21 Towards an Ontology of the Moving Image

Noël Carroll

 

 

 

Introduction

Excerpted from Philosophy and Film, edited by Cynthia A. Freeland, Thomas E. Wartenberg (1995). Reprinted with permission from Taylor and Francis.

The purpose of this paper is to address the longstanding question: “What is cinema?” However, as my title indicates, in the course of my attempt to answer this question, I shall reorient the inquiry somewhat, and, for reasons that will soon emerge, I will ultimately try to approximate an answer to the related but different question: “What is the nature of the moving image?”

Traditionally, cineastes, in the conversation that is sometimes referred to as classical film theory, have raised questions about the nature of film in the expectation that were one able to isolate the essential feature or features of cinema, then one would be in a position to know what style or stylistic choices were the most appropriate ones for the medium.1 The underlying idea here, which we will call essentialism, holds that the essence of a medium, such as film, determines what style should prevail in that medium. André Bazin wrote, for example, that “The realism of cinema follows from its photographic nature.”2

What is perhaps most peculiar about the strongest version of this brand of essentialism is that it regards artistic media as natural kinds equipped with unalterable, gene-like mechanisms that propel their destiny along one vector of stylistic development. But this presupposition is doubly wrong, because: first, artistic media are generally hospitable to multiple, nonconverging and even potentially conflictive stylistic projects (such as montage versus long-take deep focus photography); and second, artistic media are not natural kinds—they are made by humans to serve human purposes. Thus, artistic media are not unalterable. Indeed, they are frequently adapted and altered to serve stylistic purposes—exactly the opposite course of events from that which essentialism predicts.

For instance, the piano was developed at a time when composers were becoming more interested in crescendos, while it is also arguable that Beethoven’s desire for a wider range of notes in the upper register influenced piano design. In such cases as these, the reach of existing keyboards did not limit style, but rather stylistic ambitions reshaped the medium. Thus, it is not, pace essentialism, the pre-existing shape of the medium that dictates style, but style that dictates the very structure and shape of the medium.

This is not to deny that a close look at the medium by a given artist may suggest an avenue of stylistic development. It is only to deny a central premise of essentialism, viz., that the nature of a medium always determines what style is appropriate to it, i.e., that the direction of influence always only goes one way—from medium to a determinate style. For not only may an artistic medium support several divergent and even “incompatible” styles, but style may determine the very shape of the medium, rather than vice-versa.

However, even if we reject the nexus between style and the medium espoused by cinematic essentialists, we still may be interested in the residual question of whether film has an essence, or a set of essential features, or a set of necessary conditions that are jointly sufficient for defining an instance of film. Such an essence, of course, would not indicate anything about which stylistic choices are appropriate; essentialism, as described above, is false. But this does not preclude the possibility that there may nevertheless be some general features of film which, among other things, might help to distinguish it from neighboring artistic media.

In what follows, I will try to identify four necessary conditions of film.

Disembodied Viewpoints

One attempt to isolate the essence of film concentrates on the photographic basis of the cinematic image. Moreover, it is argued that photographic representation is essentially different from other modes of picture making, notably painting. For whereas the relation between a painting and the object portrayed by the painting is something like resemblance, the relation between a photograph and its object is said to be identity. Bazin claimed that the photographic image—and by extension the cinematic image—re-presents the objects, persons and events that give rise to it. He writes: “The photographic image is the object itself. … It shares by virtue of the process of its becoming, the being of the model of which it is a reproduction; it is the model.”3

Though Bazin’s concept of re-presentation is somewhat obscure and though it has been subjected to criticism,4 several contemporary philosophers, including Stanley Cavell, Roger Scruton, and Kendall Walton, have come to defend their own versions of the view, which we might call “photographic realism.”5 Since Walton’s argument is the most elaborate, let me sketch his position.

When we look through a telescope or a microscope or a periscope or into one of those parking lot mirrors that enables us to look around corners, we say that we see the objects to which we have access by means of these devices. These devices are like prosthetic devices; they are aids to vision.6 When I focus my opera glasses on the ballerina, I see the ballerina, rather than a representation of the ballerina. Such devices expand my visual powers; they enable me to see distant stars and microbes. In fact, these devices are not really very different from the corrective lenses that adjust our eyes for near-sightedness and, thereby, enable us to see the world aright. Analogously, photography is a prosthetic device; it enables me to see now my dead grandfather in his youth; or, if I am watching Triumph of the Will, it enables me to see Adolf Hitler.

Photographs are transparent—we see through them to the objects, persons and events that give rise to them. Indeed, what a photograph is a photograph of is counterfactually dependent on the objects, persons and events that cause the photograph to be. That is, if the perceptual properties of the relevant objects had been different, the photograph would have been different. In this way, photographs are like ordinary vision (where what we see is counterfactually dependent on the visible properties of the things we encounter).

