My first screenwriting theory came from David Siegel. Remember him? If you remember David Siegel, you were involved with the Web in the early-to-mid 90s. He was best known for his Casbah, the transparent-pixel trick he either invented or popularized, and an inspired rant about typography and design on the Web.
Buried inside what was, for the time, a huge, complex, and unparalleled personal web site, David Siegel described his Nine-Act Structure theory. I was probably 20, or 20-ish, and I had written at least one truly unreadable nightmare of a screenplay by that point. I thought, what the hell, I'll try it and see if it works. And it was a massive improvement, but it wasn't enough.
Years later I developed a second theory, a much simpler theory, which was this: show, don't tell. Anything you want the audience to think, believe, or observe, demonstrate it. There was a corollary to all this, which was that every scene should consist entirely of action, and scenes on telephones or in restaurants were strictly forbidden. This came from a book on screenwriting, but I've forgotten which one. I may have aggregated it from a couple books.
Anyway, this theory improved my screenwriting also, to the point where, in June of 2003, I had a script I had started on in the late 90s, and which was effective on a scene-by-scene basis. But everything I wrote broke down after about 30 to 50 pages. Since a script is supposed to be around 90 to 120 pages, this was a bad thing. Also, my characters were utterly soulless.
In 2005 I took a couple great acting classes at a great school. My new theory is based on that. Or, I guess, my new rule of thumb. I still use the 9-act structure, and I still use the "show, don't tell" rules, but my new rule is that I am not just telling a story; I am also building a role-playing game for the actors, and every scene must be an interesting game for the actors to play.
What led to this idea was an insight I had while driving. I realized that in a sense all acting is improv. The script will give you words to say; it will not tell you anything else. How you move, how you breathe, the way you stand, the tone of voice you use, the movement of your eyes -- all these things that make a character have to come both from yourself and from the other actors. That is 100% improv, and it is the main reason why movies never buried the stage; why even today, when you can have an incredible theater in your living room, people go to the theater. If you play the same scene 100 different times, you will play it 100 different ways.
If this is the case, then the job of the screenwriter isn't really to tell a story, because it won't be the screenplay which defines the story. The actors will define the story. The actors are playing a storytelling game, and when that game is complete, then the story will have been told. But because it's a game, it will have a different outcome every time. So your job as a screenwriter isn't to tell the story. It's to set up the rules of the game.
With this in mind, it's a lot easier to write a scene with real soul to it. All you have to do is consider the actors with compassion, and ask yourself, how can I make this game both genuinely challenging and genuinely rewarding? And it turns out that a good answer to that question will always, by its very nature, incorporate some fascinating drama. Make the one actor confront his fears and overcome them; let the other confess her love and change somebody's mind. In each case, there's a challenge, and a reward, and in building the challenge and the reward, you've created a context from which drama will naturally and inevitably arise. Rewards and challenges can be either personal or interpersonal; interpersonal is almost always better. You don't want one actor having his important personal moment while the other just kind of stands around waiting.
And the other thing -- smaller rewards are better. You want to match challenge to reward, but the reward should really just be a token. In fact the reward should almost always be just a little less than the challenge was worth. This comes from a story in the great book Influence.
The story goes like this: American POWs in Chinese camps, during the Korean War, were given essay contests to keep them busy. The essay contests frequently involved politics, but not always, and of course the Chinese captors were more likely to give the prize to an essay which praised communism, but every so often they'd give the prize to an essay which praised capitalism instead. So it became very normal for the Americans to make small concessions to communism in their essays, partly to win the prize, but mostly just to return the favor, even though they didn't realize it or think of it that way -- and after returning to the US, they were more likely to continue believing that communism had its merits. The reason was two things: first, they were only really required to make minor concessions; but more importantly, they were only ever given minor prizes. Like they wouldn't even get a pack of cigarettes; they'd get three individual cigarettes. If they had been given huge prizes, they would have decided afterwards that they did it for the prize. With the prizes so minor, so token, the actions taken to obtain them seemed more real.
Obviously, this is a subtle psychological trick, but I think most actors would forgive a screenwriter who used subtle psychological tricks to make his characters seem more real to the actors playing them.
Anyway, that's my new theory, and I think it has some merit.