Abstract

As a result of recent quality assurance standards in the translation industry, many providers of translation services require the systematic revision or editing of translations by a second translator and see this as necessarily beneficial to quality. Yet revision by another person can only assure quality if this person is truly competent and the translation/revision process is properly executed. Furthermore, the improvements to quality that revision may bring are not always worth the extra time, effort and cost. More importantly, when the revision process is poorly executed it can reduce and even destroy quality.

How to achieve perfect quality!

any things have been written about the merits and mishaps of having translations checked against the source text by a second translator, a practice often simply known as revision or editing. Yet few people seem to doubt that this is necessarily good for quality or that it should be systematically employed regardless of the type of document translated or the translator.

Economic forces will naturally encourage the most-experienced and competent translators to work with those translation companies that are able to appreciate their skills, trust them to work alone, and pay them more for this.

In the minds of many translation companies there is certainly no question that such revision or editing is necessary, judging from the claims they make on their websites: “All translations are revised by a second translator to ensure perfect quality..."; “To ensure quality, we require a second pair of eyes on each translation”; “Thanks to our four-eye control principle, we can guarantee premium quality and reliability”; and so on. These companies have apparently found a sure-fire solution for achieving high, if not perfect, quality: just have a second translator check the work of the first!

Now this is an interesting development. To this author's knowledge translation has become the only profession that systematically requires the checking of all work by a second practitioner. This gives pause for thought. Architects can design buildings, doctors can prescribe treatments and accountants can prepare accounts without having a colleague verify their work, but translators must have someone else make sure they have done a good job. These professionals can be trusted but translators can't. Yet if a first translator can’t be trusted to assure the quality of a translation then why should a second translator/reviser?

This new quality-assurance technique makes no distinction between documents—however important, sensitive, complex or banal—and applies to all translators, regardless of their experience or proven competence. This “one size fits all” revision rule puts everything and everyone in the same sack. The work of a translator with many years of experience, expert knowledge of the source language and a good knowledge of the areas in which he or she translates must therefore be systematically verified by someone else, just like the work of a novice. Some translators might find this rather insulting, since it implies two things: firstly that they might not know the source language or the subject of the text very well; secondly, that they may be careless and miss something, and perhaps don’t even check their work.

Things weren’t always this way. It used to be standard practice in international organizations to have “senior” translators revise the work of their “junior” colleagues, while their own work was not revised. Some institutions no doubt still do things this way. This new quality-assurance requirement is attributable to recent quality-assurance standards in the translation industry and most notably EN 15038 in Europe and ASTM F 2575 in the United States. Translation-service providers (TSP) who claim to observe the requirements of these standards, or are compelled to, are supposed to have another person compare their translations with the source text sentence by sentence and term by term to ensure they are accurate and of good quality in general.  Revision is actually not mandatory under ASTM F 2575 since the “requester” and the TSP can agree to omit this step. The U.S. standard goes on to say however that the translation “may be of lesser quality than if the entire process had been followed.”

A little terminology

EN 15038 refers to this QA operation by a second translator as “revision” and this translator is a “reviser.” When the original translator cross-checks his or her translation against the source text EN 15038  calls this “checking.” Some people refer to what the second translator does as “other-revision” and call what the first translator does “self-revision.” What EN 15038 calls “revision,” ASTM F 2575 calls “editing.” Some people use such terms as reviewing, cross-reading, comparative revision, re-reading, proofing or proof-reading to mean exactly the same thing.

When such terms as revise, check, edit or review are used without any reference to a quality standard things can become very ambiguous since this verification of the “translation” may simply mean just reading through the translated document and not cross-checking it against the original; a task best referred to as “proof-reading.” The latter is a much simpler and less time-consuming method of quality control, which when performed by a native speaker of the target language can be very effective in detecting translations that are obviously bad and poorly written passages that can be easily improved. However, proof-reading cannot of course determine whether or not a translation is accurate and therefore of good quality.

With all of these various terms in circulation there are obviously tremendous opportunities for confusion, obfuscation and disingenuousness. In this article, the term “other-revision” will be used to refer to the checking of a translation against the source text by someone other than the translator. When a translator checks his or her own work in this way this will be called “self-revision.”

Purpose of this article

This article is not saying that other-revision is useless. If done properly it can certainly be a very effective way of training inexperienced translators and also of checking whether or not a given translator is sufficiently knowledgeable, skilled and conscientious. It does not even claim that other-revision is useless for quality assurance. For some particularly important documents it can be a wise step.

