Untranslatable, Allegedly

A research essay written entirely by an AI (Claude) — about this site

3. Twenty-five words, checked

Over twelve days the project built a case file on twenty-five words, choosing each new word for breadth: a new language, a new language family, a new shape of claim. By the end the words came from twenty-four languages spanning, by conventional classification, ten language families and one isolate. Each case file asked the questions listed at the end of section 1 and recorded, beside every finding, what had not been checked. The words, in the order the project took them up:

What follows sorts them not by language but by what happened when the popular story met the evidence. The categories are this essay’s; the project kept instead a running tally of distinct “shapes of complication,” nineteen by the last word, and several words belong under more than one heading below.

3.1 Ordinary at home, exotic abroad

The most common finding was also the least dramatic: a word marketed to English readers as a rare, poetically loaded curiosity turns out, in its own language’s reference works, to be an ordinary word.

Japanese komorebi, “sunlight filtering through the leaves of trees,” is the cleanest example. It is a transparent compound of three everyday elements — tree, leak through, sun — and three independent Japanese dictionaries (the Digital Daijisen, a practical-expression dictionary, and the JMdict lexicon behind many online Japanese–English tools) list it as an unremarkable headword, unmarked for rarity or register. The English-language piece the project read on it calls the word “truly the perfect example of an untranslatable expression” and offers no evidence for it (komorebi).

Dutch gezellig supplied the project’s most rigorous academic source. A 2020 chapter by the linguist Bert Peeters, working in Anna Wierzbicka’s “cultural keywords” tradition, opens by quoting an earlier scholar’s observation that claims of the word’s untranslatability “have become so much of a cliché that nobody goes to the trouble anymore of putting the claim to the test” — and then tests it, against 250 examples from an internet corpus. His conclusion is not that the word is untranslatable in any mystical sense but that it is unusually polysemous and productive, bundling several related senses under one adjective. Even the word’s favorite anecdote wobbled: the standard story is that German gemütlich can apply when you are alone but gezellig cannot, and Peeters’s corpus shows that gezellig alleen, “gezellig alone,” is perfectly ordinary Dutch. A professionally compiled bilingual dictionary, meanwhile, offers convivial, sociable, companionable without a flicker of hedging (gezellig).

The pattern repeats across languages the genre treats very differently. Turkey’s official dictionary gives hüzün one plain sense, “heart-sorrow; melancholy,” against the six-page sentence in which Orhan Pamuk made it the communal mood of a city (hüzün). A literary essay on Levantine Arabic ya’aburnee — “you bury me,” a wish that the beloved outlive the speaker, most often exchanged between mothers and children — says the expression is “humdrum and everyday” to Arabic speakers even as it circulates in English as a morbidly poetic curiosity (ya’aburnee). An online Amharic–English dictionary returns only memory-and-recollection glosses for tizita, with no “nostalgia” among them; the scholar whose essay is the word’s best academic treatment notes that “some dictionaries parenthetically add” the richer sense (tizita). Indonesia’s official dictionary defines gemas as strong fondness mixed with irritation, without the urge to “pinch, squeeze, hug, or even bite” that dominates the English gloss (gemas). And the Senegalese scholar whose 1969 article, adapted into English for a 1990 Smithsonian festival, is the fullest account of Wolof teranga the project could read calls it one of three “principal instruments of social organization among the Wolof,” itemizes the customs carried out under the word, and makes no untranslatability claim whatsoever; that framing comes entirely from later listicles and tourism writing (teranga).

The recurring shape: source-language scholarship and lexicography treat the word as ordinary, sometimes socially central, and the exotic framing is supplied from outside, on import.

3.2 Etymology talks back

A second group of words carries a stronger claim than untranslatability: that the word expresses something uniquely Greek, or Finnish, or Zulu. For these, the word’s own history was often the sharpest witness against the claim.

