Accreditation and Standards in the Translation Industry

By Roger Chriss,
Language Realm,
U.S.A.

rbchriss@languagerealm.com
www.languagerealm.com

Get the List of 4,400+ Translation Agencies Now!
No Recurring Membership Fees!

The translation industry is slowly climbing a tortured path toward regulation and accreditation in the United States, with the ASTM meeting now to come up with nationally recognized standards for translation, LISA issuing its own ideas about what constitutes good practice in localization, and every translator, translation vendor, translation school, and translation organization adding thoughts and suggestions to this process. But has anyone stopped to ask if this is a good idea, if the industry will really benefit from accreditation or regulation, and who might suffer? That's the point of this article: to take a close look at these two closely-related issues and explore what I think are some overlooked problems.

Accreditation

For years now, accreditation has been regularly discussed, and even hotly debated, among translators, at translation and localization conferences, and even in the federal and state governments. Today there is no form of universally recognized accreditation for translators in the United States, even though we do have such options as ATA accreditation, degrees or certificates from various academic institutions, and qualifying exams from federal and state governments, and from organizations like United Nations.

There is at present slow but steady progress toward government-regulated nationwide accreditation for translators. The ASTM (American Society for Testing and Materials) Subcommittee F15.48 on Language Translation was officially established in September, 1998, and has been meeting quarterly or more often since that time toward the end of creating guidelines for translation and localization. The ASTM's Guide for User-Oriented Foreign Language Instruction was approved in March, 1995, and a guide being developed by ASTM Subcommittee F15.34 on Language Interpretation is scheduled for voting in the summer of 2000. In addition, LISA (Localization Industry Standards Association) has been meeting and discussing standards for language translation in localization. The ATA a few years ago made its accreditation exams into a two-step process, and has added more languages to its roster as well. In sum, many of the right organizations, with the participation of translators, educators, and a translation vendors, are establishing or expanding efforts at standardization or accreditation.

The goal seems to be twofold: first, a nationally recognized and accepted set of standards for translators and translation; second, an accreditation exam for translators, one that would certify a translator as being capable of translating from one particular language into another. Many translators believe that such an exam would benefit them as well as their industry, increasing among other things pay rates, ease of finding work, and respect as professionals. Many translation vendors belief that such an exam would be useful insofar as it would make finding qualified and capable translators easier.

I believe neither of these things.

First, the exam itself. In theory, an accreditation exam sounds like a brilliant idea. But just as the idea of a benevolent monarch being the best form of government has no place in reality, as amply demonstrated by John Stuart Mill, this exam is a practical impossibility. Let's take a close look at the assumptions underlying this exam, and the inevitable, and I will argue insurmountable, problems that come with them.

The exam will have to exist for every language combination in use in the United States. Even if we assume that these combinations consist of only English and another language, rather than say Japanese to Spanish or Russian to Italian, we are still dealing with a lot of exams, depending on which languages are deemed worthy of testing. At the very least, all United Nations languages (Arabic, Chinese, French, Russian, Spanish, and the two alternatives: German and Japanese) as well as obvious languages like Vietnamese, Korean, Portuguese, Italian, and so forth will have to be tested. Moreover, there will have to be two tests for each language combination, one for each direction, because translators translator from their B language into their A language. So we're already dealing with at least 22 exams.

Since most translators specialize in one or a couple of related fields, the exam will either have to be very general, and therefore not particularly useful insofar as it will not test terminology, subject area knowledge, or the ability to write in an appropriate style, or the exam will have to exist for several fields, such as law, finance, medicine, physical sciences, social sciences, computer science, and so on. For this exam to fulfill its intended purpose as a way of identifying who can really translate what, the exam will likely have to exist for various area specializations. Even if we assume there are only five subject specializations (business, law, finance, medicine, science), the number of exams is now at least 110.

