The Law of Timespans

As I blogged earlier this month, I think that both Australia and Synecdoche, New York are interesting but flawed movies, and in both cases, I liked the first part of the movie better than the second half. That's the thing: both movies divide easily into halves, with a big jump in time separating the two parts.

The first part of Australia, the cattle drove, takes place over a matter of weeks, then there's a jump of several years, then new complications in the form of the Japanese invasion. These later scenes are comparatively weaker, and unnecessary; at the end of the first part, Nullah even says "Everyone got what they want," which is usually a signal that it's time to wrap things up. Synecdoche, New York is trickier because it plays with time throughout: still, in the first half of the film, the jumps in time are not large. But after Caden wins the MacArthur grant, the next thing you know, seventeen years have gone by and the movie's tone becomes increasingly "meta" and stifling.

And that brings me to an artistic axiom I devised, the Law of Timespans: There tends to be an inverse relationship between the amount of time a movie covers and the quality of the movie. In other words, a movie that takes place over a short period of time is more likely to be good than a movie that takes place over a long period of time.

I came up with this theory a couple of years ago after my first viewings of The Godfather, Part II and Chinatown. I became annoyed that Godfather II had won Best Picture in 1974 when I felt that Chinatown was clearly a better movie. But what made it superior? Its perfectly constructed, inexorable plot where nothing is out of place. The way it covers just a few days' time yet takes Jake Gittes from cocksure private eye to tragic hero. Whereas Godfather II is actually kind of messy, toggling between Michael's life as mob boss and Vito's early days in New York City. Despite its decades-long timespan and 3+ hour running time, it doesn't hit me like Chinatown does.

When I first came up with the Law of Timespans I thought it was profound; now, I realize it is more common-sense than ingenious. After all, a shorter timespan = fewer available moments to dramatize = fewer opportunities to make the wrong choice. With less potential story material, it's easier to get a handle on what you do have, and shape it accordingly. In general, I don't believe in Unfettered Art. I believe in form, structure, working within a set of constraints to spur creativity... all of which are aided by a short timespan.

Indeed, my Law of Timespans might be just another way to describe Aristotle's "Unity of Time and Action." It's funny, you read Aristotle's Poetics so many times in drama classes that it becomes a cliche... and you grow to despise the Unities when you read too many classical plays that follow the Unities to an absurd degree, taking place all on one day in some palace antechamber... and then you see a few flawed movies that don't respect the unity of time and action, and you realize how necessary the unities are!

I guess because Homer and other epic poems are fundamental texts of Western civilization, there's a perception that epic works are inherently better or more worthy. Maybe that's true for literature, but it seems like the opposite is true for drama and film. It's become a cliche to say that epic movies are "Shakesperean in scope," but if you think about it, Shakespeare's most acclaimed plays do not take place over long periods of time. He compressed time in his history plays; his comedies (as comedies must be) are madcap and quick; Hamlet is the most famously indecisive character in literature and yet his eponymous play takes place over only about a month's time. Meanwhile, The Winter's Tale is notoriously hard to direct because of its big shift in time, tone, and action halfway through!

Or, alternatively, what made Hitchcock such a great filmmaker? Lots of things... but might one of them be that none of his movies cover much more than a year's time, and many of his best movies cover even shorter timespans?