From Perigueux to Final Fantasy to A.I: How much realism should we demand?

I found myself checking up on the parts of a horse the other day. It was after the Daily News had carried an AP story about some new prehistoric art found in the Perigueux region of France?engravings thought to predate the Lascaux cave paintings by 10,000 years. It was a burial ground of some sort, apparently: a number of human skeletons had been found in the cave as well. The version of the story Newsday carried included a quote from an official of the French Ministry of Culture: "The presence of graves in a decorated cave is unprecedented."

But the drawings in the News photograph didn't look like decorations; they looked like sketchpad studies. They were partial (a mane here, a hoof or fetlock there, an idea of musculature) and unarranged, all piled on top of one another as though the artist had been unwilling to risk taking the time to find a blank space on the wall for fear of missing a chance to capture something from life.

Only one of the figures in the photograph?a horse?was recognizable. It seemed curiously realistic?so realistic that for a moment I wondered if the drawings might be a hoax. It wasn't stylized enough for prehistoric art, I thought; and it seemed too interested in anatomical detail. This was no flat, undifferentiated geometric shape with characteristics one might interpret as equine; this was a proper horse, fully articulated and drawn (in profile) almost in perspective, complete with all the things a horse should have. You could make out every element of horse physiognomy: upper and lower muzzle, nostril, throat-latch, even the soft, fat, jowly part that covers a muscle I now know to be called the masseter.

Now, there's nothing to say that primitive artwork has to be more stylized than it is realistic. Or, to put it another way, there's no reason to think that art wasn't realistic before it was stylized?any more than there is to think it impossible that a more advanced technology than ours once existed a long time ago in a galaxy far, far away. I mention the Perigueux horse because I've been thinking about realism and views of reality in the context of some of the summer's more and less obviously cheesy movies. Mostly I've been trying to figure out why the picture of a world proposed by Steven Spielberg's A.I. bothered me so much.

When it comes to matters of realism and stylistic form, it's always interesting to find out what we are and aren't prepared to accept. Detail is what tends to create problems. We might be willing to entertain the underlying premise of An American Werewolf in London but object when the SWAT team that has the monster cornered at the end of the picture allows the heroine to step into the line of fire. That incident might seem unrealistic. We tend to hold different art forms to different standards of verisimilitude, demanding more of the narrative and dramatic, say, than the graphic arts. When the Metropolitan Museum of Art held an exhibit of late Renaissance drawings earlier this year, for instance, you didn't notice museumgoers finding a lot of fault with Correggio because some of the pictures deviated from natural truth. You didn't hear anyone saying, "Look at the way that Madonna is holding that baby! It's ridiculous! No woman would try to hold a baby that way, it would slide right off her lap!" The point was the folds of her dress, the way it draped over her leg, which would have been obscured if the artist had taken the actual real-life weight of an actual real-life baby into account.

It's artistry itself, as often as not, that leads us to ignore some discrepancy between the truth as it's depicted in a work of art and the way things are. If you go to see Kenneth Lonergan's Lobby Hero at the John Houseman Theater (it reopened there in May and runs through Sept. 2), there may come a point when you find yourself noticing something slightly unrealistic about the play. Set in the foyer of a Manhattan high-rise, it concerns the relationship between a young security guard who works the graveyard shift at the apartment building, his supervisor and two cops, one of whom is having an affair with a tenant in the building. You'd look hard to find a visual stage truth as compelling as the way the shadow of an adjacent building on Allen Moyer's set cuts off the sunlight from the sunken area just outside on the pavement?exactly the way the buildings surrounding those badly designed East Side high-rises always do. You know that building, you can visualize the whole exterior just from the way Mark McCullough has lighted that tiny sliver of stage, and the characters are equally well observed.

All the same, it's bound to occur to you that in the entire course of the two nights the play spans, no one other than the characters in the play crosses the lobby. It's unimportant. The truths contained in the characters' expectations and treatment of one another are more interesting than the convention we're being asked to accept?just as the folds in the drapery are more interesting than the bulk of the baby in Correggio's drawing.

Sometimes what prompts us to accept a glitch in verisimilitude is the arrival of a new technique, a way of expressing something that couldn't have been expressed before in a particular medium. I remember that some years back, when the Met was holding one of its exhibitions of fifth-century sculpture, there was a particularly wonderful piece of signage pointing out that the famous statue of Nike bending down to fasten her sandal is fundamentally unrealistic; at the same time it represents an important moment in the development of "realism." The way the sculpture captures the fall of the cloth over the goddess' body is lifelike beyond anything that marble had hitherto managed to express. Still, the curator noted, a cloth that fell in exactly that way, and would show the outline of the body as the one in the statue does, would have to be gossamer-like, and fabric that light wouldn't drape well. In order to express what he wanted to express, the artist had had to create another reality in which a garment and what it hides are both visible at the same time.

