Words are flowing out like endless rain into a paper cup.
Philosophers of art possess a concern with the essential that ends up renouncing any notion of generic forms, on the grounds that the singular originality of every single work entails the only possible essential genre must be the universal and individual one of art itself. The error—as Benjamin had previously charged the Early German Romantics, discussed above—is to dissolve real and important aesthetic structures or forms into an undifferentiated unity (of art), which denies their irreducible multiplicity.
—Stanford Encyclopedia of Philosophy, “Walter Benjamin”1
In the methodological prologue to Origin of the German Trauerspiel, Benjamin writes:
Through their mediating role, concepts enable phenomena to participate in the being of ideas. … As the salvation of phenomena comes about by means of ideas, the presentation of ideas comes about in the medium of the empirical. For ideas present themselves not in themselves but solely in a correlation of the elements of things in the concept—indeed, as the configuration of these elements.2
I have built this essay on the most rudimentary understanding of Benjamin’s remark. I select a set of mediating concepts—the human model, the gaze, and the field—which are not standard film-studies terms here but carry rather peculiar abstract meanings. By analyzing and developing them in specific contexts, I let these concepts break down, arrange, and combine—“reconfigure”—the elements of things, in this case images, to bring their configuration into view. This overall configuration, this virtual object I want to represent, is what I crudely understand as a question of the images’ formal style. It is the core of the essay.
Most of the images discussed here come from Yoshishige Yoshida’s Confessions Among Actresses (告白的女優論, 1971). The film tells, discontinuously, the stories of three actresses during the two days before shooting begins on a film of the same name. It falls in the final year of Yoshida’s uninterrupted run of feature filmmaking—the end of his prime—and within the period of his so-called “trilogy critiquing modern Japan”: Eros + Massacre (エロス+虐殺, 1969), Heroic Purgatory (煉獄エロイカ, 1970), and Coup d’État (戒厳令, 1973). All four share a cinematographer, Genkichi Hasegawa, and a composer, Toshi Ichiyanagi. They are also all productions of ATG,3 known for its experimental, avant-garde, uncompromising, and forbiddingly difficult films. Heroic Purgatory, the least commercial of them, ranked third from the bottom in admissions among ATG films.
With this basic background in place, I will proceed along the lines described above.
I. The Human Model
I borrow the term “human model” from Robert Bresson, but use it differently. For Bresson, the model stands opposed to the actor: “being rather than performing,”4 stripped of the actor’s theatrical, emotional performance. In Confessions Among Actresses, by contrast, the acting is sometimes exaggerated to excess. Its unreality is intensified until it becomes an outright lie—but that lie contributes to the film’s self-reflexivity. ATG films often had low budgets and sometimes used conspicuously artificial sets, as though filming a stage play. In this film, apart from the principal actors and necessary supporting players, extras have almost entirely disappeared. As Tadao Satō puts it:
Each ATG film painstakingly investigated how to make an art film on a small budget, proposing methods outside the bounds of received wisdom and thereby creating distinctive, new, and efficient styles. Since devoting resources to realistic reproduction and the depiction of atmosphere would increase production costs, filmmakers carefully explored their shooting methods. A logically conceived approach to production, one that reexamined reality from unexpected angles, became highly developed here.5
In this essay, I use “human model” only to emphasize the model’s being: they are an object occupying space. “You prescribe their actions and speech. They give you back a substance, recorded by your camera.”6 This also suggests that, as an object, the human model can be handled however the camera chooses, reduced through staging to nothing more than a being that is there. In Confessions Among Actresses, human models move between frames, continually playing at entering and leaving them, continually being cut and cropped by their edges. Yoshida also frequently arranges the models in the lower half of the image, or pushes them right into its four corners, leaving extensive negative space.
Take Clip 1. Minamikawa (Rentarō Mikuni), the manager of Shōko Ichimori (Mariko Okada), meets Dr. Hata (Kantarō Suga) for the first time. Minamikawa objects to Hata’s hypnotherapy: he believes one should not probe an actress’s inner life, since this can only destroy the commodity called Shōko Ichimori. In his view, only an actress’s exterior is real. In this scene, the camera moves with extraordinary, almost obsessive precision, using the frame to cut through the human models and creating a distinctive, firmly articulated staging.
Figure 1(a) deliberately severs the heads of the two models looking toward one another at diagonally opposite corners of the frame. In (b), the low angle is even more extreme: the vanishing point lies below the bottom of the frame, with Minamikawa’s head placed exactly on that edge and Hata’s eyes near the upper edge. The composition makes ingenious use of the positions of their eyes and the different directions of their looks.
In his well-known essay “Deframings” (Décadrages), Pascal Bonitzer writes:
Deframing is an aberration, a way of exercising the right to look that ironizes the functions of cinema, painting, and especially photography. To borrow Deleuze’s terms: deframing, displaced angles, and the radical decentering of the viewpoint—which cuts bodies off and expels them from the image, focusing instead on dead, empty, or decoratively impoverished areas—make for an art of irony plus sexual perversion. … In principle, the viewer is subjected to deprivation and the model to mutilation. The irony and sexual perversion of compositional decentering belong to a cruel mastery, a cold and provocative incitement to death: the image is wielded like a knife, pushing the living outside the frame and focusing on the scene’s gloomy or dead areas. …
Deframing, or off-center framing, thoroughly disrupts the equation “framing (cadrage) = centering,” undoing and transforming conventional composition. Like reframing (recadrage), deframing is intrinsic to cinema and hardly rare, but filmmakers often exclude it to preserve the integrity of the image and viewpoint.7 Some directors, for example, have little choice under the constraints of widescreen but to crop the actors’ bodies out of the image, as Figure 2 shows. In Figure 1(b), however, Yoshida deliberately adopts a low angle and cuts through the models’ heads, finding a new way to express their relationship in depth. This composition also becomes the core of Clip 2.
It is worth noting that Clips 1 and 2 consist entirely of low-angle shots—a feature, in fact, of many shots discussed throughout this essay. Notice especially how, in some images, the ceiling itself becomes the characters’ background. André Bazin writes in Orson Welles: A Critical View:
Low-angle shots recur throughout Citizen Kane, so that … we soon become less conscious of the technique, even while remaining under its spell. The method very likely rests, then, on a definite aesthetic intention: to impose a particular dramatic illusion. We might call it an infernal illusion, since the upward-looking gaze seems to rise from underground, while the ceiling closes over everything, making escape from the set impossible and fully expressing the inevitability of the outcome.
