Is That a Fish in Your Ear?

TWENTY-ONE Ceci n’est pas une traduction: Language Parity in the European Union

So reads a recent version of the basic language rule of the European Union. It was originally laid down in Article 248 of the Treaty of Rome, which first set up the European Economic Community in 1957: that body, and any offices under its authority, was to communicate with the governments of each of the member states in the language of the member state in question. It sounds a modest requirement, but it was actually a revolution. Unlike all previous empires, communities, treaties, and international organizations, the European Union has no one language and no finite set of languages, either. It speaks in all the languages that it needs, whatever they may be. An act of political will made the previously ungrammatical expression “a single original in Danish, Dutch, English …” an authoritative rule.

To begin with, there were six states—Belgium, France, the Netherlands, Luxembourg, Germany, and Italy—and four languages involved: French, Dutch, German, and Italian. The EU has grown meanwhile and now has twenty-seven states using twenty-four different languages. But whether we are dealing with four or twenty-four languages, the revolutionary meaning of the basic rule, ill understood when adopted and not widely acknowledged even now, is that in the whole huge mass of paper put out by the EU, there are no translations. Everything is the original, already.

Each language version of a law, regulation, directive, or letter emanating from the commission or any of its institutions has the same force, the same authority, the same validity as any other. Nothing is a translation—except that everything is translated. This has been the unprecedented language rule under which increasingly large numbers of people have now lived and worked for more than fifty years.

You might think it would have made a difference to what people say about translation, but for the most part it has not. Since it is theoretically impossible to have more than one original of a text in the long-standing traditions of literary study and language teaching, people have tended to disregard the language reality of the EU, to denigrate it as a waste of a huge amount of money, or have uttered dire warnings of the risks it incurs. However, I’ve yet to meet a translator who has turned down a well-paid job in Brussels or Strasbourg on language-theoretical grounds.

The language rule of the Treaty of Rome was obviously not thought up by philosophers, linguists, or translators, let alone by theorists. It arose from the need to make all members of this daring new venture feel they had equal respect and equal rights—to abolish what I have dubbed translation UP and DOWN. It was invented by politicians for eminently political reasons. What’s more, those politicians and several generations of their successors have been prepared to devote substantial sums of money to making the language-parity rule work. DG Translation (the translation division of the European civil service) currently employs 1,750 linguists and 600 support staff, and spends vast amounts of money to produce millions of pages of administrative and legal prose every year—probably more than has ever been spent on translation by any community ever before.[140]

From the 1960s it became fashionable to think, in a manner attributed to Michel Foucault, that language is power and that all power is language. The EU language story, like George Orwell’s polemical invention of “Newspeak” in Nineteen Eighty-Four, doesn’t invalidate that entirely—but it does go to show that, in the last analysis, power is power. Language is no less a possible object of political will than any other human activity.

The language-parity rule has many interesting consequences. It means that no official EU text can be faulted or dismissed or even queried on grounds of it having been incorrectly translated from the original, since every language version is in the original. Faced with a single original in twenty-four different languages, none of the inherited and traditional issues of translation commentary has much purchase. You could call this a political fiction. But it is not theoretical. It exists.

Paired texts in different languages each having equal force are nothing new, in fact. The Rosetta stone bears a decree written in 196 B.C.E. in honorific legalese to record a tax amnesty granted to temple priests in Egypt. It was carved on a slab of basalt in koiné Greek, in demotic Egyptian, and in hieroglyphics.

The decree was clearly intended to have the same force for three different groups of people among its potential addressees. Commonly treasured as the source of the clues that led to the decipherment of hieroglyphic script, the Rosetta stone should also be taken as proof that the founders of the EU were not seeking the impossible when they adopted the language-parity rule.