What this article is saying is that other-revision is only effective in “assuring” quality if certain principles are strictly observed, which unfortunately is rarely the case, especially in the private sector. Furthermore, when quality is improved the improvements often do not warrant the extra time, trouble and expense. What’s more, other-revision may even destroy quality. In other words, it is often unnecessary and can do more harm than good. As far as the vast majority of documents are concerned, if a translator has an expert’s knowledge of the source and target languages, knows the subject matter well, has solid experience in translation techniques and tools, has good analytical and writing skills, is conscientious and asks questions about anything in the text that is unclear, then having another person revise his or her work is a waste of time, energy and money.

When can other-revision assure quality?

What are the conditions that will enable other-revision to not simply improve quality but “assure” it? In other words, ensure that the translated document contains no significant translation error and is fit for the purpose for which it was intended. The reviser’s role is of course essential. He or she must have an expert’s knowledge of the source and target languages, know the subject well, have solid experience in translation techniques and tools, have good analytical and writing skills, be conscientious and ask questions about anything in the text that is not clear. In other words, a second translator/reviser must have the same knowledge, skills and conscientiousness that a translator needs to assure quality. If this is not the case then other-revision cannot systematically assure quality. Two inexperienced and/or mediocre translators do not equal one who is truly knowledgeable and skilled. Yet Europe's quality assurance standard for translation services, EN 15038, does not require revisers to have any more knowledge of the subject matter than the initial translator. In fact, it merely states that revisers should have translation experience in the relevant field. According to ASTM F 2575, an “editor” does not need to be any more knowledgeable or qualified than a translator and is simply “a bilingual member of the translation team....”

If the reviser is truly capable of assuring quality then the translator must of course not be allowed to reverse any of the reviser's changes.

Nonetheless, it should be noted that even when revisers are fully competent they do not begin revision with the same knowledge of the source text that the translator has, unless they first read carefully through the document, which this author suspects is rarely the case, at least in the private sector, given the time and cost constraints. Although self-revision and other-revision apparently involve doing the same thing, a translator begins self-revision from a much broader and informed perspective since he or she has already read each sentence of the text, given thought to it and has acquired an overall perception of the document. In short, a translator begins the revision process with a better understanding of the document than does the reviser. It is the translator's prior contact with the source text during the rough-draft translation that enables him or her to connect the dots, understand what was initially misinterpreted, put the pieces of the puzzle together, ensure consistency and eliminate redundancies. Self-revision is therefore an absolutely necessary step in the translation process and is much more than just checking that nothing was missed. And yet ASTM F 2575 astoundingly does not require translators to check their own work! Does the U.S. standard just assume they do? Or perhaps it assumes that revision by another translator makes self-revision unnecessary.

Revisers who simply start revising by comparing pairs of sentences one after the other will have a weaker understanding of the document than the translator, even though they may know the subject better. They should keep this in mind before making any “corrections.”

When can other-revision improve quality?

If the initial translator and his or her reviser are not thoroughly knowledgeable and skilled then neither can systematically assure the quality of the translation. Yet other-revision can still improve quality in this case and possibly to the point of acceptability, particularly when translator and reviser compensate for the gaps in the other person’s knowledge or skills. But errors and ambiguities are very likely to remain and the translated document may still be unfit for its purpose. Although other-revision certainly reduces the probability of an obviously bad translation, the final result can still be far from good enough.

The cost of other-revision must also be considered in relation to its benefits; in other words whether the improvements that a reviser may make are worth it. Any translation can always be improved one way or another—made a bit more clear, concise or readable. However, with the exception of translations that are published and for which style is very important, minor stylistic changes and other tweaks are generally not worth the additional time and expense.

It should also be kept in mind that when the translator and reviser have more or less equivalent levels of knowledge and skill other-revision will only improve translations consistently if the translator has the opportunity to validate the reviser's modifications. Otherwise, a reviser's more superficial understanding of the text, as discussed above, can easily bring inconsistencies and redundancies into the translation. If the reviser does not understand the text or its subject matter as well as the translator he or she may also make the translation more literal and “safer,” yet much harder to understand for the reader. And of course, the reviser may simply misinterpret the text and introduce errors. If the revised translation is not returned to the initial translator so that this person is able to see and either validate or refuse the changes made, then the translation may very well not be improved, to say the least. Surprisingly, neither EN 15038 nor ASTM F 2575 requires this.