François Truffaut adds in his preface to the book:
Since this is Orson Welles’s favorite angle, he puts the camera on the ground and shoots upward. But does this not also make us watch the protagonists as we would from the first ten rows of a theater’s stalls? … Welles burst into cinema from theater and radio; he was bound to consider the similarities and differences among these arts. Shooting from a low angle—the very viewpoint of an audience in the stalls, or of a theater director conducting a rehearsal from the auditorium—necessarily entails using a short-focal-length lens, with its wider field of view. From this angle, any other lens would only eliminate depth and destroy the sense of space.
Bazin, we might say, describes the sensory character of the low-angle shot beautifully and expressively, while Truffaut puts his finger on its affinity with theater. The self-reflexivity of Confessions Among Actresses—actresses playing actresses, and the depiction of the roles they play in life—fits this aesthetic perfectly: the ceiling over the stage extends outward, covering and blocking the sky of the characters’ lives.
Another feature demands attention in the staging of Clips 1 and 2, driven as it is by the frame’s cutting through heads: the head generates and sets in motion the vectors or patterns of sight lines. As Bresson describes it, the model has “a moving head on a moving body, with moving eyes inside it,” while the look is “the force radiating from the eyes.”8 The empty field thus becomes a place for sight lines to roam. The two-person arrangement in Figure 1(b), built from the relationship between heads and sight lines, develops into numerous variations throughout the film (Figure 4). Different patterns of intersecting looks divide the image into different force fields.
The models’ sight-line vectors often point outside the field. Frequently they do not even require an object to point toward; at other times they anticipate that object’s movement. The models seldom meet one another’s eyes. When they occasionally do, as in Clips 1 and 2, their looks immediately flee again. As with magnets, like poles repel and unlike poles attract, creating a peculiar air of estrangement between people.
Clips 3 and 4 come from Heroic Purgatory. In Clip 3, the head is placed in a corner, its line of sight blocked by the frame. But we should also notice the contrast between the movement and stillness of Rikiya Shōda (Kaizō Uda) and his wife Nanako (Mariko Okada). Nanako turns to look at Shōda, but he remains expressionless. Just as the area he occupies in the image has been reduced, his movement is restricted to a minimum: his only response is a slow shake of the head. Clip 4 resembles a meeting of the secret cell described in Dostoevsky’s Demons. In something like a lateral tracking shot, the camera tries to catch the slightest movement of each model’s eyes. The models’ heads evade looks, then pass them on, as though a game of catch were conveying an elusive atmosphere through the darkness.
Whereas most figures in Clips 3 and 4 remain in their places, those in Clips 1 and 2 continually shift position: they separate, share the frame, and separate again. This is a classic way of repeatedly creating and dispelling suspense through blocking and camera movement. The camera opens gaps between people that may then be closed, producing dramatic effects purely through their movements of attraction and repulsion. Stage mechanisms can also assist their entrances into and exits from the image.
In Clip 5, a sliding door effects the exchange between one actor leaving the frame and another entering it. Shōko Ichimori questions whether the aspiring actress Rie (Kiwako Taichi) is in league with Shōko’s husband and her manager, Minamikawa. Rie says nothing and makes no response. The composition places her in the extreme foreground, blurring the background to emphasize the uneasiness in her eyes. She then escapes the frame, and Dr. Hata appears to comment. The staging thus serves the narrative well, fluidly presenting the interactions among the characters.
The appearance and disappearance of figures and faces—these additions and subtractions of information—can also be achieved entirely through shifting occlusions within a shot. Clip 6 fully demonstrates how intricately four people’s reactions to one another, and the establishment and dissolution of connections between their sight lines, can be woven together in live-action filming. Rie and Minamikawa have an intimate relationship they conceal from Shōko; Shōko and Dr. Hata have one they conceal from Minamikawa. Minamikawa is the knot in Shōko’s psyche that Hata urgently needs to untie, while Hata is the lover Minamikawa has yet to discover. This is part of the same scene as Clip 5. Its limited, enclosed interior, with no view outside, creates a theatrical feeling. Within this field, the relationships among four people—and a fifth, absent person, Shōko’s husband—are played out. First Rie moves deeper into the image. Hata watches her, and Minamikawa’s back, previously hidden by Rie, comes into view (Figure 5(a)–(b)). Rie then turns, showing her face to the viewer, and joins Hata in looking at Minamikawa (Figure 5(b)). Shōko walks steadily toward Minamikawa; Rie and Hata both look at her (Figure 5(c)). In Figure 5(d), Minamikawa turns toward Shōko and meets her eyes, revealing his face to the viewer. Rie looks at Minamikawa, Hata at Shōko. With the four heads as point sources, the image presents a complex, intersecting force field of sight lines: the four-way relationship that has been gathering since the shot began reaches its compositional peak. In Figure 5(e)–(f), Minamikawa briefly meets Hata’s eyes, then turns back to Shōko, bringing the shot to an end.
With authorial intention and chance both at work, such minute staging within a shot can even yield an unexpected stroke of genius. In Clip 7, as in Clip 3, the director Nose (Isao Kimura) is placed in the bottom-left corner, his line of sight blocked by the frame. He remains motionless, sluggish and unresponsive, contrasting with the movement of the actress Aki Kaidō (Ruriko Asaoka) and letting her obscure and cut through him as she advances into the foreground. Readers who have watched Clips 3 and 7 can probably see quite easily why I chose the term “human model.” The figure is reduced to a point source emitting sight lines, or even to a placeholder. Miklós Jancsó’s The Red and the White (1967) offers a good example: people are reduced to positions in a simple sign system, and their movement consists merely of exchanging places within this static structure, circling absurdly. What is remarkable in Clip 7 is Aki’s slowing pace on the stairs. With her first step, Nose’s eyes fall exactly at the edge of her throat; with her second, the edge of her clothing slices across his face at the corner of his eye. One could even go so far as to say that Nose has been reduced to nothing but eyes. This startling suggestion about director and actress, looking and being looked at, receives still clearer expression in Figure 6.