However, painting is not transparent in the same way. Paintings are not counterfactually dependent on the visible properties of the objects of which they are paintings. They are dependent upon the painter’s beliefs about those objects. A painting offers us a representation of an object, whereas a photographic image, and by extension a cinematic image, provides us with access to the object that gave rise to the photographic or cinematographic image in the same way that a telescope boosts our perceptual powers so that we see distant towers as if they were close by.

For the photographic realist then, a necessary condition of a genuine photographic image—and, by extension, of a cinematic image—is that we see through it to the objects that cause it; moreover, this property of transparency also marks a crucial differentia between images of photographic provenance and other sorts of pictures, like paintings. For in the case of painting, we do not literally see the object portrayed by the painting.

Yet, before we accept the putative findings that the realist offers us, we must pause to ask whether it really makes sense to say that we see through photographic and cinematic pictures. In favor of this conclusion, we are offered an analogy between photographs and film, on the one hand, and microscopes and telescopes, on the other hand. If we are willing to talk about seeing through telescopes, so the argument goes, why should we draw the line with photographs? Photography is just another prosthetic device. Indeed, can we draw the line between the lenses of the telescope and the lenses of the motion picture camera in any principled way?

I think that we can. If I look through opera glasses at a ballerina or her consort, the visual array that I receive, though magnified, is nevertheless, still connected to my body in the sense that I would know how to get to the place in question if I wanted to. I can orient my body to the ballerina and her consort spatially. The same point can be made about the bacteria that I see through a microscope; I can point my body roughly in its direction.

However, the same is not true with the visual array in the photograph or the cinematic image. Suppose I am watching King Kong and I am looking at the great wall on Skull Island. It is not the case that I can orient my body to the wall—to the spatial co-ordinates of that structure as it existed sometime in the early thirties. I do not know how to point my body either towards the wall or away from it. The space between the great wall on Skull Island as it appears on the screen and my body is discontinuous; the space of the wall, though visually available through the film, is disconnected phenomenologically from the space I live in.

Francis Sparshott calls this feature of viewing cinema alienated vision.7 Ordinarily our sense of where we are depends on our sense of balance and our kinesthetic feelings. What we see is integrated with these cues to yield a sense of where we are situated. But if what we see on the cinema screen is a “view,” then it is a disembodied view. I see a visual array, but I have no sense of where the portrayed space really is in relation to my body, whereas with such perceptual prosthetic devices such as the relevant sorts of telescopes, microscopes, periscopes, mirrors, binoculars and eye-glasses I can orient my body in the space I live in to the objects these devices have empowered me to see. Indeed, I do not speak of literally seeing the objects in question unless I can perspicuously relate myself spatially to them—unless I know where they are in the space I inhabit.8

This requirement, of course, implies that I do not literally see the objects that cause photographic or cinematic images. What I do see are representations, or, better yet, displays—displays whose virtual spaces are detached from the space of my experience.9 But insofar as cinematic images are to be understood as representations or displays, they are better categorized with paintings and pictures than with telescopes and mirrors.

Photographic realism, then, is mistaken. Photographic and cinematic images cannot be presumed to be on a par with telescopes as devices through which the sight of remote things is enhanced. For authentic visuals, prosthetic devices preserve a sense of the body’s orientation to the objects they render accessible, whereas photographic and cinematic images present the viewer with a space that is disembodied from her perspective.

Though photographic realism’s candidate for a necessary condition of cinema has been challenged, it is important to note that the way the challenge has been posed suggests another candidate, viz., that all photographic and cinematic images involve alienated visions, disembodied viewpoints or detached displays. That is, all cinematic images are such that it is vastly improbable and maybe effectively impossible that spectators, save in freak situations, typically are able to orient themselves to the real, “pro­filmic” spaces physically portrayed on the screen.

Nevertheless, though this is a plausible candidate as a necessary feature of film, it is not a feature that distinguishes cinema from adjacent visual media, such as painting. In order to draw that distinction, we need to introduce another necessary feature of film.

Moving Images

Even if a disembodied viewpoint is a necessary condition of film, it is not a feature that enables us to distinguish film from paintings, since paintings like Poussin’s The Triumph of David involve a disembodied viewpoint in exactly the same way that a cinematographic restaging of that painting would. So what distinguishes a painting from a film? In order to answer that question, let us turn to an interesting suggestion proposed by Arthur Danto.

In his paper “Moving Pictures,” Danto considers the relevance of movement to an essential contrast between things like film and other pictorial techniques, like painting.10 Notice that I have said “things like film.” For, as Danto’s title indicates, he is not concerned narrowly with cinema, but with the wider class of motion pictures which, among other things, would also include video. Moreover, this expansion of the class of objects under consideration to motion pictures in general is, to my mind, all for the good, since I predict that in the future the history of what we now call cinema and the history of video, TV and whatever comes next will be considered to be of a piece.

