High standards and wide expertise

What the two categories of linguists have in common, in all the EU Institutions and agencies, is that we each strive for the greatest possible degree of accuracy, and for that reason the traditional rule has been that, wherever possible, we work only into our mother tongues. The rule is adhered to very strictly in the translation services because our output remains in permanent, written form. The interpreters can sometimes afford to be more flexible in the pressure of a meeting, but of course even if there may, in that case, be a little sacrifice in style, the requirement for absolute accuracy in rendition of meaning is always met. The quality assurance exercise begins from the very outset: we recruit our staff by competitive examination, set extremely high standards in terms of qualifications and experience and offer them training in further skills throughout their careers.
To speak for the European Commission, our thirteen hundred translators are a repository of know-how in many technical fields as well as being expert linguists. Contrary to popular belief, in fact, translation is not primarily about language proficiency - it is about understanding a complicated subject in one language and rendering it in the correct terminology in another. The Commission's areas of responsibility expand steadily with each year that passes, and range, as you know, from the financial and economic to the judicial, the social, the cultural, the scientific and the technical. So our ranks include people who have qualified and worked as lawyers, economists, engineers, accountants, computer specialists, even the occasional historian or medical doctor.
We apply this wealth of knowledge every day to the task of helping eleven different language communities in fifteen Member States understand each other, communicate with each other, work together and, in the final analysis, actually construct the European Union. This vital task, incidentally, costs the European Union only 0.8 per cent of its entire budget. Divide that by the number of people in the EU and it comes out at less than two euros per head per year. When critics of the multilingual system wonder whether even this isn't too high a price, we invite them to consider what the cost to the Union would be if its citizens and governments were unable to understand each other. The whole mechanism would grind to a halt within days, bringing to an end decades of progress towards peace, reconciliation, solidarity, development - and democracy.

Multilingualism: the legal basis

You are entitled to ask why we use so many languages, when other organisations (NATO, the United Nations, the Council of Europe) manage with a much smaller number. Those organisations function perfectly well without every citizen necessarily understanding the language or languages they work in. The answer is that the European Union is the only international organisation which, in its areas of jurisdiction, passes laws directly binding on its Member States and their citizens.
Before the countries in central and eastern Europe which have applied to join the Union eventually become fully fledged Member States, they will have signed up to the whole existing apparatus of laws and regulations which bind the existing members - what we call the acquis communautaire.
· One part of this package is Council Regulation No 1 of 1958 - this, as its name suggests, was the first regulation ever promulgated by the Council of Ministers in the then infant Community. This piece of legislation lays down that Regulations and other documents of general application shall be drafted in all the official languages. It follows that the texts concerned are equally authentic in all the language versions, which means that they must say exactly the same thing in all of them. It could not be clearer that the founding fathers' first concern was that everyone in the Community should understand what was going on - otherwise it would not be a democracy.
· Another stipulation - and one which was recently enacted explicitly in the Amsterdam Treaty - is that every citizen of the Union may write to any of the European Union institutions or bodies in one of the official languages and have an answer in the same language.
· And a third, also written into the EC Treaty, provides that: "Any citizen of the Union (...) shall have a right of access to European Parliament, Council and Commission documents." Obviously, accessibility implies providing a version in a language the citizen can understand.
These provisions are the very basis of the Union's multilingual policy, and the EU's language services, though they are not explicitly referred to anywhere in the founding legislation, exist to put them into effect.

