The Usual Suspects (1995) Review
8/10
Five criminals are pulled into a lineup that should never have happened, plan a job together, and end up working for a man who may not exist. One survivor tells the story to a customs agent in an interview room.
Everything you watch is that survivor’s account, and the account is being assembled in front of you out of whatever happens to be in the room.
Christopher McQuarrie wrote it, Bryan Singer directed it, and it was made for six million dollars by two men in their twenties who had made one film before.
The 8 accounts for a construction that works beautifully once and rewards a second viewing differently. Knowing the ending converts the film from a mystery into a performance, and watching Spacey build the story out of the furniture is better than being fooled by it. What holds it back is that the middle is a competent heist picture and little more, and the boat sequence is genuinely confusing instead of intriguingly opaque.
The Noticeboard
The reveal is not that Verbal is Keyser Soze. That is the surface of it. The reveal is where the story came from.
The names, the details, the whole architecture of the tale is lifted off a noticeboard on the wall behind Kujan’s head. Redfoot. Kobayashi, from the bottom of a coffee cup. A barber shop quartet in Skokie, Illinois, from a clipping. Guatemala. The camera has shown you all of it, repeatedly, without comment, for the entire film.
That is what makes it hold. The film never lies to the audience. It shows you the source material in plain sight for two hours and lets you assume it is set dressing, which is exactly what set dressing is for.
The manufacturer’s mark on the bottom of the coffee cup is the last beat, and it lands because it is the moment Kujan sees what the audience has been looking at all along.
The final sequence, where Verbal’s limp straightens and his hand uncurls as he walks, is the most quoted image in the film and it is the least important part of the trick. The gait is the confirmation. The noticeboard is the mechanism.
It is worth watching the interrogation scenes with your eye on the background instead of the actors, which is a thing you can only do the second time.
Hide the answer in view and let the reader’s assumptions do the concealing. A clue in plain sight that the reader dismisses as background is entirely fair, and it produces the good kind of second read. A clue withheld and produced at the end is a trick, and readers can tell the difference even when they cannot articulate it.
The practical method is to give the concealed information a plausible reason to be present. The noticeboard is in the room because police stations have noticeboards, so it never reads as planted. Ask yourself why each of your clues is where it is, in the world instead of in the plot, and if the only answer is that the reader will need it later, it will show.
The test to run is a reverse read. Go through your draft knowing the solution and mark every place the answer is visible. If there are fewer than half a dozen, the reveal will feel arbitrary. If any of them only make sense in retrospect, you have written a cheat. What you want is a set of details that are honest on both passes and mean different things on each.
An Unreliable Narrator With a Reason
Most unreliable narration comes from delusion, self-deception or damage. Verbal’s comes from strategy.
He is lying deliberately, for a purpose, to a specific listener, under time pressure, with a deadline he can see on the clock. That changes the texture completely.
Every detail is chosen for its effect on the man across the table. Kujan wants Keaton to be guilty, so Verbal gives him a Keaton who could be. Kujan is contemptuous, so Verbal is pathetic. The story is being tuned in real time to an audience of one, and the audience is getting exactly what he came in wanting.
It also solves the problem of why the story is so well shaped. A frightened witness giving a coherent two-hour narrative with foreshadowing and reversals would strain belief. An excellent liar constructing one, with a professional interest in it holding together, is exactly what you would expect.
Kujan’s own conviction is the weapon. He tells Verbal repeatedly that Keaton is behind everything, and Verbal simply agrees more and more elaborately.
The film’s cleverest line is Kujan boasting that he is smarter than Verbal and will find the truth. He says it to the man he is being told a story by.
An unreliable narrator with a motive is far stronger than one with a condition, and the difference is that a motive gives you a shape. A deluded narrator distorts in whatever direction their damage points, which is diffuse. A strategic narrator distorts in a specific direction for a specific reason, which means every distortion is a decision the reader can eventually audit.
Build the listener first. Verbal’s story only works because Kujan has arrived with a conviction, and the lie is shaped to fit it. If your narrator is deceiving somebody, work out what that person already believes and wants to be told, then construct the lie as flattery of their existing theory.
