What would it actually mean to study narrative using the cinema rather than print fiction—especially the novel—as the paradigmatic model? The ancient precedence of written narratives in general, the pervasive influence of the novel in modern Western society, and the rich and supple categories and taxonomies that have been generated by the study of novels and stories may make it impossible to answer that question in any kind of "pure" way. However, if we accept that narrative and narration are objects of study that stand independent of any specific medium, then cinema frames certain issues that deserve further investigation in other media, including the novel itself. One of the most basic issues is the question of agency: Who or what has created and/or presented this particular narrative?

In print fiction, the answer to that question is commonly the author—the creator of the written text—and the narrator—the character or non-character "voice" that "tells" the story. Even when discussing print fiction, though, such answers are far too simple, leading to studies and debates about the status of "implied authors," different types of narrators, and so forth. In discussing narrative in general, narratologists still tend to base their discussion of such literary concepts, whether they accept them or reject them. David Bordwell, for example, rejects Seymour Chatman’s use of terms like "narrator" and "implied author" (or "implied filmmaker"), when discussing film narrative, as "anthropomorphic fictions" (62). For Bordwell, film narrative is not an act of "communication," but a process that allows a film viewer to construct a film’s prompted by cues in the film itself. In turn, Chatman objects that Bordwell ignores the obvious issue of agency—the question of "Who speaks?" Bordwell, according to Chatman, does not account for how those narrative cues came to be in the film in the first place: "The viewer certainly hasn’t put them there, so it seems a bit odd to talk about ‘narration’ as if she had" (Coming to Terms 127). Still, Bordwell does admit that "the narrational process may sometimes mimic the communication situation more or less fully" (62, Bordwell’s emphasis) and Chatman grants that the cinematic "narrator is not a human being . . . . agents need not be human" (CTT 134). Both critics then point to a concept of narrative in film that is not dependent on print fiction models. Both suggest that our perceptions of the creating and presenting agents—authors and narrators—are not so much autonomous entities as functions of the text itself, no matter what form of medium it takes.

Cinema foregrounds such issues because of its own unique properties. As a collaborative medium, film undermines literary notions of the solitary author. As a multi-track medium, relays information (that is, narrates) through all the technical means available to the medium, thus undermining literary notions of a single narrator of an entire text. Finally, as a technological medium, film is not simply diegetic ("telling," in Aristotle’s sense) as the novel usually is nor mimetic ("showing") as theater usually is, thus undermining traditional notions of narration as an act. The cinema was, rather, the first major narrative medium to "tell" by "showing" and to "show" as a form of "telling."

Shlomith Rimmon-Kenan has pointed out that notions of chronology and causality "are not ‘things in the world’ but concepts we impose on the world in order to apprehend it" (160-161). "Authors" and "narrators" are also concepts that we impose on narrative worlds in order to apprehend them, sometimes in response to specific textual cues and sometimes not. Thus, the presence or lack of a narrator is a relative, not an absolute, criterion. The lack of a narrator in most theater is simply a function of that medium, just as the presence of a narrator in most print fiction is of that medium. Cinema—which usually involves both "showing" and/as "telling"—may sometimes function as though it has one or more narrators but does not always need to do so. Films may also foreground elements of their own creation that imply the agency of a filmmaker, but they do not often do so. When they are evoked, such agents are actually examples of what Edward Branigan labels "convenient fictions" (76). Cues for these "convenient fictions" in film may be absent or few, but they may also be redundant or excessive, self-referential, contradictory, or oblique. The four American films, each from a different historical periods, discussed below offer examples of such cues.

49-17, directed by Ruth Ann Baldwin (1917), is a silent film that demonstrates some of the possibilities and problems of narration without sound. Set in 1917, the year of its making, the film introduces Judge Brand, a former gold miner come East who longs to return to the freedom of the Old West. He dispatches his secretary, Tom Robbins, to California where Tom hires the company of an "Old West" show to recreate the Judge’s old mining town. In the process, Tom falls in love with one of the performers, while another performer, Jim Raynor, turns out to be the man who had betrayed the judge’s former mining partner. After attempting robbery and kidnapping, Raynor dies accidentally while trying to escape.

This seemingly simple tale is filmed in relatively static shots but has a sophisticated narrative structure created through elements of classical continuity editing. The plot is advanced through a complex structure of flashbacks and mentally subjective shots, as when Judge Brand daydreams about the West or when Tom "sees" the judge decked out in miner’s clothes. A more intrusive form of overt narration, though, occurs through the use of intertitles that offer redundant or unnecessary information, along with illustrations that are also excessive to the needs of the story.