Keeping the workload within manageable limits

Keeping the language channels open is heavy work. In the Commission alone, we produced one million two hundred thousand pages of translated text in 2000. Adding yet more languages to the existing list is certainly going to increase our workload. But, as I hope to show, the prospect is not as daunting as it may seem from outside. Let me rebut a common misconception about how far multilingualism actually goes in the European Union. By no means every single text written in all the EU institutions has to be translated into all the official languages. Because the European Parliament is
an assembly elected from among ordinary citizens, all of its official proceedings and background drafting go on in all the languages, but in the Commission, for example, a great deal of our day-to-day administrative work, including the preparation of draft legislation and the monitoring of compliance with EU law, can be done in a much smaller number, or even in just one. Looking forward to enlargement, full translation of the acquis communautaire into all the official languages is a non-negotiable, but that represents the final stage in the legislative process, the tip of a very hefty iceberg. A great deal else is never needed in all the languages. As far as the Commission is concerned, much of the preparatory work for legislation, and most of the internal, administrative paperwork, is drafted and discussed by officials in-house in only one or two procedural languages, in practice English and/or French. For meetings of the Members of the Commission, German is added. In practice there is little demand for translation between the more exotic languages: even at the moment we receive very few demands for translation between, say, Greek and Finnish, or Dutch and Portuguese. Correspondence between an EU institution and an individual Member State involving bilateral matters is generally conducted only in that Member State's language. For internal use, material which arrives in less widelyspoken languages may have to be put into something else so that the officials concerned can deal with it, but there again the demand is largely from the nonprocedural languages into English and/or French. Nor do the institutions have to use all the official languages when they communicate with each other on administrative business.
What goes for translation is also true of interpretation: multilingualism is assured where it is needed. So, administrative meetings among officials, or discussions between special-interest groups (such as scientists) who generally agree to communicate in one or two languages, and which are not intended for the general public, will not require full interpretation. But major occasions such as sittings of the
European Parliament and its committees and sessions of the Council of Ministers will, of course, need interpretation into and out of all the official languages.

Differing requirements

The various Union institutions have thus rather different language needs. As I have said the European Parliament is an assembly elected from members of the public who are required to be skilled politicians, not linguists, so it has to conduct its political work in all the official languages from the earliest stages. The members of the Committee of the Regions and the Economic and Social Committee are appointed by the national governments and will also generally need to work in their own languages.
Some of the smaller bodies working in purely technical areas, on the other hand, have been able to arrive at pragmatic arrangements which reduce the multilingual burden without a loss of transparency or efficiency: the Office for Harmonisation in the Internal Market, in Alicante, for example, operates in only five languages.

Public awareness

Of course, as far as the general public are concerned, even five languages is a fair number, and if they think of our particular language regime and its complexities at all, they may wonder why we don't save time, money and effort by using just one. I have already given the reason for this. Moreover, the answer is that the Member States, not the Commission or any other EU body, decide what the official languages are to be, and no national politician in any Member State could ever persuade the voters to abandon the use of their own language for EU business. To put it frankly, some sectors of public opinion in some Member States are already suspicious enough of the European Union as it is, and their sense of alienation would be complete if it started talking to them in a foreign language. And they would be right. People living in a democracy rightly demand that government should be conducted openly. The Union's citizens will never agree to abide by laws which they cannot understand. As for the technical consequences of working in eleven, and later more, languages, members of the public are by and large uninterested - what they want to know is that the European Union speaks their language. European politicians, on their side, also want to get their ideas and policies across to all the Union's people in as near as possible their own words, so no serious political interest in the Union has ever suggested reducing the language coverage.
We can be confident, then, that the political future of multilingualism in the European Union is secure. The Commission's Translation Service, and the Union's other language services, are not likely to be going out of business any time soon. On the contrary, we are bound to expand, but for us the challenges of enlargement will be primarily organisational, not political.

Language situation in different candidate countries

Let us take a look at the linguistic landscape before us and how it will change with enlargement. You know, of course, that at the moment there are fewer official languages than the existing Member States because some of them share the same languages. With the countries in the waiting list, however, the situation will not be so conveniently simple. Each of the central and eastern European countries which are now officially candidates negotiating for accession - Poland, the Czech Republic,
Hungary, Slovenia, Estonia, Slovakia, Lithuania, Latvia, Bulgaria and Romania - has its own distinct language. Cyprus and Malta are special cases. The requirement for Greek in Cyprus is already met, but under that country's Constitution Turkish is also an official language. Turkey has, of course, also applied to join. The Maltese Government may be content, on pragmatic grounds, to follow the Irish example and not insist on the use of its national language in every context, accepting English for
most purposes - but it would be politically entitled not to take that path.

