Can the New Penn Center for Neuroaesthetics Unlock the Science of Beauty?
Neuroscientist Anjan Chatterjee is trying to figure out why we give a damn about Mona Lisa’s smile.
Late last year in New York, Christie’s auctioned off Edmond de Belamy, an artwork generated by a computer that had been fed 15,000 portraits painted from the 14th through 20th centuries. The rather fuzzy resultant print of a man in dark clothing — a joint venture between a student of artificial intelligence and two business-school grads — was heralded by some art-world insiders as marking “the infant stages of the next great art movement.”
Richard Lloyd, the Christie’s exec who decided to offer Edmond for sale, told the New York Times he admired the way the work echoed traditional European portraits: “It looks like something you’d expect Christie’s to sell,” he said, and noted that his house set a new record in 2017 when it sold a da Vinci painting for $450 million.
Edmond was estimated to go for between $7,000 and $10,000. Instead, it brought in $433,000.
Not everyone was thrilled by the sale’s success. Columbia University art historian Frédérique Baumgartner wondered about the questions of “intention and authorship” the AI-generated portrait raises. German artist Mario Klingemann compared it to “a connect-the-dots children’s painting.” Hidebound members of the art world scoffed: What could a computer possibly have to say about love and suffering and nature and beauty — about the human condition?
But one of Edmond’s creators, Gauthier Vernier, countered that the process corresponded to that of a human artist: “We’re looking at these portraits the same way a painter would do it. Like walking in a gallery, taking some inspiration.” The only difference: An algorithm creates the final work, instead of a painter with a brush.
The Christie’s sale raised the lid on a vexing question: What exactly is art? As Supreme Court Justice Potter Stewart said of porn, it seems we know it when we see it. But figuring out how and why our brains respond to it the way they do — that’s Anjan Chatterjee’s job.
A neuroscientist who studies and quantifies the brain’s responses to visual perceptions (among other stimuli), Chatterjee last year convinced Penn to let him set up a Center for Neuroaesthetics — the first of its kind in the country and perhaps the world. The aim, its website notes, is “to understand the neural systems that underlie aesthetic experiences and choices” by investigating “the nature and neural basis of beauty, art, design and architecture.”
Chatterjee, 60, was born in India, grew up there and in the U.S., and arrived in Philly in the year of the Bicentennial to attend Haverford College, where he majored in philosophy. He thought he’d be a primary-care physician, but at Penn Med, nothing really lit him up until he got to neuroscience. He went on to study its insights into the mind and came to realize that neuroscience offered possible answers to questions that philosophy posed about why humans gravitate toward beauty, enjoy it, and try to reproduce it. Art doesn’t feed us or clothe us or shelter us. So why do we care about it?
There are, he discovered, two main positions on the question (along with lots of other theories). One holds that while we have an innate, instinctual drive to produce art, it’s merely a decorative impulse. The other posits that our enjoyment of art is a by-product of unknowable evolutionary pressures in the distant past — a “spandrel,” as evolutionary biologist Stephen Jay Gould dubbed such leftovers, after the triangular space between the tops of two arches. Chatterjee wasn’t satisfied with either of these schemata. To him, art seemed too varied and diverse to just be decorative, yet too universal to be a by-product. After all, it triggers the same reward systems that respond to food and sex — and opioids. So he went looking for answers inside the brain itself.
“Love comes in at the eye,” Yeats wrote, and it’s truer than he could have known. Above the fireplace in my living room hangs a print of Claude Monet’s Vétheuil in Summer. I love the way the reflections of the clouds and poplar trees shimmer in the waters of the Seine. Inside his office, Chatterjee, who’s super-chill for a scientist, is explaining to me why they do.
Different parts of the brain, he says, process different visual perceptions. One area responds to small details and helps us distinguish what an object is; another part tells us where it is, based on what’s called “luminance contrast,” or brightness vs. darkness. Monet exploits this by making the different colors in his painting of the Seine equi-luminant, with hardly any contrast. This effectively disables the where part of the brain and deputizes the what part to try to compensate, which it’s not very good at. The result is the shimmer that makes the river seem to move.
