Jack Nicholson and Faye Dunaway in Chinatown (1974). Photo courtesy CBS archives/Getty


The art of the plot twist

Some twists infuriate; others are brilliant. But they both use the surprise story as a self-exploding confidence game

by Vera Tobin + BIO

Jack Nicholson and Faye Dunaway in Chinatown (1974). Photo courtesy CBS archives/Getty

The thrill of a genuinely unanticipated plot twist that snaps satisfyingly into place is hard to match. Critics are protective of the good ones, to preserve their magic: ‘Because the film depends so much on the exquisite unravelling of its plot, it would be unfair to describe much more,’ as the film critic Roger Ebert wrote of Chinatown (1974). ‘It would be sinful to spoil its twists,’ said another critic, Richard Brody, some 45 years later, of Jordan Peele’s Us (2019).

When a twist lands wrong, it makes a lasting impression too. A bad twist feels like a cheap trick. It’s shoddy. It makes us feel not just disappointed but ill used. We may be in the world of fiction, but our sense of grievance is real. The finale of the TV drama Lost, in 2010, the culmination of six years of mystery, misdirection and anticipation, inspired this reaction in many of its viewers. Its co-creator J J Abrams complained: ‘For years, I had people praising Lost to death, and now they say: “I’m so pissed at you for the end of Lost.” … I’ve not yet heard the pitch of what the ending should have been. I’ve just heard: “That sucked.”’

Twists are just one kind of narrative surprise, of course. Stories can always surprise us simply by doing something sudden and unexpected. Jump scares are a staple of horror and suspense films because they are reliably surprising in a very basic and straightforward way. Absurdist juxtapositions and wild variations in style can also be surprising, and often delightful. But a surprise twist offers some rewards and faces some constraints that jump scares and absurdism don’t.

With a big reveal, a story pulls the rug out from under you in some important way. You thought one thing was going on, but it turns out it was something else, all along: Darth Vader was Luke’s father. The lost necklace was only costume jewellery. Rosebud was the sled. For a surprise to work as a twist, some new information has to override an old understanding. For the twist to be satisfying, it has to make sense – or at least to feel as if it makes sense – in retrospect. If not, you are likely to feel like Nic Pizzolatto, the creator and original showrunner of the crime series True Detective (2014-19), who once said: ‘I cannot think of anything more insulting as an audience’ than to invest hours or weeks in a story ‘and then to be told it was a lie – that what you were seeing wasn’t really what was happening.’

Stories with surprise twists involve several noteworthy kinds of labour. First, like any story, they ask their audience to put in the time and effort to build up specific ideas about what is going on. We invest our attention and affections in characters and situations. Then, in the wake of the surprise, we’re supposed to undo a lot of that work and do new work instead, reconstructing our understanding of what’s going on to fit the surprising new information.

The stories themselves also have to do special kinds of work, to forestall the kind of undesirable reaction Pizzolatto described. Storytellers have an arsenal of devices they can deploy to help prevent some of these unhappy outcomes. I’ve written about these tactics in my book Elements of Surprise (2018), which is all about how these methods take advantage of general cognitive tendencies and blind spots in the way we think. Their chief goal is to persuade the audience that, first, the pre-twist understanding was a perhaps understandable, but fundamentally mistaken, interpretation of the facts as given, and, second, that the post-twist understanding is entirely consistent with those same facts. They manage this sleight of hand with an obliging confederate: your own brain.

The first and most crucial tool for setting up a twist is the brute fact that we never depend strictly on what we see and hear as we make our way through the world, or as we make sense of what people tell us. We are continually making inferences and filling in gaps. This means that any story has many opportunities to let us draw our own conclusions, while keeping open the option of telling us something different, later on. So, for instance, (spoilers!) the surprise at the end of The Sixth Sense (1999), that Malcolm Crowe (Bruce Willis) is a ghost, is set up in just this way. In the opening sequence, we see Crowe get shot. In the following scene, we see him walking around, uninjured. Time seems to have passed. No one tells us that he has recovered from the gunshot; we infer it. Only later, we can look back and say: ‘Oh, I assumed…’

Then, like a stage magician – or a pickpocket – writers can take advantage of the way we think to keep us looking in one direction while important things are happening somewhere else, burying plot-critical information where it can be uncovered later, without drawing your attention to its significance too early. The hold that stories naturally have on our attention and imagination helps the process along. The psychologist Melanie Green and her colleagues have found that people are more open to persuasion as a function of what they call ‘narrative transportation’. When a story is gripping, vivid, engaging and immersive, we are less on guard, less concerned about whether all the details add up, and more willing to follow a narrative wherever it wants to take us.

