The Language(s) of Politics

Front matter

MULTILINGUAL POLICY-MAKING IN THE EUROPEAN UNION ase =>, a

Nils Ringe

Multilingual Policy-Making in the European Union

Contents

Acknowledgments List of Figures List of Tables

ONE The Language(s) of Politics: Multilingual Policymaking in the European Union Two Multilingualism in the EU: How It Works

THREE The EU’s Language Regime: Institutional Stability and Change

Part 1: Language Use and Language Services

There are three main reasons why the EU operates as a multilingual political community. First, multilingualism is indispensable in the EU as an “inter-national” entity, as it allows individual languages to retain their position as markers of national identity while providing the basis for building a common political community. EU multilingualism is consciously built on the recognition that languages are a means of social, cultural, and political identification and differentiation, and also of power and control. This makes the elevation of one or some languages over others fundamentally problematic in a political system that has to continuously strike the delicate balance between assuring the national sovereignty of its member states—no matter how small—while constructing a political union. Multilingualism is a key instrument in building a European community that is “anited in diversity.”

Second, multilingualism is crucial to the functioning of democracy in the EU and by extension for the EU’s democratic legitimacy (Mamadouh 1999), because “speaking the language of the state is often the critical condition enabling the citizen to participate in the political arena of that state,” as Laitin put it succinctly (Laitin 1977, 3). Offering EU citizens access to information in an official language of their choice is thus vitally important in that it helps ensure the right and capacity of citizens to participate in EU politics by having access to the information they need to understand how decisions are deliberated, negotiated, and finalized; to learn about the content and consequences of laws and regulations they are subject to; and to evaluate the performance of the representatives they elect. Even if citizens do not actively take advantage of those opportunities, EU multilingualism is critical for democratic accountability, because elected representatives are only responsible to their voters if those voters are able to get the information they need to decide if an incumbent should continue to serve in office or be replaced by somebody else. It is also essential for the legitimacy of the EU’s institutions and the decisions that are made therein. A language regime centered on a small number of official languages would make EU citizens feel even more disconnected from EU institutions than many already do and undermine the extent to which they view those institutions as rightfully making decisions that affect the individual lives of EU citizens in countless ways.

It is not only the represented who rely on a functioning multilingual regime, however, but also their representatives. As the EP put it its resolution of January 1995, “the right of an elected person to express himself and to work in his own language is an indissociable part of democratic rights and of his mandate” (cited in Athanassiou 2006, 17). Here one must make a clear distinction between EU civil servants and politicians who are either elected to represent the European people in the European Parliament, or who represent the elected national governments of the EU member states in the Council of Ministers. EU civil servants can be (and are) required to master multiple languages,' but such a requirement would be “plainly incompatible with the workings of a democratic system” if placed on elected representatives (Gazzola and Grin 2013, 105). This may seem obvious, but a number of commentators have raised the possibility of a language requirement, noting that it would not be much different from expecting elected representatives to be literate (e.g., van Els 2001, 2005; Wright 2007). It is not, however, actually a requirement for elected representatives to be able to read or write (as far as I am aware), and such a requirement would be very different from the expectation or preference that they are. While it is perhaps unlikely that an illiterate person would seek elected office, be nominated, and ultimately elected, it would be hugely problematic to implement a rule barring this possibility.

Not making language services available to the elected representatives of EU citizens would also undermine people’s equality of rights and, given current levels of language proficiency and the distribution of language skills across societal groups, come at the expense of the less educated and socioeconomically disadvantaged (Gazzola 2016b). Citizens would be prevented from selecting candidates, independent of their language proficiency, who they feel would best represent their interest, and they themselves might be prevented from pursuing one of the most important democratic rights: to seek elected office. The quality of democracy would also suffer in the absence of language services, because lower-quality candidates with foreign language skills would crowd out more qualified competitors, and because only experts with sufficient foreign language proficiency would be able to bring their knowledge to bear on policy decisions in EU meetings and hearings.

