There are conditions under which translation technology is not only worth learning but essential. Then again, it may be completely disregarded in many other circumstances without substantial loss. Localization is the paradigm of the need for technology, while interpreting and literary translation are examples of the latter. The localization business is intimately connected with the software industry and companies in the field complain about the lack of qualified personnel that combine both an adequate linguistic background and computational skills. This is the reason why the industry (around the LISA association) has taken the lead over educational institutions by proposing courseware standards (the LEIT initiative) for training localization professionals. We will discuss this and other issues connected with the formation of professional translators for today.
When discussing the relevance of technological training in the translation curricula, it is important to clarify the factors that make technology more indispensable and show how the training should be tuned accordingly. The relevance of technology will depend on the medium that contains the text to be translated. This particular aspect is becoming increasingly evident with the rise of the localization industry, which deals solely with information in digital form. There may be no other imaginable means for approaching the translation of such things as on-line manuals in software packages or CD-ROMs with technical documentation than computational ones.
On the other hand, the traditional crafts of interpreting natural speech or translating printed material, which are peripheral to technology, may still benefit from technological training slightly more than anecdotally. It is clear that word processors, on-line dictionaries and all sorts of background documentation, such as concordances or collated texts, besides e-mail or other ways of network interaction with colleagues anywhere in the world may substantially help the literary translator's work. With the exception of a few eccentrics or maniacs, it will be rare in the future to see good professional interpreters and literary translators not using more or less sophisticated and specialized tools for their jobs, comparable to the familiarization with tape recorders or typewriters in the past. In any case, this might be something best left to the professional to decide, and may not be indispensable.
However, the greater number of jobs for our students is in the localization market. Information of many types is rapidly changing format and going digital. Electronic documentation is the adequate realm for the incorporation of translation technology. This is something that young students of translation must learn. As the conception and design of technical documentation becomes progressively influenced by the electronic medium, it is integrating more and more with the whole concept of a software product. The strategies and means for translating both software packages and electronic documents are becoming very similar and both are now, as we will see, the goal of the localization industry.
The increase of information in electronic format is linked to advances in computational techniques for dealing with it. Together with the proliferation of informational webs in Internet, we can also see a growing number of search and retrieval devices, some of which integrate translation technology. Technical documentation is becoming electronic, in the form of CD-ROM, on-line manuals, intranets, etc. An important consequence of the popularization of Internet is that the access to information is now truly global and the demand for localizing institutional and commercial Web sites is growing fast. In the localization industry, the utilization of technology is congenital, and developing adequate tools has immediate economic benefits.
The main role of localization companies is to help software publishers, hardware manufacturers and telecommunications companies with versions of their software, documentation, marketing, and Web-based information in different languages for simultaneous worldwide release. The recent expansion of these industries has considerably increased the demand for translation products and has created a new burgeoning market for the language business. According to a recent industry survey by LISA (the Localization Industry Standards Association), almost one third of software publishers, such as Microsoft, Oracle, Adobe, Quark, etc., generate above 20 percent of their sales from localized products, that is, from products which have been adapted to the language and culture of their targeted markets, and the great majority of publishers expect to be localizing into more than ten different languages.
Localization is not limited to the software-publishing business and it has infiltrated many other facets of the market, from software for manufacturing and enterprise resource planning, games, home banking, and edutainment (education and entertainment), to retail automation systems, medical instruments, mobile phones, personal digital assistants (PDA), and the Internet. Doing business in an integrated global economy, with growing electronic transactions, and world wide access to products and services means an urgent need to break through language barriers. A prediction of $220 billion online spending by 2001 shows the potential of this new market. It means that product information, from purchasing procedures to user manuals, must be made available in the languages of potential customers. According to the latest surveys, there are more than 35 million non-English-speaking Internet users. Internet is thus evolving into a huge consumer of Web-based information in different languages. The company Nua Ltd. provides a good example of how the demand for multilingual Web-sites is changing the notion of translation into localization. Nua has recently won a substantial contract to develop and maintain a searchable multilingual intranet for the American Export Group (AEG), a division of Thomas Publishing International. Nua's task is to transform the existing American Export Register (AER), a directory of some 6,000 pages, into a localized database of 45,000 company listings, with information about each company, including a categorization into one of AEG's 5,000 categories. AEG's intranet will link 47,000 US firms to overseas clients. The first version of the AER register will provide access in five languages: English, French, German, Spanish, and Portuguese. Russian is due to follow, and the company hopes eventually to have an Arabic version. Any such multilingual service involves frequent revisions and updates, which in turn means a high demand for constant localization effort.
