytical units, consistent with the structures of the modern, globalized societies. As such, the recent turn away from methodological nationalism – and towards what Ulrich Beck (2000, 2004) has termed ‘methodological cosmopolitanism’ – constitutes yet another in the series of theoretical shifts suggested by social theorists trying to understand social, political and economic transformations of the late twentieth century.

Does this mean that we should simply abandon the vocabulary of the national, and embrace methodological cosmopolitanism championed by Ulrich Beck? The answer should surely be no. Instead of discarding the old vocabulary altogether, ‘the task is to expand and modify our established vocabulary rather than to assume that it has completely lost its usefulness’ (Schlesinger and Fossum 2005: 16). We share Daniel Chernilo's (2006) view that Beck overstates the contrast between what he sees as the ‘first age of modernity’ and the ‘second age of modernity’, or ‘simple’ and ‘reflexive’ modernization. His drive for a methodological cosmopolitanism is based on the belief that it is the social world itself that is being transformed and requires new methodological tools. He does not acknowledge that methodological nationalism was an inadequate tool already for describing the ‘first age of modernity’, and that methodological cosmopolitanism may be similarly unable to capture the complexities of modern societies if it dispenses with the national altogether.

For decades, class relations were downplayed in theories of nationalism, and vice versa, theories of class paid little attention to questions of nationality. We agree that the media and communication field is in dire need of ‘reconstructing the ruined tower’ of class (Murdock 2000), yet we argue that this reconstruction should not loose sight of the national. Classes and nations should be conceived as imagined communities that arose together (Mann 1993). They should be treated as ‘mutually entwined forms in which the self-consciousness of modern society is expressed’, and we should acknowledge that ‘we cannot capture their meanings unless we study them relationally’ (Fine and Chernilo 2003: 244). Due to that, classes and nations are also, by necessity, two fundamental forms of social division that underpin the structure of modern mass communication, and we cannot capture their effect on the segmentation of media spheres, markets and publics unless we examine them together. In our examination of newspaper debates on EUCT we therefore expect to find significant differences between newspaper discourse arising both along national as well as along class lines. The patterns of these differences should help us better understand the relations between class, nation and Europeanization in European mass communication.

Besides helping us avoid methodological nationalism, reintroducing class as one of the fundamental analytical distinctions has further benefits. The adoption of the new theoretical imagery, characterized by the metaphors of ‘flows’, ‘networks’ and ‘scapes’, often goes hand-in-hand with an unqualified celebration of globalization, fluidity and hybridity, and rather naïve neglect of issues of persisting inequalities and power relations. Although being network-like, the EU is not, as some would have us believe, a radically new, cosmopolitan and fluid structure. As Klaus Eder (2006: 260) reminds us, the EU is ‘not a case of deterritorialisation, of shifting boundaries or of dynamic networks with open boundaries, but rather the opposite: a pure case of territorial institutionalisation’. In a similar vein, the European communicative spaces, despite their multilayered and network-like structure, are still criss-crossed by multiple forms of exclusion. We should keep in mind that the emphasis on ‘fluidity’, ‘hybridity’, ‘flexibility’ and related terms often goes hand in hand with neoliberalism (Bourdieu and Wacquant 1999), and is increasingly becoming accepted as common sense among the global, transnational elites – precisely that social stratum, that is, that can afford being highly mobile and reap the benefits of globalization (Friedman 2000). By reducing the analysis to comparisons between national units and examinations of transnational similarities between them, we would in fact uncritically reproduce a view of the world that may well be the view of a particular social group. Paying attention to divisions along lines of class should help us avoid this trap.

And lastly, while keeping an eye on the persisting inequalities within national communicative spaces, we should not miss the persisting inequalities between these spaces. To put it differently: different nationally-bounded communicative spaces may share a number of ideological affinities, yet this does not necessarily make them equal. Are the different nationally bounded communicative spheres that share similar ideological attitudes towards Europe actually paying attention to each other? Is there a hierarchy of attention? Which of these spaces set the contours of the debate, and which are among the more passive – though attentive – followers? This is yet another layer of analysis that should help us appreciate the variegated ideological alliances and relations of power among the various European communicative spaces.

