How, then, do notions of narrative agency apply to narrative cinema? Can a cinematic model of narrative lead to a larger model of agency and narration in general? It is difficult to erase the slate and start afresh, partly because filmmakers and film critics have worked for so long (consciously or not) from models in print narrative. While traditional dramatic structure (exposition-complication-climax-denouement) has provided a powerful template for screenwriting, explanations of agency in relaying a film’s narrative have been more dependent on the models (or "schemata," as Bordwell calls them) provided by prose fiction. These schemata are not historically static but develop and change over time. Surely it is not just a coincidence that by the late 1940s the equation of the camera with the narrator, the film with the novel, and the filmmaker with the novelist was being developed in France and elsewhere.[5]

Thus, in 1948, Alexandre Astruc proclaimed the era of the camera-stylo (the "camera-pen"). Astruc prophesied that "the cinema will . . . become a means of writing just as flexible and subtle as written language" (18). But "writing" requires a writer, in this case, the director: "Direction is no longer a means of illustrating or presenting a scene, but a true act of writing. The film-maker/author writes with his camera as a writer writes with his pen" (22). By 1955, Astruc’s compatriot Andre Bazin had famously declared, "Today we can say at last the director writes in film. . . . The film-maker is no longer the competitor of the painter and the playwright, he is, at last, the equal of the novelist" (39-40). Even Sergei Eisenstein, perhaps the most original thinker about the nature of film in its first half-century, was turning in the 1940s to the novels of Charles Dickens for models that would justify his theories of montage ("Dickens, Griffith, and the Film Today").

The fullest expression of the director-writer analogy, though, would come through the "auteur theory," the doctrine of directors as the "authors" of their films proclaimed by the young French critics and filmmakers-to-be of the Cahiers du cinema and other publications. Andrew Sarris, who adapted the doctrine to a "theory" in America defined the true auteur as a director whose films exhibit the traits of "technical competence," a "distinguishable personality," and "interior meaning" (516). In other words, the auteur’s films display technical, stylistic and thematic marks that point back to their origin, the agent of their being, in this case the director-author. There are, of course, other models of creation and agency in film, emphasizing for example the contributions of the screenwriters (often disregarded by auteur critics) and even of the studio itself. In such cases, one might speak of Mr. Deeds Goes to Town as a "Robert Riskin" or "Columbia Studios" film more profitably than as a "Frank Capra" film. The power of auteurism, though, is that, in the words of Peter Wollen, it "does not limit itself to acclaiming the director as the main author of a film. It implies an operation of decipherment; it reveals authors where none had been seen before" (77), allowing artistry to emerge within the corporate products of the film industry. If Sarris’s criteria of "distinguishable personality" and "interior meaning" seem murky, one can still look for common elements within a body of films ascribed to a particular director to find out what they have in common. Thus, by Wollen’s definition, the auteur critic can be freed from the sins of intentionalism: " . . . Fuller, or Hawks or Hitchcock, the directors, are quite separate from ‘Fuller’ or ‘Hawks’ or ‘Hitchcock,’ the structures named after them, and should not be methodologically confused" (Wollen 168).

Wollen’s statement sounds like a virtual paraphrase of the definition of the implied author in literature, or perhaps, as Chatman notes, Booth’s category of the "career author": "the signifier of a certain constancy or common denominator of method among the implied authors of the various works. Its signified is the known subset of features, carried over from other, similarly signed texts, which provides readers with narratively significant information as they make their way through the new text" (CTT 88). In other words, the "career author" includes the common elements of style, subject and theme that are "typical" of a particular author and his or her work, whether in literature or film. Auteurism in general, as Chatman notes, is based on the assumption of such common elements among a director’s films (CTT 130).

As I have tried to suggest, auteurism attempted to respond to one of two main challenges that film poses for a print-based narratology: the (usually) collaborative nature of the filmmaking process. Even in the collaborative medium of theater, a playwright’s work supercedes any individual stage production of that play. A film, though, is more than its screenplay, even if that script has not been drastically altered or even dispensed with in the filmmaking process. The authority of any individual role in filmmaking also seems provisional. A single film can still have more than one director (for example, The Wizard of Oz had three directors in the course of its production). Some producers, such as David O Selznick, play a much heavier role in the actual shooting of the film than others. And the contributions of cinematographers, production designers, editors, and actors—among many others--can leave their own tell-tale marks that will undercut claims to the unique authorship of a particular director. Thus, even though it has become common practice to identify almost any film by its director, auteur critics will usually restrict "true" auteur status to a select few, thus marking "authorship" as indicative of certain qualities of creation, not of creation per se. [6]

The second challenge of film for narratology lies in cinema’s nature as a multi-track medium. If print fiction relies almost entirely on written words to convey a narrative, cinema conveys its narratives (narrates, that is) through spoken and (sometimes) written words, through gesture and action, through setting and lighting, through camera placement and movement, through music, through the editing of images, and through all the other means available to the medium. But who, if anyone, "speaks" that narration? Gerald Prince’s statement that "There is at least one narrator per narrative" (65) may be plausibly defended for print fiction, but it is much more difficult to maintain for film. Literature, as a "telling" ("diegetic," in Aristotle’s sense) medium (as opposed to an "imitating" or "showing," "mimetic medium), usually needs to have a teller. Words must be spoken or written by someone, even if they are simply a record of someone’s speech, but to what sources—what agents—do we attribute images and sounds and the act of joining them together? Any film may have one or more narrators (usually as character or non-character voice-overs), but is there a narrator of the overall text, as Chatman and others would assert for print fiction?

