A Hollywood film narrative must have stars with whom we can identify. These stars must survive at least until the concluding scenes: kill off Janet Leigh prematurely, as Hitchcock so famously did in Psycho, and you have disoriented your audience. There are other, related rules. Some people in film stories are more important than others. From the stars on down to the extras there is a delicate set of gradations controlling how much time and what sort of attention each actor/character gets. Disrupt these gradations, as Robert Altman did throughout the 1970s or as Jean Renoir did in his best American film, The Southerner, and again the audience is confused. Why should it not be? Character, theme, and pacing are controlled by the conventions of Hollywood so consistently that the control, the conventions, become invisible. Like children at bedtime, we insist that the story be told just so.

This conservatism is in many ways an advantage. If there is no rule about stars being principal characters throughout, then Janet Leigh's death in the shower does not shock us as it is supposed to. Hitchcock's perverse combination of avant-garde experimentation and commercialized suspense makes his work memorable. Nonetheless, the very power of the Hollywood conventions has often worked to our disadvantage. To take a key example, the idea of non-narrative cinema goes right past people. A film has to tell a story; if it doesn't, it is either filler or a put-on. Experimental film in America remains at the periphery as no other modern art has—not even modern music. People know who Arnold Schoenberg is, even if they don't want to listen to him. Ever heard of Stan Brakhage? Maya Deren? Joseph Cornell?

Cornell is the most famous of these three. In recent years many people have admired his boxes, enigmatic shadow-theater constructions like nothing else (despite their affinities with surrealism). It is fairly easy to accept a Cornell box, no matter how extraordinary, because there is nothing to measure it against. His films are another matter. Their distortion of narrative is not only more radical than Hitchcock's (say), it seems to be happening in a different cognitive universe. Cornell does not so much manipulate Hollywood conventions as dissolve them.¹

He loved old movies: what he did to one of them is worth telling. In the early Thirties Universal Pictures had released a mediocre jungle adventure, East of Borneo. Cornell acquired a print and cut it up into tiny little pieces. Some of them he put back together. The resulting work is called Rose Hobart, after East of Borneo's leading lady.

Rose Hobart features brief fragments of a story, of what were once causally connected events. We witness a torchlight assembly of natives, a fight between explorers and alligators, an erupting volcano, a pistol shot. These bits and pieces drift through the film like memories out of place. Released from narrative, they nonetheless generate suspense—only this suspense can have no resolution, cannot even move in the direction of resolution. The result is an atmosphere both comic and sinister. It is as though we were at a dinner party where the rules of etiquette were suspended and others, elaborate but unspecified and unspecifiable, substituted. How does one behave at a Cornell film?

One looks —and in this case laughs. The laughter is elicited by a carefully-planned incoherence. It is elicited above all by the one Hollywood convention which Cornell has lovingly, obsessively retained. Rose Hobart has no plot. It does have a star. However attractive, Rose is an incompetent actress. Cornell has built his film around her semi-somnabulistic gestures. We see her in umpteen reaction shots: tittering, looking scared, gazing down in brow-furrowed uncertainty. We seldom see just what she is reacting to. She comes, for example, through six or seven doorways, one right after another. Amidst fragments of melodrama—grand entrances, apparent assaults on her virtue, tense conversations—Rose is our one steady point of reference. Precisely because her acting range is so small, her image (especially the image of her beautiful, inflexible face) dominates the film.

Cornell has made a discovery. The removal of plot allows him and his audience to take star-worship to a point beyond narrative. The star's stardom had seemed to be implicated with story. It was either a way of carrying the tale forward or could only be expressed through it. Rose Hobart creates a moment when this wonderfully potent system self-destructs. Cornell fractures Hollywood to glorify it—or is it the other way around? In love with Rose he abducts her from east of Borneo. One can hardly imagine him going any further in this direction. Perhaps he couldn't either. He never again made a film like Rose Hobart, a point to which I shall return.

To those who love Rose Hobart, Cornell's later films often seem a bit tame. I do not think that they are. Rose is a heroine who becomes an image. Once this transformation is accomplished, Cornell finds that he can reverse it. Especially in the lyric-elegiac shorts of the Fifties, the cinematic image becomes a protagonist. We can see what this transformation signifies by looking at two of the best among these works, Centuries of June (a collaboration with Stan Brakhage) and A Legend for Fountains (a collaboration with Rudy Burckhardt).