Language consequences of enlargement

There is, in other words, a very real prospect that within the next few years the number of official languages in the Union will go up to at least nineteen, and it could even double. On a purely numerical basis, this increases the number of language combinations which have to be catered for by geometrical progression. With eleven languages, we get a hundred and ten combinations. With twenty-two, the sum comes out to four hundred and sixty-two combinations - more than four times as many. But the practical reality is less frightening, as I have said. Nonetheless, we are talking about countries which will enrich the Union with very distinct cultures and histories and which work with widely varying legal, educational, and social systems. All of these will have to be understood by our translators as thoroughly as the languages proper.

Building on experience

We veterans of the European Community, now the Union, have faced this kind of prospect before, although on not quite the same scale. In the 1950s we started off with six Member States; in 1973 the United Kingdom, Ireland and Denmark joined; Greece signed up in 1981, Spain and Portugal in 1986; and five years ago we were joined by Austria, Finland and Sweden. In every case the whole body of European Community law applying at the time - the acquis - had to be translated into the new countries' languages in advance of accession, so that they would know what they were committing themselves to. And from the very first day of membership there had to be full-time translators into and out of the new languages in the European institutions, so that business could be carried on in exactly the same way as for the existing members.
We will have to go through the same process as new countries are admitted over the next few years. Experience of the past has shown that the workload is going to rise steeply. To put a rough figure to it, every time a new language is added the Translation Service's workload rises by ten per cent. To prepare the ground, we are applying the methods we used before. What does this involve? How are the European Union and the candidate countries preparing for the linguistic effects of enlargement?

Pre-accession: translating into the new languages

At the moment the lion's share of the work falls to the candidate countries themselves. No dates for accession have yet been agreed, but they will be soon and there is no time to waste. In the pre-accession period the candidate countries' governments, not the Union, are responsible for having the EU legislation currently in force translated into their languages, and each candidate country, has set up a special unit to do this. These Translation Coordination Units have been busy for the last four years or so. This may seem like a long time, but the acquis is estimated to fill 50 standard volumes of the Official Journal of the European Communities, or, to put it another way, seventy thousand pages of standard text for translation - and it is full of specialised EU terminology and concepts which may be extremely difficult to render in countries where the political, legal, economic and social systems in place in their recent history were so different from those of the existing Member States. Meanwhile, a special department of the Commission called the TAIEX office (standing for Technical Assistance Information Exchange) has been responsible for obtaining translations into English or French of those parts of the national legislation which need to be monitored by the Commission officials working on particular aspects of the legislation, to make sure that it is in conformity with EU law. Some texts have also been translated either in-house or by freelance translators employed by the Translation Service. TAIEX also provides technical support to the Translation Coordination Units.
As this work proceeds, revision centres will need to be set up in the central and eastern European capitals some eighteen months before the date of accession. Our practice in the EU is, as you will appreciate, never to release any text with legally binding force without revision. We plan to use again the method adopted in the case of Finland and Sweden, which means that these revision centres will be staffed by locally recruited Commission employees whose task it will be to check the legal accuracy of the translation work done in the national capitals. The same people may also well form the basic catchment area for the recruitment of translators and revisers to staff the translation services of the Commission and other Institutions once their countries are full members. I believe we can look forward to some lively competition for their services.
As the date of accession approaches, the responsibility shifts to the Commission and the Council since they, rather than the acceding Member States, are responsible for publishing the authenticated versions of the legislation in the new languages in the form of Special Editions of the Official Journal. So over the last year or so of the accession process, those institutions will have to recruit legal revisers of the mother tongues concerned to carry out a final authentication of at least the most important legislative acts.

Finding enough linguists

The translating and revising of the acquis is probably the only task in the entire accession operation that is exactly the same for every would-be Member State. But they come to it from widely different starting points. Consider the relative sizes of Poland on the one hand and Slovenia on the other, and you will appreciate how much easier it is to find the necessary number of translators, revisers and legal experts to do the work in a large population than in a small one.
Another potential difficulty, which we experienced with Sweden and Finland when they first joined, is that in some countries there is very little tradition of translating the type of material the EU generates into and out of the national languages. We estimate the number of fully qualified translators who will be needed by the EU institutions once the candidate countries have joined the Union at about two hundred per new language, so that in most cases we are having to prepare the ground by inviting those countries' universities and training schools to contact us for advice on how they could slant their courses for professional linguists towards our requirements. TAIEX arranges study visits and round tables on many topics for them. The translation services themselves have taken on substantial numbers of trainees from the applicant countries for short periods, usually in the terminology department, where they are trained in our working methods and generally made familiar with the way the institutions work.
Our interpreting services, of course, will also need new staff who meet their own high standards.

