Anna Bitong at BBC Travel writes about the Yaghan of Tierra del Fuego; there’s stuff about their history and current status, but what concerns us is this passage:

That inspiration can be seen in a word that has garnered rapturous admirers and inspired many flights of the imagination. Mamihlapinatapai comes from the near-extinct Yaghan language. According to Vargas’ own interpretation, “It is the moment of meditation around the pusakí [fire in Yaghan] when the grandparents transmit their stories to the young people. It’s that instant in which everyone is quiet.”

But since the 19th Century, the word has held a different meaning – one to which people all over the world relate.

Magellan’s discovery of a ‘land of fire’ prompted more long-distance voyages to the region. In the 1860s, British missionary and linguist Thomas Bridges set up a mission in Ushuaia. He spent the next 20 years living among the Yaghans and compiled around 32,000 of their words and inflections in a Yaghan-English dictionary. The English translation of mamihlapinatapai, which differs from Vargas’ interpretation, debuted in an essay by Bridges: “To look at each other, hoping that either will offer to do something, which both parties much desire done but are unwilling to do.”

“Bridges’ dictionary records ihlapi, ‘awkward’, from which one could derive ihlapi-na, ‘to feel awkward’; ihlapi-na-ta, ‘to cause to feel awkward’; and mam-ihlapi-na-ta-pai, something like ‘to make each other feel awkward’ in a literal translation,” said Yoram Meroz, one of the few linguists who have studied the Yaghan language. “[Bridges’ translation] is more of an idiomatic or free translation.”

However, the word does not appear in Bridges’ dictionary – perhaps because it was seldom used, or possibly because he planned to include the word in the third edition of the dictionary, which he was working on before he died in 1898.

“It could be that he heard the word once or twice in that particular context, and that’s how he wrote it, because he wasn’t aware of its more general meaning. Or that it was only used in this more specific meaning that he quotes,” Meroz explained. “Bridges knew Yahgan better than any European before or since. However, he was sometimes prone to exoticising the language, and to being very verbose in his translations.”

Accurate or not, Bridges’ translation of mamihlapinatapai sparked a widespread fascination with the word that continues to this day. “The word got popularised by Bridges and was quoted and re-quoted in English-language materials,” Meroz said.

In many interpretations, the word came to signify a look between would-be lovers. On the internet, its definition is worded slightly differently as ‘a look shared by two people, each wishing that the other would initiate something that they both desire but which neither wants to begin’. Films, music, art, literature and poetry have all conjured its seemingly implicit romance and marvelled at its supposed ability to concisely capture a complex human interaction. The 1994 Guinness Book of World Records even listed mamihlapinatapai as the world’s most succinct word.

“The meaning is quite beautiful,” says a girl in the 2011 crowdsourced documentary Life in a Day, which portrays a single day on Earth. “It can be perhaps two tribal leaders both wanting to make peace, but neither wanting to be the one to begin it. Or it could be two people at a party wanting to approach each other, and neither are quite brave enough to make the first move.”

But what mamihlapinatapai actually meant to the Yaghans will likely remain a mystery.

This made me happy for a couple of reasons. Yoram Meroz has been a correspondent of mine for a decade, and it’s always a pleasure to see a Friend of LH quoted in the press. And LH covered mamihlapinatapai back in 2006, bringing out of the woodwork another linguist who has studied Yaghan, Jess Tauber; I’m reproducing his comments here, to provide as comprehensive a picture of this beloved word as is currently possible:

I’ve been studying Yahgan for nearly 10 years, collecting, analyzing, reediting and consolidating known (and unknown) materials on the language all with an eye towards its revitalization. Two years ago I accidentally discovered moldering at the Library of Congress one of the ‘lost’ grammar manuscripts of Yahgan. Its an early one, but it is much richer materially than anything else. The dictionary published in 1933 in Austria is a disaster editorially, and I’ve been working to put together a new one based on all sources, including corrections by Thomas Bridges the editors felt unnecessary to include (their mistake…). The original manuscripts showed stresses not carried over to the published version.

One of the great sadnesses of Yahgan history is that the bulk of materials produced by Rev. Bridges in the late 19th century were lost (don’t even get me started here…). Surviving works are largely earlier, less complete ones.

The Omora ethnobiological conservancy (which run a park outside Ukika and Puerto Williams on Isla Navarino, where the Yahgan village is) has produced a CD of bird-related myths with Yahgan narration by two of the then three last speakers. About a half-hour of spoken Yahgan, accompanied by translations into English, or Spanish (two versions).

My examination of these stories indicates that though the currently surviving dialect has some phonological, lexical, and grammatical differences from the one Bridges wrote about, it isn’t radically different. I doubt speakers of any of the dialects would have had too much trouble understanding each other.

There has been some interest expressed in the stress of words- based on what I’ve seen and heard, Yahgan appears to have been a mora-counting language, rather than one with fixed word stress. That is, there are stresses placed regularly and predictably along long stretches of material. Some roots attract stress, others repel it, and still others are ambiguous. They all work together to create the final product stress-wise.