The exam will also have to test skills beyond translation itself, such as the ability to use a word processor like Microsoft Word well, to work with MT or MAT software systems, to manage or create terminology databases, navigate HTML files and manipulate them as necessary, or to perform DTP in an application like QuarkXPress. To what extent are these skills necessary? Should they be evaluated? Is it realistic to accredit a translator who cannot open an HTML file, doesn't know how to create a table in Microsoft Word, or has never used a database application? If these skills are considered a part of a translator's ability to translate, and I suspect they are, at least by some people in the industry, then they should be tested. Doing so will of course augment the difficulties of creating the 110 or more exams needed.

There is also the matter of length. Most translation agencies and vendors prefer to give two or three paragraph tests, just a small way of seeing if a translator can actually translate. No one considers such tests to be comprehensive, and no one would regard passing one of them as anything remotely related to accreditation. By contrast, the graduation exams, called the Professional Exams, at the Monterey Institute of International Studies, take two full days to complete. Since part of being a professional translator, freelance or in-house, is being able to maintain a certain level of productivity for an entire day, it seems reasonable to have an accreditation exam take at least half a day, during which time a candidate might have to translate one to two-thousand words, depending on the language and subject matter. Again, the 110 or more exams will be that much harder to create the longer they have to be.

The exams will also have to be updated, if not created anew, every year or so not only to reflect changes in language and in the subject specializations, but also to avoid cheating. By cheating I refer not only to people who deliberately obtain a copy of test content before a test so as to improve performance but also to people who merely remember unconsciously the content of a test. These latter people are important because they might inadvertently leak out information about the test, or they might remember the test when they take it again, either because they failed the first time or because they have to take it again in order to maintain their accreditation. So an accreditation exam is not just a matter of creating one exam one time only, but creating more than 100 exams every year.

Who will create these exams? Obviously each exam will have to be created by a group of people based on a consistent set of standards. Some people might be able to participate in the creation of more than one exam, but for each exam I think we can safely assume that at least three people will be needed to come up with a fair and impartial test for translators. That means hundreds of people, at least. And who will grade the exams? Depending on the number of people taking the exams, large teams of graders might be necessary to evaluate the exams in a reasonable period of time. We can't have translators waiting for the results of such an exam for six months or a year. Since a translation exam will by definition have to involve translating and cannot simply be a fill-in-the-blank or a multiple-choice test, grading will be a fairly labor-intensive task. There will also have to be an organization which oversees the creation of the exams, the implementation of testing, an issue to be described at length below, the grading of the exams, and the distribution of grades so that people who claim to have passed can prove they did, and people who need to verify that a translator has passed the exam can do so, too. This organization will have to act in accordance with a set of standards defined by a separate body, such as the ASTM, and will also have to have a means for handling grievances. Inevitably someone will claim that the test was not fair in content or implementation, or was not evaluated justly, and for the exam to be considered useful overall, such grievances have to be heard.

Evaluation of the exams will be especially difficult. It is well accepted that there is no such thing as a perfect translation, that many phrases and idioms can be rendered in two or more ways, and that style is a highly subjective matter. As such, fairness in evaluation will be extremely challenging. The preferences, preconceptions, and experiences of an evaluator will undoubtedly impact at some level, conscious or otherwise, on the grading process, and could potentially affect whether or not a candidate is accredited. Some mechanism to insure fairness in evaluating the exam will be vital so as to assure the people taking the exam that they can expect to pass if they deserve to, and to assure those who use the accreditation as a benchmark for translation ability that it does in fact measure what it claims to. There may even be a need to evaluate the evaluators, but that leads to an infinite regression, which is something best avoided.

Qualifications for taking the exam will be another important issue. Who can take it? Anyone who signs up? Will there be minimum educational requirements, perhaps stemming from whatever standards for translation and translators are put into practice in this country? If such requirements do exist, who will decide how a degree or certification from another country measures against the American educational system? And how will claims by foreign-born and educated translators be verified? We all know what the CIA and FBI go through to perform background checks on American citizens who have lived abroad and then need to be cleared to work for those organizations. Will a similar system be set in place to verify the foreign education of translators, people who almost by definition have such experience? Or will the word of the person be taken at face value, opening the system to at least the possibility of abuse? Finally, will there be exemptions for the exam? Will a person who holds a M.A. from Kent State or the Monterey Institute, or who has passed the U.N. translation exams, or has a certificate from the translation program at Georgetown University, be exempted?