The additional benefit is that this makes the deception detectable on a second read without being detectable on a first. A reader who knows Kujan’s assumptions can watch Verbal serve them, which is the pleasure of rereading, and a reader who does not simply follows the story.
The Lineup Was an Accident
The most famous scene in the film is the one nobody wrote. Singer scripted the lineup as a grave identification sequence and intended to shoot it that way.
What happened instead was that the five actors could not keep composure through a full day of filming. Kevin Pollak has said on the commentary that the cause was Benicio del Toro breaking wind through roughly a dozen consecutive takes without ever admitting to it. Singer was reportedly furious, then gave up and assembled the scene from the funniest outtakes.
So the version in the film is genuine laughter from five men who found each other ridiculous, which is why the rapport reads as lived instead of performed. You cannot write that and you cannot direct it.
Del Toro had separately decided that Fenster’s dialogue was meant to be unintelligible and played it that way, with the other actors encouraged to improvise responses to lines nobody could parse. That choice survives into the finished film as a character.
The part that interests me most is what McQuarrie has said about it. He had been told the film needed a scene, could not solve the problem, and the accident solved it for him. He has been public about the debt and about the fact that it improved the script instead of merely the sequence.
There is also a structural benefit nobody planned. A lineup assembling five specialists from different cities on a charge that will not stick is a very strange event, and played as farce it never invites the question of who arranged it. Straight, an audience asks immediately. The accident buried a plot hole and built the ensemble in the same take.
The transferable lesson is not that accidents are good, which is useless advice. It is that a writer should recognize when something unplanned has solved a problem better than the plan would have, and then rewrite toward it instead of defending the original intention.
McQuarrie had a structural gap and a scene that was not working. What arrived was tonally wrong, unusable by the standards of the script he had written, and better. The professional move was to accept it and adjust everything around it, which is what happened. The amateur move is to shoot it again properly.
In prose the equivalent is the sentence, character or scene that goes somewhere you did not intend during drafting. Most writers cut those on the revision pass because they violate the outline. Read them again first and ask what problem they might be solving that you had not admitted you had. The outline was a guess made before you knew the book.
What Keyser Soze Is For
The monologue about Soze murdering his own family to demonstrate his will is the most memorable thing in the film and it is almost certainly invented on the spot.
It is also the most useful thing in the film, structurally, because it establishes an antagonist who cannot be found, cannot be predicted, and might be a rumour. That gives Verbal permission to explain any inconsistency by attributing it to a man nobody has seen.
Which is precisely what a liar needs. If the story requires something implausible, Soze did it, and Soze is by definition beyond assessment.
The line about the greatest trick being convincing the world he did not exist is doing double duty. It is the character’s boast and the film’s instruction manual.
McQuarrie has said the name came from a colleague at a law firm and the character from an idea about a criminal so effective he becomes mythological. The mythologizing is the whole plot.
The film also never confirms it. Verbal walks and Kujan is left holding a story, and there is a reading in which Soze is real and Verbal works for him. The film does not close that door.
Why the Twist Survives
Twists that depend on surprise are worth exactly one viewing. This one depends on construction, so it converts.
Watch it a second time and the pleasure is different but comparable. You see Verbal hesitate before names. You see him glance at objects. You notice that the flashbacks are shot with an authority and coverage the storyteller could not possibly have.
That last point is a small cheat that most people accept. The flashbacks show us things Verbal was not present for, in perfect detail, which is only defensible if you read them as his invention instead of his memory. The film wants both, and the second reading makes it work.
A twist that only works once was a magic trick. A twist that reorganizes the film is a structure, and the difference is whether the second viewing gives you something the first could not.
McQuarrie won the Academy Award for the screenplay and Spacey for supporting actor, and the film went from a small release to a phenomenon over several months, largely by word of mouth about the ending, which is the one thing you cannot tell anybody.
It has been imitated constantly and rarely well, because most imitations copy the reveal and not the two hours of honest groundwork that make it fair.
Should you watch The Usual Suspects?
Yes, and if you already know the ending, watch it anyway with your eye on the backgrounds. The middle is the weakest part and the boat sequence is muddled. The interrogation room is the film, and it is better on a second pass than a first.
Frequently Asked Questions
This film appears on 6 of my film lists