Project managers at U.S. translation agencies are very familiar with the challenges involved in organizing and coordinating large projects across a range of languages. My guest for this column faces those same challenges as they apply to the European Union, which consists of 28 member states with 24 official languages. His department handles translation projects for over 60 clients in more than 550 language combinations.
Thierry Fontenelle
Thierry Fontenelle is the head of the Translation Department at the Translation Centre for the Bodies of the European Union. He has a PhD in linguistics from the University of Liège, Belgium (1995). He is the author of many publications on lexicography, natural language processing, computational linguistics (in particular, computational lexicography), proofing tools, and translation and terminology, including Practical Lexicography: A Reader (Oxford University Press, 2008). His PhD thesis, “Turning a Bilingual Dictionary into a Lexical-Semantic Database,” was published in the Lexicographica Series Maior (Niemeyer, 1997). He is an associate editor of the International Journal of Lexicography (Oxford University Press) and a past president of Euralex, the European Association for Lexicography.
A native of Belgium, Thierry worked as a research and teaching assistant at the University of Liège (1986–1996), then at the European Commission Translation Service as a computational linguist. He also worked as a translator and an interpreter for NATO before moving to the U.S., where he worked as a developer and subsequently as a senior program manager for Microsoft (2001–2009) creating proofing tools (spell-checkers, grammar checkers, and linguistic technology for Microsoft Office). He returned to Europe in 2009 to work for the Translation Centre for the Bodies of the European Union in Luxembourg, where he currently manages a team of around 110 translators.
Thank you, Thierry, for spending time with us today. Your linguistic background speaks for itself. What languages do you speak and how/where did you learn them?
My mother tongue is French. I speak English, which I started learning when I was 12, and Dutch, which I started learning when I was 13. I also speak some German.
What originally got you interested in languages and then in the computational aspect of linguistics?
I graduated from the University of Liège in 1986 with a BA in English and Dutch language and literature. I also have an MA in English translation and a PhD in English linguistics. Initially, I was involved in research projects related to machine-readable dictionaries and machine translation in the 1980s. I was intrigued by the similarities between teaching language to non-native speakers and “teaching” a computer to understand language (e.g., for machine translation or grammar checking).
You were a translator and interpreter for NATO some years ago. Where were you stationed? Did you work as a computational linguist there, or were you engaged in the more traditional form of translation?
Between 1999 and 2001, I worked in Luxembourg for the NATO Maintenance and Supply Agency, as it was called at that time. (Today it’s known as the NATO Support and Procurement Agency). I worked as a translator from and into English and French (NATO’s two official languages) for about 75% of the time. I was an interpreter in the same language pair for the remaining 25%. That’s where I learned to work in a booth doing simultaneous interpreting.
Let’s talk about your current job. First, who are your clients? What types of documents are involved? The Translation Centre’s website mentions trademark translations. What exactly does that entail? What else does your team translate?
The Translation Centre’s mission is to provide linguistic services (mainly translations) to decentralized EU agencies, as well as to European institutions that use our services on an ad hoc basis. As an EU agency, we can only work for EU agencies and institutions, which are located all over Europe. We cover many subject fields, including drugs and medicine, maritime security, chemicals, education, railways, food safety, and intellectual property. Our biggest client is the European Union Intellectual Property Office (EUIPO) in Alicante, Spain. The EUIPO deals with the registration of EU trademarks and designs, and we translate the descriptions of goods and services associated with a given trademark into all official languages of the EU. For the other agencies, the documents we translate include activity reports, guidelines (e.g., in the financial supervision field), work programs, websites, leaflets, brochures, and press releases.
With this kind of diversity, presumably there are different specifications and formats to respect. How much standardization is there? How much influence does the Translation Centre wield in this area?
We receive original documents in a wide range of formats, including Word documents, PowerPoint, Excel files, InDesign documents, XML files for websites, and MP4 files for videos to be subtitled. There are de facto standards, but the variety of clients prevents us from imposing any one in particular. This is because we need to be able to deal with the formats of the documents created in all the agencies we work for.
The Translation Centre has three language groups. Please tell us what those groups consist of and how they are organized.
The Translation Centre has a little over 110 in-house translators covering the 24 official languages of the EU. These 24 languages are grouped into three language groups:
- Language Group 1: Finno-Ugric, Baltic, and Slavonic
- Language Group 2: Germanic and Slavonic
- Language Group 3: Romance and Mediterranean
Each language group is subdivided into translation teams, one for each official language. The size of each team varies according to the translation workload into each target language.
The translators in these groups translate into their native languages and have varied academic backgrounds, mostly linguistically-oriented. Please tell us about their specific duties and what sort of on-the-job training they receive.
