The shifting sands of quality

What does “quality” in translation mean as we head into 2020?

A few days ago, I read a post by Paul Dargan writing in the Japan Times about the importance of register in translation. Dargan writes:

“[…] when it comes to being an accomplished translator, your linguistic ability should be highly refined and never called into doubt, but that skill in and of itself is rarely what employers are looking for. What sets good translators apart from the rest is their ability as writers, and that ability is dependent on two qualities — how broad your foundational knowledge is, and how easily you can adapt to various kinds of writing. […] In just two years working for a translation company, I’ve handled over 700 assignments [every one of which] asked something unique of me, whether it was some level of industry-specific background knowledge or, more importantly, a manner of writing that was appropriate given the field and intended audience.”

I agree wholeheartedly with him, although I suspect certain clients are willing to sacrifice a little writing quality in a translator with truly expert knowledge in difficult niche fields. But I agree with him about adaptability. Given that our finished translations do not simply drift off into the ether (although some would beg to differ!) but are read by some kind of audience, being able to “write for the right people” is essential, and something that people who are not familiar with the translation process are often not aware of.
I would go one step further, however, and say that as translators we must be increasingly prepared to a) adapt to different kinds of quality, b) consider doing so, c) or at least acknowledge that different levels of quality exist. Changes in the translation industry with the advent of machine translation (MT) means that the definition of “quality” in translation is a lot more fluid than it once was. A translation needs to be fit for purpose, which does not always require it to be a work of literary genius.

The end of quality?
There was an interesting thread on a Proz forum recently. Called “The end of quality?”, it ignited some very interesting debate. And it all started off with the following email that the thread poster had received (re-worded slightly for brevity): “One of our clients wants very basic translations for internal use and is asking about some hybrid solution that would be better than simple machine translation. Since we currently do not offer such a service, we were wondering if you could give us your rates for a) post-editing of a machine translation, and b) lower-quality “basic” unpolished translation for internal use only (not using machine translation). What is your experience with lower-quality “basic” translations?”

The thread poster put forth the argument that he only provided “quality” translations. He knew no other way to translate, he said, nor did he have any intention of translating any other way. Many on the thread expressed similar thoughts; I too replied to the thread, comparing quality to pregnancy – something along the lines of “either a text is good quality or it’s not.” Other replies were along the same lines: “What the client wants isn’t really poor quality or “basic level” translation. What they really want is CHEAP translation.” Another saw requests for machine translation post-editing (MTPE) “merely as a pretext for further downshifting and forced compromising.”

Others, however, offered a different viewpoint. “Varying the quality level is not a new concept. We don’t all go shopping at Harrods, fly first class or buy the best and most expensive wines.” Another colleague wrote: “It seems that many colleagues believe that translation is a thing in itself, and the end goal of all clients is another Shakespeare’s play. […] Readable, mediocre translation can still be accurate in terms of the main points of the content and suffice in more cases than we would like to permit or admit.”

Although the company writing the email which sparked the thread specified that it did not want machine translation, the following MT post-editing guidelines (published on the TAUS website) give an interesting insight into what a different concept of quality might mean, and how turning out top-notch English is not always necessary. This is not news to professional translators, of course, who are all-too aware that translating high-end literature requires a very different register and approach to a user manual, a contract or a medical report. But producing texts that “sound like they were generated by a computer” is like nails on a chalkboard to those whose translations need to seem like they were written by a native speaker. Or who presume they do. The guidelines (again edited for brevity) read:

“Good enough” is defined as comprehensible (i.e. you can understand the main content of the message), accurate (i.e. it communicates the same meaning as the source text), but as not being stylistically compelling. The text may sound like it was generated by a computer, syntax might be somewhat unusual, grammar may not be perfect but the message is accurate.

  • Aim for semantically correct translation.
  • Edit any offensive, inappropriate or culturally unacceptable content.
  • Basic rules regarding spelling apply.
  • No need to implement corrections that are of a stylistic nature only.
  • No need to restructure sentences solely to improve the natural flow of the text.

I just finished reading the fantastic book Diversification in the Language Industry: success beyond translation by Nicole Y. Adams. Highly recommended even if you have no particular intention of diversifying, and written in 2012, the various predictions for 2018 are surprisingly accurate. The language professionals interviewed for the book show various levels of enthusiasm for diversification, but one very strong thread running through the book is a need to evolve and embrace change and technology in order to survive. I must admit that this book changed my views on the post-editing process, and the place it can have in the translation landscape.
Anne-Marie Colliander Lind, CEO of management consulting company, says:
“Translation resources are scarce! There aren’t enough linguists in the world to fulfil all the requirements for translation, which in turn drives the need for technology and automation. […] As weird as it might seem, technologies like machine translation drive the need for human translation. […] If volumes increase for translation companies, then volumes for individual translators will increase too.”
Indeed the book notes that the work performed by human translators is less than 1% of all the volume out there needing to be translated. Let that sink in for a moment. With the explosion of information swirling around us on the Internet, the greater drive to consume and to sell than ever before, as well as many other factors, MT actually plays an important and necessary role in making texts available to a wider audience. Professional translation is not cheap either, and provides a cost-effective solution to those clients who either cannot afford professional translation rates, or are prepared to sacrifice quality for a lower price.
One complaint levelled by translators is that MT output can be extremely poor, something that deters many from taking on this type of work. Lori Thicke, founder of machine translation firm LexWorks, makes some suggestions as to how the situation might be improved: “To expand the pool of available post-editors, it is necessary to respect their time by choosing the right engine, training it properly and updating it often; respect their talents by involving them in the process ad set their (and our) expectations correctly.”
Tineke Van Beukering, Senior Project Manager for 2M Language Services in Australia, says “… when translations are needed only to get an idea of the content of large volumes of texts… a lesser quality can be quite acceptable.” Of the MTPE process, she says: “Post-editing requires a different mindset from translating… use as much of the machine translation as possible.. [and].. only change what is incorrect or inappropriate. As a consequence, the resulting translation can be quite different from how a translator would have translated the text from scratch. That’s the part I found hard to adjust to in the beginning – letting go of a translator’s point of view and switching to a post-editor’s point of view.”

The intent of this post is not to say that we as translators must all begin doing MTPE. Nor do I suggest we should slacken off and dumb down our translations simply because the line of “quality” in translation has moved in the sand. But MT is here to stay, whether we like it or not, and it’s worthwhile reflecting on how we feel about it, the impact it will have on our work, and how big or small a part we want to play in this particular evolution of our industry.