The data were taken from digitized versions of daily newspapers in seven countries, for the period from May 1 to December 31, 2005. In the German, French, British, and Swiss cases we relied mainly on the on the Lexis/Nexis Executive database, except for Bild, which was harvested through Google site searches from Bild's online paper, which unfortunately do not replicate the print version in its entirety. For the Czech Republic we relied on the electronic database of the Anopress IT company.² The Slovak data were gathered partly from the online archives of newspapers (in the case of Hospodarske noviny and Pravda), and partly by using the electronic database of the Newton I.T. company. In the Slovenian case, we harvested the data from the online archives of five major daily newspapers.³

In choosing the newspapers, we aimed to replicate the structures of the daily press markets. Our coverage for the accession countries press markets was most complete. For Slovenia, we collected data for five major papers, two of which (Dnevnik and Večer) have clear regional strongholds, one is a financial broadsheet (Finance), and one a tabloid (Slovenske novice). They combine for a 90 percent market share in terms of circulation figures (Bašič-Hrvatin and Milosavljevič 2001: 17). In the Czech case, the market is highly nationalized. Although we collected data only from one regional paper (and five nationwide published papers), our sample covers 86 percent of the total dailies’ market.⁴ Similarly, in Slovakia four national newspapers were collected for the sample, consisting of three broadsheets (SME, Pravda and the financial broadsheet Hospodarske noviny) and one tabloid (Nový Čas). This means almost a universal coverage (86 percent of the daily press, 93 percent of the national daily press),⁵ as the only national daily missing is the Hungarian-language Új Szó. In the UK, the market is relatively nationalized, even though there are some fairly widely read regional daily newspapers. Consequently we sampled all nationwide broadsheets, mid-markets, and tabloids, which cover 61 percent of the total dailies market.⁶ In addition, the Belfast Telegraph was chosen to represent the peculiar situation in Northern Ireland. In contrast to the aforementioned countries, Germany and France have stronger regional papers and Switzerland is almost entirely regionalized (Hallin and Mancini 2004: 25). Because of the proliferation of these markets, it was not feasible to cover papers that would combine to similar market share figures. Instead, we aimed to include all papers that are (according to Adam, Berkel, and Pfetsch 2003: 99–101) quality papers, several papers from different regions in these countries and the most important tabloids. This way we covered 30 percent (Germany), 16 percent (Switzerland) and 12 percent (France).⁷ Notable gaps in our data are the French quality daily Le Monde, the French tabloid France Soir, the German financial broadsheet Handelsblatt, and the Swiss tabloid Blick. While the sample is not perfectly representative of the newspaper markets, it roughly replicates the newspaper media structures in the countries in question.

While it is certainly desirable to move beyond the usual mapping of communicative spaces, with its underlying belief in the existence of relatively distinct and homogeneous national communicative spaces, and devise a more complex mapping acknowledging also sub-national and transnational public spaces related to other social divisions, it is far less clear how such an aim could be achieved when it comes to empirical measurement of media content (Koenig et al. 2006). Even when we restrict the scope of analysis to class-based divisions, the allocation of individual newspapers to comparable taste publics across countries is not straightforward. The so-called quality papers are usually considered sufficiently similar to be comparable across countries (Peter 2004; Trenz 2004; Wimmel 2004; Beyeler and Kriesi 2005). Usually, it is assumed that ‘the’ elites consume ‘the’ quality papers, but neither the elites nor the quality papers are a homogeneous group. We might want to distinguish between three different types of elites and quality newspapers. The first type comprises well-educated professionals that work in the service sector. These are more or less the people that have been labelled new middle classes (Cohen 1985; Inglehart 1990; Eder 1993). Of these, a small fraction is politically active and will likely consume a partisan daily, while those who are less politically partisan might favour a professionalized broadsheet define professionalized broadsheet. The business elite, however, has quite a distinct type of paper at its disposal: the financial broadsheet.

