The Limitations of a Multilingual Legal System

Abstract

The Court of Justice of the European Union (CJEU) and the way in which it works can be seen as a microcosm of how a multilingual, multicultural supranationalisation process and legal order can be constructed—the Court is a microcosm of the EU as a whole and in particular of EU law. The multilingual jurisprudence produced by the CJEU is necessarily shaped by the dynamics within that institution and by the ‘cultural compromises’ at play in the production process. The resultant texts, which make up that jurisprudence, are hybrid in nature and inherently approximate. On the one hand, that approximation can lead to discrepancies between language versions of the Court’s case law and thus jeopardise the uniform application of EU law. On the other hand, that approximation and hybridity define EU law as a distinct, supranational legal order. This paper analyses the operation of the CJEU and considers whether a linguistic cultural compromise exists within that institution which exercises a formative influence on the character of its ‘output’—i.e. its jurisprudence—and what that may mean for our understanding of the development of EU law.

1 Introduction

The Court of Justice of the European Union (CJEU), once described as being ‘out of sight and out of mind by virtue of its location in the fairytale Grand Duchy of Luxembourg and the benign neglect of the media’ [14], has come increasingly into the spotlight in recent years. There is an extensive literature on the CJEU and the main bodies of literature concerning that Court (notably legal and political science literatures) focus on its role in developing the EU legal order. Legal literature is generally concerned with analysing the legal logic behind the CJEU’s rulings and discussing how that Court can affect policy changes in the EU, insofar as practice may have to change to comply with a particular ruling. Political science literature, on the other hand, is interested in ‘judicial politics’, the policy dynamics that can be inferred from the Court’s decisions and in examining the political context and consequences of those decisions. However, each of these bodies of literature remains predominantly focused on the decisions of the Court and on judicial reasoning/investigating the reasons or motivation behind those decisions. Much has been written on why the Court makes certain decisions and the effects of those decisions. This paper, however, focuses on how that Court’s multilingual jurisprudence is produced and the implications of the process of such production on the EU legal landscape.

All students of EU law will be familiar with the jurisdiction and architecture of the EU judicial order, including the procedure to be followed in various direct and indirect actions before the Court of Justice. For many scholars, that remains their understanding of, or indeed interest in, ‘how the Court of Justice works’.¹ However, there is far more involved in the working of the Court than is elaborated on in most treatises on that institution.

In recent years, anthropologists have shown great interest in EU institutions—in particular the European Parliament and Commission—and have carried out various studies of those institutions. Such studies, based on periods of fieldwork research in the services of the Parliament and Commission, provide a valuable insight into the workings of those institutions and it is interesting to note the significance of culture, identity and language in the policies, actions and day-to-day life of the institutions [1, 2, 5, 6, 8]. In spite of the fact that EU institutions are staffed by individuals from member states with diverse social and educational backgrounds, languages and cultures, each institution is, by its very nature “obliged to express itself with a single voice” [6], presupposing that it has resolved any internal conflicts deriving from technical considerations and differing political approaches to similar phenomena [8]. The question, for anthropologists, is: how exactly do the EU institutions resolve those conflicts? In their work, Abélès and Bellier make the point that process necessarily affects output. In the context of the European Parliament and Commission the ‘process’ involves a ‘cultural compromise’ through which European civil servants are able to work together in the unique hybrid environment of those institutions. The ‘output’ necessarily affected by that cultural compromise relates to the resulting ‘culture of compromise’ visible in the policies and actions of those institutions. Those anthropological studies also note the development of a mixed and hybrid ‘eurolanguage’ within the institutions, which is the linguistic manifestation of the cultural compromise by which the institution works. Bellier points out that such ‘eurolanguage’ functions perfectly well within the institution but can create problems when the Commission engages in discourse with the outside world [8]. Bellier analyses the development of a ‘eurolanguage’ within the European Commission, however the texts produced by the CJEU also have to resonate comprehensively outside of that institution in terms of an EU legal language that is applicable throughout all 27 member states.² In light of this anthropological literature, it is reasonable to presume that the process behind the production of the CJEU’s multilingual jurisprudence could have implications for the development of EU law. It follows, therefore, that understanding the situational factors of, and compromises involved in, the production of such jurisprudence could aid our understanding of EU law.