One final point for this section: Yoshida and Hasegawa shoot from unconventional camera positions, through gaps between people or things, continually enclosing figures in frames. They use foreground obstructions—frames within frames, sometimes appearing as a mass of black—to enlarge the area occupied by nonhuman, manufactured things and shrink that occupied by people. This heightens the image’s inorganic quality and imbalance, sometimes squeezing people into tiny spaces. These stylistic features resemble those of Akio Jissōji, who likewise emphasizes compositional aesthetics and unusual angles. Figure 7 shows this kinship. Figure 8 further emphasizes how widely Yoshida uses these techniques across different scenes, and the cold geometry of his compositions.
Jissōji’s style eventually entered the genetic makeup of Japanese animation, inherited, developed, and disseminated by Hideaki Anno, Tatsuya Oishi, and others. We can see this tendency toward inorganic composition in Anno’s latest Evangelion theatrical film at the time of writing (Figure 9). In “The Targeted Town,” the Ultraseven episode used here for comparison, the accumulation of unconventional shots gives viewers an extraordinary experience beyond what the script supplies. Form acts on the viewer without passing through content; it creates a direct route to emotion. The famous composition of Alien Metron and the low table (Figure 10) particularly impressed Anno for the way it places “the extraordinary amid the everyday.” But we must also notice how the frame fits tightly against either side of the two characters. With Jissōji, unnecessary, superfluous description of the surroundings is forbidden.
A particularly important feature of “The Targeted Town” is its virtual elimination of establishing shots, making it harder for viewers to grasp the relationship between characters and space. Excluding this signifier—the functional shot—also obstructs the production of meaning through the suturing and cooperation of signifiers, leaving viewers unsettled. By frequently cutting to close-ups in which objects, faces, or bodies occupy half or more of the image, Jissōji in fact reduces information about the surroundings to a minimum. An unusual angle makes the elements in the image crowd and press against one another. Viewers want to understand the surroundings but cannot get their bearings, and become anxious; backlighting further reduces the available information. These experiments with information, emotion, and sensation are the aspect of Jissōji’s direction most akin to Japanese animation.
Unlike Jissōji’s “prohibition,” however, Yoshida lays the surroundings open to view. As the camera shows us—for instance, in its playful placement of people in the four corners—the actors perform in a vast world. Yet this openness is fundamentally “useless,” “meaningless.” The opened world instead reveals the world’s “nothingness.” Perhaps, in Jean-Paul Sartre’s terms, being-in-itself does not exist for human beings and has no meaning for them. This will be one of the subjects of the next section.
II. The Field9
Figure 11 presents a sequence of shots. Shōko is undergoing hypnotherapy, though this, too, is only her playing a part as an actress. In (a)–(d), Yoshida tries four camera angles in turn: high, low, eye-level, and overhead, filming Shōko’s face from different directions. He deliberately, almost playfully, makes extreme changes in the amount of negative space. The figure seems alternately close and remote, producing a strange visual experience in a scene concerned with psychological analysis—that is, with probing beneath a character’s skin. This space, which surrounds and envelops the figures and can be occupied and organized through staging, gives direct visual expression to the latent object I call the “field”: “emptiness,” or “nothingness.” But the concept does not stop there. Quite naturally, I also use “field” for the relations between the frame and other elements of the image, and for the resulting distribution of visual tension. The “force field of sight lines” mentioned earlier belongs here. More broadly still, “field” can mean the air or atmosphere onstage, or the character of a space. Yoshida frequently uses two devices, mirrors and doors, to reorganize the space available for staging and alter the atmosphere. These occupy the first half of this section. In the second half, I will argue that the field is actually the site where meaning is generated and annihilated in cinema, containing a mechanism through which meaning operates. This is why I have gone to the trouble of devising another concept alongside existing terms such as “space” and “atmosphere.”
A. Mirrors
One feature of Yoshida’s work is its meticulous calculation of mirror images. In Confessions Among Actresses, mirrors are used particularly often in scenes involving the actress Makiko Isaku (Ineko Arima). By extending and duplicating a virtual space behind their surfaces, mirrors give scenes strange and intriguing spatial properties. They are therefore not merely reflective planes: by thoroughly transforming spatial relations, they can directly affect the staging of an entire scene.
In “Of Other Spaces,” his lecture to a group of architectural researchers, Michel Foucault describes the mirror as a mixed experience that joins utopia and heterotopia. He begins:
The mirror is a utopia, after all, because it is a place without a place. In it, I see myself where I am not, in the unreal space that opens virtually behind its surface. I am over there, where I am not; an image grants me my own visibility, enabling me to see myself where I am absent.10
In a sense, Clip 8 repeats Clip 1, virtualizing and expanding its space. Here Makiko and her costume designer Nobu (Mazuruka Kubo) revisit Makiko’s past as a suicide obsessive through recollection, questioning, and answers. As in Clip 1, they “occasionally meet each other’s eyes, only for their looks immediately to flee again,” and they pass from sharing a frame to being separated. Intriguingly, as they circle in front of the mirror, the number of people in the image changes several times—if we count the reflections in virtual space as people. The multiplied Makiko obscures Nobu twice with a single movement.
At the beginning of the film, Nobu believes Makiko’s fabricated story of a lovers’ suicide pact and feels deeply sympathetic. Having since learned that Makiko is in fact a suicide obsessive with a history of failed attempts, Nobu now tries to question her about the truth of her past. She keeps turning her eyes toward Makiko.
But the sight lines in the virtual space behind the mirror reveal their own unreality. The utopian image the mirror presents—Nobu looking at Makiko—is utterly destroyed when Makiko leaves its frame. As Figure 12(a) shows, sight lines cannot travel out of the virtual space the camera has fragmented. Their force field is obstructed and localized by the narrow mirror frame. The figures emit sight lines but cannot actually “see” anything. Consider especially the obstruction created by the cabinet and the three books deliberately placed on it, as well as the other mirror compositions that repeat this theme by bending sight lines, as in Figure 12(b) and (c). Interestingly, when Makiko admits the truth of the suicide pact—that the man with whom she tried to die was her own stepfather—the two return to a single mirror frame. But now Nobu tells her: “I’ll never believe you again. I’ll never believe any of you actresses again!” (Clip 9).
Of course, the mirror’s interest goes much further. At the beginning of Clip 9, neither the upper nor lower edge of the mirror is shown directly, misleading us into taking it for a doorway. The camera boldly follows the reflection, allowing a figure to “leave” the mirror’s frame and enter the camera’s frame, as though passing back and forth between virtual and actual space through a magic door. The mirror thus effects a transformation between an object’s image and the object itself. Second, although the figure has clearly left the mirror frame in the first shot, she is caught by the mirror frame again from the second shot’s angle. This resembles cinematic “reframing”: the mirror becomes a camera, recording singles and two-shots. Conversely, the camera reflects the world like a mirror. Finally, notice Makiko’s face reflected and caught in the pane of glass beside her when the Makiko in the mirror turns toward the departing Nobu—an image of an image. What, then, does this recording and capturing by the mirror mean?