For instance, when the Judge explains his scheme to Tom, an intertitle reads, "It was a wild dream, but Tom understood—he had dreams himself sometimes. So two days later he left for the West." In the context of the narrative itself, the commentary serves as a distraction and does not actually advance the chain of events in a meaningful way. Even less relevant—and outright excessive—is the illustration, which shows Tom apparently "flying" West on the back of what looks like a large pterodactyl! Similar excessive uses of visual humor occur elsewhere. Tom’s search for a troupe to recreate the Judge’s gold rush town of Nugget Notch is summarized by the words, "Tom soon discovered that Western types were rare birds—and shy," with a picture of a man stalking a wild goose! When the Judge first arrives at his "gold rush" town, he joyfully tosses his top hat in the air and shoots holes in it. The intertitle shows a crowd running away and the comment, apparently reflecting the crowd’s thoughts, "This was too real!" However, the picture contradicts what we had just seen on the screen, since the judge had fired at his hat in the air.

These intertitles imply a need to reinforce and clarify visual images with verbal comments, implying a lack of trust in the image alone as "narrator." The sheer excess of these titles also forces an awareness of narrational agency unlike anything else in the film. Even the illustrations draw attention to the "telling" of the story as opposed to its "showing" through the diegetic scenes and flashbacks. The fact that this narration is often redundant or unnecessary (although charming in its way) further draws attention to its overdetermination. These narratorial intrusions, however, grow fewer by the later part of the film and cease entirely by the final scenes, suggesting a narrative strategy that builds on the audience’s growing ability to understand action and motivation in the final scenes without further reinforcement.

Despite its formulaic pulp Western story, 49-17 is interesting as a film by a woman director who fell into undeserved obscurity, as a display of sophisticated structures of time and subjectivity in film, and as a mark of a style of film narration that foregrounds a narrator’s "voice" to the point of redundancy. That such redundancy was felt to be necessary at all is a mark of a type of filmic narration still dependent on print models.

The original 1937 version of A Star Is Born also displays contradictory cues of agency, this time in terms of the film’s own creation. In most respects, this familiar story about Esther Blodgett (Janet Gaynor), a young woman who aspires to stardom, achieves fame and success as movie star Vicki Lester, but pays for that success with her broken heart, is a fine example of classical Hollywood narrative and style, demonstrating how time can be advanced and delayed; how information can be given, restricted or withheld; and how humor and pathos can be generated through various means without recourse to overt extradiegetic commentary like that in 49-17.

Complications, though, exist in the film’s contradictory tones. As a satire of Hollywood as a place and institution, the film relies on an ironic distance that is generated in part by one lone intertitle, and more often by parts of Max Steiner’s score, as well as by certain characters and dialogue. The music and other elements of the film’s narration, however, also encourage the kind of emotional identification with the main character that is typical of the genre of the "woman’s picture" or "tearjerker." Such contradictions of tone, however, are not unique to this film. Other movies, such as George Cukor’s The Women, mix together such contradictory emotional cues.

However, as a film made in Hollywood about Hollywood, A Star Is Born also demonstrates a relatively high degree of "self-consciousness," as defined by David Bordwell: when the narration displays "a recognition that it is addressing an audience" (NITFF 58). The most overt cueing of such audience awareness occurs through the framing of the film’s plot by shots of the the final shooting script for A Star Is Born itself.

Even in the year 2000, such overt self-referentiality is unusual. The script is a signifier of an implied filmmaker, though it does not attribute the script to any one individual. After the high pathos of the film’s closing scene, as Vicki resolves her personal tragedy by appearing at the premier of her newest film, the dissolve to the script serves to distance the audience once again, but not for satiric purpose. Why then evoke such an implied filmmaker at all?

An answer might come through reference to one other factor that is easily overlooked today: the fact that the film is in Technicolor, a process that in the 1930s was most often used for genres such as comedies, swashbucklers, fantasies, and historical epics. Black-and-white film, however, would still be standard for films most associated with contemporary "realism," including tearjerkers. Coupled with the presence of the script, then, the use of Technicolor reinforces the "fantasy" aspects of A Star Is Born, including the sheer improbability of Esther Blodgett’s rise to fame. (Producer David O. Selznick himself called it a "Cinderella" story. [Behlmer 96]). The presence of the script and the use of Technicolor, then, are not only elements of narration but hints of an implied filmmaker, an unnamed agent offering implicit cues not to take Esther’s luck too seriously. Overt—non-verbal—marks of narration and authorship reinforce the sense of film as artifact rather than as "reality" and create a final, deliberate act of distancing.