It is possible that, over time, foreign language competence among Europeans improves such that the capacity to operate in multiple languages comes to be an expectation or that parties only nominate candidates with sufficient language proficiency to operate effectively in what would be an increasingly monolingual environment (centered, in all likelihood on English) (De Swaan 2007, 17; see also chapter 3). If this occurred organically, it might still entail some de facto exclusion, but this exclusion would not be discrimination rooted in formal rules or the absence of language services—a crucial difference indeed.

The third major reason for EU multilingualism is the principle of equal authenticity of language versions of EU law, meaning that all language versions of EU legislation are equally “legally valid” (Wagner, Bech, and Martinez 2014, 8).’ Put differently, there is no “original” version of legislative text that prevails when two or more versions are incompatible or in conflict with each other.’ This principle dates all the way back to the European Coal and Steel Community in the early 1950s, when France argued that, for the sake of legal coherence, there ought to be a single reference language and that this one authentic language should be French.* This was rejected in favor of the position advocated by Germany, which maintained that legal certainty could only be assured if legal acts were authentic in all official languages (Van der Jeught 2015, 57).

The principle of equal authenticity means that it is of crucial importance that all language versions of EU legislation are drafted so that they have equivalent effect across all member states. It also means that—formally— there are no “translations” of EU legal acts, because an equally authentic language version cannot be a translation of an original text. Paradoxically, this is the case even if one or more language versions of a legal act were, in fact, prepared by translation (Pym 2000; Doczekalska 2009; Gibova 2009; Kjær 2015). This is apparently contradictory, but it legally precludes the possibility of distinguishing between an original source text and a translated target text (Doczekalska 2009), because once a legal document has been finalized, whatever version served as the “original” throughout the drafting process ceases to exist as such (Gibova 2009). And so the EU features a lawmaking process that fundamentally depends on legal translation, but whose results cannot be called or treated as translations (Doczekalska 2009, 132). The principle of equivalence, Anthony Pym writes, is thus “is a legal fiction necessary for multilingual EU legislation to work” (Pym 2000, 6).

The reasons why it would be problematic to have an authentic text in only one or some official languages are the two most important doctrines of EU law, those of supremacy and direct effect. The principle of the supremacy of EU law establishes that EU law prevails over national law if the two conflict, while the principle of direct effect bestows rights upon EU citizens based on the EU’s treaties and secondary law, which can be invoked by EU citizens in their own language before their national and EU courts. These two core provisions of EU law, however, which are among the most important features differentiating the EU from other international organizations, could not operate effectively if all language versions of EU legislation were not equally authentic. Multilingualism is thus a necessary corollary of the two most important provisions of EU law (Athanassiou 2006, 6), because its equal application would be violated if citizens were prevented from understanding and thus invoking the rights conferred upon them through EU law (Saréevié 2013, 4; Paunio 2007, 396). It would be similarly problematic if EU legislation were only authentic in some language(s), and thus foreign and potentially incomprehensible to some EU citizens, since it prevails over national law. Finally, it is imperative that the member state institutions charged with transposing (i.e., writing into national legal code), implementing, and adjudicating EU legislation— namely national legislatures, executives, and judiciaries—be able read and understand them in their own languages. This is essential for the sake of the stated EU objective of harmonizing laws and to ensure that EU legislation confers the same rights to EU citizens across the member states in practice. It is thus no overstatement that multilingualism ensures the equality of all EU citizens before the law (Biel 2007, 145; McAuliffe 2012, 201), and thus the rule of law itself (Athanassiou 2006, 7).

In order for this to be true, however, language versions cannot be divergent, because the equally authentic language versions of EU legislation must be uniformly interpreted and applied in the member states (Baaij 2012a).