2 Methodology

Premised on the notion that the dynamics within the CJEU, and the perceptions of those who work there of their own professional environment, shape the culture of that institution, this paper focuses first on the actors at the heart of the production process and investigates some of the cultural dynamics at play in that process. In order to understand and analyse such institutional culture, one must understand the priorities and preoccupations of those who work there. This paper is based on fieldwork research, participant observation and interviews, carried out at the CJEU between 2002 and 2011. Participant observation involved observing the interactions among lawyer-linguists and between those lawyer-linguists and members of the Court and their référendaires,³ both in professional contexts such as meetings, seminars etc. and more informal contexts such as Court social functions, coffee breaks, lunchtimes etc.; engaging to some extent in those activities; interacting with participants socially and identifying and developing relationships with key stakeholders and gatekeepers. To overcome any inherent bias in the data obtained through participant observation, the findings were triangulated with existing literature concerning the CJEU, concepts developed in translation theory literature as well as with the findings of comparable studies carried out in other EU institutions.⁴ The interview sample consisted of 78 interviewees in total (56 lawyer-linguists; 5 judges; 3 Advocates General and 14 référendaires).⁵

While it is generally accepted that language cannot be divorced from culture; and the cultural or multicultural, aspect of an institution will necessarily affect its output [7], the present paper is concerned with the multilingual, opposed to the multicultural aspects of the Court’s jurisprudence. The reason for such choice of focus lies in the nature of the Court’s ‘output’: the CJEU aims to produce statements of law that have the same effect in every language in which they are published and through such statements to ensure the uniform application of EU law. Yet those statements of law consist primarily of collegiate judgments drafted by jurists in a language that is generally not their mother tongue. Moreover, those statements of law undergo many permutations of translation into and out of up to 23 different languages and they are necessarily shaped by the way in which the Court functions as a multilingual, multicultural organisation and the final ‘authentic’ judgments, as presented to the outside world, are, for the most part, translations.

3 Language at the CJEU

The importance of the multilingual aspect of the CJEU’s work has come increasingly to the fore in recent years, particularly since the ‘mega-enlargement’ of 2004 [15]. Indeed, the ‘General Presentation’ of the Court, as stated on its own website⁶ consists of only two paragraphs, the first setting out the ‘mission’ of that Court and the second stating:

As each Member State has its own language and specific legal system, the Court of Justice of the European Union is a multilingual institution. Its language arrangements have no equivalent in any other court in the world, since each of the official languages of the European Union can be the language of a case. The Court is required to observe the principle of multilingualism in full, because of the need to communicate with the parties in the language of proceedings and to ensure that its case-law is disseminated throughout the Member States.⁷

The role of language and translation in proceedings before the Court is also clearly set out in its Rules of Procedure⁸ and other literature.⁹ Article 42 of the Rules of Procedure states:

The Court shall set up a language service staffed by experts with adequate legal training and a thorough knowledge of several official languages of the European Union.¹⁰

That ‘translating service’ (Translation Directorate) makes up almost half of the Court’s total staff and is the largest ‘service’ of that institution.¹¹ The lawyer-linguists employed in the translation directorate are responsible for translation into French (the working language of the Court) of documents lodged by parties and interveners to a case and the subsequent translation into all of the EU official languages of judgments and orders of the Court (as well as, where relevant, advocates general’s opinions).

Unlike the other EU institutions, the CJEU operates using a single internal working language—French. For every action before the CJEU there is a language of procedure (which can be any one of the 23 official EU languages),¹² which must be used in the written submissions or observations submitted for all oral submissions in the action. The language of procedure of the case must also be used by the Court in any correspondence, report, or decision addressed to the parties in the case. Only the texts in the language of procedure are authentic, which means that, in most cases, the ‘authentic’ version of a judgment will be a translation of the original judgment drafted and deliberated on in French [19]. It is clear therefore, that translation plays a significant role in the working of the CJEU. However, translation is not the only language issue in the production of the multilingual jurisprudence of that court. From submission of the initial application through to delivery of the final judgment the role of language and the impact that it has on the process of production of that jurisprudence is significant.

4 The Process of Producing a Text: A Linguistic Cultural Compromise?

Producing the CJEU’s multilingual jurisprudence is a complex process: case files (including a draft judgment produced prior to deliberations) are prepared by judges and référendaires together; Advocates General and their référendaires prepare the opinion (where relevant) and finally the relevant chamber of judges prepare the final collegiate judgment in secret deliberations. In addition to the, often complicated, legal reasoning and application of EU law throughout that process, those actors are drafting the various documents in a language which for most of them is not their mother tongue [16, 18]. Those documents (judgments, orders, AGs’ opinions)—the ‘output’ of the Court—are then translated into the other 22 official EU languages, and, in the case of judgments, more often than not the authentic version of that judgment will be a translation. For an overview of the full process see Figs. 1, 2 and 3.

