CAIRN-INT.INFO : International Edition

1The 2012 French presidential election was held in a context in which European questions were particularly visible: the euro crisis, sovereign debt crisis, member states in economic recession, social unrest over austerity policies in Spain and Italy as well as Slovenia, Hungary or Romania, and last but not least Greece shaken by a social and political crisis with vigorous demonstrations against economic austerity plans. European leaders were also prominent during the French presidential campaign, viz Angela Merkel’s support of Nicolas Sarkozy before the latter had even declared himself a candidate; [1] the affirmation of the entente cordiale, linking the French president with the British prime minister, David Cameron, who “wished his friend good luck”; [2] or the stance taken by Sigmar Gabriel and Massimo d’Alema who came to champion François Hollande and support his proposal to renegotiate the budgetary treaty at the meeting of the unified European left at the Cirque d’hiver. [3] Finally, many themes tackled during the campaign, relating to finance, economics or social issues (sovereign debt, unemployment, immigration, for example) indirectly had a European dimension. [4] In other words this electoral campaign took place in a context that could be described as one of pervasive Europeanisation. But was Europe a campaign issue? Was it present in the candidates’ speeches, and if so what place did it occupy? Did it have an impact on how French people voted? After a brief overview of the recent literature which, over the last fifteen years, has focused on the European issue in the political space of member states, and in France in particular, we attempt to answer these questions by taking a threefold empirical approach. We begin with an analysis of the campaign manifestos [5] of all first-round candidates, then present the analysis of a certain number of campaign rally speeches [6] by each candidate, and finally present an analysis of the results of the post-electoral survey conducted by Cevipof (Centre de recherches politiques de Sciences Po Paris), incorporating questions relating to European integration. Combining these approaches allows us to tackle the question of the importance of the European issue (or its lack of importance), both in terms of the electoral offer and in voters’ choices. This double perspective is particularly relevant in a context where a certain number of recent studies have suggested an increased distance between the elected and the electors on European questions. [7] This article also provides fuel for current controversies regarding the reconfiguration of the dimensions of national political space within European states.

Literature review: the European issue in political space

2In 2000, a special issue of West European Politics was published under the title “Europeanised politics? European integration and national political systems”. It highlighted the importance within European studies of questions relating to the impact of European integration on national political systems since the end of the 1990s. [8] A whole literature sprang up in its wake, mainly in response to two key questions. The first relates to the ways parties integrate the European issue into their positions and strategies, and to the emergence or not of a European cleavage. The second question focuses on electors and looks at how they take European questions into account when they vote – in other words whether “EU issue voting” [9] exists, potentially leading to the development of a new dimension in national political space. These questions received many – often contradictory – responses. At the party level, many authors considered that the impact of European integration was “limited” [10] and stressed that European integration is an issue that is absorbed by traditional party structures. [11] As a result, European integration does not appear to be the source of a new cleavage within the partisan space in different member countries. [12] However, adopting a broader understanding of the notion of political cleavage, other authors have emphasised that European integration is a new source of structural conflicts between political forces that add to or even replace old ones. [13]

3At the level of voters, the existence (or not) of “EU issue voting” is also a subject of debate. Some studies emphasise the very low levels of knowledge individuals have on European issues, the low levels of importance assumed by European questions in national debates (even during European elections [14]), and the indifference of citizens to European issues. [15] Conversely, other studies from the end of the 1990s onwards have shown that the European issue did indeed play a role in certain national elections. Thus, in the United Kingdom, Geoffrey Evans has shown that European questions cost the Conservative party votes during the 1997 legislative elections. [16] Based on a study of Austria, Sweden and Finland just before and just after their membership of the EU, Erik Tillman concluded that attitudes towards the EU play a significant role in electoral behaviour at the national level when the European issue is visible in the public space. [17] More recently, Harald Schoen has shown that in Germany those who were favourable to Turkey’s entry into the EU were more likely to vote for the SPD or the Greens, and less likely to vote for the CDU/CSU or the FDP during the 2005 federal elections. [18] As early as 2000, Matthew Gabel was the first to put forward an analysis of the role of the European issue within the vote across all member states. His study, which tests the influence of support for the EU on the probability of voting for one party rather than another, concludes that this influence does exist in most European countries. [19] However, recent studies emphasise that the impact of attitudes towards Europe on the national vote varies not only according to the level of political sophistication of citizens but also according to the national context, [20] the salience of the European issue at the time of the election, [21] and the party offer on this issue. [22]

4Analysis of the French context is similarly divided. Some consider that Europe only has a weak influence on the political system, [23] describing it as “politics as usual”. The partisan schisms that occurred both on the left and the right following the ratification of the Maastricht Treaty by referendum have not fundamentally altered the French “hyper-alternation”; [24] any more than their reactivation in 2005 changed the bipartisan nature of the party space. [25] Others, however, consider that Europe has indeed been at the root of some blurring of party boundaries. [26] Illustrative of this scholarly confrontation is the battle of interpretation around the 2005 referendum and the commentary concerning it. In the 2005 “no” vote, some see a “French no”: the expression of an “irresistible nationalisation of the European vote”. [27] Others consider that voters did indeed respond to a European question, [28] but in which “a significant proportion of voters [linked] the national sources of social insecurity directly to the issue of the referendum and more broadly to European construction”. [29] In light of this controversy we can see how far we have come from the first studies conducted on citizen acceptance and rejection of European integration until today. At the beginning of the 1990s the dominant idea was that European issues were situated on the left-right axis. [30] Ten years later, new analyses identified “a new cleavage, which cut across the traditional left-right cleavage”, opposing “universalist” and “anti-universalist” [31] attitudes and shedding light on the enduring difference of the Front National. [32] Alternatively, they identified an open-closed axis articulated around values relating to the perception of the Other (immigration, Europe, globalisation), independent of the left-right axis. [33] The last generation of studies has broken with these alternatives and shows that attitudes towards Europe in fact combine two dimensions: the first is specific to Europe and independent of the left-right axis, and the second is based on the social fears that are indexed to the left-right axis. [34]

5In order to pursue this line of thought we propose to explore three closely linked research questions. Does the pervasive Europeanisation of an electoral context contribute to the Europeanisation of an electoral campaign? To what extent does this Europeanisation proceed via the “politicisation” [35] of the discourse on European integration? Finally, to what extent does this Europeanisation of discourse resonate with the opinions of citizens to the point of influencing their vote?

A threefold data set

6We propose to respond to these questions through an analysis of three sources of data: an analysis of the candidate manifestos (professions de foi) for the 2012 presidential election, the analysis of a certain number of campaign speeches by these candidates, and the analysis of quantitative data from post-electoral surveys conducted by the Cevipof the day after the second round of the presidential election. Professions de foi and campaign rally speeches are of particular interest compared to other channels of communication used by candidates because they involve direct communication (public meetings, rallies, leaflets, books) which bring the voter into “direct contact with the candidate’s message as they intended it”. [36] Indirect communication (news articles, reports etc.) do not do this, instead they pass through intermediaries and escape the candidate’s control (and that of his/her campaign staff). We have adopted a dual methodology in order to understand and interpret these texts and discourses beyond an intuitive reading: a textometric [37] process combined with qualitative analysis. Our corpus of campaign manifestos and speeches was run through the Lexico3 programme. [38] This first reading was combined with a second approach: the professions de foi were analysed using a content analysis programme (Nvivo) and the rally speeches were subject to stylistic analysis enabling us to consider political speech as a sub-genre of public speaking. Here, we particularly explored rhetoric – the art of convincing and persuading – by focusing on the discursive techniques of different candidates. [39] One of the advantages of using this discursive material is that, when analysing the research data, we retain the different messages relating to the EU to which the voters have been exposed.

7The data on voters is taken from the 2012 Cevipof post-election survey which allows us to connect the vote in the first and second rounds of the presidential election to attitudes towards Europe, measured in some depth.We offer a two-stage analysis. First, we develop a measurement model, inspired by factorial analysis, to identify the dimensions of attitudes towards Europe (via the indicators relating to it in the survey) and to verify that what we are observing is indeed the bi-dimensionality of these attitudes, as established in our previous research. [40] Second, we use a model derived from regression analysis in order to analyse the correlations between what we want to explain – the vote – and a range of explanatory variables, including our two axes relating to European integration. Thus we aim to empirically establish whether attitudes towards European integration play a role in the vote, once we have controlled for the impact of other important explanatory factors.

Professions de foi: national re-affirmation and European polarisation

8Exploring the ways in which European integration is mobilised or not in the candidates’ campaign manifestos provides a preliminary terrain for analysis. These texts are highly polished and are examples of a particular style: their content is the subject of extremely rigid norms, from which deviations are rare. The corpus we constituted for this first analysis includes the campaign manifestos of the ten candidates in the first round of the presidential election of 2012. In order to highlight the specificities of this election, we undertook a certain number of analyses of the twelve candidate manifestos from the 2007 presidential election as a point of comparison.

9The 2012 French presidential election occurred in an economic context charged with permanent references to Europe. European events were less pervasive in 2007. We would therefore expect that Europe would be a more salient theme in the campaign manifestos of 2012 than it was in 2007. If we take the number of times Europe is mentioned as a measure of its salience, we can see that this is not the case. Whether we count the references to the group of forms beginning with europ*[41]or the eighteen French terms [42] that unambiguously refer to the European integration process, they appear more frequently in 2007 than in 2012. [43]

10This masks significant divergences between candidates however. In the 2012 campaign, the a-temporal and a-contextual discourse of Nathalie Arthaud (who refers to neither France nor Europe) is matched by the discourse of Nicolas Dupont-Aignan, in which Europe appears as a structural issue (see Figure 1 below). Three other candidates are distinguished by the way they refer (or not) to Europe: Jean-Luc Mélenchon who refers to the europ* group of forms in his profession de foi more often than the average number of times across all candidates’ texts; and, conversely, François Hollande and Marine Le Pen who refer to it less than the average. Since theMaastricht Treaty, the two main French parties have avoided taking positions on Europe that are too definite. It is therefore not surprising that François Hollande, the main challenger for the presidency, stood out with a quantitatively weak “European” discourse. The low number of references to Europe in the discourse of Marine Le Pen – which is nevertheless in line with that of her father in 2007 – is more remarkable. The recurrent references to France (and its associated forms within fran*[44]) in her discourse leave us wondering if the reference to Europe isn’t in fact implicit. From this perspective it is important to understand how references to Europe work within the candidates’ arguments.

Figure 1

Occurrences of the forms fran*, europ* and mond* (world) in candidate campaign manifestos for the first round of the 2012 presidential election (out of 10,000 forms)

Figure 1

Occurrences of the forms fran*, europ* and mond* (world) in candidate campaign manifestos for the first round of the 2012 presidential election (out of 10,000 forms)

11Analysis of the most commonly used words in the candidates’ manifestos (Table 1 below) reveals that the three most commonly cited in 2012 were (in order) “France”, “Europe” and “French”. In 2007 the top three comprised “political”, “France” and “years”; “Europe” came in twelfth place and “French” fifth (see Appendix 1). Reference to Europe was thus more central in the campaign manifestos of the candidates overall in 2012 than it was in 2007, but as part of a discourse in which references to France and the French were particularly important. This national focus is to be expected given that this is the presidential election, but in 2012 it appears to be particularly strong and more widespread than in 2007. Although references to France were avoided by the Lutte Ouvrière and NPA candidates, both in 2012 and 2007, it takes pride of place in the campaign manifestos of the other candidates, which was not the case for the Greens and left-wing candidates in 2007, with the exception of Ségolène Royal. How should we understand this “nationalisation” of discourse both on the left (particularly in Jean-Luc Mélenchon’s profession de foi), and on the right (references to France being particularly overwhelming in that of Nicolas Sarkozy’s in the first round)?