The overdetermined narration of 49-17 and the evocation of an anonymous implied filmmaker in A Star Is Born still do not complicate the actual stories of either film. There is no overt reason to question the facts that are put before us in these movies. Discussions of "unreliable" narration in film usually focus on contradictions between visual representation and a voice-over soundtrack. However, other modes of handling narrative information can create patterns of "narrators" that are at odds with each other, casting one or more of them in doubt. Such a case occurs in Alfred Hitchcock’s Rear Window.

Rear Window, of course, concerns L.B. "Jeff" Jeffries (James Stewart), a photojournalist with a broken leg who is confined to his apartment. While fretting about his relationship with Lisa (Grace Kelly), a glamorous model whose lifestyle is at odds with his, he spies out of his rear window on neighbors in nearby apartments. Eventually, he comes to believe that one of them, Lars Thorwald (Raymond Burr), has killed his wife. Jeff then tries to convince Lisa and others that he is right.

The problem of narration in Rear Window hinges on how information is given to the audience. The information given ("focalized") through Jeff’s perceptions via eyeline matches and point-of-view shots (along with their Freudian implications) has received the most attention. However, there is another mode of narration in which we are given information by a moving camera, independent of what any character sees. Even though a few critics such as David Bordwell, Donald Spoto, and John Belton have noticed some of these moving shots, they do not note the way in which such shots form a significant and alternative pattern of narration (see Bordwell 57, Spoto 241, and Belton 16).

When they are portrayed as the focalizers of information, Jeff and Lisa often project their own fears, fantasies and preconceptions on the neighbors through their dialogue. At different points in the film, Jeff and/or Lisa compare their lives and their relationship with the bickering Thorwalds, the dancer "Miss Torso," the lovelorn "Miss Lonelyhearts," and a composer who seems unable to finish a song he is writing. The possibility of wedded erotic bliss suggested by nearby newlyweds hidden by their window shades is tacitly contrasted with the asexual domestic stability of an older married couple with a dog but no children.

The "rear-window ethics" of spying are questioned at first by Lisa but she soon joins in, along with Stella (Thelma Ritter), Jeff’s physical down-to-earth physical therapist, who had warned about the dangers of being a "Peeping Tom." Only the representative of the law, Jeff’s friend Detective Doyle (Wendell Corey), offers consistent objections to the other characters’ voyeurism and counters their interpretations of what they think they have seen. However, those interpretations are also seemingly undercut by information given through an alternative narration of panning shots, which for convenience I will simply label "the camera-narrator."

This camera-narrator is marked by a single panning and tracking shot (or sometimes two such shots in succession)—not motivated by a character’s act of looking—that sets up a way of knowing that is potentially, and sometimes actually, at odds with what the characters perceive and how they interpret what they see. In the first shot following the opening credit sequence (in which window blinds are raised by an unseen agent), the camera passes out of the apartment window, apparently following a cat, but it lets the cat wander out of the frame and continues to pan the courtyard, giving us our first view of the layout of the area and a few of the neighbors going about their business. By breaking away from the cat, the camera establishes its independence from the objects in the setting and "announces" its narrational function. This movement concludes as the camera returns to show us Jeff in his wheelchair, asleep and turned away from the window.

This long movement is followed almost immediately by another that shows some of the neighbors engaged in typical activities. Returning to Jeff’s room, where he is still asleep, the camera tracks across props and objects in the setting that inform the attentive viewer of a number of facts: the weather is very hot, the name of the sleeping man is L.B. Jeffries, he is a photographer, he was probably injured while on assignment, and he has a girlfriend who is a model or other celebrity. A negative of Lisa in a picture frame reinforces our perception of Jeff as a photographer (in the age of mechanical reproduction, a negative is the one truly unique image that a photographer can have) and, as a visual pun, even hints at his "negative" feelings for his love (soon to be reinforced by his dialogue with Stella).

