Oh, right, that reminds me, I was going to talk about that concept also: that film is a “director’s medium.”
Consider:
An writer writes a scene. There is one version of the scene in the script.
An actor performs the scene in front of a camera. The director requires the scene to be performed X number of times (five, or twenty, or fifty). Of those X performances, the director orders N to be printed.
The camera is then moved, and the scene performed again Z times, with the director again ordering N[sub]1[/sub] prints based on those Z performances.
Repeat as necessary, until you have completely recorded the scene.
There are now N[sub]N[/sub] recordings of that single scene. Each of these takes will be different: there will probably be more than one angle, the quality of light may have changed slightly if it’s an outdoor scene, the actor’s emotional intensity and timing will vary, the cinematographer (or camera operator) might have handled the camera in a different way to capture the variations in performance, and so on.
Extrapolate from that single scene to the whole film. For any given scene, there are available from five to thirty (on average) recordings of it. (This is, incidentally, where the term “shooting ratio” comes from, i.e. a film shot at ten to one or whatever. If you assume a 90 minute movie, a ten-to-one ratio means the director and editor go into the editing booth with 900 minutes of film. Low-budget movies try to keep the ratio down, say to four or five to one, so they aren’t wasting money. Big-budget movies can shoot at fifty to one or more.)
Now. The director looks at all of this material and, with the editor, begins cutting the film together. (Depending on schedule and budget, the editor may assemble a preliminary cut without the director’s input.) For simplicity’s sake, let’s assume the director printed five takes of every scene. In other words, for scene A, we have takes A[sub]1[/sub], A[sub]2[/sub], A[sub]3[/sub], A[sub]4[/sub], and A[sub]5[/sub], and for scene B, we have takes B[sub]1[/sub], B[sub]2[/sub], B[sub]3[/sub], B[sub]4[/sub], and B[sub]5[/sub], and so on.
In take A[sub]1[/sub], the main actor has a lot of emotional intensity, crying, shouting, throwing things, etc. In take A[sub]2[/sub], they tried an alternate version that’s quieter, softer, more still. The other takes have similar slight differences. The director makes a preliminary choice that A[sub]1[/sub] is the best one.
Now he has to decide which of the B alternatives follows best, for emotional flow, continuity (of performance as well as detail), and so on. Say that the actor’s performance in B[sub]1[/sub] is cold and detached, in B[sub]2[/sub] is weepy and weak, etc. Maybe, in trying them all out, he decides that none of them really works. However, take A[sub]4[/sub] is almost as intense as A[sub]1[/sub], and dovetails quite nicely with B[sub]2[/sub], and gives the sense that the character is quite volatile, given to flights of anger followed by withdrawal.
The point is, with five takes of scene A and five takes of scene B, there are 25 possible assemblies of just those two scenes. The effect is subtle, but by picking and choosing throughout the film, the director creates the actor’s emotional throughline in the editing room. The actor does some of it on the set in front of the camera, but the director has tremendous amount of leeway in manipulating the actor’s performance.
(Note I haven’t mentioned that longer scenes can be further chopped. Say the director likes the first half of A[sub]1[/sub] and the second half of A[sub]2[/sub]. It’s a simple matter to cut it in half and use a cutaway to another shot to bridge the transition. This gives the director and editor almost an infinite number of possible assembly options.)
In fact, Jack Nicholson, a very skilled film actor, is well-known in Hollywood for slightly varying his performance from take to take for exactly this reason. He knows he has very little control over what the director and editor will sculpt in the editing process. However, because he knows how this sculpting works, he is careful to give them as much variety of material as possible, because, paradoxically, he comes off better as a result. If the director and editor know what they’re doing, they can take the wealth of material given them by Nicholson and craft a sublime performance out of it. Obviously, as an actor, he’s managing the overall emotional arc in its broad strokes, but he gives the director a number of alternative interpretations they can use to fine-tune the overall performance.
So, to recap: The writer writes one scene. The director and actor create multiple versions of that scene on the set, and then the director and editor pick and choose from these multiple versions to construct a single overall film. Which one has greater input on the final result? In fact, the more you know about how film works, the more you suspect that the editor is just as important as the writer.
Does all of that make sense?