Chimpanzee Dialect Grows More Distinct With Age

IT STARTS as a scribble on a screen, a fraction of a second of infant chimpanzee whimper stretched into a spectrogram until it looks more like a smear of ink than a sound. Run enough of those scribbles through the right statistics, though, and a pattern climbs into view: young chimpanzees from two different homes already sound subtly different, and the gap widens the older the animals get. A study comparing infant and juvenile chimpanzees at a Zambian sanctuary and a Tanzanian national park has found that population-specific patterns in their grunts, laughs and whimpers sharpen through childhood, which is exactly the kind of timing biologists use to tell a learned trait from an inherited one. The mothers, it turns out, may have less to do with it than the neighbours.

That distinction matters because dialects, in the loose sense biologists mean by it, are supposed to be a badge of belonging: a way of telling who is from around here and who is not. Humans do this constantly and almost unconsciously, sorting accents into home teams within seconds of a stranger opening their mouth, and until recently nobody knew whether any of our closest living relatives built something similar as they grew up.

Chimpanzees have carried known population-level differences in their calls for decades, ever since researchers first compared the pant-hoots (long, whooping calls chimpanzees use to broadcast their location) of communities in Tanzania and found the fine structure did not quite match. What nobody had settled was why: genes, environment and social learning each made a case, and each new comparison favoured a different answer. The team behind the new work, at Queen Mary University of London and the University of Portsmouth, took a different angle, following the calls not across adult populations but across development, from infancy into the juvenile years. If a vocal signature is baked in at birth, it should already be at full strength in the youngest animals; if it needs experience to sharpen, it should grow louder with age.

A Signature That Grows Louder With Age

So the researchers went looking for exactly that growth curve, in 30 chimpanzees in total, 21 infants and 10 juveniles, split between Chimfunshi Wildlife Orphanage in Zambia and the Gombe Stream Research Center in Tanzania, one a semi-wild sanctuary in open miombo woodland and the other a wild population in dense rainforest. Across more than 11,000 recorded grunts, laughs and whimpers, a computer was trained to guess, from the sound alone, which site an animal called home, then checked against chance.

For every call type tested, the guesser did better than chance: about 90% accuracy for grunts and 97% for laughs, against a chance level of roughly 60% for both, and about 98% for whimpers against a chance level near 70%. Whimpers were the interesting exception on the way there, because while grunts and laughs already carried a usable signature in infancy, infant whimpers gave the model nothing to work with; only once the animals reached the juvenile years did whimpers start to give away which population an animal belonged to. Grunts showed the same climb from a different starting point, correctly sorted about 76% of the time as infants and about 92% of the time as juveniles, and laughs and whimpers each grew more distinctive with age too.

“Understanding how communication develops during early life is fundamental to understanding animal behaviour,” says Kassandra Giragosian, a doctoral researcher at Queen Mary University of London. “Our findings show that chimpanzee vocalisations are not simply fixed from birth but continue to develop as young animals interact with their social and ecological environments,” she adds, a small reframing of what a chimpanzee’s whimper is even for.

The tool doing the sorting was a permuted discriminant function analysis, a mouthful of a method that boils down to teaching a computer to draw the best possible line between two groups using fifteen acoustic measurements per call: pitch wobble, how energy is spread across frequencies, how long each element of a call bout lasts. Crucially, the method also shuffles which population each animal is labelled as belonging to, a thousand times over, generating a distribution of how well a computer could do purely by luck, so the real classification rate can be judged against chance rather than a comforting but arbitrary benchmark. Entropy, essentially how noisy or tonal a call sounds, turned up again and again as the most consistent contributor to telling the two populations apart in infant grunts and infant whimpers, while element rate (how quickly the parts of a call bout follow one another) mattered most for laughs and for the grunts of older juveniles. None of this is something a chimpanzee could consciously choose to change, any more than a person can choose an accent, but it is exactly the kind of subtle timing and texture that social exposure and repeated practice could plausibly reshape over years, rather than genes reshaping over generations, much as captive adult chimpanzees were shown, in 2015, to gradually converge on a shared food grunt after two groups were merged and got to know each other (Watson et al.). That distinction matters because this is a retrospective study, built from recordings gathered years and, in Gombe’s case, decades apart on different equipment, and the authors argue at length that a mismatch in recording gear cannot explain a pattern that shows up in some call types and stages but not others.

What Decades of Old Recordings Can’t Rule Out

The retrospective mix in equipment is not the only asterisk worth keeping in view; the sanctuary population and the wild population differ in more than which trees line the forest, since the Zambian animals were rescued and raised at an orphanage while the Tanzanian animals grew up wild, with their own mothers. The paper does not attempt to pull that rearing difference apart from habitat or genetics as a separate explanation, so some of what reads as a population signature could, in principle, be a captive-versus-wild signature wearing a population’s clothes.

The authors are not claiming anything close to language: no shared grammar, no agreed vocabulary, nothing that would let a Zambian-raised chimpanzee and a wild Tanzanian one trade calls and be understood. What they are claiming is narrower, and more interesting: a repertoire most researchers assumed was fixed at birth is still being written well into a chimpanzee’s childhood.

That has a knock-on effect for how the debate over animal dialects gets argued, because for years the central question has been binary: written into the genes, or a product of the group a chimpanzee happens to grow up in? A signal that starts weak and grows stronger with age is awkward for a purely genetic story, since genes do not usually wait until adolescence to switch a trait fully on, but it fits a trait shaped by repeated exposure to the particular grunts, laughs and whimpers of the individuals a young chimpanzee spends its days with. It also lines up with human vocal development: in 2010, a study comparing English and remote Namibian listeners found that laughter and other positive vocalisations were recognised far more reliably within a culture than across it, while distress calls like screams travelled across the cultural divide with little loss, on the logic that a call with too much riding on being understood has little room for local variation. Whimpers, the one call type that stayed silent on population membership until the juvenile years here, are also chimpanzees’ own distress call, aimed mostly at mothers rather than the wider group, and the parallel with human cries is not lost on the authors.

What remains open is the mechanism itself: whether young chimpanzees are shaping their calls through direct feedback from other chimpanzees, through years of simply overhearing the same voices, or through something else again that a spectrogram cannot show. Whatever it turns out to be, a species once assumed to arrive with its vocal repertoire fully installed is instead spending its childhood tuning it, one grunt, laugh and whimper at a time.

FAQ