Every time this kind of long moving shot occurs, Jeff is markedly not looking out of the window. Thus, one pattern gives information witnessed and interpreted by the characters, while another gives independent and apparently "objective" information. This contrast of narrators underlines a tension in the film between the characters’ stances regarding "rear window ethics." Jeff and, eventually, Lisa operate on the assumption that human behavior is comprehensible, that people’s motives can be discovered through careful observation of behavior, leading them to believe that Thorwald has acted suspiciously and inconsistently, thus offering circumstantial evidence of his guilt. Doyle counters with contradictory evidence that was unavailable for direct observation while claiming that behavior cannot be explained by mere observation and that spying on people is not only unethical but ineffectual.

Doyle’s point seems to have been confirmed by information given to us but not to Jeff by the camera-narrator. While Jeff was sleeping on the night of the supposed murder, another courtyard pan showed Thorwald leaving his apartment with a woman whom we might suppose to be the wife, as Doyle later suggests. We are likely to take such information as true because it has not been filtered by characters’ perceptions. It is the "objective" camera-narrator, though, that turns out to be unreliable. The woman was Thorwald’s mistress, leaving with him as "proof" that no murder had been committed. Jeff’s suppositions, based on observation and assumptions about behavior, turn out to be correct. In film’s final pan of the courtyard, other assumptions are validated, notably Lisa’s contention that "Miss Torso" was not a "party girl" but a woman awaiting her true love’s return. And so, in a sneaky narrative move, "rear window ethics" become validated after all. We can understand people from observing how they behave. The characters and their interpretations are more reliable than the visual information provided first-hand, as it were, by the camera.

One hint of this move is offered by the fact that the film dissociates the camera-narrator from the implied filmmaker, signified by Hitchcock’s own cameo appearance, when he is seen by Jeff in the composer’s apartment. Framed by a subjective point-of-view shot, the director looks back directly at the camera and at Jeff, the only character besides the murderer to do so. The film’s implied "author" is not associated directly with the camera-narrator; instead, his presence marks both the composer’s location and Jeff’s focalization as significant within the film. The significance of the composer is reinforced several times in the film. At one point, Lisa listens to a portion of the song he is working on and remarks that it "might be telling a story about us," even though it has no words as yet. Later, the music stops Miss Lonelyhearts from attempting suicide at a crucial moment. At the end of the film, the camera-narrator once again pans the courtyard, showing resolutions to all the mini-narratives of the neighbors, including the composer who is now playing a recording of his song for Miss Lonelyhearts. As the shot returns to Jeff’s apartment, where once again he is asleep, the camera pans right to reveal Lisa at first reading one of Jeff’s adventurous memoirs but then exchanging it for the fashion magazine Bazaar. The humorous movement is filled with ideological nuances, but the moment is capped by the end of the song, titled, it turns out, "Lisa"—it is a "story about us" after all! The implication of the film’s rigorous narrative structure is that Rear Window is also a story about us, the spectators, projecting our own lives and worries on the characters that we watch, so safely from dark anonymity.

A film like 49-17 displays a certain lack of confidence in the competency of spectators to understand film narrative without redundant narrational cues. Most films, however—including A Star Is Born and Rear Window—assume such competency, some to the point where they actively play with the perceiver’s understanding and invite the perceiver’s play with the film text. While the vast majority of contemporary American films still avoid such deliberate play, probably a greater number invite active play and interpretation than in previous eras. One recent example of the self-conscious play of narration is Paul Thomas Anderson’s Magnolia (1999).

Magnolia has been subject to both critical adulation and fierce dislike. Aside from the film’s sheer length (over three hours), three elements seem to be the focus of objections: 1) the fragmentary nature of the narrative, interweaving stories about at least nine major characters; 2) apparently unmotivated references to "urban legends," culminating in the devastating rain of frogs at the film’s end; and 3) a high degree of self-consciousness, signified by subtitles that sporadically update the time and weather, the recurring motif of the numbers 8 and 2 throughout the film, and a sequence in which the characters all mouth the words of an Aimee Mann song. Magnolia thus provides a significant challenge to viewers who view the film while expecting traditional qualities of narrative unity and coherence.

In the film’s opening sequences, a voiceover narrator tells stories about improbable coincidences, accompanied by images that illustrate them only obliquely. The first challenge for the viewer, then, is to suppose that these stories are in some way relevant to the rest of the film. Those who see no context at all, or only a tenuous one related to the equally arbitrary rain of frogs at the end, are likely to be annoyed by the move.

