The wild munia (left) tends to be less social than the Bengalese finch (right) and its song is simpler.


Can these birds explain how language first evolved?

If you want a no-fuss, no-muss pet, consider the Bengalese finch. Dubbed the society finch for its friendliness, breeders often use it to foster unrelated chicks. But put the piebald songbird next to its wild ancestor, the white-rumped munia, and you can both see and hear the differences: The aggressive munia tends to be darker and whistles a scratchy, off-kilter tune, whereas the pet finch warbles a melody so complex that even nonmusicians may wonder how this caged bird learned to sing.

All this makes the domesticated and wild birds a perfect natural experiment to help explore an upstart proposal about human evolution: that the building blocks of language are a byproduct of brain alterations that arose when natural selection favored cooperation among early humans. According to this hypothesis, skills such as learning complex calls, combining vocalizations, and simply knowing when another creature wants to communicate all came about as a consequence of pro-social traits like kindness. If so, domesticated animals, which are bred to be good-natured, might exhibit such communication skills too.

The idea is rooted in a much older one: that humans tamed themselves. This self-domestication hypothesis, which got its start with Charles Darwin, says that when early humans started to prefer cooperative friends and mates to aggressive ones, they essentially domesticated themselves. Along with tameness came evolutionary changes seen in other domesticated mammals—smoother brows, shorter faces, and more feminized features—thanks in part to lower levels of circulating androgens (such as testosterone) that tend to promote aggression.

Higher levels of neurohormones such as serotonin were also part of the domestication package. Such pro-social hormones help us infer others' mental states, learn through joint attention, and even link objects and labels—all prerequisites for language, says developmental psychologist Michael Tomasello of Duke University in Durham, North Carolina, who studies social cognition.

In recent papers and at Evolang, a biannual conference on language evolution held here this spring, researchers turned to birds, foxes, and bonobos to help understand how domestication may have paved the way for language. Constantina Theofanopoulou, a neuroscientist at the University of Barcelona in Spain who convened the Evolang workshop, calls it the "most promising" working hypothesis to account for the thorny problem of language evolution, because it "puts together evidence from different levels of biological analysis: the anatomical, the brain, the endocrine system, and behavior."

In his talk at Evolang, ornithologist Kazuo Okanoya of the Riken Center for Brain Science in Wako, Japan, focused on the munia and the Bengalese finch, which people domesticated some 250 years ago. Both birds are vocal learners, a rare trait that lets them pick up calls from adult tutors—as do parrots, hummingbirds, and humans. But their songs are wildly different, as Okanoya demonstrated by whistling examples of each.

He then presented data quantifying what the audience had heard: Munia songs tend to be shorter, simpler, and full of unmelodic segments of acoustic "noise," compared with the longer, louder finch songs, which contain peeps, chirps, and segments that often repeat and recombine in improvisational ways.

Munia song
K. Okanoya
Bengalese finch song
K. Okanoya

Okanoya says the differences likely arose from domestication, in particular the finches' relatively stress-free environment. He's shown that the finches have lower fecal levels of corticosterone—a hormone that boosts aggressiveness and blunts cognitive functioning in birds—than the munia. In his talk, Okanoya reported that high corticosterone levels inhibit the growth of neurons in the birds' song-learning system, which is larger in the finches than in the munia.

Thus, Okanoya hypothesizes, selection for tameness and gregariousness by pet owners boosted the finches' capacity for complex song. And because attention-getting songs help advertise fitness to females, the males best at learning and singing would be most likely to pass their genes on to the next generation, sparking further complexity.

If early humans somehow developed their own lower-stress "domesticated" environment—perhaps as a result of easier access to food—it could have fostered more cooperation and reduced aggression, speculates evolutionary linguist Simon Kirby, writing with James Thomas, both of The University of Edinburgh, in a recent paper in Biology & Philosophy. As with the finch, a mellow environment may have allowed for an expanded role for learning, including in language acquisition.

Kirby and Thomas point out another analog for humans: domesticated foxes. In a famous experiment, Russian geneticist Dmitry Belyaev and colleagues selected for tameness among captured Siberian silver foxes starting in the 1950s. If a wild fox did not attack a human hand placed into its cage, it was bred. Over 50 generations, the foxes came to look like other domesticated species, with shorter faces, curly tails, and lighter coloring—traits that have since been linked to shifts in prenatal hormones.

Unlike their wild counterparts, tame foxes came to understand the importance of human pointing and gazing, Thomas and Kirby note. That ability to "mind read" is key to language. Thus, even though the foxes don't vocalize in complex ways, they show that selection only for tameness can carry communication skills in its wake.

At Evolang, other researchers zeroed in on bonobos, great apes that show some signs of self-domestication, including low levels of aggression and sensitivity to the gaze of others. According to Zanna Clay, a primatologist at Durham University in the United Kingdom, bonobos also display a building block of early language: Instead of sticking to a fixed repertoire of "inherited" calls, they can improvise.

Clay and her colleagues have assembled hundreds of recordings from 18 bonobos in the wild and in zoos, showing that individuals combine set types of calls in distinct ways for different situations. She hypothesizes that self-domestication may have helped shape this communicative flexibility.

Stronger proof may come from genetic studies. Theofanopoulou and her team recently scoured the scientific literature for genes that differ between wild and domesticated species—cats, dogs, horses, and cattle—and that also show signs of being selected in the domesticated animals. The team did the same for modern humans and what they considered our nearest wild stand-ins, Neanderthals and Denisovans.

Then, the researchers looked for genes that may have evolved in the same way in more than one wild-domesticated pair. There were more than three dozen, many linked to brain plasticity, learning, and the development of the nervous system, the team reported late last year in PLOS ONE. Some, such as the gene for a receptor for the neurotransmitter glutamate, are linked to processes that could shape a language-ready brain. But there's no clear path yet from these genes to their function—or to the sweeping changes linked to domestication, cautions Antonio Benítez-Burraco, a linguist at the University of Seville in Spain.

Tomasello also cautions against trying to explain human language solely from animal models. "I think humans were selected to actually collaborate," not just to get rid of aggression, he says. "[That] fundamentally cooperative motive … is a precursor to uniquely human communication."