The following originally appeared in Science, April 17, 2009, vol. 324. Copyright 2009 Michael Erard.
After a long day in the field, deep in the mountains of southwestern China near the border with Vietnam, retired environmental health professor Gary Shook was surprised to meet another American, Jamin Pelkey, staying in the same government guesthouse. The two exchanged pleasantries.
“I’m collecting tiger beetles,” explained Shook, who had found four new species in the region. “What about you?”
“I’m collecting new species of languages,” replied Pelkey, then a graduate student at La Trobe University in Australia doing fieldwork for his dissertation. In 2006, Pelkey and his wife were gathering linguistic data in 41 villages in a 100,000-square-kilometer area of Yunnan Province. Over the course of a year, they drove 15,000 kilometers across rugged terrain in a Jeep. At the end, Pelkey had identified 24 languages associated with the Phula ethnic group, 18 of which had never been defined scientifically before. Until Pelkey’s work, these languages had been invisible because their speakers were lumped together under a single ethnic label, the Yi, which is officially considered to have one language.
At a time when hundreds of languages are disappearing because children don’t learn them and adults don’t speak them, it may seem surprising that many existing languages have never even been named (though they are not “new,” especially not to the people who speak them). Yet there are potentially hundreds of undiscovered languages in China, Burma, the Amazon, and elsewhere, linguists say.
Pelkey’s 24 are listed for the first time this month, in the latest edition of Ethnologue: Languages of the World, an authoritative, worldwide gazetteer of languages maintained and published by SIL International, a non-profit based in Dallas, Texas. This newest edition of Ethnologuelists 6909 living languages from 156 countries, including 83 “new” languages from 19 countries.
Pelkey’s new entries are the most from any single country. China is “one of the last places on earth where there are large numbers of unreported and undescribed languages,” says linguist David Bradley of La Trobe, who also works in Yunnan. The reasons have to do with geography, history, and politics. Bradley speculates that Yunnan alone may have over 150 languages, and Western and Chinese linguists are now surveying the region more thoroughly. “In the last few years, there’s been very much a heightened interest [by Chinese] in their diversity and a desire to study and work on language maintenance,” says linguist Arienne Dwyer of the University of Kansas, Lawrence. Yet this interest in linguistic diversity sometimes conflicts with the notion of a multiethnic but unified Chinese state. “The reason that language is particularly sensitive is that, in southwestern China, language was the principal way of categorizing people,” says Thomas Mullaney, a historian at Stanford University in Palo Alto, California.
How can there be so many undiscovered languages in one region? One reason is the remoteness of villages. “Yunnan has so many mountains, and transportation was so limited before the Communists started building roads, and ethnic groups have been proliferating for so many centuries there,” Pelkey says. “The astonishing thing would be to walk into the situation and find only a few dozen languages.”
Yunnan is most frequently identified by the colorfully embroidered clothes and quilted hats of the non-Han ethnic groups who have called the mountains and lowlands home for thousands of years. Because their languages were rarely written down, linguistic change went unchecked. Local and imperial governments had little interest in languages, leaving them uncounted.
Centuries of isolation widened the gap between varieties descended from the same parent tongue. Today, the 500 speakers of Alo Phola can’t understand speakers of a sister language spoken less than 8 kilometers away, says Pelkey. One of Pelkey’s main criteria for judging language separateness is “mutual intelligibility,” or how well speakers of different varieties are able to understand each other. Among speakers of the 24 Phula languages, mutual intelligibility is so low that if they ever got together, they would have to communicate in a regional variety of Mandarin, Pelkey says.
Many Chinese languages are being described only now in part because a tradition of lumping ethnic groups together has masked the extent of the diversity. Chinese social scientists of the 1930s and ’40s streamlined the number of ethnic minority groups, which were based mainly on language. “The logic was, ‘It does no one any good to have an ethnic group of 100 people,'” says Mullaney.
In the 1950s, about 50 surveyors spent 6 months in Yunnan and divided a population of 2 million into 20 official groups, even though 212 ethnic group names had been discovered. In 1991, China permanently froze the number of recognized ethnic nationalities, known as minzu, at 56: the majority Han plus 55 minority groups, 25 in Yunnan. Until the 1980s, it was forbidden to suggest that China had more than 55 languages, Bradley wrote in 2005 in the International Journal of the Sociology of Language. “Any additional linguistic entities had to be classified as ‘fangyan.'” Although the word fangyan is often translated as “dialect,” it refers more specifically to “a language spoken in a specific area,” or a “topolect,” in contrast to yuyan, or an autonomous language.