This is an important and difficult challenge and one that, if not met, has the potential to undermine the very principle it is supposed to safeguard. It is also an ideal that can be approximated at best, as “both lawyers and linguists are quick to concede that it is impossible to produce parallel texts of a single instrument which have the same meaning” (Saréevié 2013, 8); legal translation is a necessarily imperfect process (Šarčević 2012a; van Calster 1997). The system of equally authentic language versions thus relies on the inherently imperfect instrument of legal translation, which inevitably involves some “slippage” in the uniform application of EU law and thus the equality of all EU citizens before the law. Since the alternative would be a system of only one or some authentic language versions, however, and thus a system that would be fundamentally flawed with regard to that equality, the status quo is preferable as long as the equally authentic language versions are sufficiently similar so that major and consequential discrepancies are the exception rather than the rule, which is the case nowadays.

It is, of course, a reality that a term may be interpreted differently in various member states, either because of ambiguities across language versions or because legal concepts have divergent meanings in different national legal systems (De Groot and Laer 2006).° It is also the case that some, perhaps even many, such discrepancies have the potential to be of consequence but simply go unnoticed (Ginsburgh and Weber 2011, 175). But the emphasis on divergences that are actually consequential is important here because, as Robertson explains, it may well be that the precise formulations in different language versions differ in the abstract but result in the same application in practice (Robertson 2012b). For this reason, and also because every legal system involves some degree of uncertainty and potential for legal action that is not actually pursued, equating evidence of imperfection with evidence of dysfunction would be misguided.

When ambiguity or uncertainty do become consequential, the Court of Justice may be asked to interpret the law. One well-known example is the case of Stauder v City of Ulm (1969), in which the German pensioner Erich Stauder objected to his name being revealed to merchants on coupons for cheap butter that were to be distributed as welfare benefits to poor citizens. The distribution of the butter was part of an effort by the Commission to help reduce surplus stocks that had accumulated as a side effect of the Common Agricultural Policy, and butter was to be sold at lower prices to the needy through normal retail outlets (Brown 1981). To avoid abuse, however, eligibility was limited to recipients of certain welfare payments in each member state. The Commission’s decision was issued in the (then four) official languages and provided, in the German and Dutch versions, that butter was to be made available in exchange for a coupon “issued in their names” (German: “auf ihren Namen ausgestellten Gutschein”; Dutch: “op naam gestelde bon”), while the French and Italian versions merely required an individualized coupon (French: “bon individualisé”; Italian: “buono individualizzato”) (Van der Jeught 2015, 127-28). Mr. Stauder felt that his dignity was violated by his identification, by name, as somebody eligible for the benefit. The Court decided that the intention of the rule could not have been to impose a stricter obligation on one member state than on another, and that the French and Italian versions should therefore prevail (for more details, see Brown 1981; Van der Jeught 2015).°

Language-related cases before the Court illustrate, on the one hand, that the existence of different language version of legislation can result in uncertainty as to its intent; in the absence of an “original” text, it is difficult to establish the “will or intention of the legislator” (Bengoetxea 2016, 104). On the other hand, as Solan argues, the different language versions can assist in the interpretation of the law, because their comparison can help identify a law’s likely intent (although it is not clear that the Court in fact does this consistently). The “proliferation of language versions,” he writes, “appears to add to the likelihood that the court will get a case right, where getting it right means issuing a judgment that is more likely to further the purpose behind the law, and which is consistent with the intent of the enacting legislature” (Solan 2009, 52; see also Piris 2005).

In sum, EU multilingualism exists for important symbolic, political, and legal reasons. It aids the maintenance and building of a political community while safeguarding cultural, social, and political diversity; it helps ensure democratic representation; and it constitutes a cornerstone of the

EU’s legal order.

