Bonobos, our closest relatives, may not speak, but they use something strikingly similar to human language.
A new study from researchers at the University of Zurich and Harvard University reveals that wild bonobos use complex vocal combinations. These resemble how humans string words together.
Researchers observed bonobos in the Kokolopori Community Reserve in the Democratic Republic of Congo. They found that bonobo calls follow the rules of compositionality, a core feature of human language.
Bonobo calls are more than random noises. They are structured, purposeful, and, in some cases, surprisingly complex. This research doesn’t just hint at intelligence. It shows the skeleton of a system that mirrors how we construct meaning from words and phrases.
If correct, this means that the boundary between human and non-human communication is far more blurred than scientists used to think.
One feature of human language stands out – compositionality. This is our ability to string small parts together into meaningful wholes.
For example, a dancer can be blond; a dancer can be bad. In the first case, “blond” adds a trait, while in the second, “bad” modifies how we think about the dancing itself. The latter is more than a label – it’s judgment woven into structure.
Compositionality comes in two forms: trivial and nontrivial. Trivial compositionality occurs when each word keeps its independent meaning. Nontrivial compositionality occurs when the combination reshapes meaning, where one part changes the sense of the other.
Until now, evidence for nontrivial compositionality in animals other than humans was missing. Some birds and primates showed signs of trivial compositionality. But no study had proven that any non-human species could modify meaning in such layered ways.
This is where the bonobos come in. Their vocal combinations suggest they may not only understand the value of combining sounds – but can manipulate meaning like we do, through structured, nontrivial arrangements.
The researchers collected 700 vocal recordings of the Kokolopori bonobos over time.
Each utterance was annotated with more than 300 contextual features – what the bonobo was doing, who was nearby, what activity the group was engaged in. These features, called FoCs (features of circumstances), offered a way to map meaning.
To analyze the patterns, the team used a framework borrowed from distributional semantics, a linguistic method that typically applies to human languages. It maps meanings by looking at how words appear across similar contexts. Words used in similar situations tend to have related meanings.
Using this method, researchers projected bonobo calls into a semantic space. In this space, calls with similar meanings cluster close together.
This allowed the scientists to examine not just individual calls, but also what combinations formed – and whether those combinations held meaning greater than the sum of their parts.
The researchers applied a rigorous four-step process to evaluate each combination. First, they checked if individual call types had distinct meanings.
Then, they assessed whether combinations had meanings that were different from the meanings of their parts. Next, they asked whether a combination’s meaning could be derived from its components.
Finally, they looked for cases where the meaning of the whole was different from simply adding up the parts – which would be evidence of nontrivial compositionality.
“All seven bonobo call types that were considered in our analysis combine into four compositional structures, of which three exhibit nontrivial compositionality,” the study’s authors report.
This is a major shift. It means that not only are bonobos capable of combining calls meaningfully, but they also do so in ways that mirror the deeper, more abstract structures of human language.
The research found that combinations like High-hoot_Low-hoot, Peep_Whistle, and Peep-yelp_High-hoot all exhibited nontrivial compositionality. One combination, Yelp_Grunt, did not. It retained trivial structure, where the meaning was just the sum of the parts.
Understanding bonobo communication means more than just identifying call types. It means deciphering the purpose behind each sound. Grunts, for example, seem to signal “look at me,” drawing attention in various social settings like grooming or feeding.
High-hoots are louder and often used over distance or in tense moments, suggesting “pay attention to me.” Low-hoots emerge during excited moments, such as preparing a nest for the night.
The whistle plays another role entirely – it helps maintain group cohesion, perhaps saying “let’s stay together.” Peeps and yelps, though similar, differ in tone. Peeps are more suggestive – maybe “I’d like to…” – while yelps are firmer and more like commands: “let’s do that.”
Then come the combinations. Peep_Whistle, for example, may mean something like “I’d like us to stay together,” and often appears during sensitive moments such as displays or copulation.
High-hoot_Low-hoot might signal distress, urging others to pay closer attention. Peep-yelp_High-hoot could be a way to organize group movement, coordinating actions with distant individuals.
Determining whether a combination was trivial or nontrivial required more than guessing.
Researchers compared each combination’s meaning against hypothetical “additive” meanings – where A and B’s meanings were just added together. If the real combination strayed significantly from this sum, it was labeled nontrivial.
“For all compositional combinations except Yelp_Grunt, the difference was greater than zero,” the study states. That made Yelp_Grunt the only trivial structure. The other three combinations carried a new kind of meaning. They didn’t just add. They transformed.
This means that in at least three cases, bonobos created new meanings from existing calls. This aligns closely with how humans use language to express ideas. Words shift and adjust each other in context.
Bonobos, it seems, do this with their own form of language too.
What do these findings tell us about ourselves? Quite a lot, it turns out.
“Since humans and bonobos had a common ancestor approximately 7 to 13 million years ago, they share many traits by descent, and it appears that compositionality is likely one of them,” said Harvard Professor Martin Surbeck.
“Our study therefore suggests that our ancestors already extensively used compositionality at least 7 million years ago, if not more,” added Simon Townsend.
This positions language not as a sudden leap in human evolution, but as a gradual development. It suggests that the foundations for structured communication were laid long before Homo sapiens arrived.
The study also offers a method to understand communication across species. By mapping vocal signals using minimal human judgment and focusing on context, the researchers showed that meaning could be measured more objectively.
“We present a method for reliably inferring the meaning of all the signals of an animal’s repertoire with minimal human judgment,” the authors explain.
This method doesn’t rely on assigning meaning through guesswork. It uses context and data, opening the door to future studies across different groups of animals, from dolphins to elephants.
The researchers acknowledge that some utterances remain difficult to interpret. Calls might convey emotional states, not just intentions. Subtle acoustic variations may exist but go unnoticed. Some calls may serve structural roles, like those in birdsong, rather than conveying clear messages.
Also, vocalizations are not the only form of communication. Bonobos combine vocal calls with gestures and facial expressions. These multimodal signals likely shape meaning too. Future studies may show even richer layers of nonverbal language in bonobo society.
Even so, the evidence is strong. Bonobos do not just make noise. They communicate in structured, meaningful ways. They use building blocks. They rearrange, modify, and combine. And in doing so, they remind us that the roots of language run deep.
The study is published in the journal Science.
—–
Like what you read? Subscribe to our newsletter for engaging articles, exclusive content, and the latest updates.
Check us out on EarthSnap, a free app brought to you by Eric Ralls and Earth.com.
—–