Harry Cohn ran Columbia Pictures like Mussolini — he kept a picture of Il Duce in his office — and in early 1937 he had a crisis. Frank Capra, his most successful director, had just left. Profits were down.1
He thought he had the answer: a comedy called The Awful Truth, and a thirty-nine-year-old director named Leo McCarey, who had made Duck Soup with the Marx Brothers and Ruggles of Red Gap with Charles Laughton.1
McCarey said he did not like the script, but would do the picture anyway for a hundred thousand dollars — a huge sum in 1937.
Cohn exploded. McCarey got up to leave.
And on the way out he noticed the office piano.
McCarey was a frustrated songwriter. He sat down and began to play a show tune. Cohn had a weakness for such music, and he was entranced: "Anybody who likes music like that has got to be a talented man," he said. "I'll pay that exorbitant fee. Report to work tomorrow."1
Nothing about the deal changed. The number was the same. What changed was that Cohn had decided McCarey was talented, on evidence entirely unrelated to filmmaking.
Three stars were cast — Cary Grant, Irene Dunne and Ralph Bellamy — and all had problems with their roles as written, none of them wanted to do the picture, and as time went by their unhappiness only grew.1
Then the script revisions started arriving, and it became clear McCarey had junked the original and was starting over.
His process was peculiar:
he would sit in a parked car on Hollywood Boulevard with the screenwriter Viña Delmar and verbally improvise scenes with her. Later, when shooting began, he would walk on the beach and scribble the next day's setups on torn pieces of brown paper.1
His directing was equally upsetting. One day he asked Dunne whether she played the piano and Bellamy whether he could sing. Both answered, "Not very well."
So McCarey had Dunne play "Home on the Range" as best she could while Bellamy sang off key.1
The actors did not enjoy this rather humiliating exercise. McCarey was delighted and filmed the entire song.
None of this was in the script, but all of it ended up in the film.1
Sometimes the actors waited on set while McCarey messed around on the piano, then suddenly came up with an idea for what to shoot that day.
Cohn visited and witnessed the process.
"I hired you to make a great comedy so I could show up Frank Capra. The only one who's going to laugh at this picture is Capra!"1
He was disgusted and wrote the whole thing off. His irritation grew daily. And then the part that matters strategically:
he was contractually bound to pay Dunne forty thousand dollars for the film, whether it was shot or not. He could not fire McCarey at this point without creating greater problems, nor could he have him go back to the original script, since McCarey had already begun shooting and only he seemed to know where the film was going.1
Read that as a position rather than an accident. Cohn had three ways to intervene and all of them were closed: firing was expensive, reverting was impossible, and the only map of the film was in the director's head.
As the days went on, the actors began to see it.
He would shoot them in long takes in which much of their work was only loosely guided; the scenes had spontaneity and liveliness.1
And the crucial qualifier:
Casual as he seemed, he knew what he wanted and would reshoot the simplest shot if the look on the actors' faces was not loving enough. His shoot days were short and to the point.1
That is the difference between looseness and laziness, and it is the whole page. A man who will reshoot a simple shot over an insufficiently loving expression is not improvising because he doesn't know what he wants.
Then, after many days' absence, Cohn showed up to find McCarey serving drinks to the cast. He was about to explode when the director told him they were celebrating.
They had just finished shooting — ahead of schedule and two hundred thousand dollars under budget.1
Then, to his surprise as well, the picture came together in the editing room like a strange puzzle. It was good, very good. Test audiences roared with laughter.1
The Awful Truth premiered in 1937, was a complete success, and won McCarey the Oscar for Best Director. Cohn had found his new Frank Capra.
Unfortunately, McCarey had seen his boss's dictatorial tendencies all too clearly, and though Cohn made lucrative offers, McCarey never worked for Columbia again.1
McCarey was essentially a frustrated composer and songwriter. He had gone into directing slapstick — he was the man who paired Laurel with Hardy — only because he could not make a living in music.1
The Awful Truth is considered one of the greatest screwball comedies ever made, and both its style and the way McCarey worked on it stemmed from his musical instincts: he composed the film in his head in just the same loose yet logical way that he would tinker with a tune on the piano.1
Greene identifies two requirements for working that way:
To create a film this way required two things: room to maneuver and the ability to channel chaos and confusion into the creative process.1
The room was taken, deliberately:
McCarey kept his distance from Cohn, the actors, the screenwriters—in fact, everyone—as best he could. He would not let himself be boxed in by anyone's idea of how to shoot a film.1
Given room to maneuver, he could improvise, experiment, move fluidly in different directions in any scene, yet keep everything perfectly controlled—he always seemed to know what he wanted and what worked.1
And the chaos was recruited rather than tolerated. Because filmmaking this way made every day a fresh challenge, the actors had to respond with their own energy rather than regurgitating a script.
The scene he was inspired to create when he learned of Dunne's and Bellamy's lack of musical skill, for example, seems unrehearsed and lifelike because it really was. Had it been scripted, it would have been far less funny.1
That is the sharpest formulation of the principle available: the quality could not have been written. It had to be caught, which means the process had to be arranged so that catchable things happened.
Greene's generalisation is unusually balanced for this book, because he names both ways of getting it wrong.
Directing a film—or any project, artistic or professional or scientific—is like fighting a war. There is a certain strategic logic to the way you attack a problem, shape your work, deal with friction and the discrepancy between what you want and what you get.1
Over-planning:
Directors or artists often start out with great ideas but in the planning create such a straitjacket for themselves, such a rigid script to follow and form to fit in, that the process loses all joy; there's nothing left to explore in the creation itself, and the end result seems lifeless and disappointing.1
Under-planning:
On the other side, artists may start with a loose idea that seems promising, but they are too lazy or undisciplined to give it shape and form. They create so much space and confusion that in the end nothing coheres.1
The resolution:
The solution is to plan, to have a clear idea what you want, then put yourself in open space and give yourself options to work with. You direct the situation but leave room for unexpected opportunities and random events.1
Both generals and artists can be judged by the way they handle chaos and confusion, embracing it yet guiding it for their own purposes.1
You are running a project where the plan is fighting the work — either the specification has become a cage, or there is no specification and nothing is converging.