That these objects are already called motion pictures gives us a clue to the role of movement in defining them. But, as Danto realizes, one must be careful about the way in which one exploits that clue. It will not do, for example, to contend, as Roman Ingarden does, that the difference between painting and film is that in films things are always happening whereas paintings, drawings, slides and so on are static.11 For it is the case that there are films in which there is no movement, including: Oshima’s Band of Ninjas (a film of a comic strip), Michael Snow’s One Second in Montreal (a film of photos), and So Is This (a film of sentences), Hollis Frampton’s Poetic Justice (a film of a shooting script on a tabletop with a plant), Godard and Gorin’s Letter to Jane (a film of photos), and Takahiko Iimura’s I in 10 (a film of addition and subtraction tables).

A perhaps more widely known example of a film that is [almost] without movement is, of course, Chris Marker’s La Jetée, a saga of science fiction time travel that is told through the projection of still photographs. Admittedly, there is one movement in La Jetée; however, it is easy to imagine a film like La Jetée, but without any movement whatsoever.

Undoubtedly, the prospect of movies without movement may strike some readers as oxymoronic or even self-contradictory. “What,” they may ask, “is the difference between a so-called film-without-movement and a slide show?” Indeed, might not one suspect that a film-without-movement merely is a slide show mounted on celluloid for the purpose of convenience in projection?

And yet there is a profound difference between a film, without motion, of Jane Fonda’s face and a slide of Jane Fonda’s face. For as long as you know that what you are watching is a film, even a film of what appears to be a photograph, it is always justifiable to expect that the image might move. On the other hand, if you know that you are watching a slide, then it is categorically impossible that the image should move. If you know it is a slide and you understand what a slide is, then it is unreasonable—in fact, it is irrational and even downright absurd—to anticipate that the image might move.

Movement in a slide would require a miracle; movement in a film is an artistic choice and an always available technical option. Before Band of Ninjas concludes—i.e., up until the last image flickers by—the spectator may reasonably presume, if she knows she is watching a film, that there may be movement; but if she knows that she is watching a slide, it is absurd for her to entertain even the possibility of movement.

Moreover, this difference between slides and films can be applied generally across the board to the distinction between still pictures—paintings, drawings and so on—on the one hand, and moving pictures—videos, mutoscopes and so on—on the other hand. With still pictures, it, by definition, is conceptually absurd and even self-contradictory—a veritable category error—for someone to expect to see movement in what she knows to be a still picture, whereas it is reasonable to have an expectation of seeing movement in films not only because most films move, but because even in static films, it is possible that the subject might move up until the last reel runs out and the lights go up.

Once one has seen a static film in its entirety, it is no longer acceptable to anticipate that there will be movement in it a second time around (unless one has grounds for thinking that it has been subsequently doctored). However, on first viewing, one can never be sure that a film is completely still until it is over. This is what makes it reasonable to stay open to the possibility of movement throughout first viewings of static films. But to anticipate movement from what one knows to be a slide or a painting is conceptually confused.

Indeed, one can imagine a slide of a parade and a cinematic freeze frame of the self-same moment of the procession. The two images may be, in effect, perceptually indiscernible from each other. And yet they are metaphysically distinct. For the epistemic states that each warrants in the spectator when the spectator knows which category confronts him are different. With moving pictures, the anticipation of possible future movement is always logically permissible; with still pictures, never.

Like so many of Danto’s arguments, his case for the preceding necessary condition is transcendental.12 He arrays a set of two indiscernible objects—a film of the title page of War and Peace and a slide of the same leaf—before us, and he invites hypotheses to the best explanation to account for the imperceptible, yet real categorical difference between the two indiscernibilia. His own candidate is the difference between epistemic states of spectators with regard to movement when they are fully informed in each case about the status (slide versus movie) of that at which they are looking.

Thus, on the basis of this conjecture, Danto is in a position to conclude that a logically necessary condition for something x to be a moving picture is that it is reasonable or justifiable to expect to see movement in x, or, at least, it is not absurd to anticipate the possibility of movement in x. This feature of moving pictures differentiates them all as a class from still pictures, including paintings.

One putative counterexample to Danto’s way of distinguishing film and motion pictures in general from slides and still pictures in general is an artwork by Michael Snow entitled A Casing Shelved. This is a slide of a frontal view of industrial shelving packed with painting materials. As the slide is held on the screen, an audio tape is played on which Hollis Frampton describes what is on the shelves. His voice directs our glance and we move our eyes in search of the objects Frampton itemizes. Understood historically, the event is reflexive, making the spectator aware of her spectatorial activities and the way in which the voice guides the eye in the “sound film.”