What’s helped scientists pinpoint which parts of the brain take in what signals is fMRI, or functional magnetic resonance imaging, which measures the magnetic properties of the hemoglobin in blood cells. (Blood flow increases where neurons are active.) Other research methods include examining the functioning of people whose brains are damaged, and testing different regions with electronic probes. Such studies allow neuroscientists to, as Chatterjee puts it, “see how people think” — how the brain breaks sensory input down, then reassembles it.
Mark Rollins, a philosophy professor at Washington University in St. Louis, has described art as a game between two brains — that of the creator and that of the viewer. Even if artists don’t always know why we perceive things the way we do, they recognize that different effects create different responses and are able to put them to use. Consider Mona Lisa’s smile.
The 16th-century da Vinci painting that Guinness lists as having the highest known insurance value of any painting has been the subject of much discussion through the ages: What is that lady thinking? Is she smiling or not? Da Vinci, a student of optics and vision (“It was part of his scientific sensibility,” Chatterjee says), took advantage of the way our eyes process direct vs. peripheral vision to create ambiguity. He rendered the Mona Lisa’s eyes so the viewer sees lots of detail; her mouth is softer and fuzzier. If you look directly at the mouth, her expression seems neutral. Focus on the eyes, though, with the mouth in your peripheral vision, and her elusive smile appears.
Inside our heads, there’s a hierarchy to how we process visual signals (though it takes place at lightning speed). So-called “early” vision extracts simple elements like color, luminance, shape, motion and location, which register in various areas. In intermediate vision, different areas group some of these elements together and separate others out, to make sense of an otherwise overwhelming signal mass. Late vision, Chatterjee writes, “selects which of these coherent regions to scrutinize and evokes memories from which objects are recognized and meanings attached.” This is the point at which perceptions become pleasing, or frightening, or funny.
Until the invention of photography in the 1830s, most human art faithfully reproduced the world around us. (An exhibit of early photos now at the Barnes is titled, “From Today, Painting Is Dead.”) Once the camera assumed that responsibility, artists were free to explore a wider realm (see: cubism, surrealism, pop art) and stretch the definition of art so far that we now apply it to Andres Serrano’s photo of a tank of urine with a crucifix. Yet it was photography that revealed one of the first clues to just where beauty lies.
Back in the late 1800s, before “eugenics” was a dirty word, a cousin of Charles Darwin had a brilliant notion. Francis Galton, whose many hats included statistician, anthropologist, inventor and explorer, was intensely interested in heredity and improving the human race. He devised a method of blending images of the faces of criminals onto a single photographic plate. He was convinced that he could thus produce the ultimate “face of evil.”
To Galton’s surprise, the composite result proved unusually attractive instead.
Too bad for him, but good for neuroaesthetics. Galton’s exercise foreshadowed later findings that there are three components to a consistent, cross-cultural perception of human beauty: averageness, symmetry, and the exaggeration of features that distinguish women from men.
Take averageness. Looking at attractive faces activates regions in the brain that are engaged in the expectation of rewards and pleasure. (Think sex.) Why should a composite be more attractive than a single face? One argument is that the composite signals greater genetic diversity and thus better health in a mate. It’s why people of mixed race often seem particularly good-looking; when we see them, our brains register, “Deep, rich gene pool here!”
Natural selection, proponents say, doesn’t set out to favor genetic diversity. It doesn’t have goals; it’s a series of accidents. Eons ago, people who, thanks to their brain makeup, happened to be excited by genetically diverse potential mates maximized the survival potential of their offspring. That propensity got passed on more, and more, and more, until it was hardwired into humanity.
The same goes for those other two components. If you’re symmetrical, you’re not riddled with parasites. Full lips on a woman and wide shoulders on a man are caused by high levels of sex hormones. Those who took pleasure in these traits happened to have more kids who survived, who themselves had more kids, and on and on.