Centuries of June was named (ex post facto) after an Emily Dickinson poem: "There is a Zone whose even Years/No Solstice Interrupt— ." Cornell embodies his search for this timeless realm in shots of an old house about to be destroyed. We see workmen digging, children trooping past after school, trees and insects stirring. These multifarious activities revolve about the half-hidden center of the house, on which the film focuses. Cornell and Brakhage investigate the magnificent tower, the grand Victorian porches and windows, the sense of lonely but absolute identity amidst so much fleeting action. Centuries of June is elegiac in its consciousness of loss. All photographic images are to some extent traces of an irrecoverable past. For Cornell this is not just a technical condition of photography but a theme inextricably associated with it. The trace becomes a memorial, based on nothing more palpable nor less severe than the camera's automatic gaze. To view Centuries of June is to participate in a kind of resurrection, whereby the house lives again in the "even Years" of the mind.

A Legend for Fountains dwells on similar themes to a somewhat different end. Among the many versions of Legend the ones I know are Mulberry Street and What Mozart Saw on Mulberry Street (both titles referring to the street in New York City's Little Italy where the footage was shot).² Naturally enough, the subject of the film(s) is street life, as observed by the camera and as if observed by a bust of Mozart in a shop window. Like Centuries of June, this movie is an elegy, only here the scene celebrated is so evanescent as to disappear in a moment of enactment. Children stumble, crawl, and jump all over the sidewalk, getting in everybody's way. A cat prowls along the pavement. Mothers dash out of doorways. The impression—given more powerfully here in a few minutes than in many very elaborate cinematic narratives—is of community life surviving gracefully among the distractions of the metropolis. This should be a world of fragments (like the re-edited East of Borneo?). Instead, the bits and pieces of street life cohere.³

Perhaps   an   extended   technical analysis could show something of how Cornell and his collaborators transformed this mundane material: how a house, a bust, indeed an entire street came to star in a movie. It is not my ambition to provide such an analysis. What interests me most about Cornell's filmmaking is its ability to follow through, to develop from one point to another without losing force or conviction. Rose Hobart proposes the initial paradox of stardom without narrative—when stardom has been, so consistently, a narrative device. Centuries of June and A Legend for Fountains take this transformation further. These movies express an overwhelming desire to see. None of Cornell's movies go on for more than fifteen or twenty minutes. Each of them shows just how hard it is to keep on looking at something, especially at an image without overt narrative significance. To recur to the counter-example of Hitchcock one last time: movies like Notorious or North by Northwest trick us into looking at places, people, and objects covertly, almost paranoiacally. That crop-duster is dusting where there aren't any crops: I wonder why? Rose Hobart is a mechanism for dispelling the fascinations of plot. This victory accomplished, Cornell can feel his way back into everyday life, now regarded as a mystery in its own right. As against the grand novelistic narratives of traditional Hollywood cinema, he recreates the film as lyric poem.⁴

Notes:

¹ By the inevitable corollary it is not—and never will be—for his films that Cornell is famous. See, however, an excellent article by P. Adams Sitney. "The Cinematic Gaze of Joseph Cornell," in the 1980 Museum of Modern Art catalogue titled Joseph Cornell. Sitney shows eloquently the grounds on which Cornell's films can be appreciated.

² Sitney. p. 88. sets out this very complicated tale. The footage from Little Italy, it turns out, has been edited at least five different ways—by Cornell, Larry Jordan, and Rudy Burckhardt.

³ This coherence is partly a matter of sound­track. Like the other two Cornell films discussed here, A Legend for Fountains is silent, in the traditional sense that the images are to be accompanied by music (here Satie). If A Legend for Fountains is a documentary, it is a documentary of a very peculiar kind.

⁴ In fall of 1981, with the help of Joseph Cornell's sister Betty Benton, Valparaiso University sponsored a showing of five Cor­nell films. This showing was attended by about one hundred people, few of whom anticipated what they were to see and almost all of whom stayed to the end. My thanks both to Betty Benton and to that inquisitive audience.