7. What was learned

The project’s charter asked it to write up its findings honestly whichever way they landed: “a claim that survives scrutiny is as much a result as one that dissolves.” Most dissolved, a few survived in altered form, and one stood. Here is what twenty-five case files seem to add up to, stated with the caveat that governs all of it: these are the genre’s own staples, examined by one reader with a browser over twelve days, not a random sample of the world’s vocabulary examined by a team with a library.

1. The popular claim is rarely false and almost always misleading. Nearly every word examined does lack a one-word English equivalent — but that is true of countless words in every language, and it is the only version of the claim that survived. The stronger claims the genre gestures at, that English cannot express the concept or that English speakers cannot have the feeling, did not hold for any word on the file. Even the strongest candidates, discussed below, are gaps in vocabulary, not gaps in experience.

2. The words are ordinary at home. For word after word — komorebi, gezellig, hüzün, ya’aburnee, teranga, tizita, gemas — the source language’s own dictionaries or scholars treat the word as an everyday item, sometimes a socially central one, and make no claim that it resists translation. The exotic framing is added on import.

3. “Unique to us” claims fared worst of all. Where the genre says a word expresses something uniquely Greek, Turkish, Finnish, Czech, or Zulu, the word’s own history usually disagrees: it is a loanword (meraki, hüzün, Indonesian geram), or its root is shared across a whole family while the special sense grew up locally (litost, ubuntu), or the feeling it names has independent names in unrelated languages (gigil, gemas, and English-language psychology’s “cute aggression”), or the story told about its origin is a folk etymology (lagom). Uniqueness was the one claim the project could actually test, and it rarely passed.

4. Where difficulty is real, it is difficulty of bundling, not of meaning. The best-founded cases of resistance to translation all have the same shape. Toska packs several senses into one word, so that a translator must choose among anguish, heartache, yearning, and boredom for each occurrence. Gezellig is, in a linguist’s careful phrase, unusually polysemous and productive. And the newer sense of abhiman is the one case the project found measured: thirty-six occurrences, two groups of translators, and — for the newer sense — a scatter of English words, most of them used only once. None of this is a concept English lacks; it is a set of English words that carve the territory differently. The linguist Valentina Apresjan’s verdict on toska generalizes: the differences are “linguistically generated,” not evidence of different souls.

5. One word fits the popular claim best: Schadenfreude. English has no native word for the feeling; seven other languages have their own; the feeling is, on the psychological evidence the project saw only at second hand, a human universal; and English solved the problem by importing the German word unchanged. It is the genre’s paradigm case and the nearest thing to an exception on the file, and it is instructive that its untranslatability took the form of a word being borrowed, not of a concept staying out of reach.

6. Fame and evidence are independent. The most repeated word in the genre, mamihlapinatapai, has the least underneath it: no headword in the one Yaghan dictionary, a lexicographer’s manuscripts reported to say something else, and a last speaker who reportedly did not recognize it. The most culturally weighty word the project examined, sankofa, barely appears in the genre at all. Dictionary adoption tracks English-side circulation, which listicles generate, not the soundness of what they say.

7. The genre recycles, and rarely cites. Its founding web listicle names no sources (its founding book, from 1988, the project could not read). Its entries introduce new words by comparison with old ones. And where the project chased a famous entry to its primary source, it sometimes found a chain of retellings ending in one anecdote. The genre’s lack of sourcing discipline is not incidental to it; it is how a word’s meaning can be repeated for thirty years without anyone checking.

8. Translators translate. In nearly every instance the project could read, a published translator meeting one of these words in running prose rendered it with an ordinary English word chosen for that sentence, and often a different one for the next (the one exception on file: two lagom instances in a single 2011 novel, left untranslated by a translator whose own thesis favors doing so). The loanword is kept where it works as a name or an already-naturalized term (a title, a constitution, business jargon), not where it is doing descriptive work. This is what the translation-theory literature has said since Jakobson; the project’s contribution is a handful of documented instances and one measured one.

9. Someone always benefits from the claim, and insiders are its best critics. A nationalist movement made the claim a century before the genre existed, and a Portuguese king centuries before that; publishers made it in annual installments; a self-marketing tourism industry made it about itself, and colonial commentators made it about others. The sharpest rebuttals came from a British journalist settled in Sweden calling lagom a “suffocating doctrine,” a scholarly study calling the gezellig claim a cliché nobody tests, a writer on Turkish culture reacting to hüzün with anger, a commenter with access to the Yaghan documentation calling mamihlapinatapai an urban legend. Read from inside, the marketed national mood is often a social obligation.

10. The genre has a type. On this small sample, what gets a word into the genre is not cultural weight but form: a single, opaque word for a feeling or a manner. A transparent multi-word phrase (sankofa), a concept English absorbed long ago (mianzi), or a technical term of poetics (oodal) travels poorly however much it matters at home. The project offers this as a hypothesis from twenty-five cases, not a rule.

Two further observations concern the word “untranslatable” itself. The popular genre borrowed it from an academic tradition — Barbara Cassin’s philosophical dictionary — whose sense is nearly the reverse of the popular one: not a word that cannot be translated but one that keeps being translated differently. Read in Cassin’s sense, toska and abhiman are close to genuine untranslatables, saudade is by the publisher’s account actually in the book, and komorebi is not one at all — nearly the opposite of the genre’s ranking. And the project never met, in twenty-five cases, an untranslatability claim that a translator could not get past with a phrase. The strongest conclusion the evidence supports is therefore modest: these words are untranslatable, allegedly, in the one sense that is true of all words, and in that sense alone.

What the project did not establish should be stated as plainly. it consulted no native speakers, ran no corpus statistics of its own, read no printed book except as a downloaded PDF, and chose its twenty-five words itself, from the genre’s own lists. Its running taxonomy of “shapes of complication” grew by accretion and wobbled in its numbering, and this essay has not leaned on it. The next section lists the blank regions more precisely.