The first move is to take room, and to notice that room is taken rather than granted. McCarey kept his distance from the studio head, the actors and the writers, all of whom had legitimate claims on the process. Nobody offers autonomy; you occupy it and then have to survive looking irresponsible for a while.
The way you survive that period is worth being precise about, because it is not confidence. Cohn could not intervene for three specific structural reasons: firing was expensive, reverting was impossible once shooting started, and the only complete picture of the work existed in McCarey's head. Those are positions, not attitudes. If you want the freedom, arrange the equivalents before you need them — commitments that make reversal costly, work that has started, and a whole that nobody else can assemble.
Then hold the two disciplines simultaneously, because either alone fails. Know exactly what you want — McCarey would reshoot a simple shot over a facial expression — and decide how to get it as late as possible. Most people can do one. The plan is about the destination and the looseness is about the route.
Then treat the constraints that arrive as material rather than as problems. Dunne couldn't play and Bellamy couldn't sing, which is a scheduling obstacle in a scripted production and became one of the film's best scenes. The question when something goes wrong is not how to work around it but what this makes possible that was not available before.
And know what it costs. McCarey delivered an Oscar-winning picture early and under budget, and never worked for that studio again — because operating this way required seeing exactly how his employer behaved under pressure, and what he saw was disqualifying. The autonomy is purchased with the relationship, and that bill arrives after the work is finished and praised.
The Awful Truth, McCarey's Best Director Oscar, the improvisational method, the parked-car script sessions and the film's status among screwball comedies are all documented Hollywood history. [POPULAR SOURCE]
The account is told entirely from the director's side and the costs land elsewhere. Three major stars were unhappy for the duration, were directed into a scene explicitly described as humiliating, and had no idea where the film was going. Cary Grant reportedly tried to buy his way out of the picture. Greene reads their discomfort as the raw energy the method required — which is one description of it, and the other is that a group of professionals were made anxious for months so that one person could keep his options open. Both are true, and the chapter records only the first.
And the survivorship problem is severe here. Improvisational directing with no locked script and daily setups scribbled on paper is also an exact description of productions that collapse — the method is indistinguishable, from outside and during, from a director who has lost control. McCarey came in under budget and won an Oscar, so his process is a strategy. The identical process by someone with less craft is a disaster, and the chapter offers no way to tell which you are looking at until the edit.
Greene half-acknowledges this with the "too lazy or undisciplined" failure mode, but treats the distinction as one of character. It is really one of capability — McCarey could compose a film in his head because he had spent twenty years learning to, and the room to maneuver was only valuable because of what he could do with it.
Against Hitchcock and the Detached Buddha, this is the exact opposite method producing the same result, and the pairing is the most useful thing in either page.
Hitchcock locked everything before shooting — the script complete, the shots designed, the film effectively made in pre-production — which produced total control and famous on-set calm. McCarey locked nothing and decided daily. Both achieved complete authorial control of the finished picture. Both kept the studio out.
What they share is not a method but a position: in each case the whole of the film existed in one head and nowhere else, so nobody else could assemble or alter it. Hitchcock got there by finishing early; McCarey by never finishing until the edit. The control came from the exclusivity, not from the timing — which suggests the real variable in creative autonomy is not planning versus improvisation at all.
Against Mae West and the Shifted Battlefield, another Depression-era studio figure who won by making her contribution structurally impossible to remove. West's changes lived in the footage; McCarey's lived in his head. Same principle, different hiding place.
Against Strategy 6: The Controlled-Chaos Strategy the title is nearly identical and the content complements it — that chapter is about an organisation designed to produce controlled disorder in an enemy; this is one person doing it to their own process on purpose.
Creative Practice → Planning the Destination, Not the Route. The chapter's resolution — know exactly what you want, decide how to get it as late as possible — describes how a great deal of good creative work actually gets made, and it explains why both common pieces of advice ("plan it properly" and "just start and see") are half-right and jointly useless.
The insight the pairing produces: the two failure modes are not opposite errors on one axis, they're failures of two different things that people mistake for one. Over-planning is usually not excessive rigour about the goal but premature specification of the method, and under-planning is usually not freedom of method but absence of a goal to converge on. Which means the diagnostic question for a project in trouble is not "did we plan too much or too little" but which of the two we specified. A rigid script with an unclear purpose and a vague process with a precise aim look similar from outside and require opposite corrections.
Business → Autonomy Is a Structural Position, Not a Permission. Cohn could not intervene, and the reasons were three specific irreversibilities rather than any respect for the director's process.
The insight neither domain gives alone: creative autonomy inside an organisation is almost never granted and almost always constructed out of sunk commitments, started work, and information asymmetry — which is why it correlates so poorly with seniority and so well with a certain kind of operator. It also explains the reliable pattern that follows: people who take autonomy this way tend not to stay. The same structural moves that protect the work from interference are legible to the organisation as evasion, and the relationship rarely survives the success. McCarey delivered early, under budget, with an Oscar, and never came back — which is the normal ending rather than an unfortunate one.
Cohn couldn't stop the film because firing was expensive, reverting was impossible, and the only complete version of it existed in one man's head. That's a position, not a personality — which means creative freedom inside any organisation is something you build out of irreversibilities before you need it, and the reason so few people have it is that it looks like insubordination right up until the picture is good.