Fig. 1

References for a preliminary ruling

Fig. 2

Judgments

Fig. 3

Translation of documents at the Court of Justice

From the participant observation and interviews carried out for the purpose of the present paper, it is clear that, similarly to the ‘cultural compromise’ evident in the Commission and Parliament (see supra), there is a similar linguistic cultural compromise within the CJEU, which exercises a formative influence on the working of the institution and the character of its ‘output’ (i.e. its jurisprudence). It appears, in fact, that there are two types of linguistic cultural compromise at play in the working of the Court.

First, there is the linguistic cultural compromise involved in producing the Court’s case law. The collegiate judgments of the CJEU are, by their very nature, compromise documents. The wording chosen in a (French language) judgment has very often been the subject of painstaking deliberations. However, because the deliberations are secret it is impossible for anyone other than the judges involved to know where compromises lie in the text. This has implications both for the subsequent translation of such judgments and also for the drafting of future judgments (insofar as those drafting subsequent judgments will not know if the wording of certain sections of the text in question was a compromise and therefore important to retain in any reference to the text). Furthermore, the case law of the CJEU is shaped by the language in which it is drafted—i.e. French.

Although most référendaires claim to work solely in French, when interviewed, the majority actually admitted to drafting ‘half in [their own mother tongue] and half in French’:

I tend to translate what I want to say into French instead of really working in French (référendaire, interviewee’s emphasis)

Part of the linguistic cultural compromise in the production of the jurisprudence of the CJEU involves the amalgamation of legal reasoning and method from many European legal orders due to the fact that those producing that jurisprudence are working in a language which is not their mother tongue or that through which they may have learned EU law. That particular ‘cultural compromise’ necessarily affects the Court’s ‘output’:

While it is more difficult for someone to draft in a language that isn’t their own, it can also be helpful since writing in a foreign language ‘formalises’ the text;

While it can be difficult to find terms in a foreign language that meet your exact thinking, working in a foreign language can also help you to find answers to legal problems that you wouldn’t have found in your own language.

Furthermore, because French is rarely the mother tongue of those drafting that case law, there is a tendency to repeat expressions and to ‘cut and paste’ from previous case law or source documents. Together with the difficulties of manipulating a language that is not one’s own, the result is often a stilted and awkward text. In addition, those drafting the case law of the CJEU are constrained in their use of language and style of writing (owing to pressures of technology and in order to reinforce the rule of law being developed). The resulting texts, as this author has previous submitted, are hybrid in nature and this hybridity itself shapes the development of EU law:¹³

It is often difficult to say exactly what you want to say in a judgment… often the Court will want to say X but in the very rigid and hybrid language of the Court that is used in the judgments you have to get around to X by saying that it is not Y! …such use of language necessarily has implications for the way in which the case law develops… (référendaire).

Secondly, the case law of the CJEU is ‘filtered out’ through the linguistic cultural compromises involved in translation. Translation itself is a ‘linguistic cultural compromise’ and all translation, including legal translation, involves an element of approximation. The difficulties of translating the, already hybrid, case law of the CJEU have been well documented by the present author [15–20]. While the role of the lawyer-linguists who produce the translations may be “the perfect synthesis of a lawyer and a linguist”, there are nonetheless many difficulties which arise due to the ambiguity and approximation inherent in translation [16, 19, 20]. In addition to those classic problems of translation (and legal translation in particular) such as ambiguity, translation of ‘untranslatable’ legal concepts, the effects of translated legal texts etc., translation at the CJEU is also affected by the role perceptions of lawyer-linguists who struggle to balance a dual professional identity. The loyalties of those lawyer-linguists are divided between ‘the law’ and ‘language’. On the one hand, they have a responsibility as lawyers to ensure that their translations are legally sound and represent the statement of law that the Court wishes to make. On the other hand, they are linguists, and as such must accept the inherent approximation involved in translation. The lawyer-linguists work at the interface of law and language, dealing with the relationship between those two concepts on a daily basis. The necessary compromise resulting from the struggle to reconcile the notions of ‘law’ and ‘translation’ is reflected in the process whereby the case law of the Court is ‘filtered’ out to the wider EU.¹⁴

5 Issues of Translation

Everybody involved in producing the case law of the CJEU acknowledges that translation necessarily involves an element of approximation and that there are also frequent discrepancies, both avoidable and unavoidable, between translated documents. A number of judges, interviewed for the purposes of the present paper, pointed out that, as a result of the significant time pressures on the French translation division at the Court, the quality of procedural documents into French varies; and admitted that:

…if I feel that a document has been rushed through translation, and where I understand the language of the original document, I tend to work from that original.