Besides Internet, another emerging sector for the localization industry is the introduction of the e-book (electronic book) in the literary market. Microsoft, Bertelsmann, HarperCollins, Penguin Putnam, Simon & Schuster, and TimeWarner Books have launched a new association for standardizing the format of electronic books. Although there may be doubts about whether we will ever be able to bring the electronic page into line with the printed page in terms of readability and ease of use, it is clear that for a new generation of console and video-game users, who are more than adapted to reading on screens, literature on the console may be more than appealing.
To understand the relevance of the localization market we can look at some figures provided by companies in the field. AlpNet, for example, who claims to be the largest publicly owned dedicated supplier of worldwide translation and product localization services, with over 375 employees in 13 countries, has recently reported sales of US$10.4 million in one quarter of 1997, with net income of US$619,000. In addition to AlpNet, here are some more names of buoyant companies in the localization business: International Software Products, EnCompas Globalization, Lernout & Huaspie, Flanders Language Valley, Vertaalbureau Bothof , Intertrans, Bowne Global Solutions, LionBridge Technologies, Language Management International, International Language Engineering, Techno-Graphics & Translations Accent Software International Ltd.. The specialized magazine Language International, with six issues a year, is a good source of information to find out more about these companies. Many claim to have problems recruiting people. The General Manager of LionBridge, Santi van der Kruk, for example, declares:
The profile we look for in translators is an excellent knowledge of computer technology and superb linguistic ability in both the source and target languages. They must know how to use the leading CAT [computer assisted translation] tools and applications and be flexible. The information technology and localization industries are evolving very rapidly and translators need to move with them.
Vand der Meer, president of AlpNet, puts it this way:
Localization was originally intended to set software (or information technology) translators apart from 'old fashioned' non-technical translators of all types of documents. Software translation required a different skill set: software translators had to understand programming code, they had to work under tremendous time pressure and be flexible about product changes and updates. Originally there was only a select group--the localizers--who knew how to respond to the needs of the software industry. >From these beginnings, pure localization companies emerged focusing on testing, engineering, and project management.
This shows that the localization market is requiring an expertise that the vast majority of academic centers is not properly providing. This state of affairs explains why the localization industry itself, around the LISA association, has seen the need to promote an educational initiative.
LISA Education Initiative Taskforce (LEIT) is a consortium of schools training translators and computational linguists that was announced in 1998 as an initiative to develop a promotional program for the academic communities in Europe, North America, and Asia. The initial mandate of LEIT was to conduct a survey among academic and non-academic programs that offer courseware and training for internationalizers and localizers and to query the market players to determine their needs with respect to major job profiles. LEIT's main objective is to stimulate more formal education in skills beneficial to the localization industry that complains of a labor shortage. The academic institutions involved in the first release of LEIT are: University of Geneva (Switzerland), Brigham Young University (Utah), Kent State University (Ohio), University of Cologne (Germany), City College of Dublin (Ireland), Monterey Institute of International Studies (California), and National Software Center in Bombay (India).
Professor Margaret King of Geneva University described the first step of the project as consisting of the "clarification of the state of affairs and to plan courses that are comprehensive enough to cover all aspects of interest of the localization industry, to review all aspects of the localization industry, from translation and technical writing through globalization, internationalization, and localization". The definition of the critical terms involved was a contentious topic, although there seems to be a consensus with the following:
The main goal of the LEIT initiative is to introduce localization courseware into translation studies, with versions ready for the start of the 1999 academic year. However, this must be done with care. Bert Esselink (1998), from AlpNet, for example, argues against separating localization from other disciplines and claims its basic principles should be covered in all areas of translation training. Furthermore, it would be useful to add the trainers not only need constant feedback and guidance from the commercial sector, they also need to maintain close contact with the software industry. So, perhaps, one of the best features of the LEIT initiative is its combination of partnership from the academic as well as from the industry world. LISA offers the first version of this courseware on its Web-site and users have the possibility to contact the LEIT group and collaborate through an on-line questionnaire.
