Mamoru Oshii, Author of Tachigui

Editor’s note: The author prefers to remain out of the public eye and authorized Tata-jun Minkun to post this essay on Zhihu on their behalf. Lab on Roof subsequently received permission to republish it in Bilibili’s article section.

“Ramen lovers are more like frustrated progressives, because ramen is full of the flavor of city streets.” Such was the curious insight that Shuji Terayama, a ramen enthusiast, offered into this humble food. Whether Terayama influenced Mamoru Oshii’s invention of the equally curious profession of tachiguishi—fast-food grifters who eat soba without paying—is, of course, impossible to know. Oshii himself has said that, as a child, he often saw old men and children standing by the roadside to eat; now this has become something people think deserves a scolding. Making a film about what happens in these standing-room eateries was simply a pleasure. We might believe this professed nostalgia coming from someone else. But once we have seen a standing-room soba shop through Oshii’s many layers of filters, it is hard to imagine that mere nostalgia could require going quite so far. In fact, consciously or otherwise, Oshii was already inserting grifters into works he contributed to in the early 1980s. By now, their appearances in “Duck films” seem almost as frequent as masked men in Gundam.¹ Clearly, anyone interested in Oshii cannot avoid the grifters. Seen another way, his willingness to release their long-gestating world in this “collected edition” has done his fans a favor. Tracking down and assembling the grifter puzzle across his many anime, manga, and films is no easy task. Oshii himself had been plotting Tachigui: The Amazing Lives of the Fast Food Grifters (Tachiguishi Retsuden, hereafter Tachigui) for more than a decade. When Innocence competed at Cannes, Mitsuhisa Ishikawa announced that, if it failed to win a prize, Oshii could make a film he wanted to make. At last, Oshii had his chance. Budget constraints, however, meant that his friends and relatives had to appear in it themselves—not even his daughter and son-in-law escaped. The film has long been regarded as “a fairly unimportant little diversion Oshii made in his spare time.”

Perhaps this explains why, more than ten years after its release, so few responses or interpretations have appeared. Admittedly, another factor is that, since the turn of the century, Oshii has felt himself reaching the limits of his own directing and creative system, and has gone years without offering anime fans and cinephiles another work on the order of Ghost in the Shell. Yet perhaps the obscurity of his cinematic language also confines what audiences recognize—or would prefer to see—to an Oshii proficient in film theory. The Oshii trying to articulate his personal views feels utterly unfamiliar. The obscure metaphors of Patlabor 2 leave us like blind men feeling an elephant; Jin-Roh, mediated through someone else’s hands, gives us the frustration of scratching an itch through a boot. Then again, Oshii is not formally trained, and his thinking may have little depth or originality. Numerous textual analyses have criticized him on precisely this point. What, then, makes Tachigui worth analyzing?

First, from the perspective of auteur criticism—the politique des auteurs—the answer should need little elaboration. Oshii’s remarkably continuous pursuit of reflexivity and film ontology can be glimpsed throughout his personal statements, interviews, and works. Fans also know that this pursuit extends beyond form and permeates content. Following earlier analytical models, I would venture that it is indeed connected to his fascination with the expression of ambiguity. Formally, ambiguity can dissolve the inevitable conflict over reality between art cinema’s realist and expressionist aesthetics. Yet Oshii’s ambiguity is generally understood to be less concerned with resolving this conflict than with his own wavering between the real and the fictional. For a fuller discussion, readers can turn to the extensive interpretations of his dream-labyrinth works. His interest in this problem has lasted for most of his career, which can only make us intensely curious about its reasons. One major question for Oshii auteur studies is therefore: “How can we uncover the truth, or the inside story, of Oshii’s ambiguity?” It is reasonable to hope for some answers from Tachigui, a work he himself takes very seriously. And objectively, its dedramatized text, with no corresponding loss of ambiguity, does lend itself to close analytical dissection.