The written history of the two main languages of the original EU also began with a bilingual edict. The Oath of Strasbourg was sworn in 842 C.E. by two grandsons of Charlemagne who ganged up on a cousin they suspected of trying to elbow them out of their inheritance. Charles and Louis spoke different languages—the one having an early dialect of German, the other an early dialect of what would become French. Each swore allegiance to the other in the language of his ally. This was not just feudal politeness. The oath was written down so it could be copied and taken around and read out to the armies of Charles and Louis. Louis did not need it to tell his own people not to fight Charles, nor did Charles need it to tell his own people not to fight Louis’s men. Each needed to give assurance to the other side that he was no longer an enemy but an ally in the common fight against the cousin, Lothaire. That is why they produced a bilingual screed, with the texts in the two languages in parallel columns, each intended not to say exactly the same thing, but to have exactly the same force when read aloud to bands of illiterate soldiers. The Strasbourg Oath, the founding document of two languages and also the key to the geographical shape that European nations have taken since then, is also the founding document of the EU’s language norm.

But there is a catch. It’s unlikely that the signatories of the oath actually spoke to each other in either of the languages written down. They probably used Latin for face-to-face negotiation of the terms of the treaty and then left their scribes to find a way of writing down the agreement in the (previously unrecorded) languages of their troops. So although there is no explicit original of the Oath of Strasbourg, it is very likely there was an implicit master text that would have been the out-turn of a bargaining session in learned Latin that was probably translated by scribes or educated slaves into Old High German and Old French, respectively.

It’s an open secret that the EU also possesses an interlanguage for most practical uses in the corridors of the Berlaymont building, in the canteens and private meeting rooms—and it’s English. However, it is definitely not the case that EU texts are first written in English and then translated. Things work in an altogether more interesting way. A panel or subcommittee meets to draft a regulation. It uses one of the four official working languages of the EU—German, French, English, Italian—but there are always other language drafters present. The first draft is argued over not only for content but also for how it is going to be expressed in the other working languages. The draft is then translated and the committee reconvenes with the drafters to smooth out difficulties and inconsistencies in the different versions. The drafters are indistinguishably language professionals and civil servants participating in the development of the substantive text of EU regulations. The back-and-forth movement of the draft between the committee and the drafting departments produces, in the end, a text all consider equal in all its versions, and in that sense the “language fiction” of the EU’s rule of parity is not fictional at all.

The European Court of Justice in Luxembourg (ECJ), which resolves questions of law that cannot be answered by any of the national appeals courts of the states that make up the EU, is run in a slightly different way. It has a single working language, which is French. All documents used at every level by the court are either written in French or translated into French by members of the army of language professionals who work there.

However, plaintiffs—who may be member states or authorities within a national jurisdiction—may bring cases in whichever language they wish, which is normally the language of that state. The language of the state becomes “the language of the case,” and all documents in the file, whatever their source, must be translated into that language. More work for (different members) of the law-and-language team. But that is only half the story. The legal decisions of the court are made by all or some of the twenty-seven European advocates general, one appointed by each of the member states. The ultimate authority consists of a set of distinguished judges collectively speaking and writing all twenty-four languages of the EU. They use French for lunchtime conversation, informal consultations, and committee discussion, but they give their all-important opinions on the cases before them in their home tongues. For example, a case brought by the Portuguese government against a Bavarian dairy-farming consortium that is judged by the Estonian advocate general involves translations in five directions—PORFRE, FREPOR, FRE GER, GERFRE, and ESTFRE, allowing four additional transmissions by relay from French POR[FRE]GER, GER [FRE]POR, EST[FRE]POR, and EST[FRE]GER. The three remaining directions (see here for the math), from French, Portuguese, and German into Estonian, are not needed because, except when giving his opinion, the advocate general operates in the language of the court, which is French. However, since the rulings of the ECJ have force in the entire EU, legal opinions are not released and do not come into effect until they have been translated into all twenty-four of the official tongues. Every section of the 750-strong corps of lawyer-linguists at the ECJ becomes involved at some stage in every decision that is made.

Euroskeptics treat this lavish provision of multidirectional translation at the ECJ as a scandal of waste—a mere job-creation scheme. It’s true things didn’t happen that way in the appeal courts of the multilingual Ottoman and Hapsburg states, and the ECJ does cost a lot to run. It’s also true that the law of unintended consequences means that language parity as implemented in the steel-and-glass palace on the Kirchberg Plateau creates very awkward disparities of its own. If you are a trained lawyer from Malta, Estonia, or Hungary with excellent French and a good knowledge of one other European language, the job opportunities in Luxembourg are very attractive indeed. The effect is that Malta, Estonia, and Hungary have difficulty recruiting such individuals for their own national civil services, where their skills are very much in need. But if you are a British lawyer with excellent French and one other European language under your belt, you have far more lucrative careers waiting for you in London and New York, and the ECJ thus has a chronic shortage of translators into precisely those languages it most frequently needs.

