In the past, I was thinking that Google Translate just analyzes the lexical meaning of each word. That’s why there are many translations that does not make sense at all. I supposed the words are not translated in a certain context; however, I learned that Google Translate uses the official papers of EU, which have been translated into many different languages for more than 50 years. In this way, the software finds similar context and translates accordingly. I still do not think that they can replace us, but we are getting close to be titled as ‘translation editors’ step by step. Here is the source that I learn all these. I wanted you to know how Google Translate works. Maybe there are still some translators who do not know the working principles exactly-like me 🙂 You can read the rest of the text by clicking link at the bottom:
“Using software originally developed in the 1980s by researchers at IBM, Google has created an automatic translation tool that is unlike all others. It is not based on the intellectual presuppositions of early machine translation efforts – it isn’t an algorithm designed only to extract the meaning of an expression from its syntax and vocabulary.
In fact, at bottom, it doesn’t deal with meaning at all. Instead of taking a linguistic expression as something that requires decoding, Google Translate (GT) takes it as something that has probably been said before.
It uses vast computing power to scour the internet in the blink of an eye, looking for the expression in some text that exists alongside its paired translation.”
For the rest of the text, click here.
Visit our Facebook page.
AIM Danışmanlık | Consulting 
Posted by José Carlos G. Ribeiro on October 3, 2011 at 20:38
Searching for data/information IS what Google does…
Their success is due to the extraordinary computer system they developed to ‘find’ things in a collection of data. They have created the ultimate Translation Memory system, which searches the WEB for a fuzzy match and uses that as a translation.
Whenever you use Google and provide a translation, the segment (source + target) is indexed and goes into this gigantic TM.
The limit is that story about training a monkey to hit typewriter keys. If you leave the animal pecking at the keys long enough, someday you may be able to extract a Shakespeare play from the output…
Posted by João Esteves on October 18, 2011 at 23:19
This your page provides a bit of insight on how GT works, whith is commendable if you consider the feelings professional translators have about machine translation. I always thought it could be something of the sort but had never met it so clearly explained – if explained at all – before.
Posted by gavinwheeldon1 on November 1, 2011 at 17:29
Machine translation provides an opportunity for all translators, certainly from a productivity perspective. The problem with publicly available ones is both the data privacy issues and the fact it is broad domain. If you use something like http://www.smartmate.co then you can build your own engine with your own data and the results are pretty spectacular.
Posted by Website Translator on December 10, 2011 at 16:59
A gigantic TM is not sufficient for proper translation, as any translator can attest. For example, the translation of a SAME sentence can have a different gender in the target language, depending on context. Genders of different words may be also different, depending on the language.
And I do not see all technical documents, literature or poetry being translated based on the translations of EU documents….