One judge commented that in some cases, where a judge has been working from the original documents and the référendaire has been working from the French translations of those documents:

…[the judge and référendaire] will end up understanding different things about the case in question – it is almost as though [they] are reading two slightly different cases…

Another issue that highlights the members’ awareness of approximation and discrepancies in translation is their reaction to the introduction of pivot translation at the Court in May 2004. That system is actually a mixed translation system—whenever possible, direct translation is used instead of translation through a ‘pivot language’. There are five pivot languages in use at the CJEU: French, English, German, Spanish and Italian. Those five languages are ‘partnered’ with a number of the ‘post-2004’ languages. Documents are translated into the ‘partner’ or pivot language and from that translation can be further translated into all other EU official languages¹⁵ (for an overview of the pivot translation system at the CJEU see McAuliffe 2008).

While the members of the CJEU accept that some form of pivot or inter-lingual translation is necessary in order to successfully produce case law in 23 languages, those members interviewed for the purposes of the present paper remained somewhat uneasy about the use of that system of translation by the Court. A number of those interviewed described pivot translation as “dangerous” and claimed that it “exacerbates any problems already existing within the translation system”. Thus, most members of the Court do have a sense that “translation is an approximation”.

All of those interviewed commented that the Court’s judgments are “shaped by the fact that the working language at the Court is French” and those judgments are drafted in French. Most lawyer-linguists, référendaires and even judges feel that the Court’s judgments are too formulaic, stilted and pompous in style. Interestingly, most of the judges interviewed blamed this on “the translators adhering too literally to the French originals” and not on the fact that those documents are drafted, for the most part, by non-francophones, or that they are quite often the product of a compromise in deliberations (during which the precise wording of a particular phrase may be discussed for days or even weeks), or indeed that the nature of those judgments makes it almost impossible for them to be drafted in a free-flowing or easily readable style.¹⁶

The compromise which results from the reconciliation of two sets of norms (of translation and law) by the Court’s lawyer-linguists is thus widely acknowledged and accepted by the small legal community within that Court. However, as has been shown here, the linguistic cultural compromise in the production of the Court’s case law is inherent in all stages of that production process—not just the translation stage. As mentioned above, such a hybrid eurolanguage does not tend to cause difficulty within the institution itself, but can create problems when that institution engages in discourse with the outside world. Such ‘engagement’ on the part of the CJEU involves the interpretations and application of EU law and consequently the development of a legal order based on the principle of uniformity of that law across all member states.

6 The Problems: Some Examples

Language versions of a judgment of the [CJEU] cannot be identical because [those drafting and translating them] have so many different ways of working… the vast majority of those who actually use the judgments are not aware of this, which, in a supranational, supposedly uniform legal order, is hardly an ideal situation (lawyer-linguist, interviewee’s emphasis).

Some approximations between different language versions of CJEU judgments and other EU legal instruments simply cannot be avoided. For example, the translation of the English “cartel”, into French as either “entente”, “accord” or even “cartel”, none of which have exactly the same meaning as “cartel” does in English. One lawyer-linguist interviewed used this example to highlight the relationship between European law, language and translation:

…as well as the linguistic problem there is also a legal problem: we all accept that cartels are illegal – but since ‘cartel’ doesn’t translate exactly from on language to another [in this case, English to French] how do we know what behaviour is illegal under European law? In [the French language version of Article 104 TFEU] words such as ‘entente’ and ‘accord’ are used, and [from that language] it seems that even discussions and negotiations can be illegal: a concurrence of wills to act on the market in a specific way will be contrary to the provisions on competition in the [TFEU]. This is much more punitive than would be the case in many member state legal systems where a practice would have to be much further down the line of a cartel to make it illegal…¹⁷

Another example raised by a large number of those interviewed for the present paper is the unavoidable approximation involved in rendering the legal notion of a “trust” in a language other than English. Even the word “contract” in English does not fully correspond to “contrat” in French or “vertrag” in German since (among other distinctions) a “contract” in English relates to an agreement based on the concept of “consideration” and “contrat” in French or “vertrag” in German do not imply the same notion of consideration.¹⁸

According to all of the lawyer-linguists and the majority of référendaires interviewed for the purposes of the present paper, the process of producing case law at the CJEU can potentially lead to problems of a legal nature:

Where an original [judgment] is not readily comprehensible… and lawyer-linguists working to a tight deadline do not want to lose time in lengthy discussions with the cabinet concerned, there is a risk that [the lawyer-linguists in question] will adopt their own individual solutions. …This can lead to divergence among language versions [of judgments] and potentially to the misapplication of [EU] law.