And what about re-testing? Will translators be required to take the exam again after a certain number of years pass? Will sufficient work experience or continuing education be required to take the test again, or might it replace the need to be re-tested? Translators certainly need to maintain their language skills and keep up with the subjects they work in. But who will decide to what extent either is necessary? Should translators be expected to prove their ability again every so many years, even if all their clients are satisfied with their performance?

Even assuming that a fair and impartial test could be created and evaluated, there is still another category of difficulties. Like so many things in life, the idea is the easy part, and the implementation is where the problems really begin.

Where will these accreditation exams be held? Washington D.C. might seem like a logical choice, though as I live in Monterey, California, I personally would prefer, oh, say, Monterey, though readers living in Boston might not like my preference. If there is only one testing center, at least some translators are going to incur serious costs, both in time and money, to take this exam. The testing center itself may pose serious problems, too, as it will have to accommodate a large number of people and their translation resources; it will probably have to have a computer station of some sort for each person taking the test; and it will have to be set up so that people cannot cheat. I've taken and proctored enough standardized tests in my life to be aware of the difficulties associated with this kind of testing.

There is also the question of what constitutes a fair and reasonable testing environment for such an exam. Having translators write out their exam by hand while using print dictionaries they bring themselves may seem practical, but it is unrealistic. Translators work on computers, they word processes their translations, they use the Web to find terminology, they even work with MAT software. Again, standards might help answer these questions, but the matter of fairness, real and perceived, will linger.

What will these exams cost to take? How much is accreditation worth and who will pay for it? Clearly the cost or creating, evaluating, and implementing this exam will be formidable, and though translators may not be asked to pay but a small portion of this, I suspect that translation vendors will pass on the costs indirectly through reduced rates. And what about the expenses mentioned above for the translator who has to travel across country, perhaps stay in a hotel for a day or two, and then take the exam? Accreditation had better have a significant impact on a translator's income to justify the travel expenses, not to mention lost work time. And this holds for in-house as well as freelance translators. If I were the president of a translation agency and were satisfied with my translators, why would I bother sending them out for a couple of days or so to take a test? Why not just give them a bonus for a job well done, save money, and make my translators happy?

Next problem: how do we keep track of who has taken the test, who has passed, and who has failed? Will such a list be available to the public on the Internet, or will one have to file a request for such information about a translator? What will happen when some translator forges accreditation, as inevitably someone will? If an individual can forge a medical license and practice medicine without ever having been to medical school, which people have managed to do in the U.S., why not a translator? Will there be consequences? If so, what? If not, then one of the values of accreditation is lost.

All right, enough of the problems associated with an accreditation exam. I hope I have suggested that there are as yet more than a few fundamental hurdles associated with nationwide accreditation, enough so that I believe an accreditation system is a futile exercise that will merely complicate our industry. But there is another side to this, one that we should look at briefly before leaving this topic.

Results

Will anyone care if a translator is accredited? Will that really lead to greater opportunity in terms of word rates for freelance translators or salary for in-house translators? Will accredited translators get more interesting jobs, survive corporate restructuring or streamlining more readily, get better benefits or enjoy more rapid career advancement? Of course this question is impossible to answer with certainty without implementing an accreditation exam, but I think we can make some educated guesses as to what would happen.

The development we can virtually count on will be a gray market for translators. Ultimately in the translation industry what counts is the ability to translate. Translation vendors have never expressed much interest in how or where a translator acquired translation skills, as long as the translator can provide some form of proof of competence. The most widely accepted form nowadays is experience, after that comes some form of formal education, and following that stands an accreditation exam like the ATA's.