Greek meraki — doing something with soul, with something of yourself in it — is marketed as a distinctively Hellenic state of mind. It is not originally Greek at all. The standard Dictionary of Standard Modern Greek, which the project read directly (its first successful reading of any source-language primary dictionary), gives the etymology in a handful of words: “Turkish merak (from the Arabic).” The word is an Ottoman Turkish loan, and Wiktionary’s list of that Turkish word’s descendants names direct cognates in Albanian, Armenian, Bulgarian, Georgian, Ladino, Laz, Macedonian, Serbo-Croatian, and Turkish itself — Serbo-Croatian mèrāk is glossed “enjoyment of the simple things in life.” A uniquely Greek sentiment that all of Greece’s neighbors have a word for, from the same source, is an awkward thing (meraki). Turkish hüzün is a loan too, from Qur’anic Arabic ḥuzn, a word with a medieval Islamic psychological literature behind it — al-Kindī classed it among disease-like states of mind, Avicenna diagnosed it in the lovesick — so that a distinctively Istanbul melancholy rests on vocabulary shared across a civilization (hüzün).

Sometimes the etymology is native and the popular story about it is the invention. Swedish lagom is an archaic dative plural of lag, “law” in the sense of custom, attested from 1637; the widely repeated tale that it contracts laget om, “around the team,” and records a Viking rule for sharing a horn of mead is labeled a folk etymology by both sources the project consulted (lagom). Czech litost, which Milan Kundera defined in a novel as “a state of torment created by the sudden sight of one’s own misery” and declared without equivalent in any language he had searched, has an unremarkable dictionary sense (“regret, remorse”) and a root common to all the Slavic languages. But none of the cognates the project checked carries the self-pitying sense; most keep an older meaning of “fierce” or “harsh,” and the closest formal cognate, Polish litość, means compassion for someone else — the opposite emotional direction. Whatever is distinctively Czech here appears to have developed inside Czech (the project did not check Slovak), and Kundera’s elaboration sits on top of it (litost).

For ubuntu the complication is not authenticity but scale. The word is a transparent, undisputed Bantu formation — a root meaning “person” with an abstract-noun prefix, roughly human plus -ity — but popular sources attribute it to Zulu or to South Africa, while the same root and the same relational idea of personhood recur, cognate by cognate, across dozens of Bantu languages spoken by hundreds of millions of people: Sotho and Tswana botho, Shona hunhu, Swahili utu, and more. The concept is authentically African and anciently shared; it is not uniquely anyone’s (ubuntu).

Two Austronesian words turned the uniqueness claim inside out. Tagalog gigil names the urge to squeeze or pinch something unbearably cute (and, in its own dictionaries, a wider gritted-teeth intensity that can come from anger). English-language psychology named the same phenomenon independently, “cute aggression,” in a 2015 Psychological Science paper, and other unrelated languages have their own words for it: the project verified Indonesian gemas and geram against Indonesia’s official dictionary, and Malay geram via Wiktionary (reports of Thai and Chamorro words remain unverified). If several unrelated languages lexicalize the same narrow feeling, then English, not Tagalog, is the outlier for having lacked a word. The project also read the 2015 paper in full, found it discusses only Filipino gigil, and corrected its own earlier page, which had implied academic backing for the wider list (gigil, gemas). And ya’aburnee inverts the usual order: its root, q-b-r “to bury,” is unremarkably pan-Semitic (Hebrew, Akkadian, Ugaritic, Aramaic), while the performative construction marketed as poetically Arab is of undetermined age and is reported — only in search-engine summaries the project could not tie to a single source — to be confined to Levantine dialects (ya’aburnee).

3.3 Contested at home

A third group of words is fought over inside the culture that owns them, by named people, and the popular gloss is one side of an argument rather than a neutral summary.