After describing the mirror as a utopia, Foucault continues:
Yet the mirror is also a heterotopia insofar as it really exists and acts back upon the position I occupy. From its standpoint, I discover my absence from where I am, since I see myself on the other side. From this gaze directed toward me, from the depths of the virtual space beyond the glass, I return to myself, look at myself again, and reconstitute myself where I am.11
When I look into a mirror, I see myself seeing myself. This is precisely the form of the cogito—I think that I am thinking—and of the displacement, division, or differentiation that necessarily exists within the subject. Before the mirror, I move back and forth between the virtual and the actual, watching my own performance:
Thus two different egos [moi] one of which, conscious of its freedom, sets itself up as independent spectator of a scene which the other would play in a mechanical fashion. But this dividing-in-two never goes to the limit. It is rather an oscillation of the person between two points of view on himself, a hither-and-thither of the spirit …12
This goes some way toward revealing the performative nature of Makiko’s suicide attempts. Near the film’s end, she again tries to kill herself before a mirror, and again fails. She throws the scissors she meant to use and shatters the mirror, declaring an end to this suicide performance in which the two subjects, actress and spectator, coincide. The mirror remains, then, a familiar symbol of the subject’s division and reconstitution.
Jacques Lacan, however, would point out that this consciousness of “seeing oneself see oneself” is an illusion: a mode of looking I imagine to accommodate, evade, and ignore the gaze. It is not, in fact, that I am looking, but that I am being looked at—uncannily, by an object.
We are not primarily conscious subjects viewing the world, but rather we are always-already “beings that are looked at.” There is a fundamental separation between the eye and the gaze. While “I” see from only one point, I am looked at from all sides. There is a gaze that pre-exists my subjective view—an all-seeing to which I am subjected. Žižek summarizes these notions well:
The eye viewing the object is on the side of the subject, while the gaze is on the side of the object. When I look at an object, the object is always already gazing at me, and from a point at which I cannot see it.13
Foucault’s “gaze directed toward me,” then, is emitted neither by “me” nor by “my” reflection. It is the gaze of the mirror that records like a camera and can always catch the figures. From a point we cannot see, the mirror is always watching the actresses perform on the stage of life. To some extent, this repeats the earlier claim about the low-angle shot: there is “no escape from the set.” This relationship between looking and being looked at can also extend to the relationship between spectator and film:
As Merleau-Ponty points out, in the “spectacle of the world,” it is the human being who is looked at. We think we are looking through our own eyes, but what we cannot see is that we are already under a certain gaze. We think we go to the cinema to “watch” a film, but we are not really “watching”: through editing, mise-en-scène, and so forth, the film has already determined that we must be under a certain gaze. Film is like fodder for our eyes, feeding them so that we will not look. …14
Cinema makes us lay down looking as we would lay down our weapons. Yoshida criticizes this evasion of others’ looking on the filmmaker’s part, a point developed further in subsection C. In my analysis of sight lines in Figure 12(a), readers may have overlooked the wooden carving on the cabinet. It stands in exact contrast to the figures who “emit sight lines but cannot actually ‘see’ anything.” These false eyes may be a reminder to viewers: it is not only the people in the film who are being looked at, but also we who imagine ourselves to be watching it. Some of Yoshida’s deframings in fact serve more or less related purposes: resisting the forced convergence of the spectator’s surplus visual attention on a single point; using the peeping effect of foreground obstructions to make spectators conscious of their own act of looking, rather than allowing their gaze to be “tamed” (dompte-regard); and emphasizing the frame so that viewers exercise their own imaginations. As Shigehiko Hasumi says in an interview, spectators must reclaim the freedom of their eyes from the centering imposed by cinema as a cultural institution:
… There is an enormous amount of information in a single instant of a film image. Within it, a centering takes place. Through composition, lighting conditions, and distance from the subject, the filmmaker indicates a point: “Look here.” Certainly, we do need to pay attention to such places, but this is ultimately only a cultural institution. By resisting that centering, our eyes can see even things squeezed into the edges of the image. Sometimes something the filmmaker never intended can shake our sensibility.15
Here we have already anticipated a good deal of Section III. I will discuss the concept of the gaze in Yoshida’s cinema more fully there. A slight, elementary borrowing from Lacan is useful to us here, since this mode of looking—“things looking at people”—is so important to Yoshida, even though he does not necessarily discuss it in psychoanalytic terms.
Let me conclude this part with one further mirror sequence from Heroic Purgatory.
The staging resembles Clip 9: a figure leaves the camera frame, enters the mirror frame, then exits through a frame within that mirror frame—a “frame within a frame within a frame.” The choice of angle also places two people looking at one another across the depth of the scene in diagonally opposite corners of the image. The center becomes the virtual space behind the mirror. Spatial relationships are displaced and complicated, producing an estranging effect.
B. Doors
A door connects two spaces and can be opened or closed. It readily becomes a vehicle for metaphor and is therefore especially prominent in image analysis. Many films also shoot through doorframes or use doors to conceal. Here I will briefly introduce Yoshida’s use of doors to manipulate space—the field.
Makiko wants to know who told her mother (Yumeji Tsukioka) that she had visited the snow country where she once tried to die with her stepfather. She does not yet know that her father is still alive, that it was he who informed her mother, or that her mother has concealed his survival from her for twelve years. In classic fashion, Clips 11 and 12, parts of the same scene, move through the opening and closing of doors from connected, continuous space to divided, split space. To evade her daughter’s questioning eyes, the mother goes through a doorway into another room, trying to create an interruption. Knowing the question cannot be avoided, however, she waits just beyond the sliding door. As in Clip 5, the door serves as a mechanism for bringing a person into the frame. It establishes and opens a small space in which the daughter can question and communicate with her mother at close quarters. But the mother’s answer is an unconvincing lie and plainly fails to satisfy her daughter. In the end, the two women each close a door, Makiko shutting out her current lover, Karasawa (Gō Itō). Clip 11’s use of this stage mechanism is already conspicuously calculated: the mother is deliberately made to wait behind the sliding door and the daughter to open it, their psychological advances and retreats played out like a theatrical routine. In Clip 12, Karasawa’s look points toward an offscreen source of suspense, which the camera dispels by continuing its movement to show the mother closing her door.