The Centre’s business model is based on extensive outsourcing because we don’t have enough in-house translators to handle the volume of work we’re expected to process. Around 80–85% of our translations are outsourced to external contractors. Our in-house translators then assess and revise all outsourced translations. This is an essential step because the in-house translators are familiar with our clients’ preferred terminology, their specific requirements, and the linguistic and stylistic conventions used across all EU institutions and agencies. The in-house staff translates urgent or very urgent documents and confidential documents that cannot be outsourced. They are also involved in creating client glossaries that are eventually fed into the Inter-Active Terminology for Europe (IATE) database, the joint terminology database for all the institutions/bodies of the EU. The Translation Centre manages the technical aspects of the IATE project on behalf of the project partners.
Translators are offered training throughout their careers at the Centre, both in terms of language learning (developing existing language skills and acquiring new languages) and client-specific activities (seminars and workshops on specific thematic domains, such as financial supervision or border guard training).
The Translation Centre also has a section that liaises with its clients. Please tell us something about this process.
We have a small external relations and communication section that serves as a link between the in-house staff and our clients. We process the feedback we receive from our clients about the quality of our translations. We also organize seminars and workshops with our clients, and they bring their experts to provide our translators with information about the key terms they use in their documents and explain the basic concepts of their specific mission. These experts can tell our language teams why they prefer to use this or that variant in a particular context. The knowledge we acquire at these seminars is incredibly useful to our translators, who then use it when they revise translations done by our freelance contractors.
All this is part of a global holistic approach to quality assurance. If we didn’t have this internal pool of in-house revisers and translators or organize these seminars, we wouldn’t be able to integrate our clients’ preferred terminology and specific requirements into our workflow and processes. This mechanism is a crucial aspect of our activity. I’m sure we would fail in our mission if we were “just” a mailbox between hundreds of freelancers and our clients. We would miss many opportunities to capture and transmit this specialized knowledge and incorporate it into our translations.
At the heart of the Centre there is a Translation Support Department. What does this group do?
This department is at the hub of the Centre’s operations. Its Workflow Management Section oversees the entire translation process, from the reception of clients’ requests to the final delivery of translated material. This is where requests are analyzed and documents are pre-formatted and pre-processed with translation memories, then assigned to external contractors or to in-house translators. This section also manages the deadlines for the various actors along the production chain. In addition to the Workflow Management Section, the Translation Support Department also includes a Business Development and Workflow Support Section, which deals with several important aspects of our activities, including examining technology and the maintenance of language resources (e.g., corpora, translation memories, term bases, and glossaries). It also manages the workflow for the translation of EU trademarks. The department is also responsible for managing the IATE project on behalf of the institutions that contribute to what is estimated to be the largest terminology database in the world today (over eight million terms covering all 24 official languages).
What are the other components of the Translation Centre?
The Centre has an Information Technology (IT) Department, which manages the IT assets and infrastructure, a service desk, and a Development Section that develops and maintains an entire range of state-of-the-art computer applications and equipment. The Centre’s Project Management Office is also integrated into this IT Department. The Administration Department provides administrative and technical support in the areas of human resources, budgetary and strategic planning, legal affairs, calls for tenders, accountancy, facilities, and security.
How about deadlines? Are there standard turnaround times for regular or rush projects?
The Centre has a grid of tentative deadlines with standard turnaround times that depend on the number of pages in the source document. Clients can request shorter deadlines for urgent (or sometimes even very urgent) translations, in which case the price will be higher. Rates are lower for “slow” translations with much more comfortable turnaround times.
Is the Centre fully self-sufficient or is it funded by another source?
The Translation Centre is entirely self-sufficient; it receives no subvention or subsidy. Via its charge-back system, the Centre receives payment from its clients in exchange for the linguistic services it provides. It’s also not allowed to make a profit.
You manage a team of about 110 translators. Is there much turnover or is the team pretty stable?
The department is pretty stable with rather low turnover.
Tell us about an average day in your life as head of the Translation Department. What do you actually do?
My department performs a rather impressive range of activities. Our structure is shallow, so I attend regular strategic meetings with our director and the other department heads. I meet regularly with the section heads and the heads of language groups in the department to identify new needs, but also to monitor the progress of the development of some crucial computer applications. Because quality assurance is our key concern, I spend a lot of time on the various processes that allow us to collect client feedback, measure their satisfaction, and respond to their concerns, if any. I travel all over Europe to meet with clients to better identify and understand their needs and concerns, but also to market our services and explain how we work. I also frequently meet with representatives of the translation services of other EU institutions.
Your department has worked with different IT tools over the years to handle the workflow. What are you using these days? What are the benefits of this system compared to what the Centre has used in the past?
