When Ahana Fernandez and her colleagues trek through the rainforests of Central America, they keep their ears tuned for an unusual sound: high-pitched, repetitive chirping and squeaking. The noises come from greater sac-winged bat pups (Saccopteryx bilineata). Though they sound nothing like the babbling of human babies, the animal behavior researcher and her colleagues at the Museum of Natural History (MNH) in Berlin suspected the two might have something in common.
Now, after analyzing more than 55,000 of these bat babbles, the team has found that the sounds share significant similarities with the babbling of human infants, right down to their endless repetition of a single syllable. That means the chatty bats could serve as a useful model species for understanding how some mammals—including humans—turn their early vocal play into adult “speech.”
It’s “beautiful work,” says Clara Levelt, a developmental linguist at Leiden University who was not involved with the study. “It will form the basis for a lot of new research.”
Adult male sac-winged bats sing complex songs, much like songbirds, to defend their territories and attract mates. These sounds are made up of strings of 25 different distinctive “syllables.” During infancy, the pups produce the syllables over and over again, seemingly practicing for the day they’ll string them together into full songs.
But as much as this behavior seemed like human babbling, scientists hadn’t formally compared the two, says co-author Mirjam Knörnschild, a behavioral ecologist also at MNH. So the scientists sought advice from researchers studying language development in human babies to draw up a list of the central features of human babbling. These include when such sounds begin—typically early in infancy—and how they are structured—often with syllables that repeat over and over again. Finally, most baby babbles seem to lack communicative purpose. Human babies babble not to ask for food or express discomfort, Knörnschild says; in fact, it seems more like play, something they do even when they’re alone.
Next, the researchers spent long days in the rainforests of Costa Rica and Panama, recording 20 bat pups from eight different colonies for weeks on end. Because human ears aren’t well-attuned to bat calls—and some of the frequencies are beyond our range of hearing—the researchers relied on a laptop showing visual representations of the sounds in real time. Some patterns are nice, distinct streaks—a bit like a heartbeat on a heart rate monitor, Fernandez says. Others look more like dark clouds.
It was the first time anyone had done research like this, so all 55,000 syllables in the recordings had to be classified by hand. But now, Knörnschild says, “We have this awesome database that we can use for machine learning approaches in the future.”
The bat pup chatter shared all the major features of human babbling, the researchers report today in Science. As in humans, it begins early in development, and it contains a large number of repeated syllables—like a human baby saying “bababababa.” And unlike the “isolation calls” that attract their mothers’ attention, the vocalizations didn’t seem to be a form of communication. A babbling pup is “a relaxed and happy pup,” Fernandez says, “sitting in the day roost and just practicing and playing around.”
The work is an “amazingly focused new body of information,” says D. Kimbrough Oller, a developmental linguist at the University of Memphis. Some human development researchers might object to it, given all the differences between bat and human vocalizations—and “there are some massive differences,” he says. The bats had a far more sudden onset and end to their babbling phase, for example, and they learned a much more limited set of adult syllables than do human infants. But the similarities are still important, he says.
The work opens up exciting new options for studies of vocal learning, says Sonja Vernes, a bat researcher at the University of St. Andrews. It’s possible to study bats in far greater numbers than other species like primates, she notes, making them an ideal animal model for researching vocal learning and development in mammals. “You can start asking questions that you can’t ask in humans.”