Google Translate may make for a passable travel dictionary for simple phrases, but how well can it handle the complexities of real life?
Everyone loves Google Translate. It now covers over 100 languages and, according to Google, there are over 500 million users. For traveling, it’s a hugely popular app, offering real timetranslation via voice, as well as the ability to snap a photo of a street sign and get an on-the-spot translation quick enough to navigate your way around strange cities like a pro.
But what, exactly, are those 500 million users actually doing with Google Translate? Chances are, not much more than extracting only the simplest of phrases… enough to get around a foreign city but not much else.
Although that’s hardly the end goal for the engineers who work tirelessly on the algorithm that drives the app, it’s pretty much all Google Translate good for at the moment, and here’s why.
Nuances? Complexities? Context? Not There Yet
Take, let’s say, a simple idiom that’s used often in business: “The ball is in your court”. If someone tells you this, it means you have a decision to make, an action to take, a responsibility to fulfill. It’s your move.
Feed it through the Google Translate mill a few times and it becomes:
The ball is in the court.
Somewhere in its formula, Google removed the meaning of “your” and the whole phrase lost its meaning. Now, this is one of the simplest of idioms, so much so that even a foreigner who’s never heard it before can sort of detect the meaning.
Google Translate, however, seems to revert to mere dictionary-level translation (word-for-word), resulting in the vague phrase: “the ball is in the court”.
If it Can’t Handle Simplest Idiom…
If it butchers even that simple idiom, just imagine what would happen to something like ” In 15th century Florence, lowered eyes signified modesty and obedience in women.” (from a V&A Museum description of Botticelli’s “Smeralda Bandinelli” painting).
So, as far as how well the app handles the complexities of real life with all its socio-cultural contextualisation, we’re just not there yet. For now, we’ll have to leave the tough stuff to human translators.