This article by William O. Beeman (Department of Anthropology, Brown University) fuels my worst suspicions about anthropologists’ shaky grasp of the concepts of linguistics; the first two sentences indicate a distilled confusion such as is rarely achieved: “One of the bedrock principles of linguistic analysis since the nineteenth century has been the principle of the regularity of cognate borrowing. It forms the basis of the ‘comparative method’ not only in linguistics, but in all of social science.” A little later comes a paragraph with a more expansive form of the confusion:

However, there is a limited, but powerful countervailing tendency in language behavior—words that absolutely resist borrowing even from their closest linguistic relatives. These words seem to be coined anew by each population group. Because we expect cognate borrowing as a norm, it is surprising when we encounter these fascinating examples. It makes us wonder about the cultural processes that govern the development of communication systems, and the functional differences between segments of vocabulary.

This leads into his main point, the often-observed fact that most languages, even closely related ones, have entirely different words for ‘butterfly.’ He then lists all those he has collected, which is what makes the page worth linking to. Ignore the miasma and enjoy the variety of (often semi-onomatopoeic) words, to which I will add cipelebesha (Bemba), balanbaalis (Somali), ekiwojjolo (Luganda), ihe n’efe-efe or uru baba (Igbo), lolo (Malagasy), vannatti pucchi (Tamil), titernig (West Armenian), peperuga (Kalderash Romanes [“Gypsy”]), fepule or minni or tirtirk (Kurmanji Kurdish), metelik (Ukrainian), palomma (Neapolitan Italian), kubelek (Tatar), kapalak (Uzbek), göpölök (Kyrgyz) (notice the interesting variations among these three and Turkish kelebek), khovagan (Tuvinian), pepela (Georgian), polla (Chechen). A few corrections: according to my Basque dictionary, the first word should be tximeleta (rather than txipilota), the Hausa word is malam-bude-littafi (first and last a‘s long), and the Latvian word is taurinš (not tauriųš); furthermore, there is no such language as “Senegalese” (anybody know what language lupe lupe is from?). And a couple of alternative Zulu terms: ijubajubane and itwabitwabi. What fun!

For further amusement, I reproduce here the last thing on the page, a footnote quoting a mind-bogglingly lunatic theory of “universal word derivation” that Prof. Beeman apparently takes seriously:

2. However, Isaac Mozeson, author of The Word, a treatise on common word origins, contributed this commentary based on his own theories of universal word derivation:

I had in my “PYRALIDID” entry (appendix A) the PR Greek, the PPL Latin, the Malay PPL and the Nahuatl PPL terms for butterfly. All should be influenced by Hebrew PaR PaR (butterfly) and the PR root of PiRPooR (to twitch). I am grateful for the Tagalog paruparo, and would like to credit the contributor. As for the Paiwan/Taiwan term, two phonemes are at work. One, kali, could be like Hebrew KAL (light, swift), and the other is a duplicated dungudungul, which appears to be a nasalized DIGDAIG (Hebrew for the tickle-like wavering motion of DAG (fish) and DeGel (flag). Needless to say, TICKle itself is a form of this Daled-Gimel root from Edensprach. Lastly, the Autronesian KUPO root could be a form of Ayin-Peh, KHuPh (to fly—see “AVIATE” in THE WORD, p. 26).

(Via wood s lot.)


  1. Right you are about the addictive properties of searching the text in Amazon — Basque guru Larry Trask has some comments about tximeleta.
    It is fun reading lists like this. And it would be great fun to try to build a new one collaboratively with all your readership, don’t you think? I imagine the collective collection of dictionaries on the shelves of Languagehat readers is formidable indeed!

  2. Great idea! All contributions welcome; put ’em in the comments.
    And a great link—the Amazon effect strikes again!

  3. …furthermore, there is no such language as “Senegalese” (anybody know what language lupe lupe is from?)
    Well, the principle language of Senegal is of course Wolof. (I don’t know why I say “of course”, it’s only by coincidence that I know this, although it would not be difficult to find out). The only online dictionary I could find doesn’t include an entry for butterfly, but I did turn up some vocabulary resources for students of the language, which give the Wolof for butterfly as lëpp-lëpp bi. Looks pretty similar to me, although whether it’s the same word distorted by orthography and/or dialect, or a cognate/borrowing in a seperate language, I can’t tell.

  4. Yeah, I know Wolof is the main language, and I even have a (French) dictionary, but I couldn’t lay my hands on it, and “lupe lupe” didn’t sound like Wolof. Lëpp-lëpp bi is more like it—thanks for the contribution!

  5. Some anthropologists have a pretty decent grasp of linguistics …

  6. I think that it is worth noting that Beeman is not himself a historical linguist, and that some of the work he cites, such as that of Joseph Greenberg, is well respected – even if controversial. This particular post of his seems to be more of a hobby project than anything to do with his main research (on discorse in Iran). Beeman’s writings on Afghanistan have been invaluable during the past few years.
    I know there is a running theme on these pages attacking any attempt at historical linguistics, but there are some interesting (even if often over-interpreted) relationships between historical linguistic work and research in genetics. While I agree that attempts at constructing an “ur-language” are misplaced, there is much to be gained using linguistic data to supplement other (cultural and genetic) data for the study of migrations and other related questions.

