Robert Towne’s Chinatown Screenplay: How the Greatest Spec Script Got Made

The origins

Where the Chinatown screenplay started: four unrelated ideas

Towne didn’t set out to write a noir about the Los Angeles water wars. The screenplay assembled itself from four separate sources over several years, none of them obviously connected until Towne recognized they were all about the same thing.

The first was a conversation. Towne was buying a dog from a vice cop in Los Angeles, and the conversation turned to working Chinatown. The police were told to do as little as possible there — the dialects, the tongs, the internal power structures were impenetrable. You couldn’t tell if your interventions were helping or making things worse. That phrase — “as little as possible” — lodged in Towne’s mind as a governing metaphor. The futility of good intentions. He gave his detective a background in Chinatown, and the neighborhood stopped being a location and became a state of mind.

The second was a photograph. Around 1969 Towne saw a copy of Old West Magazine — a Los Angeles Times supplement — containing a photo essay called “Raymond Chandler’s L.A.” Photographs taken in 1969, meant to represent the city in the 1930s. One image stopped him: a Plymouth convertible under a streetlamp outside Bullock’s Wilshire. The present reconstructing the vanished past. He was reading Chandler at the time, studying the genre he intended to work in.

Water as the crime

The third was a history book. At the University of Oregon, where Nicholson was filming Drive, He Said, Towne borrowed a book from the library called Southern California Country: Island on the Land. A chapter titled “Water, Water, Water” laid out how Los Angeles acquired its water supply through systematic deception of the Owens Valley farmers — dumping their water, starving them out, buying their land at distressed prices. “Why not do a picture about a crime that’s right out in front of everybody,” Towne said later. “Instead of a jewel-encrusted falcon, make it something as prevalent as water faucets.”

The fourth was the vice cop’s phrase coming back to him, connected now to all the rest: corruption hidden in plain sight, good intentions that make things worse, a place you can’t understand until it’s too late.

What Towne did — and what makes his process worth studying — was recognize that these unrelated experiences were all about the same thing. Power taking what isn’t theirs. The rest of us barely noticing until it’s gone.

The refusal

Why the Chinatown screenplay almost didn’t happen

The script might never have been written if Towne had taken the easy money.

In 1971, Robert Evans — head of production at Paramount — offered Towne $175,000 to adapt The Great Gatsby. An enormous sum. Towne said no. His stated reason: he didn’t want to be the unknown Hollywood screenwriter who ruined a literary classic. The real reason was more strategic. He had spent over a decade doing uncredited work — polishing Bonnie and Clyde, fixing The Godfather, doing emergency script surgery across a dozen productions — and he had watched what happened when writers attached their names to high-profile adaptations that went wrong. He wrote the Chinatown screenplay instead, as a spec, for Jack Nicholson.

“I wrote the part for him, in his voice,” Towne said. “I could not have written that character without knowing Jack. We had been roommates, and we’d studied acting with Jeff Corey for years, so he was, in a very real sense, a collaborator.”

The drafting

Drafting the Chinatown screenplay: nine months of struggle

The writing nearly broke him.

Towne was going through domestic difficulties and had to leave his house. He took himself and his dog to Catalina Island and worked at The Isthmus for several months, then moved around the city — Curtis Hanson loaned him an apartment, he worked in rented rooms, kept moving. The circumstances weren’t the problem. The writing was.

“Writing scenario, after scenario, after scenario was just so complicated that after a certain point, I thought I’d never get through it,” he said. The structural challenge was genuine: a detective story where the detective never fully understands what he’s investigating, where the real crime is hidden inside a false one, where the femme fatale is innocent and the real villain is the charming grandfather in the fishing hat. The first draft ran approximately 178 pages. Evans called it brilliant but incomprehensible. Towne later admitted that if it had been shot as written, “it would have been a mess.”

The clean structure only emerged through extended revision — eventually involving Polanski himself.

The architecture

The first structural decision: which revelation comes first

The most important insight in Towne’s Chinatown screenplay was recognizing that the plot and the subplot were two expressions of the same crime, committed by the same man, at two different scales.

The water scandal was the plot — the civic crime, moving outward to encompass farmers, politicians, city hall, the entire infrastructure of Los Angeles. The incest was the subplot — the personal crime, hidden inside a family, protected by silence and shame. The two were, as Towne put it, “intimately connected, literally and metaphorically: raping the future and raping the land.”

The sequencing call

The structural question was sequencing. Which revelation comes first? “And of course, it was the water scandal,” Towne said. “The water scandal was the plot, essentially, and the subplot was the incest. That was the underbelly.” By revealing the civic crime first and the personal crime second, Towne ensured both landed with full force — the water conspiracy establishing a framework of evil that makes the incest revelation both shocking and, on reflection, inevitable.

This is not how dual-track mystery plots typically work. The conventional approach is to surface the deeper secret earlier, giving the detective something personal to push against. Towne inverted it. The deeper secret arrives later, and it’s worse than anything the audience had imagined.

The restriction

The POV lock: why the second draft failed

The single most consequential decision in the Chinatown screenplay is the POV lock — and the evidence for its importance is what happened when Towne abandoned it.