But the EU cannot exist without the ECJ. If the ECJ were to abandon its own version of the language-parity rule, it’s not obvious how European law could have force in any of its twenty-seven member states. That’s why for the last fifty years, all supposedly commonsensical or budgetary objections to its translation regime have been overruled. The political will to make Europe work is too strong and too grand to let translation issues stand in its way. Europe really has built a radically new kind of translation world.

What’s quite specific to the ECJ, however, is that it does not employ translators as such. Language professionals in the Kirchberg complex are all also lawyers, and they are involved in the work of the court at many levels beyond strict language transfer.

Lawyer-linguists have access to confidential material and work under the same procedural rules as lawyers; they also advise on drafting, down to the small details that might produce ambiguities when expressed in other languages. The work of a lawyer-linguist is much more than translation—it is the manipulation of the law as language and language as law.[141]

Many of the cases brought before the ECJ arise from conflicting interpretations in different member states of regulations made by the European authorities—in effect, clashes between different interpretations of different language versions of what is held to be the same text. Given that all language versions have force of law, how does the court deliver the judgment of Solomon that this version is to be preferred over that one?

They can’t call either version a translation since all versions are originals; because the court’s working language is French, moreover, there are almost always three texts or formulations involved. On rare occasions, the taboo term translation mistake has been used—for example, when the German version of a regulation about the import of sour cherries used the term Süßkirschen, “sweet cherries,” instead.[142] Such an easy judgment is untypical of the court’s work. More frequently the court has to decide what the law was intended to achieve, over and above any one of its linguistic expressions. In monolingual national cultures of law, the best evidence of the legislator’s intention lies in the words of the law, and much traditional legal argument is about the meanings of words. In European law, you have to go one further than that. Questions of legal interpretation in the appeals court of the EU are also always questions about language in twenty-four different forms.

Let’s suppose in some practical circumstance not foreseen by the drafters of an EU directive there is a substantive difference in the force of the French and the German texts, and that this has given rise to a complaint by France that Germany is not applying EU law correctly. The ECJ has to decide whether France is right. But since there is no master text (in Latin, for instance) to provide a higher authority or a standard of judgment, the court has basically only two ways of working out what it thinks. Using the skills of its divisions of lawyer-linguists, it can list all the language versions that support the French interpretation and all those whose sense in the context of the case leans more toward the German interpretation—and grant victory to the larger group, whichever it is. But the ECJ does not have to proceed by this kind of “majority verdict.” It may identify one language version that it considers to have expressed the legislative intention of the directive more clearly, or more precisely, than any of the others.

Both these procedures hark back to the tools developed by the Church Fathers for establishing the “word of God” through comparisons of the different translations of the Bible (principally, the Greek and Latin ones). What has been called the “Augustinian approach” to the interpretation of European law effectively seeks to establish a meaning that transcends any one of its language versions but which animates them all. It runs into some fairly obvious problems.

In Peterson v. Weddel & Co., Ltd., the issue was a criminal prosecution within the U.K. for violation of a regulation setting limits on the operations of trucks. An EU regulation exists that allows member states to make exemptions from the general rule for the “transport of animal carcasses or waste not intended for human consumption.” The firm that had been fined in the U.K. had been transporting animal carcasses to butchers’ shops, which clearly intended to sell them for human consumption. But the trucking firm claimed it was exempted from the rule by the EU clause just quoted and was appealing against an English court’s refusal to allow it to get away with its behavior. The trucking firm’s lawyers claimed that waste not intended for human consumption and animal carcasses in general (whether intended for human consumption or not) were exempted, whereas the English courts had considered that the exemption applied only to waste and animal carcasses not intended for human consumption. It may sound arcane, but the issue was clear enough: Was the trucking firm cheating on the rules or was it not?

