This article is part of a new regular feature at Film Comment in which university scholars share concise essays bringing our readers into the rich and varied conversations occurring in the fields of film and moving-image studies. The following abridged excerpt comes from Chapter 2 of the book The Dynamic Frame: Camera Movement in Classical Hollywood by Patrick Keating (published by Columbia University Press). In the book, Keating charts a fascinating history of film history through camera movement filled with rich readings of both shot sequences and filmmakers’ ideas, featuring Murnau, Mamoulian, Hawks, and many more. The below text is abridged from a passage on the point-of-view shot and what it means.

Gary Cooper and Helen Hayes in A Farewell To Arms (Frank Borzage, 1932)

Click on select titles to view the clips from the films under discussion.

As in the silent period, in later periods the idea that the camera might move like a person took different forms in the 1930s. Most literally, the camera’s movements might represent a character’s movements, as in a point-of-view shot. More loosely, the camera might mimic the movements of an observer—a hypothetical onlooker who watches the story unfold.

Let us start with the most literal form of the camera–person analogy: the point-of-view shot. Several films of the early 1930s experimented with the technique: Dr. Jekyll, of course, but also Inspiration (1931), A Farewell to Arms (1932), Guilty as Hell (1932), The Sport Parade (1932), The Story of Temple Drake (1933), and The Kennel Murder Case (1933). At first, we might suppose that the point-of-view shot serves an obvious function, encouraging identification by asking the spectator to see the world through a character’s eyes. But the process of identification was never so simple. Some point-of-view shots encourage identification; some do not; and some ask us to confront the logic of identification itself. In Michael Curtiz’s film The Kennel Murder Case, detective Philo Vance (William Powell) explains his theory of the murder, even though he has not yet identified the specific suspect. While Vance’s explanation plays on the soundtrack, the camera represents the murderer’s point of view, stalking through the house and looking for the victim. Cinematographer William Rees sells the trick by casting the man’s shadow on a nearby wall. Notice that the primary purpose of the shot is to block identification. We do not know whose point of view is being represented, and the shot serves to conceal the murderer’s identity.

Compare this example with the extended subjective sequence in Frank Borzage’s film A Farewell to Arms, where the technique produces a rich feeling of empathy for the protagonist. Recently wounded, Frederic (Gary Cooper) is wheeled into a makeshift hospital in an Italian church. The camera adopts his point of view as he helplessly stares up at the ceiling. Heads pop into the frame occasionally to exchange dialogue with Frederic, but for several seconds we simply watch the vaulted ceiling pass by, giving way to the inside of a dome. Notice how the lack of information encourages identification with Frederic. He is unfamiliar with this new location, and so are we. He is desperate to know if he may meet his beloved Catherine (Helen Hayes) at this hospital, and so are we. Of course, we know that we are watching a film, so our identification can never be total. But the filmmakers produce an affinity between Frederic’s exceedingly limited state of knowledge and the spectator’s. Trapped inside Frederic’s consciousness, the camera offers no reassuring information. Instead, it stares into the circle of the dome, beautiful but empty.

Turning the argument around, a filmmaker might foster identification by refusing to use a point-of-view shot, conspicuously detaching camera from character. Another scene from A Farewell to Arms shows a pregnant Catherine writing a letter describing her new apartment in Switzerland to her faraway lover, Frederic. As Catherine reads her letter aloud, the soundtrack plays her voice, describing the apartment in the most luxurious terms she can think of; meanwhile, the camera wanders about the apartment, isolating details that belie her words. When Catherine says that the bed is “fit for an empress,” the camera shows a bed that is small and Spartan; when she praises the “maroon velvet carpet, ankle deep,” we see a rug that is wrinkled and tattered. As Catherine jokes about her dire circumstances, we are asked to share in her knowledge and emotions, without looking through her eyes.

