A century ago, it would have been the stuff of science fiction. Using a machine to translate a foreign language into English? It would have sounded crazy.
Today, though, it’s a reality. But not a perfect one.
One problem with this technology is that there are so many different ways a word, phrase or sentence can be used. It depends on the context and let’s face it, machines (so far) aren’t that great at figuring out exactly how a sentence in one language is supposed to sound in another.
For example, in Spanish some phrases have words reversed from how they would be translated into English. A human translator would be able to figure that out, whereas a machine might just translate them as they are written in Spanish.
In a recent study pitting a human translator against the Google Translate “state of the art” machine, the ability to translate coherently doesn’t turn out favorably for the app when it comes to important document-type information.
A human translator is able to understand the written context of a document and can interpret it and relay it in a manner that can be understood. A machine doesn’t have that luxury. It translates it as it sees it and sometimes that can translate into a confusing jumble of words.
The overall idea behind the document being translated comes through okay, but the grammar and broken up sentences can be distracting and potentially annoying enough to turn away a reader.
When faced with a Spanish documentation spoken into the Google Translate app, it does okay at translating it into English. However, it can’t understand different inflections of words and how those words are intended to be used. So again, for important documents, spoken human translation seems to prevail.
So far, and until there’s a way to put a human element into a machine, they just cannot compare to humans in perfectly translating languages. When a person writes or speaks, he or she puts emotions, inflections and ideas within the sentences. Machines can’t do that. They can’t read a phrase and think “okay, this is what they’re trying to say” and then translate that into English or some other language.
The Google Translate app and other machine-based translation technologies are still useful in specific situations. This would include ordering food in a restaurant or reading street signs – important aspects of traveling in foreign countries.
Another use would be for skimming over a foreign document or page, not to get the exact interpretation, but just an idea of what’s written.
In summary, when it comes to translating languages, humans are still a few steps ahead of machines.