Using a pioneering approach to machine translation (MT), search behemoth Google now provides translations from 52 languages through its Google Translate service. Google has capitalized on its access to unfathomable amounts of data, largely in the form of transcripts from the proceedings at the United Nations, which have been rendered into some 23 languages by professional human translators. Google Translate trawls this invaluable source of data, along with text from the Google Books scanning project and additional Internet resources, for likely translation matches. Internet users access the tool tens of millions of times each day to translate information as they surf the web.
While Google Translate has made impressive strides in our ability to understand and communicate with the rest of the world, what do the future prospects look like for the service and other machine translation programs? According to the leader of Google’s machine translation team, Franz Och, “This technology can make the language barrier go away.” Other linguistics experts contend that MT will strengthen linguistic diversity by freeing the world from the need to focus on dominant languages such as English. Ironically, one potential consequence of the widespread use of tools like Google Translate is decreased incentive for individuals to learn English and/or become multilingual.
Though some experts claim that Google Translate’s results will better with time, researchers and computer scientists working on the project note that the system is unlikely to dramatically improve with the addition of more data. “We are now at this limit where there isn’t that much more data in the world that we can use,” notes Andreas Zollmann, a Google Translate team member, “so now it is much more important again to add on different approaches and rules-based models.”
Of course, detractors state that regardless of the technological advances made, machine translation will never learn to pick up on the cultural undertones and subtleties at play in language. Jokes, idioms and wordplay are largely lost on Google Translate, which fails to capture the “flavor” of the text. According to author Douglas Hofstadter, “There is no attempt at creating understanding, and therefore Google Translate is doomed to the same kind of failure forever.”
Machine translation is nowhere close to a usable standard. For languages such as German to English, you can get the main idea, but it’s really bad at other languages. What bugs me is the people that say, “I use Google Translate on my website to gain foreign readers.” Little do they know foreign readers are like WTF? 🙂