One of this summer's cinematic talking points is a movie that uses computer-generated images of actors instead of real actors. It's fascinating for the space of about 10 minutes because of the precise way in which it doesn't work. The moving figures that act out the story don't seem like either actors or animations, they're merely an attempt to ape a simulation of life. Animation takes nonhuman-looking images (illustrations) and breathes life into them. Its wit historically resided in its ability to assign human attributes to nonhuman creatures, thereby commenting on humanity. But the suggestion of life is dependent on spontaneity. The creators of Final Fantasy didn't have that to work with, so they had to fall back on facial and gestural cliche: this expression for fear, that pose for anger or grief. For all its technical prowess, Final Fantasy turns out to be a throwback to silent movie acting.

Of course, it's caused a certain amount of consternation in the entertainment industry. The fear is that if such methods are found to be "successful," computer images will gradually come to replace real actors on the screen. Interestingly, this real-life development actually mirrors the major plot point of A.I., Spielberg's long-awaited movie, about a boy-robot who develops mortal longings. The film, which Spielberg developed from an idea that Stanley Kubrick had researched for years before turning the project over to the younger director, posits a postapocalyptic future (some polar icecaps have melted, drowning all of the world except for a significant portion of New Jersey) in which human beings have so perfected the art of simulating humanity that the only thing left for a self-respecting Promethean to explore is whether a robot can be programmed to love and thereby become more "human."

It's odd that in the role of the boy-robot David, Spielberg chose to cast Haley Joel Osment, the child actor whose passion in The Sixth Sense was so moving and played so well against Bruce Willis' trademark lack of affect. In A.I., the young actor is required first to simulate lack of affect himself and then, as David's adoptive mother utters the words that program him to love her for all time, to simulate recently acquired, artificial affect.

Actually, there are a number of curious things about A.I., not least of which is the widely noted "schizoid" quality that critics have enjoyed attributing to the Spielberg/Kubrick dichotomy. The movie keeps presenting us with recognizable tropes, situations arising out of the singular plot, which we think will develop in a way that explores what it means to be human. (That's Spielberg the Bard, King of Genre, Arch-Storyteller.) But these setups keep petering out, wandering off into tough-minded existential gloom. (That's Kubrick the Genius and Redoubtable Intellect.)

Watching A.I., I found myself prey to the American Werewolf syndrome, willing to entertain the premise but stumbling over details. I was prepared to accept a world of punishingly planned parenthood serviced by a race of humanoid robots created to lick the resources problem (it beats rationing). But I kept wondering why the couple in the movie, David's adoptive parents, are so inexplicably wealthy. They live in a huge, beautifully appointed house, miles from anyone else, and can afford to have their birth son cryogenically frozen until such time as a cure is found for whatever it is that ails him. And why is it that they appear to have no friends? Where are all the other people in this world? It seems inhabited entirely by people who work at the robot plant. Apart from them, the only human beings are the rabble?the sweaty, ugly crowds that frequent the roving demolition festivals called Flesh Fairs (carnivals; get it?) where antiquated, damaged or otherwise unwanted robots are ritually trashed. They're part theme park, part slave market, part revival meeting, part public execution, and the unkempt folk who attend them are there to exorcise their fears of extinction.

A friend pointed out that the mob that turns on the carnival manager, rallying to defend the robot child because he is a child, is acting is out of sentiment, not humanity. There's nothing noble or uplifting about the scene; it simply substitutes mawkish savagery and brutality for the self-interested sort. I doubt that was Spielberg's intention, but then the whole movie is sort of one big glitch in verisimilitude. It's a portrait of a society trying to make lifelike beings, drawn by a man who has been so removed from real life for so long that he doesn't remember what it looks like. Or, rather, two men unusually removed from life?one who conceived the project and one who carried it out.

At least the movie based on a computer game knows that it's junk. Ironically (or perhaps predictably), it carries the same message as the Spielberg epic: what makes us human is our dreams. But Spielberg is being either disingenuous or naive: his point, surely, is that what exalts the human race is movies, not dreams but dreammakers like himself and Kubrick. The whole movie is a series of allusions to Spielberg and Kubrick, their achievement, their work. I think it's telling (and more fraught with worrying connotations than anything in Final Fantasy) that the most lifelike and compelling performance in A.I. comes from a computer-animated teddy bear.