Anyone interested in the language should look at the Yahoo Group Waata Chis (Old News). I’m slowly creating updated Yahgan info at Wikipedia, and there should eventually be permanent linguistically oriented posted at Dartmouth. This week I’ll be scanning more material for the CDROM I’ll be making of all extant documentation.
Hala yella (gotta go…)
* * *
To answer some of the comments and questions above- Yahgan is NOT polysynthetic, though some of the words can be very, very long. Though ‘polysynthesis’ was used a century ago to mean any long word stem composed of many different morphemes (some of which go beyond the merely derivational and inflectional to include pragmatically oriented ones), in recent years linguists use the term to refer to only the inflectional criteria (both subject and object pronouns on the verb, incorporation of generic nominals), often with ‘normal’ free lexemes relegated to adjunct status (not necessary to interpret the sentence)- the inflected word is equivalent to the sentence, or even more than one.

Yahgan, on the other hand, is an agglutinating, serializing language- that is, several verb roots (plus some derivational materials) are strung together to make a compound, but only the subject pronouns are bound. Thus only intransitives could be construed as full sentences. Agglutinating means that the morphemes are just added one to the other, each bringing their meaning to mixture. Yahgan does though have an interesting specialization of the serial verb stem — one where there is a prefix delineating the instrumental means before the main root, and/or a suffix following the main root that describes the path of motion or static position of the action.

So for instance one could say hat-ak-u:-vnggu:ta-kvn-ude: ‘I broke it in the boat or floating by hitting it with a blow’: hat– is first person subject (t epenthetic), ak– means ‘by striking’, u:– causative/let, vnggu:ta– ‘break/split’, kvn(a) ‘floating/in boat’, -ude: past tense. Yahgan has dozens of elements to choose from in both prefix instrument/bodypart and suffix position/pathway terms. All are relatively generic. Specifications can be made by adding in other elements (for instance ak-isiu: by hitting in a very narrow or controlled way, such as superficially.

A very large percentage of the dictionary consists of stems made up of just such combinations with normal roots- the dictionary is often touted as having ‘32000’ words (actually it has 23000- the final lost one had the larger number), but Bridges could easily have created a document with millions (he notes he’s suppressed quite a few).

The stem mamihlapinatapai (or ma(m)-ihlvpi-:n-at-a:pai = passive/reflexive-‘be at a loss which way to go’-state-achievement-dual is neither serializing nor contains the complex described above. It is a simple root ihlvpi plus derivational elements. So much for concise. As for ‘looking at each other, etc.’ which author after author cites or elaborates upon without even consulting the actual dictionary texts or grammars (the form was listed in the PREFACE- is that how deep people actually are willing to go? Wow.), Bridges tended to use illustrative examples rather than simple definitions, in order that the reader should get a better feel for the meaning. Apparently latecomers to the table haven’t caught this particular trick, and so we end up getting a linguistic urban legend, self-perpetuating and little more than party chit-chat. Who knows, maybe this particular tidbit helped someone get lucky, somewhere…..

So there you have it. I thank Trevor Joyce for the BBC link, and reading the old thread reminds me how much I miss Jimmy Ho’s invariably lively and informative comments. Come back, Jimmy!


  1. David Eddyshaw says:

    Good on Anna Bitong for at least consulting an expert and casting some doubt on the ever-popular Meaning of Tingo drivel.
    In some ways this is the saddest bit:

    According to Vargas’ own interpretation, “It is the moment of meditation around the pusakí [fire in Yaghan] when the grandparents transmit their stories to the young people. It’s that instant in which everyone is quiet.”

    As the language dies, and the culture dwindles to mere folklore, even well-intentioned descendants of the speakers start with the exoticism and the nonsense.

    Still, personally I’m happy that Yahgan can make a word that means “two have got into a state of not knowing where to go.” It’s Enough.

  2. David Marjanović says:

    Some roots attract stress, others repel it, and still others are ambiguous. They all work together to create the final product stress-wise.

    So, is it possible to tell which roots do what by counting their morae? Because if not, that sounds like Russian or PIE and not mora-counting at all.

  3. Yeah, I wish he had come back to answer my question about stress.

  4. Jess Tauber says:

    I am here.

  5. Welcome back!

  6. Yoram Meroz says:

    I realize I had made an error, which made it to Bitong’s article. -ta is not a causative, but an aspectual marker. iłapinata is not ‘to cause to feel awkward’ but rather something like ‘to get awkward’. I will write her and see if she can correct it.

    Stress in Yahgan is complex, and I have not fully analyzed it. It depends on phonological as well as morphological factors, and in some cases may be lexically dependent. Primary stress often falls on the antepenultimate syllable of a long word. It often falls on the penult of the verb stem.

    In the case at hand, Bridges records iłápi, but the stress moves in iłapína and iłapínata. I would guess the stress in the full word is mamiłapinátapai, perhaps with a secondary stress on the -ła-. I could be wrong.

  7. Yoram Meroz says:

    Bridges earliest manuscript dictionary (1865) translates iłapinata as “To be unable to to help thro’ the parties being near of kin.”. In his final but partial dictionary manuscript (1879) the gloss to both iłapina and iłapinata is “To feel awkward or unable to help because of some affinity with both parties.” As in many other cases, Bridges’ glosses record some particular uses of a word, but not necessarily its complete range of meaning.

  8. marie-lucie says:

    Hi Jess. nice to see you here. Can you tell us more?

  9. Jess Tauber says:

    Hi folks. -ata NEVER takes stress, based on stress-marked examples in the dictionary manuscripts. It will (nearly?) always appear before the -n- in -Vna- (the vowel shape is somewhat dependent on preceding materials). And -a:pai seems to take stress on the first vowel, again from Bridges stress-marking. He doesn’t make any indication as to primary versus secondary stresses.

  10. Thanks!

  11. marie-lucie says:

    Thank you Jess!

Speak Your Mind