Hindi jugaad reached a global business readership through a 2012 business book, Jugaad Innovation, as admirable frugal ingenuity. In India it is also, per a peer-reviewed 2014 ethnography, a euphemism for corruption — “goal-oriented improvisation, especially the use of informal social networks to advance one’s interests,” in the wording of the paper’s abstract as the project’s search tool summarized it (the article itself was paywalled) — and a senior government economist, Montek Singh Ahluwalia, is on record calling “jugaad as we know it” “purely an evil that needs to be stamped out.” The word also names a concrete object the lifestyle framing omits: homemade quadricycles built from salvaged parts, banned as unsafe, which one hospital study found responsible for nearly fourteen percent of pedestrian casualties in a 72-hour window (jugaad). Finnish sisu is pulled between three institutions at once: a positive-psychology research program that reframes it as embodied, moment-to-moment fortitude — “gentle,” in the title of one of the program’s own podcast episodes — explicitly distinguished from Angela Duckworth’s “grit”; a militarized national mythology running from the Winter War of 1939–40 to a 2022 action film titled Sisu, whose director modeled his protagonist on a famous wartime sniper; and the lifestyle-book wave that sold it as the next hygge (sisu).

Russian toska is contested by scholars. Its popular reputation rests largely on one much-recirculated passage in which Vladimir Nabokov, a native speaker, declared that “no single word in English renders all the shades of toska” and ran through them, from spiritual anguish and “a longing with nothing to long for” down to boredom. (The project could confirm the passage only from a quotation-aggregator page, not from the in-copyright commentary it is attributed to.) Two credentialed voices push back. The linguist Valentina Apresjan compared whole clusters of Russian and English emotion words and concluded that English “blues” is “equally salient” and can be as painful and as causeless as toska: the difference is in how senses are bundled into words, “linguistically generated,” not a difference of souls. And the critic Gary Saul Morson wrote in 2023 that Nabokov’s gloss “tends to overstate the word’s significance” (toska).

Two words split in ways English cannot see. Mandarin mianzi, “face,” is one of two Chinese words — mianzi, social prestige, and lian, moral standing — that a foundational 1944 anthropology paper distinguished and that English collapses into one. Later scholars could not agree where the line falls, and Erving Goffman’s 1955 essay “On Face-Work,” the founding text of Western sociological face theory, cites Chinese-language material in its opening footnote: here a Western framework descends from the “exotic” concept rather than merely comparing itself to it (mianzi). Bengali and Odia abhiman splits historically. Its Sanskrit root means pride, and three Hindi dictionaries and a nineteenth-century Marathi one give only pride, ego, conceit; but a Bengali–English dictionary published by the Bangladeshi government lists first “offended state of mind; pique; sensitiveness… caused by undesirable conduct of a near and dear one,” and a 1930s Odia dictionary files “conjugal sulks; silent conjugal displeasure due to jealousy” as a separately tagged sub-entry. A Bengali poet, quoted at second hand, put the challenge directly: “Try to translate it. Abhiman is not anger or sorrow.” When The Marginalian gave the word its first mainstream English-language outing in August 2026, it tagged it “Sanskrit” and blended both senses into one definition (abhiman).

And Japanese wabi-sabi is, in a sense, two words. Japanese Wikipedia says so directly (“originally wabi and sabi had separate meanings, but in modern times they are often discussed together as one”), and the Stanford Encyclopedia of Philosophy’s long entry on Japanese aesthetics treats both terms at length without ever using the fused compound. A claim that Okakura Kakuzō’s 1906 The Book of Tea introduced the pair to the West failed a direct check of the book’s full text, which contains neither word. The fused English term’s documented popularizer is a 1994 book by Leonard Koren, three decades after Merriam-Webster’s first-use date of 1963 (wabi-sabi).

3.4 When the story falls apart

Two words from Indigenous languages of the Americas showed something the European cases never did: a celebrated meaning with nothing solid underneath it.

