麻豆传媒

Test yourself: 鈥�What emotions are written all over their faces?鈥�

IN 1868, while working on his latest book on evolution, Charles Darwin liked to show visitors to his house a series of ghoulish photographs of people鈥檚 faces.

The pictures, taken by French physiologist Guillaume-Benjamin Duchenne, showed people whose facial muscles were being zapped by electric shocks, contorting them into strange and often eerie arrangements. Darwin was fascinated by how a twitch of the mouth here or furrow of the brow there could conjure up the impression of an emotion 鈥� fear, say, or surprise. He wanted to know whether his guests perceived the same emotions in the pictures. They usually did.

Darwin came to the conclusion that emotional facial expressions were universal: people all over the world made the same ones, and could easily and automatically recognise them in others. He didn鈥檛 claim to know what the expressions were for 鈥� he thought they were probably 鈥渘ot of the least use鈥� 鈥� but he did suggest that they were innate and rooted in our shared ancestry with other animals. He presented this argument to the world in his book , published in 1872.

Darwin wasn鈥檛 the first scientist to investigate the meaning of facial expressions, but his enormous influence brought them to wider attention and initiated a debate that has ebbed and flowed ever since. Are emotional expressions universal and innate, or do they vary from culture to culture?

By the late 20th century, the pendulum appeared to have swung decisively in favour of Darwin鈥檚 view. But today the debate is alive and well again: while some researchers are claiming we are close to a full understanding of emotional expressions, others argue that we are nowhere near. This is much more than an intellectual row: it has implications for our concept of human nature and even emotions themselves.

The modern orthodoxy on emotional expressions was largely secured by a classic experiment that took place a century after Darwin was spooking his visitors. In the late 1960s, a team of psychologists led by Paul Ekman of the University of California in San Francisco travelled to Brazil, Japan, Borneo and New Guinea and showed people photographs of six stereotypical emotional expressions: happiness, fear, anger, surprise, disgust/contempt and sadness. They also did the experiment at home in the US.

Winning smiles

Ekman鈥檚 team found that everybody they tested, regardless of culture, recognised the same six basic emotions 鈥� even people in Borneo and New Guinea who鈥檇 had no prior contact with the outside world. 鈥淔or most of these people, I was the first outsider they鈥檇 seen,鈥� Ekman recalls. The research, published in , strongly supported the idea that expressions of emotions are common to all people, regardless of culture, because they have a common evolutionary origin.

Since then, dozens of studies have seen similar effects. In addition to the basic six (or seven, with contempt and disgust sometimes treated separately), the model has been extended to include pride, indicated by a tall posture and puffed-up chest, and shame, with a downturned head and bent posture. This all supports the view that emotional expressions are hardwired into the human brain.

Other support comes from studies of people who were born blind and have therefore never seen an emotional expression. For example, psychologist David Matsumoto at San Francisco State University analysed the facial expressions of judo contestants in the 2004 Olympics and Paralympics, including athletes who were blind from birth or later became blind. He found that all three groups produced the same faces when they won a bout. The expressions included so-called Duchenne smiles 鈥� big, beaming smiles involving the eyes as well as the mouth 鈥� which are considered to be authentic expressions of happiness. 鈥淭he evidence for the universality of facial expressions of emotion is overwhelming,鈥� says Matsumoto.

Some researchers are now talking about tackling two big questions that Darwin and Ekman ducked: what are emotional expressions for, and how did they evolve?

In a recent review paper, Jessica Tracy and Azim Shariff claim that we are close to answers. New research 鈥渁llows us to explain why facial expressions look the way they do,鈥� says Tracy, a psychologist at the University of British Columbia in Vancouver, Canada ().

The starting point is that emotions serve a biological function. An emotion like fear is a 鈥渃ompendium of different bodily responses鈥�, says Shariff, a psychologist at the University of Oregon in Eugene. The heart begins to race, the breath quickens and blood surges to the legs. All of these can be explained as preparation for fight or flight.