Chatterjee says we show a preference for these three characteristics all around the world, and from a shockingly young age. Babies just a few days old look longer at faces adults find attractive. One-year-olds play with pretty dolls nearly twice as long as ugly ones — way before culture can groom their tastes. The current human brain underwent most of its development in the Pleistocene Epoch, from around 2.6 million to 11,700 years ago. We don’t know for sure what evolutionary pressures were at work back then. But we can observe the palimpsest they’ve left in our brains.
What’s true for bodies also holds for landscapes. Whether young people grow up in New York City or the steppes of Russia or the Amazon rain forest, they prefer a certain type of landscape — a gently sloping grassland dotted with trees. That happens to describe an African savannah. The preference exists whether or not you’ve ever seen an actual African savannah. We’re drawn to that environment as surely as we are to averageness or symmetry in faces, and for similar reasons: It offers water, trees to climb and hide in, nourishing vegetation, changes in elevation, open space in which to see predators coming — all factors that would have helped our ancestors survive.
Monet’s Vétheuil in Summer has these elements, too.
Not every scientist buys into this. The New York Times Magazine recently featured a Yale ornithologist, Richard Prum, who argues that animals prefer fancy feathers and mating calls simply because they’re pretty, without any relevance to survival. The article’s author, Ferris Jabr, termed the African savannah theory “ludicrous.” But Jabr also acknowledged that Prum’s “indifference to the ultimate source of aesthetic taste” is problematic. Even if our preferences for beauty are arbitrary, they exist. Chatterjee would argue that there has to be a cause — one that’s rooted in neurobiology.
Chatterjee’s probing of the brain has convinced him that our aesthetic judgments rely on the “core triad” of sensation, meaning and emotion. There’s no discrete area of the brain that registers beauty; the electrical impulses generated when I look at my Monet activate the same brain structures involved when I eye a cheesesteak. And they’re rewarded the same way — with pleasure hormones like dopamine.
For most of human history, Chatterjee notes, artists were anonymous contributors to joint cultural efforts: Byzantine icons, Romanesque cathedrals, Dogon masks, China’s terra-cotta warriors. Their unsigned work highlights another role of art in human life: its use in rituals that solidify societal bonds. The concept of the artist as a solitary (often tortured) genius is quite recent. Naturally, our opinions on what is or isn’t art are affected by cultural context (see: Serrano’s Piss Christ). Our brains’ responses to art are apparently infinitely flexible and adaptable and are strengthened by exposure and knowledge. Architecture students show a greater neural response to buildings than to faces.
Like Mona Lisa’s smile, how we view art depends on how we’re looking at it. In his 2014 book The Aesthetic Brain, Chatterjee writes: “When we emphasize the universality of art, we slide into thinking of art as an instinct. When we acknowledge the sheer diversity and cultural fashioning of art, we slide into thinking of art as a spandrel.” He offers a third way of envisioning art’s role, using the example of a small Asian bird.
Like many birds, the male munia sings to attract its mates. Several hundred years ago, Japanese bird breeders began to selectively mate wild munias for colorful plumage, resulting in a domestic bird called the Bengalese finch. Through hundreds of generations, the controlled breeding made the finch’s song irrelevant to its mating success.
You might expect the song to wither and die under such circumstances. Instead, the finches’ songs grew increasingly variable and more complex. In the absence of evolutionary forces that drove the birds to sing a certain way, their songs — their art, if you will — became freer, more improvised.
It’s precisely in the relaxation of prehistoric evolutionary pressures to adapt and survive, Chatterjee says, that human art has flourished. Our brains rely for our appreciation of it on systems that evolved to process sensations, emotions and meaning in other contexts entirely: social cohesion, natural selection, environmental safety. Once we were sufficiently at ease in our surroundings, we could create the Sphinx, Homer’s poems, Michelangelo’s David, the Mayan murals at Bonampak …
And, at Penn, ENIAC, the “Giant Brain” unveiled to the world in 1946 that kicked off the Computer Age.