Until comparatively recent times, there has been a gulf between what translation software developers provided for the translator and what in fact the translator needed. This has brought about many confronted views in connection with translation technology. Mark Homnack, president of SimulTrans poses the following questions (Language International 10.6):
How much money has been lost in the pursuit of translation technology? Ask the developers, and their investors. How much benefit has been gained? Ask the clients who have bought their tools. To what extent should we believe the analysts' fantastic forecasts about translation technologies? Ask the people who have managed significant translation activities for five years or more.
He goes on to say:
I have just given up hope of any universal translation technology. After listening to the eternal and forever broken promises of Weidner 20 years ago, Systran 15 years ago, Logos 10 years ago, XL8 five years ago, and now Mendez's Internet ploys, I have grown cynical about the sprite hopes of Spring.
Yet machine translation is never plug-and-play. It requires a huge effort in preparation, evaluation, and maintenance. Suitability of technology depends on many factors, but fundamentally text type. Without these considerations, the technology may be seen as a fiasco. Few informed people still see the original ideal of fully automatic high-quality translation of arbitrary texts as a realistic goal. Translation technology suppliers are now working under the assumption that, rather than batch processes, man-machine interaction together with the integration of tools into the translator's working environment is the solution.
There are recent briefs reporting achievements of faster turnaround, and substantial cost reductions resulting from an adequate use of translation technology. The Canadian Mitel company, with the Logos machine translation (MT) system, or the Dutch Baan, with an intelligent combination of Logos and the translation memory (TM) manager Transit, are both claiming time and quality improvements in their document production and translation cycles, with cost reductions of 40% to 60%. These companies have well established translation and document management departments which have incorporated new technology with extreme caution. However, despite the difficulties, the demand for translation technology is constantly growing. The British consulting company Ovum Ltd reported in 1996 that the world market for translation products in 1995 had been $200 million, half of which had been in the Japanese market.
Leaving behind the old conception of a monolithic compact translation engine, the industry is now moving in the direction of integrating systems: "In the future Trados will offer solutions that provide enterprise-wide applications for multilingual information creation and dissemination, integrating logistical and language-engineering applications into smooth workflow that spans the globe," says Trados manager Henri Broekmate. Logos, the veteran translation technology provider, has announced "an integrated technology-based translation package, which will combine term management, TM, MT and related tools to create a seamless full service localization environment." Other software manufacturers also in the race are Corel, Star, IBM, and the small but belligerent Spanish company Atril. This approach for integrating different tools is largely the view advocated by many language-technology specialists. Below is a description of an ideal engine which captures the answers given by Muriel Vasconcellos (from the Pan American Health Organization), Minako O'Hagan (author of The Coming Age of Teletranslations) and Eduard Hovy (President of the Association of Machine Translation in the Americas) to a recent survey (by Language International 10.6). The ideal workstation for the translator would combine the following features:
Eduard Hovy underlines the need for a genre detector. "We need a genre topology, a tree of more or less related types of text and ways of recognizing and treating the different types computationally." He also sees the difficulty of constantly up-dating the dictionaries and suggests a "restless lexicon builder that crawls all over the Web every night, ceaselessly collecting words, names, and phrases, and putting them into the appropriate lexicons."
Muriel Vasconcellos pictures her ideal design of the workstation in the following way:
The above lines depict a view of a translation environment which is closer to more traditional needs of the translator than to current requirements of the industry. Many aspects of software localization have not been considered, particularly the concepts of multilingual management and document-life monitoring. Corporations are now realizing that documentation is an integral part of the production line where the distinction between product, marketing and technical material is becoming more and more blurred. Product documentation is gaining importance in the whole process of product development with direct impact on time-to-market. Software engineering techniques that apply in other phases of software development are beginning to apply to document production as well. The appraisal of national and international standards of various types is also significant: text and character coding standards (e.g. SGML/XML and Unicode), as well as translation quality control standards (e.g. DIN 2345 in Germany, or UNI 10574 in Italy).
