Se7en (1995) Review
9/10
Official trailer
It rains for the entire film except at the very end, when the sky clears and they drive out to open country, and by then you know that can’t be good. Two detectives, seven murders arranged around the deadly sins, and a killer who’s several steps ahead throughout and never once in doubt about what he’s doing.
I think Se7en is the best-made thriller of the nineties, and what interests me is how little it cares about being a thriller. It spends its energy on two men arguing about whether any of this is worth it, and by the last ten minutes the mystery has dropped away and I’m watching one of them lose the argument in the worst way there is.
What separates it from the hundred imitations is that it’s a film about whether the world can be borne, with a procedural attached. The murders are the delivery mechanism for an argument between two men who have looked at the same evidence and reached opposite conclusions.
The 9 is for a film that commits to its argument and refuses the exit it’s offered in the final ten minutes. The gore is almost entirely implied, the killer is barely present, and the ending is one of the very few in mainstream cinema where the antagonist gets exactly what he wanted. The point withheld is for a second act that leans on coincidence more than the first.
The Opening Nobody Skips
The credit sequence is scratched film, obsessive handwriting, close-ups of hands cutting and pasting and stitching, all cut to a degraded remix of a Nine Inch Nails track. It runs about two minutes.
Kyle Cooper designed it and it changed title sequences permanently. Before this, credits were typography over an establishing shot. After this, they were a place to do character work.
The economy of it is the achievement. By the time John Doe appears in person, more than ninety minutes in, he needs no introduction at all, because the audience has spent the whole film inside the evidence of his method. You’ve seen his hands and his handwriting and you know what kind of patience he has.
It also announces the film’s atmosphere before a single scene. Everything will be wet, dark, decaying and hand-made. Nothing will be clean.
I rank this sequence above anything else of its kind, because it does real work. Most films give you credits to wait out, and this one gives you the killer’s handwriting. That’s a far creepier introduction than any scene with him in it could manage. There’s a practical benefit too. A killer who’s absent for most of a film is a hard sell to an audience, and the sequence buys Fincher the room to keep him offscreen.
An opening should establish the rules of the world instead of the events of the plot. Two minutes of somebody’s hands and handwriting told the audience who they were dealing with more efficiently than any scene could have, and it bought the film ninety minutes of the villain’s absence. Ask what your first page teaches the reader about how this book will behave, and whether it earns you anything later.
The practical version is to decide what your first page is buying. Se7en’s title sequence buys the villain’s absence for ninety minutes. That’s a structural purchase instead of an atmospheric one. Fincher could keep Doe offscreen because the audience already had a sense of his patience, and an audience with that sense will wait.
So ask what a reader must believe by the end of your opening for the middle of the book to work, then build the opening to establish exactly that and nothing else. Openings fail when they try to establish plot, because plot is what the following chapters are for. Rules, atmosphere and character are what only a beginning can do cheaply.
The City Has No Name
It’s never identified. There are no landmarks, no local references, no license plates that resolve, nothing to place it. Fincher and the writer Andrew Kevin Walker have both confirmed this was deliberate.
A named city invites the audience to think about that city and its particular problems. An unnamed one becomes every city. That’s what the argument requires, because the film is asking whether urban indifference is a survivable condition instead of whether one place has gone wrong. It’s also why the film hasn’t dated. There’s no technology, no politics and no geography to tie it to 1995.
The rain does similar work. It rains without explanation for the entire running time, and it makes the city feel like a condition instead of a location. When it finally stops, the film has moved somewhere with no shelter at all.
Darius Khondji shot it, and the technique was to underexpose and then use a silver retention process in the printing. That crushes the blacks and makes the highlights sit oddly. It looks like nothing else from that decade and it still looks deliberate instead of dated.
Somerset’s line about apathy being a solution because it’s so much easier is the thesis, and it’s delivered quietly, in a bar, by a man who has already decided to leave.
Two Detectives, One Argument
Somerset has thirty years of this and is retiring in a week. Mills has just transferred in, requested the city, and believes the work is worth doing. The film is scrupulously fair to both. Somerset isn’t a burnt-out cliché; he’s the more intelligent man, reads Chaucer and Dante in the library to understand the case, and his position is thoroughly argued. Mills isn’t naive; he’s competent, brave and correct that somebody has to care.
Their disagreement is the actual content of the film. Every scene between them is the same conversation from a different angle: whether looking squarely at this material obliges you to withdraw from the world or to fight it.
The dinner scene at the Mills apartment is the hinge, and it’s the only warm sequence in the picture. Tracy invites Somerset, the train shakes the building, they laugh about it, and for ten minutes there are three people who like each other. That scene exists so that the ending will work.