The fragmentary main narrative presents a similar challenge. Of the nine main characters, two large groups are clearly interrelated. There is one set of stories about a long-running television quiz show, featuring a panel of young contestants, titled What Do Kids Know? The host Jimmy Gator (Philip Baker Hall); Stanley, a current "quiz kid" (Jeremy Blackman); and Donnie (William H. Macy), a former quiz kid, form the core of this group. A sub-group is headed by Jimmy’s drug-addict daughter Claudia (Melora Walters) and a sympathetic policeman played by John C. Reilly. The second major group includes the rich and dying Earl Partridge (Jason Robards, Jr.), his "trophy wife" Linda (Julianne Moore), his male nurse Phil Parma (Philip Seymour Hoffman), and his estranged son, a chauvinistic male-motivational guru now going by the name of Frank T.J. Mackey (Tom Cruise). Even the relationships within each group, though, are revealed only gradually.

One working assumption that viewers might form, especially given the context of the opening vignettes, is that there is little or no connection among these stories, that they simply form an "omnibus" of narratives linked by a particular time and general sense of place. The relationships are gradually revealed, but it requires more active attention to discover some of them than others. The most important—and perhaps most obscure—link is between two major groups—the ones connected with Jimmy, his job, and his family and the ones connected with Earl. Only when we see the final credits for What Do Kids Know? roll by on a television screen do we learn that the show is an Earl Partridge Production, a source of the dying man’s wealth and a cue to the film’s overall theme. The show’s title can be, and often is, a dismissive comment: Kids really know nothing compared to adults. But it also implies that kids know love and anger, shame and pride, and that they know what they really need. The closest that the film comes to explicitly pronouncing such a "meaning," though, comes toward the end when Stanley tells his overbearing father, "You have to be nicer to me." That meaning is further reinforced by the most outrageous bit of self-consciousness in the film, when most of the characters are shown lip-synching Aimee Mann, who sings, "It’s not going to stop until you wise up," implying that the sins of the fathers are literally visited upon the children.

There is one other clue to the structure (and the narrative agency) binding these stories together. When Donnie goes to a gay bar, he talks about how he had once stood out as a Boy Genius. "Now," he continues, "I’m just another . . . another . . . " As we grasp for appropriate cliches ("cog in the wheel"? "face in the crowd"?), an aging barfly (Henry Gibson) provides an apt but unusual metaphor: "Another spoke in the wheel." Here is the key to the film’s structure. All of the major narratives are linked not directly or as subsets of each other, but radiating as it were from a central point, in this case the television quiz show, What Do Kids Know?

Understanding the centrality of the TV show as the film’s narrative hub also helps to clarify the motif of the numbers 8 and 2. During the show’s taping, a member of the studio audience holds up a card reading "Exodus 8:2," like those sports fans who constantly advertise "John 3:16." In this case, though, the biblical reference is to the Plague of Frogs:: "And if thou refuse to let them go, behold, I will smite all thy borders with frogs" (King James Version). The actual rain of frogs itself is the culmination of a series of subtitles announcing the time and weather conditions. These various forms of "narration" in the film, then-some coming from within the mise-en-scene, some nondiegetic, imply a narrational agency that leads to the final inference. The last of the biblical Plagues was the death of the firstborn of Egypt. The final answer to the question What Do Kids Know? is that they know their own expendability.

Magnolia’s overt forms of film "narration" present us in the end with the implication of a creator deliberately manipulating the elements of the text to portray a moment in contemporary society and to call attention to the very act of that portrayal. Its closest fictional counterpart is not some modernist or postmodern text, but Thackeray’s Vanity Fair, with its puppetmaster-narrator who takes the characters out of their box, displays them, and puts them back at the end. It was precisely this kind of overt manipulation that writers and critics in the realist tradition, following Henry James, have objected to. Whether the individual film viewer will accept—or even find—such agency in Magnolia will probably continue to be equally controversial.

As I hope these examples suggest, narratology, if it is to take in the scope of narrative in films, television, comics, and electronic and other media, must look beyond the novel or print fiction in general for its own analytic authority. The precedence of the novel was historically inevitable, yet the novel can no longer account for, or even stand as the principal model of, all narrative forms in all media. We need to be able to account for the specific functions available to narrative in each medium, for the specific elements that empirical creators will "play with" in crafting their narratives. As the first major medium to interpose technology as a source of narration and to combine "telling" and "showing" as dual substances of narrative, cinema provides a likely starting place for any future narratology.