The best interpretation of A Casing Shelved situates it in the tradition of film history, specifically in the history of modernist or reflexive film, as a comment upon the phenomenology of spectatorship and its related structures. And this may tempt some to count A Casing Shelved as a film, even though such a classification contradicts the conclusion of Danto’s transcendental argument. Do we have a counterexample here or not?

The grounds for considering A Casing Shelved to be a film are historical; we get our most accurate interpretation of the piece if we regard it as a station in the evolution of the historical conversation of reflexive film. That is, it secures its best explanation as an artistic contribution or comment to the modernist dialogue in film history. Thus, we have explanatory reasons for classifying A Casing Shelved along with (other?) reflexive art films. Do these explanatory considerations outweigh our categorical or transcendental reasons for discounting its candidacy as a motion picture? I do not think they do for the simple reason that we may make the historical points we wish to make about the place of A Casing Shelved in the tradition of film modernism without maintaining that A Casing Shelved is a film. The only reason to suppose that this is not an alternative is if one assumes that the only reflexive comments that can be made about an art form are in the idiom of the medium in question. That is, for example, reflexive comments on film must be made by films, and artistic comments on the nature of painting must be made in the medium of painting.

But this is clearly false. Happenings, which are in the nature of theatrical performances, make comments on the politics of modernist painting.13 There is no problem with artworks in one medium, like Happenings, making comments on artworks in another medium, nor is their belonging to the history of another art form paradoxical. Thus, of A Casing Shelved, we can concede that it is not a film, while, at the same time, we may agree that its best interpretation is as an episode in film history—i.e., as a comment (from an adjacent artistic medium) on the structure of film spectatorship.

Though Danto’s conception of moving pictures has much to recommend it, it does appear to me to call for one small adjustment in order to successfully accommodate the phenomena in question. For where Danto speaks of “moving pictures,” it would be more appropriate to speak of moving images. “Picture” seems to imply the sort of intentional visual artifact in which one recognizes the depiction of objects, persons, situations and events. But many films and videos traffic in what are called nonrepresentational or non-objective imagery. Films and videos may comprise nonrecognizable shapes and purely visual structures. This, of course, may be moving imagery, or, if it is static imagery, the films and videos in question nevertheless afford the possibility of movement.

X, then, is a moving image only if x possesses a disembodied viewpoint (or—to state it less anthropomorphically—only if x is a detached display), and only if it is logically justifiable to expect movement in x when the spectator of x is informed about its nature. The latter necessary condition gives us the conceptual wherewithal to differentiate films and videos from paintings and slides. But since movement is a legitimate expectation in theater, how will we differentiate the relevant sort of moving images from dramatic representations?

A Difference in Performances?14

Theatrical performances are detached displays insofar as we cannot orient ourselves in their space; and the viewpoint may be equally said to be disembodied in the sense that when we see a dramatic re-enactment of Lee’s surrender at Appomatox, we receive no information about the relation between the spatial co-ordinates of our bodies and those of Appomatox, Virginia. Similarly, though there may be still works in theater—such as Douglas Dunn’s performance piece 10115—throughout the event it is reasonable for us to suppose that movement may be forthcoming. So even if we have already delivered some necessary conditions for the moving image, we have not differentiated things like films and videos from dramatic performances.

Earlier philosophers, such as Roman Ingarden, attempted to draw the line between theater and film by claiming that in theater the word dominates and the spectacle is ancillary (shades of Aristotle here?), while in film, action dominates and words serve only to enhance the comprehension of the action.16 However, there are significant counterexamples to this view, including: History Lessons and Fortini-Cani by Jean-Marie Straub and Daniele Huillet, and Journeys from Berlin by Yvonne Rainer, not to mention Robert Benchley’s once popular comic shorts, Burns and Allen sit-coms, and Groucho Marx’s You Bet Your Life.

Photographic realists, in contrast, try to get at the distinction between film and theater by focussing on the performer.17 Because of the intimacy of the photographic lens to its subject, some, like Cavell, think of the quintessence of film acting in terms of stars, whereas stage performers are actors who take on roles. For Erwin Panofsky, stage actors interpret their roles, whereas film actors incarnate them.18 In the case of movies, we go to see Bogart, whereas in the case of theater we go to see Paul Scofield playing or interpreting King Lear. But is this contrast really compelling? People go to the theater to see Baryshnikov dance and to hear Callas sing no matter what the role, just as they once flocked to see Bernhardt act. We may say “Sam Spade just is Bogart,” but only in the sense that people once said that Gilette was Sherlock Holmes or O’Neill was the Man in the Iron Mask.

