Can you tell if a man is dangerous just by looking at his face?

The state of the universe.
Oct. 14 2009 10:36 AM

Facial Profiling

Can you tell if a man is dangerous by the shape of his mug?

Illustration by Charlie Powell. Click image to expand.

On Nov. 27, 2008, Indian police interrogators came face to face with the only gunman captured alive in last year's bloody Mumbai terror attacks. They were surprised by what they saw. Ajmal Kasab, who had murdered dozens in the city's main railway station, stood barely 5 feet tall, with bright eyes and apple cheeks. His boyish looks earned him a nickname among Indians—"the baby-faced killer"—and further spooked a rattled public. "Who or what is he? Dangerous fanatic or exploited innocent?"wondered a horrified columnist in the Times of India. No one, it seems, had expected the face of terror to look so sweet.

The notion that a man's mug reveals his character is an age-old bias. Since Aristotle, people have thought it possible to infer personality traits from the face and body, an art known as physiognomy. The practice grew popular in the years after the American Revolution, when a Swiss enthusiast published a series of illustrated pocket guides to help readers interpret faces on the go. Soon, it was plain to everyone that a man's greatness was prefigured in his face. (George Washington's big schnoz, for example, signaled strength and foresight.) Over the next 150 years, a gang of enterprising physiognomists set about using the new "science" to identify society's bad apples, too.

In the late 19th century, the Italian criminologist Cesare Lombroso ran autopsies on convicts and cataloged features that might identify "born criminals," such as jug ears and overdeveloped canines. In the 1930s, Harvard's Earnest Hooton examined 14,000 prisoners and observed that first-degree murderers tended to have straight hair, while the hair of second-degree murderers was unusually golden. A few years later, Columbia psychologist William Sheldon studied delinquent youth and invented a human taxonomy consisting of three types—ectomorphs (thin-faced, skinny, brainy), mesomorphs (broad-faced, muscular, aggressive), and endomorphs (round-faced, fat, sociable). He further divided these groups into 88 subtypes named after animals, such as the Herons (very often Phi Beta Kappas, he wrote) and the Foxes and Coyotes (Jesus Christ's type, per Sheldon). Overall, he concluded that the meaty-faced mesomorphs were most prone to criminality.

Much of this work fell apart under scrutiny. Lombroso's statistical methods stunk. Hooton chucked data that didn't fit his hypothesis. Sheldon had not examined very many criminal delinquents, and no one much understood how he distinguished a Great Cat (such as King Arthur) from a Great Saber-tooth Bobcat (e.g., Bronko Nagurski) in the first place. In the wake of the Nazi death camps, theories of "criminal anthropology" fell from favor, and researchers emphasized social explanations for behavior.

But today, physiognomy is making a comeback. In the last decade, breakthroughs in 3-D modeling and animation software have opened up the field. At the same time, ideas from genetics and evolutionary psychology are reanimating old debates about biological determinism, race and gender differences, and why humans possess the faces and bodies that we do.

The new research suggests we are more skilled at "reading faces" than we knew. People are surprisingly adept at assessing sexual orientation from headshots. Five-year-olds can predict election outcomes based on photos of the candidates. We can even guess whether a face belongs to a Democrat or a Republican at a rate better than chance, according to a forthcoming study out of Princeton.

Now some of the "new physiognomists" are resurrecting an old claim: that you can gauge a man's penchant for aggression by the cut of his jib. Last fall University of California-Santa Barbara psychologist Aaron Sell reported that college students could accurately estimate the upper body strength of unfamiliar men after viewing their faces alone. (The men's necks were obscured.) The students did equally well with fellow undergraduates and men from South American indigenous groups—all of whom had had their strength measured using gym equipment. Interestingly, the toughest-looking undergrads also reported getting in the most fights. Another study by Sell suggests that such formidable men are more prone to use violence—or advocate military action—to resolve conflicts.

Many animals employ similar systems. Male orangutans grow fatty cheek pads that reflect group status. Lions with long, dark manes tend to rule the pride. From an evolutionary perspective, these advertisements may be a convenient way of saying, "Hey bro—btw, I can kick your ass" without having to go through the risk of combat.

