I was giving a talk once with a writer I’d translated, and he was telling the audience one of those stories about untranslatable words – you know the kind – in this case a word from an indigenous language in southern Patagonia, and the word means, well, when a man and a woman are in a bar, and he looks at her, and she looks at him, and they look at each other and their looks say okay I’m interested in you but you need to make the first move and come over to me? The word means that. Everyone laughed, the writer is charming and tells a good story. He’d told the story because I had just said that as a translator I didn’t like to admit anything was untranslatable. Now I said, “See, you translated it! You told us in English, and everybody laughed!” He said, “But you can’t translate it in one word—” and I said, “Well, what matters more to you, how many words it has or whether everybody laughs?” 

Translators read to decide what’s important and what’s less important, then re-create what they’ve decided is important; the debates about whether to translate literally or for the overall sense, whether to domesticate a term or leave it foreign sounding, whether to translate rhymed and metered poetry with rhyme and meter, and all the rest are pointless, because – obviously – it depends. These decisions tend to get made unconsciously and instinctively, even if the translator likes to rationalize them afterward, and the decisions are different in every novel or story or poem or instruction manual, and at every moment of every such utterance. Now what matters is a rhyme, a joke, a repetition, a specific description of a particular object, a verbal tic, verve, historical accuracy, sentence structure, now it’s something else – in instruction manuals it’s mostly the literal content – and translators snake through the text carving off what they can live without from what they can’t. They gerrymander unscrupulously. Because if what matters is that look in a bar, then the translator needs to spend a sentence or two describing it, or include a footnote (a footnote isn’t “added to” the translation, it’s part of the translation). If what matters is keeping the text smooth and lively with no footnotes, then that’s what you’re translating for and that’s what you’ll get. If a character says something that would take fifteen words or inappropriately technical language to express in English so you substitute something else to keep the conversation moving, that doesn’t mean the word for that thing was untranslatable – it means that you’re translating for naturalness of dialogue. Anything left untranslated is what you’ve decided doesn’t matter anyway. 

In a deeper sense, nothing can be left untranslated. Since translation is realignment, nothing will stay fixed – nothing can stay fixed, even if you want it to. When you look at the world from a different location, aiming your gaze in a different direction, every angle and occlusion and vector will change. To say something is untranslatable is to say that every reader in all of space and time would read it the same way, and that is not what reading is. (It would be like saying a chair is always seen from the same angle.) The limiting case is described in Borges’s amazing story of Pierre Menard, who tried to write a book that would be word for word the same as Don Quixote. Once he succeeded, for a couple of chapters, the result was a different book despite every last word being exactly the same. It was different because realigned: Don Quixote written by a French Symbolist poet for modern readers. 

A class of examples that makes the impossibility of untranslatability especially clear is when a word or passage is in a foreign language in the original text. Even if you leave it in the other language, you can’t leave it untranslated. For example, Tolstoy’s novels have a lot of dialogue in French: the characters’ use of French was meaningful information, and he also expected his readers to know French. Some of Tolstoy’s translators into English keep the French and put a translation in a footnote or endnote; some translate the French, with or without slipping in “he said in French” instead of “he said.” Since twenty-first-century American readers cannot be expected to have the same relationship to French as Tolstoy’s nineteenth-century Russian readers, simply keeping the French without a note is not automatically the right translation choice. And even if that is the choice a translator makes, those French utterances have been realigned – translated into things characters say in an English-language book. They are as different from Tolstoy’s original French dialogue as Menard’s words are from Cervantes’s.

Several books I have translated use English foreign words – that is, English words in the German or French original book. This is clearly a special case of the situation we saw with Tolstoy’s use of French: the English-language reader’s relationship to English cannot possibly be the same as the original reader’s; the relationship between the English phrase and the rest of the text cannot be the same when the rest of the text is also in English. When the author chooses English because the dialogue is spoken by Americans or because he or she wants to call attention to an ideological implication or other quirk of the English language, I have often indicated that to the reader, for example with a note in the front of the book: “All text in SMALL CAPS is in English in the original.” 

When the author’s ear for English isn’t perfect and he or she has a character say something they would never actually say, for instance a New York cabdriver saying “Madam” instead of “Ma’am,” I translate the author’s English-language dialogue into English: the taxi driver who in the German original says “Yes madam” now says “YES, MA’AM.” There is nothing to be gained by incorporating a gotcha on the original author’s competence in English. Sometimes, when I feel that the English language has just permeated the other language and the author is using English words (“laptop,” “sweet sixteen”) alongside his or her other vocabulary, not making any point about English, it works to simply leave those words un-marked in the translation. In effect, the word is just a cognate, indistinguishable from all the other German or French words of the original. 