But maybe there is another way to get at the difference here. Let’s begin by focussing on the difference between theatrical performances and things like film performances. We might go to a film performance tonight at seven or to a theater performance. In either case, we are likely to be seated in an auditorium, and perhaps each performance starts with a rising curtain. But whatever the similarities, there are also deep differences between a theatrical performance and a film performance.

This assertion may at first strike some philosophers as strange. For it is very common nowadays to divide the arts into those arts that involve unique objects (some paintings, some sculptures) and those that involve multiple copies of the same artwork—there are probably over a million copies of Jane Austen’s Pride and the Prejudice. Furthermore, having distinguished some arts as multiple in some sense, philosophers frequently go on to characterize the multiple arts—like novels, plays, and movies—in terms of the type/token relation.19 But on this account, film performances and theatrical performances don’t appear very different; they are both tokens of types. Tonight’s film performance is a token of The Piano by Jane Campion, while tonight’s dramatic performance is a token of The Frogs by Aristophanes. So, it might be concluded that there really is no profound distinction between theatrical performances and film performances.

However, though a simple type/token distinction can be useful as far as it goes here, more needs to be said. For even if theatrical performances and film performances may both be said to be tokens, the tokens in the theatrical case are generated by interpretations whereas the tokens in the film case are generated by templates. And this, in turn, is related to a crucial aesthetic difference between the two, viz., that theatrical performances are artworks in their own right which, thereby, can be objects of artistic evaluation, while the film performance itself neither is an artwork nor is it susceptible to artistic assessment.

The film performance is generated from a template—standardly a film print, but it might also be a videotape or a laser disk or a computer program. These templates are tokens; each one of them can be destroyed and each one can be assigned a spatial location. But the film—say Broken Blossoms by D. W. Griffith—is not destroyed when any of the prints are destroyed, including the negative or master. Indeed, all the prints can be destroyed and the film will survive if a laser disk does, or if a collection of photos of all the frames does, or if a computer program of it does whether on disk, or tape or even on paper or in human memory.20

Moreover, to get to a token film performance tonight we require a template which itself is a token of the film type. Whereas the paint on the Mona Lisa is a constitutive part of the unique painting, the print on the page of my copy of the novel The Green Knight conveys Iris Murdoch’s artwork to me. Similarly, the film performance—the projection or screening event—is a token of the type, which token conveys Broken Blossoms, the type, to the spectator.

The story is somewhat different and a little more complicated when it comes to plays. First, plays may have as tokens both objects and performances. That is, when considered as a literary work, a token of Major Barbara is a text, in the same sense in which my copy of Emma is a text. But considered from the viewpoint of theater, a token of Major Barbara is a performance which occurs at a specific place and time. Unlike the film performance, however, the theatrical performance is not generated by a template. It is generated by an interpretation. For when considered from the viewpoint of theatrical performance, the play, by Shaw, is rather like a recipe which must be filled in by other artists—such as the director, the actors, the various designers, and so on.

This interpretation is a conception of the play and it is this conception of the play that governs the performances from evening to evening. Actors may change while the interpretation stays the same. Other casts may repeat the same interpretation. For the interpretation is a type, which, in turn, generates performances which are tokens.21 Thus, the relation of the play to its performance is mediated by an interpretation, suggesting that the interpretation can be construed as a type within a type. What gets us from the play to a performance is not a template, which is a token, but an interpretation.

One difference between the performance of a play and the performance of a film, then, is that the former is generated by an interpretation and the latter is generated by a template. Moreover, this difference is connected to another, which is perhaps more interesting, viz., that performances of plays are artworks in their own right and can be aesthetically appreciated as such, whereas performances of films and videos are not artworks, nor does it make sense to evaluate them as such. One may complain that there is a focus problem at a film performance or that the video is not tracking, but these are not artistic failures; they are mechanical or electrical failures or shortcomings with respect to routine procedures.

In theater, the play, the interpretation and the performance are each arenas of artistic achievement in their own right. Though ideally integrated, in our theatrical practices, we recognize that these represent different, separable strata of artistry. We regard the play by the playwright as an artwork, which is then interpreted like a recipe or a set of instructions by the director and others in the process of producing another artwork.22

But in our practices with regard to motion pictures, things stand somewhat differently. If in theater, the play-type is a recipe that the director interprets, and the recipe and the interpretation can be regarded as different though related artworks, in film both the recipe and the interpretations are constituents of the same artwork. When the writer produces a play, we appreciate it independently of what its theatrical interpreters make of it. But in the world of motion pictures, as we know it, scenarios are not read like plays and novels, but are ingredients in moving pictures (or, more accurately, moving images).

