Before he chose science over dance, Erich Jarvis took a moonwalk into magic. “I wanted to be a magician,” he says. “My cousin Sean and I put on shows, picked locks, invented tricks. Looking back on it now, I realize that my curiosity about how things work sort of led me into science.”

So let’s imagine his new lab at the Rockefeller University, where he’s unpacking boxes and firing up his enigmatic contraptions, as a modern-day magician’s workshop. And on this mild, sunny Wednesday in Manhattan, with the East River winking at us through the windows, we can slip behind the curtain and hear how the magic gets made.

Songbirds do it. Bats, elephants, whales, dolphins, and sea lions do it. Humans do it, too. We learn to sing. In nature, this is risky, because singing burns energy, and songs that court a mate alert the predator as well: Here I am. I can sing.

Singers pay the price and take the risk because it’s in our code. But in nature, we’re rare. Sure, a dog can learn to recognize some words, maybe croon a reply. But only vocal learners, as Jarvis calls us, truly learn to sing. We announce our presence, defend our territory. But mostly, we’re singing for love.

So the science of vocal learning wears the sheen of romance. But behind the curtain, the mechanics get murky. You can’t pick the locks of genetics and neurobiology without meticulous, mind-bending work. And that’s where Erich Jarvis thrives.

Over the course of two decades, Jarvis, with colleagues and students, has used tools such as molecular mapping to study the circuits that fire in the brain as a bird hears a tune, recreates the pattern, pulses signals to its brainstem, and jolts its feathered body into song. The goal of this work is bodacious enough for Houdini: Jarvis plans to engineer, in the brain of a non-learning species, the circuits of vocal learning. He expects to try this first in a mouse or a chicken, and work his way up to a non-human primate someday.

As sensational as that prospect might seem, for Jarvis it’s all about medical science, not stagecraft. He isn’t rigging up a talking mouse or an ape that can sing for its supper. Vocal learning is the “substrate,” he says, on which language is built. If he can engineer that substrate into a brain where it didn’t exist, then maybe he can help restore some human voices silenced by disability or disease. But for now he needs to understand birds, and what they can say about us.

For years, the lead vocalist in the Jarvis lab has been the zebra finch, a boisterous Australian songbird that earns its chops as a youngster, tutored by its kin. But not every species of bird displays the same knack. Why do some birds learn to sing and others don’t? And why do some, like the parrot, also learn to imitate the human voice and talk?

Jarvis couldn’t answer such questions without parsing the genomes of many kinds of birds. So he ventured out to the wild frontier of genomics and big data, connecting with a group called Genome 10K, an international consortium that aims to sequence the genomes of 10,000 vertebrate species. Before long, Jarvis was co-leading the bird work. He wasn’t a bird guy, but he needed the data. “So I kept pushing,” he says.

Applying the considerable brainpower of more than 200 scientists from 20 countries the consortium sequenced the genomes of 45 bird species and claimed serious ink in more than 50 publications, including Science, which in 2014 published no fewer than eight articles based on the work. The project reconfigured a goodly portion of the avian family tree and confirmed that vocal learning evolved at least twice among birds – once in hummingbirds and once in a common ancestor of songbirds and parrots. 

As the effort continues, Jarvis is angling for better tools to sequence genomes faster, cheaper, and in greater detail. But can the project really crank its way through 10,000 genomes? “I have no idea,” says Olivier Fedrigo, director of the Vertebrate Genome Laboratory and a member of the Jarvis team. “So far, there aren’t enough instruments in the country to do this. But instead of waiting, we’ll get started, and in two years maybe we’ll have the technology we need.”

This is quintessential Jarvis: No pausing, no waiting around. He learned his management style at home, growing up in Harlem, Queens, and the Bronx. “I had to look after my brothers and sisters because often we didn’t have money for daycare, especially after my mother divorced. So I was a little kid looking after a bunch of little kids. That experience taught me how to manage people, and how to try and take care of people.”

To take care of himself, he dances, because dance helps him balance the pressures of work. Before he left Duke University – where he invested 18 years, ran a high-profile lab, and earned tenure – he performed with the Cobo Brothers Dance Company, a sassy little salsa troupe. In New York he’ll keep dancing, for pleasure and self-preservation. And each time he dances, he’ll embody his work, because dance is the brain’s body-double for song. 

“We think vocal learning evolved by brain pathway duplication in songbirds and in humans,” Jarvis says. “So you take a pathway that all animals use when they learn to walk or fly, duplicate it, and then hook it up to organs that produce the sound.”

The genetic changes that made this possible may have invaded the adjacent motor-learning circuits. Only vocal learners can synchronize their body movements to music, Jarvis says. “The purpose wasn’t dance, but it has given us the ability to dance.” And we hear what we utter in real time, so we can tailor our tunes on the fly. In courtship, swapping songs and adding complexity grabs attention – an insight that chirped into mind as Jarvis sat in a garden to read, distracted by a songbird that kept changing its tune.

