How easily can we read a face?

Who's the most famous living psychologist? Right up there would have to be Martin Seligman, Daniel Kahneman, Steven Pinker, Jonathan Haidt, Aaron Beck. But only one psychologist has a prime-time TV drama based on his work...Yes, it's Paul Ekman, the world expert on facial expressions.
Ekman made his name in the 1970s and 1980s with his work on the basic theory of emotions, which suggested that emotional experience can be divided into six discrete emotions - anger, disgust, fear, joy, sadness and surprise - each of which has a correspondent facial expression. And these expressions, Ekman has suggested, are universal and cross-cultural. So emotions are, according to the Ekman school, innate, universal and embodied in set physical responses. One could say that emotions are facial responses - our face takes on the universal shape of the emotion.
Ekman's theory of emotion recognition has become widely accepted. It's even taught in most English state schools today, as part of the subject Social and Emotional Aspects of Learning, in which children as young as three are shown various faces and taught to shout out 'disgust!' or 'surprise!' or whatever emotion is being taught that day. Though if these emotional responses are innate and universal, why do school-children have to be taught to recognise them? Shouldn't they just know?
Recently, Ekman has taken his work further, suggesting that you can tell if someone is lying by their facial expressions, movements and language. His work got picked up by Malcolm Gladwell, the Oprah Winfrey of psychology research: once he's mentioned you, you're made for life.
Gladwell's article on Ekman in the New Yorker led to Fox TV making a prime-time series about a professional lie detector, called Lie To Me, in which Ekman is played by Tim Roth. Ekman admits that the show is somewhat fast and loose with his research (the first episode revolves around the unproven idea that touching your nose is a 'tell' that you're lying), but Ekman writes a column for the show, so he can't mind that much.
Ekman's work on lying has also led to what must be a highly lucrative contract with US airport security, to work on a controversial programme called Screening Passengers for Observational Techniques (SPOT), which 161 US airports use to screen potential terrorists (it's also being tried out in the UK). It has led to over 232,000 passenger interrogations in the US, but only around 1,700 arrests - none of them for terrorism. Nonetheless, Ekman claims he can tell a lier close to 100% of the time.
Official analyses of SPOT have been unimpressed. An article in Nature magazine this year says:

"No scientific evidence exists to support the detection or inference of future behavior, including intent," declares a 2008 report prepared by the JASON defense advisory group. And the TSA had no business deploying SPOT across the nation's airports "without first validating the scientific basis for identifying suspicious passengers in an airport environment," stated a two-year review of the program released on May 20 by the Government Accountability Office (GAO), the investigative arm of the U.S. Congress.

Still, even if we think his work on lying may be a bit Hollywood, Ekman's basic theory of emotion recognition still stands up, doesn't it?

Last night I went to see James A. Russell, professor of emotion at Boston College, who is probably the leading critic of Ekman's ideas. He gave a talk at the Centre for the History of Emotions at Queen Mary College (where I'm working as a researcher), in which he challenged the whole theory that there are discrete entities called 'shame', 'fear', 'disgust' and so on, which exist within us and upon us, like Platonic Forms or Jungian archetypes etched into our faces.
In this I have much sympathy with him. Is happiness one emotion? Is fear? What's the difference between the fear we feel when we see a dangerous animal, and the fear we feel when we approach a girl we are in love with? Are emotions always expressed with the same facial expression? Russell showed us a picture of a girl having an orgasm (yes, it was quite a racy evening in Mile End), and pointed out that it appeared she was in pain rather than joy.
Ekman's most famous study involved showing photos of posed facial expressions to tribesmen in Papua New Guinea, and finding an above chance recognition of the six basic emotions. But, as Russell has written, the experiment was quite flawed. The conditions were far from controlled, and the tribespeople may simply have been learning what response was expected from them rather than genuinely recognising what Ekman thought they were recognising. Indeed, his experiments show much greater recognition when it is explained to participants what is expected of them, and when it is explained that they must choose one of the six discrete emotions. Their responses roam much further from the six basic types if they are allowed to respond however they want.

Russell noted that the photos which participants are shown in Ekman's experiments are typically posed. They're not genuine, they're exaggerated, like Greek masks. If the photos are taken of spontaneous emotional expressions, the recognition rate falls from 84% (for posed expressions) to 31% among westerners. And recognition of most of Ekman's six facial expressions is lower than 50% among non-western, non-literate societies, according to a meta-study put together by Russell.

So where does this leave us? Well, Russell went on to suggest that perhaps emotion is a folk concept. Perhaps it is a very complex set of thoughts, feelings, words, images and physical responses, and we're doing violence to nature if we try to fit them into narrow boxes. If that is so, then of course psychology is also 'folk', and the Diagnostic and Statistical Manual will have to be thrown into the bin (as some psychologists wish it to be). How can you prove a therapy gets rid of depression, if you aren't even sure what sadness is?
Russell reminds me somewhat of Daniel Everett, the anthropologist who is challenging Noam Chomsky's universal theory of grammar. Like Everett, Russell is interested in the particular, in the exception, rather than trying to shoe-horn all different cultures and experiences into some universal system.

But while Everett aspires to be a serious challenge to the cognitive theory of language, I don't think Russell is (or wants to be) a serious challenge to the cognitive theory of emotions. Cognitive emotional theorists like Beck, Ellis or Arnold Lazarus are, like Russell, constructivists - they believe that emotional experiences are constructed, through an individual's beliefs and also through their culture's beliefs and values. So each person feels something individual and unique. We can label our experiences 'grief' or 'joy' or other socially-accepted terms, but my grief is likely to be different to your grief, both verbally and physically.

Nonetheless, science and philosophy have proven a lot about how emotional episodes and patterns arise, and how we can change them if they are causing us suffering. So even if emotional patterns are highly person-specific and culture-specific, the ways to change them are, I believe, testable and to some extent cross-cultural - although how we want to change them and what we think are appropriate and 'healthy' emotional responses will also be culture-specific, and therapist-specific.
To read more, here is a good article from the APA bulletin on emotions and facial expressions.