Table 1

The 12 words most frequently used per candidate in 2012

Table 1
Overall Arthaud Bayrou Cheminade Dupont-Aignan Hollande Joly Le Pen Mélenchon Poutou Sarkozy France Workers France World France Change Years French France Billions France Europe State Life Europe Years France Rights France Europe Politics Country French Economy Work France Europe Justice France Immigration Rights Population Businesses Years Society Law Credit Prices Republic Equality Country Independence Crisis Values Politics Businesses Banks French Businesses Republic Economy Country Euros Europe World Capitalists Employees Work Banks Future Politics Employment French Measurements French Country Classes Businesses Justice Politics April Power Jobs Front Banks Employment Businesses Struggle French Projects Loans Housing Europe People Sarkozy Debt Nicolas Work Shareholders Country System Referendum Service Corruption Politics Years Energy Sarkozy Life Crisis Politics Bank Brussels Europe Jobs Power Prohibition Peoples Jobs Rights Interests Recovery Development Small Businesses Purchasing Justice Service World Service Taxes Banks Part Europe Markets Freeways Benefits Future System People Society World

The 12 words most frequently used per candidate in 2012

12Content analysis enables us to demonstrate that a discourse on European integration, and globalization more broadly, is developed between the lines of the persistent references to France and the French people. We can see this, for example, when Nicolas Sarkozy stated that “France is a great country” but then went on, “it is able to take advantage of globalisation; in many sectors our country is very successful”. What matters, then, is that the candidates reaffirm the ability of national actors to respond to problems that transcend the national frame. From Jean-Luc Mélenchon (“A vote for Jean-Luc Mélenchon is … a message for finance: ‘French people will not allow themselves to be fleeced’”) to Marine Le Pen (“I will organise the re-industrialisation of France to create jobs, by reinstating differentiated and adapted border protection”), via François Bayrou (“I will mobilise consumers around a ‘made in France’ label”) and François Hollande (“I will take steps to produce goods in France […] I will allocate public subsidies and tax breaks to businesses that invest and recruit in France and which launch export offensives”), the objective was the same: to emphasise the capacity of French state actors to get their voices across and impose their solutions in a Europeanised and globalised context. These discourses were not fundamentally new but in the 2012 presidential elections they took on a particular magnitude, both in their frequency and the extent to which they crossed partisan divisions, from Jean-Luc Mélenchon to Marine Le Pen via the two main presidential candidates.

13There is little difference between the candidate manifestos in their affirmation of the capacity of the government and the French state to make France’s voice heard at the European level and impose her decisions. Yet the solutions the candidates propose to resolve the problems identified within a European and globalised framework are clearly not the same. Thus Marine Le Pen does not need to reiterate the reference to Europe because it is designated – through the market, major industrial groups, immigration, “globalism” and “communitarianism” – as being the source of all France’s woes:


“I will give French people back their legitimate rights and recover our national sovereignty by overturning the European treaties. Laws will once again be voted in Paris and not in Brussels.”

15Le Pen’s rhetoric is the same as her father’s, with the markets and the industrial groups added to the list of those responsible for France’s “deteriorated” situation. Jean-Luc Mélenchon’s text also reflects a closeness to Marie-George Buffet’s manifesto in 2007, with slogans such as “change Europe”, “no to liberal Europe” and “Europe is ours too!”

16Certain themes around the EU and the integration process appear well-oiled and revolve around the same statements from election to election, particularly for the candidates at both extremes. Others however, appear to be specific to this presidential election, proof of a certain radicalisation, or at least a hardening of the candidates at the “centre” of the political chessboard. For example, whereas François Bayrou declared in 2007 that he wanted to “reconcile […] France with Europe and Europe with France” particularly by making Europe more “legible” for citizens – thereby implying that the result of the 2005 referendum was due to a problem of communication and comprehension – in 2012 he suggested “recreating Europe”, thus establishing a certain distance from Europe as it exists today.

17Above all, the 2012 presidential election saw the two main candidates distance themselves from one another on European issues. The referendums for the Maastricht Treaty in 1992 and the Constitutional Treaty in 2005 had demonstrated the divisive potential of European questions for the two major parties, the PS and the RPR (which became the UMP). It is therefore not surprising that, in 2002 as in 2007, reference to Europe appeared in the manifestos of the two main candidates as a necessary rite of passage, albeit one that was not very ideologically charged. The terms used by Jacques Chirac in 2002 to designate Europe were thus particularly vague: “France has given itself a new horizon: Europe. Europe stimulates us. Europe protects us. It is an opportunity for our country.” Lionel Jospin’s formulation was no less vague, calling for a “European federation of nation states”. In 2007, positions appeared to be no clearer cut and were formulated in almost the same terms. Nicolas Sarkozy thus promised to act “so that Europe protects us within a global world […] particularly against offshoring”, and by re-establishing “Europe’s right to protect its products and businesses”. Ségolène Royal stated her desire to “build a Europe that protects from offshoring” and that is “at the service of employment and growth, and not just money”.

18In 2012, however, the positions of Nicolas Sarkozy and François Hollande, as set out in their campaign manifestos, were very different both in content and in style. François Hollande’s approach appears very much to take the offensive. First he condemns the European leaders who “have resigned themselves to austerity and […] have been incapable of controlling finance.” He then puts himself forward as a leader capable of creating a new European order:


“I will change the direction of Europe: renegotiate the austerity treaty towards growth and employment; launch major projects for the future; develop business policy against financial, social and environmental dumping.”

20As the outgoing head of state, Nicolas Sarkozy’s position is initially more defensive, because he is justifying the policy decisions taken during his five-year term: “the implosion of the euro, which would have been a catastrophe for every one of you, has been avoided”, “a strong France is a France which carries weight in a protective Europe”. At the same time, he takes a more critical stance towards Europe:


“If Europe does not regain control of its borders […] France will suspend her participation in the Schengen zone […]. If […] Europe does not reserve its public contracts for businesses that produce in Europe, France will do so for her own markets[…]”

22He thus adopts a national discourse that calls for a certain distance from Europe, whereas his adversary proposes a national discourse as a model for Europe, closer to the traditional discourse of the French elites who assigned France a leading role in the EU. Moreover, these discourses tended to become more radical on European questions between the first and the second rounds of the election. François Hollande declared in his manifesto for the second round, “I will renegotiate the European treaty to include major new projects for the future and protect us from unfair competition within a global world”. Nicolas Sarkozy reiterated his threats of withdrawing France from certain policies and concluded by clearly identifying Europe as responsible for some of France’s woes:


“Choose controlled immigration and respect for our identity, do not choose a Europe open to all winds.”

24The analysis of these campaign manifestos allows us to observe that although references to Europe were not more numerically frequent in the 2012 campaign than in previous elections, the context of the economic crisis and the tension it created between member states constituted a particularly strong framework for this election. The particularly insistent references made by the majority of candidates to France and the French people should be understood within this context – not only on the right, where Nicolas Sarkozy’s discourse appears even more nationalist than that of Marine Le Pen, but also on the left, from Jean-Luc Mélenchon to Eva Joly. It was important to underline the ability of leaders at the national and state level to confront a global problem, if not alone then at least by taking their own decisions. Moreover, from the point of view of European questions, this presidential election was different from previous ones in that the PS and UMP candidates both set out a real difference in the future role of France within the integration process. For François Hollande, the French president had to “change the direction of Europe”, whereas for Nicolas Sarkozy he must hold France back from Europe if the latter adopted measures which France did not support. Thus, although in former presidential elections the two main candidates had shared similar positions on Europe, in the 2012 elections their positions diverged significantly. This polarisation by the leaders of the two main political parties testifies to the politicisation of European questions at national level.

Campaign rally speeches: denunciation, nationalisation, politicisation

25To identify the relative place of Europe in the speeches the candidates gave at rallies during the pre-campaign and campaign for the 2012 presidential elections, and to compare these speeches with the manifestos, we constructed a corpus of speeches – “Presidential Speeches 1st Round 2012” [45] – which is subject to certain methodological choices themselves governed by technical constraints. We selected the most general and most emblematic speeches delivered by the candidates; the most general because we wanted to record the “heart” of the candidates’ arguments rather than an ad hoc speech written for a well-informed public, [46] and the most emblematic so that we could reconstruct the rhythm of the campaign. To achieve this, and to allow us to investigate potential changes in the arguments, for each candidate we selected the speeches given at the first major meeting (M1), [47] the last major meeting (M3), and the speech considered as the “most important” (M2) of the campaign (for the first round). [48] Although we have these three speeches for François Bayrou, François Hollande, and Nicolas Sarkozy (thanks to their party websites) the speeches of the other candidates are only partially accessible in electronic or written form. The current composition of our corpus “Presidential Speeches 1st Round 2012” thus reflects these constraints (see Appendix 2).

26The study of the prominence of European questions via the tally of occurrences of the europ*[49] group of forms in the fifteen speeches in the corpus produces suggestive and partially counterintuitive results (see Figure 2). The candidate with the fewest references to europ* was Eva Joly, whereas Nicolas Dupont-Aignan had the most. Nicolas Sarkozy made many more references than François Bayrou, and also than François Hollande and Jean-Luc Mélenchon. Although for several decades the European issue has been considered a “fundamental” aspect of centrist and environmentalist references, it wasn’t mobilised by the candidates from these two political movements in this election. [50] As for Marine Le Pen, she made reference only sparingly to Europe and, unexpectedly, did not mobilise an (anti)-European discourse, in either her manifesto or in her speeches.

Figure 2

Frequency of fran*, europ* and mond* (world) group of forms in campaign speeches by candidate (frequency out of 10,000)

Figure 2

Frequency of fran*, europ* and mond* (world) group of forms in campaign speeches by candidate (frequency out of 10,000)

27We then individualised the different speeches in the corpus in order to see whether the candidates had talked about Europe throughout their campaigns or whether it was only evoked at one particular time (see Figure 3). It is important to remember that, in the corpus as it currently stands, we are comparing the three speeches (M1-M2-M3) by François Bayrou, François Hollande, and Nicolas Sarkozy.