While the categorization of quality newspapers into professionalized broadsheets, partisan dailies, and financial broadsheets across our sample was fairly clear-cut, the different newspaper market structures made it extremely difficult to establish similarly viable cross-country categories for other types of newspapers. Whereas tabloid newspapers generally cater to lower classes, the readership of regional newspapers – a sizeable category in several countries included in our sample – cuts across several classes. The problem of comparability increases even further if we consider the differences in market penetration: While in Germany almost two-thirds of the population read a daily newspaper, in France less than one-third do so. The population addressed by the press market in Germany is thus considerably broader. In view of the vast differences in newspaper markets and cultures, the only commonality among the non-quality press seems to be exactly that: It is quite distinct from the quality press in terms of readership and format, but serves very different strata in different countries. We therefore decided to follow Sartori's (1970) advice to increase the level of generality and treated all non-quality papers as a heterogeneous group that is distinct from the quality-press in the sense that it is much less concerned with political debate and does not cater primarily to the upper social classes. Table 1 displays our sample sorted by the different types of papers we use to identify taste publics.

Drawing on claims-making analysis (e.g., Koopmans and Statham 1999) and newer methodological developments in frame analysis (Koenig 2004, 2006), our method consisted of four main steps. We started by identifying the main themes and positions in small saturation subsamples for each of the countries. Drawing on discourse analytical methods, we proceeded by identifying keywords that regularly appeared in connection with particular themes and could therefore serve as unambiguous indicators for these themes. For instance, the term ‘citizens’ and its translations were used as an indication for theme of democratic participation and legitimacy (see below). As languages structure meaning and as different national publics draw on different cultural repertoires, often different keywords were utilized in different languages. Between six and 14 keywords per country and position were identified,⁸ and the data were then automatically coded using computer-assisted qualitative data analysis software (QDA Miner). Finally, we validated the thematic model via latent class analysis.

In order to gain insight into the unequal relations of power among different nationally-bounded communicative spaces, we identified the 30 most frequently quoted actors (individual politicians) in each country. We then automatically coded the articles, counting occurrences of actors in each country's press via CAQDAS. This enabled us to compare the various nationally-bounded communicative spaces with respect to the amount of coverage mentioning actors from individual countries, thus identifying the unevenness in direction and intensity of the communication flows across these spaces.

In existing literature on Europeanization, the EU and European identity, two features are repeatedly mentioned as distinctive traits of Europe. One is usually summarized in the slogan ‘unity in diversity’, and refers to the diversity of European (national) cultures united by the attachment to common values and principles (Shore 2000; Sassatelli 2002). The other distinctive trait, which has become a matter of increasing public and academic debate only in the late 1980s (Delanty and Rumford 2005: 108–9), is Europe's peculiar social model that despite the existence of several distinct national variations, makes Europe fundamentally different from the United Stated of America (Hutton 2002; Martin and Ross 2004). Both featured prominently in the newspaper debates we surveyed, and constituted the two main lines of disagreement:

• 1.

  The first line of disagreement arose in response to the question of whether the EUCT, because of strengthening the supranational level dimension of the EU, is good or bad. In most cases, both those who thought it was good (federalist position) as well as those who thought it was bad (isolationist position) started from a (banal) nationalist position (cf. Billig 1995), i.e., they started by considering what is good or bad for the (their) nation-state. The position of those who thought the EUCT is a threat to the nation-state, its sovereignty, or for national identity, was often referred to as ‘nationalist’ – yet it is important to underline that neither of the two positions were outside nationalist discourse.

• 2.

  The second lineof disagreement concerned the question of whether or not the Constitution is a threat to the welfare state, and more broadly whether the welfare state or the ‘European social model’ is something that should be protected on a pan-European constitutional level. Within the welfare state position there were those who argued against ratification on the grounds that the EUCT would legitimize the establishment of a common market while at the same time not providing adequate measures for protection of social rights, while others argued the EUCT was the best way of safeguarding the future of social Europe in the context of globalization. Within the neoliberalist position those in favour of ratification argued that the EUCT represented a welcome liberalisation while others argued that the EUCT did not liberalise at all or did not go far enough in liberalising Europe.