Second, Oshii is a representative figure in Japanese animation and cinema. Research that moves beyond the auteurist perspective to incorporate social and historical circumstances places greater value on more explicit works of this kind. Oshii has important connections with numerous major figures in animation. He once said that part of the reason he and Yoshiyuki Tomino’s Gundam were not a good match was that “we explore similar concerns, but pursue different directions.” These “similar concerns” are by no means confined to the two of them; they recur among many prominent figures in the Japanese industry. This surely cannot be unrelated to their social environment. Consider a few names routinely discussed alongside Oshii: Hayao Miyazaki, Satoshi Kon, Katsuhiro Otomo, Toshio Suzuki, Mitsuhisa Ishikawa. Between them, they cover most of the industry’s map. Or we might approach the matter from the other direction: do the themes of writers Oshii often reads—Takaaki Yoshimoto, Yukio Mishima, Haruki Murakami—echo or answer his own concerns at the level of values, ideology, and so forth? This, too, would offer a route into the historical study of literature and cinema in “postwar Japan.”

Finally, we might simply take the film as an old geezer’s ramblings. As Oshii puts it: “I know it can be awkward for everyone when somebody keeps going on about himself. But opportunities like this don’t come often, and I wanted to get it across.” The film is no more than a stray dog howling in the distance. Still, someone who hears it will always choose to turn around.

I believe that statement is sincere. After completing Tachigui, Oshii underwent a major shift in both creative form and subject matter. The conventional account, which overlooks this film, places the turning point at either Innocence before it or The Sky Crawlers after it. As the endpoint of the less visible line running through Oshii’s “own works,” Tachigui naturally offers plenty to look at; the overt line is Cerberus. This essay takes the first of the approaches outlined above as its axis, attempting a small contribution to Chinese-language Oshii auteur studies. As for the second, let me first make clear that I have no intention of championing the importance of analyses built around “gossip” about personal relationships. Film history lies at the intersection of historical research and film analysis, and requires a more multidimensional perspective than I can offer. To avoid compounding errors by venturing into unfamiliar territory, I will stay away from questions of historical research, even saying as little as possible about the historical significance of the events to which the film refers. I ask readers’ indulgence.

Let us begin with this animation’s distinctive style—yes, animation. Figure 1 shows a shot from its trailer. Most readers may find nothing surprising here, but I suspect a few will be startled to hear the film called animation. In fact, quite a few auteurist accounts that mention it classify it as live action.

This is animation, not live action. To Oshii, who was already saying in a 2001 interview about Avalon that “all films will become animation,” my repeated insistence on the distinction might seem narrow-minded. But form matters enormously to him, and I do not think we can overemphasize its influence on the content of his works. This film is no exception. Its production technique, superlivemation, combines photographs of real people with postproduction, rather like filming a paper-puppet play. Anyone who has seen MiniPato should have little trouble accepting Tachigui as animation: MiniPato uses hand-drawn material, while this film uses photographs. That is essentially the difference.²

Enough narrow-mindedness. Oshii jokes that this film is certain to go down in history, because nobody else has done anything like it. As cinema, it does indeed break down the boundary between animation and live action. That boundary may seem clear-cut, but attempts to distinguish their essential natures have provoked fierce theoretical debate. There is no need to dwell on the old fallacy that “live action is more real and substantial, animation more fake and exaggerated.” Today’s more scientifically inclined consensus distinguishes the two through an ontology of motion: “actual motion” versus “apparent motion.” In practical terms, many filmmakers believe that live action generally contains more information, whereas animation, containing less, makes the information it presents more prominent and directs the viewer’s attention more effectively. Oshii has rejected this view. Having worked in both media, he may be more receptive—perhaps because of his failures in live action—to a deceptively simple proposition: animation adds; live action subtracts. Readers can doubtless find countless discussions of this idea. What I must reiterate is that it is not a methodology restricting what animation and live action can do; it emphasizes where creation begins. Animation builds its information from zero; live action builds from a hundred. Oshii recognized this long ago. There is no need to list the additions he has attempted in live action, some barely comprehensible to ordinary viewers, but he has also done a good deal of subtraction in animation. Figure 2 shows a “subtraction game” played through editing in Patlabor 2.

Subtitles, from upper right to lower right: “It was essentially designed as a closed system”; “Through a British university’s network, then infiltrating the U.S. military’s base system in Japan”; “The current system was completed in 1991 by updating the old one.”