In response to these new challenges, localization packages are now being designed to assist users throughout the whole life cycle of a multilingual document. These take them through job setup, authoring, translation preparation, translation, validation, and publishing, besides ensuring consistency and quality in source and target language variants of the documentation. New systems help developers monitor different versions, variants and languages of product documentation, and author customer specific solutions. An average localization package today will normally consist of an industry standard SGML/XML editor (e.g. ArborText), a translation and terminology toolkit (Trados Translator's Workbench), and a publishing engine (e.g. Adobe's Frame+SGML).
Unlike traditional translators, software localizers may be engaged in early stages of software development, as there are issues, such as platform portability, code exchange, format conversion, etc. which if not properly dealt with may hinder product internationalization. Localizers are often involved in the selection and application of utilities that perform code scanning and checking, that automatically isolate and suggest solutions to National Language Support (NLS) issues, which save time during the internationalization enabling process. There are run-time libraries that enable software developers and localizers to create single-source, multilingual, and portable cross-platform applications. Unicode support is also fundamental for software developers who work with multilingual texts, as it provides a consistent coding format for international character sets.
In the words of Rose Lockwood (Language International 10.5), a consultant from Equipe Consortium Ltd, "as traditional translation methods give way to language engineering and disciplined authoring, translation and document-management methods, the role of technically proficient linguists and authors will be increasingly important to global WWW. The challenge will be to employ the skills used in conventional technical publishing in the new environment of a digital economy."
Having said all this, it is important to reassess the human factor. Like cooks, tailors or architects, professional translators need to become acquainted with technology, because good use of technology will make their jobs more competitive and satisfactory. But they should not dismiss craftsmanship. Technology enhances productivity, but translation excellence goes beyond technology. It is important to delimit the roles of humans and machines in translation. Martin Kay's (1987) words in this respect are most illustrative:
A computer is a device that can be used to magnify human productivity. Properly used, it does not dehumanize by imposing its own Orwellian stamp on the products of human spirit and the dignity of human labor but, by taking over what is mechanical and routine, it frees human beings over what is mechanical and routine. Translation is a fine and exacting art, but there is much about it that is mechanical and routine, if this were given over to a machine, the productivity of the translator would not only be magnified but this work would become more rewarding, more exciting, more human.
It has taken some 40 years for the specialists involved in the development of MT to realize that the limits to technology arise when going beyond the mechanical and routine aspects of language. From the outside, translation is often seen as a mere mechanical process, not any more complex than playing chess, for example. If computers have been programed with the capacity of beating a chess master champion such as Kasparov, why should they not be capable of performing translation of the highest quality? Few people are aware of the complexity of literary translation. Douglas Hofstadter (1998) depicts this well:
A skilled literary translator makes a far larger number of changes, and far more significant changes, than any virtuoso performer of classical music would ever dare to make in playing notes in the score of, say, a Beethoven piano sonata. In literary translation, it's totally humdrum stuff for new ideas to be interpreted, old ideas to be deleted, structures to be inverted, twisted around, and on and on.
Although it may not be perceived at first sight, the complexity of natural language is of an order of magnitude far superior to any purely mechanical process. To how many words should the vocabulary be limited to make the complexity of producing "free sonnets" (that is, any combination of 6 words in 14 verses) comparable to the number of possible chess games? It may be difficult to believe, but the vocabulary should be restricted to 100 words. That is, making free sonnets with 100 words offers as many different alternatives as there are ways of playing a chess game (roughly, 10120; see DELI's Web page for discussion). The number of possibilities would quickly come down if combinations were restricted so that they not only made sense but acquired some sort of poetic value. However, defining formally or mechanically the properties of "make sense" and "have poetic value" is not an easy task. Or at least, it is far more difficult than establishing winning heuristics for a color to succeed in a chess game. No wonder then that Douglas Hofstadter's MT experiment translating 16th century French Clément Marot's poemMa Mignonne into English using IBM's Candide system should have performed so badly (see Sgrung's interview in Language International 10.1) :
Well, when you look at [IBM's Candide's] translation of Ma Mignonne, thinking of Ma Mignonne as prose, not as poetry, it's by far the worst. It's so terrible that it's not even laughable, it just stinks! It's pathetic!
