Posted by: patenttranslator | November 20, 2016

Globalization Blowback in the Translation Industry?

“Globalization”, “Globalizierung”, “globalizace”, “mondalization” – I hear or read this word in various languages whenever I watch news or read a newspaper online.

Globalization has had a major impact on our world: it made some people much richer, mostly those who were already very rich, and some people poorer, mostly those who were not very rich to begin with. I am told that it also lifted some people out of grinding poverty, mostly people in what is still called the third world. I have no way to verify the statement in the previous sentence, but it probably is true to some extent.

Globalization not only made billionaires out of people who only had a few hundred million before it started reshaping the world in the mid 1990s. This after Bill Clinton, who felt our pain so convincingly in 1992, signed NAFTA (North American Free Trade Association) a few years later.

Anger at the results of globalization also resulted a few decades later in Brexit and in the election of another president who, although he lost the popular vote by more than two million votes (while 45 percent of eligible voters stayed home!), managed to paint most of the states in the United States with an angry red color in places where the ignored and underpaid (former?) middle class is seething with rage at a neoliberal establishment that brought it nothing but humiliation and suffering.

Globalization also proved to be a perfect medium for various types of criminal activities. Most of the junk mail in my mailbox, and yours probably too, which has links asking me to click on them to verify my credit card information or sign up for a new warranty on my refrigerator or car, is sent from somewhere in Europe, Asia, or Africa, because it is very difficult for police located in one country to go after criminals who are located on another continent.

For several months I have been receiving threatening recorded calls from the “Internal Revenue Service” asking me to contact them immediately because the “IRS” filed a suit against me. I had to get a new telephone number that I use only for direct communication with my clients, and I don’t answer calls to my business number that is on the internet unless the callers seem legit, which is almost never the case. Many people do this these days.

I read in the Washington Post a few days ago how crooks pretending to be calling from the IRS created a sophisticated organization in India and the United States that swindled many people out of many millions of dollars, I forgot whether it was tens or hundreds of millions.

During the last two decades globalization has also reshaped what is now called the “translation industry”. Because unlike for example hair cutting, which is a highly regulated professional activity requiring a specialized license that can be obtained only after hundreds of hours of professional training, translation is an unregulated, semi-professional activity in most countries – just about anybody can do it.

And just about anybody does do it.

When a client looking for specialized translation service clicks on an advertisement on the internet, there is no telling who this client will be dealing with and who will in fact be translating the document that needs translating, whether it is a birth certificate, a college diploma or transcript, a software manual, a patent application, or an article from a technical journal.

Years ago when the internet was still kind of new, there used to be a popular joke:”When you’re on the internet, nobody knows you’re a dog.”

When clients click on links to translation agencies on the internet, they have absolutely no idea what kind of dog they might be dealing with and who will actually be doing the translation. The client may be dealing with an “entrepreneur” working out of his kitchen on a laptop somewhere in India who put together a website created from radiant Photoshop images. That website may have no physical address, or if there is an address, it may be just a post office box in New York or London, while the office may be located in Thailand or Bangladesh.

Or the link click could lead to a large agency with offices in half a dozen major cities, in which kids fresh out of college willing to work for low wages are being trained in how to parcel out translation jobs to people calling themselves “freelance translators”, who again may live anywhere on this troubled planet.

Whether the people calling themselves “freelance translators” are highly educated and highly experienced specialized professional translators or rank amateurs who have no business translating anything (let alone important and complicated documents, although they made it their business), is not really all that important to the guy working on a laptop from his kitchen, or to the monolingual kid working for a large translation agency whose arms are stretching like an octopus’s tentacles across several continents.

What is much more important in our globalized world to most translation agencies is how much the “freelance translators” are charging for their work. And that is why the translations may be done by the same person, whether a customer is dealing with a huge agency or with a guy working on a laptop out of his kitchen – namely the cheapest warm body available at the moment.

From the viewpoint of the client, the results of this “bottom line” approach to translation are often pretty ugly.

I have been in this “translation business” for so long now (more than three decades) that I seem to be able to recognize certain trends that may not be obvious to most people.