It is easy to connect this “addition and subtraction” with the “breaking down of boundaries” mentioned earlier, and then ask: what is this boundary—the boundary between addition and subtraction? Suppose animation and live action are two people setting out from opposite ends of a street. Developments in CG have thoroughly demolished the old settled distinction between “reality” and “exaggeration,” while increasingly detailed 3D modeling has made the prejudice about “how much detail” each contains scarcely worth examining. Could further technological advances allow these two people to move freely wherever they wished along the street? If not, what threshold must they eventually reach? Oshii’s answer, which also provides a working answer to the question of how animation differs from live action, is this: the crucial distinction is whether the actor participates as a member of the production team or is used as material.

In other words, the distinction concerns characters, an important component of the vast majority of films. A friend once saw Tachigui’s cast list and joked to me: with all these “nonprofessional actors,” could Oshii’s film count as neorealism? Had he actually watched it, his jaw would surely have dropped. Oshii has subjected the photographs of these “actors” to exhaustive selection and restriction. Forget the spontaneous outpourings of feeling so common in neorealist cinema: even natural expressions have been excluded from the material finally used. “Grotesques in human form” would probably describe the finished result more aptly. Generally, in live-action cinema, an actor’s enactment involves committing everything she is as a real person to the mimetic process—what we call acting, “getting into character,” and so on. In fact, a trend toward leaving actors’ performances alone had already emerged in Hollywood in the 1950s and 1960s. Oshii deliberately goes the other way. His “material method,” selecting only parts of a real person, may be difficult to understand for viewers, and even some filmmakers, accustomed to classical narrative schemata or most art-cinema schemata. After all, creating characters who resemble real people—or even surpass them—is an ambition shared by live-action filmmakers and animators who construct characters through drawing or modeling. Most creators pride themselves on creating more fully developed characters, and critics praise their abilities by calling their characters lifelike: “Yoshiyuki Tomino is a master of characterization,” for example. Yet Oshii says: “Creating characters the way Tomino does is simply impossible for me.” And: “…I don’t understand Noa Izumi at all. I don’t understand what goes on in the head of a policewoman in her twenties who likes piloting robots…”

Oshii’s method may recall the familiar practice in film criticism and interpretation of treating characters as material. Critics reduce the material supplied by characters to certain features. Their feelings and actions may at first seem those of living people, but critics extract and refine them into a succession of propositions, then use these to interpret a theme or pursue an argument of their own. This interpretive method does have something in common with Oshii’s thinking. At bottom, however, its critical premise may arise from the critic’s wish, at some level, to make interpretive meaning independent of authorial intention. Unlike the critic, Oshii as a creator can address the problem at its roots.

Those roots lie in human cognition: “Humans are predisposed, biologically and culturally, to attend to humanlike agents in representations. The character-as-person schema seems obvious because it is ours; it is us.”³ As a basic mode of human cognition, the person schema naturally becomes indispensable to our understanding of moving images. This is precisely the schema Oshii tries to challenge. Taking the final scene of episode five of his 1989 experimental anime Gosenzo-sama Banbanzai! as an example, I will examine how he dismantles it from the creator’s side.

As Gosenzo is a meta-anime, its many incongruities are easily understood as meta elements. But matters are not so simple. In Figure 3, the heroine Maroko delivers her lines in two vocal registers. In the lower register, panel a, she says: “After all the trouble of taking this family story so far, we end it with this absurd turn of events. Is that really all right?” This cannot function as the character’s own dialogue. Viewers could certainly fall back on the meta elements that have run throughout the first five episodes—Oshii’s statements about how stories are made—and understand it as the familiar device of Oshii speaking through a character. In the higher register, panel b, she asks the man: “What for? Why on earth do I have to see Grandpa?” Lines in this register fit her established character. The division is clear enough to serve as a signal. But what does it signify? As another line tells us in panel c, the heroine is the one who “brings destruction and confusion to the Yomota family,” the protagonists. An ordinary family, living an ordinary, peaceful life, has thus been subjected to a sudden intrusion of “destruction and confusion”—dramatic force. What is dramatic force? Fundamentally, it is conflict. And the one who introduces conflict is, fundamentally, neither the enemy within the story nor the protagonist’s destiny, but the dramatist. This passage brilliantly punctures the “fictitious premise of reality” conventionally agreed upon by audience and dramatist. That, however, is not my point here. What I want to emphasize is Maroko’s status as material: structurally, she exists as “plot conflict” itself.

