Tongue Twisters

Weekly Article
Luca Lorenzelli /
Oct. 31, 2019

This article originally appeared in Future Tense, a collaboration among Arizona State University, New America, and Slate.

In 1882, linguists were electrified by the publication of a lost language—one supposedly spoken by the extinct Taensa people of Louisiana—because it bore hardly any relation to the languages of other Native American peoples of that region. The Taensa grammar was so unusual they were convinced it could teach them something momentous either about the region’s history, or the way that languages evolve, or both.

The reconstruction of the Taensa grammar was the painstaking work of a French teenager named Jean Parisot. He claimed to have stumbled upon a manuscript in his grandfather’s library in the Vosges region of France, and to have realized that it was notes made by unknown explorers who had passed through Taensa territory while the now-extinct people inhabited it.

Parisot’s glory was short-lived. The linguists soon became suspicious about his Taensa grammar: The verbs seemed too regular, the relative clause structure too European. And there were anachronisms in the stories and songs he claimed to have transcribed from the manuscript: They contained references to sugar cane, for example, which had only been introduced to Louisiana, by Jesuits, around the time the Taensa disappeared.

Within a couple of years, Parisot’s grammar had been outed as a hoax and he had retreated to a monastery to take up the religious life. His story might have been forgotten, except that in 2017 Oliver Mayeux, a young linguist at Cambridge University, saw a passing reference to the Taensa hoax and became intrigued. He investigated and, without condoning Parisot’s lie, came to recognize a kindred spirit: an early constructor of languages, or “conlanger,” who took delight in linguistic play. “He must have enjoyed it to produce such a detailed grammar,” Mayeux says. “The only problem is, it’s not real.”

Mayeux presented Parisot’s story to a conference of the Language Creation Society in Cambridge in June. The society’s vice president, Dutch linguist and translator Jan van Steenbergen, told the same conference that language invention was experiencing a golden age. Though “conlangs” are not a new phenomenon—Esperanto and Volapük were 19th-century examples—they have exploded since 1990. There are no hard numbers where conlangs are concerned because no one is really counting, but according to van Steenbergen—a regular visitor to online conlanger forums—they probably number in the thousands by now: “Sooner or later, they are bound to outnumber the natural languages.”

Most conlangers do what they do for pleasure. J.R.R. Tolkien admitted in later life that he had created Middle Earth to give his languages somewhere to be spoken—and something to speak about. He referred to conlanging as his “secret vice,” and conlanger David Adger agrees that there is something “curiously satisfying, and almost addictive” about it. “As you create [a grammar], solving problems about how to express different thoughts, you can begin to see how the sinews of the language work, how they mould the thoughts into ways of speaking that make sense for that language,” he wrote in a recent blog post.

The recent proliferation of conlangs has been driven by the internet, as resources became more accessible and people who were initially ashamed of a nerdy pastime discovered like-minded others and came together in online communities. That in turn meant that producers of sci-fi movies and TV series knew where to turn when they wanted a now obligatory alien-sounding conlang built, and some conlangers—like David Peterson, the inventor of Game of Thrones’ Dothraki—have turned professional. There is another category of conlangers, however, who couple their love of linguistic creativity with serious scientific investigation.

Mayeux, for example, isn’t just interested in Parisot’s psyche. The fake Taensa grammar is also of scientific interest to him. Natural languages—languages that humans speak—obey certain rules, few of which had been identified by Parisot’s time, so the Frenchman inadvertently broke them. “The hoax would have been detected much more quickly today because of what we now know,” says Mayeux. But we still don’t know exactly what it is that makes a language easy for humans to learn, or why some languages are more popular than others.