Obviously, Hofstadter's experiment has gone beyond the recommended mechanical and routine scope of language and is therefore an abuse of MT. Outside the limits of the mechanical and routine, MT is impracticable and human creativity becomes indispensable. Translators of the highest quality are only obtainable from first-class raw materials and constant and disciplined training. The potentially good translator must be a sensitive, wise, vigilant, talented, gifted, experienced, and knowledgeable person. An adequate use of mechanical means and resources can make a good human translator a much more productive one. Nevertheless, very much like dictionaries and other reference material, technology may be considered an excellent prothesis, but little more than that. As Martin Kay (1992) argues, there is an intrinsic and irreplaceable human aspect of translation:
There is nothing that a person could know, or feel, or dream, that could not be crucial for getting a good translation of some text or other. To be a translator, therefore, one cannot just have some parts of humanity; one must be a complete human being.
However, even for skilled human translators, translation is often
difficult. One clear example is when linguistic form, as opposed to
content, becomes an important part of a literary piece. Conveying the
content, but missing the poetic aspects of the signifier may considerably
hinder the quality of the translation. This is a challenge to any
translator. Jaime de Ojeda's (1989) Spanish translation of Lewis Carroll's
Alice in Wonderland illustrates this problem:
Twinkle, twinkle, little bat |
Brilla, luce, ratita alada |
Manuel Breva (1996) analyzes the example and shows how Ojeda solves the "formal hurdles" of the original:
The above lines are a parody of the famous poem "Twinkle, twinkle, little star" by Jane Taylor, which, in Carroll's version, turns into a sarcastic attack against Bartholomew Price, a professor of mathematics, nicknamed "The Bat". Jaime de Ojeda translates "bat" as "ratita alada" for rhythmical reasons. "Murciélago", the Spanish equivalent of "bat", would be hard to fit in this context for the same poetic reasons. With Ojeda's choice of words the Spanish version preserves the meaning and maintains the same rhyming pattern (AABB) as in the original English verse-lines.
What would the output of any MT system be like if confronted with this fragment? Obviously, the result would be disastrous. Compared with the complexity of natural language, the figures that serve to quantify the "knowledge" of any MT program are absurd: 100,000 word bilingual vocabularies, 5,000 transfer rules.... Well developed systems such as Systran, or Logos hardly surpass these figures. How many more bilingual entries and transfer rules would be necessary to match Ojeda's competence? How long would it take to adequately train such a system? And even then, would it be capable of challenging Ojeda in the way the chess master Kasparov has been challenged? I have serious doubts about that being attainable at all. But there are other opinions, as is the case of the famous Artificial Intelligence master, Marvin Minsky. Minsky would argue that it is all a matter of time. He sees the human brain as an organic machine, and as such, its behavior, reactions and performance can be studied and reproduced. Other people believe there is an important aspect separating organic, living "machines" from synthetic machines. They would claim that creativity is in life, and that it is an exclusive faculty of living creatures to be creative.
Away from such metaphysical dilemmas, what I personally expect are systems that learn while they are exposed to translations like Ojeda's; systems that are capable of memorizing any bilingual chunk which may be considered a translation unit. Sometimes the translation unit will correspond to just a word or a phrase, like "bat" and "ratita alada", but more often whole paragraphs or even entire literary works could be taken as translation units. One can think of systems that, when confronted with a text which contains an occurrence of Lewis Carroll's parody of Jane Taylor, would be clever enough to resort to Ojeda's translation, and not only use "ratita alada" instead of "murciélago", but provide the whole verse if needed.
More remarkably than Carroll, Shakespeare may be taken as the literary author who is most frequently quoted or paraphrased in English. There are two established translators of Shakespeare into Spanish, Astrana Marín and Ángel Luis Pujante (Rupérez 1998). Astrana had been the main reference until Pujante's versions were published in 1998. Astrana translated Shakespeare in prose, with frequent paraphrases and explanations of the source text. Pujante tries to maintain as much poetic effect as possible not only in rhyme and rhythm, but also with the archaic flavor of the original. One would like to see a system with expertise in Shakespeare's translations into Spanish, where both Astrana and Pujante's versions were registered, together with other known alternatives, including Carmen Criado's work in the Spanish dubbed version of the film Shakespeare in love. It would be an attractive content for an e-book, as is in fact Francisco Rico's CD-ROM with collations and the final revised version of El Quijote in Spanish. Such an ideal MT system, rather than competing with Astrana's, or Pujante's translating skills, would just be able to reproduce their versions in one's own working environment, word by word, through the simple stroke of a key. And this is completely within the state of the art in translation technology.
