Making the world a continuous place

Aristotle wrote that drama must be guided by three principles, the Unities. All aspects of a good play must take place in the same location, within a short time period, and contribute to a single plot. Otherwise, forced to stretch their imagination, the audience wouldn't be able to suspend disbelief, and the play would cease to be a reasonable imitation of reality. The ideal play would take place over the same amount of time it took to perform (say, two hours), would be set in the same place, and would have a single course of action.

It is indeed an interesting feat when a play or a film manages to accomplish such unity. But surely it's not necessary: we routinely watch films that cover periods of hundreds or even thousands of years, spanning dozens of locations. What is necessary, from a psychological perspective, to make viewers believe a film is a single, continuous action? If we are willing to accept discontinuities in film, is it because we perceive film in a different way from the real world?

Daniel T. Levin of Kent State and Daniel T. Simons of Harvard attempt to answer those questions in their article "Perceiving Stability in a Changing World: Combining Shots and Integrating Views in Motion Pictures and the Real World" (Media Psychology, 2000).

Film makers expend a lot of effort to make sure their movies are continuous. Each member of the crew makes sure that for every shot, the cast is costumed identically, holding the same items, standing in the same position. Yet they also know that the audience typically doesn't notice continuity errors. Fan Web sites will often compile lists of errors (I've linked to a "Jurassic Park" fan site), the vast majority of which are never noticed by anyone except obsessive repeat-viewers. When you saw this movie, did you notice that the hinges of the door move from the left to the right when Dr. Grant enters the trailer to meet Hammond? Or that the flashlight on Ellie's leg "disappears" after she goes through a door?

Some have argued that the fact that we often don't notice continuity errors in film suggests that we perceive film differently from the real world. Levin and Simons argue the opposite: that we don't notice continuity errors in film because we wouldn't notice them in real life, either. Consider the following movie, which will flash a well-known painting at you. Play it for a while before reading the text that follows:

Did you notice anything changing as it flashed? If you're like most people, it took you a while to notice the dog disappearing and reappearing. It would have taken even longer if we had changed a less salient detail, such as the flowers in the vase in the background. (You can pause the movie now to stop that annoying flashing. It gives me a headache, too.)

Levin and Simons have shown people simple films in which an actress hears a phone ringing, gets up, and answers it. When the film cuts to the woman answering the phone, a new actress, wearing different clothes, is substituted in. After showing the film, they asked viewers if they saw anything "unusual," and 66 percent of the viewers said they did not. Only when they were alerted to look for an unusual change were they able to spot the changing actress.

Levin and Simons went even further: they staged an experiment in the Harvard Yard where an actor would stop a random passerby to ask for directions. Halfway through their conversation, two men carrying a door would shove their way inbetween the actor and his unsuspecting victim. As the door passed by, the original actor would grab onto one end and walk off, and one of the door-porters would continue the conversation as if nothing unusual had happened. Almost half of the victims carried on with the conversation and tried to answer the actor's questions. They were genuinely surprised when, after revealing the experiment, the interviewer explained that he was a different person. The video of this experiment (which, unfortunately, we don't have) is hilarious.

This result would explain how the movie That Obscure Object of Desire was still watchable, despite the fact that the leading female character, Conchita, was played half the time by Angela Molina and half the time by Carole Bouquet. The film's director, Luis Bunuel, observed that a fair number of viewers didn't notice the shift.

Why don't we notice these changes? Wouldn't it adversely impact our ability to function in the world? It's perfectly possible to imagine a brain capable of instantaneously noting the slightest change in anything in our field of vision. Wouldn't we be better adapted if we could do that? Not necessarily: after all, the apparent shape of an object changes depending on the angle we view it from. Levin and Simons give the example of a coffee mug. Would we get along better in the world if whenever a mug was rotated so that the handle disappeared from view we assumed that it had been removed and substituted with a new object? By far most common occurences in the real world are for objects to stay where we put them, or to continue along the path they were following. By assuming that continuity is maintained, we can keep track of more different things, such as the cars at a busy intersection, or a group of animals we are hunting.

What's good for us is better for the filmmaker or dramatist, who can leave out key details and count on us to fill them in. And what of Aristotle? Why did he argue for the clearly unnecessary unities of time, space, and action? I suspect it's because writing a play that follows the three unities is more difficult that one that doesn't. A play composed following the three unities probably had a more skilled playwright, and that is why it's a better play.

More like this

[...] n, 2003). They gave participants a change blindness test similar to the one I demonstrated here (scroll down to see the movie), but instead of changing an innocuous detail in the picture, they changed an [...]

This is wonderful research that sheds considerable light on the nature of consciousness and perception.

One of the best papers that I ever read on this subject is by Peter Lynds, controversial New Zealand scientist and philosopher.

Although not formally an Academic, nor formally trained in Psychology, he has linked the nature of consciousness with the perception of time, and homed in in the philosophical puzzles of discrete versus continuous, going back to Zeno's paradoxes, and a novel (and as I say, controversial) solution at the heart of Physics.

This is a very nice summary of those change blindness findings, but one thing that should be pointed out is that during most change blindness tasks (though not inattentional blindness tasks), the change needs to co-occur with lots of other "stuff." In films, the 'stuff' is a cut between camera angles. In change-blindness tasks, it's either a blank screen in between views or a series of 'mudsplashes.'

Most changes are accompanied by a large motion transient in the visual field, which is VERY easy to detect (hence people are generally overconfident in their change-detection abilities, see Levin, Momen, Drivdahl and Simons, 2000). In most every-day situations we DO notice pretty much every change that occurs in the environment provided we're not totally distracted at the time. It's the failure to explicitly compare representations across these massive disruptions that makes change blindness so robust.

Anyway, that's more of a minor point since I think you're getting at the idea of continuity in general. The assumption of a stable world is critical (as are most ecologically valid assumptions!), and the perception/understanding of movies is, in my opinion, one of the most bizarre human abilities.

Very cool stuff.