Foreign Language Use in the EU Institutions

English is by far the most dominant language used inside the EU institutions today, with the exception of the Court of Justice, which continues to rely primarily on French. Deliberation and negotiation, both formal and informal, most commonly take place in English, whether in the preparatory phrases of the legislative process, in intra-institutional bargaining, or in trilogue meetings involving representatives from the Council, the EP, and the Commission.’ English is also the most frequently written language in the EU institutions. Draft legislation is usually written in English, which puts the rest of the policymaking process on a mostly English track, even if translations are available. Most amendments are tabled (i.e., introduced)

in English and, as is the case with other policy-relevant documents, most quickly translated into English if they are not. Legislation is almost always finalized in English, both inside the institutions and interinstitutionally. One EP official estimates that “95 percent of the time, it is the English version that is approved” (#4).

Since English is the language most commonly spoken and written by participants in EU policymaking, the focus of this book is disproportionately on the role of English. The dominance and increasing role of English over time do not mean, however, that other languages are irrelevant for communication inside the EU institutions. French, in particular, continues to be an important and frequently used language. It may have lost its status as the primary vehicular language in the EU, and French politicians and officials may be forced to rely on English more than before, but having French as a native or foreign language continues to be of value. There are a number of reasons for this:

* meetings, especially in the Commission (#55), often rely on a mix of English and French (#11, 55);

e French remains a language of administration inside the institutions (#38), again especially in the Commission (#55);

* native French speakers tend to be less willing or able than speakers of other languages to use English (#20);

* many emails directed at EU officials are in French (#12);

e French terms have been incorporated into EU jargon, so even when somebody speaks English it may be “tricky to understand if you don’t know any French” (#61);

e French continues to be a language that many EU actors have in common, so numerous direct interactions inside the EU institutions still take place in French.

French is also the second most frequently used drafting language for written documents, even if it trails far behind English. This can be a challenge for EU actors who do not have a strong command of the language (#59), the number of whom has increased following the eastward enlargement of the EU. The same is true when, for example, a country holding the rotating Council presidency insists on operating in French, as was the case for Luxembourg during the second half of 2015 (#20). Because of the difficulties this may cause, some member states have a strong preference for English; one even has an informal guideline in place to promote the use of English over French inside the EU institutions.*

While less common today than it used to be (#47, 67), many EU actors capably mix English and French, especially in the Commission (Ban 2013). At the extreme, such “code switching” results in “Franglais,” as one Member of the EP (MEP) recalled; people “say things like ‘so, I saw this document, je dois dire que je ne comprends pas le premier paragraphe, but I agree with the purpose of the document. Mais le titre, il faut changer’” (#47). Alternatively, an English speaker may address a French speaker in English and receive a reply in French. Code switching happens for a variety of reasons and is ultimately dependent on the speakers’ language competence and the ability of others to understand them (Wodak, Krzyzanowski, and Forchtner 2012, 162). My respondents described a variety of reasons for why and when it occurs, including when actors use somebody else’s language as a courtesy (#3, 55, 56), when something is better said in another language (#14, 70), when responding to a question asked in another language (#3, 14), or because a speaker who said something important in her mother tongue repeats it in English to ensure everyone understood (#13, 70, 72, 73). Code switching may also occur when a general discussion gives way to the consideration of particular amendments that are only available in another language (#65), or when technical discussions involve particular terminology in English, for example (#70). Finally, code switching may simply reflect a speaker’s mood. One long-term EP official recalled that when his German superior “was in a good mood he spoke to me in English, and when he was in a bad mood he shouted at me in German” (#79).

Because of the continued importance of French, member states today still have an incentive to send representatives to Brussels “who at least have enough knowledge of French to be able to understand it . . . [even if] they will not be able to express themselves in French” (#20). Yet French is still losing ground overall, even in the Commission, as a high-ranking official in that institution whose native language is French acknowledges. He described that in one recurring high-level meeting, “I’m the only one who speaks French from time to time. The only one. And if I speak French [some of the other high-ranking officials] will reply to me in English, although they speak perfect French. . . . Yesterday [we had] the first meeting without the [British], and still we all speak English” (#55).

