In my talk at the Narrative Conference in Atlanta last year, I presented a reading of the last three chapters of the biblical book of Judges that argued that this episode is for at least a certain class of readers radically ambiguous or indeterminate, like that image of the rabbit-duck that shifts in our gaze from one configuration to another. In Emma Kafalenos’s terms, the sjuzet gives rise to two inconsistent fabulas; in James Phelan’s terms, the text is "stubborn" because it demands to be read two ways that are intellectually or emotionally incoherent with each other. On one level, the text is "sacred history" presenting the rape and murder of a nameless concubine, a crime that causes a widespread and horrendous civil war. This narrative is moralized, or rather politicized, by the last line: "This was when there was no king in Israel, and everyone did what was right in his own eyes." Surely if these events were typical of what went on in the days of the judges, then Israel cannot have kings too soon. But about a third of the way into the story the events of the narrative begin to strain credulity, and their locales—Bethlehem, Gibeah, Ramah in Ephraim—begin to suggest that the story should be read as a satirical allegory directed against the first king of Israel, Saul of Gibeah. The politics of the lampoon, however, contradict that of the sacred history for it satirizes the moral defects and inanities of monarchs rather than judges.

As a narrative, the episode in the last three chapters of Judges simultaneously generates two authorial readings that are not only inconsistent but emotionally incoherent with each other. Flesh and blood readers are at liberty to refuse either reading but both audiences are welcomed within the text, and even a reader who refuses to join one of the audiences may on some level be aware of the temptation. For one authorial audience, the concubine is a real person with a place in history; and this audience experiences the horror of her death and dismemberment. For the other audience, she is a symbolic agent in a parodic or satirical narrative, no more to be taken as real (and no less vivid) than the Yahoos about whom Lemuel Gulliver told us in the fourth book of his Travels. Rimmon presents these texts as undecidable, and of course I agree with that. But her way of analyzing narrative hasn’t any way of distinguishing prospective from retrospective configurations of the sjuzet, or analyzing in any really effective way the phenomenology of indeterminate texts. Perhaps the reader, deciding whether to opt for the a+ or b+ interpretation is always like a hungry donkey placed equidistant between two bales of hay, starving because unable to decide in which direction to turn.

But perhaps not. For many texts, like my own text from Judges, it seems to me that the reader attempting these texts cold is pretty much forced to start out in one fabula and only gradually becomes aware of the existence of the other after a particular event or series of events which calls that fabula into question and suggests a different configuration of the narrative, one that operates retrospectively as well as prospectively. For my text in Judges, I posited that the point at which the "sacred history" reading of the Concubine narrative receives a possibly-fatal shock is the moment when the Levite addresses the recumbent corpse of his wife with a cheery "Up, and let us be going" as though he were in a hurry to beat the rush hour traffic on the road from Gibeah to Shiloh. Similarly, the crucial moment in my own first reading of "The Turn of the Screw," when I became aware that there was perhaps another way of configuring the narrative, was the moment when Mrs. Bread failed to see the ghost of Miss Jessel.

[Of course there is always the question of whether one can ever read some of these texts cold any more. On one side biblical texts like Judges are presumed to be sacred history by most readers coming from either Judaism or Christianity—whether the values of that history are ones the readers want to spread or undercut—so positing the existence of literary genres like satire already has a strike or two against it. And on the other side, "The Turn of the Screw" is so thoroughly positioned as the indeterminate text par excellence that there cannot be too many readers coming to it without knowing in advance the figure that is supposed to be in its carpet. Nevertheless those who might take Rimmon’s rabbit/duck analysis as an analysis of gaps and the way readers fill gaps that would be true without regard to the audience or the era of reading need to explain why it was that a novella published in 1890 was not read as a story of a hysterical governess with hallucinations until Edith Kenton’s article in 1926. Were ducks invented during the ‘20s? The history of reading, though, is a can of worms I don’t want to get into right now.]