Indeed, insofar as moving picture types always involve recipes in conjunction with interpretations, they are more like what are called theatrical productions than they are like play types. Perhaps this is why movies are also called “productions.” But, in any case, it should clarify why people say things like “Many actors can play Prospero and the performance will still be a performance of the play type The Tempest, but it would not be an instance of the movie type To Kill A Mockingbird, unless Gregory Peck played Atticus Finch.” For Gregory Peck’s performance of Atticus Finch is part of the 1962 film To Kill A Mockingbird because Peck’s interpretation, in concert with the director Robert Mulligan’s, is a nondetachable constituent of the film. Saying “No To Kill A Mockingbird without Gregory Peck” loses its paradoxical air when the movie type is no longer compared to the play type, but to a theatrical production. For it is not paradoxical to say: “No instance of the Burton-Gielgud production of Hamlet without Burton,” since theatrical productions, like movie types, have interpretations (by directors, actors, etc.) as constituent parts.

Whereas film performances are generated from templates which are tokens, play performances are generated from interpretation types. Thus, whereas film performances are counterfactually dependent on certain electrical, chemical, mechanical and otherwise routine processes, play performances are counterfactually dependent upon the beliefs, intentions and judgments of people—actors, lighting experts, make-up artists, and so on. Though there is an overarching directorial interpretation of the playwright’s recipe, the realization of the token performance on a given night depends on the continuous interpretation of that plan, given the special exigencies of the unique performance situation. It is because of the contribution that interpretation makes in the production of the performance that the performance warrants artistic appreciation, whereas the performance of the film warrants no artistic appreciation, since that is merely a function of the physical mechanisms engaging the template properly, or, to put the matter differently, it is an issue of running the relevant devices properly.

A successful motion picture performance—i.e., the projection or screening of a film, or the running of a video cassette—does not merit aesthetic appreciation, nor is it considered as an artwork. We do not commend projectionists as we do actors or violinists. It is true that we do complain when the film dissolves in the projector beam, but that is a technical failure, not an aesthetic one. For if it were an aesthetic failure, we might expect people to cheer when the film does not burn. But they don’t. For the happy film performance only depends on operating the apparatus as it was designed to be operated. Since that involves no more than (often minimal) mechanical understanding and savvy, running the template through the machine is not an aesthetic accomplishment and, therefore, is not an object of artistic evaluation. On the other hand, a successful theatrical performance involves a token interpretation of an interpretation type, and inasmuch as that depends on artistic understanding and judgment, it is a suitable object of aesthetic evaluation.

Furthermore, if the argument so far is correct, then it seems fair to surmise that a major difference between motion picture (or moving image) performances and theatrical performances is that the latter are artworks and the former are not, and, in consequence, that performances of motion pictures are not objects of artistic evaluation, whereas theatrical performances are. Or, yet another way to state the conclusion is to say that, in one sense, motion pictures are not a performing art—i.e., they are not something whose performance itself is art.

This, undoubtedly, will sound like a peculiar conclusion, one that will certainly call forth counterexamples. Three come to mind immediately. First: before motors were installed in projectors, film projectionists hand-cranked the performance, and audiences supposedly came to prefer some projectionists over others; weren’t these projectionists performers who deserved artistic appreciation? Second: the avant-garde filmmaker Harry Smith accompanied some of his film screenings by personally alternating colored gels in front of the projector lens; wasn’t he a performing artist? Third: Malcolm LeGrice presented a piece called Monster Film in which he walked—stripped to the waist—into the projector beam (his shadow becoming progressively larger like a monster), while a crashing din sounded loudly. If Monster Film is a film, surely its performance was an artwork.

However, these counterexamples are not persuasive. Since the early projectionists who are cited usually are also said to have cranked the films they thought were boring in such a way that the films were comically sped up, I wonder whether their performances were actually performances of the film types that were advertised, rather than travesties or send-ups thereof—that is to say, comic routines in their own right. On the other hand, both Smith and LeGrice seem to me to be engaged in multi-media artworks in which film or film apparatus play an important role, but which cannot be thought of simply as motion pictures.

What seems so bizarre about denying that moving pictures (and/or images) are instances of the performing arts is that motion picture types are made by what we standardly think of as performing artists—actors, directors, choreographers and so on. But, it is key to note that the interpretations and the performances that these artists contribute to the motion picture type are integrated and edited into the final product as constituent parts of the moving picture type.

We do not go to see Gregory Peck’s performance, but a performance of To Kill A Mockingbird. And while Gregory Peck’s performance required artistry, the performance of To Kill A Mockingbird does not. It requires only the proper manipulation of the template and the apparatus. A performance of a play, contrariwise, involves the kind of talents exhibited by Gregory Peck prior to the appearance of the first template of To Kill A Mockingbird. That is why the performance of a play is an artistic event and the performance of a moving picture is not.