German, in contrast to English and French, does not play a major role as a vehicular language, although it is recognized across all institutions as the EU’s “third language.” It derives this status, first, from Germany’s position as Europe’s political and economic powerhouse; second, the size of the German-speaking population in the EU (in Germany and beyond); and, third, the resulting size of German-speaking delegations, especially in the

Council and the EP. Moreover, German serves as the language of procedure in the Court of Justice particularly often, which makes it “an extremely important language” in the Court (#24) and when dealing with legal texts more generally (#56). Nevertheless, German is not frequently spoken in meetings in the EU institutions, except by some German speakers when interpretation is available, and it does not come even close to rivaling English or French as a shared language between EU actors (#11, 12).

The use of other languages inside the institutions largely depends on the size of national delegations representing various language communities. The size (and also the power) of the member states thus matters in determining the relative importance of a national language, but it does not translate one-to-one. One EP staffer, for example, suggests that “Slovakia is less important than Luxembourg because [Luxembourgers] speak German, English, and French.”!° The other main factor that determines the use of languages other than English, French, and German is the particular configuration of participants in a given meeting or interaction. One Council official, for example, emphasized that “a lot of people speak Italian, you’d be surprised” (#58). Another one described in more detail how we participated in drafting sessions during the Italian Presidency. . . . And it happened, it just happened this way, that the Commission representatives, four of them, were all Italians. And of course the Italian representatives, also three or four. . . . [And the representative from the] Council Secretariat, she was French, but she understands Italian, so she replied in French but she assumed all the information in Italian. .. . So 70 percent of the drafting session was Italian, while the text was in English on the screen. (#61)

In an attempt to generalize, one official described the use of languages in the Council as “the big three plus an opportunity language,” by which she meant a fourth language that is the national language of the Council presidency or a function of the policy area under consideration: “it may be that in Fisheries sometimes Spanish is an easy language between various people, there are Spanish people in the Commission, Spanish people here, and so on.... That would be a matter of context” (#58).

There is some pushback against the dominance of English inside the institutions, but it is rare. The EP’s former Secretary General, Klaus Welle, attempted “an intensive push” for greater use of German, according to one EP official; this proved futile, however, because “you cannot add a third [vehicular language] and you will not replace English or French, that’s not realistic” (#12). Individual actors might also attempt to counteract the prevailing use of English. One German MEP thus insists on “speaking in German 90 percent of the time, consciously so, because I want to force the Brits to put on their headsets” (#45).

But the issue seems to have become less contested over time, at least according to one high-level Council official who has been serving in the institutions for several decades, who said that from time to time the Spaniards are sort of getting a bit upset about it, because after all Spanish is a world language. But . . . within Europe it’s smaller. . . . It’s a status issue, it’s an issue of pride... . [But] we used to have very different debates about language, it has calmed down, more people are a bit more relaxed about it. (#52)

The main focus and concern for member states when it comes to the status of their language in the EU seems to be relative to “peer” languages. Sweden, for example, does not try to have Swedish rival English, French, or German, but merely wants to ensure that their language is “not discriminated against in relation to other comparable languages such as Danish, Finnish, Greek, and Portuguese” (#74).

Some EU institutions are, at times, described as more or less multilingual than others. Van der Jeught (2015), for example, maintains that the EP and the Court of Justice are the most multilingual institutions. This is certainly correct when we consider the extent to which the language regimes of the institutions are (formally or informally) more or less restricted, since the representational and legal dimensions of language (which, respectively, provide the basis for popular participation in EU politics and ensure the uniform application of EU law) are of greater importance in the EP and the Court of Justice and tilt their language regimes in favor of greater equality between languages. As a result, it is also correct that the EP and the Court of Justice are more multilingual in the extent to which different languages are used regularly in these institution. The EP, in particular, sees much greater language variation than the other institutions in its day-to-day operations. This, however, masks the extent to which individuals inside the different institutions rely on multiple languages in their day-today work (Wodak 2014, 141). By that measure, the Commission is more multilingual than the EP because Commission officials more frequently rely on multiple languages in their work and use them with greater facility. This is in part because foreign language proficiency is not a formal job requirement for lawmakers and turnover is much greater than it is for officials working in the Commission, but also because the representational role played by the directly elected delegates of the European people entails strong incentives to use native languages when communication is directed at national constituencies. In what follows, I briefly outline specific patterns in language use in each of the EU’s main institutions, to qualify or add nuance to the general tendencies in language use just described."