For example, I remember that after 2003, when the phenomenon of globalization and its effect on translation as a professional activity was still relatively new, many new customers were finding my website at PatentTranslators.com and asking for price quotes for patent translations. This happy period (for me) lasted for a number of blessed years during which I kept adding new customers, many of which later became repeat customers, mostly patent law firms.

My busiest and most profitable years were 2006 through to 2008, and new business kept rolling in from new clients who found my modest website until the year 2010, when the demand suddenly dropped by about 25 percent.

I think that the drop may have occurred not only as a result of a worldwide economic recession, but also as a result of maturing of “the globalization of the translation industry”, which in concrete terms meant that the market ultimately became saturated with offers of translations at rock-bottom prices, prices so low in fact that they would not be survivable by an educated and experienced translator living in a high-cost Western country.

But after a period of dropping demand, I seem to have experienced an uptick in the demand for my services again last year, with a more dramatic increase this year. I am only guessing, but I think that the increased demand is the result of what might be called “translation industry fatigue”. Customers are tired of paying for substandard translations necessarily resulting from the “bottom line” approach of the “translation industry”, translations that are sometime so bad as to be almost unusable.

So, as they are trying new alternatives, new customers are again discovering my website. A desire for change, change at almost any cost, whose distinct component is what might be called a globalization blowback, or disgust with the results of globalization, is clearly the leitmotif of the second part of the present decade.

This globalization blowback has resulted so far in Europe in Brexit, and in this country, it will soon install in the White House a candidate who, although he lost the popular vote by many votes, still managed to somehow beat his opponent, the second time that this happened so far in the last 16 years.

If I am right about this, the globalization blowback in what for lack of better term is called the translation industry also creates new opportunities for independent translators who are disgusted with working for the “translation industry” and for specialized translation agencies who, unlike a large part of the “translation industry”, possibly most of it, are in fact able to provide the kinds of services that clients actually want and need.

Will they be able to take advantage of this development and make a connection with clients who need their services? Or will globalization continue to grind on, producing layers and layers of inferior but relatively cheap translations on top of even more affordable abominations such as “post-edited machine translations”?

Only time will tell. But as I learned in my Latin classes in high school, (before globalization, there was a time, decades ago, when kids in high school were taught a foreign language called Latin), “Dum spiro, spero“, which means “While I breathe, I hope.”


Responses

  1. I wish I could agree re “translation industry fatigue”. In the UK, a previous Minister of Justice decided to destroy the livelihoods of professional court interpreters by handing over the entire provision of court interpreters to an obscure agency, run by a monolingual “businessman” and based in the wilds of the Yorkshire Moors. The agency was offered a lump sum in exchange for which it had to provide an unlimited number of interpreters to the British courts for three years. Needless to say this was an unmitigated disaster. The UK government trumpeted delightedly that this would save 18 million sterling, in fact it probably cost the taxpayer a fortune. That is because the agency, naturally enough, decided to offer interpreters a fee that amounted to less than the minimum legal wage, so qualified interpreters left the profession in droves. What was left was a rag-tag of barely trained or untrained semi-literates. Those that bothered to turn up in court when summoned were inadequate. So the agency (which had done a deal with a much bigger company that eventually bore the brunt of the disaster) lost the contract and it has now been handed over, lock,stock and barrel – to another, equally awful agency!

    Liked by 1 person

  2. If I were a millionaire who found myself opening a translation agency (as has happened a couple of times in the UK) I would throw money at internet advertising and all other channels until they were saturated with my name. Then I would employ sales managers at rock-bottom prices (and in some cases not paying them at all, one agency calls them “interns” and doesn’t pay them) to reel in clients. I would use as many translators in the third world as I could and pay peanuts. This is system has worked brilliantly for a few agencies world-wide who have hoovered up clients from naive individuals to major law firms, corporations, etc. Quality is, of course, no consideration and if the client complains, the agency simply pays the translator nothing at all.

    Like

  3. The system you are describing, Josephine, is precisely what has been happening for more than a decade in the “translation industry”. The question I am asking is whether the globalized approach of “translation industry” to producing translated words at low cost in large quantities also creates an opportunities for translators and translation agencies who want to offer a different approach to translation.

    I believe such an opportunity exists because I myself have been taking advantage of it for quite some time. But I don’t know whether it will be taken advantage of also by many other people, although I hope so.

