There’s no word in English for the Spanish term friolero, a person who is especially sensitive to cold weather, or in Spanish for smirk. The German word Waldeinsamkeit, the feeling of being alone in the woods, does not have a direct translation to English or Spanish. While humans can at least come close to approximating translations for tricky words like these, machine translation hasn’t caught up. Why so?
Robots in trouble
Things go smoothly as long as the machine translator is able to find a suitable candidate for the word in question, and – if it is sophisticated enough to manage computational linguistics – a functioning structure to make it work in its new language. But untranslatable words are not in databases, and if they are, any resemblance to the original term is purely coincidental. It’s also possible that none of the equivalents the machine is considering, works in the destination language’s syntax. Machines cannot even use a circumlocution, because they really aren’t so sure about what needs to be circumvented.
Language and thought may be more closely linked than you might realize. For one thing, languages are more similar to each other than some people used to believe. Take the Eskimos, for example. They do not have hundreds of words for snow, but a dozen at the most, which is not far from English (snow, sleet, slush, blizzard, avalanche, hail, hardpack, powder, flurry, dusting).
How is this relevant for our discussion? Our innate language instinct that we all share, helps us understand each other even when we don’t want to. For humans, it is always possible to translate between two languages or, for that matter, between two cultures.
As the experimental psychologist and Harvard professor, Steven Pinker, put it in his book How the Mind Works: “When English-speakers hear the [German] word Schadenfreude for the first time, their reaction is not, “Let me see . . . Pleasure in another’s misfortunes. . . What could that possibly be? I cannot grasp the concept; my language and culture have not provided me with such a category.” Their reaction is, ‘You mean there’s a word for it? Cool!”
There are a few reasons why one particular language may have a larger vocabulary than another: new words might have been acquired from contact with other languages, its culture may boast the existence of influential word creators, there could be more lax rules for forming new words, or the literacy rate in the countries where it is spoken might be particularly high. That’s why humans, even when there is no word for a concept in their language, can make a circumlocution or use a metaphor when they encounter a word they’ve never heard before.
Wikipedia offers an extensive appendix with terms considered difficult or impossible to translate into English from 46 languages. The list makes a very entertaining read for humans – but if you are a robot you may want to skip it. Currently machine translators that are used to automatically translate website content or written documents don’t know what to do with such beautiful words as Waldeinsamkeit, let alone how to write a whole poem about it, as Ralph Waldo Emerson did, or design a series of illustrations like Anjana Iyer in Found in Translation.