The Commission

The Commission’s procedural languages are English, French, and German, as the institution indicated in proceedings before the EU Ombudsman in 2007. This was a rare formal acknowledgment of its internal language practices, which are not spelled out in the institution’s Rules of Procedure (Van der Jeught 2015, 136-7). The main languages used in the College of Commissioners are English and French, as Kruse and Ammon found in their investigation of 996 oral contributions by all 27 EU commissioners over the course of one year: 55.92 percent of interventions were in English and 16.67 percent were in French. Some commissioners consistently spoke only their native language, however, which explains fairly high proportions in the use of Italian (6.1), Czech (5.7), Greek (5.5), German (5.22), and Spanish (4.58) (Kruse and Ammon 2013, 167).

Languages other than the procedural ones are only used when it is legally required, which are instances when all languages are equally authentic (i.e., legal texts or instruments of general application), when the addressee of a communication is entitled to it in another language, or on formal occasions and in public meetings with participants from the member states (Athanassiou 2006, 20). Overall, English has established itself as the main procedural language in the Commission, a status that used to be held by French. The use of German lags far behind both English and French; it is used when a sufficient number of participants in a given conversation share the language, which is also the case (on occasion) for some other languages like Spanish or Polish (Krzyzanowski 2014, 112). The fact that Commission officials speak multiple languages allows for such flexibility in language choice and the efficiency gains that come with it; as one Council official observed, the reality in the Commission is that the officials “always speak in languages other than their own” (#76). Commission officials are de facto multilingual, as both their recruitment and promotion in the institution depend in part on their language repertoires (Wodak 2013, 96).

The tide turned decisively against French as the main internal language of the Commission after the 2004 eastward enlargement, when few of the new officials that entered the institution spoke French, even though their language repertoires on average exceeded those of their colleagues from the old member states (Ban 2009). Today, one can still reasonably assume that everyone in the Commission has at least passive knowledge of both French and English and that most can comfortably speak either (Ban 2009). Moreover, some enclaves remain where French plays a more active role, such as the Commission’s Legal Service and Secretariat-General (Van der Jeught 2015, 138). French also still dominates in a small number of Directorates-General (DGs), such as DG Human Resources and Security (Gravier and Lundquist 2016, 80). The EU’s quintessential supranational institution has thus adopted a rather restricted language regime, which reflects the Commission’s “composition and operational requirements” and is “dictated by the imperatives of speed and efficiency” (Athanassiou 2006, 20). At the same time, however, new requirements that Commission staff members must master a third language prior to their first promotion also signal the institution’s continued commitment to linguistic diversity (Ban 2013, 224).

The Council of the EU (Council of Ministers)

‘The Council’s language regime is, on the surface, centered on the principle of the equality of all official languages, which stands to reason given that it is the institution in which the member states and their interests are represented and which adopts legislation, jointly with the EP, that is equally authentic in all language versions (Athanassiou 2006, 18). In practice, however, the Council of the EU relies on a full language regime only in its formal meetings and in a very limited number of preparatory meetings financed entirely from the Council budget."

Language services are critical at this level because national politicians who come to Brussels to take final decisions on legislative matters often do not possess the foreign language skills to negotiate complex legislation (#16, 72); moreover, as delegates of their member states, they have an incentive to use their national language because they often address a national constituency, especially in meetings that are broadcast (Van der Jeught 2015, 135). Many of the detailed discussions, however, have already happened at the working group or Coreper (Committee of Permanent Representatives) levels, meaning that “language capacity and interpretation do not matter that much” because “in meetings with ministers in the Council there usually isn’t much left to decide” (#59).