The director must know the truth, but that includes knowing whether or not there’s an answer. I would argue that the “reality” behind Memento is ultimately unknowable, given the evidence in the film; based on this or that shot you can make a case for either hypothesis (Sammy = Lenny and Lenny killed wife, vs. Sammy is real but Lenny is confused). But the thing is, that’s the point of the movie. The central theme of the film is the slippery nature of memory, and how things you’re absolutely certain you remember happening may have happened in a different way, or may not have happened at all. The film is constructed, I believe, to make it appear that it can be untangled, but that it resists a clean and unambiguous restructuring; in order to support one or another hypothesis, you must interpretively discard certain evidence (“the scene at the end with him and his wife is a fantasy”) while relying on equally valid evidence as being “true.” But Nolan knows all of this, and he knows what the movie is about, so the film is constructed to illuminate the theme, and not necessarily the “reality,” whatever that is.
To see a “puzzle movie” where the filmmakers don’t give a shit whether or not everything hangs together, look at Basic. Random plot twists thrown in for the sake of confusing the audience, and a final “gotcha” that makes no sense. They’ve looked at the success of The Usual Suspects and The Sixth Sense and completely misinterpreted why they were successful: “Hey, people must like being surprised and confused! Yeah, we can do that!” But that’s typical in Hollywood: imitate the most obvious elements of a successful movie. “Hey, look, American Pie was successful, and it had lots of teenagers, gross-out humor, and goofy sex gags! Greenlight Road Trip!”
I’m getting off the subject. My point is, the director absolutely, definitely needs to know what the film is about, but that doesn’t necessarily mean the movie has to reveal what actually “happens” in “reality.” Cinema is a much more flexible medium than that. For example, during the recent Seattle film festival, I saw a pretty good movie called Spring Subway, from China. Regarding this film, I wrote:
In other words, this is an example of a film where trying to figure out “reality” isn’t just a waste of time, it’s actively wrong given the motives of the film.
However: Regarding “director’s reality,” it’s sometimes true that what the director tries to impose, usually in comments after the fact, will undercut and/or minimize the movie. Look at Blade Runner, for example; Ridley Scott has lately been saying that Deckard is supposed to be a Replicant. Now, while that’s a supportable interpretation of the film (the director’s cut, anyway), I don’t think it’s nearly as interesting as the ambiguity of the actual film minus his comments. The movie certainly has a preoccupation with the nature of identity; note the scene where Deckard wakes up on the piano and slowly looks across all the portraits and snapshots arranged in front of him. He doesn’t say anything, but we’re supposed to recognize that he’s thinking about his own history, and how the Replicants are pitied because they collect photographs as false but reassuring evidence of a nonexistent life but that he’s got a collection of photos with much the same emotional motive. This is a case, I would argue, where asking the question is much more interesting than answering it.
Some people like ambiguity in their films; they like having to think about what happened, and they recognize that the answers they come up with say more about themselves than about the film. Other people don’t like ambiguity; they want the movie to tell them everything that happened and why. I fall into the first group; people in the second group didn’t like The Blair Witch Project because it never explicitly explained that the witch was making these kids wander around in a forest they should have found their way out of. Neither side has a superior position. It’s just two different ways of watching movies.
