Bonobo Calls Are More like Human Language Than We Thought

Bonobos’ Complex Calls Share an Extraordinary Trait with Human Language

By Cody Cottier edited by Allison Parshall

We humans concoct never-before-heard sentences with ease, embedding phrases within phrases to express the wildest ideas we can dream up (“the purple pangolin that waltzed across the ballroom had a flaming pineapple on its nose”). Such abilities seem unrivaled in the animal world, but a new study suggests they’re not entirely absent: bonobos, our closest living relatives, create combinations of calls that seem to share key aspects of human language.

In a new study published on Thursday in Science, researchers report that bonobo communication is rich in a feature that linguists call compositionality. This refers to the way we string words together to compose larger structures with more complicated meanings. Linguists divide compositionality into two categories, a simple version and a more sophisticated one, and researchers have long thought human language stands alone in the higher tier. Previous studies have found that some primates and birds are capable of “trivial” compositionality, in which words that each have a specific meaning on their own can be added together to create a fuller, more meaning-rich picture (“bake pie”).

But the new study shows that bonobos, like us, seem to do something a bit more advanced than that. In “nontrivial” compositionality, certain parts modify others. An example is the sentence “they baked a pumpkin pie.” Here “pumpkin” and “pie” join to form a new composite idea. This strategy gets more bang for your communicative buck, according to the new paper’s co-senior author Simon Townsend, who studies comparative communication at the University of Zurich. “That’s what we’ve evolved it for,” he says, “to add this important nuance and complexity to the meaning that we convey.”

On supporting science journalism

If you’re enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.

Townsend notes that nontrivial compositionality in bonobos is orders of magnitude less complex than what we see in human language. Still, he argues, it represents “another layer chipped away” from the apparent wall of human uniqueness that separates us from the rest of the animal kingdom. And because this crucial capacity may be present in such a close evolutionary relative, it could add to our understanding of how language arose in Homo sapiens. “That’s really exciting,” Townsend says. “It allows us to go back in time and work out what our last common ancestor—who was living seven million years ago in the forests of Africa—was doing.”

To gather data for the study, lead author Mélissa Berthet, a postdoctoral researcher in Townsend’s group, turned to those very same forests. She spent five months recording bonobo calls in the Kokolopori Bonobo Reserve in the Democratic Republic of the Congo. These calls were labeled with categories that ranged from “peeps” and “whistles” to “grunts” and “yelps.” For each of the 700 recordings, Berthet took meticulous note of 300 different contextual details: Was another group of bonobos nearby? Was the caller feeding, grooming or resting? How did others react?

These details allowed the researchers to infer the potential meaning of different calls based on context. Peeps, for example, seem to help coordinate activities with other bonobos—“I would like to …” may be a rough translation, Berthet says. Whistles seem to be used to maintain group cohesion, with the sounds expressing something like “let’s stay together.” The scientists used these contextual data to plot each bonobo call on a five-dimensional “map” of meaning; calls that were closer to each other on the map had more similar meanings. Using this technique, which was borrowed from linguistics, Berthet and her colleagues built what she describes as “a kind of dictionary” of the seven most common calls in the bonobo repertoire.

Then the team analyzed combinations of those seven calls. That’s where compositionality emerges: for example, a peep followed by a whistle makes a “peep-whistle,” a nontrivial composition that bonobos use in sensitive social interactions such as sex or dominance displays. By plotting the combinations on the meaning map, the researchers calculated that four showed compositionality and that three of these had a meaning beyond what would be expected by just adding the meanings of two calls together—indicating that the three combinations were examples of nontrivial compositionality. What’s more, all seven of the common call types appeared in at least one combination, revealing more extensive compositionality in bonobos than in any other species that has yet been studied.

“Bonobospeak” likely does not map neatly onto human concepts. It’s not clear what the peep-whistle means or how it’s derived from the meanings of its component parts (à la “pumpkin pie”). But because the researchers’ method comes from linguistics, where it has been used to determine compositionality in human communication, they’re “confident [the results] look like nontrivial compositionality, at least in the mathematical sense,” Berthet says. “We’re just not sure yet what it means exactly.”

Shane Steinert-Threlkeld, a computational linguist at the University of Washington, who wasn’t involved in the new paper, suggests another interpretation: the nontrivial combined calls could be more like idioms. Maybe their meanings aren’t a function of their parts at all, like when someone tells you to “break a leg” to wish you good luck (though the study’s researchers think that’s unlikely because their calculations indicate that the meanings of the nontrivial combinations are still related to the meanings of those combinations’ parts). Steinert-Threlkeld also notes that this linguistics-based method, while innovative, is “a little bit too new to fully endorse.” But he lauds the unprecedented scale of data collection, calling the study “valuable in showing us what can be done and what more needs to be done [for the results to be] fully convincing.”

Thom Scott-Phillips, a cognitive scientist at Central European University’s location in Hungary, who was also not involved in the study, found its enormous dataset and novel method impressive as well. But he’s not sold on bonobo calls being comparable to language. He argues that even bacteria—which signal to one another using combinations of molecules—seem to meet most of this method’s criteria for nontrivial compositionality, raising the question of whether it’s actually measuring something else entirely. “If they go and do the same work with [bacteria], and they don’t find it,” he says, “that would be a challenge to someone like me.”

Townsend says nontrivial compositionality may indeed be more widespread than previously assumed, though he doubts it will be very extensive in nonprimates, let alone bacteria. He and his colleagues hope their new observational approach—which is far more efficient than traditional callback experiments, in which a researcher plays a recording and uses animals’ reaction to it to determine meaning—will encourage other scientists to test nontrivial compositionality in a wide range of species. “We don’t know yet if bonobos are special,” Berthet says. “We developed this method, we used it on bonobos, and we found very cool results. But maybe you could do that on other animals.”