So if a reviser is not an expert capable of systematically assuring quality, other-revision can only improve a translation consistently if the translator is able to review all changes, is willing to accept those that are justified and refuses those that are misguided. This additional step is absolutely necessary to ensure improvement. However, it makes the process even more cumbersome, slows translation delivery even more and further increases the cost of other-revision, with no guarantee that the improvements will be worth it.

When does other-revision reduce or destroy quality?

As just discussed, sloppy or incompetent revision that the translator does not have an opportunity to correct can easily reduce quality, possibly to an unacceptable level. But other-revision can be bad for quality in other and more subtle ways.

The perverse effects of other-revision

Assuming that other-revision will systematically  assure quality or even simply improve it is to neglect some very basic things about human nature. Various problems can arise, particularly in the private sector where translator and reviser are often competitors working for the same customer. For example, a reviser may be overly critical and make changes that are unnecessary and even harmful to the translation. Or a translator may not be willing to accept changes that are justified.

But there is a larger and less obvious threat to translation quality—the dilution of responsibility. Neither EN 15038 nor ASTM F 2575 explicitly states whether it is the translator or the reviser who is responsible for the quality of the translation, although it may be assumed that the reviser has responsibility if the translator cannot validate changes. If no one is clearly responsible for the translation then responsibility will be implicitly and insidiously shared.

For one thing, a reviser does not “own” a translation in the same way the translator does, or at least should. Even when a reviser clearly has the last word it is not really his or her translation. Considering how poorly paid revision generally is in the private sector, even the most conscientious revisers will not take the time to rewrite all of the various things they consider to be awkward or poorly expressed, particularly when such changes might often be seen as unnecessary. As a result, revisers do not have the same personal commitment toward the translation. It's not their baby.

As far as the translator is concerned, when a reviser is considered to be an expert and has final responsibility for any modifications, the translator may neglect to do the necessary research and assume that the reviser will detect and correct any deficiencies. This is especially likely when the translator is pressed for time, as is often the case. Some translators may even commit the unpardonable sin of skipping self-revision.

When the reviser is not a subject-matter expert but simply another translator, i.e. just a “second pair of eyes,” then both translator and reviser may count excessively on the other person, assuming that he or she knows what they are doing. A translator may guess about the meaning of a phrase or the translation of a term and expect the reviser to make the correction if necessary. Or perhaps the reviser isn't really sure about something but just accepts the translator's interpretation because it seems logical. This can easily happen, particularly when the clock is ticking.

When can other-revision be useful?

Provided that the reviser is an experienced and competent translator with good knowledge of the subject matter, other-revision may be useful for quality assurance and may also serve other purposes. It can, for example, enable a heavy workload to be shared between a “senior” translator/reviser and one or more competent “junior” colleagues. Tight customer deadlines can thus be met while enabling less-seasoned translators to gain valuable experience and feedback in a specific field of translation.

Other-revision by a qualified reviser is also the only way to determine whether a document has been properly translated and is the best way to test a translator's knowledge and skill.

Conclusion

Other-revision can assure quality only if the reviser is a first-rate professional who knows the subject matter of the document well, is conscientious and takes the time to understand the source text and not merely correct sentence by sentence. This leads to some chilling conclusions since this is often not the case, particularly in the private sector where the economics of translation are generally stacked against effective other-revision, as the low per-word rates that many translation companies pay for this service tend to discourage the most experienced, specialized, skilled and conscientious translators from revising their colleagues. To put it bluntly, skilled revisers with specialist expertise do not want to be paid peanuts for correcting other people’s work when they can generally earn significantly more translating.

Furthermore, if a translator does not have the opportunity to validate the reviser's changes, it is quite possible that other-revision will not only not improve the translation but actually reduce or destroy quality. Even when improvements are made their contribution to quality might be insignificant and not worth the extra time and cost.

Although quality standards have enshrined other-revision as an effective quality-assurance technique, there is a good chance that many providers of translation services will come to reconsider this practice. Some translation companies will no doubt adopt a more discriminating approach to other-revision, reserving it for highly critical documents or when deadlines require the use of less experienced translators. In any case, economic forces will naturally encourage the most-experienced and competent translators to work with those translation companies that are able to appreciate their skills, trust them to work alone and pay them more for this.