Doors can also create an unsettled atmosphere. In a sense, Yoshida’s work has something of the thriller about it. A row of Japanese sliding doors resembles the closed doors along the long corridors of Stanley Kubrick’s The Shining. The unstable field partitioned off by a door may take form at any moment, as though wandering, overflowing fear and tension were suddenly to gather. Yet we do not know at which door this will happen, or what will happen: a sudden change, a turn, can occur anywhere.
Clip 13 is Aki Kaidō’s recollection of being molested by a teacher (Yūsuke Kawazu). The camera films her response through a doorway and advances toward her, expressing a threat invisible to the spectator but drawing near. Yet no one is behind Aki as she runs away. Only when a sliding door suddenly opens at the corner does the diffuse field take visible form.
There is a similar example in Eros + Massacre, where Japanese sliding doors form an intricate labyrinth. After Itsuko Masaoka (Yūko Kusunoki) stabs Sakae Ōsugi (Toshiyuki Hosokawa), he lies motionless on the floor. Masaoka staggers away, only for Ōsugi unexpectedly to appear at a sliding door and seize her. They struggle, and she breaks free again. Of course, Yoshida does not use doors this way only in the extraordinary situations I have chosen: he also creates surprises in everyday life.
Clip 15, at the end of the scene, stages Ōsugi’s death theatrically. The sliding doors fall one after another, and the spaces they have constructed and partitioned collapse and merge. The field’s earlier instability and tension reach a critical point and finally move toward calm.
Heroic Purgatory has a stage device similar to that in Clip 15: an electric roller shutter. Doors are repeatedly used as interfaces between times, entangling events from 1952, 1970, and 1980—past, present, and future. Confessions Among Actresses inherits these surreal qualities from Eros + Massacre and Heroic Purgatory, turning its attention to the intermingling of reality and dreams. In the scene shared by Clips 5 and 6, Dr. Hata proposes projecting Shōko’s dream into reality. Rie is to play the Rie of Shōko’s dream, and Hata her husband, as a means of analyzing it. But once questions about Rie’s relationship with Shōko’s husband begin, the scene gradually goes off the rails. Rie’s untenable account and exaggerated performance make it increasingly chaotic.
In Clip 17, after giving a false account, Rie flees into a room. The reenactment of the dream continues with one participant absent. … Shōko approaches the room as she did in the dream and opens the door—Rie is lying naked on the bed, exactly as in Shōko’s dream (Figure 13). Unlike the deliberate role-playing that preceded it, here, beyond the door, dream and reality actually coincide. A surreal situation emerges, and the atmosphere abruptly changes. Within this enclosed interior, the characters begin to act irrationally and inexplicably. Even the previously indifferent Minamikawa joins in, acting out his own desires in an almost deranged manner within the positions and relationships supplied by Shōko’s dream.
C. The Centrifugal Field
As we have seen, Yoshida’s cinematography appears heretical: it departs from and dismantles a certain cinematic tradition—framing = centering—producing unexpected visual shocks, “irony plus sexual perversion.” Yet by continually using deframing to remind viewers that the frame and its edges exist, freeing their eyes and calling the act of looking back into play, Yoshida returns to cinema itself. Comparing cinema with theater, Bazin writes:
… Cinema’s principle is to deny every boundary that limits action. The screen is not a picture frame but a mask revealing only part of an event to the viewer. … Unlike stage space, screen space is centrifugal. … Baudelaire compared theater to a crystal chandelier. This dazzling, intricately fashioned ring of artificial crystal refracts the surrounding light toward its center, bathing us in its magical radiance. If we wanted an opposing symbol for cinema, we could choose the usherette’s little flashlight. Its wavering beam passes like a flickering comet through the night of our waking dream: the limitless space that extends in every direction around the screen.16
Screen space is centrifugal, yet the center of the screen precisely accommodates the tendency of viewers’ eyes to settle in the middle. By organizing elements at the center, as many commercials do, images feed and tame the eyes, preventing them from “looking.” Yoshida observes: “The center of the screen is also the center of power. … Isn’t this cinema’s totalitarianism? … Without realizing it, we follow the story as its maker tells it, and end up being shown the film on the maker’s terms alone.”17 A thoroughgoing centrifugality, the freedom of “denying every boundary that limits action,” or, as Deleuze puts it, “just because it lacks a centre of anchorage and of horizon … [it] could go back up towards the acentred state of things …”18—this is what cinema ought to be, but it has been constrained and bound by the so-called cultural institution. Yoshida’s field seems to possess theatrical qualities, yet its opposition to a centripetal magnetic field—ordered and organized around a center—returns it to cinema’s nature. Here I call it the “centrifugal field.” This centrifugality also entails a rejection of story-centrism.
Confronting films that privilege story, Yoshida criticizes “the creative side—the filmmaker—who refuses to open himself outward, refuses to let himself be caught by another’s eyes, and constantly settles for a self-enclosed form. Afraid of being wounded and bleeding amid the circumstances around him, he wants to shut himself inside his cave and survive that way.”19 In this sense, deframing is the form Yoshida adopts to receive the spectator’s look openly and prevent both director and spectator from becoming trapped in the world of the story. Moreover, the emergence of surreal situations, as in Clips 15 and 16 and Figure 13, banishes the film’s supposedly rational space, keeping spectators at a distance from the story. Discontinuous narration likewise attempts to resist the story’s magnetic field—for example, the intercut collage of the three actresses’ stories in Confessions Among Actresses.
Contrary to the nature of live-action filming, Japanese commercial animation often displays a centripetal tendency, whatever the apparent centrifugality of its images—that is, its frequent use of deframing or off-center composition. It presents a centering imposed by a cultural institution, a “giving-to-be-seen” (donne-à-voir). Animation, after all, is an intentional construction: an additive process that brings things into existence around an intention. The principle resembles that of live-action cinema’s centripetal movement. What we call “centering” is already present in the choice and intention of framing; a human element intervenes between camera and thing, and humans are animals afraid of the centrifugal. The conventional definition of storyboarding as “turning the script’s words into images” may help readers understand why animation’s centrifugal movement is inhibited.
