Just What Is the Most Untranslatable Word in the World?

In her column Untranslatable, Veronica Esposito considers why various words are so difficult to translate. Here, she looks at the famously difficult Tshilubà word ilunga and wonders what you think is the most difficult word to translate.

Just what does “the most untranslatable word in the world” really mean?

In 2004 a BBC article raised to prominence the Tshilubà word ilunga, citing it as “the world's most difficult word to translate,” based on the opinions of a group of one thousand linguists. At the time, the word circulated in translation circles with some fanfare, and it still gets cited occasionally as the most difficult to translate word in the world, mostly on the strength of that original BBC article.

But is it really? Just where did this title of world’s most untranslatable word come from, and how valid is it?

First of all, let’s break down ilunga. The purported translation of this word is: “a person who is ready to forgive any abuse for the first time, to tolerate it a second time, but never a third time.” Clever participants on the social networking website reddit supplied many of their own, more colloquial and concise translations: “three strikes rule,” “forgiving to a point,” “once forgiven, twice tolerated, thrice shunned,” “only two cheeks to turn,” or even “fool me once, blame me, fool me twice, blame you.”

Although the original BBC article is short on just what makes this word so untranslatable, it does state that, “although the definitions seem fairly precise, the problem is trying to convey the local references associated with such words.” That’s fair enough, although the article doesn’t explain why that goes more for ilunga than for the many other foreign words enmeshed in a web of local references.

Ilunga comes from the Tshilubà language, also known as Luba-Kasai, which is one of four national languages spoken in the Democratic Republic of the Congo. In the most recent estimate I could find, for 1991, it was estimated to have around six million native speakers, and it is primarily spoken in the southeastern part of the DRC. The language is a part of the larger family of some six hundred Bantu languages, spoken by about 350 million people in Africa, or one-third of that continent’s population. Bantu words that have become widespread in English include simba, ubuntu, jenga, safari, mamba, Kwanzaa, bongo, and chimpanzee.

Even when the BBC article came out, some were already raising a skeptical tone. First of all, the survey was commissioned by a professional translation company named Today Translations, and it seemed to be more about marketing than edification. At the popular linguistics blog Language Log, Mark Liberman wanted more information about just who were these one thousand translators surveyed and how it was done. Others pointed out that the survey had never considered the question of what the target language was, rightfully arguing that different target languages would make for an easier or harder translation. An unsigned article on the Greenwood Calendar website further noted that “not all of the words on Today Translations’ list were even legitimate. Some of them turned out to be mistakes and hoaxes.” Looking at a 2008 posting on the Today Translations website, the whole thing seemed like an exercise in public relations (which apparently worked very well). Interestingly, nowadays Today Translations seems to have come to rely less and less on professionals, instead touting its “instant AI-translations” that are “powered by pure data.”

So, what exactly does make a translation difficult?

So, what exactly does make a translation difficult? The answer to that is, of course, very complex and multifaceted—humor may be a good place to start, since humor is notorious among literary translators for being one of the things they find the most challenging to move from one language to another. Writing about just that topic, Jascha Hoffman of the New York Times noted that “when complications do arise [with translating humor], they are usually caused by one of two tricky areas: cultural references and wordplay.” That is, jokes are funny because of the web of unspoken cultural referents that they draw on, but if you pick up a joke from one culture and move it to another, it often just seems nonsensical. And wordplay tends to draw on the particular construction of a word, or similarities between one word and other, which of course can vary greatly from language to language, making it nigh impossible to translate.

Another notion often drawn on when considering difficult translations is the idea of compression: one language may put lengthy and complicated ideas into concise colloquial form, whereas another language may not have an elegant solution for reproducing this. Ilunga would seem to implicate this idea, as the standard English translation of the word requires many, many English words, whereas ilunga is just one word in Tshilubà. In the various suggested idiomatic translations of the word, one can see attempts to get at that compression and approach an elegance that can be had with the original.

Compression tends to come up a lot in conversations around untranslatability. For instance, another frequent favorite on hard-to-translate lists, mamihlapinatapai, tends to show up there because it compresses such a complicated idea into just one word: “a look that without words is shared by two people who want to initiate something, but that neither will start.” Such compression seems to delight and seduce English speakers, perhaps because our language is not a very inflected one, and thus tends to rely on phrasing to explain complicated ideas rather than collect it all into a single, highly inflected word. Mamihlapinatapai is also popular because it just seems like English should have a word for this, and people seem to delight in knowing that there is a word out there that means what it means. Thus words like ilunga or mamihlapinatapai will tend to seem almost mystical in the amazing compression that they achieve compared to English.

Lastly, there is the notion of “iconicity,” that is, the degree to which the form of the word itself is a part of its meaning. Standard examples of this would be onomatopoeia, like cuckoo or sizzle, and religious words like inshallah or hallelujah also show a high degree of iconicity for their performative significance. Of course, this can be very challenging to translate because a word’s iconicity can be very connected to how it is pronounced, or the space it holds within a religio-cultural framework, all things that aren’t so simple to include in the context of a seamless translation. As a measure of its difficulty, words with high iconicity are often said to occur frequently in poetry (think, for instance, of e. e. cummings), which is of course notoriously difficult to translate.

I’m curious to know what the readers of this essay would consider to be a challenging translation.

By these measures, it would seem that ilunga, while a challenging translation, is not particularly difficult, certainly not hardest-word-ever status. We may never know the precise logic behind why this word attained brief fame as the “hardest to translate word” and still occasionally gets described as such. Seemingly, the whole notion of an “impossible translation” is full of so much complexity and so many different, conflicting sides that I’m not sure it makes sense to even have a ranking like “hardest word to translate.” Nevertheless, I’m curious to know what the readers of this essay would consider to be a challenging translation—if you are reading this now and have a suggestion, please do share it with the #untranslatable hashtag on WLT’s Twitter account and tag me @la_doppia_vita.

Oakland, California