In the 1910s and 1920s, Lev Kuleshov was a famous Russian filmmaker curious about how audiences responded to film. This was a time when the art form was extremely new, but audiences were already going nuts for stars. One such star was Ivan Mozhukin, a matinee idol and beloved actor. Kuleshov shot a short film of Mozhukin looking into the camera. Nobody is entirely sure what else was on that piece of film, because it has been lost.
Most accounts say that the film showed a shot of the actor's face looking into the camera, and then a bowl of soup. Then the film showed another shot of Moszhukin's face, and then a child in a coffin. Lastly, there was another shot of his face, and shot of a beautiful woman. The audiences, according to Kuleshov, talked about how well the actor expressed his hunger, and then his sorrow, and then his love. But every shot of the face was the same - used over and over. The film is, metaphorically, shoved in the face of every film editor ever. But the Kuleshov Effect is also discussed by psychologists. It shows that people will bring their own emotions and assumptions to what they see.
Or so it is claimed. Subsequent recreations have shown that people will classify (or mis-classify) certain faces depending on context. Neutral faces can look happy or sad. Screaming faces can look joyful or agonized. Fear and anger are confused. That's understandable.
But creating any feeling out of a neutral face? That, argues a group from the Society for Cinema and Media Studies, goes to far. For one thing, no one knows what Kuleshov told the audiences before the montage, or how he judged their reactions. Moreover, their reactions weren't as diverse as they might be. The audiences might infer that Mozhukin was in love with the woman, or in lust with the woman, but not indifferent to or repulsed by the woman. No one thought that he had negative feelings for the soup, either. While some films have actors looking at a child playing and some have them looking at a child in a coffin, no audience mistakes the intense look of the actor for anger or impatience with the child. And in no clip does the actor seem bored or disinterested to audiences. So while audiences might need context to understand the look that the actor is giving, the film editing only goes so far. The intense emotion is a given, and because of the performance it's shown not to be an intense negative or hostile emotion. The audience knows that, and then forms a more specific opinion once they're shown what the object of the intense emotion is.
Today, it's hard to recreate the effect, in part because people know it so well. A group of French students who were subjects of such an experiment, named the experiment in their reaction surveys. Slightly more successful (in getting unbiased results) was an experiment done by two people from the Society for Cinema and Media Studies. An actor's face was shown, then either soup, a child playing, or a woman in a coffin. This film wasn't shown as a montage. Instead it was three different segments. When shown only the actor's face, eighty-eight percent of people said it was neutral. When new subjects were shown one of the three sequences in the film, the majority still said the actor's expression was neutral. Sixty-eight percent believed the expression to be neutral when they watched the soup sequence. Sixty-one percent said it was neutral when they saw the child or the coffin sequence.
There are still questions, of course. Kuleshov wasn't just dealing with early film audiences, he was dealing with fame. This was a known actor. Would so many of the modern-day watchers - who had been asked to evaluate the man's acting performance - have considered the face neutral if it had belonged to as talented an actor as, say, Meryl Streep? (That would be an amazing experiment all on its own, though probably beyond the budget of most psychology or film studies departments.)
So how much are we manipulated by editing in film? Can we recognize emotion without context? Maybe not - but we can probably rule quite a few emotions out.