This scholarship essay asked applicants to write on how technology is changing the translation industry. It might be helpful to know who was offering the scholarships, so I’ll add that in any future posts in this series. This one is offered by Architekst, a translation service that has been around for 20 years. You can access the scholarship information and learn more about Architekst here.
Not that long ago most of us traveled to non-English (or whatever your language of choice is) countries carrying a paperback translation dictionary. Slow, cumbersome, but effective enough to get you fed without too much embarrassment.
Old school translation saw some mind-boggling interpretations. Here are some Chinese translation examples from the movie industry: “Risky Business” was translated to “Just Send Him to University Unqualified,” “As Good as it Gets” to “Mr. Cat Poop,” and “Pretty Woman” to “I Will Marry a Prostitute to Save Money. (Catron, 2013)”
I was an AirBnB host for almost five years and translation technology (technology) made it possible for me to host non-English-speakers. At first it was more reliable if you were translating to other Latin-based languages, but you entered unknown territory outside of Spanish, French, and Italian. I remember inadvertently asking some Asian ladies for a sexual favor as the device used algorithms and social bias to wrongly decide what I was trying to say. Fortunately, they saw it as hilarious. In that five years the technology improved exponentially, evolving from the rusty knife to a surgeon’s scalpel of technological utility.
Those are funny examples, but at times precise, clear transfer of ideas is critical. Consider how unfortunate the above type mistakes could be in cases of legal documents, medical procedures, or business contracts.
So how has technology evolved since the days of the dog-eared translation dictionary, how does it work, and where might it go in the future?
According to Alexander Waibel, a computer science professor at Carnegie Mellon University, the technology initially used rule-based methods to decide what was trying to be said, upstaged by statistical algorithms, and finally neural machine translation. Which, according to Waibel, was such a huge improvement that it was adopted industry-wide within a year (Jeremy Hobson, 2018).
In laymen’s terms, rule-based translation is done by setting linguistic rules with enough detail that the translation program can pick the likely word for you (Robin, 2009). This first technology type was often problematic for The Machine, especially when it ran across homonyms. What do I mean when I say bark, nails, or mine? Don’t worry, The Machine often didn’t know either.
Next up was statistical technology. How likely is it that you mean to say tree bark instead of the dog’s bark? The Machine ran the likelihood and voilà! You had your translation.
Neural machine translation is an enhanced form of statistical translation. The difference is there are multiple layers, or neurons, of statistical abstraction (Jeremy Hobson, 2018). This allows for an almost human-like ability to distill meaning from one language and deliver it to another.
So, back to that business translation train of thought. On their home page, Architekst has this maxim: “My product is going to change the world. It needs to launch in a month. In 17 countries (Arkitekst.com, n.d.).”
According to economicshelp.org, worldwide exports of goods and services increased from $2,500 billion in 1970 to nearly 100 times that by 2016 (Pettinger, 2017). We now live in a world that is an ever-growing global marketplace.
Many factors have gone into this exponential growth. Travel has become easier, leading to increased movement of people and good around the world. In 1970 the use of steel shipping containers made it cheaper and more efficient to ship goods. There has been an increase of companies that operate in multiple nations, a reduction in tariffs, growth of trading blocks such as NAFTA, ASEAN, and the European Union, and the growth in the number of increasingly media connected people. And then there is the role technology has played. Technology has made it easier to hire internationally, bid on jobs internationally, find competitive wages for freelance work, and, with the use of translation technology, to talk to people anywhere about anything (Pettinger, 2017).
With today’s technology companies no longer have to rely on the second-language skills of people to ensure that transcription, voiceovers, subtitles and documents are precise and correct. Along with the growth of globalization has been the need for professional translation services. At Architekst their professional translation team recently translated 700 pages of Chinese legal documents into English, a market research paper from Dutch to English, and the inventory of a beauty product manufacturer into 18 separate languages (Arkitekst.com, n.d.). These would have been daunting undertakings without the incredible ability of modern translation technology.
Translation technology has made astounding strides in enabling better communication, and in so doing, has granted the open-minded, the curious-minded, and the business-minded entrance to a world of new possibilities.
Arkitekst.com. (n.d.). Retrieved from Architekst Tanslation Solutions for your Business: https://www.architekst.com/index-en.aspx
Catron, E. (2013, August 29). Longform. Retrieved from ScreenCrush.com: https://screencrush.com/movie-titles-lost-in-translation/
Jeremy Hobson. (2018, July 19). Here and Now. Retrieved from wbur.org: https://www.wbur.org/hereandnow/2018/07/19/translation-technology-future-language
Pettinger, T. (2017, April 6). Econmics help blog. Retrieved from economicshelp.org: https://www.economicshelp.org/blog/401/trade/what-caused-globalization/
Robin. (2009, December 13). Magazine. Retrieved from WorldofComputing.net: http://language.worldofcomputing.net/machine-translation/rule-based-machine-translation.html