Adger isn’t only the inventor of a number of languages—including, most recently, one for fictional early modern humans; he is also a linguist at Queen Mary University of London who tries to answer those big questions about how we learn languages using conlangs as experimental probes. A few years ago, with Jenny Culbertson of the University of Edinburgh and others, he tested the choices that speakers of English and Thai made when asked to build sentences in a conlang called Nápíjò. Using a conlang allowed the researchers to alter the word order they presented to study participants, while keeping the words the same, to see how they responded to that order—knowing that their responses had to be pure guesswork. To lend Nápíjò some authenticity, however, they told participants that it was a language spoken by about 10,000 people in rural southeast Asia.

The fact that so many conlangers want to both invent and preserve languages attests to a peculiarly human passion for language.

Adger and his team recruited native speakers of English and Thai for the study. When it comes to nouns and modifiers, English and Thai use two of the most common word orders found in human languages, which happen to be perfect opposites—the English phrase “those three big oranges” would be the equivalent of “oranges big three those” in Thai. Adger taught the study participants different versions of Nápíjò, in which the noun placement in this word order was the opposite of in their native tongue: first for English speakers, last for Thai. When speakers of the two languages had to guess the rest of the modifier order in their version of Nápíjò, they did not follow their own language’s rules, but those of the other one—ignoring the vast majority of theoretically possible alternatives and defaulting to the other common word order.

According to Adger, that suggests that once you remove the learned component of language, the human brain still shows certain linguistic biases. Noam Chomsky’s concept of universal grammar has come under attack in recent years, but to Adger—a Chomsky fan—this is evidence that at least some components of language are universally hard-wired. He and his colleagues are now testing that hypothesis with a third language—Kîîtharaka, which is spoken in Kenya—that contains a highly unusual word order, very different from that of Thai or English.

While conlangs are useful for linguistics research, their creators generally recognize that they are unlikely to add to or replace natural languages. They are not going to reverse the erosion of linguistic diversity that is currently underway, for example, that has reduced the number of spoken languages to around 6,500—about half the number estimated, very approximately, to have been spoken at the peak of linguistic diversity around 10,000 years ago. “The languages that are dying are products of thousands of years of evolution,” says Mayeux. “They have cultural knowledge and a collective memory inside them that is being lost for reasons that are very different from those that would lead someone to create a language.”

There have, nevertheless, been cases of conlangs that have escaped the virtual realm. Star Trek fans will no doubt cite Klingon, which was created in the 1980s, but in her 2010 book In the Land of Invented Languages, American linguist Arika Okrent estimated that only a couple dozen people speak Klingon fluently. More interesting, perhaps, is Interslavic, which was invented to allow speakers of the 15 or so Slavic languages to easily communicate with each other.

There have been several versions of Interslavic through history, including an early one invented by a 17th-century Croatian missionary named Juraj Križanić, but the modern version was created by Jan van Steenbergen and others. Modern Interslavic is only about 10 years old, but it already has thousands of fans and hundreds of fluent speakers, van Steenbergen says. It is practical because it allows speakers of the Slavic languages—not all of which are mutually comprehensible—to communicate with each other. All Slavs can understand it, and they can learn to speak it very quickly—thereby reducing translation costs. “You see hotel notices in Interslavic now,” says van Steenbergen, who has been awarded a medal by the Czech Senate for his efforts. “And I have been called upon to interpret in court using Interslavic.”

Mayeux, meanwhile, splits his spare time between creating new languages and reviving old ones—or attempting to—and he says that there is a good deal of overlap between the people who engage in the two activities. He grew up in Louisiana, and he belongs to an online community that is attempting to revive Louisiana Creole—the language spoken by the colonial inhabitants of Louisiana who replaced the Taensa—which is itself now under threat. He is part of a community translating local stories, songs, and other materials into Louisiana Creole. Mayeux doesn’t know if these efforts will succeed, since experience suggests that language revitalization only works when political will swings behind it and children start learning the endangered languages again. (Only about 4,000 native speakers of Louisiana Creole remain, and most of them are elderly.) Nevertheless, the fact that so many conlangers want to both invent and preserve languages attests—like the story of Parisot—to the human passion for language. And that, itself, is of scientific significance.