Point-of-view plays a different role in the horror genre. Silent-era horror films such as The Cat and the Canary featured “ambiguous” point-of-view shots, playing on the camera’s machinelike qualities to produce a sense of the uncanny. This tradition continued into the 1930s, as in Roland West’s The Bat Whispers (1930). Based on a play West had adapted once before (with minimal camera movement), this early sound film appeared in two versions: 35-millimeter and the still-rare widescreen format 65-millimeter. The story concerns a masked supervillain who terrorizes a house full of oddballs until he is exposed in a last-minute twist. Though the Bat turns out to be an ordinary human, for much of the film it appears that this disguised rogue has the gift of omnipresence, suddenly appearing where we least expect him.

The camera evokes this quasi-superpower by executing several dazzling dolly and crane shots, zipping through space and entering rooms as if by magic. In one sequence, a shot of a round clock face dissolves to a graphically matched image of a round train wheel spinning on a track. Next, the camera dollies laterally past a cityscape, rendered in miniature, as if the train were dashing by. The film then dissolves to a shot representing the point of view of the train itself, speeding toward the sign of a sanitarium. So far the movements are rapid but not magical; they evoke the velocity of a rushing train. The next shot shifts the imagery toward a more supernatural register. As the camera careens toward a stately bank, the shadow of a bat envelops the facade, suggesting that the camera has been representing the moving point of view of the Bat himself all along. A final dissolve introduces an overhead view, looking over the Bat’s shoulder as he peers through the window into the bank itself. Astonishingly, the camera cranes down as if passing through the glass of the skylight, quoting the moment in The Last Laugh when the camera dollies forward and dissolves through glass. Throughout this sequence, the film develops an extended comparison between the camera and the Bat: both have the power to fly through space, to climb walls, to peer through windows, to see without being seen. But the analogy does not humanize the camera so much as it renders it uncanny: like the Bat, the camera is eerily human and nonhuman at the same time.

Here, the film takes advantage of an obvious but important fact: the camera is not a person. When the camera dollies, it moves with a straight-ahead smoothness that seems alien to human motion. Most of the time, this smoothness is not a problem. Nobody requires the camera to move like a person because the camera simply is not a part of the fictional world. But what if a filmmaker does want to locate the camera’s movements within the storyworld, as in a moving point-of-view shot? In The Kennel Murder Case, it could be argued that the camera does a poor job representing a character who is walking through a house; its movements are too strange, too inhuman. The Bat Whispers takes this strangeness and turns it into an asset, evoking the movement of some supernatural entity. Another early horror film, Chandu the Magician (1932), is even more explicit about the supernatural strangeness of the moving frame. In one scene, the protagonist (Edmund Lowe) and the Princess (Irene Ware) are looking into a crystal ball in order to spy on Roxor (Bela Lugosi), an evil genius who is building a death ray under the ruins of an old Egyptian temple. After a close-up of the crystal ball, a slow dissolve introduces a flurry of dynamic shots: toward the temple, through the doors, down a hallway, left, right, down another hallway, left again, down a third hallway, and down to reveal an overhead view of the maniac’s laboratory. The camera’s inhuman ability to fly through space evokes the crystal ball’s mystical powers.

Although the vogue for subjective shots soon subsided, the camera–person analogy remained persistent. Many Hollywood filmmakers insisted that the camera should always move the way a person moves—not just in point-of-view shots, but in every shot. According to this line of thinking, the camera represented the eye of an unseen witness to the storyworld’s action. But what sort of observer was this camera-eye to be? Like a criminal stalking through space? Like a powerless patient, wheeled around by others? Like a mysterious being endowed with quasi-superpowers? Or like an utterly ordinary person?

Excerpted from The Dynamic Frame by Patrick Keating. Copyright (c) 2019 Columbia University Press. Used by arrangement with the Publisher. All rights reserved.

Patrick Keating is associate professor of communication at Trinity University, where he teaches courses in film studies and video production. He is the author of Hollywood Lighting from the Silent Era to Film Noir (Columbia, 2010).