    Like

  4. The “no one knows you’re a dog” meme reportedly originated from a guy who got a credit card for his dog in the wild web 1.0 days. When he tried to tell the company what happened, they gave him a canned response and another credit card.

    Like

  5. I thought it was a sexist put-down for unattractive women.

    Like

  6. […] https://www.youtube.com/watch?v=lY2yjAdbvdQ "Globalization", "Globalizierung", "globalizace", "mondalization" – I hear or read this word in various languages whenever I watch news or read a newspaper online. Globalization has had a major impact on our world: it made some people much richer, mostly those who were already very rich, and some people poorer, mostly those who were not…  […]

    Like

  7. For a relevant quote you might also use Melville to describe the struggle with bottomfeeders: “To the last, I grapple with thee; From Hell’s heart, I stab at thee; For hate’s sake, I spit my last breath at thee.”

    Liked by 1 person

  8. This is OT, and you might already know:
    Neural Conquers Patent Translation in Major WIPO Roll-out
    https://slator.com/technology/neural-conquers-patent-translation-in-major-wipo-roll-out/

    Like

  9. What does neural mean here? I don’t see a concrete explanation, only verbiage.

    Like

    • The big new thing is to use neural networks for machine translations Instead of purely rule-based, purely statistical, or a hybrid approach; this is called Neural Machine Translation: NMT. Neural networks are fun because they “learn” from the input. The advantage is that the neural network (NN) gobbles up large amounts of data and “learns” quickly. However, what they don’t tell you is 1) the NN has to be configured properly; 2) the data NN learns from has to be curated to an extent, or you have to have a post-process for weeding out noise. Fascinating stuff for me as a computer scientist, but it ain’t no golden ticket.

      See: https://en.wikipedia.org/wiki/Neural_machine_translation
      (Nvidia, surprising) https://devblogs.nvidia.com/parallelforall/introduction-neural-machine-translation-with-gpus/
      (Google, of course) https://research.googleblog.com/2016/09/a-neural-network-for-machine.html
      (Standford, NMT for J2E) https://cs224d.stanford.edu/reports/GreensteinEric.pdf

      I do think that discerning clients (those who care beyond the bottom line) are seeking out skilled professionals that they can reach with one email or one phone call. But there is another reality — clients whose demands (by volume, time constraints) cannot be met will investigate MT and will incorporate it in their process. I don’t think that’s necessarily a bad thing though.

      Like

      • Yeah, right, Shakespeare used neural machine translation, did he? Like most “big new things” NMT is a load of hogwash.

        Like

      • This news story gained traction because a small, modestly funded team at WIPO beat Google Translate by a small margin at the patent translation game. At least that’s what the techies and their metrics say.

        This factoid generated some hyperbolic coverage of the tech-is-oh-so-cool variety. I’ve already seen it in Slator and in a general newspaper (don’t remember which). Slator’s first story was more sober, but this latest headline is high-test clickbait.

        Liked by 1 person

      • nedwards, I think you’ll find this side development interesting: https://research.googleblog.com/2016/11/zero-shot-translation-with-googles.html

        Like

  10. Reblogged this on Translator Power.

    Like

  11. Well, yes, a robot can translate a patent and also a weather forecast. But not the usual stuff we translators have to deal with.

    Like

    • A robot CANNOT translate a patent, Josephine, don’t believe that.

      Like

  12. “Like most “big new things” NMT is a load of hogwash.”

    I think so.

    I was recently translating a patent while looking at a Google Machine translation, probably done with the latest NMT, and the machine translation was really, really good.

    Except it was factually wrong because the original text said something else. There are often several different versions of machine translations of patents: version (A), which is an unexamined application, version B, which is an examine patent application, these version are often changed and amended, especially in claims, and an actual patent may be issued much later, a decade later.

    Unless somebody tells the big machine brain which version is relevant, the big machine brain has no idea, which is why I often see GoogleTranslation versions of patents that look very good, but are factually wrong.

    Like

    • Surely, that is the fault of the person selecting which version of the patent to translate, not the fault of the machine? Even a human translator could make such a mistake.