Thus, there are important differences between the performance of a motion picture and the performance of a play. Two of them are: that the play performance is generated by an interpretation, whereas the performance of the motion picture is generated by a template; and, the performance of a play is an artwork in its own right and an object of aesthetic evaluation, whereas the performance of the motion picture is neither. Moreover, the first of these contrasts helps to explain the second. For it is insofar as the performance of the motion picture is generated by engaging the template mechanically that it is not an appropriate object of artistic evaluation in the way that a performance generated by an interpretation or a set of interpretations is. These two features of film performances are enough to differentiate performances of moving images from performances of plays, and, furthermore, the two differentia under consideration apply to all films and videos, whether the film and video types are indexed as artworks or not.

Conclusion

So far we have identified four necessary conditions for the phenomena that we are calling moving images. X is a moving image: (1) only if x possesses a disembodied viewpoint (or is a detached display); (2) only if it is reasonable to anticipate movement in x (on first viewing) when one knows what x is; (3) only if performance tokens of x are generated by templates; and (4) only if performance tokens of x are not artworks. Moreover, these conditions provide us with the conceptual resources to discriminate the moving image from neighboring artforms like painting and theater.

Of course, with these necessary conditions in hand, it is natural to wonder whether or not they are jointly sufficient conditions for something to be what we typically call a motion picture. I think they are not, because treated as a set of jointly sufficient conditions, they are overly inclusive. For example, the upper right hand page corners of Arlene Croce’s The Fred Astaire and Ginger Rogers Book contains photographs of the eponymous couple dancing.23 If you flick the pages quickly, you can animate the dancers after the fashion of a flip book. And though condition #3 above eliminates hand-made, one-of-a-kind flip books from our catch basin, the Astaire/Rogers example meets condition #3, as would any mass-produced flip book, whether it employed photographs or some other mechanically reproduced illustrations. Similarly, Muybridge photos of horses animated by the nineteenth-century device known as the zoetrope fit the formula. However, these do not seem to be the kind of phenomena that one has in mind when speaking of moving pictures in ordinary language or of moving images in slightly regimented language.

One might try to block this species of example by requiring that motion pictures (and/or images) be projected. But that would have the untoward consequence of cashiering early Edison kinetoscopes from the order of the motion picture. Obviously it will be difficult to draw any firm boundaries between motion pictures and the proto-cinematic devices that led to the invention of cinema without coming up with hard cases; indeed, we should predict problematic boundary cases exactly in this vicinity. But, in any event, it does not seem obvious to me that we can turn the four preceding necessary conditions into jointly sufficient conditions for what is commonly thought to be a motion picture without doing some violence to our everyday intuitions.24

Thus, the characterization of moving pictures (and images) that I have proposed is not essentialist in the philosophical sense that supposes that an essential definition of motion pictures would be comprised of necessary conditions that are jointly sufficient for picking out the extension of motion pictures. Nor is my characterization essentialist in the sense of that label used in the opening section of this essay. For the conditions I have enumerated have no implications for the stylistic directions that film and/or video should take. The preceding four conditions, it seems to me, are compatible with any motion picture style, including ones that may conflict with others. Thus, if I have indeed managed to set out four necessary conditions for moving pictures (and images), then I have also shown that, contrary to the previous tradition of film theory, it is possible to philosophize about the nature of moving images without implicitly legislating what film and video artists should and should not do.

Notes

1For an account of the tradition of classical film theory, see Noël Carroll, Philosophical Problems of Classical Film Theory (Princeton: Princeton University Press, 1988).

2André Bazin, What Is Cinema?, vol. 1, trans. H. Gray (Berkeley: University of California Press, 1967), 108.

3André Bazin, What is Cinema?, vol. 1, 14. See also 96–7.

4For criticism of Bazin, see Noël Carroll, “Cinematic Representation and Realism,” in Philosophical Problems of Classical Film Theory.

5See Stanley Cavell, The World Viewed: Reflections on the Ontology of Film (Cambridge: Harvard University Press, 1979); Roger Scruton, “Photography and Representation,” in The Aesthetic Understanding (London: Methuen, 1983); Kendall L. Walton, “Transparent Pictures: On the Nature of Photographic Realism,” Critical Inquiry, vol. 11, no. 2 (December 1984). A similar view has been defended by Patrick Maynard in several essays including “Drawing and Shooting: Causality in Depiction,” Journal of Aesthetics and Art Criticism 44 (1985). In “Looking Again Through Photographs,” Kendall Walton defends his position against Edwin Martin’s “On Seeing Walton’s Great-Grandfather”; both articles appear in Critical Inquiry, vol. 12, no. 4 (summer 1986). Two other defenders of photographic realism include: Roland Barthes, Camera Lucida (New York: Farrar, Straus and Giroux, 1981) and Susan Sontag, On Photography (New York: Farrar, Straus and Giroux, 1977).