Theater can usually dispense with the issue of a narrator because, by and large, it is a mimetic form, "with all the persons who are performing the imitation acting, that is, carrying on for themselves" (Aristotle 18). But even though narrative films have persons (or non-human characters, as in cartoons, animal films, science fiction, and so on) acting, those actions are always mediated. Staging, lighting and so forth can affect our perceptions of characters and narrative in a live play, but in film those same elements come to us through the camera lens, further mediated by manipulations of film stock, editing, the sound track, and other film elements. In theater, every production of a play is unique and every performance within that production is at least slightly different from all others. Aside from exigencies of exhibition or transferal to video, however, every film screening is the same. The medium may not be the message, but it is a substantive element of the text itself. All of these forms of mediation have the potential to shape our perception of the narrative, its diegesis and the characters within it. In other words, cinematic mediation is itself narration, "telling" and/by "showing." But the question remains, Who or what "tells" and "shows"?

While Chatman and others posit an overall filmic narrator as well as an implied author for each film, David Bordwell sees little use for such concepts in cinema: "Since any utterance can be construed with respect to a putative source, literary theory may be justified in looking for a speaking voice or narrator. But in watching films, we are seldom aware of being told something by an entity resembling a human being. . . . As for the implied author, this construct adds nothing to our understanding of filmic narration. No trait we could assign to an implied author of a film could not more simply be assigned to the narration itself. . . . To give every film a narrator or implied author is to indulge in an anthropomorphic fiction" (NITFF 62).

One crucial distinction between Chatman and Bordwell is that the former sees film, like other forms of narrative, as a medium of communication. Bordwell, however, does not think that such a model works well in film and other media: "Literary theories of the implied author, such as Seymour Chatman’s, take the process of narration to be grounded in the classic communication diagram: a message is passed from sender to receiver. . . . I suggest, however, that narration is better understood as ‘the organization of a set of cues for the construction of a story.’ This presupposes a perceiver, but not any sender, of a message" (NITFF 62).

Chatman, in turn, claims that Bordwell "goes too far in arguing that film has no agency corresponding to the literary narrator and that film narrative is best considered as a kind of work wholly performed by the spectator " (CTT 124). For Chatman, agency is the crucial distinction: " . . . it is not that the viewer constructs but that she reconstructs the film’s narrative (along with other features) from the set of cues encoded in the film. Bordwell . . . does not explain [these cues’] mode of existence in the film, only in the viewer. . . . The viewer certainly hasn’t put [those cues in the projected film], so it seems a bit odd to talk about ‘narration’ as if she had" (CTT 127).

Still, for all their apparent differences on this issue, Bordwell and Chatman are not that far apart in how they describe film narration, only in the terminology that they assign to it. Bordwell admits "the possibility that the narrational process may sometimes mimic the communication situation more or less fully. A text’s narration may emit cues that suggest a narrator, or a ‘narratee,’ or it may not" (NITFF 62, Bordwell’s emphasis). On the other hand, Chatman seems to say much the same thing (but with a difference in emphasis) when he admits that the "narrator is not a human being. . . . agents need not be human" (CTT 133-134). If the narrator need not be human, then what form of agency is it? What exactly has placed the cues within the text? For Bordwell, film narratives may mimic narrators. For Chatman, the film narrator need not be human and is cued by the narrative. Both scholars still appeal to an agency within the narrative process that allows spectators to engage in the re/construction of the fabula/story.

A more serious question and apparent contradiction lies in Chatman’s claim that the film narrator "shows" the film (CTT 134). Obviously, Chatman is not referring to the projector or projectionist; he is referring to an agent that presents the text to a perceiver, not necessarily congruent with the film’s implied author. But just how is the narrator different from the implied author? What is the difference between "showing [presenting] the film" and creating it? How do various forms of film narration (voice-over narration, "purposeful" camera movements, editing, etc.), which may compete with and contradict each other through both sound and image, differ from the overall narrator, the "show-er" who has placed those elements together? Sarah Kozloff offers a similar version of the film narrator, which she distinguishes from other, voice-over narrators in film by referring to it as the "image-maker" (44). Nonetheless, Kozloff admits that "even at his most manifest, the image-maker is an odd, bodiless construction, difficult to assimilate to the rest of our experience of narratives and narrators" (48). It is such insubstantiality that leads Bordwell and others to dismiss such a concept altogether. In addition, the term "image-maker" suggests a reason for the difficulty, since an image-maker would seem to be more akin to an implied author/filmmaker than a narrator as such. The levels of agency (extratextual, intratextual, and intradiegetic; creation of the text and presentation of the text) are being confused.

return to top