A useful teaching exercise I’ve tried on occasion is to ask everyone in a class to raise their hand if they know more than one language, then tell the students who didn’t raise their hands that “you’re bilingual too,” because of the split Latinate/Germanic vocabulary of the English language and its roots in history. They already know the difference between “Help!” and “I request assistance,” between a ceremony formal enough to be called “commencement” and a “fresh start”; they can appreciate why science sounds intellectual in English because of words imported during the Renaissance, like sulfur, hypotenuse, marine biology, as opposed to if these were called “stinkstone” or “longline” or “sea lore”; they can quickly grasp the political history of these different registers going back to the Norman Conquest (on the Anglo-Saxon peasant’s farm, it’s cow, lamb, pig; on the French nobleman’s table, it’s beef, mutton, pork; the words estate and property are Latinate, while the rent that’s due is Anglo-Saxon).  

English is not only half-Latinate half-German either. As a mercantile language, stripped of many endings and other grammatical markers as a way of reconciling the initial clash of Norman French and Anglo-Saxon, it has long been famously adept at borrowing and acquiring vocabulary from other world languages. 

While it isn’t a case of translation per se, let me end by describing a beautiful short story by the Indian writer Vishwapriya L. Iyengar originally written in English, “No Letter from Mother,” about translation, English communication, and the cultural translations and discontinuities of the main character, a girl at an international boarding school in India in the 1960s. It reinforces the idea that no language is monolingual. 

We hear in the story about rajnigandha growing at the Madonna’s feet, or someone eating “English toffees from Kenya,” or “Zarina from Zambia … passing a stick of Wrigley’s Spearmint to Bina from Calcutta.” The first quoted dialogue in the story is: 

“Holy Mary Mother of Christ, blessed is the fruit of thy womb …” I was late. 

“Always late,” Sister Francis screamed, “in her last janam she must have been a tortoise.” 

This is funny – the Catholic nun screaming about Hindu reincarnation – and it mixes languages in the literal sense, too, not just discourses or voices: janam is Hindi, not English, or at least not in the OED. But English has always done this: other words Iyengar uses, like dhobi for a washerwoman, and dhow, and brinjal, are now “inside” “English” “proper.” One girl, who is key to the plot of the story, has a dress with “French chiffon pleated ruffles, but they looked shabby with washing, the dhobi had burned the fabric a bit, so it wasn’t ‘foreign-foreign.’” This French chiffon dress – something foreign, burned by its contact with another culture, and no longer foreign-foreign – is just like the borrowed Hindi word dhobi, and the borrowed French word chiffon, and the Anglo-Indian grammar of foreign-foreign. 

Central to the story is that all the girls look forward to letters from their parents, including the narrator. Her father’s letters are a somewhat dry anthology, “full of anecdotes from Birbal, Tenali Raman and Kautilya” or “Gladstone or Disraeli”; they sound “a little boring, like Nehru.” Still, they’re all she has, because “Mother never wrote me letters. Mother was always reading Kannada novels by women late at night.” (Kannada is one of the official languages of India, in the same family as Tamil.) As the narrator says near the end: “I was too young to realize that colonisation had cut the bond twixt mother and child.” Yet the story rebuilds these broken bonds. Her mother would read these Kannada novels out loud late at night, “when the house had been quilted to sleep in a soft blend of mixed breathing.”  

When the narrator finally does get a letter from Mother, belying the story’s title, it has none of her mother’s trademark spicy metaphors – “I cannot write in English,” the letter bluntly says. “Your teachers will make fun of my English so I will not write again to you” – but by the time the narrator tells the story, she has learned how to speak in her mother’s voice, or with it, in a soft blend of mixed breathing. 

Iyengar’s point, and mine, is that language is not pure and cleanly delimited, even in a monolingual context or in a so-called source language or target language. And with this we are back at the idea that we translate utterances: not words of a language, not anything that can be described as purely linguistic. Naturally, Iyengar’s story isn’t evidence for anything except that she describes or constructs the English language as hybrid and polylingual. But this lesson is true to my experience as a translator: everything I do is a case of believing that one language can and always will carry traces of another, that the boundaries are permeable, and that (in my case) English is capacious enough for “other” languages. The idea of making English be something non-English is alien to my experience as a translator, and to that of every translator I know of. When I read like a translator, I see what the original author is doing with and to their language; when I write, I am trying to activate and exploit the capacities of English; by yoking the two, I deny the isolated monolinguality of both languages. The original text is inherently able to enter a newly enriched and expanded English, and English is inherently able to welcome it in. 

When it comes to translation, nothing is foreign-foreign. 

This is an edited extract from The Philosophy of Translation by Damion Searls, published in the UK by Yale University Press