This legacy has led to some disagreement between Chinese and Western linguists over what counts as a language. “We are very strict, while foreign researchers are very loose,” says linguist Sun Hongkai of the Chinese Academy of Social Sciences in Beijing. Sun began doing fieldwork in 1953 in Sichuan Province and Yunnan and has helped identify 19 languages. He promotes a method different from that of Western linguists, saying that the boundary between a language and a dialect should be determined by comparing grammatical patterns, vocabulary, and sound rules. If they are similar, the varieties are dialects of the same language.
Other Chinese scholars add that varieties that come from the same parent language and have the same writing system must be fangyan, not yuyan, and reject mutual intelligibility as unscientific. Using such criteria, the roughly 230 European languages would be fangyan of a handful of languages. Chinese linguists “are still constrained by political realities as well as the traditional macro- categories imposed by the Han Chinese majority on their minorities,” says Bradley.
For example, in a 2008 report to UNESCO of endangered languages in China, Sun listed a single language for the Yi minzu. Although some of the Phula languages Pelkey described are endangered, they cannot be identified as such because the Yi officially have only one language. So it may be harder to target those languages with revitalization resources.
All the same, since the 1980s, Chinese linguistic diversity has become an open secret, and Chinese researchers have become freer to identify new languages as yuyan, says Bradley. In 1992, Sun helped establish an academy project on new languages, for example. Overall, Chinese linguists have identified a total of 134 languages, and the 80 identified in the last 25 years are called yuyan, not fangyan.
The Chinese have also opened their doors to foreign researchers such as Pelkey, who studied under Bradley. In 2005, Pelkey joined SIL International, the world’s biggest player in describing minority languages. SIL has a Christian goal: It describes and analyzes languages to aid in Bible translation and literacy projects. In the past, a Christian organization might have had difficulties in China, but such survey work has been encouraged recently because it helps to provide education in mother tongues and coordinate language revitalization, Pelkey says.
Pelkey did his research using his affiliation with SIL, which has been registered as a nongovernmental organization in Yunnan since 2004. Pelkey stayed 3 to 5 days at a time in Yunnan villages, interviewing 10 or so local people. Using a list of 1200 words, he would say a word in southwestern Mandarin Chinese and show a picture of the object, then record people saying the word in their language. On breaks, he recorded people telling stories and played recordings of people from other villages in order to determine mutual intelligibility.
Originally, Pelkey had hypothesized that the languages associated with the Phula ethnonym were related to each other. They are all tonal languages, have a default subject-object-verb word order and very simple word structure. However, he found that although they have the same ancient ancestry, they’re not siblings or even distant cousins. Using a distance matrix, a tool from evolutionary biology that is new to historical linguists, Pelkey determined that Azha (spoken by 53,000 people) and Pholo (spoken by 30,000) don’t share the recent ancestry of the other 22 Phula languages. Thus these two would not be fangyan even by Chinese criteria. The speakers of all these languages have been subsumed under the Yi minzu.
For some communities, linguistic description and discovery is welcomed, but others are uncomfortable with losing traditional affiliations, linguists say. In Sichuan, Bradley says, speakers of 20 to 25 languages in the Tibetan minzu strongly reject any claim that they’re anything but Tibetan and so don’t want distinct languages to be identified as such.
The 24 new Phula languages included in Ethnologue have now acquired something of an official status internationally because they have been assigned identification codes by the International Organization for Standardization (ISO). Such language codes are used in software, digital archives, and library collections and are an official recognition that a speech variety meets ISO’s definition of a “language.” It remains to be seen how the Chinese government will react to this recognition. Says Mullaney, “When people start to talk about there being new languages out there, it really starts to pull the thread out of this idea that there are a set number of minzu.”
Pelkey hopes a discussion will ensue. “You start out with assumed categories, then you find a lot of diversity inside them, and then you use a scientific approach to modify your understanding,” he says. “The two don’t have to be in dissonance, and they don’t have to be in consonance, either.” Otherwise, defining a language invites so much controversy, discovering species of beetles looks like a walk in the park.
With reporting by Chen Xi in Beijing.