Figure 3

Evolution of the europ* group of forms in the campaign speeches of Bayrou, Hollande, and Sarkozy (frequency out of 10,000)

Figure 3

Evolution of the europ* group of forms in the campaign speeches of Bayrou, Hollande, and Sarkozy (frequency out of 10,000)

28We can see that the three candidates adopt different strategies, but that all converge towards fewer references to Europe at the end of the campaign. As in the campaign manifestos, it is clear that the pervasive Europeanisation of the electoral context does not necessarily contribute to the Europeanisation of the electoral campaign. François Bayrou progressively emptied his speeches of references to Europe over the course of the campaign, François Hollande opened his series of campaign meetings by talking a little about Europe, then references stabilised at a much lower level, whereas Nicolas Sarkozy made it a recurring theme in his Villepinte speech (M2), but invisible in Marseille (M1) and discreet in Nice (M3). This came as a surprise because he transformed what should have been a speech unveiling his programme into a protectionist speech, thus consolidating his campaign’s shift to the right by borrowing themes from the FN (immigration, economic protectionism) disguised as “European issues”. His demand for reform of the governance of the Schengen area was an excuse to describe Europe as a “sieve”, made up of technocrats and tribunals, and the proposal to create a Buy European Act was the opportunity for him to state “Free-trade, yes. Unfair competition, no!”

29Analysing the position of Europe in these speeches means observing its connection with other spaces of reference (“France”, “the world”), and we have already seen the position of these nouns in the campaign manifestos. We consider both Europe’s relative position in the speeches (Figure 2) as well as the way it is used in relation to the textual specificities of each candidate (Table 2).

Table 2

The 12 most frequent words in the “major speeches” (M2)* in 2012

Table 2
Bayrou (Paris Zénith) Dupont-Aignan (Paris) Hollande (Vincennes) Joly (Paris Cirque d’hiver) Le Pen (Toulouse) Mélenchon (Toulouse) Sarkozy (Villepinte) Country France France Politics France France France France French Republic Campaign People Freedom Europe Republic Europe French Planet Left Republic French Hope Country President Environmentalism President World World French World Country Courage French Rights Republic People Power History Candidacy Republic Accountability Life President Nation Victory World State People President Choice Freedom Left Things Freedom Sovereignty Work Ideas People Change Environmentalists Friends Years Country Years Republic Europe Left Markets Country People Law School Greatness Hope Order Nation Years Europe Change Majority Country Law Life Crisis

The 12 most frequent words in the “major speeches” (M2)* in 2012

*: speeches not available in written and/or video form for Arthaud, Cheminade and Poutou.

30From one candidate to another, the connection between these spaces changes, but for all candidates (except Eva Joly), France comes before Europe and the world (based on the groups of forms “fran*”, “europ*” and “mond*”, see Figure 2). As was the case in the campaign manifestos, “France” is the most commonly cited word in the meeting speeches, often closely followed by “French” (see Table 2). Europe is absent here, however, or comes at the bottom of the table, except in the speeches of Nicolas Sarkozy (2nd place) and Nicolas Dupont-Aignon (3rd place). The “nationalisation” of the campaign occurred to the detriment of explicit references to Europe, which, from being a compulsory reference in the manifestos, became secondary in the rally speeches. Moreover, three candidates (Jean-Luc Mélenchon, Eva Joly and Marine Le Pen) use the mond* group of forms more than that of europ*, which is an invitation for a more in-depth reading of these speeches on the basis of stylistic analysis.

31In our corpus it is Nicolas Dupont-Aignan who talks the most about Europe. His speech, delivered during a meeting at the Gymnase Theatre in Paris, 22 January 2012, was based around a single theme (“the nation has lost its freedom”). It was made up of two parts: France must be liberated from the “abandonment of sovereignty”, and from “internal feudalism”. He made extensive use of two stylistic devices in the first part of his speech, which was entirely devoted to the European issue: anaphora and synecdoche. [51] In the first case, repetition leads a cumulative denunciation:


“A Europe like a sieve, a passage between two oceans. A Europe of unemployment. A Europe accelerating the crisis. A straitjacketed Europe. A Europe of ruin, debt and bankruptcy.”

33The second case allows for the iteration of scapegoats: “Brussels”, “non-elected technocrats”, “European Commission”, “Constitutional Treaty” etc. In both cases the use of these stylistic devices establishes a syntactical and semantic saturation in relation to Europe and “its chains which bind France”. Here, talking (a lot) about the European Union, is a way of (more effectively) condemning it.

34Eva Joly’s speech at the Cirque d’hiver in Paris on 18 April 2012, on the other hand, made almost no mention of the European Union. The two identifiable occurrences refer to Europe as a common territory embedded in a global space:


“We are at home [in France], we Bretons, Corsicans, Occitans, we Polaks, Eyeties, and Dagos, we Yids, coons, and darkies, we menopausal Norwegians, we Europe, we the world, we the planet. Because we are the freedom to love, we are equality before the law, and fraternity in the Republic.”

36The final anaphora, “vote for the planet” (eight occurrences) does not mention Europe and is focused on environmental issues (pesticides, local consumption, nuclear issues). Here, not talking (at all) about the European Union is a way of negotiating the (differing) sensibilities within a party that is now divided on the question.

37Jean-Luc Mélenchon refers to Europe in his speech in two ways. Like Eva Joly he demonstrates concern for the “destiny of human civilisation” and asks “what can embody the general interests of humanity?” (M1). In this case the political project extends beyond borders, and Europe has a simple transitive function, as a link in a chain which stretches from the individual to humanity. Calling for a “citizen revolution”, Mélenchon uses Europe metonymically. His first speech at the Place Stalingrad was dedicated to “our Greek brothers and sisters and their obstinate struggle”. Here the Greek people stand for both “the whole of the [French] people” and all oppressed peoples. In the same way, institutional Europe, as a symbol of “limitless capitulation before the banks” (M1) and the abandonment “of freedom and sovereignty” (M2), is a condensed version of national compromises and global selfishness. The second use of Europe is more partisan and clearly iterative. Condemning “austerity Europe” after the European Summit in Brussels on 8 December 2011, Mélenchon politicised European questions in order to rekindle the idea of the left as being capable of capitalising on the victory of the “no” vote in 2005 to set sail for a presidential win in 2012. Here, talking (a lot) about globalisation (and secondarily about Europe) is intended to allow him (as far as possible) to forge an extreme-left coalition, and even reach beyond it.

38This is precisely the trap that François Hollande tried to avoid throughout the campaign: becoming a prisoner to those on his left on European issues. This probably helps explain the three characteristics of his speeches studied here: the lack of any occurrences relating to Europe in the major oratorical moments of his speeches, and the weak presence of Europe in his speeches generally, but also Europe’s contribution to the “imperative for correction” [52] attributed to the left. François Hollande makes extensive use of anaphora; the now notorious “I, President of the Republic…” during the televised debate on 2 May 2012 was just the final touch. Yet the vision of the presidential role that is conveyed through anaphora does not seem to have a strong European perspective. Not overly in evidence from the outset, Europe tended to fade further from the picture as the campaign progressed (see Figure 4). It is as though François Hollande, in bringing his supporters together, sought to avoid difficult subjects that were potentially divisive via a “politics of muffling” [53] which aimed to minimise the electoral costs of an issue seen as a source of intra-party division. In this context the EU could be viewed as the object of a “narrative ellipsis”, which involved remaining silent about a certain period of time and not discussing its events. [54] Yet at the same time, talking about Europe was possible as long as it was included in the left-wing’s narrative of struggle for social reform. After diagnosing Europe as “incapable of protecting its currency from speculation” and increasingly an object of “defiance”, came the announcement of the overall project of “a return to justice”: “France must find the ambition to change the direction of Europe […]: a movement towards a Europe of growth, towards a Europe of solidarity, towards a Europe of protection” (M1-Le Bourget), “Everyone knows that if I become president of the republic tomorrow, there will be a renegotiation of the treaty” [55] (M2-Vincennes). We might wonder whether the socialist candidate’s energetic and ostentatious “no” to the future treaty was not there to compensate for (or balance out) his “yes” to the Constitutional Treaty in 2005, when he was first secretary to the PS, as a reconciliatory gesture to bring together a divided left around the theme of “a social Europe”, promising support for the “deglobalisation” and “European protectionism” proposals put forward by Arnaud Montebourg and others on the left of the party. Here, talking (as little as possible) about the European Union doubles as a (selective) politicisation of the issue.

39François Bayrou dedicated the second quarter of his speech in Pau to Europe to highlight two elements. The EU, which was constructed by the founding fathers as “a desire, a harmony and a power” [56] is now, according to him, a space without solidarity and an “oligarchy”. Whilst declaring himself “profoundly European” the centrist candidate drew three conclusions from this: Europe must be used as the megaphone for France; it must be reformed; and, above all, France’s recovery must come from within. He declared:


“Hence we believe that these problems do not lie elsewhere, they are here, and it is here that we have to deal with them, and that is dependent on the mobilisation of French people like you and me.”

41This threefold framing of Europe as a projection of France, an idea which has been led astray, and a “secondary” priority can also be found in Bayrou’s speeches at the Zenith in Paris and in Lyon: the three speeches are coherent and tally with with the campaign manifesto. Although a centrist candidate, Bayrou speaks relatively little about Europe, and when he does he employs a rhetoric of denunciation and (re)nationalisation. One final quotation illustrates the ambiguous relationship that Bayrou constructed with Europe throughout his campaign:


“I want to speak to you about Europe, our Europe. I love nothing more than sovereignty. I love it in my personal life: I don’t like to be dictated to. I love it in my life as a citizen: I expect to play my part in the sovereign body. I love it passionately for my country, which I want to remain independent whenever possible. But I love real sovereignty, not apparent sovereignty which is in fact a lack of sovereignty, a dependency. That is why I am European.”

43Beyond the asymmetry between the number of occurrences of “Europe/sovereignty” what is striking is the oxymoronic [57] construction of the phrase which enables him to declare himself European in the name of his love of sovereignty… It reflects the very “national” tone of Bayrou’s 2012 campaign, dominated by the mantra “produced in France”. Here, talking (a little) about Europe, enables him to talk (a lot) about “France, the country”.


“Tonight France is back in Europe […] I call on [our European partners] to not remain deaf to the anger of the peoples who see the European Union not as a source of protection but as the Trojan Horse of all the threats that carry within them the transformations of the world.” [58]

45This speech, delivered by Nicolas Sarkozy on the night of his election in the 2nd round of the 2007 presidential election, inaugurated a five-year term which was marked by the contradictions of proactive Europeanism and the temptation to extend sovereignty. Another speech, delivered in Toulon on 1 December 2011 concluded along ostensibly more Gaullist lines:


“Re-building Europe is not a march towards more supra-nationality. The crisis has pushed heads of state and governments to take on increased responsibilities, because fundamentally only they have the democratic legitimacy which allows them to make choices.” [59]

47Nailing voluntarism, pragmatism, protection and sovereignty to the European flag should, in fact, be understood as a way of “giving control of her destiny back to France”. [60] This embodied vision explains (or even justifies) the lone ranger strategy within the EU, illustrated by the announcement on 29 January 2012 of the decision to unilaterally tax financial transactions, because France is invested with a mission to “lead by example and provoke a shock”. [61] The political Europe that the French president called for could simply be France writ large. Sarkozy’s speeches as a candidate were no less ambiguous. From one speech to another, Europe was either forgotten (M1-Marseille), or subject to contradictory demands (M2-Villepinte), or reinterpreted (M3-Nice). There was no trace of Europe (or almost none) in the Marseille speech which was structured around a triple anaphora: “When we forget France, we forget…”, “Loving France means…” and “When you love France…” On the other hand, Europe occupied the central third of the Villepinte speech and was treated ambivalently. Although the passage dedicated to Europe opens with an affirmation of de facto solidarity, it closes – after questioning the validity of changes in Europe – with the unilateral threat of the use of (return to?) the “empty chair” policy.


