Translation quality assessment (TQA) for Chinese-English translation almost always results in unsatisfactory outcomes. The most common outcome is that the assessing translator will make some random changes, often just shifting things around and playing with synonyms. For those translators with more expertise, a counterproductive and confrontational vicious junkyard dog type response is also common. This latter outcome has gotten so bad that many companies have begun adding training policies on civility. However, the problem for clients is not the lack of civility, but rather that the actual translation quality is often worsened due to the reviewer’s bad attitude toward their work. In this article, I’ll explain the problem and pitfalls within the Chinese-English translation quality assessment process and share strategies on how to overcome these challenges.
Junkyard Dog Behavior
A typical problem that occurs when an assessing translator assesses the quality of a translation completed by someone who understands how to do translation, is something that I would term as “junkyard dog behavior.” If you’ve seen a junkyard dog in action before, they are usually very fierce creatures that attack anything that gets in their way. Such is often the case with TQA done in Chinese-English translation. When reviewing another translator’s work, reviewers go into full-on attack mode and describe almost everything they encounter in the translation as being somehow bad or wrong, possibly even shameful!
A surprising flaw in these reviewers’ behavior emerges whenever you interact with them. They tend to have a lot of great reasons about why the other translator’s work is bad and wrong, however they don’t really have any reasons to back up why their own translations are superior. At most, their TQA analysis largely boils down to “my preferred translation is the best, because that is the way it is.” An interesting experiment to try is to have a translator revise something they worked on two years ago, as if it were done by someone else. Astonishingly, many translators will also go full junkyard dog on their own work. What’s happening here is that translators’ opinions and beliefs shift over time, and they are simply attacking any disagreement with their current opinions and beliefs. Yet, their conclusions are largely random, in part because of their limited grasp of effective translation quality assessment.
Getting TQA Right
Most translators and clients working with Chinese-English translations are unfamiliar with the concept of translation quality assessment, yet understanding its significance is paramount. TQA is the process of measuring the accuracy, completeness, and consistency of a language translation against a set of predefined criteria, and it’s an essential part of ensuring that the translation is of the highest quality and accuracy. Moreover, TQA acts as a crucial checkpoint in the translation production process, as it helps identify any potential problems before they materialize into costly mistakes.
TQA itself is really a blend of different assessments of the linguistic, functional, and cultural requirements for the translation. The linguistic assessment focuses on the accuracy of the language used in the translation, and is typically done manually by comparing the source and target texts side by side to ensure the translated version is correct in terms of grammar, syntax, and meaning. The functional assessment focuses on the how accurately a translation aligns with its purpose, ensuring that the translation is suitable for its intended audience and purpose, and that it is complete and accurate in terms of content. Lastly, the cultural assessment is done to guarantee the accuracy of the translation in terms of cultural context and intended audience, ensuring that the translation is suitable for the intended audience and does not contain any cultural or linguistic inaccuracies.
Linguistic validation techniques will also sometimes be applied as part of a translation quality assessment, particularly in the medical regulatory field, to ensure documents accurately reflect the meaning of the original version.
The linguistic validation process goes through several stages. First, the original version of the document is reviewed to ensure it is conceptually accurate and complete. Then, the translated version is compared to the original version to ensure that the meaning and content of the original version have been accurately transferred to the translated version. Next, the translation is reviewed by an expert panel to ensure that it is linguistically and culturally appropriate. Finally, the translated version is tested with a representative sample of the target population to ensure that it is understandable and acceptable. While relatively expensive, it’s worth noting that linguistic validation should be used by translators in all fields, at least occasionally, to ensure that their translation approaches and methodologies are getting accurate results.
Special Considerations for Chinese-English Translation
The unique challenge faced in Chinese-English TQA is rooted not so much in the translations being critiqued, but rather the extremely poor alternatives presented. I’ve encountered plenty of translators doing assessments defending machine translations, word-for-word dictionary translations, and illogical nonsense that defies reality. As an assessing translator, it’s not enough to merely say that someone else’s translation is bad; you need to show that your work is correct too. If you are doing work that is fundamentally incorrect, showcasing that someone else is wrong becomes irrelevant. If either side of the disagreement’s reason their translation is good and reliable is simply “because I say so,” a business employing these two people to do the translations will have to decide who is right purely based on seniority. But, in this case, seniority is as good as flipping a coin because if the senior translator lacks substantive reasoning to justify why their translation is correct, then the odds of their work being correct are no better than the junior translator’s.
A second common problem in Chinese-English translation is that a translator will come up with a defective alternative, then begin entrenching and trying to come up with reasons why correcting the defect is wrong. The translator will start focusing on fighting off sources of potential criticism and deflecting and rejecting any possible competing ideas. This is not a good attitude either. As a translator, other translators’ ideas, even if you disagree, offer valuable insight into where your translation may potentially be flawed or at least why someone else may think it’s flawed. This gives you a chance to ensure that your work does not have any common defects in it.
An effective solution I have devised to both of these problems is evidence-based translation, which I describe in more detail in a blog post here, and also my related guide to analyzing Chinese legal terminology for translation. Paradoxically, the reason translators entrench so firmly is their lack of a strong foundation. These translators are not thinking critically about why their translation is correct. They are not using the basic tools of linguistics—such as evidence, corpuses, and lexicogrammar analysis—to use logical reasoning to arrive at a conclusion. Instead, they are arriving at snap judgments intuitively.
Automated TQA of Chinese-English translation causes special problems when glossaries are involved. Inevitably, an organization using glossaries for Chinese-English translation will come up with a list of one-to-one translations for each of the terms being used on the product. Doing so is almost always a terrible idea, and will force the translation quality to be lower. There are very few words in Chinese that have a simple, one-to-one correspondence with words in English; even if you look at bilingual learners’ dictionaries, not even intended for the specialized purpose of translation, for many words you can find a dozen or more possible English correlates, and vice versa. My previous article on lexicogrammar you can find here lays out the reasons this may be the case, but the very short answer is that meaning in language is not comprised of words alone, but also by grammar, genre, and context.