One such example can be found in the order of the General Court of 2 June 1997 in Case T-60/96¹⁹:

This case concerned an application for the annulment of certain Commission decisions refusing authorisation by a number of EU member states to take protective measures with regard to Spanish pharmaceutical products. The language of the case (i.e. the authentic version of the order) was English. According to the normal procedure of the then Court of First Instance, the order in question was drafted in French and subsequently translated into other EU official languages. Paragraph 44 of that order referred to “un droit subjectif préexistant des titulaires des brevets en cause”. However, “un droit subjectif” is a legal concept that exists in civil law jurisdictions but not in common law jurisdictions and thus has no equivalent in English (the authentic language of the order in question). This problem was brought to the attention of the judges in the relevant chamber, who deliberated over it for a considerable period of time and eventually decided that the phrase should be changed and that, in English, it should refer simply to a “pre-existing right of the patent holder”. However, it appears that the original French language version of that order was never amended and, to this day, refers to “un droit subjectif”.

The danger in that case, according to the lawyer-linguists and référendaires interviewed, is that, since only the English language version of the order was amended, the right referred to in that order could be understood differently in member states with common law legal orders than in member states with civil law legal orders, thereby potentially jeopardising the ‘uniform application’ of EU law.²⁰

Another example of a translation problem that, according to those lawyer-linguists and référendaires, could possibly have far-reaching legal consequences, can be seen in an order of the Court of Justice regarding waste management:²¹

In Council Directive 75/442/EEC,²² the word “réemploi” is used in French, and “reuse” is used in English. That term is not defined in the directive itself but is defined in various waste-related legislation and papers on the EU waste management hierarchy as referring to a substance or object that us used again for the same purpose as that for which it was originally used. The primary meaning of the term ‘re-use’, as found in the EU waste hierarchy, is the repeated use of non-hazardous wastes, such as, for example, paper, used clothing and glass, in their original form. Waste can also be ‘re-used as part of a recovery operation, such as operation R1 in Annex II B of the Directive—“use principally as a fuel or other means to generate energy”. Case C-235/02 centred on the “reutilisation”, as fuel, of petroleum refining by-products, which are classified as hazardous waste and therefore cannot be “re-used” (except as part of a recovery operation).²³ It seems that the industries in question in that case were burning those by-products and using the energy produced from burning them as fuel-claiming, therefore, that it was a “recovery action”. In the French language version of the order in question (i.e. the language in which it was originally drafted), the term “reutilisation” is used to describe such use. Since that word is different from “réemploi” as used in the French language version of Directive 75/442, in the opinion of the person drafting the order in question there was no problem. However, while the word “reutilisation” can be translated into English as “recuperation” or “recovery”, the far more usual translation would be “reuse”, and therein lies the problem (in particular since the authentic language version of most cases before the CJEU concerning waste management is English). The “réutilisation” of the toxic waste referred to in the order is not in fact “reuse” as meaning a product being used again for its original purpose: rather, if it is considered a waste, it would be re-use as part of a recovery operation. However, if the Court were to use the word “reuse” with reference to certain toxic substances that would normally be considered hazardous, one might reasonably assume that the substances in question are not to be considered hazardous (and can therefore be disposed of or dealt with without having to conform to any special criteria under Council Directive 75/442/EEC or Council Directive 91/689/EEC of 12 December 1991 on hazardous waste).²⁴

In the order in question, “réutilisation” was, in fact, rendered in English as “further use”.²⁵ However, that particular translation resulted from the fact that the lawyer-linguist responsible for translating the order into English was an expert in waste management! That lawyer-linguist deliberately avoided using the term “reuse” because of what she regarded as the potential consequences of such use:

If a référendaire who is weak in French and doesn’t understand the technicalities of waste issues, drafts an order or judgment [in French] using words which have a very specific technical meaning, without understanding the nuances of those words; and the judgment is then translated… by a lawyer-linguist who also has no technical knowledge of the subject, toxic waste, which cannot under any circumstances be ‘re-used’ but which must be subject to either a disposal or a recovery operation, might wind up described as being ‘re-used’. That, in turn, would create a feedback loop whereby hazardous wastes – by virtue of being capable of re-use – are no longer hazardous but are merely waste. And while they are being used, they will not even be waste any longer, until further discard takes place. Whereas, if they are hazardous waste, they remain waste no matter how they are used and are subject to strict conditions of use.