  7. Kerim: I’m being somewhat tongue-in-cheek about anthropologists; of course I’m aware there are ones who know their linguistics (and indeed American linguistics was founded by anthros), but my anthro teacher in college was a complete ignoramus, so I take pleasure in poking them once in a while. No offense meant. And I certainly wouldn’t assume anything about Beeman’s main work based on this sideline; if you say his Afghan stuff is valuable, I’ll definitely check it out.
    I would never attack “any attempt at historical linguistics” (that having been my field), I just think many people have no idea how limited its range is and try to use its concepts (or concepts that they imagine apply) to reach farther back than is possible or to back up theories about (say) cultural development or migrations that linguistics can’t really help with.
    Pat: Why don’t you put up a page where we can keep ‘butterfly’ words? Then we won’t have to visit the one with all the silliness (and we can control the entries — I’m dubious about some of his).

  8. Sorry if I overreacted – I was feeling a little defensive this morning …

  9. Here are a few more butterfly words (all from the net – I’m afraid I don’t have as much in the way of exotic dictionaries as I’d like).
    First off, another existing list. It’s shorter than Beeman’s, but it does include some languages missing from his, and disagrees on some common to both, so it might be useful. A much shorter list here provides a little etymological information.
    In a post on the Constructed Languages list, Dirk Elzinga (mentioned in this Languagehat entry) provides butterfly words in several Uto-Aztecan languages (and Navajo):
    Shoshoni: waayapputunkih
    Kaibab Paiute: aïcïvïtsi (ï = barred-i, c = long-s)
    Luiseño: avéllaka
    Hopi: povolhoya (monarch: hookona)
    Navajo: k’aalógii
    (Note that this not what Beeman gives for Navajo.) See this followup post for a morphological breakdown of some of these.
    (Finally, on one of the Wolof sites I linked to above, I see that ë is a schwa, and pp is aspirated. It’s noted thaat some of the other doubled stops are followed by a short epenthetic vowel when they occur finally. This being the case, I can imagine an orthography that would render lëpp-lëpp as lupe lupe.)

  10. A butterfly-like word in the medical-supply world is “tape”. One international 8-language box of tape was labelled “tape/ruban/cinta/sparadrap/pfleister”. There were only 8 languages and Japanese was one of them; pfleister and sparadrap, as I remember, each had one cognate.

  11. Thanks! We’re accumulating quite a swarm of butterfly words here.

  12. Back when I was a budding polyglot, I noticed the similarity of the Hebrew/Arabic, Romance and Latin words for butterfly mentioned in the quote above. An “intermediary form” that the author has missed: Welsh pili-pala (if I remember correctly).

  13. Here’s a list I compiled for my book “A Desert Bestiary” (Johnson Books, 1997). It’s got a few languages not represented in the other lists.
    Acoma, buh’rai
    Arabic, farasha
    Dutch, vlinder
    German, Schmetterling
    Greek, petalou’da
    Hebrew, parpar
    Hungarian, lepke
    Italian, farfalla
    Japanese, chou
    Kyaka Enga, maemae
    Latin, papilio
    Lushootseed, yubec
    Maltese, farfetta
    Nahuatl, papalotl
    Ndumba, kaapura’rora
    Osage, dsithato’ga
    Polish, motyl
    Portuguese, borboleta
    Romanian, fluture
    Russian, babochka
    Spanish, mariposa
    Swahili, kipepeo
    Tohono O’odham, hohokimal
    Turkish, kelebek
    Yaqui, vaisevo’i

  14. nutrition supplements says:

    Right on my man!

  15. Once again a piece of spam I can’t bring myself to delete, only to erase the URL; how can I deprive myself of “Right on my man!”?

  16. In Chinese – “hudie”
    In Russian except of “babochka” also can be “motyliok” =D

  17. Charles Perry says:

    The people of Harer in SE Ethiopia have a charming word for butterfly. They speak a Semitic language related to the Amharic of the highlands, but unlike the Amharas, they are Muslims and use the Arabic alphabet. Their word for butterfly is amhara kitab “Amharic book,” because to them the Amharic alphabet resembles the markings on a butterfly’s wing.

  18. That is nice—thanks!

  19. Late to the party but I can add a term common to the Jívaro languages – wámpishuk
    Also I note the Maori word given is pulelehua, which seems pretty unlikely given that Maori has no /l/. Sure enough, my dictionary gives ‘purerehua’ (with the first /u/ long).

  20. michael farris says:

    Mvskoke (Creek): tvffolopv, tvffolope (nb. v = [a] and e = [i], both short)
    grasshopper is tvffo so -lopv / -lope looks like it should mean something, (e)lope is “liver” but that doesn’t look promising. There’s a verb lopicetv (nb. i = [e(y)]) which looks like lop + causitive but it’s intransitive and means ‘to be nice, kind, well-behaved’.

  21. Kobelek in Kazakh
    Zimerfoigele in Yiddish
    Motyl in Yiddish

  22. Tatar should be күбәләк (approximately kybäläk), as is Bashkir.

Speak Your Mind