There is no reason for a translation vendor not to use an unaccredited translator whose competence is otherwise established. In fact, if the unaccredited translator is willing to work at a slightly reduced rate, or if accredited translators are somehow in a position to demand a premium for their accreditation, then that ultimate business consideration will motivate many translation vendors to choose the unaccredited translator: cost savings. Unless and until accreditation becomes widely recognized and accepted as a sign of superior quality, as for instance ISO-9000 certification is within some industries, many vendors will likely have little preference regarding accreditation, and some may choose unaccredited translators for the cost savings.

There is the related issue of how to monitor translation vendors and their hiring practices. Just as a homeowner can choose to hire a union or non-union plumber with no consequences other than the potential risks in terms of quality of work performed, a vendor would likely not incur any risks beyond those associated with the translation ability of the translator in question, freelance or in-house. Even if laws were enacted, on a state and federal level, requiring vendors to use accredited translators when and where possible, they could never be enforced, no more than prohibition, union hiring practices, or speed limits can be consistently and thoroughly. So translation vendors can make the choices they prefer, accepting the risks associated with each as they like.

Let us also remember here that there is no effective way for translators to collectively demand higher pay, either in the form of increased salary for in-house translators or higher word rates for freelancers. Translators are barred from unionizing, and the ATA is barred from setting rates for translation, both as a result of decisions made by the Justice Department in the early 1990s. It is unlikely that this state of affairs will change, regardless of the quality accreditation guarantees.

So if there is no more or less money to be had from accreditation, perhaps a freelance translator can expect more work, and an in-house translator can expect faster promotion. In the former case, most freelance translators with a few years experience have as much work as they can handle on average, so more work would be irrelevant. And though word rates might rise for accredited translators, if the vendors are bearing some of the costs associated with accreditation, then word rates could just as easily fall. In the latter case, the differences in pay between a new translator and a senior translator, or for that matter a senior translation manager, are between $10,000 and $20,000 on average, though sometimes considerably less. Thus, if accreditation costs a translator $1,000 in terms of the fee for the exam, travel expenses to take the exam, and other direct and indirect expenses and losses associated with taking the exam, then the promotions have to come quickly enough to justify this. Quantifying this to a precise degree is not possible in this article, but the example above does suggest that accreditation could not have too much value in a corporate environment.

To sum up, there is no particular motivation for translation vendors to prefer accredited translators unilaterally. Given the choice between an accredited translator and a translator with no training, experience, or credentials, most vendors would pay the extra for the former individual. Given the choice between a seasoned translator with formal academic training but no accreditation and a translation newbie with no experience but a newly-minted accreditation, I suspect most vendors would work with the former individual. Combine this reasoning with the practical impossibility of creating and implementing an accreditation exam and the system to back it up, and I am convinced that a nationwide accreditation system run by the government is a bad idea with few if any possible benefits to translators or the translation industry.

Can We Avoid Accreditation? An Alternative

So if you agree with the above argument, or if you are curious to see if I am just criticizing the system, then I am happy to report that I do have a few thoughts as to how we can avoid accreditation while still improving the translation industry for translators and translation vendors, and raising translation quality.

The translation industry at present is a meritocracy. Those who can do the work get more work; those who cannot do the work are given the chance to learn if they so choose, and once they can do the work, they get more work. All others are rapidly dismissed from the profession. In the translation industry, all translators, regardless of background, education, or accreditation, start at the bottom. You have to prove yourself and your ability to each new client or employer, though this process becomes ever shorter the longer you stay in the profession. You prove your ability by taking vendor translation tests and by doing quality translation work. Degrees, certificates, and accreditation seem mostly to help accelerate the speed with which you get your first translation job and become one of the primary translators for an organization if you are a freelancer, or the ease with which you land your first in-house position and perhaps how quickly you are promoted.

