In the 10th century, an Arab geographer described the Caucasus region as a "mountain of tongues." The nickname has stuck to this day, likely because of how well it captures two of the area's main features: its dramatic cliffs and its array of languages.
But new and controversial research by a U.S. linguist suggests that the "mountains" may have more to do with the "tongues" than anyone has guessed.
In a study published last month in the journal "Plos One," Caleb Everett, an anthropological linguist at the University of Miami, claims that a special class of sounds occurring in almost all of the languages of the Caucasus may be due to "the direct influence" of the region's high altitude.
Everett's conclusion applies beyond the Caucasus as well. He offers apparent proof that the rare sounds, known as "ejectives," are far more likely to occur in regions of high elevation worldwide.
That argument -- and the physical and biological factors that Everett says may explain the correlation -- are generating plenty of buzz, and a range of reactions, among linguists.
"The assumption made by linguists is that geography can impact language, but in sort of superficial ways: If you live in the Amazon, you're not going to have words for 'snow-capped mountains,'" Everett said. "But linguists have traditionally been skeptical that geography can affect the structure of language, things like phonology -- the sound system -- of a language. I think this is really good evidence [of that]."
Ejectives, which occur in approximately 20 percent of the world's languages, are consonants produced when air is compressed in the mouth and pharyngeal cavity and then released in a burst. Unlike most sounds, they are not produced using air from the lungs.
For Caucasian languages, many of which are known for their rich inventories of sounds, ejectives are a characteristic feature.
Linguists and befuddled visitors alike have described these sounds as hisses, spits, and even mini-explosions in the mouth.
In his research, Everett considered the distribution of 567 world languages in relation to six high-elevation "zones." Those zones were defined as major regions greater than 1,500 meters in altitude, plus the surrounding 200 kilometers. They included the Caucasus range and the Javakheti plateau, the Rocky Mountain region in North America, the southern African plateau, and others.
Of Everett's sample of languages, only a small portion contained ejectives. However, he found that nearly two-thirds of those that did were located in the high-altitude zones.
The only zone in which ejectives were absent was the Tibetan plateau. "In fact it strikes us as remarkable that only one region presents an exception," Everett writes.
Languages occurring in each zone, from the Caucasus to the Andes, were also from multiple, often unrelated language families. The Caucasian language Abkhaz contains ejectives, but so do several dialects of Armenian, a language from the entirely distinct Indo-European family. Such evidence, Everett says, goes far in arguing that geography, and not genetic relations, is behind the trend.
Explaining why such a correlation might exist is a more challenging proposition. Everett offers his best guesses, albeit tentatively.
"Hypothetically, these sounds should be easier to make at high altitudes because they require the compression of ambient air," Everett says. "Since air pressure is lower at higher altitudes, the sounds should be easier to make. That was my first hypothesis."
He also suggests that use of the sounds may be a biological adaption to the dryness of high-altitude locations. "Because you don't have to expel air from the lungs to produce ejective sounds, they should theoretically reduce the amount of water vapor lost during speech," he says.
Bernard Comrie, the director of the linguistics department at the Max Planck Institute for Evolutionary Anthropology in Leipzig, Germany, says Everett may be on to something, at least in showing that a real correlation exists between altitude and ejectives.
Like others, however, he is more skeptical when it comes to Everett's suggested explanations, but says they are not "unreasonable."
Comrie also says that the controversy may give linguistic science a shake-up it needs.
"I think [this research] is important in that it really suggests a way in which one could seriously investigate a kind of claim that has largely been neglected by linguists," Comrie says.
"This is a direction they are going to have to consider. So far, linguists are rather negative toward such generalizations, but more because their ideological background leads them to be negative, rather than because there is strong empirical evidence against a particular claim."