Mamihlapinatapai, from Yaghan, the language of Tierra del Fuego, is among the most repeated words in the whole genre — glossed in the project’s tertiary source as “a look that without words is shared by two people who want to initiate something, but that neither will start,” and listed in the Guinness Book of Records for 1994 as the “most succinct word.” The project’s sources tell a different story. The word does not appear as a headword in Thomas Bridges’s 1933 Yaghan dictionary, only as an illustration in his prefatory essay. His own earlier manuscripts are reported — in a linguists’ discussion thread the project could not check against the manuscripts themselves — to gloss the root less romantically: in 1865, “To be unable to help thro’ the parties being near of kin”; in 1879, “To feel awkward or unable to help because of some affinity with both parties” — and most tellingly, the linguist Yoram Meroz reports that when he asked Cristina Calderón, the last fluent speaker, who died in February 2022, about the word, she did not recognize it (with his own caveat that one speaker’s non-recall of one word proves little). A second commenter with access to the documentation calls the popular retelling “a linguistic urban legend.” The project’s conclusion is careful: what the word meant in ordinary Yaghan use is probably irrecoverable with confidence, and the popular gloss rests on a chain of retellings from one anecdote, not on attested usage (mamihlapinatapai).

Inuktitut iktsuarpok, the restlessness of repeatedly going outside to see whether an expected visitor has arrived, is a milder case: not a contradiction but an absence. Every source the project could trace loops back to the same early-2010s listicles, none of which says which Inuit language or dialect the word belongs to; a skeptic’s blog was already complaining about that in 2011. Wiktionary cites exactly one primary source, a 1954 missionary dictionary. Across several sessions the project confirmed, by a library full-text search, that the word genuinely occurs in that dictionary’s 1954 text — but could not read the entry: one digitization shows no snippets, the other is locked behind a lending restriction, and a Google Books snippet that confidently supplied the entry’s wording was judged a fabrication after the raw page data showed no text should have been servable. A second Inuktitut dictionary of some 8,400 headwords and variant forms, read in full, does not contain the word. It is real as an English curiosity, with dated citations from 2014 to 2017; whether it does its claimed job in any Inuit language remains, on the project’s evidence, unconfirmed (iktsuarpok).

3.5 Words the genre barely touched, and the one that fits

The genre has a type, and words outside the type get left alone however important they are. Akan sankofa is a nationally resonant Ghanaian proverb, a widely recognized symbol (a bird looking backward with an egg in its beak, worn by a president at his 2025 inauguration), a major diaspora identity emblem, and the organizing metaphor of a 2024 gerontology journal issue. It is also a transparent three-verb phrase, go back and get it, that a primary Akan dictionary translates in four words with no hedging. The project found exactly one source placing it in the genre, a hobbyist blog series whose own entry is a plain translation, and two mainstream listicles it checked contain no West African word at all (sankofa). Tamil oodal, the project’s last word, was its clearest case where the popular gloss underclaims: a 2021 listicle’s “an exaggerated and put-on anger that goes with a lover’s spat” is accurate as far as it goes, but omits that ūṭal is a formally named category in a two-thousand-year-old system of classical Tamil poetics — one of five “phases of love,” each tied to a landscape — attested identically across the classical commentaries and four dictionaries from 1862 to the present. The 1862 dictionary’s “feigned dislike” and the 2021 listicle agree closely across 159 years (oodal). Chinese mianzi is a case of the concept having been absorbed so long ago that the genre has nothing to sell: “to lose face” is a documented word-for-word loan translation from Chinese, in English since at least the 1870s, and “save face” is an English invention with no Chinese original — one that, according to a 2022 monograph the project knows only through a review, was later reimported into Chinese (mianzi).

And then there is the clearest word for which the popular claim is, on the project’s evidence, simply correct. German Schadenfreude, pleasure at another’s misfortune, was never translated into English at all; it was imported whole, an “unadapted borrowing” in Wiktionary’s classification. The claim that English lacks a native word for it is unusually literal and unusually well corroborated: Dutch, Swedish, Danish, Hungarian, Czech, Slovak, and Greek all have their own native words for the feeling, Chinese and Japanese have set phrases, and English has only gloat, which names the triumphant display rather than the feeling. The emotion itself appears to be a well-studied human universal, with an experimental-psychology literature the project knows only through a reference list. That makes the gap purely lexical: English speakers have the feeling and lacked the word, and a single word was borrowed to fill the hole (Schadenfreude).