Facial muscles also play their part. When people make a classic fear expression, they increase their field of vision, make more rapid eye movements, and open their airways 鈥� all of which would allow them to better monitor danger and respond.

Plausible functions have also been proposed for other emotions. The scrunched-up expression of disgust, for example, may serve to constrict airways to stop the entry of contaminants, while the cringing posture of shame may hide vulnerable parts of the face from attack.

Not all emotional expressions have obvious functions, however. The original biological purpose of happy smiles, angry scowls or sad frowns have so far eluded psychologists.

But the origin of facial expressions goes beyond simple physical responses. Humans are social animals who need to communicate, and facial expressions are a very powerful way of doing so. Being able to transmit and receive emotional states would have been advantageous to our ancestors. For example, displaying fear and reading it on another face helps both of you to respond to danger.

鈥淗uman are social animals, and facial expressions are a very powerful way of communicating鈥�

In this scenario, emotional expressions started out as something evolutionary biologists call a 鈥渃ue鈥� 鈥� they revealed information about an inner state or behaviour but they didn鈥檛 evolve for this signalling purpose, much as chewing is a reliable signal that somebody is eating. Then over time they evolved into signals for expressly conveying information.

Exaggerated and distinctive

鈥淚t鈥檚 a process called ritualisation,鈥� says Tracy. Expressions became more exaggerated and distinctive, to make it easier to communicate non-verbally. This stage of the evolution of expressions represents what鈥檚 called an exaptation 鈥� a feature that initially evolved for one purpose but is co-opted for another.

This process may explain why it is hard to discern a function for some expressions: the original purpose has long been masked by ritualisation. Another possibility is that some emotional expressions only ever served a signalling function. Pride and shame, which are particularly social emotions, are likely candidates. The expressions resemble the dominance and submission postures of other social primates, suggesting they are signals of status inherited from distant ancestors.

Despite the progress, the theory of emotional expressions is not complete. And even as researchers work on it, others are challenging the entire enterprise.

Some question the methods used to demonstrate universality. For example, study participants are often given a list of emotions and asked to pick the one that best matches the facial expression they see. Ekman鈥檚 influential research was done this way.

James Russell, a psychologist at Boston University, has argued that providing such prompts artificially inflates identification of the 鈥渃orrect鈥� emotion. If subjects know that they are trying to identify happiness, sadness, anger and so on, then that is probably what they will see. When people are asked to come up with their own words, they find it much harder to hit the target. In one experiment, removing the list reduced accuracy from over 80 per cent to about 50.

Lisa Feldman Barrett, a psychologist at Northeastern University in Boston, has expanded Russell鈥檚 critique. In a response to Tracy and Shariff鈥檚 review, she argues that the dogma of universal emotional expressions has been accepted too uncritically and that the findings can be explained without it. She suggests that, rather than being biologically based, they are culturally learned symbols 鈥� a form of 鈥渂ody language鈥� that we learn to communicate emotions to others. And, like spoken languages, the expressions share commonalities but also vary from culture to culture ().

Experiments in Barrett鈥檚 lab and elsewhere find that people鈥檚 recognition of an emotion depends heavily on context. In the real world, faces are rarely seen in isolation. Posture, voices, other faces and the wider context are also available for inspection, and they influence how expressions are perceived. For instance, a scowl 鈥� usually associated with anger 鈥� can be perceived as disgust if the person is holding a dirty object, or fear if it鈥檚 paired with a description of danger. The disgust face can even be seen as pride if attached to a body with arms raised in triumph. Similarly, viewing the same expression labelled alternately with the words 鈥渁nger鈥�, 鈥渟urprise鈥� and 鈥渇ear鈥� changes how people perceive it ().

Barrett likes to further illustrate her point with a picture of tennis star Serena Williams. When you can only see her face she appears angry or in pain (see above). But when you see her body too 鈥� on the front foot, racket in hand, fist clenched in front of her face 鈥� it is obvious that she is actually delighted (in fact, she has just beaten her sister Venus in the quarter-finals of the 2008 US Open). 鈥淎 face does not speak for itself,鈥� says Barrett.