Beyond fitting a character to the person schema, then, Oshii tests the possibility of treating a character as inanimate material called “conflict.” Only “conflict” can make a remark such as “So magnificent a conflict, and it ends with a turn of events like this.” As though announcing this function, she also says, in the lower register in panel d: “Finally, an unanswered riddle remains. And now, the ending.” “An unanswered riddle” naturally refers to “suspense” at the level of story, or a “gap” at the level of narration; “ending” needs no explanation. These terms name material at a level prior to the finished film, bringing out the possibility of a character not constituted as a person. The final episode is entirely consistent with this. It breaks completely with the theatrical format of the preceding five episodes to recount the disappearance of conflict—the heroine’s absence—and the end of the story’s characters as material. The family wanders homeless because the author no longer needs the material of a finished story.

What interpretation cannot do, creation can. Readers will have noticed that the clear division described above announces a schema of material, not one of a character’s inner life, desires, or established image. The Maroko we first encounter is indeed Maroko as a person. But, with a touch of trickery, another Maroko parachutes in as material belonging to “story structure,” and her different manner of speaking clearly marks the boundary. Viewers may sense this only dimly, but they cannot simply force the two into a single identity because they share the same visible body—persona. This design thus overturns the person schema and establishes a method for treating characters as material.

In Gosenzo, Oshii gives a character “story structure,” separating it from her constitution as a person. In Innocence, the same problem appears in the purer doublet of doll—“shell”—and human. Figure 4 shows several frames presenting this pairing.

外見上は生きているように見えるものが本当に生きているのかどうかという疑惑。その逆に、生命のない事物がひょっとして生きているのではないかという疑惑。人形の不気味さはどこから来るのかといえばそれは、人形が人間の雛形であり、つまり、人間自身にほかならないからだ。人間が簡単な仕掛けと物質に還元されてしまうのではないかという恐怖。

Author’s Chinese gloss, translated:

Doubt as to whether something that looks alive really is alive; conversely, doubt as to whether a lifeless thing might actually be alive. Where does the uncanniness of dolls come from? From their being models of human beings—in other words, human beings themselves and nothing else. This is the fear that humans might be reduced to simple mechanisms or matter.

As this “soliloquy” makes plain, humans can—and only humans can—fear “things that reduce them to matter.” Characters are different. Characters considered as material, as dolls, do not feel that fear. What difference, then, is there between a character created by humans and a doll? The character designs in Gosenzo even have ball joints. The significance hardly needs spelling out.

Oshii’s thinking about “material” does not stop here. If this extraordinary device is one step toward the threshold, an alert viewer can glimpse its mirror-image counterpart in Innocence. Oshii elaborated on it in his 1992 film Talking Head:

人はより多く無防備で無意識に突き動かされそれと気づかぬところでイデオロギッシュな立場にたたされるからほかなりません……通常観客はペルソナと呼ばれる絵画的身体すなわち俳優によって演じられる具体的な役という新旧を通し主人公、敵役などの物語における役割や弁別可能で内面的人格をもつ個人としての新旧を了解するわけですが、アニメにあってはこのペルソナの実態をなす実在する俳優が存在せず……

Author’s Chinese gloss, translated:

Most of the time, people are unguarded and unaware when they find themselves presupposing an ideological position in something else… Through changes in the pictorial body called the persona—the concrete role played by an actor—viewers distinguish such story components as protagonist and antagonist, as well as changes in an identifiable individual’s inner personality. In animation, however, the actor who would exist as the actual substance of this persona is absent…⁴

(Oshii is referring here to body, character, and person all at once. It is hard to imagine him doing so without considerable thought.)

具体的な人物はもちろん動物だろうが植物だろうがモンスターだろうが宇宙人だろうがいや机だって椅子だってメロンパンだってただのぐっちゃぐちゃだって内面的人格と特定の物語の具体的役割を担う存在として創造することが可能なのです。

Author’s Chinese gloss, translated:

In animation, whether it is a person, an animal, a plant, a monster, or an alien—or, for that matter, a table, a chair, a pineapple bun, even an undifferentiated mess—you can create it as a being with a particular inner personality and a concrete role in the story.

