Imagine you’re in a foreign country using a software device that can translate any language into your native language. No more walking around with a bilingual dictionary trying to communicate your basic wants and needs! But is this future really possible?

Technology policy expert Alec Ross believes it is. He believes we’ll be able to communicate with one another via earpieces with built-in microphones in as little as a decade. However, there are many who believe that language is far too nuanced to ever be interpreted by a computer, including Professor David Arbesu, Associate Professor of Spanish at the University of South Florida.

In a recent article, Arbesu argues that while replacing a word with its equivalent in the target language is the ‘easy part’ of a translator’s job, it’s no piece of cake for a computer. This is because computers are only capable of translating the literal meaning of words.

To demonstrate his point, Arbesu shared a personal experience in which one of his students accidentally wrote that he “sawed his parents in half,” when he meant to say he “saw his parents.” The true intention behind the words was easily caught by Arbesu, but translation software, incapable of understanding the intricacies of the language, was unable to render it meaningful.  

Why is language so hard for computers to understand?

“The reason why language is so hard for computers to understand is because language is much more than words put into the right order. Volume, pitch, situation, and context conveys just as much meaning as the words we use.”   – David Arbescu

Languages use metaphors and tropes, adjectives and adverbs to communicate meaning.

“Abstract and figurative speech or concepts are next to impossible for a computer to understand and adequately interpret.” – David Arbescu

The way we speak may in fact have nothing to do with the reality around us, making human communication doubly confusing for the machine. 

Computers are rational and logical, so for example, if you search for a translation of  “I’m eating my heart out,” it will be literally translated to “Je mange mon coeur” — which means “I’m eating my heart” — because computer software does not understand the emotional feeling behind these words.

People, on the other hand, are both rational and logical, but also emotional and illogical. Therefore, human communication is both logical and emotional. But computers are always logical. This, describes Arbesu, “is the paradox of computers and languages.”


Because of its idiosyncrasies, emotion and abstracts, one can reasonably assume that computers will never replace human beings as the masters of human speech.

In the same way that machines help lawyers, doctors and teachers, but have not replaced them, machines can help translators work better and faster, but — unless an incredible technological breakthrough is made — they cannot be replaced by them. 

Because, writes researcher Ben Screen, “while machines have become quite good indeed at translating the text, when it comes to text-behind-text, they need help.”