Sightlines · a reading
A Deleuzian reading of Indiana Jones and the Temple of Doom
1984 · Steven Spielberg · Indiana Jones and the Temple of Doom on Sightlines
A face lit from below. The only lamp in the shot is a brazier on the floor, so the light climbs the jaw, floods the nostrils, and leaves the eye sockets as two pits. Mola Ram has that face when he works the crowd. Later in the film, so does Indiana Jones, after the cult makes him drink.
Douglas Slocombe didn't invent the setup. He is reproducing, almost exactly, Karl Freund's under-lit close-up of the hypnotist in The Mummy (1932), where a brazier below the actor is the frame's only key. Fifty-two years, one optical figure, carried over intact. Slocombe had been lighting faces by single flame since Ealing; Dead of Night (1945) taught him that a face modeled by flicker instead of fill goes strange, and the Pankot corridors get that lesson from the same operator's eye, thirty-nine years on.
Keep the underlight in mind. It falls on the villain and on the hero, and it does the same thing to both. That is the most interesting fact about this film.
Deleuze split cinema in two. In the movement-image, somebody perceives a situation and acts to change it, and the world answers back; the circuit from eye to hand to world holds, and cutting exists to keep that circuit legible. In the time-image, the circuit breaks. Action stops working, people watch and endure, and you feel time itself rather than movement. Deleuze thought the second regime arrived because the first one broke, historically, after the war. What makes Temple of Doom worth this lens is that it is the movement-image deliberately rebuilt in 1984 by two men who went back to the archive for the parts. The dossier names them. Gunga Din (1939) supplies the subterranean temple and the raised gallery above a prostrate congregation. The Republic serial Spy Smasher (1942) supplies the method: William Witney rehearsed a fight to a numbered beat sheet, then shot each blow as its own setup with screen direction locked. Spielberg storyboards that procedure straight into the rope bridge and the conveyor belt, along with the chapter-play's habit of ending a unit on an unresolved fall.
The lens earns itself because the standard complaint about this film, that it's all set pieces and thin between them, turns into a description of the engineering. Deleuze's large form runs situation → action → transformed situation. A village stripped of its stone and its children is a situation that presses on whoever walks into it. Deleuze calls a milieu that presses like that the encompasser, and Elstree built the largest one anybody could afford: a temple lit almost entirely by apparent firelight, crowds held in silhouette, orange pools picking figures out of black. Then force meets force. Indy against Mola Ram, finishing on a severed span over a drop, which is the binomial reduced to its purest architecture. Two men, one line, nothing else in frame.
Deleuze has a finer term for what passes between a hero and that kind of world: the imprint, the bond by which situation and character stamp each other. Usually it's a metaphor. This film makes it literal. Indy is made to drink the Black Sleep, the milieu takes him over, and his face drops into the same brazier-light as the celebrant's. That face is what Deleuze calls a potisign, an intensive face: not a settled expression but micro-movements crossing a threshold, one quality turning into its opposite while you watch. And the film cannot finish until it gives him back to himself, because everything after depends on a working sensory-motor circuit and for a while he doesn't have one.
Between the big units the film breathes. Matte-painted Pankot and the mountain vistas open the space out; the mine cart pulls it back into a single taut line. Deleuze calls that alternation respiration-space, and Michael Kahn is why it functions: however fast the cutting goes, you always know where everyone is and what they want.
Underneath all of it the film is running a second kind of image entirely, which is why the MPAA had to build a new box. Deleuze places the impulse-image between affection and action. Its world is made of drives instead of situations: a social surface, the derived milieu, turns out to be the crust over an originary world of pure appetite, and the film's business is showing the crust crack. Stroheim, Buñuel, Losey. Temple of Doom builds this vertically, as set design, which no naturalist film had quite done. The Pankot banquet is the derived milieu: a palace, place settings, manners. On the plates, eyeball soup and chilled monkey brains. Food as fetish, in Deleuze's strict sense of a part-object charged with compulsive force. Once you notice the film's appetite for part-objects it doesn't stop. A tunnel packed with live insects, real ones, in enormous quantities, on a stage in Hertfordshire. Then a hand going into a chest and coming out with a heart still beating in it.
The naturalist world has a law: it runs downhill toward exhaustion and death. Deleuze calls it the line of slope, and it's why films built this way normally end in ruin. Temple of Doom goes the whole way down and then declines the ending. Indy climbs back out. That collision is what the film actually is, an impulse-image world loaded into a movement-image machine, with the machine still running when the lights come up. PG-13 is what it looks like when an industry notices a form has outgrown its container. Very few films visibly restructure a classification system. This one and Gremlins did it in the same year.