That narratives can be radically ambiguous is nothing new. Indeed narrative ambiguity has been analyzed in considerable detail in Shlomit Rimmon’s The Concept of Ambiguity: The Example of James, in which she presents as a coherent structure of gaps and inferential walks the mutually exclusive fabulas underlying the sjuzets of "The Figure in the Carpet," "The Sacred Fount," and that president-in-perpetuity of the club of ambiguous narratives, "The Turn of the Screw." Each of these texts is a rabbit/duck in that it can be coherently read in different ways: either there is a secret code at the heart of Vereker’s work, or there is not; either the ghosts at Bly are nefarious spirits in league with the children or they are hallucinations fabulated by a hysterical governess.

Nevertheless, the reconfigurations of narrative raise a host of interesting questions for reader response. The one I want to take up here is the question of timing. Just how far into a narrative can the audience be made aware that there is another fabula consistent with the events narrated? The closer to the denouement one places the event which causes the alternative configuration to come into focus the greater the danger of driving the gentle reader beyond the bounds of patience. Narratives in Chaucer’s day were frequently framed as a dream, but the narrative sin that every schoolgirl knows to be lame beyond redemption is to work up an exciting situation only to resolve it with the words "And then I woke up." <Apart from Keats’s "Ode to a Nightingale," the only canonical text I can think of that solves the problems of closure by having the protagonist awaken from a dream is Alice in Wonderland.>

Of course once the quintessential ineptitude of the "And then I woke up" ending has been sufficiently established, it becomes subject to what Peter Rabinowitz would perhaps call the Rule of Retroactive Revocation, so it can be used with a wink to the audience as a way of abruptly ending a plot, especially when that abruptness is known to stem from external circumstances. At least two television series that had been cancelled by their networks ended that way. The hospital drama, St. Elsewhere, ended with a suggestion that the whole series had been merely the dream of an autistic child. <During the final credits, the MTM kitten lies on a heart monitor, flatlines and expires.> And the comedy series Newhart (the one in which Bob Newhart is a Vermont innkeeper), when it was cancelled in 1990 ended with an episode in which Bob wakes up in the same bedroom he had shared with Suzanne Pleshette throughout the earlier Bob Newhart Show (the one where he is a Chicago psychiatrist), blinks his eyes and says: "Emily, I’ve had the craziest dream!"

"And then I woke up" has admittedly limited utility as a way of reconfiguring a sjuzet into a different fabula, but a fair number of recent Hollywood films have chosen to force the audience to reconfigure the narrative starting at a point close to the end of the last reel. I shall be discussing three films that do this, each with different strategies: The Usual Suspects, The Sixth Sense and Fight Club. None of these is a rabbit/duck in the sense that "The Turn of the Screw" and the Concubine of Gibeah story are: they are all rabbits that suddenly turn into ducks. What I thought was particularly interesting was the different ways audiences seem to have reacted to these films and their surprises.

The Usual Suspects is an extraordinarily complex "caper flick" in which five criminals, having successfully pulled one heist in New York, are then forced into another in L.A., and finally into a bloodbath aboard a freighter in San Pedro at the behest of a mythic supercriminal known as Keyser Söze. Much of the film is told in voice over flashback by the only surviving perp, known throughout as Verbal (Kevin Spacey) and the moment of awakening occurs after Verbal leaves, when the customs agent who had been listening begins to notice how details we have seen on the screen and heard in voice-over line up with the contents of the borrowed office being used for the interrogation. The events we took to be fact at the level of the narrative audience, must be reconfigured as fictions improvised by Verbal. WHO IS KEYSER SÖZE??? was one ad for the film, and we certainly have enough information independent of Verbal's story to feel certain that he exists and that he is Verbal. But as soon as the film is over we discover we have little solid information about who did what and why they did it.

Just how disturbing is it to the audience to discover that much of what it has been taking as "real" must be reconfigured as a fiction within the larger fiction? I read through a large number of print and web reviews, and the answer seems to depend on how well one has been following the intricate plot. Paradoxically, perhaps, those who registered irritation and scorn at the ending of the film (including Roger Ebert) were also generally annoyed at the complexity of the events, at how hard it was to take everything in. <Ebert said: "When I began to lose track of the plot, I thought it was maybe because I'd seen too many movies that day…. Verbal lives up to his name by telling a story so complicated that I finally gave up trying to keep track of it, and just filed further information under ‘More Complications.’"> Whereas those who were delighted by the fact that they were watching a movie that defied one to go outside for popcorn were knocked over—positively—by the surprise ending, despite the fact that the actual events leading up to the shipboard slaughter thereby become completely indeterminate.

The fabula as originally configured made sense. It was an example of a well-known plot-form that might be titled "Getting In Over Your Head," where each successive heist puts the group more into the power of the mysterious Keyser Söze, whose aim is to kill the one man that can "finger" him to the FBI, even if his five adventurous stooges all get killed in the process. But if Verbal is Keyser Söze it isn't clear why he has done any of it. The initial premise -- that Verbal has been given immunity from prosecution for murder and armed robbery because of the Prince of Darkness's pull with the Powers That Be -- makes nonsense of any notion that any witness, whatever he might have to say, was going to endanger a gangster so powerfully positioned. So it suggests that all Verbal/Keyser is doing for the two hours during which he narrates his story is having his own version of fun with the intense customs agent.

But if that's what is going on, the sort of fun he has, interestingly enough, is thematically consistent with the portrait of Keyser Söze delivered in the flashback. What is said to be unique about him is that he knows no limits--he is willing to do what others are not, including killing his own family to illustrate how impossible it would be to control him. Verbal's storytelling performance culminates in a similar total askesis: he makes himself out to be a coward, a weakling and a rat who has leaked the entire story of the crimes rather than holding his peace. The performance has the agent's insisting that Keaton (played by Gabriel Byrne) was Keyser Söze -- and that he left Verbal alive because he was no threat to him -- though it isn't clear how a bent New York cop with an Irish accent could possibly have been identical with a mysterious Hungarian gangster.... The bravura plot and the bravura of the improvised story connect up so that instead of considering the story a mere lie we think of it as a fiction within the fiction of the film as a whole, and Keyser Söze as an author, rather than a mere liar.

The Sixth Sense has a different sort of reversal which, ideally, one needs to learn at the same time the psychiatrist does, a few minutes before the end of the film. In expiation for being unable to help a young schizophrenic patient who commits suicide after attempting to murder him, a psychiatrist has been giving therapy to a similarly disturbed child, who thinks he "sees dead people." So the psychiatrist thinks: but in fact he died of the wound we see him receive at the beginning of the film and is one of the "dead people" the boy senses "want something" from him. The Sixth Sense follows one of the standard "rules" of ghost stories: that ghosts appear because they are pursuing unfinished business and retreat into the Great Beyond once that business is completed. Part of the psychiatrist's unfinished business involves helping the boy learn how to cope with the ghosts, and helping involves, among other things, allowing the boy to help him with the rest of his own unfinished business. The psychiatrist's realization occurs after he helps the boy find the courage to help two other ghosts accomplish what they need to, and he is helped in his turn, since for the first time in the film he is able to communicate with his widow who—up to this point—we had seen as a joyless, estranged, and possibly adulterous wife.

Is second sight a blessing? Not always. In two dozen pieces of reception I read, many viewers claimed to have guessed the truth long before the ending. This is not surprising since the script plays as scrupulously fair a game with the audience as a classic murder mystery. (For example, the boy remarks to the psychiatrist that the dead people he sees don’t know that they are dead.) But although this was (at least for me) a film one can see again with pleasure, those who had "guessed" where the film was going the first time around also lost the poignance of the denouement and found themselves waiting impatiently during a series of meandering moves for the film to reach that foreseen conclusion.

There was also considerable hostility directed against what some viewers thought of as "cheating." One viewer was bothered by the fact that it would be impossible for the psychiatrist to remain unaware that he is dead since, if one were invisible, one would quickly notice that fact. (The central action of the film begins six months after his death, which even in Philadelphia would be plenty of time to notice that taxicabs aren’t stopping for one anymore.) One may be skeptical whether plots involved with the supernatural can ever cheat, since there are no hard and fast rules that can be broken.

The rule that is broken in The Sixth Sense is not a rule about ghosts, it is rather a narratological rule. In effect one focalizing character’s awareness has been limited to what the audience is allowed to be aware of – which is another way of saying that the fabula that the character must be experiencing corresponds the same sjuzet we are experiencing (rather than an experientially different fabula underlying or abstracted from that sjuzet). Normally we assume that the sjuzet is in mathematical terms a dependent variable: it is a sequence of scene, dialogue and summary crafted to optimize the potential impact of the fabula. But here the fabula depends on the sjuzet; that is, the psychiatrist's experience is limited to our experience of him, which itself has been carefully crafted to avoid showing anything that would require us to discern that he is a ghost.

The director-writer M. Night Shamalayan doesn't seem to have been aware of the problems caused by this sort of cheating. His concentration was on a different matter of internal consistency -- the way his storyboards arranged people so as to give the impression that the psychiatrist was interacting with people other than the boy without actually showing any such interaction. For example, he is seated in an identical chair opposite that on which the boy's mother sits when the boy walks in after school, so that the two adults seem to have been conferring with each other while they never actually exchange a glance. Similarly, when the psychiatrist sits down at a restaurant with his wife, for example, the director took care to place the chair so that he doesn't have to move it to sit down--ghostly movement that would be perceptible to other diners. It is possible that we may be cued that the psychiatrist is dead by the fact that the table is set for one rather than two, though the first time we see the film we are more likely to pay attention to his guilt over arriving late for an anniversary dinner.

What the focalizing character knows and when he knows it should be an issue even more intensely provoked by Fight Club. Fight Club is reconfigured when it is revealed that the nameless narrator (called Jack in the film) and the central figure, Tyler Durden, are two personalities of one individual in the grip of multiple personality disorder. Unlike "Dr. Jekyll and Mr. Hyde," and other literary and filmed representations of MPD, however, the two personalities appear at the same time and are played by different actors, the heartthrob Brad Pitt and nearly anonymous Edward Norton. <A closer resemblance to Fight Club might be the representation of the two Conchitas in Bunuel’s Cet Obscur objet du désir, although the cool Carole Bouquet and the hot Angela Molina never appear at the same time—and the joke is that Fernando Rey cannot see any difference between them.> The scenario insists on the literal identity of Tyler and Jack, and replays as flashback bits of a revised sjuzet with Tyler absent. At one point Jack, at Tyler’s insistence, hits Tyler as hard as he can—so after the revelation the episode is replayed with Jack alone in a parking lot, absurdly smiting his own neck.

This sort of revelation is easier to take in a book than in a film, because the mind's eye is more forgiving than real vision. But, although some people were hard on Fight Club, usually for its violence, few reviewers objected to this clearly outrageous aspect of the film. Revisiting the film one sees that the screenwriter threw out broad hints about the identity of Tyler and Jack, but nothing that could help the viewer discount the ocular effect of two leading men visible at the same time playing the same person.

The reason this reconfiguration causes so little trouble is genre: Fight Club is a satire whose progression depends on an argument being taken to its furthest point rather than on a plot being taken to its conclusion, so our communion with the authorial vision is not primarily narrative. To put it another way, within the narrative Jack and Tyler merely represent two attitudes towards what Chuck Palahniuk (who wrote the book) and screenwriter Jim Uhls portray as the soullessness of postindustrial society where people hold down jobs they hate to buy stuff they don’t need. The attitudes might be paired as passive-aggressive and aggressive, and since the passive-aggressive already contains the aggressive, Jack in effect already contains Tyler. We have come to understand this long before the film literalizes the point. Plot-oriented progressions are indeed used for closure, e.g., via the deepening relationship between Jack and Marla, who are finally left holding hands as lovers while the skyscrapers of Wilmington collapse all around them. But this seems less in keeping with the narrative than a final joke aimed at that dimwitted part of the audience that needs a happily-ever-after. <There were trailers that advertised Fight Club as though it were a romantic comedy. I assume nobody asked for their money back.>

My own conclusion is one in which nothing very much can be concluded. Forcing the reader to reconfigure the plot at the very end of the representation hurts least when plot isn’t the most important factor in the progression. It hurts most when the plot is all-important but the viewer hasn’t been paying very much attention to it, which may be a corollary to the rule that those who don’t much care who killed Roger Ackroyd, as Edmund Wilson didn’t, aren’t apt to be terribly impressed by Agatha Christie's transgressive ending either. And it also hurts a lot if the audience has been paying too much attention—as with those who guessed too soon that the psychiatrist in The Sixth Sense was a ghost. There are beautiful stories in which we watch a focalizing character learn what we have known all along (such as The Beast in the Jungle) but not when what he learns is a literal fact--even if that fact should be a matter of life and death.