The auteur theory says the director. Writer-driven critics say the screenplay. The structural reality is more complicated.
The most-contested question in serious film criticism — and one of the most-contested questions in film labour relations — is who actually makes a film. The auteur theory, developed in the 1950s by the French critics who would later become the French New Wave directors, argues that the director is the author. The screenwriter-centric tradition, particularly strong in working-screenwriter circles, argues that the screenplay is the foundational creative work and the director is, in effect, an interpreter.
Both positions can be defended. Both are wrong. The structural reality is more complicated than either side acknowledges.
The auteur theory was articulated in the French film magazine Cahiers du cinéma in the 1950s, primarily by François Truffaut, Jean-Luc Godard, Éric Rohmer, Jacques Rivette, and Claude Chabrol — all of whom would become significant directors of the French New Wave from 1959 onwards. The argument was that the director, even working inside the constraints of the studio system, was the author of a film's visual style, pacing, performance direction, and tonal register. The screenplay was the raw material; the director was the artist who transformed it.
The theory was developed partly to argue for the seriousness of Hollywood B-pictures and genre cinema. The Cahiers critics wanted to celebrate Howard Hawks, John Ford, Alfred Hitchcock, Anthony Mann, and other working studio directors whose individual stylistic signatures the theory could now identify. The intellectual achievement was to provide a vocabulary for taking these directors seriously as artists.
The theory crossed the Atlantic in Andrew Sarris's 1962 essay 'Notes on the Auteur Theory in 1962,' which translated and elaborated the French argument for American audiences. By the late 1960s and into the New Hollywood era, the auteur theory had become the dominant interpretive framework for serious film criticism.
The counter-argument — that the screenplay is the foundational creative work and the director is its interpreter — has been made consistently by working screenwriters across the auteur-theory era. William Goldman's Adventures in the Screen Trade (1983), the canonical screenwriter memoir of the period, makes the argument explicitly: 'Screenplays are structure. That's all they are.' The director's job, in Goldman's framing, is to render the structure the screenwriter has built.
The counter-argument has historical evidence on its side. Many of the films most-often discussed as 'director's films' were, in fact, primarily written by their screenwriters: Chinatown (1974) is Robert Towne's screenplay more than Roman Polanski's direction. Network (1976) is Paddy Chayefsky's screenplay; Sidney Lumet has said as much. The Apartment (1960) is the Wilder-Diamond screenplay. The screenwriters who provided the structural skeleton of these films were, in the working economy of film production, doing significant authorial work.
One way of resolving the question is to look at the directors who write their own films. The auteur-theory canon includes a significant number of directors who wrote — or co-wrote — most of their work. Quentin Tarantino has written almost every word of his filmography. Paul Thomas Anderson the same. The Coen brothers the same. Christopher Nolan co-writes with his brother Jonathan. Bong Joon-ho writes most of his films. Greta Gerwig co-writes with Noah Baumbach. Wes Anderson co-writes with various collaborators. Jane Campion writes her own films.
These directors are also, almost by definition, the directors whose work is most-consistently identified as having a recognisable authorial signature. The auteur theory's strongest cases are, in fact, also the cases where the director is also the screenwriter. The theory's weaker cases are the directors who have worked entirely from other people's scripts.
Most films are made by directors who did not write the screenplay but who, in production, made significant authorial decisions that shaped the final film. Stanley Kubrick adapted other people's novels but, in adaptation, made structural choices that fundamentally reshaped the source material (The Shining's deviations from King's novel; Barry Lyndon's tonal departures from Thackeray; Eyes Wide Shut's pacing of Schnitzler).
Sidney Lumet, who directed 12 Angry Men, Dog Day Afternoon, Network, and dozens of other films, almost never wrote his own screenplays. But his recurring concerns (New York, institutional failure, the moral weight of professionals doing their jobs) constitute an authorial signature across forty-four films. The signature is the director's; the words are the writers'.
Some directors are openly collaborative writers in practice. Robert Altman's films were famously written more on-set than on-page. Mike Leigh develops his films through extended improvisation rather than scripted dialogue. Steven Spielberg has worked with multiple screenwriters across his career but the films are, recognisably, Spielberg films.
The practical answer is that authorship in cinema is collaborative in a way that authorship in most other artistic disciplines is not. A novel has one author. A poem has one author. A film has, at minimum: the screenwriter, the director, the cinematographer, the production designer, the editor, the composer, the lead actors, and the producer who shaped the project's framing. Each of these contributes authorial decisions. The 'author' of a film, properly, is the team.
The auteur theory's value is in identifying which member of the team most-shaped any given film. For films where the director did write the screenplay and supervise the production closely, the director is the appropriate author. For films where a strong screenplay was largely interpreted faithfully by a less-distinctive director, the screenwriter is the appropriate author. For films where a strong producer drove the project (David O. Selznick on Gone with the Wind, Bob Iger on the contemporary Disney slate), the producer can be the appropriate author.
The reductive version of the auteur theory — that the director is always the author — is a misuse of the theory's original insight. The reductive version of the screenwriter-centric counter — that the screenplay is always the work — is the same kind of mistake from the opposite direction. The honest answer is that authorship in cinema requires asking, for each film, whose decisions actually shaped it. The answer changes from film to film.