When something unexpected happens, however mundane, it kicks off a particular kind of remembering

Once our guard is down, we take more cognitive shortcuts. This fact opens the door to many tactical manoeuvres. In the 1970s, the psychologists Amos Tversky and Daniel Kahneman observed that people often take an approach to reasoning that can be described as ‘anchoring and adjustment’. Presented with some piece of information – the asking price for a condo, early results from exit polls, a sample answer to a trivia question – people tend to use that information as a starting point in formulating their own estimates, hypotheses and predictions. If I ask you to guess the lifespan of the average tortoise, after first asking you whether most tortoises live past the age of nine, you’re likely to name a lower number than your friend who was asked whether they often live past the age of 70. The same cognitive logic is what makes red herrings so effective. They exert a gravitational pull on our speculation. Information can also act as an anchor when we know it’s false or even irrelevant. If you recognise some false trail as a red herring, it can still affect what Tversky and Kahneman described as the ‘availability’ of other ideas. It will constrain your imagination and make some lines of reasoning much easier to see than others.

Characters make particularly useful vessels for misleading information, because they add deniability. A character can lie or be mistaken without making the story as a whole inconsistent or dishonest. This handy fact works nicely with another predictable feature of human cognition: our relatively bad memory for source information. Especially if we’re distracted (say, by some compelling, emotionally engaging story content), we can encounter claims in dubious settings, or even see them being debunked, but later remember only that we’ve heard them somewhere before. Then if we encounter the claims again, that sense of familiarity can make us more likely to take them seriously. The psychologist Daniel Schacter, a major figure in the cognitive and neurological study of memory, calls this failing and others like it the ‘sins of memory’, and stories can use them to plant information in places where we are likely to remember the misleading bits while losing track of the caveats. Fortunately, despite our tendency to lose track of this information, we never fully lose track of the distinction, so if a story shows us that a source we trusted was unreliable, we get it. The source is blown, but the story itself can still be on the up-and-up.

If stories play their cards right, they can even introduce entirely new information at the dénouement but finesse the introduction so that it doesn’t seem to be new. A very common and surprisingly dependable example is the case of the underspecified clue: Sherlock Holmes observes the knees of a suspect’s trousers, but ‘reminds’ Watson of the fact that they were dirty and worn only as the solution to the mystery is revealed. Miss Marple is told only that there was a maid in the house where a murder took place but explains later that she knew what happened as soon as she ‘heard there was a pretty young girl in the house’.

This strategy takes advantage of the fact that, in general, when something unexpected happens, however mundane, it kicks off a particular kind of remembering. Conveniently for storytellers, surprise biases this process by enhancing the retrieval of surprise-congruent information. So we remember – yes, we’ve heard about trouser knees, or yes, we’ve heard about a maid named Gladys – and the pieces fit together. The fact that different revelations would have been equally consistent with the partial information we had at the beginning is harder to see, because that’s not how we’re built to think about outcomes and predictability. Once something has happened, that’s important data about what we should have predicted on the basis of whatever partial information we had available to us.

This way of thinking about the past is a ‘sin’ in some respects, but useful in everyday experience. Imagine that you’re walking in the park, enjoying the sunshine, effortlessly navigating the terrain, when, unexpectedly, you slip. What happened? One theory is that your sensorimotor systems were doing predictive processing, making models on the basis of past experience to anticipate what would happen next. When the model is wrong, we’re surprised. (When the surprise is too small to register with our conscious minds, we borrow a term associated with the great information theorist Claude Shannon and call it ‘surprisal’.) When that happens, the system needs to revise its appraisal of the relevant recent past in light of what it turned out to mean. It makes sense to give less attention to the bits of evidence that were also consistent with other outcomes, because they didn’t serve us well. A story can leverage this tendency to give us loose connections that feel convincing and tight.

It’s fun and interesting to identify the cognitive principles that explain how story structures produce their effects, but authors and screenwriters don’t have to be cognitive scientists to make use of these strategies. Good storytellers know by instinct and from experience that they are effective. And the central principle, above all, is to do what will produce the desired effect. The point is to build a device that works.