      Like

      • It was not a person who selected the document, it was an algorithm that grabbed an existing document, originally written by a human, from the internet, and substituted it for human translation of sentences in a document in a different language instead of trying to really translate the sentences, i.e. simulate the thinking of a translator, because machines are not capable of such a simulation of human thinking.

        But in the real world, the relevance of documents such as patent applications or legal briefs is constantly changing and it is not possible to write an algorithm enabling a critical evaluation of which document is relevant and which is no longer relevant in each and every case. This, again, is a determination that only a human brain can make.

        It just goes to show that the search for “a machine translation that is just as good as human translation” is an exercise in futility also based on the approach that Google Translate is using.

        And all this talk about “neural” this and “neural” that sounds to me just like more marketing propaganda, especially in the absence of a concrete explanation of what this “neural” really means.

        Like

      • Something similar happened to me earlier this week. I entered some claims in Google Translate just to see how well the new neural machine translation method worked. The translations of the claims were perfect. Funny thing, though–the translations included the claim numbers. I hadn’t entered the bracketed claim numbers from the source document so GT had clearly retrieved recycled text from an earlier human translation of the same patent application.

        Like

  13. It’s called neural because it uses artificial neural networks, an AI algorithm. After rules-based and statistical MT, they’re having another go using a third algorithm.

    The latest wave in AI is deep learning, which uses networks like those described in https://en.wikipedia.org/wiki/Artificial_neural_network, but massive ones with thousands of artificial neurons. Some of Google’s servers actually have dedicated chips that simulate deep learning networks.

    Google announced that its tests were successful and that NMT has been deployed in Google Translate, but I noticed no difference in the quality of its output, at least from/into Portuguese.

    I think the ‘neural’ in neural machine translation isn’t a marketing gimmick yet, but it will become one quite soon.

    Like

  14. Thank you, that was very helpful.

    This sentence from your linked Wikipedia article probably summarizes the state of the art of Google Translate with neural networks best:

    “Modern neural network projects typically work with a few thousand to a few million neural units and millions of connections, which is still several orders of magnitude less complex than the human brain and closer to the computing power of a worm.”

    I had to laugh when I read it.

    Like

  15. “”Modern neural network projects typically work with a few thousand to a few million neural units and millions of connections, which is still several orders of magnitude less complex than the human brain and closer to the computing power of a worm.”

    I had to laugh when I read it.”

    Notice there is no citation for anything in this wikipedia introductory section. A $1500 laptop is expected to have in terms of petaflops the power to simulate a human brain by 2022, just 8 years away, but it doesn’t have to be that powerful to continue to lower wages in the translation world.

    Real wages for J>E translation is down over 30% from 2000.

    Liked by 1 person

  16. Ditto “real wages” in all language combinations.

    Like

    • I have lowered my rates to agencies by about 15%, but I have not lowered my rates to direct customers, who are my main clients. But I stopped increasing the rates about 10 years ago, and I have less work now from direct clients, and much less work from agencies, then 10 years ago.

      Fortunately, I need less money now because my children live on their own and I am receiving also a retirement income.

      So the Universe always tends to somehow even things out, in the long run.

      I’ve noticed that.

      Like

  17. “Funny thing, though–the translations included the claim numbers. I hadn’t entered the bracketed claim numbers from the source document so GT had clearly retrieved recycled text from an earlier human translation of the same patent application”.

    I see this all the time because I usually print out a machine translation when I translate patents, and it’s usually Google Translate. It is so good sometime because it really is human translation.

    But the flip side is that it can be completely wrong because GT simply substitute a similar text for a translation and ignores the difference. And if you don’t know the original language, you have no idea what the differences are.

    Another problem with GT is that it does poor job on documents that have no prior versions, such as my posts on this blog.

    My brother in Europe, who does not speak English, told me a number of time that he can’t understand what it is that I am saying in my posts when he runs them through GT.

    Like

  18. The cases where MT delivers human translations that come from a slightly different source document probably occur because Google has been using WIPO’s databases of human translations to train their algorithms. WIPO and Google signed an agreement to that effect a few years ago–around 2010 if I remember correctly.

    Presumably the human translation is better than whatever the machine could churn out on its own, even if it’s made from a slightly different source document,

    Like


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Categories

%d bloggers like this: