Jalees Rehman in The Next Web:
Bilingual or multilingual friends can be quite annoying. Especially if you’re stuck at a social gathering with the ones who repeatedly mention their language skills and utter phrases such as ”Well, if only you could read this novel in the original, you would have a much more profound understanding of what the author wanted to express…..”. Or the ones who like to cite French, German and Arabic language newspaper articles and then remind you with a thinly veiled pomposity that you may have a very narrow view of the world if you only rely on English-language news.
However, this latter group is becoming more rare, possibly because a formidable foe is taking the wind out of their sails: Google Translate. The excellent book “Is That a Fish In Your Ear” by David Bellos has a chapter entitled “The Adventure of Automated Language-Translation Machines”, which is especially thought-provoking, because it explains some key concepts about Google Translate and the future of automated translation.
If a user enters a text into Google Translate, the linguistic search engine scours the Internet for multilingual texts, ranging from official documents posted by the European Union to articles and books that are available online in bilingual or multilingual versions. Using pattern recognition algorithms and statistical methods, Google Translate matches words and phrases contained in the user-entered text with those found in the large online repository of previously translated texts.
The underlying assumption of Google Translate is that any new text requiring translation contains phrases and word patterns that have been adequately translated in the existing online collection of bilingual or multilingual texts. Anyone who has used Google Translate can appreciate the success of this approach.
More here.