But a story can play by all these rules, it can be engineered with structures that seem to its authors to be up to code in every critical respect, and still infuriate its audience. Ebert, for instance, loved both Ang Lee’s Life of Pi (2012) – a sentimental fable that uses an existential twist to raise questions about faith and trauma – and David Mamet’s House of Games (1987) – a tightly structured, sentiment-free thriller with a deceptive sting in its tail – but loathed the twist ending of The Usual Suspects (1995). ‘To the degree that you will want to see this movie, it will be because of the surprise,’ he warned, but its ‘blinding revelation … filled me not with delight but with the feeling that the writer, Christopher McQuarrie, and the director, Bryan Singer, would have been better off unravelling their carefully knit sleeve of fiction and just telling us a story about their characters.’

At least McQuarrie and Singer can comfort themselves that Ebert had a minority view. The writers, cast and crew of Lost are in a less enviable boat. In a recent retrospective, Henry Ian Cusick, the actor who played Desmond, insisted: ‘The show is not about the ending.’ But there are plenty of viewers out there who are, as we say now, going to die mad about it. What went wrong? For some viewers, at least, Lost and The Usual Suspects fumbled their work at the final stage. They let their audience walk away without first thoroughly teaching them to appreciate and accept what has happened. They neglected to cool out the mark.

The more invested you get that there is something wonderful and magical inside the box, the worse it gets

Three years into Lost’s six-year run, Abrams gave a TED Talk about his storytelling philosophy. He was inspired, he said, by the ‘mystery box’ he bought from a magic shop as a child. The box was sold as a deal: $50-worth of magic tricks for $15. In exchange for the savings, of course, you didn’t get to choose what specific tricks would be inside. Lots of vendors offer this sort of lucky-dip discount. It’s a way of unloading some under-performing merchandise while offering customers a fun little see-what-you-get experience. But Abrams never did open his mystery box. He kept it around, unopened, for years. Why? Because it reminds him of his grandfather, but also because ‘it represents infinite possibility. It represents hope. It represents potential.’ And so, as he tells it in the TED Talk, anyway, he had a realisation about his job as a storyteller: ‘What are stories but mystery boxes? … And the mystery box, in honour of my grandfather, stays closed.’

Another way of looking at the mystery box comparison, though, is this: the real reason you might be better off not opening that box is that the contents are very likely to be disappointing. If you open it, you’ll find out that you didn’t get $50-worth of magic, if you count worth by what you would have been willing to spend if you knew what you were buying. You probably won’t even have $15-worth, by that reckoning. You’ll feel cheated, and that won’t feel good. The longer you wait, and the more invested you get in the idea that there is something wonderful and magical and compelling inside the box, the worse it gets. The same goes for stories. When a mystery box turns out to be a pig in a poke, you have a problem.

As I worked on Elements of Surprise, it seemed to me that stories with a twist in their tail were aiming to be a kind of self-healing confidence trick. You can con someone even if that person never realises they’ve been had. Indeed, for a con artist, that outcome can be ideal. But it won’t fly for the twisty tale. A story with a twist has to include not just an act of deception, but also the mechanism by which the deception is unmasked. That combination of elements makes a surprise story what you might call a self-exploding confidence game. What makes it self-healing, if all goes well, is what happens to that explosion. You are supposed to walk away feeling delighted and edified, not tricked and frustrated. The story should contain the engine of its own reconstruction. When I first began studying the self-healing aspect of narrative surprises, I thought of it as a fundamental difference between stories and actual con games. But that was a mistake.

Psychologists, economists, sociologists and anthropologists – not to mention gamblers and con artists – have all observed in their various investigations that human beings have a very finely tuned sense of whether other people are playing fair or cheating. Some evolutionary psychologists have even theorised that humans have neurocognitive architecture that was specially evolved for ‘cheater detection’, so pervasive and automatic is our attention to the possibility that someone, somewhere, is promising more than they have delivered. As it happens, our judgment in these matters is often wrong, but it is passionate, regardless. Other people can do lots of things that we don’t like, but most of us really, really don’t like a cheat. And this seems to be true whether the ripoff is a matter of fact or of fiction – because in either case, it is our real time and goodwill that have been ill used. So a story that wants you to feel good about it afterwards should take a page from the people who know how to manage the same kind of risk in the real world.