So which features might hint at belligerence? Sell suspects the brow ridge and jaw, two structures that are shaped by testosterone in puberty. (High testosterone has been linked with masculine looks as well as with aggression.) Other scientists propose a different measure: the width-to-height ratio of the face, as measured from cheek to cheek and lip to brow. Last year, a team of Canadian psychologists showed that men with wider faces (think Ernie) score higher in lab tests of aggression than slender-faced men (think Bert). They also found that wide-faced hockey players rack up more penalty minutes. Now, two studies in Psychological Scienceone from August and another forthcoming—reinforce the notion that stout-faced men appear tougher and are more likely to behave in aggressive and untrustworthy ways.

The idea is not far from what William Sheldon proposed in the 1940s.   Certain excitable meat-headed mesomorphs, he wrote, were prone to "muscular unreasonableness" and could be "as destructively dangerous in a human setting as a great gray owl in a colony of Snowshoe rabbits." (These, of course, were the Great Gray Owls.) Sheldon suggested not only that mesomorphic men were prone to aggression but that they were more likely to be criminals.   So, was he right?

The psychologist Robert Deaner recently went through the database of mug shots of the Michigan Department of Corrections and measured the facial width-to-height ratios of 688 white convicted criminals. He classified each inmate as either violent or nonviolent based on FBI standards and compared their faces with their crimes. There was no difference in the average width-to-height ratio of violent vs. nonviolent offenders, he says. However, Deaner did find that prisoners' faces were significantly wider than a population of undergraduates. Curiously, pro hockey players had even wider faces than the prisoners. His conclusion: "Face width does not predict violent crime. … [W]e believe the cliché, 'never judge a book by its cover' remains sound advice."  

We seem to have a strong tendency to ignore that advice. A number of studies have demonstrated that most people hold similar stereotypes about what criminals look like and believe that "the face fits the crime."   This can play out in court: The psychologist Leslie Zebrowitz has shown that "mature-faced" defendants are more likely to be found guilty of certain kinds of crimes. And when baby-faced defendants are found guilty, they tend to get more lenient sentences. She calls this form of discrimination "face-ism" and argues that defendants shouldn't be required to show their faces in court. But if it is proved that the male face does indeed reveal "honest" signals about aggressiveness, jurors might deserve access to that information. (Then, too, defense attorneys might want to adopt a novel legal strategy: the meathead defense. "My client can't be blamed for his actions because he suffers from high testosterone. Just look at his face!")

To be clear, no one is saying cops should start rounding up meatheads. But it wouldn't be unprecedented in the annals of physiognomy; Earnest Hooton, an avowed eugenicist, proposed permanent lockup for people of poor genetic stock. A major aim of William Sheldon's work was to identify for eugenic elimination that which he termed "PPPP": poor protoplasm, poorly put-together.

Today's physiognomists share none of these racist attitudes. But racial issues inevitably emerge from the science. For example, if there are racial differences in upper body strength, then Aaron Sell's theory predicts there might be racial differences in aggression. (Jon Entine, author of Taboo: Why Black Athletes Dominate Sports and Why We're Afraid To Talk About It, claims that what evidence there is on the topic suggests that Eurasian whites have the strongest upper bodies.)

Bias may also seep into the lab. Study participants may be responding to subtle racial cues, for example, when a psychologist asks them which of two faces looks more "dominant." This kind of feature-based discrimination does occur: A 2004 study (PDF) showed that people with more Afrocentric facial features receive harsher prison sentences— even when the individuals are white.   In this respect, racism and face-ism are one and the same.

Physiognomy faces other tough questions. For example: How do we know that tough-looking dudes don't act tough simply because society treats them like thugs? How do we know that a penchant for aggression doesn't also incline men to pump iron, which alters their bodies and their faces? In any event, even if it is proved that manly-faced men tend to be naughty by nature, a policy of facial profiling won't guarantee our safety. India's baby-faced killer—and the many baby-faced killers who have come before him—would seem to prove the point.

David Merritt Johns is a doctoral student with the Center for the History and Ethics of Public Health at Columbia University’s Mailman School of Public Health.