• 3.

  A third theme emerging prominently in the coverage – and, unlike the first two, virtually never contested – was the theme of democratic participation and legitimacy. A similar consensus is visible in scholarly writings on the topic as well; analysts tend to agree that that the EU is suffering from a democracy and legitimacy deficit (Beetham and Lord 1998; Majone 1998).

While we can find each of the five positions in all of the communicative spaces, we can clearly identify also two groups of nationally-bounded communicative spaces sharing considerable similarities with respect to the distribution of particular positions.⁹

While concern with the potentially damaging effects of the EU Constitution on the welfare state is predictably relatively prominent in France and Germany – traditional strongholds of welfare state ideals – it is also high in one of the new member-states: Slovenia. On the other hand, neoliberalism, while unsurprisingly widely present in the UK, is also strong in Switzerland. While its fairly pronounced presence in the Czech Republic and Slovakia fits the usual fears of economic liberalism ‘taking over’ the whole of Central and Eastern Europe (Kelly-Holmes 1998), its comparatively lower occurrence in Slovenia clearly shows that the situation is not as clear-cut as some would maintain.

The distribution of positions with regard to the second major line of disagreement, i.e., the question of whether or not the EU Constitution is potentially harmful to the nation-state, reveals a similar pattern (Figure 2). Again, two groups of fairly similar communicative spaces can be discerned; the first one containing spaces in which federalist positions were at least twice more common than isolationist ones (France, Germany and Slovenia), and the second one including communicative spaces where the amount of federalist and isolationist positions was closely similar (the Czech Republic, Switzerland, Slovakia and the UK). It should be noted, however that these differences are differences in degree. All themes are present in all countries under analysis.

It is interesting to observe that higher levels of neoliberalist positions tend to coincide with higher levels of isolationist arguments, while higher amounts of welfare state positions are likely to go hand-in-hand with higher amounts of federalist positions As the patterns of distribution for both major lines of disagreement are very similar – in both cases, the two groups are the same – two major trans-state ideoscapes can be identified: a federalist/welfare-state ideoscape that is dominant in France, Germany and Slovenia, and an isolationist/neoliberalist ideoscape that is dominant in the UK, Switzerland, the Czech Republic, and Slovakia. Roughly speaking, while the first ideoscape gave more prominence to the understanding of European Union as a more integrated federal entity, possibly with its own foreign policy, army and system of taxation, based on ‘social values’ such as an extensive welfare state and state regulation of the economy, the second one most often discussed Europe as a loose organization of nation-states that share a common, free market for goods and services.

A further point should be noted. While the federalist/welfare-state ideoscape is remarkably strong on federalist positions – in all cases, the amount of federalist positions is at least twice as high as the amount of isolationist positions – welfare state positions are clearly far less dominant. In the isolationist/neoliberalist ideoscape, the debate seems to be open on the federalist/isolationist front, but largely resolved on the welfare state/neoliberalist front – with the exception of the Czech Republic, where the amounts of welfare state and neoliberalist positions are much closer together. While any further discussion of these results would of course need more extensive investigation, it is safe to conclude that as far as the debate on the EU Constitution is concerned, considerably different ideological battles are being waged in these two ideoscapes. In the isolationist/neoliberalist ideoscape, the neoliberalist/welfare state contest is largely over – with the exception of the Czech Republic – while the federalist/isolationist discussion is still open. In the federalist/welfare-state ideoscape, the federalist/isolationist debate seems to be largely resolved, while the contest of neoliberalist and welfare state positions is still ongoing.

We leave to one side explanations concerning why we find different proportions of the two ideoscapes in different geographical locations across Europe. There are obviously particular national and regional histories within Europe that would help to explain the relative narrative fidelity and empirical credibility of national and federalism and of social democracy and liberalism. Perceptions of present lived experience and circumstance, however, also clearly contribute to the strengthening and weakening of ideological currents and are obviously central to explaining contemporary ideological change within European communicative spaces.