Take episode 19 of Mobile Suit Zeta Gundam (1985), storyboarded by Jun’ichi Satō. Through photographic techniques such as focus, character blocking and the relation between movement and stillness, framing, directional lines such as eyelines and perspective lines, lighting, and color, space as a unified whole can be “structured”20—organized into a hierarchy that invisibly guides the viewer’s attention. Figure 14(a) and (b) are adjacent shots. Three elements in (a) are easy to identify—eyeline, framing, and color—and I have marked them more clearly in (c). Together they emphasize a particular part of the space: what the direction wants us to see and read.
The space occupied by the couple is precisely what Four sees and registers, as her line in the following shot, (b), confirms. Conversely, “What am I thinking?” psychologizes the space in the preceding shot. The couple’s spatial position represents the desire to fall in love like an ordinary person, a desire that Four—now a Cyber-Newtype, subjected to mental reconditioning—continually pushes into the background of her psyche. Yet it still breaks away and exposes itself so clearly in this space, laying the groundwork for her later meeting with Kamille. Through a brief shot lasting about two seconds, the direction deftly constructs a subject position within the text: viewers identify with Four, entering her emotions and predicament.
The warmth of the colors here also contrasts with the episode’s later scenes, when the Psycho Gundam turns Hong Kong into a sea of fire. This is how a good animation director constructs and shapes images around dramatic logic. Even in Mamoru Oshii’s Patlabor 2: The Movie (1993), where drama is reduced to a minimum, the images remain centered on the depiction of a situation. As Satoshi Kon remarks in a lecture:
With animation, if you don’t draw it, it isn’t there. In live-action film, you sometimes catch something by accident. Of course you have an intention when you shoot, but even then, you can’t say the shapes of the clouds drifting through the sky were intentional. In animation, everything is intentional. Nothing without intention appears. … After all, you can’t draw without an intention.21
It is on the basis of this decisive difference between animation and live action that Hiroki Azuma argues that the critical logic of Hasumi and his disciples cannot be applied to animation:
Through a director’s shots, we encounter “events” and “things that cannot be named” beyond intention. Criticism means grasping and conveying this moment. … The critic’s task is to capture the instant when a creator encounters something exceeding his own plans, and convey it to readers. … In animation, we cannot capture the moment of a creator’s encounter with something beyond intention. Live-action cinema has an outside called “reality,” from which it cuts out images. Animation has no such outside.22
Only by understanding these points can we understand why Oshii both invests a shot with excessive meaning, preparing an excess of directorial intention, and places birds without any symbolic meaning in Patlabor 2. The birds “are noise mixed into the ‘world’ and ‘story’ depicted by this work. They can only be called heterogeneous signs inserted to disrupt the sign system.”23 They disturb animation’s centripetal order and invade the film’s rational space. As a disintegrating force of centrifugal movement, they are the image’s death drive.
In Figure 15(a), Oshii deliberately adopts a decapitating composition and holds the shot for a considerable time so that the poster on the wall will make an impression. Yet it is noise with no connection to the story.24 This puts Oshii’s method into practice:
Among the shots making up a film, shots that contribute nothing whatsoever to the story’s development are also necessary. More than that, to put it strongly, a film also needs “shots, sequences, and settings of uncertain meaning, used to estrange—to distance the story itself.” If every shot fulfilled its necessary and sufficient function as a component of the story, and the shots were arranged according to strict calculation, the result might be an ideal film for its director. But if every element of a film functions only to realize the story, the film ultimately remains no more than a “container for a story.”25
The same argument appears in Oshii’s Talking Head (1992). What the filmmaker asks is “not about film as a container for a story, but how to tell faithfully the story called film.” In Patlabor 2, the birds, as empty objects, are cracks embedded in the network of signs. Together with other estranging shots, sequences, and settings, they call upon the spectator to look. Because artistic experience lacks a transcendental signified—a foundational “reverse” behind the visual object’s “surface”—spectators seeking to release their desire to look can let their eyes slide across a meaningless optical surface. Only then can the act of “looking” take place. Yet anxiety before this meaninglessness, this peril of “the pleasure of doing nothing” = death, makes us seek by every means a ground for the object before us. We think it is something other than what it is, tame it, and integrate it into the story’s sign system. We thereby produce a “misreading”: we manufacture a “reverse” that is not there.26 The deframing in Figure 15(b) ought, by rights, to show something behind the figure. But there is nothing behind him; even an empty object such as the bird has been eliminated. Better to say there is only indifferent white and rigid straight lines: a centrifugal field of “nothing” urgently awaiting something to fill it. Or it has become a surface that prohibits any interpretable depth and seeks to capture looking—a virtual curtain that cannot be unmasked. It is what it is, yet it is nothing. The very word “surface” implies my inability in relation to an object that cannot be integrated. The frame’s cutting of the human model calls forth looking, yet the freedom of the eyes is directly bound to meaninglessness and uselessness—what Yoshida calls the anarchism of looking. This, too, is why deframing is “irony plus sexual perversion”: through an intention working against intention, it lays bare the fundamental lack that framing = centering tries to conceal. There is no ultimate guarantee of meaning.
Yoshida describes the experience of a filmmaker looking through the camera for the first time, the so-called “zero hour”:
This reality, this world seen through a movie camera for the first time, is probably far removed from the images that stimulate our imagination as spectators. It is something matter-of-fact, infinitely meaningless. The image captured by the camera merely shows reality as it is, things, people, or events as they are. No new image comes into being there. Can this matter-of-fact quality of the movie camera, this mechanism that shows reality as it is, really be called creation?27
What Yoshida first saw through the camera was just such a despair-inducing, disordered world: existing reality and its sheer contingency. It is therefore hardly surprising when he presents viewers with disordered, floating fragments of signifiers, partly detached from context and stripped of function. The centrifugal field continually “radiates” outward the symbols and meanings that take form within it, rather than tethering things to a center and structuring them around it. First, a repetition of Figure 15(b)’s theme.
Here a mirror frames a figure, separating them from the surroundings, but this fragment is itself placed in a corner of the image. Compared with centering the figure, this makes viewers keenly aware of meaning suspended, with nowhere to take hold. This is also, of course, the effect of the deframing in Figure 12(a): if the figures were centered, we would not feel the wooden carving’s gaze.
In Clips 18 and 19, we see another form of this “radiation.” The door is stripped of its function, becoming an opening that swallows figures and spits them out. This formalizing process itself becomes a floating signifier. Drawing on what lies behind the set—what Deleuze calls the “absolute out-of-field”—the opening connects with a dark stretch of duration, giving the image a brief pause, like a breath.