The issue at the heart of this case is a familiar problem in the language of law and in language in general: When you have a list of nouns followed by a qualifying or restricting phrase, where do you put the punctuation? Does the restricting phrase restrict every member of the list, or only the last one? Does the expression “children and women with babes in arms” include children with babes in arms or does it not?

In daily usage, we leave disambiguation of this kind to common sense and context. In law, it’s fertile ground for persnickety legalese. When this issue came before the ECJ, however, the lawyers, the linguists, and especially the lawyer-linguists began by reviewing and comparing all twenty-four language versions of the exemption. They found one among them—the Dutch text—where the restriction to goods “not intended for human consumption” precedes both “animal carcasses” and “waste.” It does so for almost exclusively grammatical reasons. The court treated it as a godsend, however, not as a grammatical variant of the same ambiguous text. It chose to regard the Dutch order of words as a clearer and more precise expression than all the others of the true intention of the law—and turned down the appeal. The trucking firm had to pay the fine.[143]

Let’s assume that the EU body that first thought of exempting certain classes of trucks from general rules was thinking about trucks full of potentially rotting and smelly flesh. What’s of interest here is not the ECJ’s final judgment, with which we can easily agree, but the reasoning it used to justify itself. The reasoning is of a very simple grammatical kind: it says that qualifications preceding a list of nouns apply to all nouns in the list. This semantic principle is made manifest in the Dutch version, but all the others, which for grammatical or stylistic rules put the qualification at the end of the list, must be taken as expressing the same thought.

The reasoning does not make sense in most of the languages of the EU and especially not in the court’s working language, French, where all kinds of qualifiers, including simple adjectives attached to single nouns, follow and do not precede the noun. Where does the ECJ’s insight into the clarity of Dutch come from? The most likely answer is the grammar of English. It is English, not French, Spanish, or Hungarian, that lends intuitive support to the view that “not-intended-for-human-consumption animal carcasses and waste” is a less ambiguous expression than “animal carcasses and waste not intended for human consumption.” Despite the huge and conscious efforts it makes in precisely the opposite direction, the ECJ cannot resist the slow but steady homogenization of the languages it uses to uphold European law.

I don’t mean to snipe at this particular judgment or to undermine the important work that the legal-cum-linguistic contortionists of Luxembourg do. However, the comparative method used to establish the ultimate intention of a law—a method that can be likened to Saint Augustine’s practice of biblical exegesis—must itself be conducted in a language. Suppositions and assumptions about the meanings of words, grammatical structures, and rhetorical turns are necessarily rooted in one language, not suspended on a hook from a supralinguistic legal sky. In the polyglot corridors and canteens at Kirchberg, however, where you may start a conversation with a Spanish judge in French and switch to German to say hello to a nice person from Prague, it’s a truth that is easy to forget. As one lawyer working there said to me when I visited, he never really thinks about which of his four languages he is speaking or writing at any given time—he switches without conscious effort, as if he were just shifting the weight of his shoulder bag from the left to the right side. The outcome of such unconscious linguistic determinations of legal finagling is that the meaning and grammar of twenty-four languages have begun to merge into an ECJ language culture that is all its own—sui generis, in Saussure’s terms, or “Euro-speak” in common language. As one of the few scholars to have studied the language maze of Luxembourg closely puts it, “The unique situational factors in the production of European jurisprudence have led to a hybridization of law and language.” It seems to me—admittedly an outsider and an amateur in this field—that the underlying structure of this new hybrid, even though it is formally expressed through the medium of French, is provided by the English tongue.

Some people from both “Europhile” and “Euroskeptic” factions think that it would be better if European institutions were run in English anyway. This is because the language-parity rule of the EU is a constant cause for delay and also tends to make official decisions and opinions more contorted and obscure than they really need to be. As stated earlier, the rulings of the ECJ come into force when they are published, simultaneously, in all the official languages of the EU. Judges are therefore under constant if discreet pressure from their permanently overworked lawyer-linguists to keep it short. European jurisprudence is thus typically tight-lipped and does not provide the many pages of argument and justification that normally accompany a ruling from the House of Lords or the U.S. Supreme Court. The laudable political aim of treating all the languages of Europe as equal produces the unwanted but perhaps inevitable result that ECJ rulings are sometimes so pithy as to defy comprehension in any of them.