Finally, it is important to note that although the newspaper coverage of the EU Constitution shows clear disagreements and sharply different priorities between states, there is one prominent position which is shared by all, and left virtually uncontested: the theme of democratic participation and legitimacy. While the amount of articles highlighting this theme varied considerably across the states (Figure 3), we could hardly find any articles which were explicitly denying its legitimacy. However, clearly the financial papers were relatively unfazed with this issue, as they on average reported less than half as frequent about this point than both quality and other newspapers (14 versus 29 and 32 percent, respectively. This difference brings us to our next point, the difference in reporting across class/habitus boundaries.

Although no systematic data on the habitus of the readership exists (see, however, Bourdieu 1977: 519 for France or Sparks 2003 for the UK), it seems safe to assume that the addressees of these newspapers are not random sections of the population. Our data show that the different taste publics do consume different very different coverages.

If the differences between states are marked, differences within them are also pronounced. In a first step to analyze these differences, we pooled all articles from the so-called ‘quality press’, which is considered to represent the views of the opinion leaders and thereby drives public opinion, in particular when it comes to EU issues (Trenz 2004), into two different categories: one comprises the ‘proper’ quality papers, the other one financial papers, primarily read by the business elites. The remainder of the vastly different papers with undoubtedly very different styles and readerships was lumped into a residual category that is purely defined by its commonality of not catering to political, cultural, or business elites. While this might be considered a fairly crude category, the vastly different structures of the national markets across our sample necessitate such simplifications (Sartori 1973).

The results of this pooling show differences across the three paper types that even exceed the differences between countries. Although there is no standard measure for the assessment of the strength of a relationship between two non-nested models, the BIC criterion, which is based on the Chi-square test and replaces that goodness of fit criterion, allows for a comparison of the goodness of fit between two models (Raftery 1995; Hauser 1996). Based on the BIC differential at –20, intra-state differences are more pronounced than cross-state ones.

As one might expect, the omnibus quality papers with their on average much longer articles contain more themes than other papers. The latter are also more inclusive in the debate than the financial papers, which on the whole present the least complete picture of the debate, as the welfare state position was almost completely ignored, with only 11 percent of all articles taking up this position. The relative strength of the two sets of antagonistic themes (welfare state versus neoliberalism and federalism versus isolationism) remains almost invariant across quality and non-quality papers, which is good news for a coherent public sphere. The argument between isolationists and federalists is, however, much more intensive in the ‘popular’ press than the one between welfare state and neoliberalist positions. This finding seems to resonate with Friedman's (2000) arguments about increasing ideological polarization between classes: while the rising elites are becoming increasingly cosmopolitan, the lower socio-economic classes are becoming more and more nationalist, localist and xenophobic. Last but not least, it is particularly worrying that the debate between proponents of the welfare state and advocates of economic liberalism is less prominent precisely among newspapers catering to social strata that are likely to be most directly affected by the outcome of this debate. Obviously, more research is needed to confirm these trends, yet it is beyond doubt that these differences – revealed only once we make the analysis of European public communication more sophisticated by introducing class – should be at the forefront of research on the European public sphere.

Even more startling differences are revealed if we compare the distribution of positions in financial papers to the distribution in quality papers. This comparison shows that there is a strong divide within the elites themselves: with respect to the relationship between welfare and neoliberalist positions, the reporting strength from roughly two to one in the omnibus papers is reversed to one to four (!) in the finance papers, which hints at a hegemony of neoliberalist ideas among the business elites,. With respect to the isolationism-federalism axis, the differences are not that dramatic, but still a two to one predominance in favor of federalism is canceled in the financial papers.

According to our analysis, the mediated debate on the EUCT reveals not only class-based and ideological divisions described above, but also substantial unevenness in direction and intensity of the communication flows across Europe. Adopting a method of counting occurrences of political actors in each country's press, we were able to outline a structure of communication patterns and detect both ‘the leaders’ and ‘the followers’ of the EUCT debates as well as the transnational hierarchy of influence on newspapers’ agenda within the spectrum of institutionally nationally-bounded communicative spaces which we examined (see Figure 4).