There is also what I call “casting forward”: the spectator can grasp meaning only retrospectively, sometimes after a long interval, and even then the effect is not clear-cut. Yoshida identifies such images, still ambiguous and suspended as we watch them, in Yasujirō Ozu’s work. When Tokyo first appears in Tokyo Story, for example, a sign reading “Hirayama Clinic” happens to be visible on a utility pole. But before we have been told the elderly couple’s surname, it is merely a meaningless sign.28 In the scene from Clip 18, Makiko learns the truth her mother has concealed: her father did not die. Furious, she cries out: “Now I understand perfectly. You’ve tied everything down, Mother. My whole life!” Only now can viewers retrospectively recognize the meaning of the mother’s shadow in the earlier scene, Clips 11 and 12. Later in the film, the musician Hayato Watari makes it even more explicit when confessing to Shōko: “Her mother’s shadow is always behind her.”
Figure 17(b) and (c) correspond to Clips 11 and 12; (a) comes from the same scene. When Makiko first comes home and steps inside, her mother, who greets her, is positioned behind her. As in (b), there is a shadow Makiko cannot see but which is already watching her. This is also the mother’s first appearance in the film. In (c), the scene’s final image, the mother conceals herself, shutting her shadow outside the door, in a design that echoes its beginning. Yet we rarely go back to look again while watching a film, especially at such visual elements, which seem not to have made much of an impression and may never have been noticed at all. On watching again, viewers discover that the mother’s shadow had already been watching them secretly the first time. Let me also give an example involving a mirror.
From an unusual angle, the glass bottle of medicated mouthwash held by Nose’s wife Kiriko (Kazuko Inano) is doubled in a mirror. Its significance only becomes apparent near the end, after a very long interval, when Kiriko, suspecting an affair between Nose and Aki, throws a glass bottle of concentrated sulfuric acid at the actress. For Kiriko, the mouthwash is a warning and a threat to Aki. For the spectator, its doubling is unquestionably an as-yet-unrecognized, ambiguous hint or omen.
When meaning is exhausted, the centrifugal field develops into an inert field—a field that cannot be activated, an absolute “nothing.” Any meaning dissipates in this void without being reflected back. This is the situation I call “annihilation.”
Just as Jissōji uses backlighting to reduce the information in an image, Yoshida achieves the same effect through overexposure. Its pallor becomes part of the visual beauty of Eros + Massacre and Heroic Purgatory. Compared with Jissōji’s, however, Yoshida’s camera generally stays farther from the figures. They occupy less of the image, surrounded by an enormous field. The field in Figure 19 possesses no religious or transcendent sanctity. It is utterly banal, exposing in its glare the futility and stupidity of the revolution enacted within it. This inert expansion of space, or expansion of absolute “nothing,” is also Yoshida’s chosen path through the contradiction between the centripetal movement that human involvement makes unavoidable in cinema and the centrifugal movement inherent in cinema’s nature. To film the world’s purposeless, disordered space, excluded by the story, is the result of an intention working against intention. He writes:
This is the foolish confusion of making a film while insisting that one is not making it. Yet by placing myself within this contradiction, might I release before the spectator, just as it is, the weight of the existence surrounding me? Perhaps one could say I wanted to wager on impossible love. And in response to that love, my space expanded without restraint, dispersed, and hollowed out.29
This “existence surrounding me” is the subject to which this section has repeatedly returned: the field. In cinema, it can be manipulated through mirrors and doors, by the various means described in the first two subsections. Through the radiation and casting forward of meaning, it becomes centrifugal and inert—the subject of this subsection. In the next section, I will explain how the field is woven from innumerable gazes.
III. The Gaze
What is the gaze? It is an object that invades the network of signifiers as a hole, a lack: an obstruction that disturbs the network.30 What is the field? It is an existence that has not yet taken form, drifting, escaping, impossible to capture, containing innumerable gazes. More simply, it is what Yoshida calls the “disordered world.”
In the film within the film, Aki Kaidō plays an adulterous married woman, and she also seems to be having an affair with Nose in reality. In fact, however, she fears men and sex because of the incident in which a teacher molested her. Her story in the film revolves around this traumatic past, the molestation that continues to haunt her and invades her life again.
In Figure 20(a), Aki is acting in a film. Her inner monologue says: “The camera is rolling now. Everyone is watching me. Eyes, countless eyes are watching me.” Yet neither camera nor eyes appear in the image. This ought to be a space upon which others’ sight lines converge, but those lines dissipate in the vast space and cannot be caught. It has already become a field filled with the gazes of things, without Aki noticing. In (b), she is at home. A “frame within a frame within a frame” surrounds her with an inorganic environment. This defamiliarization of “home” is another means of showing that she is “being looked at.” It is also worth pointing out that the shot comes immediately after the doorbell rings. The freelance photographer (Yoshio Harada) who enters after the bell, invading the safe space of Aki’s home and pursuing her traumatic past, is, for her, the gaze in human form. Figure 20(c) shows his first appearance. His defining trait is that he is always photographing and recording; the mirror’s disclosure of his secret photographing reveals, by reversal, the furtiveness of photography itself. In Aki’s nightmare he appears in another guise, an all-seeing figure: a stranger who nevertheless knows all her secrets, walking around her home and commenting on her hidden life. Her union with him at the film’s end is her way of maintaining the stability of her reality, suturing its breach. Paradoxically, it is through this that her distrust of men is healed.
Makiko Isaku takes her friend, the costume designer Nobu, to her hometown in the snow country. There she invents a past and performs for Nobu, wanting Nobu to understand that Makiko herself has no right to love anyone, so as to bring Nobu together with her own boyfriend, Karasawa. Figure 21(a)–(d) shows four consecutive shots. As in Figure 20(a), it is precisely the town’s surreal absence of other people that reveals the gaze of things, a gaze without a subject. The uncanniness does not come from the camera peering at the women from different angles, although it may seem to. It comes from the omnipresent world surrounding them, spying on them: things watch the characters from every direction.
