For most of the past two millennia, the primary purpose of portraiture was remembrance. But since the few people who could afford to commission portraits - rulers, merchants, clergy - usually had interests extending beyond the merely personal, these were also expected to be edifying and exemplary in character. Then, from around the 15th century, a more intimate ideal of portraiture gradually took hold in western art: an expectation that portraits should provide insight into the character of the sitter.
Of course, in more recent times, the idea that character can be pinned down in an image has been thrown into doubt. Yet the assumption persists. To read about any well loved portrait, be it a Hockney, a Hogarth or a Holbein, is to be subjected to a stream of near random guesses about the personality and temperament of the person depicted. Art history is still mired in spurious suppositions about the relationship between facial appearance and identity.
A current exhibition at the Mus?e de Luxembourg in Paris called "Moi! Autoportraits de XXe si?cle" presents self-portraits by some of the great artists of modernism, starting with Degas, Picasso, Giacometti, and going through to Kahlo and Baselitz. In their variety, the works bring home the fact that the 20th century saw a spectacular breaking apart of old notions about the relationship between portrait and personality. You can see this in Picasso's two-faced heads, or Giacometti's wretchedly isolated individuals. And now, at the Hayward Gallery, a photographic exhibition curated by William Ewing called "About Face" aims to question further the conventions of portraiture, casting doubt "on the idea that a photographic portrait can be a credible likeness of an individual or a 'window on the soul.'" Ewing's selection has been prompted by the new technological possibilities of manipulation both in reality (cosmetic surgery, personality-altering drugs) and in representations (digital manipulation). The underpinning of both exhibitions is clear: "essences" and "truths" about a person cannot be defined by an image of that person.
And yet, despite all such developments, discussion of traditional forms of portraiture still takes its cue from the discredited 18th-century science of physiognomy. Witness Casanova, who wrote in the introduction to his memoirs: "Only there, in the face, is a man's character in plain view, for that is its seat." What's surprising is that this continues to sound acceptable, and even obvious, to us today.
Modern science acknowledges the importance of the face as the locus of expression and communication. According to Paul Ekman, the scientist who in the 1970s co-developed the facial action coding system (FACS), the face is "our badge of identity." He says: "We have such specialised circuitry in the brain for responding to faces, because it is the input for most of our senses. We don't hear, see, taste or smell anywhere except in our face. Touch is the only sense we have elsewhere. It's where speech occurs, and it's where food comes in. There's an amazing number of things concentrated in this part of the head."
All this certainly helps to explain our intense interest in artistic depictions of the human face. But science also tells us that there is a great difference between the human face at rest and the human face in motion. And although great portrait artists have all sorts of ways of suggesting the engagement of the sitter - from the basic illusion that his or her eyes are following you around the room to more sophisticated combinations of facial expression and body language - in the end, a portrait is a static presentation.
According to Ekman, "It seems unlikely that static facial signs are actually related to temperament or personality. The wrinkle that one person interprets as a sign of wisdom may be interpreted by another as a sign of dissolution." This is one of the reasons physiognomy lost scientific respectability long ago. But people continue to share a variety of beliefs about the relationship between facial appearance and personality, including the ability to tell whether someone is lying.
Why the persistence of these beliefs? The most likely explanation is that we are susceptible to stereotypes. Research in the 1950s showed that Americans routinely associated dark-complexioned, oily-skinned faces with negative traits - hence movie villains tended to be dark and swarthy.
More pertinently for art history, people tend to mistake a part for the whole. In other words, seeing a smiling face is enough to make people assume that happiness is an enduring characteristic, when it may be no more than a fleeting response to a stimulus. Such prejudices infect people's responses to portraiture, too. This may be why, since the advent of modernism, artists have been more interested in subjective states of mind, and in the mechanisms of vision, than in anything so hard to pin down as a person's character.