Post-accession: translating from the new languages

Once the new countries have joined, material coming from them will have to be translated out of their languages, which will then be the responsibility of the EU institutions. However, in the Commission we know from experience with previous enlargements that this will mainly involve the English and French-language translation communities. The example of Swedish and Finnish shows that in the Commission some seventy-five per cent of the total number of original pages in each new language will need to be translated into English and about fifteen per cent into French. Only the ten per cent remaining will need to be put into the other official languages. The reason is that the bulk of incoming material from the Member States is addressed to the Commission for information or evaluation purposes and is not intended for public release or discussion, so most of it need only be available in a
language widely understood within the Commission, such as English or French. The problem will be greater for the Council and the European Parliament, which, as representatives of the governments and the public have to operate more multilingually.

New solutions

The scale of the enlargement this time round exceeds anything we have known in the past, and the new languages we are talking about are relatively little known in the existing Member States, so, here again, finding enough native speakers of English, French, German and the other existing official languages who know the new ones to the high standard required is likely to be a particularly difficult exercise. In the Commission we propose to pursue an active policy of recruiting translators with a knowledge of the new languages before enlargement, and we will be asking our administrative authorities to depart from previous practice by organising recruitment competitions including tests in languages which are not yet official languages. We are also looking at the possibility of temporarily recruiting specialists, such as university lecturers on sabbatical leave, who are not interested in permanent employment with the Translation Service. Meanwhile, ten per cent of our existing staff translators - one hundred and twenty of them - are attending in-house courses in the languages of the front-line countries and are encouraged to do further, intensive study in the countries concerned. Thirty staff translators have already completed the courses in some of the languages and are already operational.
There is one principle which has so far been inviolate in the European Union's language services but which may finally have to be breached: for the first time, we may also be looking for translators who can work both ways - who are, that is, able to translate out of their mother tongue as well as into it. Those most in demand will be able to work into English or French. This is a surrender to practical reality, but also a recognition of the happy fact that in the eastern European countries most professional linguists are trained to work both ways from the outset, so that for them it would not
be the innovation which it would be for us. Informally this has already happened once, with Finnish when that country first joined and before the Commission had had time to recruit or train speakers of other languages who could translate from Finnish to a sufficiently high standard. It is an adequate way of dealing with certain categories of documents which have no legal force and are not intended for public release.
Translation out of the mother tongue inevitably involves some compromise on style, but documents produced in that way would only be used internally by the institutions, and then only as a way of getting a basic understanding of a text. This kind of translation can be compared with the routine drafting in English and French that is already being done every day, in all the institutions, by civil servants who are not native speakers of those languages.

Huge expansion? No.

It might sound from what I have said as if staff numbers in the language services are going to grow in the same proportion as the language combinations needed. The media also sometimes ask if working with twenty-two languages instead of eleven will mean at least a doubling of the costs. We don't think either of these things will happen, for a number of reasons in addition to those I have already given:
· We can make savings on manpower and costs by categorising documents so that simpler approaches (summarising, for example) can be adopted towards less sensitive or more ephemeral material.
· Another cost-saver is further investment in, and use of, modern technology. We have several dozen linguists and information technology specialists working fulltime in translation support units on the development of translation tools. The Commission's Translation Service has put substantial human resources and funds into developing Eurodicautom, an enormous multilingual database containing terminology from all our fields of work. We have also pioneered the use of textrecognition and replacement programs which save countless man-hours by enabling translators in any of the official languages to find and incorporate into their work phrases, sentences, even whole paragraphs which have been translated before. Our oldest research and development project is Systran, a machine translation system which we have been working on for some thirty years. It provides rough-and-ready translations between several of the official languages, including English, French, German and Spanish, working in eighteen language combinations out of the one hundred and ten possibles. We do not think it will ever replace human translators, since it understands nothing, but for getting the bare gist of a clearly written text it is useful. Voice-recognition software, too, is being tested at the moment; if a reliable system can be found, we hope translators will be able to dictate directly onto the screen.
· We can economise on the need for unusual language combinations by working, as our interpreter colleagues often do already, through relay languages. You will have heard the expression "Traduttore, traditore", and certainly there is an added risk of distorting the original meaning when an utterance has to go through two gates before it reaches its destination. But it is a risk worth taking when the translation cannot otherwise be done at all because no one individual has a sufficient grasp of the two languages concerned - or of the subject matter.
· Making greater use of freelances is another possibility. An average of forty per cent* of the interpreters used by the Commission's interpreting service at any given time are freelances, and this allows the service to respond very flexibly to language needs which may change every day. In our Translation Service the amount of work sent outside is just short of twenty per cent at the moment, and we hope to increase it if the funds are made available by our political masters, as we expect they will be. Texts which are not legislative, confidential or urgent can be farmed out, provided that competent translators can be found on the market, leaving the established staff to concentrate on a core category of material which will always have to be done in-house. Whether this would materially reduce costs is still a moot point, but it would certainly rein in the tendency for established staff numbers to expand.
· Teleworking is being studied in the Commission as a whole, as a way of reducing overheads. Pilot projects for translators have been run in some of the EU institutions, and in the Commission's Translation Service we are considering the implications of allowing some staff translators to work from home.