There is of course, no guarantee that all waste management judgments or orders will be translated by lawyer-linguists who are technical experts in that field! All of the lawyer-linguists interviewed for this paper felt that is was “extremely likely” that “réutilisation” in French would “usually be translated as reuse” in English. One lawyer-linguist commented that:

…the use of that one little word could completely change the hierarchy of waste management in the European Union: industries could potentially bring an action claiming that the substances in question in those cases cannot be hazardous because they are being ‘reused’ within the meaning of [Council Directive 75/442/EEC].

In their book The Court of Justice of the European Communities, Neville Brown and Tom Kennedy highlight Case 131/79²⁶ as an example of approximation between different language versions of judgments of the Court [9]. In that case a preliminary ruling was sought on whether a lapse of time could render a recommendation to deport invalid under Council Directive 64/221.²⁷ The Court, in the English language version of the judgment (which was the authentic version in that case), provided a criterion in terms of whether the lapse of time “is liable to deprive” such recommendation of its validity. Brown and Kennedy point out that the French language version of that judgment (i.e. the original version drafted) was more precise, employing the words “est de nature à priver”;²⁸ and that the ambiguity of the English language version of that judgment “may have misled the Divisional Court and the Court of Appeal in their application of the ruling”.²⁹

A similar issue more recently arose in a series of cases before the Competition Appeal Tribunal in the UK³⁰ (the replica sports kit cases) in which the applicants sought to rely on the wording of the English language judgment of the General Court in Case T-25/95,³¹ which sets out the requirements for a concerted practice. Paragraph 1852 of the English language version of that judgment states:

In order to prove that there has been a concerted practice, it is not necessary to show that the competitor in question has formally undertaken, in respect of one or several others, to adopt a particular course of conduct or that the competitors have colluded over their future conduct on the market. It is sufficient that by its statement of intention the competitor should have eliminated, or at the very least, substantially reduced uncertainty as to the conduct to expect of the other on the market (my emphasis).

However, if one considers the French language version (that is, the original judgment drafted), which states:

Il suffit que, à travers sa déclaration d’intention, le concurrent ait éliminé ou à tout le moins substantiellement réduit l’incertitude quant au comportement à attendre de sa part sur le marché (my emphasis)

It seems that, for a concerted practice to exist, it is sufficient that two competitors (A and B) meet and that A receives information about B’s likely conduct; whereas the English language version implies that A has to indicate his own conduct to B.³² That case was relatively unusual in that there were nine languages of the case and therefore nine equally authentic language versions of the judgment (Danish, Dutch, English, French, German, Greek, Italian, Portuguese and Spanish). As a result, the UK Competition Appeal Tribunal compared four of those language versions of the judgment (French, German, Italian and Spanish) and concluded that they were indeed “translated slightly differently” from the English language version,³³ the correct rendering of which should be:

In order to prove that there has been a concerted practice, it is not… necessary to show that the competitor in question has formally undertaken, in respect of one or several others, to adopt a particular course of conduct or that the competitors have colluded over their future conduct on the market… It is sufficient that, by its statement of intenion, the competitor should have eliminated or at the very least, substantially reduced uncertainty as to the conduct [on the market to be expected on his part].³⁴

In other words, the UK Competition Appeal Tribunal agreed with the respondent that the French language version was correct.

It appears therefore that discrepancies and approximations in translation at the CJEU can indeed have consequences for the application of EU law at a Member State level. In the latter example there were a number of authentic language versions of the judgment in question, however, would the ruling of the UK Competition Appeal Tribunal have been any different had the English language version of Case T-25/95 been the only authentic version of that judgment? How realistic is it to expect member state courts and tribunals to compare up to 23 different language versions of a CJEU judgment before interpreting that judgment, in particular where that Court official declares only one of those language versions authentic?³⁵

7 Teleological Interpretation and a New EU Legal Language

All of those interviewed for the present paper agreed that not only is it:

…not possible to render exactly the same meaning from one legal text in one language to another legal text in another language (référendaire)

but that:

Approximation [in EU law] goes on at the legal level as well as at the linguistic level (référendaire);

There are approximations throughout the entire body of European law – legislation, case law, every part of it… (judge).