In “What You Dare Not Ask Ozu, Never Ask Yoshishige Yoshida,”31 Fei Nei fiercely attacks Yoshida’s Ozu’s Anti-Cinema: “Yoshida’s writing in this book unmistakably exposes his grave, fatal weakness: he does not actually understand cinema.” The severity and absoluteness of this criticism compel me to defend Yoshida here. Addressing the absence of the inflatable travel pillow and of images of Tokyo discussed in the book’s first two sections, Fei Nei writes: “Yoshida thinks this is because Ozu did not believe cinema could faithfully reproduce what it filmed—which, of course, makes it ‘anti-cinema’!” This is unmistakably a misreading. First, Yoshida says no such thing. Second, this is not what Ozu’s “anti-cinema” consists in. In the book’s first section, Yoshida identifies the gaze concealed and revealed by the inflatable pillow at the beginning of Tokyo Story. Gazes are not on the characters’ side, but on the side of what surrounds them—things, absence, the other side—and there are countless numbers of them. In fact, Yoshida calls this practice of “looking at human beings from the side of things,” the reverse of ordinary cinema’s “looking at things from the side of human beings,” “anti-cinema.”32
In Figure 22, the ending of Tokyo Story, (c) and (d), almost repeats the compositions involving the inflatable pillow at the beginning, (a) and (b). The unnatural position of the old man in (c) makes us feel something strange and unsettling. Through this composition, “the absent space behind the old man is unquestionably filled with the gaze his dead wife casts from the other world.” The gaze of the inflatable pillow at the beginning, which “silently listens to the elderly couple’s conversation and knows more than we do about how events will unfold,” thus becomes the dead wife’s gaze from the other side at the end: the gaze of what is absent. This is Yoshida’s argument.
In fact, Fei Nei mentions Yoshida’s account of the pillow’s gaze several times in his book Seeking Cinema’s Gaps Between the Baroque and Zen. He writes:
At most this is an interesting idea. Examined closely, the claim does not hold. The inflatable pillow’s “gaze” can only be called a rather poetic interpretation, belonging entirely to the creator, and cannot be taken seriously. After all, the pillow is absent and appears only briefly in an altered form, compressed and folded. How can it become a thing’s viewpoint that guides our looking?
Yet, as though recognizing the power and convenience of this “poetic interpretation,” he says of a scene in Max Ophuls’s The Earrings of Madame de…:
The earring, temporarily absent here, performs essentially the same function as the inflatable pillow in Tokyo Story: a gaze belonging to an object, yet an “absent” gaze, quietly watching fate unfold.
This reveals an inconsistency in Fei Nei’s own account. Moreover, his objection fails to recognize the gaze’s further implication: it is an irrational element invading the space of our reality and disturbing the ordered structures of everyday life.33 For Yoshida, only when we notice the pillow’s gaze can we discover what is amiss beneath the elderly couple’s domestic conversation and utterly ordinary, unconscious responses. Only then do we perceive that this event has become a sign we have failed to grasp.
In Looking Awry, Žižek comments on Hitchcock’s continual alternation in Psycho between objective shots of Lila climbing the hill and her subjective view of the house:
Why does this formal procedure as such generate anxiety, why does the approaching object (the house) become “uncanny”? What we have here is precisely the above-mentioned dialectic of eye and gaze: the subject sees the house, but what provokes anxiety is the indefinable feeling that the house itself is somehow already gazing at her, gazing at her from a point that totally escapes her view and thus makes her utterly helpless.34
In a note to this passage, Žižek adds:
Since the gaze is on the side of the object, it cannot be subjectified: as soon as we attempt to do so, as soon as we try to add a subjective shot from the house itself, for example (the trembling camera looking at the approaching Lilah from behind the curtains), we fall to the level of the ordinary thriller, i.e., we would be concerned with the point of view of another subject, not with the gaze as object.
This prohibition on subjectifying the gaze follows the same logic as my examples in Figures 21(a)–(b) and 22: subjectification removes the uncanniness of the object’s gaze. Another of Yoshida’s methods—shot/reverse-shot between people and things that cannot be subjectified, or filming people past such things—also displays, in a certain sense, the character of a thriller.
While acting in a sex scene, Aki Kaidō says: “I’m being watched. Someone is watching me. … Their eyes are violating me!” The line’s exaggeration within the performance corresponds exactly to its truth outside it—perhaps there is no inside or outside. But that is not the only interesting thing. After the scene, as Aki chats with the director, the recording of this line suddenly starts playing by itself (Clip 20). What appears to be a film organized around actresses’ lives has its internal rationality collapse and disintegrate through this peculiar automatic playback. The words also correspond precisely to what the image shows at that moment: Aki being watched by someone else. But that is not the point either. The point is that both gaze and voice are on the object’s side, not the subject’s. I spoke these words. Their carrier is this machine, though no one knows who switched it on. Yet however I think about it, this voice is estranged from me. It came from my mouth, but it no longer belongs to me. It sounds utterly unfamiliar, cannot be properly understood, and begins to question my identity. Its strange separation from me reveals precisely my own division and noncoincidence with myself.
At the swimming pool where she made her debut as an actress, Shōko Ichimori discovers a plastic ball. She gives it as her reason for not wanting to appear in the next film: “In my dream it got stuck in my throat, and I couldn’t make a sound. The plastic ball really does exist—and it’s right here in this pool!” Figure 23(a) follows. This image erases Shōko’s viewing angle and with it the basis for the camera’s line of sight. The resulting image—anonymous and impersonal, as Yoshida calls it—hangs suspended, profoundly out of balance. The plastic ball is extremely hemmed in, yet possesses a powerful presence, floating discordantly and glaringly in this world. It is, in fact, a metaphor for an eyeball. What catches in Shōko’s throat is just such a dead eye, incapable of emitting a sight line. In (b), this “eyeball,” at an exaggerated scale, directly disrupts the image’s order. It watches Shōko and looks back at the spectator.
In Story of the Eye, Georges Bataille supplies a series of metaphors: eyes, eggs, the testicles of a bull killed in the ring, and an eyeball dug out of a corpse. These signifiers without a signified move through the narrative, interacting with and borrowing from another series of metaphors,35 forming the surface without a reverse discussed in Section II(C).
With the example of Figure 17 and the revelation of the mother’s shadow in mind, we can quickly discover that suggestive spherical objects continually appear in the image from Shōko’s very first entrance: Figure 24(a) and (b). Near the end, they appear when she parts from Dr. Hata, (c) and (d), and from her manager Minamikawa, (e). Or, to put it another way, the disordered field gives these spherical objects form. Their recurrence likewise makes up a network of metaphors, suspended and spread out.
End.