As the saying goes, we are what we eat—but does that aspect of our identity carry over to the languages we speak?
In a new study in Science, a team of linguists at the University of Zurich uses biomechanics and linguistic evidence to make the case that the rise of agriculture thousands of years ago increased the odds that populations would start to use sounds such as f and v. The idea is that agriculture introduced a range of softer foods into human diets, which altered how humans' teeth and jaws wore down with age in ways that made these sounds slightly easier to produce.
“I hope our study will trigger a wider discussion on the fact that at least some aspects of language and speech—and I insist, some—need to be treated as we treat other complex human behaviors: laying between biology and culture,” says lead study author Damián Blasi.
If confirmed, the study would be among the first to show that a culturally induced change in human biology altered the arc of global languages. Blasi and his colleagues stress that changes in tooth wear didn't guarantee changes in language, nor did they replace any other forces. Instead, they argue that the shift in tooth wear improved the odds of sounds such as f and v emerging. Some scientists in other fields, such as experts in tooth wear, are open to the idea. (Today, many scientists are racing to save languages that are dying out.)
“[Tooth wear] is a common pattern with deep evolutionary roots; it’s not specific for humans [and] hominins but also present in the great apes,” University of Zurich paleoanthropologists Marcia Ponce de León and Christoph Zollikofer, who didn't participate in the study, say in a joint email. “Who could have imagined that, after millions of years of evolution, it will have implications for human language diversity?” (Another study shows how ancient cave art may be linked to language.)
While the study relies on various assumptions, “I think the authors build a very plausible case,” adds Tecumseh Fitch, an expert on bioacoustics at the University of Vienna who wasn't involved with the work. “This is probably the most convincing study yet showing how biological constraints on language change could themselves change over time due to cultural changes.”
But many linguists have defaulted to skepticism, out of a broader concern about tracing differences in languages back to differences in biology—a line of thinking within the field that has led to ethnocentrism or worse. Based on the world's huge variety of tongues and dialects, most linguists now think that we all broadly share the same biological tools and sound-making abilities for spoken languages.
“We really need to know that the small [average] differences observed in studies like this aren't swamped by the ordinary diversity within a community,” Adam Albright, a linguist at MIT who wasn't involved with the study, says in an email.
Teeth might feel solidly embedded in the skull and jaw, but as anyone who's worn braces can tell you, teeth can shift and drift in the jawbone quite a bit as people age. Humans are often born with a slight overbite, but as teeth naturally wear down, they tilt to a more vertical orientation. To compensate, the bottom jaw shifts forward so that the top and bottom rows of teeth are in an edge-on-edge alignment.
For much of our species's history, this edge-on-edge configuration was the norm in adulthood, as seen in many prehistoric skulls studied over the last three decades. But when societies adopted new agricultural techniques, such as the cultivation of cereal grains and raising cattle, diets changed. Once porridge, cheese, and other soft foods dominated ancient menus, people's teeth saw less wear, which let more people keep an overbite into adulthood.
A more common overbite, the thinking goes, set the stage for sounds such as f and v, which you make by tucking your bottom lips beneath their top teeth. If your top teeth jut out slightly more, it's theoretically easier to make these sounds, which linguists call labiodentals.
Blasi and his colleagues aren't actually the first to make this case. Influential linguist Charles Hockett suggested a similar idea in an essay published in 1985. But Hockett's case rested on a particular claim by C. Loring Brace, an influential anthropologist at the University of Michigan. A year after Hockett's essay, Brace replied to say that he had changed his mind—causing Hockett to dismiss his own idea.
For decades, Hockett and Brace's back-and-forth was taken as the final word on the matter. So when Blasi and his colleagues revisited the issue several years ago, it was for demonstrative purposes. But when the team started statistically analyzing databases of world languages and their distribution, they started to see a stubborn relationship they couldn't explain.
“We tried for months to show that this correlation didn’t exist ... and then we thought, maybe there’s actually something there,” says study coauthor Steven Moran, a linguist at the University of Zurich.
The team then conducted follow-up analyses, including some that made use of a computer model of the face's bones and muscles. The models found that it takes about 29 percent less energy to make labiodentals with an overbite than without.
Once f and v became less energetically expensive to make, Blasi's team says, the sounds became more common—perhaps only accidentally at first, as people mis-vocalized sounds made by both lips touching, such as p or b, or what linguists call bilabials. But once labiodentals appeared, they stuck around, presumably because they're usefully distinct. In English, the phrases “fever has gone global” and “Bieber has gone global” have very different meanings.
In English, the phrases “fever has gone global” and “Bieber has gone global” have very different meanings.
When Blasi's team compared language records with data on how different societies acquire food, they found that languages used by modern hunter-gatherer societies use about a quarter of the f sounds that that agricultural societies do, suggesting a possible correlation with diet. And when they looked at the vast family of Indo-European languages, they found that the odds of labiodentals popping up were worse than 50 percent until 4,000 to 6,000 years ago.
The time of labiodentals' rise roughly matches up to when their speakers first started using dairy products and cultivating cereal grains. Blasi's team argues that this is no coincidence.
“The landscape of sounds that we have is fundamentally affected by the biology of our speech apparatus,” says study coauthor Balthasar Bickel. “It's not just cultural evolution.”
That said, everything from social structure to short-term fads can also shape language—and the rise of agriculture brought with it profound societal changes. Linguists also stress that even within a single population, people's speech can vary widely. (Does geography influence how a language sounds?)
University of Southern California linguist Khalil Iskarous, who wasn't involved with the study, is willing to entertain the paper's probabilistic arguments. But he points out that human speech organs don't use that much energy relative to movement, and they're so flexible, they often can compensate for differences in bone structure. Sounds made more difficult by an overbite, such as bilabials, might be expected to decline—but many languages clearly keep them around.
What's more, if energy expenditures really play a driving role in languages, many difficult speech sounds would face an uphill climb to adoption. For instance, Iskarous points to the clicks that are still integral to many of the Khoisan languages of southern Africa.
“If extremely small amounts of effort should make a difference between whether you're likely to have a speech sound or not, you would predict, for instance, that no language should have clicks. And clicks not only exist, they've spread into many languages that didn't have them,” he says. “These are extremely effortful, but it doesn't matter: There are cultural forces that decided that clicks would spread.”
But Blasi continues to stress that his team's claims don't preclude culture.
“The probabilities [for making labiodentals accidentally] are relatively low, but given sufficient numbers of trials—and by this, we mean that every utterance you make is a single trial—over generations, that leads to the statistical signal we see,” he says. “But it's not a deterministic process, right?”
As scholars continue to debate, Blasi's team has ideas for where to go next. For instance, they say their methods could help better reconstruct how ancient written languages were spoken aloud and so catalog language's endless phonemes most beautiful.