“It is urgent because there is no question of us putting up with insufficient controls at Europe’s borders. I have taken the decision to commit totally to having these principles implemented because it is the only way to avoid the implosion of Europe. But if I see that in the year ahead that there is no serious progress in this direction, then France will suspend her participation in the Schengen agreements until negotiations succeed.” [62]

49The speech in Nice, whilst leaving much less space for Europe, accentuated this critique and re-evaluated borders as playing a calming role and as a response to the “errors” of the European Union:


“And it is true that the European project has lost its way, lost its way in technocracy, lost its values and diverted our compatriots from this ideal […] What mistakes have been made to get to this point? […] The second major mistake was in removing the founding words ‘nation’ and ‘border’ from our vocabulary. I know it seems strange that I’m talking to you about that and yet I feel it very deeply. I really think it was a mistake. Borders. Border are portrayed as a barrier by those who haven’t really considered the matter. But I think that borders are reassuring, borders are calming, that borders are comforting, and that a country, a society, that is reassured, comforted and confident is a society that is open; while a society that is anxious is a closed society.”

51Once again, paradoxisme is the most important stylistic device for understanding this extract, which here consists in affirming that an “open” society is “reassured” by its borders. This tension between being open and closed, between Europe and the border, in and of itself sums up the development of Sarkozy’s discourse over the course of the campaign, and his nationalistic shift to the right after Villepinte.

52Marine Le Pen stuck to the national-populist position which the National Front has espoused since the 1990s, and reserved a special place in her discourse for Europe, from her inaugural speech as party president on the 16 January 2011 in Tours. One fifth of this speech was dedicated to castigating “the Europe of Brussels” and “Euro-globalisation”, the freedom-killing agents of all France’s woes. A strong state, the restoration of monetary sovereignty and the abandonment of the euro, the on-shoring of production, and economic and social patriotism would henceforth be the tokens of national, social and identity renewal. According to Le Pen, in 2012 the choice would be binary: “either globalisation or the nation”, [63] “a free country or a nation enslaved” (M2-Toulouse), “the oligarchs or the people” (M1-Metz). This discursive structure would be woven throughout the campaign, structuring semantic pairs: “the elites/those forgotten by French politics”, “globalism/patriotism”, “virtual world/return to the real world”, “renunciation/ national sovereignty”. In this framework, Europe is presented as instrumental in “handing over the keys of France to the financial markets” (M2-Toulouse): “It was Jacques Delors, Pierre Bérégovoy and Pascal Lamy, the director of the WTO, who gave way to the ‘wall of money’ with the Single European Act” (M2-Toulouse). In this respect Europe is one of a number of things that acts as a metonym for figures of defeat (as corrupt oligarchies and transnational profiteers): Nicolas Sarkozy, senior ministers and public servants, the UMPS [= Union pour un mouvement populaire (UMP) plus the Parti socialiste (PS)], the European Union, the World Trade Organisation, and so on. Here, talking (a little) about Europe facilitates the (deep) condemnation of globalisation, in order to better return to France.

53The European theme was more or less present and explicit depending on the candidate, but in all cases it was organised around three leitmotifs: denunciation, nationalisation, and politicisation. First, a frame used by all candidates established the theme of the powerlessness and injustice of the European Union. Even if (major or minor) variations can be identified, they all converged in condemning the corruption of the European ideal, the democratic disengagement, and the lack of protection perceived by the peoples of Europe. Second, nationalisation was a trans-partisan theme; it can be seen in each candidate’s obsession with finding a French solution to Europe’s woes and proclaiming France’s universalist vocation. In the face of crisis and fears, the national territory becomes once again a space of refuge, which probably explains the importance of the theme of borders. [64] Third, and finally, we can see the outline of a politicisation of European issues throughout the campaign. The dynamics of denial (Dupont-Aignan) and displacement (Mélenchon, Joly, Le Pen) exist alongside a dynamic of bipolar confrontation: based on either solidarity and redistribution (Hollande) or national protectionism (Sarkozy). This triple framework helps to demonstrate how, in terms of the electoral offer, the relationship with Europe is shot through with double-talk, structural tensions, and conjunctural instrumentalisation. In any event, the relationship feeds various ideological wellsprings and resonates with the ambivalences of each party’s electorate.

The vote: bi-dimensionality and tripartition?

54Although Europe seems to have played a more divisive role in the messages of the campaigning candidates in 2012 than it has in the past, it is still not clear whether the politicisation of this issue was converted into an electoral choice and thus into a vote. Identifying the right indicators to allow us to measure the effect of Europe on the vote is not an easy task. Our previous results enabled us to identify two aspects of attitudes towards European integration using indicators of fears about the consequences of such integration, which were originally part of a battery of Eurobarometer items. [65] We thus first reproduced these analyses with the 2012 data, using the same indicators, in order to see whether this bi-dimensionality still persisted (see Box 1).

Box 1. Indicators of fear in relation to European integration used for factorial analysis

  • fear that the construction of Europe means there will be less social protection in France
  • fear that the construction of Europe means there will be more unemployment in France
  • fear that that the construction of Europe means we lose our national identity and our culture
  • fear that that the construction of Europe means that there will be an increase in the number of immigrants
  • fear that that the construction of Europe means France will pay for other countries
  • fear that that the construction of Europe means France plays a less important role in the world
Source: 2012 Cevipof post-election survey

55We conducted a factor analysis to synthesise the 2012 data and to operationalise our model for measuring the attitudes of French voters towards European integration. This is a statistical method that is particularly well-adapted to the identification of latent aspects of a concept – here the fear inspired by developments in European integration. Multiple correspondence analysis conducted on these six indicators of fear towards European integration first produces a three-dimensional analysis which explains 70.57% of the total inertia of the scatter graph of individuals. In this article, we will look at the first two dimensions, which explain more than 58% of the total variance. [66] The first factorial dimension (44.16% of the total inertia) broadly covers the first dimension identified in our previous work, both in 2002 and 2007. It represents a general opposition between all the categories of response indicating fear and those indicating the absence of fear. To qualify this dimension we verified that the expression of fear or absence of fear did indeed indicate a more general attitude towards European integration: that those who expressed fear, on whatever grounds, were indeed those who thought that France belonging to the European Union was “a bad thing” and who feel that the word “Europe” has very negative or quite negative connotations. Those who don’t express fear regarding European integration have a profile of opinions on Europe that is exactly the opposite. The expression of fear is most structured around those fears that that the construction of Europe means that there will be “a loss of national identity and our culture”, that “France will play a less important role in the world” and that there will be “an increase in the number of immigrants”, than by the other fears; whilst the lack of fear is structured less directly in relation to the different indicators. The second dimension (which accounts for 14.06% of the inertia [67]) of the factorial analysis opposes two types of fear concerning European construction (as was the case in 2002 and 2007): on one hand, there are those who say they fear that there will be “less social protection in France” but who don’t fear “that there will be an increase in immigrants”; and on the other hand, those who fear the contrary. The fear or non-fear of a decrease in social protection is the variable that most structures this dimension. The 2012 data thus confirm that in France attitudes towards the EU are bi-dimensional. We therefore propose to create two indexes from the scores of the first and the second dimensions of this factorial analysis in order to understand their impact on the vote in the first round of the presidential elections. These two indexes measuring attitudes towards European integration will be considered here as explanatory variables of the vote and integrated into a larger explanatory model. The empirical analysis will be conducted using a regression analysis model.

56The model we have developed is a multinominal logistic regression model in which our dependent variable, the vote, is a categorical variable with categories corresponding to the ten candidates for the presidential election. In order to simplify the data analysis and avoid major methodological problems linked to the low sample sizes for the minor candidates (Nathalie Arthaud, Philippe Poutou, Nicolas Dupont-Aignan, Jacques Cheminade, Eva Joly), we have limited our analysis to the five major candidates: Jean-Luc Mélenchon, François Hollande, François Bayrou, Nicolas Sarkozy and Marine Le Pen. These five candidates together account for a little more than 76% of votes cast in the first round and cover more or less the entire political spectrum of the first round of the presidential election. [68]

57Given that the dependent variable in the regression is a categorical variable, we have to consider one of the categories as the “reference category”, a “benchmark” category to which the votes for the other candidates will be compared. Multinomial logistic regression analysis proceeds by a series of binary logistic regressions for which we organise a systematic comparison between the reference category and the other categories in turn. Accepting the premise that the political space of the first round was “tripolar”, [69] the vote for Marine Le Pen will be considered as the reference category. Multinomial logistic regression will compare a series of contrasts between the votes for each of the other four candidates and Marine Le Pen.

58We aim to test the hypothesis that the two attitude dimensions regarding European integration identified by factorial analysis express specific effects on the vote in the first round, which are not simply already expressed by voters’ ideological preferences and systems of attitudes that meaningfully structure the vote. These include the scales of economic liberalism, cultural liberalism, ethnocentrism, opinions regarding globalization, and, of course, the left-right scale. [70] The research hypothesis was thus formulated as follows: all other things being equal, and once other factors have been taken into account, do the two attitude dimensions measuring fear about the developments of European integration have significant effects on the vote in the first round? We also introduced control variables: age (in five categories), gender, socio-professional categories (a series of dummy variables with the working-class category as the reference category) and level of education (measured by a series of dummy variables with the “no qualification” category as the reference). [71] Table 3 presents the regression coefficients expressed on an additive scale. [72] These coefficients each measure the effect of the explanatory variable on the “logit”, i.e. the logarithm of the likelihood of voting for Jean-Luc Mélenchon rather than for Marine Le Pen, then for François Hollande rather than for Marine Le Pen, then for François Bayrou and Nicolas Sarkozy rather than Marine Le Pen. [73] The vote in favour of the FN candidate is the reference category here: the regression coefficients on Table 3 are thus interpreted according to this reference.

59The attitude scales taken into account – the economic liberalism scale, cultural liberalism scale, ethnocentrism scale, and left-right scale – are firstly always connected in a very statistically significant way with the vote. [74] This result is not very surprising considering the links generally observed between attitudes and political behaviour. Once we have taken the attitude scales into account, the control variables have a much smaller impact than they would if we were testing a purely sociological model, the “class voting” or “cultural voting” model for example. [75] Concerning the two attitude dimensions relating to European integration more specifically, the analysis shows that the first attitude dimension, which reflects the most general feeling of fear, maintains its effects whichever binary choice our model places voters in. In other words, the fear of the consequences of European integration, made up of our six indicators of fear, produces effects that are always significant regarding the propensity to vote Marine Le Pen, whichever candidate she is compared to. This is an important result. Given that the correlations are quite strong between this first dimension of general fear and the cultural liberalism or ethnocentrism scales, we might have hypothesised that the effects of these explicative factors would absorb and encompass voters’ fears about European integration and the “integration/demarcation” cleavage more generally. [76]But this is not the case.Our first attitude dimension in relation to European integration produces its own significant effects on the vote in favour of Marine Le Pen.