Gwyneth Paltrow’s Tracy is given one serious scene, the meeting with Somerset in the diner where she tells him she’s pregnant and doesn’t want the child raised here. It’s the most important scene in the film and almost nothing in the plot depends on it.
Paltrow’s diner scene is the one that stays with me. She barely raises her voice, and she tells you exactly what’s at stake so quietly that it doesn’t fully register until the box. I think giving her so little screen time and making this one scene count is the smartest choice in the picture.
Doe understands the argument better than either of them, and constructs an ending that resolves it by making Somerset right.
If your story contains a debate, give the position you disagree with your best writing. Somerset’s case for withdrawal is more articulate than Mills’s case for engagement, and the film is stronger for it, because a reader who can dismiss one side isn’t watching an argument, they’re watching a lecture. The scene that makes an ending work is usually the one where nothing happens.
The method is to write your strongest prose for the position you disagree with. Somerset’s case for withdrawal is more articulate than Mills’s case for engagement, and the film is stronger for it, because a reader who can dismiss one side is watching a lecture instead of an argument.
The related move is to build the scene that makes your ending work well before the ending. The dinner at the Mills apartment is warm, funny and irrelevant to the plot, and without it the final ten minutes are merely unpleasant. Find the scene where your characters simply like each other, and put it in.
The Ending That Was Nearly Cut
New Line wanted it changed. There were versions discussed in which the box contains a dog, in which Mills doesn’t fire, in which Somerset intervenes successfully.
Fincher and Pitt refused, and Pitt’s contract reportedly made the script’s ending a condition of his participation. The version that exists is the one where Doe’s plan completes exactly as designed and the detective becomes the seventh sin. It works because Doe isn’t caught in a struggle. He surrenders, walks into the police station with his hands up and blood on his sleeves, and directs the final act from the back seat of a car. He was never going to be stopped, because being stopped was never a risk he’d taken.
It also never shows you the contents of the box. That restraint makes the scene unbearable. You learn what’s in it from a man’s face and another man’s voice. I think it’s the best ending a studio thriller was ever allowed to keep, and the idea of New Line putting a dog in that box makes me want to throw something. The scene is almost unbearable to sit through, and I want exactly that from it.
Somerset’s closing narration is the only comfort on offer and it’s thin on purpose. Hemingway wrote that the world is a fine place and worth fighting for, and Somerset says he agrees with the second part. That’s the correct ending for the argument the film has been having. He doesn’t agree that it’s fine. He’s going to stay anyway, and staying is the most the film will grant.
The strongest version of a bleak ending is one the antagonist chose and the protagonist completes voluntarily.
Nobody is overpowered here. Doe doesn’t win by being stronger or luckier; he wins by understanding Mills better than Mills understands himself, and he builds every element of the trap out of the protagonist’s own character. Construct your antagonist’s victory from your hero’s temperament and it’ll read as inevitable instead of cruel.
The mechanism to steal is that your antagonist’s plan should be constructed from your protagonist’s temperament. Doe doesn’t exploit anger in the abstract. He exploits Mills himself. Mills requested this city, believes the work matters, and has a wife he’s told nobody about.
Test it by asking whether the trap would work on somebody else. If it would, your antagonist is merely competent and the ending reads as bad luck. If it only works on this person, for reasons the reader watched you establish, the ending reads as inevitable, and inevitability makes a bleak conclusion feel earned.
What Fincher Learned From Alien 3
This was his second feature. The first was Alien 3, which was taken away from him, recut, and which he’s publicly refused to acknowledge as his work. That experience is visible in every decision here. The insistence on the ending. The refusal to show the box. The complete control of the look. Fincher arrived at Se7en having already learned what happens when a director negotiates. It’s also why the film is so tightly built. There’s very little in it that a studio could remove without collapsing something else. That’s a structural defense as much as an artistic choice.
Against a thirty-three million budget, the film made over three hundred million dollars, established Fincher permanently, and made the ending non-negotiable in retrospect. He was thirty-three when he made it, and he built the reputation for control that now precedes him by having lost it once. I find that the most interesting fact about his whole career.
Should you watch Se7en?
Yes, and it’s less graphic than its reputation suggests.
You see aftermath and evidence instead of acts, and that’s more disturbing than any depiction would be. I give it a 9 because it has the nerve to follow its argument all the way out to that open country, and it would get the 10 if the second act didn’t lean so hard on coincidence. That’s a small crack in a film I admire enormously.