When David Maurer’s The Big Con (1940) was published, he characterised the book as a ‘by-product’ of his research on the slang and technical language of the American underworld. His linguistics career ranged over the social dialects of moonshiners, pickpockets, prostitutes, credit fraudsters, morphine peddlers, racetrack touts and many others, but The Big Con is of course about confidence artists. Because their argot was a technical language of their profession, he couldn’t study it without studying the ways and means of the profession itself.

Straight society knew something of those methods before The Big Con – from newspaper reports, from the fiction of Ring Lardner, or in some cases from unfortunate personal experience. But The Big Con was an approachable, thorough and affectionate portrayal of a golden age of ornate con-artistry (already in decline at the time of publication) that went right down to the nuts and bolts, and that was something new. Its cultural influence was considerable. You have probably not read Maurer yourself but, if you do, you’ll find that many of the terms and tactics he documented are already familiar from the scores of movies and novels that have benefitted, directly and indirectly, from his work. George Roy Hill’s movie The Sting (1973) drew so comprehensively and directly from The Big Con that Maurer sued the film studio and the screenwriter; they settled out of court. I learned recently from an essay by Barbara Wyllie that, in 1950, the critic Edmund Wilson even sent a copy to his friend Vladimir Nabokov, telling him that it ‘throws a good deal of light on certain aspects of American life’.

Erving Goffman, one of the most prominent sociologists of the 20th century, read it, too. His classic essay ‘On Cooling the Mark Out’ (1952) draws on Maurer’s account of the later stages of a big con: the first work of grifters is to identify a target, or mark, and gain his confidence. Then the mark can be roped into the criminals’ fraudulent scheme, drawn further in with some small successes, encouraged at last to make a really large investment, and fleeced accordingly. But simply getting the mark’s money is not enough. Having got it, the mob gang needs to make a clean getaway, without reprisals from either the victim or the law.

Life is full of people in need of cooling out: the rejected lover, the employee passed over for promotion

For this reason, the ability to ‘cool out’ a fractious mark is a vital skill. Clever ‘blowoffs’ (sometimes including colourful devices such as a ‘cackle-bladder’, or police officers who are in on the con) can be deployed to prevent a mark from panicking or calling the police until the gang is well out of range. The mark may need cooling here and there at earlier points in the proceedings as well. ‘Suspicious marks are not unusual,’ Maurer reports; ‘in fact all marks are suspicious at first.’ The skilful cooler pours oil on those troubled waters, redirecting, mollifying, distracting and reassuring as needed so that, when it is time to make the big play, the mark goes quietly and willingly, a lamb to the slaughter.

As a sociologist, Goffman points out that even though it’s pretty rare to be the victim of a big con, social life is full of people in need of cooling out. People in this situation include the would-be lover whose advances are rejected, the long-time employee who is passed over for promotion, the convicted felon who must submit to the indignities of imprisonment, and the defeated party in any contest, game or battle. Any time someone is ‘caught out on a limb’, when their expectations are dashed and they feel that their status is precarious and their face is threatened, there is a crisis. If not pacified, they could become violent, make a scene, file a lawsuit, or otherwise lash out and spin out of control.

Institutions and individuals in a position to set off this kind of disappointment often make preparations ahead of time to handle the blow-up. One option is to arrange that things never get hot enough to require cooling: for example, contriving some reason why a sub-par employee is not eligible to apply for an upcoming promotion in the first place. But often this is not possible. The crisis is unavoidable. If so, afterwards the mark must be shown how to accept the new state of affairs and move on with dignity intact. Goffman explains:

For the mark, cooling represents a process of adjustment to an impossible situation – a situation arising from having defined himself in a way which the social facts come to contradict. The mark must therefore be supplied with a new set of apologies for himself, a new framework in which to see himself and judge himself. A process of redefining the self along defensible lines must be instigated and carried along; since the mark himself is frequently in too weakened a condition to do this, the cooler must initially do it for him.

Why is such delicate handling necessary? Goffman sees cooling as the art of helping people accept failure, and suggests that we need this help because failure presents an existential threat to our sense of self. We work together to hide the terror of failure from one another and from ourselves, to usher one another across the mortal threshold of failure into a sort of post-failure afterlife. The person in need of cooling out, Goffman says, is ‘a person who is losing one of his social lives and is about to die one of the deaths that are possible for him.’