Both of the film's great pieces of staging are performances with an audience inside the frame. The Shanghai prologue is Berkeley pastiche down to the overhead crane resolving the chorines into a rotating kaleidoscope, and a stage that impossibly deepens past its own proscenium into pure soundstage. That's Deleuze's attraction-image: action transposed into spectacle, staging that figures rather than acts. The Kali ritual is assembled the same way. Raised gallery, uplit celebrant working a congregation, a mechanized altar descending, all of it inherited from Cedric Gibbons's sacrificial chamber in The Mask of Fu Manchu (1932). The yellow-peril framing of Asian ritual as a horror attraction arrives in the same crate. Which brings the film's cost into view in its own formal terms. Deleuze's name for where a film locates a person is the centre of indetermination, a living centre that hesitates between perceiving and acting. Temple of Doom has Indy. Short Round and Willie get fractions of one. Everyone else is a mass. The chanting worshippers in silhouette and the starving villagers are a dividual, a crowd rendered as a single affective unit, and the film needs them that way, because a cinema tuned for maximum sensory-motor power has to distribute the capacity to act very unevenly. Somebody has to be the milieu. India declined to host the production. The film has been a postcolonial case study ever since. Neither fact is separable from the grammar.
One more thing, and it's the technical jewel. Deleuze opens Cinema 1 by arguing that cinema broke with the older arts by delivering movement as any-instant-whatever, ordinary equidistant slices, rather than the privileged instant, the posed climactic form of sculpture and history painting. Stop-motion is a regression to the posed instant: every frame sharp, held, decisive, which is exactly why old stop-motion strobes and never feels alive. ILM's go-motion moves the miniature during the exposure so each frame carries blur. It is the any-instant-whatever manufactured by hand. Dennis Muren's mine-cart chase is the payoff, cut from miniature tunnels, motion-control rigs and actors on gimballed carts, assembled while Harrison Ford was off the film with a back injury, and you cannot find the seam. The most physically convincing movement in a famously physical film is movement nobody ever photographed at speed. Less than a decade later, on Jurassic Park, Spielberg would break the ceiling this film establishes.
So watch the key light. When it comes from above or the side, Indiana Jones can act, and the oldest engine in the medium runs hotter here than it ever ran for Stevens in 1939 or Witney in 1942. When the key drops to the floor, something older than plot has the room, and the whip is no help. Temple of Doom is the movement-image at full power discovering there's a basement underneath it. The question it leaves open is whether an adventure film can go down there without turning the people who live there into furniture. This one didn't manage it. That's part of why the descent is still worth taking, and worth arguing with all the way down.
Concepts in play
- the movement-image — The whole film is a deliberate reconstruction of the sensory-motor circuit from archival parts: Gunga Din's temple architecture and Witney's numbered-beat-sheet fight grammar from Spy Smasher, storyboarded into the rope bridge and conveyor belt.
- action-image, large form — A village stripped of its stone and its children presents a global situation; Indy descends, acts, and restores it — situation → action → transformed situation, across the palace, temple, mine and bridge.
- the encompasser — The vast Elstree temple set lit almost entirely by apparent firelight, with shafts and pools of orange picking figures out of blackness and silhouetted crowds — a milieu built to call forth action.
- binomial (force against force) — The rope-bridge standoff: Indy against Mola Ram on a severed span over a drop, the duel reduced to two men and one line.
- empreinte / the imprint — The Black Sleep of Kali Ma literalizes the bond between milieu and hero: Ford is lit by the same brazier-from-below figure Slocombe reproduces on Mola Ram, and the film cannot resume its action until he is given back to himself.
- icon of power (intensive face) — Freund's under-lit close-up from The Mummy (1932) reproduced on Mola Ram and on Ford under the Black Sleep — a face crossing a threshold rather than holding a settled expression.
- respiration-space — Matte-painted Pankot Palace and the mountain vistas open the space out; the mine-cart chase contracts it to a single taut line, with Kahn's cutting preserving orientation throughout.
- impulse-image (naturalism) — The subterranean Thuggee sequence: appetite, cruelty, child slavery and sacrifice erupting beneath the palace's social surface, staged as a world of drives rather than of situations.
- derived milieu — The Pankot banquet — palace, place settings, manners — serving eyeball soup and chilled monkey brains, the social surface already showing what it covers.
- originary world — The catacombs beneath the palace: the pit, lava-glow underlighting turning faces into masks, chanting congregation, the heart-extraction ritual — the palace revealed as crust over a savage world.
- fetish (part-object) — Food as compulsive part-object at the banquet (eyeball soup, chilled monkey brains), the tunnel of live insects shot with enormous quantities of real insects on stage, and the still-beating heart pulled from a chest.
- expressionism — Slocombe's firelight temple: hard sculpted key, large areas of frame left black, silhouetted crowds, lava-glow underlighting — expressionist lighting at blockbuster scale, doing most of the work of the horror.
- attraction-image — The 'Anything Goes' prologue as shot-for-shot Berkeley pastiche (overhead crane resolving chorines into a rotating kaleidoscope, a stage deepening past its own proscenium), and the Kali ritual staged as spectacle before a diegetic congregation from Gibbons's Fu Manchu chamber.
- dividual — The chanting worshippers rendered in silhouette as a single affective unit, and the village of starving supplicants presented as one collective body to be acted upon.
- the interval / centre of indetermination — Indy is the film's only full living centre hesitating between perceiving and acting; Short Round and Willie carry fractions, and the Indian crowd carries none — the formal shape underlying the film's standing postcolonial critique and India's refusal of permission to shoot.
- any-instant-whatever — ILM's go-motion, introducing blur during exposure so miniature movement reads as photographic, used across the mine-cart chase's miniature tunnels and motion-control rigs — the ordinary continuous instant manufactured by hand, seamless against live-action gimbal work shot while Ford was injured.