Understandably, we found politicians from other countries than the six ones surveyed among them – with Luxemburg (7.7 percent of articles on average),¹⁰ The Netherlands (4.1 percent), Italy (2.3 percent), Austria (2.2), Poland (1.7 percent) and Denmark (0.9 percent) getting the highest media attention. These are not pictured in the figure.

Looking at the structure of references to European politicians across the surveyed countries, the sample can be clearly divided into two distinct parts, characterised by different levels of symbolic power to shape newspapers’ agenda in other countries. The first group consists of the United Kingdom, France and Germany, which cumulatively dominate in the coverage of EUCT throughout the countries in the second group – Czech Republic, Slovakia and Slovenia – without paying any attention to these new EU member states themselves. Similarly, the media from the second group of states display only little interest in their ‘fellow’ Eastern European countries when reporting about this topic.¹¹ These results broadly coincide with results of other studies, which showed that the UK, Germany and France are most often the subject of reporting in other EU states (see the overview of existing studies in Machill et al. 2006). On the other hand, the position of Slovenia, the Czech Republic and Slovakia, seems to be comparable to that of Finland, Denmark, Sweden, Portugal, Ireland, which are hardly ever mentioned in other EU states at all (ibid: 72).

Not surprisingly, we found out that French politicians were the most quoted of all the foreign political leaders in every country surveyed. The outcome of the French referendum was widely perceived as a decisive one, not only for the future of the EUCT, but for the process of European integration in general. As our study confirmed, the French referendum was profusely reflected upon and to a significant extent shaped the character of the newspapers’ agenda throughout the sample. In the cases of Slovenia and Slovakia it even overshadowed debates on the topic within their own national political scene.¹²

What is more striking in this context is the virtual absence of the Dutch politicians on the European newspapers’ pages. Even though the public debates in The Netherlands before the referendum clearly indicated the possibility of its refusal, and even though the Dutch ‘nee’ formally had exactly the same political consequences for the continuation of the EUCT ratification process as had the French ‘non’ (which happened only three days earlier), the percentage of articles dealing with Dutch political representatives did not exceed five percent in any country, with the exception of Slovakia (6.7 percent). Apart from the Prime Minister Jan Peter Balkenende, the European public was practically uninformed about what the Dutch politicians thought about the EUCT. By contrast, the French political scene was portrayed as much more heterogeneous and the debates were presented in a much more personalised way than those in The Netherlands.

To interpret this unevenness in press coverage of the EUCT across the spectrum of newspapers in our sample, we cannot simply turn to the theories of news values and the omnipresent bias of the media towards the ‘elite nations’, as conceptualised by Galtung and Ruge (1965). Even if we accepted such a label for the UK, France and Germany,¹³ as a metaphor for the power of these states to shape politics on the European level – and it is indeed obvious that they are favoured by the press throughout the sample – we would still be faced with the question of why does the strength of these communication patterns differ quite substantially from country to country. This is perhaps most obvious in the case of relatively high coverage of British politicians in Czech (17.5 percent) and Slovak (22.2 percent) dailies – the highest in the sample – especially when compared to the amount of attention devoted by them to Germany as a geographically more proximate neighbour (the difference is particularly striking in the case of Slovakia). The Slovenian press, on the other hand, shows quite the opposite figures for British (11.3 percent) and German (17.7 percent) politicians’ occurrences. The same pattern can be detected in the French press, mentioning political leaders from Germany almost twice as often as those from the UK.

In an attempt to explain this, we look at the clusters of ideological positions within the EUCT debates that were outlined in the previous section of the paper. The frequency analysis of political leaders quoted in each country seem to support the above-stated claim about the division of the sample into two parts; one tending towards a more neoliberalist view of Europe and displaying smaller difference between nationalistic and federalist positions in context of the EUCT (with the UK media in the lead, followed by the Czech Republic and Slovakia), while the other part, consisting of France, Germany and Slovenia, emphasises rather the welfare-state approach and unambiguously favours federalist viewpoints on the future of the European integration process. The data allow us to presume that these shared ideological positions correspond to the tendency of the press to pay more attention to representatives of countries whose political orientations are in line with their own. Table 3, in which the averages of frame presence for the different paper types are presented (which hence correct for the different intensity of the debates across countries), summarizes these results.