There has, as a result, been a general dearth of level-eyed critical appraisal of portraiture. Echoing the conclusions of science about the relationship between emotions and the face in motion, the 20th century's fascination with the human face has been concentrated in the movies and on television (one of the most profound investigations of the matter is found in Ingmar Bergman's Persona). Some contemporary artists, such as Bill Viola, whose disappointingly slight videos of actors in slow motion were exhibited recently at the National Gallery, have tried to switch their attention to expressions in motion. But for the most part, portraiture, remaining static, has lost critical, if not popular, credibility.
Is science a good place to start to find new ways of talking about portraiture? At first I thought so, which is why, on a cold day in London last year, I invited Paul Ekman to the National Gallery. A trailblazer in his field, Ekman has advised the FBI and CIA (his knowledge of the human face makes him a connoisseur of deception) and has written and edited many books, including, most recently, Emotions Revealed. FACS, developed with Wally Friesen, describes the roughly 10,000 facial expressions of which the human face is capable, with a number given to every muscle. I glimpsed its sophistication when Ekman stopped us in front of a group portrait by one of the Le Nain brothers, who worked in France in the mid-17th century.
An old woman dominates the group, surrounded by her children. "She's seen difficult times," Ekman began. "You can see the triangulation of the upper eyelid muscle at the left-hand side and the bunching of her two eyebrows, which is a wonderfully reliable sign of anguish and sadness. Her lips are interesting, because of course gravity continually causes one's mouth to droop, so the question here is, is it muscle 15, or is it just an age effect? Her gaze is not terribly focused, which suggests a vague melancholy. But she's admirable, this woman. She's stoic."
I couldn't wait to hear what Ekman had to say about Rembrandt, whose greatness as a portrait painter is surely beyond argument. The first Rembrandts Ekman and I came to were the two portraits of Jacob Trip's wife. They are tremendously moving pictures, thought to have been painted in 1661. Of the first, Ekman observed that her face was not "in play." "She's off in her own world," he said. The second, he said, was different: "She's more in command of the situation. There's a sense of anticipation in her slightly parted lips." Such comments, however, came between long and slightly puzzled pauses. After a few minutes, Ekman had to admit: "I get very little from either of these faces other than the ravages of age. You can see the broken blood vessels and a lot of bagging to the right of the lips. The eyes are quite expressionless. The area around them is totally smooth compared to the lower area of her face."
Common sense tells us that there are many problems with the idea that a person's character is revealed in his or her face. Ekman's description of the old Dutchwoman's face as "in play" or "not in play" pointed the way to perhaps the most serious obstacle facing portrait painters: that expressions are mobile. Some of the most telling - Ekman calls them "micro-expressions" - are those which pass across the countenance in a split second, and they have a habit of giving away the true feelings of people who are working hard to control their expressions. How is a painter adequately to convey these?
Besides which, isn't character too complicated a phenomenon to sit so glibly on the surface? Aren't we too deep, too variable, too mysterious? Ekman told me that several years ago he had a correspondence with the late art historian EH Gombrich on this subject. Gombrich's belief was, in Ekman's words, that "the job of the portrait painter is to capture the essence of a person, the part that doesn't change over their life cycle." But is character really such a stable phenomenon? For a large portion of Gombrich's lifetime, research had been suggesting that personality was profoundly unstable, casting doubt on the kind of essence Gombrich had in mind; never mind whether it could be captured visually. "The research, however, has come back to the view that personality is a reasonably stable feature," Ekman now told me. "Once an extrovert always an extrovert, and so on."
So was Gombrich right? Are we really learning about someone's true character when we look at a great portrait? Ekman takes a deep breath, and then: "I think most of the time we're not... We're reaching for prejudices, like, for instance, thinking someone's cruel because their lips are thin. (Everyone's lips get thinner when they're angry). Most of the time they're stereotypes." But, he said, "you can be a great portrait painter because you get the 'impression' of a person - regardless of whether it's accurate. A great performance by Laurence Olivier doesn't tell me who he is as a person. But it's still a great performance, because I get a sense of the person he's presenting. In the same way a great portrait is a presentation. It's irrelevant whether it's true or not."
