Thursday, June 19, 2014

Why So Many Translators Hate Translation Technology

In general, technology should serve to make our lives better. Software can automate manual steps, reducing the human workload. Translation, like most language-related tasks, is complex. Machines have not yet gotten to the point where they can use language the way people can. Translation is, at a minimum, two times more complex than just writing in a language, and it's several layers deeper too.

So, with such complexity, it would seem that every professional translator on earth would be eager to use technology to streamline their work, speed up the translation process, and make their lives better. This is not always the case.

Here's why:

Google has changed society's definition of "translation."
In many ways, Google Translate has been both the best and the worst thing to happen to professional translators in our lifetime. On the one hand, Google's work in this area has put translation more squarely on the radar of society in general. Suddenly, people know that translation is available and possible. This fuels market growth and demand, because people quickly learn that free, online machine translation quality is insufficient for most purposes, especially for business settings.

On the other hand, while Google Translate has raised awareness of the need for translation at large, it has also made high-quality translation seem easier and more accessible than it actually is. Even though Google Translate is not actually free beyond a limited usage volume, many people now mistakenly think translation is "free and easy." This makes it that much more difficult for professional translators to help people understand the value of their work.

Translation tool development has largely stagnated. It's incredibly frustrating for translators that so little development has taken place when it comes to professional tools. I find it incredible that the interface for most translation tools today looks nearly identical to the interface that I used as a translator in 1996. Two decades on, there are hardly any tools available that give translators a "what you see is what you get," in-context translation view for translating websites and other such translation projects of a web and digital nature.

Perhaps even more frustrating is the fact that translators have to work in so many different systems to actually obtain the files to translate, to do the translation, and to deliver them back. Even though translation process automation software exists, adoption in the market is not widespread by any stretch. This means that the chances of, say, attaching the wrong version of a file, or accidentally inserting errors at the time of saving, or other such common human errors, are still quite high.

Translators prioritize quality, whether the customer does or not. Ask a translator what kind of quality is acceptable, and a professional will tell you, "Only the best." Would a musician be happy with a sub-par performance? No, and neither are translators. Professionals hold the bar very high for what they consider a good translation. But ask a buyer what kind of translator is "best," and the answer will often be, "One that delivers the translation to me on time." Buyers are sometimes willing to sacrifice quality for on-time delivery.

Some buyers, because they lack knowledge in translation and don't usually speak the language they are translating into, put cost first and foremost, depending on the type of information they are translating and how it will be used. Where professional translation is concerned, you often get what you pay for. The same is true of other professional writing services. You won't get the quality of a bestselling author if you are only willing to pay the rates of an unpublished, first-time writer. Buyers are often pennywise and pound-foolish in this regard, but in other cases, they are simply making a decision driven by very real business needs.

What does this have to do with technology? Many buyers want a quick fix, low-cost solution for translation thinking that technology can offer this, and translators see the extreme danger of that. It's wrong to think that you can get the same level of quality while cutting corners. The disconnect between buyers and translators is often that buyers have priorities that are not in line with translators' perspectives, or their objectives. For example, a client might have some content that will not be published anywhere, that is for internal use only, but that must be delivered on time, no matter what the level of quality. It's nearly impossible for a professional translator to translate at a lower level of quality, just like a professional musician would struggle to perform at a lower level of quality that what they have trained for their entire life.

Translators are usually isolated from the rest of the translation process.
Buyers of translation, especially business buyers, purchase the majority of translations from agencies. What agencies offer that individual translators do not is very simple but very important for many business buyers - project management. Agencies are capable of sourcing translators for many languages, while translators typically specialize in just one or two combinations, due to the level of mastery needed in even just one combination. Also, agencies can take on larger volumes than individuals can.

Yet, when agencies are involved, translators often feel cheated. They are typically underpaid, when one considers their level of skill and education. Most translators have a graduate degree, but earn below average for what someone with a graduate degree would earn in other industries. They see agencies charging nearly double what they are paid, and they wonder why the agency is making so much money off their backs. This seems unfair, given that without them, the translation simply would not be possible.