60The effect of our second dimension of fears regarding European integration is less strong. With the vote for Marine Le Pen still our reference category, it is first and foremost the vote for Jean-Luc Mélenchon that is statistically significant here. The values and signs of the regression coefficients seem to indicate that the second dimension of fears about European integration orders votes according to the left-right scale. This interpretation is statistically valid at the two extremes of the scale; on this second dimension the votes for the Front de Gauche candidate and the Front National candidate are strongly opposed. The fact that this second dimension of fear towards Europe is less statistically significant than the first dimension in our regression is not surprising, given that the second dimension has quite a strong relation with the left-right scale, which is present in the model. [77] Table 3 shows however, that the effects on the vote are not null when the different attitudes scales are taken into account. This analysis shows that these two dimensions of fear regarding European integration seem to have had an impact on the vote in the first round of the presidential election. The first dimension plays an important role in the decision to vote for Marine Le Pen or not. The second dimension plays a less important role but nonetheless reinforces the division between the vote for the left and the vote for the right.

Table 3

Model of the first round vote according to factors of political attitudes, controlled by socio-demographic variables (vote for Le Pen is the reference category)***,**,*

Table 3
(1) (2) (3) (4) Mélenchon Hollande Bayrou Sarkozy Economic liberalism scale -1.833*** -1.178*** 0.243 1.012*** (0.301) (0.256) (0.280) (0.237) Cultural liberalism scale 0.450 0.439 -0.0573 0.491* (0.271) (0.226) (0.256) (0.209) Ethnocentrism scale -2.082*** -1.823*** -1.754*** -1.215*** (0.249) (0.218) (0.243) (0.207) Left-right scale -0.663*** -0.581*** -0.251*** 0.127** (0.0577) (0.0497) (0.0559) (0.0443) Opinion on Europe scale (dimension 1) -0.837*** -1.019*** -1.056*** -0.960*** (0.244) (0.211) (0.232) (0.197) Opinion on Europe scale (dimension 2) -0.731* -0.475 -0.317 0.0765 (0.348) (0.305) (0.336) (0.281) Opinion on globalisation: a danger for France -0.234 -1.164*** -0.711*** -1.320*** (0.272) (0.233) (0.258) (0.216) Constant 3.512*** 5.177*** 0.798 0.841 (0.713) (0.612) (0.724) (0.580) Number of observations 1787 1787 1787 1787

Model of the first round vote according to factors of political attitudes, controlled by socio-demographic variables (vote for Le Pen is the reference category)***,**,*

The standard errors are given in brackets
***. p < 0.01,
**. p < 0.05,
*. p < 0.1
To improve the readability of the data, the table does not include the regression coefficients for a range of control variables included in the model (age, sex, socio-professional category, level of education). The complete results are available from the authors upon request.

61It might be objected that taking the vote for Marine Le Pen as the reference category has exacerbated the statistical significance of the effect of fear about European integration. The same objection could be made about the effects of the second dimension to the extent that, in relation to fears about increased immigration and a decrease in social protection due to the EU, the vote for Le Pen also represents an extreme point. To address these objections we have modified the reference category and run the regression model for all the binary pairs that can be constructed from the five candidates, a total of ten. Table 4 presents only the coefficients that are significant at the 0.05 level for the two dimensions of attitudes towards European integration. It is remarkable to note that the only significant coefficients for the first dimension are those which oppose the four other candidates to Marine Le Pen. In other words, the effects of the first dimension of attitudes regarding European integration are above all felt in that configuration of the political space which opposes the FN candidates to the other four candidates in our analysis. For the second dimension of attitudes towards European integration, the regression coefficients are significant only when the binary choice is between Jean-Luc Mélenchon or François Hollande rather than Marine Le Pen or Nicolas Sarkozy. Expressing a fear about European integration in the social domain rather than in terms of immigration is indeed based on an ideological orientation in left-right terms, and this result persists when we control for the left-right scale. It is nonetheless remarkable to observe that the coefficient is not significant in the case of a binary choice for Nicolas Sarkozy rather than Marine Le Pen. The electorates of the UMP and the FN candidates are thus distinguished more in their general attitudes towards European integration (first dimension) than in terms of social fear or fear of immigration (second dimension).

Table 4

The effects of the two attitude dimensions towards European integration for all voting choices in the first round of the presidential elections. (Regression coefficients significant to 0.05 for all logits estimated)

Table 4
Scale of general fears regarding European integration B Z-test p-value Exp(B) Mélenchon-Le Pen -0.83725 -3.435 0.001 0.4329 Hollande-Le Pen -1.01872 -4.838 0.000 0.3611 Bayrou-Le Pen -1.05602 -4.549 0.000 0.3478 Sarkozy-Le Pen -0.95953 -4.868 0.000 0.3831 Scale of social fears regarding European integration Mélenchon-Sarkozy -0.80738 -2.673 0.008 0.4460 Mélenchon-Le Pen -0.73089 -2.099 0.036 0.4815 Hollande-Sarkozy -0.55135 -2.237 0.025 0.5762

The effects of the two attitude dimensions towards European integration for all voting choices in the first round of the presidential elections. (Regression coefficients significant to 0.05 for all logits estimated)

Key: In the first column the negative coefficients corresponding to the first attitude dimension indicate that when fear of European integration increases the chances of voting Mélenchon, Hollande, Bayrou or Sarkozy rather than Le Pen decrease. In other words, fear increases the relative chance of voting Le Pen. On the second attitude dimension, the negative coefficients indicate that when fear regarding social protection decreases, the chances of voting Mélenchon rather than Sarkozy or Le Pen, or Hollande rather than Sarkozy, also decrease.

62These first analyses clearly show that, all other things being equal, the two dimensions of attitudes towards European integration have an impact on the vote. Table 4 demonstrates that these effects are particularly strong in certain segments of the electorate and thus encourages us to extend our analysis, which we have done by analysing the probabilities predicted by the model for certain voter profiles. [78] In particular we have sought to establish what influence each of the major factors recognised as impacting on the vote – cultural liberalism, economic liberalism, ethnocentrism – and our two dimensions of attitudes regarding Europe, have on the probability of voting for the different candidates. [79] Our analyses are now conducted in terms of the marginal effects of explanatory variables on voting probabilities (see Table 5). They show that attitudes towards economic or cultural liberalism, ethnocentrism, or the position on the left-right scale have stronger effects on the vote for one candidate or another, than the two scales of attitudes towards Europe – except when it comes to the probability of voting Le Pen (the general dimension of attitudes regarding the EU can increase the probability of voting for Marine Le Pen by up to 14%). The decision to vote for the FN candidate thus seems to depend more on the position of these voters on our first attitudinal scale relating to Europe than on economic liberalism or even cultural liberalism and the left-right scale. It is the attitudes of these voters in terms of ethnocentrism and towards Europe that have the most impact. This result reinforces our previous results concerning the decisive role of attitudes towards Europe in the vote for Marine Le Pen.

Table 5

Marginal effects of attitude variables on the vote in the first round of the presidential elections(a),(b)

Table 5
Mean (a) Mélenchon Hollande Bayrou Sarkozy Le Pen Economic liberalism Min->max (b) .33348532 -.3250665 -.50864683 .15386023 .63996003 .03989303 Marginal effect .18587523 -.14424312 -.32044495 .11314876 .3196769 .03186241 Cultural liberalism Min->max .11038707 .03650778 .14327076 -.20471805 .09618914 -.07124964 Marginal effect .03490627 .01140931 .04357291 -.06383293 .03228345 -.02343274 Ethnocentrism Min->max .19668126 -.13567911 -.28114378 -.07488027 .14355654 .3481466 Marginal effect .7523109 -.05261205 -.10806827 -.02739741 .0795452 .10853253 Left-right scale Min->max .35447944 -.2356665 -.6505321 .02908158 .7400131 .11710392 Marginal effect .05610336 -.0336942 -.1065642 .01474527 .10195729 .02355584 Attitudes towards globalisation Min->max .06876473 .07206454 -.09244469 .03800285 -.0794671 .06184445 Marginal effect .07120374 .07523218 -.09737165 .03780675 -.08063769 .06497041 EU1 attitude scale Min->max .06271836 .01654238 -.09276637 -.0451469 -.01888265 .1402535 Marginal effect .02892526 .00902786 -0.4307384 -.02130191 -.0079374 .06328529 EU2 attitude scale Min->max .07317393 -.06874754 -.11384511 -.00034218 .14622167 .03671314 Marginal effect .04413954 -.04283115 -.06751769 .00122234 .0868656 .0222609

Marginal effects of attitude variables on the vote in the first round of the presidential elections(a),(b)

(a) This is the mean of absolute values of marginal effects;
(b) variation of the predicted possibility associated with a variation of the explanatory variable from its lowest value to its highest value.

63In terms of the probability of voting for Nicolas Sarkozy or François Hollande, our first dimension of attitudes towards Europe only has a very weak impact on the vote for the outgoing president, which seems above all influenced by his potential voters’ level of economic liberalism and position on the left-right scale. Attitudes towards Europe weigh a little more heavily in the vote for Hollande, particularly in the degree of variation which is nearly 10% when we move from the lowest value in the scale to the highest value (marginal coefficient -0.0927). François Hollande’s potential electorate – such as the model predicts it – thus appears to have been quite receptive to the candidate’s tone in favour of a new European dynamic; more so than Nicolas Sarkozy’s electorate was to his generally favourable stance towards European integration – a stance necessary for an outgoing president of the Fifth Republic. The probabilities of voting for Jean-Luc Mélenchon or François Bayrou seem less strongly influenced by attitudes towards European integration than by other explanatory factors. For Mélenchon these are: low economic liberalism, weak ethnocentrism and a leftwing position on the left-right scale. For Bayrou these probabilities are not very sensitive to the position on the left-right scale but are instead influenced by strong economic liberalism and weaker cultural liberalism.

64The marginal effects of our second scale of attitudes towards Europe appear less strong than those concerning economic or cultural liberalism or the left-right scale. This is not surprising given that this scale is more marked by the left-right dimension than the first one. However, its effects can be seen in the probabilities of voting for François Hollande or Nicolas Sarkozy. For the latter, the attitude of his electors on this scale has a positive influence that is stronger than their opinions on globalisation. For these voters, if globalisation is considered an opportunity, it is above all an economic perception of globalisation. [80] European integration is not seen as a threat to social protection in France but it is seen as a threat in terms of immigration. The combination of a vision of the French economy as positively exposed to globalisation but of French society as negatively exposed to immigration is one of the important results of Table 5. In this respect, Nicolas Sarkozy’s campaign speeches, and notably those where he expressed his very clear opinions on the Schengen area, allowed part of his potential electorate to buy into this two-fold representation of France as being open/closed to the world and to Europe. But this posture was also a handicap in his strategy for conquering Marine Le Pen’s electorate.