But along with the singing and dancing, some of us learn to speak. How did that happen? In parrots, Jarvis says, a second brain pathway duplication seems to have created a “shell system” around the vocal-learning core, and the two together may have endowed parrots with the ability to imitate human speech. The feature is a clue, perhaps, to how human language might have evolved, through layered duplication. 

When he dares to speculate this way, some of his critics take umbrage. “We get serious pushback from people who think birds are too different from humans to teach us anything about the human brain,” Jarvis says. “And yes, of course we are very different animals. But in our lab, we see convergence, and we can learn from the convergence.”

Convergence. In evolutionary biology, the term refers to distantly related species evolving similar traits under similar conditions. Bats and birds, for example, both developed wings. In the case of vocal learning, convergence is not superficial. Hundreds of specialized gene expressions in the brains of songbirds also appear in the brains of humans, and they are notably absent in creatures that lack vocal learning.

While the lab has found strong convergence in the motor functions that drive the vocal organs, other parts of the pathway in humans remain hidden for now. Gregory Gedman, a graduate student in the Jarvis lab, is searching through the songbird brain region that may be convergent with human Broca’s area, a region of the frontal lobe that’s been linked to language.

Meanwhile, Jarvis pursues the big whys and what ifs. What if vocal learning is not a trait we gained but one we preserved? What if other species dropped it because full-throated courtship was getting them killed? Did hundreds of specialized genes evolve independently in birds and humans, or did a master gene summon and conduct these orchestras of genetic virtuosity? And what if the rudiments of vocal learning are present in non-learners? Mice, for example, have neurons that link the motor cortex directly to the larynx, but the connections are skimpy compared to our own.

Jarvis wants the answers, and he’s building a team to pursue them. He’ll keep pushing toward his big, audacious goal, with support from HHMI, where he’s been an investigator since 2008. At Rockefeller, where he earned his PhD, he’ll reopen a field station where his mentor, Fernando Nottebohm, pioneered the neurobiology of vocal learning in birds. The stakes are high, and not just for the potential applications in medicine. Speech may be entwined with human consciousness, Jarvis says, so learning its secrets might help us fathom who we are.

And who are we? This is the sort of question Jarvis’s father pursued to the near edge of madness before he was shot and killed, apparently at random, in a park in Washington Heights in New York City. James Jarvis, once a prodigy in music and science, lived for years in a cave in Inwood Hill Park, curating his collection of fossils and rocks, and watching the stars. 

The son has blazed a different trail, but the father’s question persists: who are we? This is where Erich Jarvis’s scientific quest converges with his personal quest. He has interviewed hundreds of relatives, calculating degrees of racial and ethnic mixing. With the peacemaking impulse his mother instilled, taken from the example set by Martin Luther King Jr., Jarvis would like to resolve some disputes about identity that have rankled his family for ages.

And what does his racial identity mean to him, as a scientist, beyond the ever-present obligation to serve on committees? He is, without doubt, a success. But sometimes he still has to fight the implied accusation, like a hum of static in the culture around him, that because he’s a person of color he is somehow less than.

On his computer, he displays a genetic self-portrait – a chart of his personal genome, color-keyed to the dozens of regions from which it arose. Jarvis has commissioned the sequencing twice, with similar results: equal parts European and African, with a few threads of Native American. The graph says as much about us as it says about him, loomed as we are in the warp and weft of our genes. And it speaks to what he values in science, where diversity is not just a buzzword but a necessary condition of success. He aspires to engineer in the culture around him a substrate on which new ways of thinking will grow. 

“We will erase the borders and make the world our lab,” he says. “We’ll make it diverse and inclusive, because in that mix of cultures you find your best ideas.”

In that vision, even a white guy finds room. Wayne Kuenzel, a soft-spoken poultry scientist from the University of Arkansas, might seem a world apart from his younger, hipper host. And yet here Kuenzel is, in the Jarvis lab, searching a chicken brain for signs of a feature Jarvis found in the zebra finch. The men met at Duke in 2002, when Jarvis co-directed a forum that revised and brought more accurate order to the nomenclature of the avian brain, and added to our understanding of vertebrate brain evolution. They’ve been in touch ever since.

“Erich is a leader,” Kuenzel says. “He’s progressive, always looking for a challenge. And he’s always been fair.”

So Jarvis keeps faith with his mother’s belief in a King-like inclusion. He listens, he takes care of people, and he leads. But sometimes, when he feels pushed toward a cage that feels less than, he consults Malcolm X on the theme of achieving a goal “by any means necessary,” as long as no one gets hurt. So he’s willing to put himself out there, to raise his voice above the leafy treetop. Here I am. I can sing. ■

Story by Neil Caudle
Photography by Peter Ross
Banner title art credit: Wikimedia