The first draft was almost entirely locked to Jake Gittes’s perspective. The audience knows only what Jake knows, sees only what Jake sees, understands the mystery at exactly the pace Jake does. This is what Towne identified as one of the very few detective films, including The Maltese Falcon, to commit to a truly singular POV.

The second draft opened it up

In the second draft, he tried opening the perspective — adding scenes from Evelyn’s or Cross’s point of view to clarify the plot. It made the screenplay weaker. Polanski agreed with the first draft. The opacity was the point. The confusion the audience felt was Jake’s confusion. Breaking the POV broke the identification, and without that identification, the ending is just sad instead of devastating.

The POV lock stayed. The audience connects the dots only when Jake does — which means they arrive at the full horror at exactly the same moment he does.

The collaboration

The six-week rewrite with Polanski

The draft Evans called brilliant but incomprehensible needed to become shootable. What followed was one of the most productive — and combative — collaborations in American cinema.

Polanski came to Los Angeles with reservations. He had left after Sharon Tate’s murder in 1969 and had no desire to return. But Evans was persuasive, the material was compelling, and Polanski had genuine investment in what the script could become. He gave Towne a book before they began: How to Write a Screenplay, signed “To my dear partner with hope, Roman.”

Strips of paper on the door

They worked for six weeks in Evans’s house during a California heat wave. Polanski came every day. They sat in a small room — Towne and Polanski on opposite sides of a desk, the dog on Towne’s feet, Towne’s pipe smoke in the air — and wrote one-line descriptions of each scene. Then they cut those into strips of paper and pasted them on the door. Every day they moved the strips around, resequencing scenes, collapsing redundant beats, identifying what the story needed versus what Towne had written as exploration.

The fights were about the ending. Towne wanted Evelyn to survive. Polanski wanted something more Greek. They couldn’t agree, and with production approaching, they started shooting without a finished ending. The ending was eventually written under duress — with Polanski’s vision prevailing over Towne’s and producing, despite everything, the only ending the film could have. That story has its own chapter. For the full account of what each version argues, see Chinatown ending explained.

The mechanism

What makes the Chinatown screenplay structurally unprecedented

Most detective screenplays are structured around information — the detective acquires it, the audience acquires it at the same rate, and the mystery resolves when everyone has enough. Chinatown is structured around the protagonist’s blindness. Jake doesn’t just fail to solve the mystery. His competence, his professional instincts, his drive to understand — these are what make him the instrument of the tragedy.

Every choice Jake makes is correct

Every choice Jake makes is professionally correct. He identifies the fake Evelyn Mulwray immediately. The watch goes under Mulwray’s tire exactly where it should. Land purchases get tracked. The bifocal clue resolves before Evelyn sees it. By the end, he has solved the case. And none of it is enough — not because Jake lacks perception, but because he can’t conceive of evil at Cross’s scale. His professional formation as a divorce detective has calibrated him for a certain magnitude of human corruption. Cross operates at a different order of magnitude entirely.

This is the structural innovation no other screenplay of the era matched. The protagonist is neither blind to the truth nor corrupt enough not to care. He is exactly competent enough and exactly limited enough to cause the specific outcome he was trying to prevent. That’s not a plot. That’s a mechanism. And Towne built it from nothing, in a borrowed apartment, in nine months, from four ideas that turned out to be the same story.

The standing

Behind the WGA’s top-three ranking

The WGA’s top-three ranking isn’t sentiment. It reflects what working screenwriters actually believe Towne’s Chinatown screenplay does that others don’t — and what every subsequent Chinatown screenplay analysis keeps returning to.

The exposition is the most efficient in American film — every piece of information serves multiple purposes simultaneously. The first scene establishes Jake’s character, his profession, his relationship to money, his ethics, and his relationship to his clients, all in under three pages, through action and dialogue that never announces what it’s doing.

What it changed for good

The protagonist design is the most precise in the genre — Jake’s flaw is inseparable from his competence in a way that makes both tragic rather than correctable.

The plot/subplot architecture — the civic crime and the personal crime as two expressions of the same predatory appetite — established a structural relationship that has been imitated ever since and matched rarely.

And it permanently changed what the femme fatale was allowed to be. After Chinatown, she didn’t have to be guilty. That inversion created a genre schism that filmmakers are still navigating fifty years later.

The myth and the man

What Towne said about it

Towne was never entirely comfortable with the mythology around the screenplay. He knew where the problems were. The first draft, by his own admission, was a mess. As for the ending — that had been forced on him. And the “perfect script” story, he understood, was being built on the finished product rather than the process that produced it.

What he said, consistently, was simpler: “We weren’t trying to write a screenplay that was perfectly-structured. We were just trying to make it make sense.”

The script that became the standard was not born from mastery. It was born from struggle — nine months of scenarios that wouldn’t resolve, a six-week argument in a California heat wave, and an ending dispute that was never really settled. The perfection is real. So is the chaos that produced it. It’s one chapter in the complete Chinatown analysis — the same craft runs through every part of the film.

The making, reassembled

The door: the whole struggle in six strips

In Evans’s study, Towne and Polanski wrote every scene as one line on a strip of paper, pasted the strips on a door, and moved them around until the story held. Here is the making of the script, told the same way.

Questions readers ask

The Chinatown screenplay: questions readers ask

The writer

The script