Language, the ability to learn words and combine them into meaningful sentences, is at the heart of what it means to be human. We start our lives like our great ape cousins (e.g. chimpanzees), able to do little more than cry. But quickly, we learn to copy the words of our parents and peers. This ability to mimic sounds, called vocal learning, is surprisingly rare. Other great apes can’t do it. Neither can dogs or cats. In fact, the only mammals that have been shown to learn their vocalizations are cetaceans (dolphins and whales), pinnipeds (seals and sea lions), elephants, and bats.

Since our closest relatives can’t modify their sounds, our last common ancestors probably couldn’t either. What is different about humans that caused us to evolve our linguistic powers? How did the brains of our ancestors change to enable us to speak?

These questions motivate my research, but are difficult to answer. Scientists can’t study the vocalizations of our hominin ancestors (e.g. ancient Homo sapiens) because spoken language doesn’t fossilize like bones do (though our descendants millions of years from now will have it easier when they dig up our YouTube videos and podcasts). Brains decay very quickly so we can’t look at the neural structures of Homo habilis or Homo erectus.

So, where can researchers like me look for answers? One word: birds. Vocal learning is widespread in songbirds, parrots, and hummingbirds, with abilities ranging from songbirds (e.g. zebra finches) that can learn only one song to parrots (e.g. African greys) that can mimic hundreds of human words. And even though birds and humans are separated by over 300 million years of evolution, the brain areas in birds controlling vocal learning are strikingly similar to language regions in the human brain. That means birds and humans have converged on similar neural mechanisms for mimicking sounds.

In my work, I study how vocal learning birds perceive patterns of sounds. Patterns can be thought of as sequences of basic units, whether they are words, bird sounds, or geometric shapes. With this framework, we can look at language and birdsong on the same plane. But since I don’t speak bird, I ask the birds to tell me what patterns they can hear. I train them to perform the equivalent of a human hearing test: instead of raising their hand when they hear a sound, they peck a key. Using this approach, our lab (the Avian Behavioral Neuroscience lab at University of Maryland) has shown that parakeets can hear when a single element of their long rambling song is played out of order, meaning they know grammar-like rules about how their song should be arranged.

And I am just one of many scientists worldwide exploring the cognitive abilities of vocal learning birds. Exciting research using these feathered linguists is pushing us closer than ever to understanding how a brain without language could evolve into one with it.

Scientists from the University of Texas Southwestern and Duke University recently discovered a new pathway in the songbird brain by making firing neurons glow. They injected a virus into neurons that connect a region important for producing song to a region important for hearing song. The virus causes a fluorescent protein to be expressed every time the brain cells fire, and the researchers found that the neurons glowed every time the birds sang. This means that the neurons are transferring instructions about how to sing to neurons that hear what is being sung. This feedback from the motor system onto the auditory system allows a learning bird, like your pet cockatiel, to predict what they’ll hear coming out of their beaks, and, as a result, better mimic your favorite song. Similar pathways in the human brain likely help us learn and maintain spoken words.

Researchers at Leiden University have identified important differences in how a songbird (i.e. zebra finches) and parrot (i.e. parakeets) listen to patterns of sounds. They trained birds to peck a key when they heard a sequence of sounds arranged in a specific pattern, XXY, and to withhold pecking when they heard a different pattern, XYX. Once the birds learned the difference, the researchers plugged new sounds into the same patterns and the parakeets still pecked for the correct XXY arrangement, while the zebra finches did not. This means that parakeets can learn an abstract pattern of sounds, which is a very rare ability among animals and fundamental to human language acquisition. We use abstract patterns, for instance, to understand the relationship between words in a subject-verb-object sentence, no matter what words are plugged into the frame.

These differences in how zebra finches and parakeets listen to patterns of sounds might be explained by intriguing differences in songbird and parrot brains. After comparing the brains of many parrot and songbird species, a large international group of researchers hypothesized that parrots share a “core” song system with songbirds, but then evolved a unique “shell” song system beyond that, enabling enhanced vocal learning ability. In a similar way, human brains may have developed new language-related circuits on top of ancestral hominin circuitry.

After decades of research, we now know more than ever about vocal learning in birds, providing invaluable clues about how we, Homo sapiens, became equipped for language. Yet, there is still so much we have yet to discover, and I think birds are the key to understanding the evolution of language. Without them, the trees and skies would be less colorful and melodic, and our chances of discovering how human language evolved would be slim. So, when the story of the evolution of language is finally complete, be prepared to thank the birds.