Our translators use an off-the-shelf computer-assisted translation tool that we licensed via an interinstitutional call for tenders (SDL Trados Studio). We also use an in-house workflow management tool that the IT Department developed, called eCdT. This tool manages the entire workflow, from the reception of requests to the final delivery, via the various intermediate stages—analysis, pre-formatting, pre-processing with translation memories, outsourcing to contractors via framework contracts or negotiated procedures, and assessment and revision by in-house translators. We also integrated SDL WorldServer into the eCdT workflow tool to manage our translation memories and automate a number of pre-processing tasks.
Given your background in computational linguistics, are you directly involved in the design and implementation of the Centre’s language processing systems?
Not directly, because the natural language processing systems are part of the software packages we license via our interinstitutional calls for tenders. My knowledge of computational linguistics and experience with the development of natural language processing systems allow me to better understand the possibilities and limitations of these systems.
Is all the Centre’s translation work processed through an IT tool of some kind? Is there any sort of literary translation that’s done the old-fashioned way?
Most of our translations are done with our computer-assisted translation tool and we use translation memories whenever we can. We’ve also started offering a new service for video subtitling that’s very useful for EU agencies interested in publishing videos about their activities on YouTube, Facebook, and other social media. Our translators are trained to use a sophisticated subtitling software package. We don’t really do any kind of “literary” translation because of the highly specialized nature of the activities of EU decentralized agencies and the fairly technical nature of the documents they produce.
Given your huge databases, how do you handle storage?
Our IT infrastructure is quite complex, with lots of servers and a data center located outside of our building to ensure business continuity in the event of major problems onsite.
You were recently invited to the Colloquium on Lexicography in Santiago de Compostela, Spain. One of the discussions there claimed that: “Because simple equivalence between terms in multiple languages is not enough, terminologists who create term bases need to be able to structure the data in such a systematic way that they are usable not only by humans, but also by the authoring or translation tools used in the translation process.” Please tell us more about this process.
Creating a term base requires some thought concerning the structure of the linguistic data one wishes to represent. Twenty years ago it was not uncommon to find glossaries compiled as Word documents, where the various fields corresponding to the various pieces of linguistic information would simply be concatenated, without any kind of separators. Such glossaries were meant exclusively for human use. Today, it’s important to make sure that the data can be exploited not only by a human user (the translator) who looks up a given term, but by computer applications. This means that different fields must be used to store the various bits of information (e.g., the term proper, its abbreviation, the part-of-speech, definition, context, usage note, reliability code, and equivalent in another language). For a term to be recognized automatically in a text, the application needs to know from what field the information should be retrieved.
What do you consider the greatest challenges your department faces?
Our clients expect us to translate greater volume more quickly, with fewer resources and no mistakes. I believe this is a goal we should all have. We should never rest on our laurels, but strive to make the best use of what technology can offer. We must also educate our clients so that they understand that translation is an intellectual activity that requires time. Quality will inevitably suffer if everything has to be done in a rush. Sometimes clients remind me of Alain Delon, the actor in Man in a Hurry (L’Homme pressé), a movie about a man who doesn’t realize that it takes nine months to make a baby and that there is no way to shorten his wife’s pregnancy, which he thinks takes way too long. That said, our clients have their own constraints, so we must constantly strive to improve our processes to respond to their increasingly complex requests. Finding the right mix between technology (to automate what can be automated) and a human approach to our operations (based on a thorough awareness of the limitations of tools and technology) is always a precarious balance.
What else would you like to say about the Centre and the work you and your team do there?
We should remember that people are always the most valuable part of any organization. My colleagues are incredible and we have an amazing pool of talent. We produce around 750,000 pages of translation a year for clients from all over Europe, from 24 different linguistic backgrounds, with over 550 possible language combinations (not counting combinations involving nonofficial languages such as Russian, Arabic, Chinese, Norwegian, and Icelandic, which we also offer to our clients). Thanks to what we do, Europeans can read—in their own languages—about the work being done by several dozen highly specialized European agencies. I’m proud to be part of the team that makes this possible!
Notes
- Annual budget of the Translation Department at the Translation Centre for the Bodies of the European Union,
http://bit.ly/Centre-budget. - European Economic Community Council: Regulation Number 1 determining the languages to be used by the European Economic Community, http://bit.ly/EU-Regulation-1.
Tony Beckwith was born in Buenos Aires, Argentina, spent his formative years in Montevideo, Uruguay, then set off to see the world. He moved to Texas in 1980 and currently lives in Austin, Texas, where he works as a writer, translator, poet, and cartoonist. Contact: tony@tonybeckwith.com.