This guest post was written by Verbal Ink Translation.
Google Translate is the latest in a number of services to attempt to automate the process of translating text between languages. But given the complexity of human languages, the question that most people must be wondering is how effective a machine can be at performing translation?
A human translator translates by relaying concepts from the source language into the target language. Obviously, this is only possible if the human translator understands both languages. The key term here is “understand.”
Computers have a hard time translating words and sentences between languages as they are not able to truly understand the meaning behind the words - they may be able to recognize a pattern of letters and words, but this does not equate to understanding. And the main reason for this is ambiguity.
Language is innately ambiguous. Almost every word in almost every language has many different definitions that apply to it. This adds up to a great many different translations are possible for almost any sentence as this study from verbalink illustrates. As humans, we can quickly see which definitions make sense because of the context and the meaning of the sentence. We reject the very improbable meanings as “nonsense.”
But computers lack this ability. This makes it impossible to know which definition is correct. It is a problem that programmers have been struggling with for decades. And the very poor results of machine translators have only served to illustrate this point.
In 2005, Google translate appeared on the scene, without great ceremony or publicity. But in the past decade, it has grown significantly in popularity. It seems to be satisfying a very real need that users have.
And as time passes, the service actually seems to be improving. It is able to “learn” from its mistakes and also from the huge quantity of text data that Google is collecting as the world’s largest search engine.
As yet, Google translate is not as reliable as a human being. Translations can be pretty bad at times, with distortions in meaning and incorrect phrases popping up from time to time. Still, when compared to the efforts of the past, Google’s service seems to eclipse all others in terms of accuracy, speed and flexibility.
Are we heading towards an age where computers can tear down the language barriers that separate us? Only the future can tell. But if the state of technology today is any indication, it may not be such a far-fetched dream, after all.
[Click on the image to get a closer look!]