Keeping demand under control

All these points relate to increasing the supply of translation without inflating the cost. Just as important is reducing the demand. Our only customers are the Commission departments, and for some time we have been pressing our administrator colleagues to exercise a little more self-discipline when they start writing. To put it plainly, we have asked them to write shorter, clearer texts - and slowly we are beginning to see some return. For effective communication in any language, a useful watchword may often be "Less is more". A system of internal billing which is to be introduced in the Commission in 2003 will also, we think, concentrate their minds on sending us only what really needs to be translated.

A new horizon

* Ian Andersen of SCIC has been asked to come back with confirmation of this figure direct to Mr cCluskey. One area of potential expansion in demand which no one could have foreseen a few years ago is the Internet. The EU institutions cooperate to produce a huge, interinstitutional website called Europa, which now contains information on every aspect of the EU's policies and activities, running to thousands of pages. It is a multilingual site, of course, but selectively so, since certain types of information donot have to be put into all languages, only the ones needed by their target audiences.
Nevertheless, keeping the material up-to-date is a serious linguistic challenge, and not just because of the volume. Increasingly, Europa aspires to be interactive. It is the most heavily consulted publicly-run website in Europe and scores huge numbers of hits. An administrator in the Directorate-General for Health and Consumer Protection told a meeting of the Europa team the other week that her department's page on the Commission's response to the BSE crisis had received nearly a million hits in the previous month. In eleven languages, of course, and many of them requiring a rapid answer. Another web-based activity which calls for even more immediate responses in the official languages is internet chats. Several Members of the Commission have had on-line, multilingual discussions with the public, in real time, using interpreters to mediate between themselves and their interlocutors as though they were all attending a traditional meeting in the same place. In the interests of democracy, these contacts with the public will clearly have to be available in the candidate countries' languages as well after enlargement, and we in the language services will have to devise imaginative ways of meeting that challenge along with all the others.

Maintaining the standard

To conclude, I hope I have shown you that the European Union is fully committed to its traditional policy of multilingualism and will not flinch before the very serious budgetary and organisational challenges of coping with a new influx of languages after enlargement. For us, the word multilingualism goes beyond its dictionary definition of an ability to speak a number of languages. We extend it to encompass the fundamental principle that all our Member States' official languages, however familiar or unfamiliar they may be outside their own borders, enjoy equal rights. The whole philosophy of the European idea is that best way to bring people together is not to iron out their differences but to respect them and build on them. Our diversity of languages is a positive asset, not a burden. That's what the Treaty means, and that's what we in the language services are there to uphold. For us, enlargement will create what is essentially a management problem. As support services in the EU set-up, we don't make the policies, but through our work for the policy-making departments we are indirectly at the service of every citizen of the Union, and we will, as we have done before, do everything we can to make sure that with nineteen, twenty-two or even more languages the standard of that service will be as high as it was when there were only four.


Brian McCluskey, acting Director-General,
Translation Service of the European Commission