After hearing hundreds of hours of ape calls, a team of scientists say they have detected features of human language. It is the ability to bring together the strings of sound to create new meaning.
The provocative findings published in Journal Science on Thursday elicited praise from some scholars and skepticism from others.
Federica Amich, a primatologist at the University of Leipzig in Germany, said the study helped us to go back further back to the roots of languages millions of years before our species' emergence. “The difference between humans and other primates, including communication, is much clearer and clearly defined than we have long assumed,” Dr. Amichi said.
However, other researchers said the study, conducted on bonobo, a chimpanzee close relative, has revealed little about how we use the word. “The current discoveries say nothing about the evolution of language,” said Johann Borhuis, a neurobiologist at Utrecht University in the Netherlands.
Many species can communicate with sound. But when an animal makes noise, it usually means only one thing. For example, a monkey can refer to a leopard and make another warning call for an eagle's flight.
In contrast, we humans can connect words in a way that combines individual meanings into new ones. Let's say, “I'm a bad dancer.” Combining the words “bad” and “dancer” doesn't mean independently. I'm not saying, “I'm a bad person who happens to dance.” Instead, it means I don't dance well.
Linguists have invoked this construct and have long considered it an integral part of language. “It's the power behind language creativity and productivity,” said Simon Townsend, a comparative psychologist at the University of Zurich in Switzerland. “In theory, you could come up with a phrase you've never spoken before.”
For decades, scientists have found no clear indication of the constituency of other species. But a few years ago, Dr. Townsend and his colleagues discovered a hint of it in a chimpanzee.
In the Ugandan forest, Dr. Townsend's team recorded over 330 hours of chimpanzees working in their daily lives, identifying dozens of different phone calls. To an untrained ear, the recording may sound like a random dissonance. However, Dr. Townsend and his colleagues noticed that certain calls followed other calls more than they could expect by chance alone. All in all, they identified 15 distinctive phone pairs.
Scientists thought that a pair of calls might convey more meaning than the meaning of two individual calls themselves. To test that hypothesis, they spent two years studying one pair in particular. A call known as “WAA-BARK” is followed by another pair known as “Alarm-Huu.”
Chimpanzees make the WAA Burke call as a way to bring other chimpanzees to them. Apes can call, for example, while hunting, or summon allies during battle. They raise alarms and surprise when they are terribly surprised. It probably corresponds to the unexpected sight of a scientist's raincoat.
Dr. Townsend and his colleagues wondered if “alarm-w” meant something else when “waa-bark” followed. They noticed two opportunities for the other chimpanzees to be within their ear range when they paired Cole when they encountered a snake. Perhaps the scientists meant that the two calls came together and something like, “Get ride here and help deal with this snake!”
The experiment continued. For one, as the chimpanzees passed, researchers pulled a fake snake across the trail. Apes were often followed by “WAA-BARK” as predicted.
The researchers then played a pair of calls via speakers to see how chimpanzees respond. Apes have long tended to look at speakers. Almost one minute. If you played only “Alarm Who” or “Waa-Bark” on its own, the chimpanzee glanced for just a few seconds.
Additional clues suggested that the two calls combined to form a snake alarm. When the chimpanzees heard the pair's call, a typical reaction (between apes) jumped into the tree when the snake was around.
Just as these ideas are interesting, I've been slow to test them. To expand and speed up research, Dr. Townsend began working with Martin Sarbeck, a behavioral ecologist at Harvard University, a species of apes that split from chimpanzees two million years ago. Dr. Sarbeck and his colleagues have spent years following the apes in the Cocolopoli Bonobo Reserve in the Democratic Republic of the Congo.
In 2022, Melissa Bursett, a postdoctoral researcher in Dr. Townsend's lab, joined them to eavesdrop on apes. She made 400 hours of recordings, capturing 567 single calls and 425 pairs. Dr. Berthet also made a note of what happened just before Bonobos called. Has the tree fallen? Did the monkeys build nests for the night or groom their friends? Dr. Berthet completed a 336 checklist for each call.
Shane Steinert-Strelkeld, a computational linguist at the University of Washington, said he was not involved in the study. “So I'm very excited about it,” he said.
Back in Zurich, Dr. Bursett listened to the recordings and divided the call into dozens of types. To analyze the meaning of the call, she analyzed the checklist. She and her colleagues used some of the mathematical techniques that artificial intelligence systems like ChatGPT use to learn how words relate to each other. This analysis allowed scientists to visually map bonobo calls. The closer the Cole gets to each other on the map, the more similar their meanings are.
Researchers also found that bonobos frequently used 16 specific calls, with most pairs appearing on the same neighborhood map as two separate sounds containing them. This suggested that their combination did not convey any special meaning.
However, four pairs of call stood out. These landed on maps far away from the arrangement of two separate calls. Together, they seemed to have meaning, unlike either one of the call alone. For example, one such pair combined two calls. A hyfoot made when a bonobo is trying to attract other people's attention far away, and a low foot made when a bonobo is excited by emotions.
Combined, the two calls seem to represent rescue pleas for a perhaps distant bonobo when under attack. “It says, 'I'm suffering so please pay attention to me,'” Dr. Bursett said.
Dr. Berthet said the new results should address skepticism about Dr. Townsend's previous research into chimpanzees. “Linguists always say, 'Yeah, OK, but that's just one combination – what does it really teach us?” “We show here that bonobos actually have some compositional structures, and they use a lot of them.”
Together, two studies on bonobos and chimpanzees suggest that our common ancestors with these apes are also compositional, researchers argue.
However, Dr. Borghis questioned whether new research could indeed detect the constitutive nature of bonobos. “Constituity isn't just about combining two words,” he said. It also said about following the rules of the syntax for constructing words into phrases and units of greater meaning.
Dr. Townsend retorted that the act of pairing calls was probably the first step into a full-scale composition that later emerged in early humans.
The next step, Dr. Steinert-Strelkeld said, is for researchers to analyze Bonobo data in a more sophisticated way and see if these results continue. Maybe you can train your computer to learn the meaning of individual calls and then test it to see if you can predict the meaning of a pair of calls.
“It's incomplete,” he said of the new research. “But that's a good first step.”