What should students learn about translation technology? As we now know, there is no one single answer to this question. Technological skills will depend on how students see their own future as translators. Those with good aptitudes for interpreting or literary translation could leave technology on a secondary level. However, it is clear that the vast majority of students should be prepared to satisfy the growing demand for specialists in technical documentation, and in particular the demand from the localization industry. Thus, training centers should seriously consider introducing the LEIT initiative into their training curricula.
Apart from a basic common computational background these would include official and industrial standards in offimatics (with word-processing, data-base maintenance, spreadsheet management, Internet browsing, emailing, etc.); students should have realistic knowledge of some specific translation technology, ideally in the form of a translation workstation. However, it is important to realize that software is constantly evolving, software and hardware up-dating is expensive, and that key concepts and skills may be equally well acquired with tools which are two or three years old. What is most important is becoming competent with the basic functional operations such as file and window management, editing, and net interaction. More specialized operations will be easily acquired on top of the basic ones, and will largely depend on the student's natural sympathy for the computer. I would recommend at least one year of basic computer training before attempting any specialization.
It is thus important to tune training courses to the expectations of the students. Out of the following six options, any person with a University degree in Translation should be qualified at least to be able to carry out the first three of the following:
Thus, it can be seen that the traditional role of the translator will be changing very quickly and a direct consequence of this is that there will be more career opportunities for the graduate in Translation Studies than ever before. A recent survey, done by the Department of Languages at the University of Applied Sciences in Cologne provoked the following comments by Michael Grade, Professor of Technical English (Language International 10.4):
Career prospects are favorable for technical translation graduates with further job qualifications. The chances of finding a language-related job in scientific, medical, or technical fields appear high. The result of the survey indicated that the most important area of activity was the international export and sales sector for technical products and services. The job description now includes a diversified range of activities such as commercial and specialized technical tasks, customer relations, clerical, and organizational responsibilities.
So this, together with the fact that the market is beginning to
recognize the value of the translator in the world today, seems to augur
well for the future.
* El autor agradece la oportunidad para concebir y escribir este artículo a Richard Samson, sin cuyos continuos y tenaces requerimientos no hubiera sido concluido. Larry Muies ha contribuido de manera decisiva a que el texto tenga un nivel de inglés tolerable, introduciendo además geniales toques de cohesión y estilo. Si todavía quedan errores, serán retoques míos de última hora. Inés Jacob, JosuKa Díaz Labrador, Andoni Eguiluz y Manuel Breva han aportado sabios consejos y comentarios. Los datos que no han sido debidamente documentados provienen de los números de Language International de los dos últimos años, a cuyo editor y redactores debo poder mantenerme al día.
Manuel Breva. 1996. Some pragmatic, form, and content difficulties in the field of translation. ES. Revista de filología inglesa. 20:119-131.
Lewis Carroll. 1970. The Annotated Alice. Alice's Adventures in Wonderland and Through the Looking Glass. Edited by Martin Gardner. Penguin Books. Spanish version 1989. Alicia en el país de las maravillas. Translation and commentary by Jaime de Ojeda. Seix Barral.
Deusto's Club of Linguists, Computer Scientists and Engineers (DELI). Complejidad: partidas de ajedrez vs. sonetos libres. http://orion.deusto.es/~abaitua/kanpetzu/recicla/JORGEW.HTM
Bert Esselink. 1998. A practical guide to software localization. John Benjamins.
Douglas Hofstadter. 1998. Le Ton beau de Marot: In Praise of the Music of Language. HarperCollins Publishers.
Martin Kay. 1992. Foreword. In W. John Hutchins & Harold L. Somers. An Introduction to Machine Translation. Academic Press.
Martin Kay. 1997. The Proper Place of Men and Machines in Language Translation. Machine Translation 13:3-23.
Localization Industrial Standards Association (LISA): http://www.lisa.unige.ch
LISA Education Initiative Taskforce (LEIT): http://www.ttt.org/leit/
Vand der Meer. 1997. On localization. Language International 9.5:43.
Ángel Rupérez. 1998. Shakespeare Revisado. Babelia-El País 8.8.1998:12.
Robert Sprung. 1998. A word with Hofstadter. Language International 10.1:32-36.