65* * *

66The study of both the campaign and the vote show that the European Union was indeed a factor – either directly or indirectly – in the French presidential election of 2012. The candidates’ arguments were structured by a dual frame of denunciation and politicisation, echoing citizens’ concerns. The EU was repeatedly condemned, from simply questioning the direction of French participation in the European integration process, to challenging the very principle of this participation. Politicisation was patent: structured according to a left-right opposition between the welfare state and the sovereign state. Above all, for the first time in a major national election, Europe appeared as the object of polarisation between the two main candidates. Such denunciation and politicisation resonated with the various fears citizens have regarding Europe – fears which are expressed through the vote both in their general dimension (strong fears clearly favouring a vote for Marine Le Pen), and in their specific dimension, with the opposition between fears about a loss of social protection and fears about immigration reinforcing the division between the left-wing and right-wing vote. The position regarding Europe weighs more heavily on the vote for Hollande than on the vote for Sarkozy, which seems to suggest that François Hollande’s message that he would construct a new European dynamic was heard within a potential electorate beyond the partisans of the PS. If he was able to achieve this “synthesis” from the first round, it was largely because the left-wing electorate appeared less divided on European questions in this election than the right-wing electorate.

67More generally, our results provide a threefold contribution to the current debates on the role of the European Union in the political spaces of the member states – in this case in France. First, what the candidates offer discursively appears to be linked to voters’ demands. The candidates’ treatment of the themes of defence and sovereignty and/or the protection of welfare benefits is indeed a mirror held up to the (major or minor) fears that citizens project onto the European object. The disconnect between elite discourses and citizens’ expectations that is often invoked to explain the lack of legitimacy and support for the EU (among other things) was not observed in this presidential election. Second, fears about Europe, measured through our first dimension, enable us to clearly distinguish Marine Le Pen’s electorate from those of the other candidates, appearing as one of the essential factors which increase the probability of voting for her over the other candidates. This leads us to explicitly defend the thesis of the tripartite structure of the 2012 presidential election. Finally, the bi-dimensional nature of attitudes towards the EU, their complex relationship with the left-right scale, and their influence on electoral behaviour, demonstrate that reflection on European integration as a cleavage is not a closed debate as some would like to believe. Instead it constitutes an invitation to pursue this question further. [81]


Appendix 1

Table A1

The 12 words most commonly used by each candidate in 2007*

Table A1
All Bayrou Besancenot Bové Buffet De Villiers Laguiller Le Pen Nihous Royal Sarkozy Schivardi Voynet Politics French Rights Rights Left Villiers Classes France CPNT* France Work Towns Years France France Businesses Rights Politics France Businesses French Politics Place Country Mayors Greens Years Years Profits World Employment French Working-Class Years Europe Democracy France European Union Creation Country Bayrou Debt Social (sing. adj) Life Vendée Power Candidate Planning Employment President (Publics) Reinforcement Work François Youth Social (pl. adj) Social Politics Vote Le Pen Rights (Just) Sarkozy Rupture Research Life Country Work Vote Rights Philippe Housing Future EnvironmentalismCreation Crisis Politics Biodiversity (Social) Children Life Right-Wing Salaries Europe Problems Promises Freedom Royal Employment Services Justice Business Politics Euros Women Vote Years Revenues State Territory Ségolène Politics Gérard Law Public President Law Life France Patriotism Population Republic Frédéric Education Businesses Maastricht Politics Rights Research Millions Bové Right-Wing Power Profits Candidates Future Jobs Europe Schivardi Protection French Retirement Retirement Access Women Republic Salaries Choices Campaign Salaries Years Article Stopping (Europeans) State Europe April Training Right-Wing Life (French) Development State Unemployment Defence Extension

The 12 words most commonly used by each candidate in 2007*

* CPNT Party = Chasse, pêche, nature et traditions (Hunting, fishing, nature and tradition)

Appendix 2

Table A2

Corpus “Presidential Speeches 1st Round 2012”*

Table A2
Rallies* Date Town Place Medium: video + text Bayrou M1 10.12.11 Pau Video+text M2 24.03.12 Paris Zénith Video+text M3 16.04.12 Lyon Video+text Dupont-Aignan M2 25.03.12 Paris Balard Video+text Hollande M1 22.01.12 Le Bourget Video+text M2 15.04.12 Vincennes Video+text M3 19.04.12 Cenon Video+text Joly M3 18.04.12 Paris Cirque d’hiver Video+text Le Pen M1 11.12.11 Metz Video+text M2 07.02.12 Toulouse Video+text Mélenchon M1 29.06.11 Paris Place Stalingrad Video+text M2 05.04.12 Toulouse Video+text Sarkozy M1 19.02.12 Marseille Video+text M2 11.03.12 Villepinte Video+text M3 20.04.12 Nice Video+text

Corpus “Presidential Speeches 1st Round 2012”*

* M1: first rally, M2:“major rally”, M3: last rally

Appendix 3. Constructing the attitude scales

68Several models can be used to construct attitude scales. These scales combine the scores which reflect the degree of acceptance of a so-called “latent” attitude – which cannot be observed directly but is measured by a number of “manifest” indicators. Two major types of analysis are most commonly used: on one hand techniques which suppose that the indicators can be accumulated to form patterns of responses (such as the summative Likert scales); on the other hand the so-called “French” [82] factorial analysis, and particularly multiple correspondence analysis (MCA). We have adopted the latter technique given that MCA is particularly appropriate for our types of variables (nominal categorical or ordinal categorical) and is a more powerful and suitable technique for obtaining individuals’ “optimal scores” on latent variables. The first dimension of this kind of analysis provides an optimal coding of the individuals. The attitude scales used in this article were thus constructed via a series of specific multiple correspondence analyses, conducted separately for each scale. We then analysed the first factor in each of these analyses in order to verify that it was indeed evaluating the attitude we were looking at in the data, and we used the optimal coding of the individuals as explanatory factors of the vote.

69The indicators used for each attitude scale are listed below, along with the coding proposed to respondents at the time of the survey. These indicators were selected following several multiple correspondence analyses for each attitude scale, with the items that did not contribute (much or at all) to the first factor of the MCA being systematically removed in order to obtain the most “pure” and optimal factorial coding possible. The other explanatory variables in our multinomial logistic model were directly available in the 2012 Cevipof postelectoral survey.

Economic liberalism scale: 8 indicators, eigenvalue of the first axis (1 = 0.2965, or 13.05% of the inertia)

70To establish social justice, we should take from the rich to give to the poor

71Unemployed people could find work if they really wanted to

72The risk of unemployment benefits is that they may encourage people to be satisfied with that and not look for work

73The right to strike should be more limited for public transport workers

74Bosses should be able to fire staff more easily so that they are not afraid of hiring

75(Strongly agree/ Mostly agree/ Mostly disagree /Strongly disagree/No response-Don’t know)

76In order to address economic difficulties, do you think that… the state should trust businesses and give them more freedom; or, on the contrary, control and regulate them more closely?

77(Refusal-No response-Don’t know)

78Do you think that in the years to come we should give priority to… the competitiveness of the French economy; or the improvement of conditions for workers?

79(Refusal-No response-Don’t know)

80If you had to choose between… working more and earning more money; or working less and earning less money; what would you choose?

81(Refusal-No response-Don’t know)

Cultural liberalism scale: 7 indicators, eigenvalue of the first axis (1 = 0.2817, 10.33% of inertia)

82The death penalty should be reintroduced

83Parents have no authority these days

84Homosexuality is an acceptable form of sexuality

85When there is an employment crisis men should have priority over women in finding work

86A pre-school child is likely to suffer if his or her mother works

87(Strongly agree/ Mostly agree/ Mostly disagree /Strongly disagree/No response – Don’t know)

88For each of the following words, tell me if it evokes something very positive, quite positive, quite negative or very negative for you: feminism

89If you think about school, which of these two statements do you agree with most? Above all, school should provide a sense of discipline and hard work; above all, school should teach people to be enlightened and critical

Ethnocentrism scale: 5 indicators, eigenvalue of the first axis (1 = 0.4609, 15.2% of inertia) [83]

90There are too many immigrants in France

91French people should have priority over immigrants in terms of employment

92We don’t feel safe anywhere

93(Strongly agree/ Mostly agree/ Mostly disagree /Strongly disagree/No response-Don’t know)

94In your opinion, might respecting the following Muslim religious practices be problematic for life in French society?

95Public prayer

96The wearing of the burqa

97(Yes, strongly agree/Yes somewhat/No, not really/No, Not at all/Refusal/No response/Don’t know)