In reality, some agencies do hike up the price beyond what may be considered reasonable for the value they add, with profit margins in excess of 200%. That's extremely rare. The industry norm for a translation agency profit margin hovers around 20 to 30%. In other words, if you add up the costs of the translator's rates, plus the editor's and proofreader's rates, and if you build in the project management costs, an agency makes about 25 cents per dollar spent. While this might seem like a hefty profit from a translator's perspective, from a business perspective, this is not necessarily that attractive. While it's better than most services businesses, it's a lackluster profit margin when compared to other industries. Still, for most of the tens of thousands of agencies, the vast majority of which are very small businesses, often family-run, this profit margin is respectable and good enough to keep their doors open.

The translation market is changing faster than the translator's role and tools. Communication methods and technologies have evolved, but the role of the translator has not changed very much. In today's real-time world, where people are "live-tweeting" and instant communication is more common, you would think that there would be more "live-translating" and more tools to facilitate real-time translation. In actual fact, many translators would reject the idea of doing "real-time translation," because they are accustomed to having more time to carefully research terminology and review their own work, just like any good writer would do.

To provide just one of many examples of how market demands are shaping the translator's role faster than translators would like, some companies are using machine translation - not the free online kind, but a more robust type of machine translation that provides better quality within a confined domain. They often want translators to "clean up" the machine-translated output. This has created a new role that did not exist previously, called a "post-editor." The human "post-editors" serve as the clean-up crew, and are expected to fix all the mistakes made by the computer-generated translation tool.

Here's the problem. If you're a professional translator, why would you want to clean up a big mess that you didn't create? It's like asking a professional musician to take a recording that was done by a synthesizer, or by a less-skilled musician, and go back and fill in all the sour-sounding notes. It's a hacky way to produce a translation from the translator's point of view, it does not generally produce great quality, and it takes all the joy out of translation. Many translators know that they can produce a better translation from the start, so they will quite naturally refuse to do what they may rightly view as linguistic janitorial work.

Translators have often been sidelined by translation technology companies.
Many of the companies that created translation technology in decades past failed to take an interest in the very real needs and concerns of their largest group of users, translators, even if they derived the majority of their technology revenue from them. Very few technology companies have hired people with deep knowledge of the translator's work to help guide their development.

Perhaps due in great part to this disconnect, translation technology companies have often failed to treat translators well, to recognize that they are professionals who deserve to earn a fair living wage, to stand up for their perspectives, and to incorporate them into the product development process. Translators distrust many translation technology companies because they believe that these companies do not have their best interests at heart. Often, they are right. The lack of communication between these two groups only exacerbates the problem.

"Hate" is a strong word, and of course there are many translators out there who don't exactly hate translation technology - they can live with it, but they don't love it by any means. Other translators are ambivalent and roll with the changes as they come. But a certain percentage of translators truly do despise and distrust translation technology. They feel that they have been burned, and thus are skeptical that any technology company, let alone the tools these companies produce, will ever make their lives any better. These concerns are rooted in their personal experiences -- these are important and should not be discounted.

Because I am a former professional freelance translator who is still an active literary translator, I appreciate and respect the viewpoints of my colleagues who have completely lost faith in translation technology and its ability to help them. But I don't fully agree with this negative outlook -- my view of the possibilities and the progress is more optimistic. Like them, I don't believe that the greatest advances in the professional translator's world will come from machine translation, at least, not in the near term. Simply showing a translator how something can be translated is not the be-all, end-all.

However, do I believe that technology can support translators to make quality better from the start? Absolutely. Quality improvement is the next phase of translation technology, the next problem to be solved for translators, the next challenge to be undertaken by cutting-edge technology companies, and something that I believe all translators can ultimately get behind.

No comments:

Post a Comment