However, those interviewees were evenly divided regarding their opinions on the problems that such approximation causes, or may potentially cause.³⁶ Those who feel that such approximation is unproblematic claim that this is largely because “how you deal with that approximation is a matter of convention and common sense”:

…of course, any word will have a different semantic field in another language – 98 % will be the same, 2 % different – but when the reader has the whole context he can hone in on the exact meaning… the context can make things 100 % precise (référendaire; interviewee’s emphasis);

Approximation in translation at the Court of Justice isn’t a problem because at the end of the day everyone knows what is meant by a particular translation (référendaire).

It appears that, while it is accepted that there is approximation involved in producing the jurisprudence of the CJEU, it is also accepted that those who use that law will acknowledge that exact transpositions of concepts are impossible to achieve, yet will understand the “EU meaning” of those concepts. It seems that that notion of an ‘EU meaning’ is the key to how EU legal language copies with its multilingual nature and with the approximations that necessarily occur within it; it is quite simply, a new legal language:

There are European concepts that may be new to a reader and may not translate easily or sound right in the target language but it is for the case law then to ‘fill out’ those concepts… and thus ‘eurospeak’ has evolved (référendaire);

Concepts of EU law are expressed in an EU legal language and while some terminology might be the same as that used in individual national legal systems, when used in an EU law context, that terminology has a meaning distinct and separate from the ‘national’ meaning (judge).

That new EU legal language was, famously, expressly acknowledged by the CJEU in CILFIT, in which the Court stated that:

…even where the different language versions [of EU legislation] are entirely in accord with one another… [EU] law uses terminology which is peculiar to it. Furthermore, it must be emphasised that legal concepts do not necessarily have the same meaning in [EU] law and in the law of the various member states.³⁷

That ‘EU language’, or ‘eurospeak’, is used throughout all of the EU institutions, and those who use it are aware that it does indeed have a distinct and specific EU meaning:

Everyone [who works in the EU institutions] knows that the phrases and terminology used in EU law are really just labels that are stuck on new concepts in a new legal order (référendaire).

However, the question arises whether those who do not work within the EU institutions but who use and apply EU law are equally conscious of the fact that they are working with an independent and distinct EU legal language. As Barents points out:

It is natural that lawyers confronted with a term of [EU law] in their own language, tend to interpret this term within the framework of their own way of legal thinking, which in turn is often influenced by concepts, approaches and features which are proper to the legal culture they belong to [4].

It has been submitted that the very nature of the distinctive, synthetic language of EU law may actually help to ensure the uniform application of that law since the hybrid character of the case law of the CJEU highlights the distinct nature of EU law and the EU legal order [16]. The subsequent (re)translation of that case law into ‘hybrid’ or ‘stilted’ forms of member states’ national languages can serve to alert readers and those using such case law to the fact that they are dealing not with their own national legal language but with a new and distinct EU legal language:

…it’s like putting sleeping policemen in the text – little words which will nudge the reader into the awareness that he’s not dealing with his own national legal order (lawyer-linguist).

One judge interviewed stated, where there are discrepancies between the language versions of judgments, and in the rare cases where the CJEU may be called to decide between two or more different language versions of a judgment, that:

[the CJEU] will not necessarily look at the authentic version of the judgment (which is, after all, a translation), but will read all of the language versions and then decide which version (or versions) is to be considered correct and only at that stage provide reasoning for that decision, taking account of the legal order as it stands at that particular time.

That approach would appear to be in line with the Court’s approach when faced with issues arising as a result of linguistic differences between language versions of EU legislation, first set out in a 1969 judgment in which the Court stated that EU legislation must be interpreted

on the bases of both the real intention of its author and the aim he seeks to achieve, in the light in particular of the versions in all [official] languages.³⁸

That method of teleological interpretation advocated by the CJEU became the established method for the interpretation of EU law, and has always been linked with the language and translation issue. In a 1973 judgment, the Court reinforced that link between language/translation and its teleological method of interpretation—playing down the relevance of linguistic discrepancies between different language versions of legislation:

No argument can be drawn… from any linguistic divergences between the various language versions [of EU legislation], as the meaning of the provisions in question must be determined with respect to their objective³⁹

In the CILFIT judgment, in 1982, the Court acknowledged the importance and validity of all language versions of EU legislation, stating that teleological interpretation necessarily begins with a comparison of the different language versions of the relevant legislation:

An interpretation of a provision of [EU] law… involves a comparison of the different language versions… every provision of [EU] law must be placed in its context and interpreted in the light of the provision of [EU] law as a whole, regard being had to the objectives thereof and to its state of evolution at the date on which the provision in question is to be applied.⁴⁰

The CJEU reiterated its interpretative approach in Case 100/84:

…in the case of divergence between the language versions the provision in question must be interpreted by reference to the purpose and general scheme of the rules of which it forms a part.⁴¹

The CJEU continued to reiterate that approach in its case law throughout the 1980s 1990s and continues to the present day.⁴² As stated in the General Court judgment in joined cases T-22/02 and T-23/02:

According to settled case-law, whilst the need for a uniform interpretation of [EU] regulations means that a particular provision should not be considered in isolation but, in cases of doubt, should be interpreted and applied in the light of the other official languages, in the case of divergence between language versions, the provision in question must be interpreted by reference to the purpose and general scheme of the rules of which it forms part.⁴³

In the same way that it has acknowledged the existence of a new EU legal language in its case law, so too has the CJEU specifically acknowledged that, in certain cases, that EU legal language, while it retains an overall ‘EU meaning’, may not necessarily refer to exactly the same thing in various member states—thereby tacitly acknowledging the approximation that exists in the translation of EU law from one language to another. For example, in Case 327/82,⁴⁴ the CJEU considered that a single term (“thin flank” in Commission Regulation No 2787/81)⁴⁵ could be assumed to refer to different cuts of meat in different member states, it had no uniform precise anatomical definition but depended on various cutting and boning methods, themselves varying according to consumer habits and trade practices. In those circumstances the Court decided that it was not for it to provide a uniform EU definition—but this did not preclude the Regulation from bring ‘uniformly binding’ throughout the EU.⁴⁶

Thus, it seems that the CJEU functions in the way that it does, producing case law in 23 different languages to be ‘uniformly’ applied in 27 different member states, quite simply because the actors within that Court are aware, if not of the relationship and compromise between law and language, that:

…the entire system of EU law is a legal system built from approximations of law and language from different legal cultures and different legal languages, come together to form a new supranational legal system with its own language (référendaire, interviewee’s emphasis).

8 Conclusions

The teleological interpretative method employed by the CJEU assumes a normative, platonic notion of ‘EU law’ which is expressed in one language that exists in many linguistic versions. The Court itself seems to claim that while those linguistic versions may differ from each other on a purely linguistic level, at the legal level they express the same concepts (i.e. each linguistic version of EU law draws from the same EU legal concepts and therefore forms the same, new EU legal language). As mentioned above, the concept of hybridisation supports such arguments—it is only through a ‘new’ hybrid language that EU law may be properly understood. Similarly, approximation in language and translation in the EU can actually fulfil a positive role in ensuring the effectiveness of the legal order—indeed one could argue that the continued effectiveness of EU law is in fact dependent on its hybrid nature. The EU legal order functions precisely because of the implicit understanding among those who work at the EU level of the indeterminate and imprecise nature of language and law.

It is certainly true that many of the language and translation problems arising in the jurisprudence of the CJEU can be overcome through teleological interpretation or by reference to a ‘new’ EU legal language. In spite of that, however, there are certain instances where language and translation do cause problems in relation to the case law of the CJEU. The examples highlighted in this paper can be taken as an indication of a wider trend within that case law. Each of those examples came to light only in an incidental manner; the mere fact of their existence, however, points to the probability that there are a significant number of cases in which such problems are never discovered.

The approximation inherent in the production and translation of the case law of the CJEU is illustrative of the limitations of a multilingual legal system. The feature that distinguishes EU law from international law (and the very reason behind the EU’s language policy) is the fact that EU law is applicable to individual citizens in individual member states and therefore must be accessible (and effective) in all of the official languages of the EU. The method of teleological interpretation developed by the CJEU and the evolution of the notion of a new EU legal language do ensure the effectiveness of EU law to a large extent. However, the fact remains that different languages offer different accounts of reality. The approximation and imprecision inherent in language and translation do have implications for the case law produced by the CJEU. The concept of a single EU legal language that allows EU law to be uniformly applied throughout the Union is, in fact, necessarily based on a legal fiction. That fiction is a workable one, since EU law does function reasonably effectively. It is nonetheless a fiction, and an awareness of the problems of language and translation should therefore condition our understanding of the multilingual EU legal order.⁴⁷