Christopher Nolan gave an interview in Wired where he touched on the issue of ambiguity in fiction (http://www.wired.com/magazine/2010/11/pl_inception_nolan/). He said:
“I’ve always believed that if you make a film with ambiguity, it needs to be based on a sincere interpretation. If it’s not, then it will contradict itself, or it will be somehow insubstantial and end up making the audience feel cheated. I think the only way to make ambiguity satisfying is to base it on a very solid point of view of what you think is going on, and then allow the ambiguity to come from the inability of the character to know, and the alignment of the audience with that character.”
This, I’m aware, is the conventional wisdom. CW sez there’s “good ambiguity” (where the Author knows the Truth of “what really happened” in the story), and “bad ambiguity” (where he doesn’t). But what it leaves out is, How can you tell the difference? If the Author tells you what’s what in an interview, okay, but most stories do not come with Authorial Cliffs Notes (and they would not be counted very good stories if they did). Generally the only way to know what a writer thought is by looking at what he wrote. So any text clear enough that the writer’s intentions can be divined but not so clear as to remove all doubt about the facts of the story is therefore “good ambiguity,” whereas texts where neither the writer’s intentions nor the facts can be divined is “bad ambiguity.” You could make a chart:
1) Writer’s Intent clear, Facts clear = Normal Mainstream
2) Writer’s Intent clear, Facts unclear = “Good Ambiguity”
3) Writer’s Intent unclear, Facts unclear = “Bad Ambiguity”
4) Writer’s Intent unclear, Facts clear = Unknown
The problem is that Case 2 essentially does not exist. If the writer’s intent is clear, then the facts are clear. Knowledge of the writer’s intent makes the facts known. If Nolan’s interpretation of the end of Inception is known (and if we’re to take it as Gospel), then the facts of the end of Inception are known. If you had a story where neither the facts of the case nor the writer’s intentions can be divined (Case 3), that’s “bad ambiguity,” so let’s say you began adding clues that push the story in the direction of revealing your intentions (toward Case 2)—once you’ve added enough clues that the reader can discern your intentions, at that moment the story ceases to be ambiguous (and you’ve reached Case 1). Case 2 can be asymptotically approached but never touched–if you touch it, you are immediately transported through to Case 1. So what CW calls “good ambiguity” is actually not ambiguity at all; it’s certainty. Only what CW calls “bad ambiguity” actually is ambiguity.
I’m reminded of another thing I read maybe 10 years ago, a profile of the writer Penelope Fitzgerald (after research, it turns out to have been very close to exactly 10 years: the New Yorker, Feb 7, 2000). She ended one of her novels with a boy jumping off a wall, practicing for a performance of Shakespeare’s King John. The last sentence is “Meanwhile he went on climbing and jumping, again and again and again into the darkness.” The profile writer interpreted that image as a “hymn to professionalism” of the artist. Fitzgerald, however, said that the ending was meant to have implied that the boy is killed by the fall. Then the profile writer calls it “ambiguous.” But how can the boy have died when he “went on climbing and jumping, again and again”? How can he have jumped a 2nd time if he died the 1st? Regardless of the fact that she wrote it, the text does not support Fitzgerald’s position. What she was thinking when she wrote it is irrelevant; only what words she actually put down matter. My feeling is that it’s not ambiguous, but Fitzgerald believed* ambiguity was essential to the “literary enterprise.” Somewhere along the line people got the notion that ambiguity is superior, more literary, than certainty, and any opportunity to inject ambiguity into things is to be seized. Except people don’t really like ambiguity, so they came up with the further notion of good vs. bad ambiguity, which allows them to call ambiguous things that are really not, that are really only slightly unsure†. They want something where they can know the truth but leave it unproven. That’s fine to want, but it’s not ambiguity.
The alternative is that the Authorial Proclamations are not taken as Gospel, as definitive. But that merely transmutes the problem from one of “How can you discern the writer’s intentions from the text?” to “How can you discern the writer’s intentions from his proclamations?” Or, “Even if the writer tells you what he really thought, can you trust him?” What if, like Fitzgerald, he told you something at variance to what was actually on the page? For instance, what if the creators of that (bad) movie K-PAX said that their interpretation was that Kevin Spacey’s character “prot” is actually just an ordinary man, a mentally ill Homo Sapiens, and not the intergalactic traveler he claims to be and the film very strongly implies he is? Should we believe them over what we actually see on the screen? But the greater problem is that if the Authorial Proclamation is taken as definitive it eliminates ambiguity, but if the Authorial Proclamation is taken as non-definitive then it lacks the power to transform “bad ambiguity” into “good ambiguity.” If the Proclamation is definitive we’re transported into Case 1, and if it’s not we remain stuck in Case 3.
A third alternative is that the writer neither makes his intentions plain in the text, nor in other media, but lets it be known that he has a single sincere interpretation but declines to tell anyone what it is. This is the sole case in which a work may be counted ambiguous in the sense of Case 2, but, honestly, what good is it? What writer couldn’t claim to have been guided by a single sincere interpretation, no matter how contradictory or absurd the narrative? Some extraneous facts, some action taken by some character off-screen, can always be invented to explain even the most outrageous coincidences or other offenses against narrative sense; witness, e.g., the incredible contortions of logic that some Star Wars fans have created in order to justify the events of the prequels.
Moving on, to deal with what Nolan feels are the consequences of not basing an ambiguous story on a single sincere interpretation: that it will contradict itself or be somehow insubstantial and leave the audience feeling cheated. Will ambiguity that is not based on a single sincere Authorial interpretation necessarily contradict itself? It may, but there seems no reason to believe it will necessarily do so. After all, is it possible that, simply by chance, such a work will just not happen to contain any contradictory elements? And if it could happen by chance, it could surely happen on purpose. Non-ambiguous stories may contain contradictory details also; whether a story is ambiguous does not seem to be a determining factor. It is probably true that a story with ambiguity that contradicts itself will be seen as not having been based on a single Authorial interpretation, but that it is self-contradictory is not sufficient proof. The contradictions may have been accidental, or only apparent. There are many works that contradict themselves without even bringing ambiguity into it. See http://tvtropes.org/pmwiki/pmwiki.php/Main/PlotHole for examples. Perhaps an ambiguous story not based on a single Authorial interpretation is more inclined to contradiction, but perhaps we are simply more inclined to think so, as only those stories that are both ambiguous and self-contradictory come to our attention for the latter failing; those stories with ambiguity not based on a single Authorial interpretation that, by intent or chance, managed to avoid self-contradiction, easily masquerade as stories with ambiguity that is based in a single Authorial interpretation. Likewise, it may be insubstantial or leave the audience feeling cheated, but non-ambiguous works may be similarly insubstantial and leave audiences feeling cheated. In any case, if Nolan believes that audiences can tell the difference between ambiguity based in a single Authorial interpretation and ambiguity that’s not so based, then he’s just repeated the fallacy that fictions can make plain their authors’ intentions while simultaneously leaving unclear the facts of the story. Any amount of textual clues sufficient to tell an audience what the Authorial interpretation is, is enough to remove the ambiguity.
*To clarify, Fitzgerald wasn’t “just saying that” because she believed ambiguity is a highly prized literary value; she was “just feeling that” or “just believing that” because it is a highly prized literary value.
†Obviously there are degrees of uncertainty, but simply leaving open a faint possibility is not the same as being ambiguous in the sense employed in literary critique. Hemingway does not actually show Robert Jordan dying at the end of “For Whom the Bells Tolls,” so in that sense it could be called ambiguous, but it seems quite clear that he will die.