It also turns out that, contrary to Ekman鈥檚 classic paper, emotional expressions are not culturally invariant. 鈥淭here are differences,鈥� says Rachael Jack at the University of Glasgow, UK. In a study published in April in the , she used a graphics package to generate thousands of expressions by randomly combining facial muscle positions.

Her team generated 4800 faces and showed them to 15 European and 15 Chinese volunteers. Their task was to categorise the faces as expressing one of the six basic emotions or to say 鈥渄on鈥檛 know鈥�, with no predetermined correct answers. The Europeans (who were shown European-looking faces) reliably sorted the expressions into clusters representing the six basic emotions, with high levels of agreement between them. But the Chinese (who saw east Asian faces) produced much more overlapping categorisations, and disagreed much more.

Jack is now looking more closely at how Chinese people interpret emotional expressions. Rather than asking them to choose from among six, she will have them label faces with their own words to see which groupings naturally emerge.

The results suggest that culture does have an influence. 鈥淚 wouldn鈥檛 disagree with people who suggest that there is a biological origin to certain facial expressions,鈥� Jack says, 鈥渂ut people have had culture for about 80,000 years.鈥� These once-hardwired signals have been extensively reshaped by cultural evolution for use in social communication, she says, allowing regional variations to arise.

So if emotional expressions turn out to be less universal than Ekman and others claim, what鈥檚 the alternative? Barrett鈥檚 position is that when we observe emotions in others, the categories we use are culturally constructed, learned and context-dependent. 鈥淓motion recognition isn鈥檛 so much recognition as it is perception,鈥� she says.

In visual perception, incoming information is combined with 鈥渢op-down鈥� knowledge based on experience and expectations. That is what allows us to effortlessly categorise objects and interpret their meaning. Barrett argues we do the same with emotional expressions. We easily categorise them into 鈥渁nger鈥�, 鈥渟adness鈥� and so on, but we should not mistake this for evidence that those categories are innate.

There are practical implications for this seemingly esoteric debate. If expressions are innate and universal, then they can be used to read people鈥檚 inner states like a book. And that is potentially useful information.

Matsumoto heads a company called Humintell, which provides workshops and training programmes to police departments, government agencies, businesses and individuals who want learn to read other people. The training emphasises reading subtle expressions and microexpressions 鈥� facial expressions that pass across the face in a fraction of a second, which he says can be signs of concealed emotions.

Human nature

Similar ideas have made their way into training programmes for the FBI and the US Department of Homeland Security. But if the production and perception of facial expressions is much more fluid and contextual, then such programmes could be of limited value.

The debate also raises more fundamental questions. If emotional expressions are universal, they give us insights into something etched deep in the biology of all of us 鈥� something you might call human nature. Ekman鈥檚 1969 research is often put forward as some of the best evidence we have for a universal human nature.

But if expressions are learned and contextual, that evidence is significantly weakened. Variation in the way people express and perceive emotions isn鈥檛 a phenomenon to be explained away 鈥� it鈥檚 the fundamental phenomenon to explore.

Scientists on both sides of the debate agree on one thing: more research is needed to tease apart the roles of biology and culture in emotional expressions. 鈥淧eople love to think in black and white, but it鈥檚 really not like that,鈥� says Matsumoto. He points out that there are many things we do with our faces that aren鈥檛 about conveying emotion, such as raising your eyebrows when greeting someone. Those should be characterised as well. 鈥淚 think that we鈥檝e just started to scratch the surface on the complexity of faces.鈥�

At issue is not just the meaning of expressions. They鈥檙e one arena in a larger debate over emotions themselves. A 2010 survey of emotion researchers found widespread disagreement over the meaning of the word 鈥渆motion鈥�. The study prompted Peter Zachar, a philosopher of psychiatry at Auburn University at Montgomery, Alabama, to suggest that the word may eventually meet the same fate as 鈥済ene鈥� in biology 鈥� a once-useful concept that has become less meaningful as science has uncovered ever-greater complexity. The larger question is not just how emotions get written on the face, but what they are.