But surely things are not as dire as all that for readers or moviegoers who have been disappointed by a lacklustre surprise twist? And, anyway, why should the audience need the kind of consolation society offers to failures? If anyone or anything has failed here, isn’t it the story with the malfunctioning gimmick?

The clever devices I described above – burying information, finessing misinformation, encouraging emotional involvement at key moments, taking advantage of the ‘sins’ of memory – are a kind of risk management. They serve to insulate stories from the charge of deception. But the way they do it is to put the blame on you. That’s a risky play. And as Goffman says, ‘the disappointment of reasonable expectations, as well as misguided ones, creates a need for consolation’ and a process of adjustment. This is why, no matter how annoying and irrational the viewers’ complaints may seem to Abrams, it might not have been wise to leave them reeling from their final revelations with only the end credits and some B-roll footage of the original crash scene for company.

In fact, that final shot of Lost was actually meant to be a cool-out of sorts. Co-showrunner Carlton Cuse told Vulture magazine that, after the initial cut of ‘The End’ was ready to air, Barry Jossen (then head of ABC Studios) was concerned that the finale was too abrupt. ‘I’m worried that we’re going to come out of this incredibly emotional ending of this show,’ Cuse remembered Jossen saying, ‘and then slam into a Proctor & Gamble commercial and that isn’t going to be good. Is there any way to soften that or ameliorate that?’

Once we know something, it becomes impossible to fully remember or imagine what it is like not to know it

The footage they put together, a montage of shots of plane wreckage on the beach, accompanied by the sound of crashing ocean waves nearby, ‘was just to create a narrative pause,’ Cuse said. ‘But it was too portentous … I think we could have done some things to make it clear that that wasn’t what you were supposed to take away. But one of the big intentions of the show was intentional ambiguity and giving people the opportunity to digest and interpret Lost as they want to if they wanted to. And at some level, you know, you can’t have it both ways.’

After my book Elements of Surprise came out, two different writers contacted me with roughly similar requests: they had both been scammed by talented con artists and wanted to talk to me about how the cognitive science of plot twists could help them understand their experiences. One, Stephanie Wood, has written a book and several articles about being taken in by the elaborate fake life of a man she met online. The other, Bruce Grierson, was hooked by an imposter scam, the kind of thing where you get a phone call with a bad connection, lots of noise on the line, but you can hear the distress in the voice as someone says: ‘Grandpa? I’m in trouble. Please don’t tell mom and dad!’ In Grierson’s case, it was an email from – it seemed – the minister of his church, with a story about iTunes gift cards for an ailing friend. He lost $300 and gained a story for The Walrus magazine.

The con artists in The Big Con are fond of the maxim ‘You can’t cheat an honest man.’ That claim holds true for some kinds of confidence tricks, perhaps, but not these. It’s also a nice thing to tell yourself if you happen to cheat people for a living. I felt a little fraudulent myself, asked to provide any kind of insight into the aftermath of real-world confidence tricks. My research is about exactly how stories provide a safe space for playing with the kinds of tricks and deceptions that do real harm when they’re deployed in earnest. My correspondents were suffering from the genuine article. What good could they get from poking around in the machinery of the toy version? Both said that they found it reassuring to learn how efficiently fictions pull our levers, to see the parallels between the lines they swallowed and the garden paths audiences go down.

Most of all, they were comforted by the fact that the most effective twists capitalise on what the behavioural economists Colin Camerer, George Loewenstein and Martin Weber called ‘the curse of knowledge’: once we know something, it becomes impossible to fully remember or imagine what it is like not to know it. Our earlier selves seem gullible and stupid by comparison. How could we have failed to see what was coming? When we have this reaction to a story, it makes the plot seem brilliant and revelatory. When we have it to events in our own lives, we are more likely to feel guilty and ashamed, culpable in our own downfall.

I see now that Wood and Grierson were expertly cooling themselves out, in Goffman’s sense of the term. As Goffman says, when our dignity is compromised, we ‘can seek comfort in one role for injuries incurred in others’. And in this way, it turns out that surprise twists have two kinds of value. One is the charm of the surprise itself. When a story serves up a convincing surprise, we get to have an ‘aha’ experience, an (engineered) epiphany about a deeper, truer interpretation of what went before. The other, less obvious, benefit is the opportunity to see ourselves as suckers without shame. We’ve been dupes, then learned that we were duped (and how). We know better than to think we’ll never be fooled again; the important thing is to find a good use for that wooden nickel.