  • [1]
    Joint interview with Nicolas Sarkozy and Angela Merkel, 6 February 2012 on French (France 2) and German (ZDF) television following a Franco-German Council of Ministers.
  • [2]
    Joint press conferences held at the Elysée Palace on 17 February 2012.
  • [3]
    European meeting, “Une renaissance pour l’Europe: vers une vision progressiste commune”, organised by the Fondation Jean Jaurès, the FEPS, the Fondation Italiani Europei and the Fondation Friedrich-Ebert, 16-17 March 2012.
  • [4]
    Renaud Dehousse, Angela Tacea, “The French 2012 presidential election. A Europeanised contest”, Les Cahiers Européens de Sciences Po, 2, 2012, Paris: Centre d’études européennes de Sciences Po, 7.
  • [5]
    Translator’s note: specifically, what is analysed here are the candidates’ “professions de foi”. These are small leaflets (two pages maximum) from each candidate which every voter receives by mail a week before the election. In this leaflet, the candidate presents him/herself along with the main ideas of his/her programme.
  • [6]
    Translator’s note: in this article the expressions “campaign rally speeches”, “rally speeches”, and “campaign speeches” are used interchangeably to refer to the speeches given at the meetings and rallies during the campaign.
  • [7]
    Cf. in particular Olivier Costa, Paul Magnette (eds), Une Europe des Élites? Réflexions sur la fracture démocratique de l’Union Européenne (Brussels: Éditions de l’Université de Bruxelles, 2007); Sophie Duchesne, Elizabeth Frazer, Florence Haegel, Virginie Van Ingelgom, Citizens’ Reactions to European Integration Compared. Overlooking Europe (Basingstoke: Palgrave Macmillan, 2013); Dieter Fuchs, Raul Magni-Berton, Antoine Roger (eds), Euroscepticism. Images of Europe Among Mass Publics and Political Elites (Opladen & Farmongton Hills: Barbara Budrich Publishers, 2009).
  • [8]
    Klaus Goetz, Simon Hix (eds), special issue, “Europeanised politics? European integration and national political systems”, West European Politics, 23(4), 2000, 1-231. See particularly the contributions by Peter Mair on political parties and Matthew Gabel on voters: Peter Mair, “The limited impact of Europe on national party systems”, 27-51; Matthew Gabel, “European integration, voters and national politics”, 52-72.
  • [9]
    This expression was coined by Catherine de Vries in her article “Sleeping giant: fact or fairy tale? How European integration affects national elections”, European Union Politics, 8(3), 2007, 363-85.
  • [10]
    Peter Mair, “The limited impact of Europe on national party systems”.
  • [11]
    See in particular Nick Sitter, “Euro-scepticism as party strategy: persistence and change in party-based opposition to European integration”, Austrian Journal of Political Science, 32(3), 2003, 239-53; Thomas Poguntke, Nicholas Aylott, Elisabeth Carter, Robert Ladrech, Richard Luther (eds), The Europeanization of National Political Parties: Power and Organizational Adaptation (Abingdon: Routledge, 2007).
  • [12]
    Robert Harmsen, “L’Europe et les partis politiques nationaux: les leçons d’un ‘non-clivage’”, Revue internationale de politique comparée, 12(1), 2005, 77-94.
  • [13]
    For a good summary of this debate at the end of the 2000s, see the chapter by Antoine Roger, “Clivages et partis politiques”, in Céline Belot, Paul Magnette, Sabine Saurugger (eds), Science politique de l’Union européenne (Paris: Economica, 2008), 197-213.
  • [14]
    See Peter Jochen, Claes H. de Vreese, “In search of Europe – a cross-national comparative study of the European Union in national television news”, Harvard Journal of Press/Politics, 9(4), 2004, 3-24; Mathieu Petithomme, “L’absence de conflictualisation des débats politiques sur l’Union européenne au sein des presses nationales: une analyse de contenu qualitative comparée des cas français, britannique, irlandais et espagnol (2005-2006)”, Politique européenne, 33, 2011, 215-53.
  • [15]
    Virginie Van Ingelgom, “Mesurer l’indifference. Intégration européenne et attitudes des citoyens”, Sociologie, 3(1), 2012, 1-20; Florian Stoeckel, “Ambivalent or indifferent? Reconsidering the structure of EU public opinion”, European Union Politics, 14(1), 2013, 23-45.
  • [16]
    Geoffrey Evans, “Euroscepticism and Conservative electoral support: how an asset became a liability”, British Journal of Political Science, 28(4), 1998, 573-90.
  • [17]
    Erik Tillman, “The European Union at the ballot box? European integration and voting behaviour in the new member states”, Comparative Political Studies, 37(5), 2004, 590-610.
  • [18]
    Harald Schoen, “Turkey’s Bid for EU membership, contrasting views of public opinion, and vote choice. Evidence from the 2005 German federal election”, Electoral Studies, 27(2), 2008, 344-55.
  • [19]
    Matthew Gabel, “European integration, voters and national politics”, West European Politics, 23(4), 2000, 52-72.
  • [20]
    Catherine de Vries, Erik Tillman, “European Union issue voting in East and West Europe: the role of political context”, Comparative European Politics, 9(1), 2011, 1-17.
  • [21]
    Catherine de Vries, Wouter van der Brug, Marcel van Egmond, Cees van der Eijk, “Individual and contextual variation in EU issue voting: the role of political information”, Electoral Studies, 30(1), 2011, 16-28.
  • [22]
    Catherine de Vries, “EU issue voting: asset or liability? How European integration affects parties’ electoral fortunes”, European Union Politics, 11(1), 2010, 89-117.
  • [23]
    “Europe presents nothing new”: Jocelyn Evans, “The European dimension in French public opinion”, Journal of European Public Policy, 2007, 14(7), 1098-116 (1112).
  • [24]
    Jocelyn Evans, Gilles Ivaldi, “Quand la crise du consensus profite à l’extrême droite”, Le Figaro, 18 May 2002, 14.
  • [25]
    Gérard Grunberg, Florence Haegel, La France vers le bipartisme? La présidentialisation du PS et de l’UMP (Paris: Presses de Sciences Po, 2007).
  • [26]
    Nicolas Sauger, Sylvain Brouard, Emiliano Grossman (eds), Les Français contre l’Europe? (Paris: Presses de Sciences Po, 2007), 85.
  • [27]
    Pascal Perrineau, “Le referendum français du 29 mai 2005. L’irrésistible nationalisation d’un vote européen”, in Pascal Perrineau (ed.), Le vote européen, de l’élargissement au referendum français (Paris: Presses de Sciences Po, 2005), 229-44.
  • [28]
    Alain Mergier et al. Le jour où la France a ditnon” (Paris: Plon, 2005).
  • [29]
    Nicolas Sauger, Sylvain Brouard, and Emiliano Grossman (eds) Les Français contre l’Europe?, 115.
  • [30]
    Annick Percheron, “Les Français et l’Europe: acquiescement de façade ou adhésion véritable? Note de recherche”, Revue française de science politique, 41(3), 1991, 382-406.
  • [31]
    Gérard Grunberg, Etienne Schweisguth, “Vers une tripartition de l’espace politique”, in Daniel Boy, Nonna Mayer (eds), L’électeur a ses raisons (Paris: Presses de Sciences Po. 1997), 179-218.
  • [32]
    Gérard Grunberg, Etienne Schweisguth, “La tripartition de l’espace politique”, in Pascal Perrineau, Colette Ysmal (eds), Le vote de tous les refus, Les élections présidentielle et législatives de 2002 (Paris: Presses de Sciences Po, 2003), 339-62.
  • [33]
    Jean Chiche, Brigitte Le Roux, Pascal Perrineau, Henry Rouannet, “L’espace politique des électeurs français à la fin des années 1990. Nouveaux et anciens clivages, hétérogénéité des électeurs”, Revue française de science politique, 50(3), 2000, 463-87.
  • [34]
    Céline Belot, Bruno Cautrès, “L’Europe, invisible mais omniprésente?” in Bruno Cautrès, Nonna Mayer (eds) Le nouveau désordre électoral. Les leçons du 21 avril 2002 (Paris: Presses de Sciences Po, 2004), 119-41; Bruno Cautrès, Sylvie Strudel, “Les traces du referendum du 29 mai 2005 dans la campagne présidentielle de 2007”, Les Cahiers du Cevipofle Baromètre politique français (2006-2007), 46, April 2007, 141-8.
  • [35]
    Here we refer to the open controversy launched in 2006 by Simon Hix and Stefano Bartolini on the Politicisation de l’UE: remède ou poison? Here we focus on one element of this discussion, which questions the potential left-right polarisation of the European process.
  • [36]
    Thierry Vedel, Comment devient-on président(e) de la République? (Paris: Robert Laffont, 2007), 122.
  • [37]
    For a defence of logometric approaches to political discourse, see Damien Mayaffre, Le discours présidentiel sous la Cinquième République (Paris: Presses de Sciences Po, 2012), 14ff.
  • [38]
    This programme was designed by Cédric Lamalle, William Martinez, Serge Fleury and André Salem (SYLEDCLA2T university team at Université Paris III-Sorbonne Nouvelle). We would like to thank the latter in particular for his help in using the programme.
  • [39]
    See Ruth Amossy, L’argumentation dans le discours (Paris: Nathan, 2000); Paul Bacot, La construction verbale du politique: études de politologie lexicale (Paris: L’Harmattan, 2011); Constantin Salavastru, Rhétorique et politique. Le pouvoir du discours et le discours du pouvoir (Paris: L’Harmattan, 2005).
  • [40]
    Céline Belot, Bruno Cautrès, “L’Europe, invisible mais omniprésente?”; Bruno Cautrès, Sylvie Strudel, “Les traces du referendum du 29 mai 2005 dans la campagne présidentielle de 2007”.
  • [41]
    A group of forms includes occurrences of the different graphical forms linked by a common characteristic in French. It combines the singular, the plural, the conjugations of the same verb, and forms connected by a semantic link. Here the group of forms europ* corresponds to the sum of all the terms beginning with the letters “europ”: Europe, European (Europe, l’Europe, européen, européenne, européen, etc.).
  • [42]
    These terms are as follows: Europe, referendum (only when the term is relative to a referendum on European questions), directives, directive, European, Maastricht, Bolkestein, ECB, stability pact, Brussels, Central Bank, CAP, common agricultural policy, treaty. (Translator’s note: there are fewer than eighteen terms in English, where adjectives – such as “European” – do not vary in gender and number as they do in French.)
  • [43]
    43 out of 10,000 in 2007 compared to 39 in 2012, for the former; 65 out of 10,000 in 2007 compared to 53 in 2012 for the latter. A search for the occurrences of the terms “Europe” and “European Union” in the corpus of Le Monde, Figaro, and Liberation in the three months preceding the election showed that these references were no more present in the media in 2012 than in the three previous presidential elections.
  • [44]
    The fran* group consists of the terms France and (the) French (France, Français, Française, Françaises, français, française, françaises).
  • [45]
    We would like to thank Odile Gaultier-Voituriez, who runs the Documentation Centre at the Centre de recherches politiques de Sciences Po (Cevipof), and who facilitated our research and made useful suggestions for the construction of the corpus.
  • [46]
    Such as that delivered by François Hollande at the meeting entitled “Une renaissance pour l’Europe: vers une vision progressiste commune” at the Cirque d’hiver, 17 March 2012, as part of the European meeting organised by the Fondation Jean Jaurès, the FEPS, the Fondation Italiani Europei and the Fondation Friedrich Ebert. Obviously exploration of all the candidates’ speeches at a later date and within a longer-term research project will enable us to identify not only convergences but also divergences within their discourse, taking both general and thematic speeches into account.
  • [47]
    For Nicolas Sarkozy and François Hollande, we have considered that their first two meetings were actually “warm ups” before the first real meeting: Mérignac on the 5 January 2012 for François Hollande before Le Bourget (22 January), and Annecy on the 16 February 2012 before Marseille (19 February) for Nicolas Sarkozy.
  • [48]
    Whether by the size of the audience and/or by the media comments it provoked.
  • [49]
    See note 1, p. 56.
  • [50]
    The differences between the manifestos and rally speeches may stem from the fact that the former are more standardised exercises, with a certain number of obligatory features. The higher number of references to Europe in Eva Joly’s campaign manifestos were thus part of an obligatory nod in that direction.
  • [51]
    Anaphora consists in beginning verse or prose with the same word or the same phrase (for example “Paris! Paris outraged! Paris broken! Paris martyred! But Paris liberated!” Charles de Gaulle, extract from his speech on 25 August 1944). Synecdoche consists in using a part to stand for the whole (for example: “I see sails ahead”, here the sails (part) signify boats (whole)). See Pierre Fontanier, Les figures du discours (Paris: Flammarion, 1977).
  • [52]
    Michel Hastings, Sylvie Strudel, “Gauche indivise et gauches singulières”, in Pierre Bréchon, Annie Laurent, Pascal Perrineau (eds), Les cultures politiques des Français (Paris: Presses de Sciences Po., 2000), 181.
  • [53]
    Craig Parsons, “Puzzling out the EU role in national politics”, Journal of European Public Policy, 14(7), 2007, 1135-49.
  • [54]
    There is obviously no trace of any mention of or allusion to the 2005 referendum.
  • [55]
    François Hollande is referring here to the new treaty on stability, coordination and governance within the EMU, which was finalised by the member states (excepting Great Britain and the Czech Republic) on 30 January 2012.
  • [56]
    The following citations are extracts from his speech in Pau, 10 December 2011.
  • [57]
    In rhetoric, an oxymoron is a figure of opposition which consists in bringing together two terms which have opposite meanings within the same phrase (for example: “light darkness”, “hurry slowly”). Moreover, here we are close to a “paradoxisme”, a stylistic device identified by the grammarian Pierre Fontanier, which combines contradictory attributes in such a way as to make an impact (for example: “property is theft”).
  • [58]
  • [59]
  • [60]
    But also: “Europe does not mean less sovereignty but more sovereignty, because it means a greater capacity for action”, Toulon speech.
  • [61]
    Television interview with Nicolas Sarkozy, Sunday 29 January 2012.
  • [62]
    He made a second, parallel commitment: “So I solemnly ask, why should Europe deny itself what the United States, the most liberal country in the world, allows itself. France will demand that Europe adopt a ‘Buy European Act’ along the lines of the ‘Buy American Act’. Thus European businesses whose production is based in Europe will benefit from public money. I intend to commit totally to this struggle. If in the next twelve months, no serious progress on the need for reciprocity with our main partners is observed, then France will unilaterally apply this rule until negotiations succeed.”
  • [63]
    Marine Le Pen’s inaugural speech as president of the Front National, 16 January 2011 in Tours. (<>
  • [64]
    “Borders’ is the term that occurs most often in association with the europ* group of forms. In other words it is the noun most often associated with Europe in these speeches.
  • [65]
    Given that a large proportion of European citizens continue to declare a certain amount of support for the EU in the Eurobarometer surveys, even though at the same time the results of the referendums relating to European integration are much more uncertain than they were in the past, it seems more sensible to explore the question of attitudes towards European integration through the obstacles and fears it generates.
  • [66]
    The third dimension is not interpreted because the histogram of eigenvalues shows only a slight decrease between the second and the third dimension, suggesting that the analysis can be based on the two first dimensions alone. Moreover the third dimension of factorial analysis is characterised by a single variable (the opposition around the fear that the construction of Europe means that France will pay for other countries).
  • [67]
    The eigenvalue is slightly inferior to the average eigenvalue (1/6=0.165) and could lead us to not investigate this second dimension. However, it is statistically acceptable to continue the analysis if we can interpret this dimension. The interpretation of the second dimension of our analysis is particularly significant in terms of the left-right scale. Moreover, our further analysis has shown the statistical significance of this second dimension where it is introduced into a regression model to explain the vote. It is clear that all results of the factorial analysis could be further consolidated if we had many more indicators.
  • [68]
    Analysing the discourses of the smaller candidates has nonetheless enabled us to understand the Europeans discourses voters are subjected to overall.
  • [69]
    This tripolar dimension, which has constituted the heart of the electoral order in France since the 1984 European elections, has been empirically analysed and demonstrated by Gerard Grunberg and Etienne Schweisguth, “La tripartition de l’espace politique”. See also, Jean Chiche, Béatrice Le Roux, Pascal Perrineau, Henri Rouanet, “L’espace politique des électeurs français à la fin des années 1990s”. The thesis of tripartition has more recently been the object of complementary analyses which both nuance it, and, at least in part, challenge it. See Robert Anderson, Jocelyn Evans, “The stability of French political space, 1988-2002”, French Politics, 3, 2005, 282-301. See also Florent Gougou, Simon Labouret, “La fin de la tripartition? Les recompositions et la transformation du système partisan”, Revue française de science politique, 63(2), 2013, 279-302. Here we consider it as the premise for the basis of our analysis and we empirically test its relevance further on in our article.
  • [70]
    We have controlled for possible co-linearity of explanatory variables. The values of the VIF coefficients (Variable Inflated Factor) fall between 1.13 and 2.40, which means that we do not have co-linearity problems.
  • [71]
    Appendix 3 explains in detail how these scales were constructed, notably using specific multiple correspondence analysis.
  • [72]
    The statistical tests (chi square test) allow us to verify that the model (Table 3) is better adjusted to the observed data than if the two attitude scales relating to the EU were withdrawn
  • [73]
    We thank our colleagues Flora Chanvril at the Cevipof and Hyungsoo Woo, PhD student at Sciences Po and the LSE, for reading our analysis and for their comments and help on this part of the text.
  • [74]
    The fact that in Table 5 the cultural liberalism scale appears to exert far less of a significant effect than the economic liberalism scale is due to the fact that the vote for Le Pen is the reference category here. This makes the effects of the ethnocentrism scale appear more significant. If we were to take other votes as the reference category then the effects of the cultural liberalism scale would once again become more significant.
  • [75]
    It is common to observe that when independent opinion or attitude variables and socio-demographic variables are simultaneously introduced into an explanatory model, the effects of the latter become much less (or even not at all) significant. This does not invalidate exclusively sociological analyses of the vote (for example the “class vote”) which are an important aspect of explaining voting behaviour.
  • [76]
    The effects of this new cleavage, linked to European integration as well as globalisation, have notably been analysed by Hans-Peter Kriesi, Edgar Grande, Romain Lachat, Martin Dolezal, Simon Bornschier, Timotheos Frey, West European Politics in the Age of Globalization (Cambridge: Cambridge University Press, 2008). On the effects of European integration in terms of “winners” and “losers”, see Neil Fligstein, Euroclash. The EU, European Identity, and the future of Europe (New York: Oxford University Press, 2008).
  • [77]
    By removing the left-right scale from the regression model, the effects of the second dimension of fear about European integration become much more significant. They confirm the opposition that this dimension establishes between the votes for left-wing candidates and those for Nicolas Sarkozy or Marine Le Pen.
  • [78]
    According to our model, based on their attitudinal and socio-demographic profile each voter has a particular probability of voting for one candidate rather than another. Focusing on the probabilities of voting predicted by the model can be seen as part of electoral research’s recent interest into voting probabilities. See Vincent Tiberj, Bernard Denni, Nonna Mayer, “Un choix, des logiques multiples. Préférences politiques, espace des possibles et votes en 2012”, Revue française de science politique, 63(2), 2013, 249-78.
  • [79]
    Given that these are complex analyses we cannot present them fully here; the detail of our analyses is available upon request.
  • [80]
    This is also an effect of the wording of the indicator that measures opinion on globalisation: “For a country like France, globalisation is more an opportunity because it opens up markets overseas and pushes France to modernise.”
  • [81]
    We would like to thank Isabelle Guinaudeau and the readers at the Revue française de science politique for their comments and suggestions.
  • [82]
    Particularly Principal Component Analysis (PCA) and Multiple Correspondence Analysis (MCA) which, although their origins are not exclusively French, did develop extensively in France through the impetus of Jean-Paul Benzécri. Specific multiple correspondence analysis was developed by Jean Chiche, Brigitte Le Roux and Henri Rouanet: see Brigitte Le Roux, Henri Rouanet, Multiple Correspondence Analysis (Thousand Oaks: Sage Publications, 2009).
  • [83]
    For this scale we made sure that the coefficient of internal consistency (Crohnbach’s alpha) was not stronger if we removed the two opinion indicators on Muslim religious practice. The correlations between these indicators and the three others are moderate or even weak (the strongest correlations are with the opinion that there are too many immigrants in France, with the values of 0.274 and 0.218 respectively). However, the coefficient of internal consistency is 0.670 with the five indicators and 0.622 or 0.641 if we remove either the first or the second item relating to religious practice. The empirical argument alone is not sufficient however to validate the choice to maintain these indicators in the scale. It is clear that if our explicative model was intended to test the specific effects of an attitude towards Islam in particular we would have to identify what was due to this attitude and what was due to ethnocentrism more generally. The two attitudes could be due to different logics, but the correspondence analysis does not clearly illustrate this result on the subsequent factors. The first factor in the correspondence analysis only shows the five items that contribute to the same axis; the opinion on public prayer being even more strongly correlated with this dimension than opinion on the wearing of the burqa.