Which brings us back to Edmond de Belamy. To most of us, art generated by computers is clearly less — less real, less important, less meaningful, dammit — than art made by humans. But is it fundamentally any different from a da Vinci that exploits our peripheral vision?
“This is a wide-open question,” Chatterjee says, leaning back in his chair. But Edmond does show that aesthetic experimentation can generate big bucks. And that provides a clue to why Penn — a nonprofit, granted, but hardly a charity — gave Chatterjee the nod to open his new center. He mentions an experiment he ran in which subjects look at two faces on a computer screen — one attractive, one neutral. There are numbers next to each face — say, 21 by the hot face, 27 by the so-so one. The screen shows a third number beneath the faces — 25, perhaps. The participant is supposed to move the cursor next to whichever choice the 25 is numerically closer to; the faces are irrelevant. But Chatterjee found that if the 21 is beside the attractive face, the participant’s hand actually slides in that direction before correcting and moving to the other side. “Our limbs themselves are more attracted to beauty,” he says.
There are obvious implications for any website where you buy stuff by clicking. Chatterjee also brings up our universal conflation of beauty with “goodness.” Attractive people are perceived to be more intelligent; they’re given lesser punishments, higher pay, better grades. People with facial abnormalities suffer because of this bias, which is more pronounced in men than in women. Research shows that when we see such anomalies, the parts of our brains that govern empathy become less active. You can understand the evolutionary impetus for this: Don’t mate! But such holdovers have unfair consequences. For instance, men occupy more positions of power and make more hiring decisions. If they don’t realize they harbor this bias, they can’t correct for it. “These triggers have deep roots,” Chatterjee acknowledges, “but you have a huge frontal lobe that can counteract them.”
The center, housed in Penn’s Goddard Laboratories, is still a work in progress, but Chatterjee is attracting scholars from all over the world. He’s drawing on Penn’s disparate strengths — a professor of plastic surgery, Wharton faculty who study consumer decision-making, the school of architecture, a psychologist studying Tanzanian tribes that haven’t been exposed much to our cultural overlays. He wants to partner with institutions, like the Barnes, that complement the center’s agenda. To set up an artist-in-residence program. To form community connections to help make neighborhoods more beautiful and not just tidier. To explore mixtures of emotion like “the sublime,” that catch-your-breath sensation we experience inside a mighty cathedral: “Beauty plus grandeur plus anxiety — that reflection of our existential insignificance … ”
Um. There are so very many problems in our world. Does he ever get pushback that studying aesthetics might be a little … frivolous?
“All the time,” Chatterjee says promptly. “Funding is hard to get. But if you ask people, they say beauty gives meaning to their lives.” There’s also, he notes drolly, no mass market for cosmetics or products or surgery to make us look wiser or more sincere.
Fair enough. But it’s not hard to envision a future in which computers and algorithms blowtorch the human part of art that reaches out to us across time and space. Look at what analytics have done to baseball, for chrissake.
“I’m a basketball guy,” Chatterjee says. “I’m not sure analytics have ruined sports.” But, he acknowledges, too much focus on quantifying the sensation part of the triad could blunt our emotional response to art.
Then again, he says, consider what happens when an artwork is discovered to be fake: “It may be just as good as a Rembrandt or Vermeer, but once it’s known as a forgery, it loses worth.” He mentions a Yale psychologist whose experiments with children have shown “a basic, chemical core desire in the developing brain for authenticity. Another human being produced this to communicate something — it’s still a big piece of what people value.”
Only time will tell where AI and algorithms take our aesthetic senses. We like to think of art and beauty as mystical wonders. But we could so easily have turned out not to notice sunsets or admire the shape of an ankle. There’s wonder in the long, blind crawl of evolution, too.
Published as “You Call That Art?” in the March 2019 issue of Philadelphia magazine.