Meritocracies are good, I believe, because not only do they recognize and reward quality work while punishing inferior efforts, but also because they give fair access to all who want to enter, and as a result, force those who are in to keep their skills sharp. Just because I am an established translator with a clientele who seems to appreciate my ability doesn't mean I can slack off. I still have to maintain and improve my language skills, keep up with changes and developments in the subjects I work in, and upgrade or extend my arsenal of business tools. I am competing with both other established translators who can do what I can, and would, I presume, be perfectly willing to step in and take my place with my clients, as well as newcomers who are itching to get a chance to prove themselves, who are hungry to do work, and who will not hesitate to pick up any slack I let out. This is a powerful motivation for me to maintain and improve myself as a translator, and it is a strong motivation for newcomers insofar as they know, or should know, that they will be given a chance soon enough.

Meritocracies do have their limitations. I know some translators who complain that vendors do not treat them with respect, accuse them of low-quality work, and even try to rip them off. I know people who work at translation agencies who complain about the difficulties of sifting the wheat from the chaff among the 50 or more translation résumés or brochures they receive each week, who decry the lack of good translators who submit quality work on time, and who long for some consistency in quality and professionalism from translators.

I agree with all of the above, though I do not think accreditation would do anything to change any of it. An accredited translator may still be unprofessional, an agency may choose to ignore one's accreditation. What's more, agencies will have to verify the claims of accreditation for each and every translator they choose to work with, something that will likely take time and cost money. And translators may still feel as though they have little recourse for dealing with bad vendors.

As for the matter of unprofessional translators or bad translation vendors, I have made suggestions in previous articles as to how to handle specific situations, such as late payment or no payment, when they arise. Translators can cooperate to keep track of information about vendors among themselves so as to avoid bad vendors; agencies can keep good translators by paying them promptly and fairly. Translators can increase repeat customers by doing quality work and submitting it on time; agencies can ask for sample translations and start new translators off with small assignments as a way to sift the wheat from the chaff, as well as perhaps placing more value on education or formal training for translators.

In essence, the current system, such as it is, in the translation industry may not seem to work particularly well, but it is preferable to accreditation, and can be improved in a few simple ways. All of the complaints in the preceding paragraphs really have to do with a lack of understanding and communication regarding expectations. This stems primarily from two areas: One, a lack of understanding regarding what translation is and how it is to be performed; and two: a lack of well-defined standards for the translation industry in general, and the various kinds of translators and translation vendors in particular.

Standards

The translation industry is in desperate need of some fundamental definitions. The ATA has attempted on a few occasions to issue standards, but for reasons too complex to explore here, has failed to convince translators, translation vendors, or the industry overall to accept their efforts. Other organizations have made various attempts, all without success.

This is changing, however. As mentioned in the opening of this article, the ASTM and LISA are in the process of developing standards for the translation profession. I applaud this effort and hope that the results will finally bring to this industry what it has needed for so long: some simple, clear-cut, straight-forward definitions of what a translator is, what a translator does, how a translator should translate, what constitutes a good translation, what a translation agency is and does, and how translation agencies and translators, or translation employers and translators, should interact with each other, to name a few possibilities.

Translators who want to avoid accreditation, who would prefer that our industry police itself and improve itself from within, should make every effort to become more professional in the translation work, to contribute, whether in writing or otherwise, to increasing understanding about the translation profession and the translation process, and to do quality work in a timely fashion at all times. Translation vendors similarly have responsibilities to fulfill, including making the effort to work with those translators whose abilities justify their respect, offering fair market rates for the work being done, and being willing to bear an equal share of the responsibility for quality translation work, whether through providing technology, training, or information to in-house and freelance translators, or working with organizations like the ASTM and schools like Georgetown University, the Monterey Institute of International Studies, and Kent State University, to develop better training programs for translation professionals.

Many translators already do the above things and then some. Many translation vendors already do the above and then some. Obviously there are enough who do not that the industry as a whole is considering accreditation as a way to resolve existing problems and avoid new ones. I hope I have shown how accreditation will create far more problems than it will solve, and how a reasonable effort from all members of the industry along with standards and definitions from a respectable, independent organization will allow the translation industry to move forward successfully into its future.