What role did the European issue play in the domestic policy programmes and in voter choice during the 2012 French presidential election? Using three different datasets (manifestos, campaign speeches and the Cevipof post-electoral opinion poll), our results unveil the disturbing impact of the European issue. The candidates’ speeches show a double focus of denouncing and politicising (revealing a left-right structure opposing welfare state andsovereign state) which echoes citizens’ worries. Our article confirms the bi-dimensionality of attitudes towards Europe. As these two dimensions do not have the same influence on electoral behaviour, our results suggest that the debate over the role of European integration in the emergence of a new political cleavage is far from settled.

Céline Belot
Céline Belot is a CNRS researcher at the PACTE Research Centre and lectures at Sciences Po Grenoble. She recently published: “Backing our men. Comprendre le soutien des Britanniques et des Français à leurs forces armées”, Gouvernement et action publique, 2(4), 2013, 595-620; “Moving towards a more tolerant society? Attitudes towards immigrants in French politics”, French Politics, 10(3), 2012, 209-32; (with Pierre Bréchon) “Gouverner par les sondages? Nicolas Sarkozy et l’opinion publique”, in Jacques de Maillard, Yves Surel (eds), Les politiques publiques sous Sarkozy (Paris: Presses de Sciences Po, 2012), 71-90. Her research deals with the process of legitimation of the EU, the political values of Europeans, and the relationship between public opinion and public policy, in particular in the domain of foreign affairs (PACTE, Sciences Po Grenoble, BP 48, 38040, Grenoble cedex 9).
Bruno Cautrès
Bruno Cautrès is a CNRS researcher at the Centre de recherches politiques de Sciences Po (Cevipof) and lecturer at Sciences Po. He has recently published: “Le vote Mélenchon: le poids de l’héritage communiste et les limites d’une dynamique”, in Pascal Perrineau (ed.), Le vote normal. Les élections présidentielle et législatives d’avril-juin 2012 (Paris: Presses de Sciences Po, 2013), 110-31; (with Flora Chanvril and Nonna Mayer) “Retour sur l’hypothèse de ‘l’homologie structurale’: les déplacements des catégories sociales dans l’espace politique français depuis La Distinction”, in Philippe Coulangeon, Julien Duval (eds) Trente ans après La Distinction de Pierre Bourdieu (Paris: La Découverte, 2013), 327-37; (with Mariette Sineau), “Les attentes vis-à-vis du nouveau président”, in Pascal Perrineau (ed.), La décision électorale en 2012 (Paris: Armand Colin, 2013), 229-42. His research focuses on the analysis of the vote and elections, and the sociology of political attitudes and behaviour, particularly regarding European integration and globalisation (Cevipof, 98 rue de l’Université, 75007).
Sylvie Strudel
Sylvie Strudel is professor of political science at the Université Panthéon-Assas (Paris-II), researcher at the Centre d’études constitutionnelles et politiques (Paris II), and associate researcher at the Centre Marc-Bloch in Berlin and at the Centre de recherches politiques de Sciences Po. She has recently published: “La marginalisation centriste”, in Pascal Perrineau (ed.) Le vote normal. Les élections présidentielle et législatives d’avril-juin 2012 (Paris: Presses de Sciences Po, 2013), 175-91; (with Bruno Cautrès) “Les flux à l’intérieur de la droite: d’une victoire centripète à la défaite centrifuge” in Pascal Perrineau (ed.) La décision électorale en 2012 (Paris: Armand Colin, 2013), 109-28; (with Karolina Koc Michalska), “European citizenship in action: EU movers as voters”, in “All Citizens Now”. Intra-EU Mobility and Political Participation of British, Germans, Poles and Romanians in Western and Southern Europe, report of the MOVEACT Project, April 2012. Her research focuses on the analysis of the vote and political behaviour, the theory and uses of citizenship (particularly European), and the impact of European integration on national political systems (Université Paris II-Panthéon Assas, 12 place du Panthéon, 75005 (Paris).
Translated from French by
Katharine Throssell
Latest publication on cairn or another partner portal
Uploaded on on 09/10/2014
Distribution électronique pour Presses de Sciences Po © Presses de Sciences Po. Tous droits réservés pour tous pays. Il est interdit, sauf accord préalable et écrit de l’éditeur, de reproduire (notamment par photocopie) partiellement ou totalement le présent article, de le stocker dans une banque de données ou de le communiquer au public sous quelque forme et de quelque manière que ce soit.
Loading... Please wait