Linguist Michael Erard, the author of Um. .. Slips, Stumbles, and Verbal Blunders, and What They Mean who we recently interviewed, graciously sent us this article, which he first wrote and published in the magazine Lingua Franca:

In a recent issue of the moderated e-mail list Linguist, Brown University anthropologist William O. Beeman addressed an odd phenomenon: Apparently, there is a different word for butterfly in every language, even though historical relationships and geographic contacts often suggest the words should be similar. Beeman called it "the butterfly problem."

"We expect to find cognates," Beeman says, "but here we don't. What seems to happen is that groups come up with a term that is particular to them." By contrast, words for cat resemble each other closely in most Indo-European languages. "Moth" is also frequently borrowed. What about "bat"? "I know," Beeman concedes. "It's like eating peanuts -- once you find out about one word, you want to go off and find out about more."

Beeman says his interest in butterflies began in an idle chat with mathematician friends. After a query to the Linguist list elicited fifty replies in two days, his involvement ballooned into a full-scale study. Beeman soon discovered he wasn't the first linguist to notice the phenomenon: In an unpublished 1982 paper, University of North Texas linguist Haj Ross rhapsodized about butterflies as such "perfect symbols of transformation" that "each language finds its own verbal beauty to celebrate the stunning salience of the butterfly's being." (Ross credited linguistic anthropologist Emmon Bach with having first observed the phenomenon.)

In his post to Linguist, Beeman provided a list of words for butterfly in about ninety-seven languages, including Gujarati (popti), Lao (maingkabula), and Masai (osampurumpuri). But even if the words don't sound alike, are they as different as they seem? Beeman noted that a large number have one thing in common: They are related to "cultural metaphors." For example, the folkloric belief among Russians that butterflies are witches seems to be responsible for their butterfly word, babochka, or "little grandmother." The English "butterfly" and the German Schmetterling derive from the insect's supposed attraction to cream.

More important, many butterfly words share features such as the repetition of syllables (as in Malay ramarama, Senegalese lupe lupe, and Welsh pili pala). These sound patterns, Beeman provocatively argued, echo the symmetrical structure and movement of butterfly wings. Many of the words also repeat their consonants. "In... the many cases of reiterated b's, p's, l's and f's (in widely separated language families)," Beeman wrote, "one can almost hear the gentle rustle of butterfly wings and see their repetitive motion."

Until a few years ago, Beeman's appeal to sound symbolism -- the idea that certain sounds are naturally connected to certain meanings -- would have been unfashionable, perhaps even derided. Ever since Plato's Cratylus, people have wondered whether the form of a word might be motivated by its content. But scientific debate on the question all but ended in 1916, with the posthumous publication of Ferdinand de Saussure's Course in General Linguistics. Saussure declared that linguistic signs were conventional and arbitrary. "There is no internal connexion, for example, between the idea 'sister' and the French sequence of sounds s-รถ-r [soeur] which acts as its signal," he wrote. According to Saussure, if a word such as "cuckoo" mimics a cuckoo's call, it's no more than coincidence. The few genuine onomatopoetic words that do exist -- such as glas, the French word for knell -- Saussure dismissed as marginal cases. By mid-century, Saussure's theory had become linguistic dogma.

"Sound symbolism is hard to study," says Leanne Hinton, professor of linguistics at Berkeley. "It's easier to study things that have definite rules to them, something that's more definitely patterned or governed in some predictable way." Hinton believed that Saussurean dogma needed to be challenged, and so, with fellow linguists Johanna Nichols and John Ohala, she co-edited Sound Symbolism (Cambridge) in 1994. Unfortunately, Hinton says, "a lot of people have cockamamie ideas about sound symbolism, so it gets a bad reputation."

"In the back of every linguist's mind, there's always been some question about sound symbolism," says Beeman. He points to a 1929 experiment by Edward Sapir in which Sapir's subjects were asked to match nonsense words with small and large versions of the same object. The subjects tended to match words with a high vowel (such as ee) to the small object and those with a low vowel (such as the o in "cot") to the larger object. British linguist J.R. Firth later called these links between sound and meaning "phonesthemes." One English phonestheme, the consonant cluster /gl/, is generally taken to mean "light" or "shining," as evidenced in words such as "glass," "glisten," "glow," "glare," "glimmer," "glimpse," "glitter," and "gloss." (The English poet Gerard Manley Hopkins loved to compile lists like this in his journals.)

"These things are fun, but very few people want to put in the work to go about it in a rigorous way," says Terence Hays, an ethnobiologist at Rhode Island College who has tackled what one might call the "frog problem": The sound /r/ or /g/ recurs in over two hundred words for frogs and toads in New Guinea. (John Beatty of Brooklyn College has also playfully suggested what might be called the Godzilla problem: The syllable /ra/ seems to symbolize large size and monstrousness when used to name creatures in Japanese movies.)

Unfortunately, to establish the sound symbolism of butterfly terms to a linguist's satisfaction would require a Herculean effort, Hays says. After all, the repetition of syllables and other elements is a common feature of many languages. "Only if you show me a language that never reduplicates sounds in nouns but does it with butterflies, then I'll sit up and listen," he says. After he read Beeman's post, Hays consulted his database of the Mailuan language family in New Guinea. Among the five words for butterfly in the Mailuan languages, he found two cognates.

But Hays doesn't discourage the search for sound symbolism altogether. He cites work by ethnobiologists such as Brent Berlin, who has found that in Huambisa, a language spoken in northern Peru, some vowels are more often associated with elongated animals than with flying animals. "So there are some interesting things going on with sound symbolism," Hays says.

"Actually, if you go to the field," he adds, "you notice fireflies most of all. Here are spots of light, appearing, disappearing, all around you. They're much more salient than butterflies." But there's not likely to be a firefly problem: In the languages Hays studies, the word for firefly is almost always the same as that for star. Explains Hays: "Just because something really strikes you perceptually doesn't mean you'll treat it in a special linguistic way."