Ekman seemed to be saying that a great portrait is less about likeness and more about the artistry, the invention, involved. "Looking at Rembrandt," Lucian Freud told me, "we'd all have to believe that the people who sat for him, who were bankers and merchants and probably unremarkable people, were all filled with spiritual grandeur. I don't think they were. Rembrandt may have been. A great portrait has to do with the way it is approached. If you look at Chardin's animals, they're portraits. It's to do with the feeling of individuality and the intensity of the regard and the focus on the specific."
In front of the Rembrandts, Ekman made various astute observations, only to concede: "You can read almost anything into it," (in front of Rembrandt's Self-portrait at the Age of 34), or "I can read very little in her expression," (in front of the portrait of Rembrandt's wife Saskia). One of the world's best readers of faces was studying some of the greatest portraits of all time - and this was all he could come up with?
It seemed to me that this did not point to a failing in Ekman, but it did say something about the nature of portraiture. The greatest portraits are more often about the painter's response to a sitter, a response which can take any number of forms, than about the sitter himself or herself. Comparing portraits of the same person by different artists brings this home. Artists often say that every one of their paintings is really a self-portrait: ogni dipintori dipinge se, "all painters paint themselves." This strikes me as an over-simplification, but the point is clear enough.
We are all capable, it turns out, of reading faces which express obvious states of mind, such as depression, surprise, anger or contempt. Some people - and Ekman is clearly one - are even fluent at reading more complicated expressions: a smile masking sadness, for instance; politeness masking imminent aggression; contained joy, and so on. But all this is as nothing when set against the true depths and contradictions of human character. Consider this description of what Ekman calls the "nuances of personality," from a novel by Philip Roth, and ask yourself whether a portrait could ever tell you so much: "Her laugh was very shy now, a delightful surprise. A delightful person suffused by a light soulfulness that wasn't at all juvenile, however juvenile she happened to look. An adventurous mind with an intuitive treasure her suffering hadn't shut down... The sliding about of her self-possession was practically visible as she spoke. Self-possession was not her centre of gravity, nor was anything else of hers that was on display."
One could probably find even better descriptions in Bellow, or Austen, or Shakespeare. The point, though, is that you need to have access to a living, breathing, interacting human being to get anywhere near this level of insight into character. Which may be why Shakespeare himself wrote: "There's no art to find the mind's construction in the face." According to the criteria we have been discussing, painting rates as a poor cousin to literature and film. Which may be why, just as modern literature has made the mining of consciousness its chief concern, recent art has largely abandoned the portrait. Even the man many regard as the greatest painter of portraits today, Lucian Freud, seems to consider psychological insight somehow tangential to what he does. His sitters are often asleep, and even when they are awake, their faces are rarely "in play."
Great portraits attain their status for any number of reasons. But many of the greatest seem to remind us of the unknowability of other people, even those we love. This apprehension - funnelled through great art - can be electrifying. I think of Titian's portrait of the young Ranuccio Farnese, Velazquez's portrait of the dwarf Don Diego de Acedo, Manet's of Berthe Morisot, Freud's of his granddaughter Frances Costelloe. "I can take you this close to another human being," these artists seem to be saying. "I can pay this much attention, be this responsive, distil more information into this concentrated essence of illusionism than you thought possible from a mere painter... and still this human being will break your heart."
The important portraits of recent times tend to be about intimacy, and intimacy takes shape over time. Inherent in the painted portrait, unlike the photograph, is this sense of duration, of elapsed time distilled into a single image. In the process some residue is left behind, some literal or poetic evocation of the simple act of attending to someone. But of course, true intimacy with other human beings seems to require an apprehension of their essential difference, too: their otherness. Perhaps that is why there is nothing quite so moving as the human face attentive to its own inner drama. "There's nothing more fascinating," as the Australian photographer Bill Henson once told me, "than to have someone stare out of an image into your eyes, yet never allow you to know anything about them."