6See David Lewis, “Veridical Hallucination and Prosthetic Vision,” in his Philosophical Papers, vol. 2 (Oxford: Oxford University Press, 1986). See also E. M. Zemach, “Seeing, ‘Seeing’ and Feeling,” Review of Metaphysics 23 (September 1969).

7F. E. Sparshott, “Vision and Dream in the Cinema,” Philosophic Exchange (summer 1971): 115.

8Arguments against photographic realism like this one can be found in Nigel Warburton’s “Seeing Through ‘Seeing Through Photographs’,” Ratio, New Series 1 (1988) and in Gregory Currie’s “Photography, Painting and Perception,” Journal of Aesthetics and Art Criticism, vol. 49, no. 1 (winter 1991).

9I prefer to use the term “display” here, rather than “representation,” because the latter implies an image in which we see recognizable objects, persons and situations, whereas “display” allows that the subject of a cinematic view might be what is called “nonrepresentational.” In this I differ from the idiom Currie favors in his “Photography, Painting and Perception.”

10Arthur Danto, “Moving Pictures,” Quarterly Review of Film Studies, vol. 4, no. 1 (winter 1979).

11Roman Ingarden, “On the Borderline between Literature and Painting,” in his Ontology of the Work of Art: The Musical Work, The Picture, The Architectural Work, The Film, trans. Raymond Meyer and J. T. Goldwait (Athens, Ohio: Ohio University Press, 1989), 324–25.

12For a discussion of Danto’s use of transcendental arguments, see Noël Carroll, “Essence, Expression and History: Arthur Danto’s Philosophy of Art,” in Danto and His Critics, ed. Mark Rollins (Oxford: Blackwell, 1993).

13For a fuller explanation of this, see Noël Carroll, “Performance,” Formations, vol. 3, no. 1 (spring 1986).

14This section develops some suggestions that Danto makes in “Moving Pictures,” but it differs from Danto’s major line of analysis of the difference between film and theater because the main line of analysis seems confused. Danto spends most of his time examining the distinction between documentaries of plays and screenplays proper, but this does not help him to isolate the difference between drama and film because the differences that Danto finds between documentaries of plays and screenplays proper turn out to be the same differences that one finds between documentaries of plays and plays. For example, both screenplays proper and plays are about characters and locales, not about actors and sets. Therefore, the contrast between documentaries of plays and screenplays proper does not get at what Danto wants—a distinction between screenplays proper and plays. However, despite this criticism of Danto’s central argument, his more castaway asides have led me toward the analysis in this section.

15For descriptions of this piece see Sally Banes, Terpsichore in Sneakers (Boston: Houghton Mifflin Company, 1980), 189; and Noël Carroll, “Douglas Dunn, 308 Broadway,” Artforum 13 (September 1974): 86.

16Ingarden, “On the Borderline,” 328–29. See also Rudolf Arnheim, Film as Art (Berkeley: University of California Press, 1957), 210–30.

17See, for example, Cavell, 27–28. Also, see Erwin Panofsky, “Style and Medium in the Motion Pictures,” Film Theory and Criticism, eds. Gerald Mast and Marshall Cohen (New York: Oxford University Press, 1985).

18Cavell, World Viewed. Panofsky, “Style and Medium.”

19Richard Wollheim seems to be responsible for popularizing the use of this distinction amongst aestheticians. See his book Art and Its Objects (Cambridge: Cambridge University Press, 1980); see especially sections 35–38. The type/token distinction itself, of course, derives from C. S. Peirce, Collected Papers, vol. 4 (Cambridge, Mass.: Harvard University Press), 537.

20If you can print the code out, then it is theoretically possible for it to be memorized, if not by one person, then by a group—like the population of China.

21R. A. Sharpe presents an argument for regarding interpretations as types in his “Type, Token, Interpretation, and Performance” in Mind (1979): 437–40. Throughout this section, I have benefitted immensely from reading parts of David Zucker Saltz’s unpublished manuscript The Real Stage: Philosophical Foundations of Performance (though, of course, Prof. Saltz is not responsible for any of my errors).

22This is not to preclude the possibility that the playwright may also be the director of and an actor in his/her own play. Nevertheless, his role as an artist-author is different from his/her role as an artist-interpreter, whether director, or actor, or both.

23Arlene Croce, The Fred Astaire and Ginger Rogers Book (New York: Vintage, 1972).

24In “Moving Pictures,” Danto tries to block flip books and the like by claiming that moving pictures in his sense possess the possibility of moving viewpoints (e.g., camera movements), while flip books do not. But I see no reason to think that flip books can’t have moving viewpoints. Think of a flip book like the Astaire/Rogers example where the scene involves a camera movement. Moreover, where the image in the flip book is drawn, the artist can still convey the impression of moving closer or further away from the object depicted.