In other words, the communication network consisting of the seven countries in our sample reflects differences in symbolic power of particular actors within this network just as much as the political and ideological alliances between them. However, the answer to the question of whether the established communication patterns foster the creation of ideoscapes, or it is instead the pre-existing ideological bias of the press which influences its choices to cover some countries more than other, lies beyond the explanative scope of this paper.

Despite embracing the rhetoric of transnational flows and networks, comparative research on media content continues to fall prey to methodological nationalism. When it comes to empirical measurement, researchers often, despite their best intentions, fall back on techniques that assume that the discourses circulating within particular nationally bounded communicative spaces are homogenous. In this article, we developed a set of propositions and analytical approaches that should help to overcome this impasse, and used them to examine the newspaper debates on the EUCT in seven European states. We started by acknowledging the precarious and unfinished nature of national communicative spaces. We suggested that instead of focusing solely on comparisons between nationally bounded communicative spheres, we should also look at differences between class-related communicative spaces. By adopting such an approach, we can acknowledge both sub-national segmentations of communicative spaces and transnational linkages, while at the same time not losing sight of the importance of the national.

Applied to the examination of the newspaper debates on the EUCT in seven European states, this approach proved illuminating. Our results confirm the existence of significant ideological divisions in the conceptions of Europe. In our sample, two major trans-state ideoscapes can be identified in Europe: a federalist/welfare-state ideoscape dominant in France, Germany and Slovenia, and an isolationist/neoliberalist ideoscape dominant in the UK, Switzerland, the Czech Republic, and Slovakia. We are confident that the two ideoscapes that we have outlined are present beyond the countries included in our sample, though more research on a bigger sample is needed to ascertain the exact patterns of these ideoscapes in the rest of Europe. Also, longitudinal research would be desirable to chart the ideological flux within these spaces, and the shifting balances between the two ideoscapes over time.

The measurements of the direction and intensity of the communication flows across Europe were also instructive. The structure of communication flows was markedly unequal and one-way: while the newspapers in the Czech Republic, Slovakia and Slovenia regularly referred to actors from the United Kingdom, France and Germany, the reverse was hardly ever the case. Furthermore, the newspapers from the three Eastern European states had rather different preferences with regard to the actors from Western European states. The patter of these ideological affinities coincided with the two ideoscapes identified earlier: while Slovak and Czech newspapers usually quoted actors from the UK, Slovenian newspapers preferred those from Germany.

Lastly and perhaps most importantly, the comparisons between class-related communicative spaces revealed important differences. These were most manifest when comparing financial newspapers with other newspapers, as the former were ideologically far more one-sided, almost completely ignoring the welfare state position. The comparison between quality and popular newspaper proved revealing as well, suggesting that newspapers addressed at lower social strata were less likely to report on the debate between the proponents of neoliberalism and the advocates of the welfare state – a particularly worrying result given that the readers of these newspapers are most dependent on a strong welfare state. Obviously, more research is needed to confirm these trends, yet it is beyond doubt that these differences – revealed only once we supplement the analysis of European public communication by introducing class – should be at the forefront of research on the European public sphere. This clearly supports our initial contention that the research on European mass communication ought to move beyond comparisons between national units and the levels of their respective Europeanization, and examine how European issues are conveyed in media catering to different social classes.

Sabina Mihelj is a Lecturer in Media, Communication and Culture at Loughborough University. Her research interests include issues of identity and communication, with particular reference to nationalism and religion in multinational settings and border regions.

Thomas Koenig is a Research Associate at the Department of Social Sciences, Loughborough University. His areas of specialism include social movements, quantitative methodology and textual analysis.

John Downey is a Senior Lecturer in Media and Communication at the Department of Social Sciences, Loughborough University. His areas of interest include critical theory, cinema, information and communication technologies and European media.

Václav Štetka is a Lecturer in Media Studies at Masaryk University. His research interests include media representations of cultural identities, particularly in relation to the nation state and the nationalism discourse, and the role of media in globalisation processes.