Indeed, few people think that a character must be Homo sapiens: “The critic constructs the characters by means of a schema she also applies to real people (and animals, teeth, computers, and whatnot).”⁵ Take a fixed mahjong program, one not based on machine learning. It does not have good and bad days like a person, nor improve with experience: it does not “think.” Yet its human opponent will often imagine a rational adversary, even remarking when it wins with a high-scoring hand, “This computer really knows how to hold out for a big hand.” We take it for granted that animals or nonliving things such as Tigger and SpongeBob can be characters. More fundamentally, drawings—even the lines that make up drawings—are “material” endowed with life: animated. Oshii is considering the limits of that endowment, in both breadth and depth. In one scene in Talking Head, he experiments by creating “a character who keeps appearing yet has no visible form.” Having no visible form, the character has no body. Yet a character is established all the same.

In animation, Oshii attempts to dismantle the person schema; in live action, he attempts to establish an indeterminate body. The freedom of these choices seems to declare that, in his mind, there really is no boundary between animation and live action. Yet it is clear that neither Talking Head nor Gosenzo satisfied him. The device in Gosenzo still depends on how it is understood, while Talking Head itself has already been denounced as a “perverse work.” Its design lasts for only one passage and cannot be extended across the whole film, making that passage easier to understand as an “incongruous element” or a “meta element.” This design is actually used in Gosenzo too; readers should have little difficulty identifying the character concerned. If these two works touch the threshold from opposite sides, Tachigui aims to break straight through it: a work that dispenses with settled bodily identities and with characters constituted as persons. If the attempt succeeds, what viewers see should not be “a film whose information has been compressed to the absolute minimum,” taking subtraction to its limit, but “a work that, like a spring, permits large-scale addition and large-scale subtraction.” The avant-garde form this challenge produces naturally invites us to examine how Oshii’s distinctive character system is implemented in Tachigui, and what particular effects it achieves. Further, as the introduction suggested, we can ask whether this system bears some relation to his ambiguity of the real and the fictional.

Let me first point out something relatively obvious. The eight grifters in Tachigui have no distinctive temperaments or personalities of their own. They exist as features of the ideologies of successive periods in postwar Japan. As the film’s dialogue explains, the grifters are approaches taken by wanderers on the margins of the system toward “what belongs to the postwar,” using their own particular wills and methods. Oshii thus constructs a textual schema out of a kind of method, a current of thought. The individual instances belonging to it are more like pieces in a puzzle, completing characters organized by this schema rather than by the person schema.

This method may readily suggest the concept of the stock character. There is indeed a resemblance: each individual appears as an unsurprising example possessing the characteristics of its class. But the two should not be equated. First, stock types exist to a large extent independently of the text, determined by an audience’s shared cultural understanding. If people in a particular country knew nothing of “samurai,” or any comparable stock type, they would inevitably have some difficulty understanding the behavior of a character based on that model. The stock type, in other words, already exists before viewing, and the viewing experience depends on it. In this film, however, all these “approaches” are being proposed by Oshii for the first time. Real-life freeloaders and the grifters are probably worlds apart. Viewers can only analyze and understand the film on its own terms, constructing a model internal to this particular work—its “internal norms.” This model can only be acquired after viewing. Second, a stock type usually triggers information beyond what is directly indicated. If viewers identify a character as a “scoundrel,” for instance, they immediately entertain a series of possibilities: he may betray someone, inform on them, or come to a bad end. The ultimate effect of the stock type is to communicate certain character traits more quickly; it returns us, in the end, to the character as a person. Here, by contrast, once the internal model has been established, there is no further step from a class of figures to an individualized person. Their approach is directly identified, and the process stops there. Personhood is barely present throughout the process.

Nor do the film’s characters appear with settled bodily identities. Earlier analyses have mostly centered on specific descriptions of their personalities, appearances, or actions. I have to regard this as a misreading. I am not accusing those writers of carelessness. On the contrary, the more practiced the critic, the harder this film may be to accommodate. Audiences and filmmakers generally share a tacit agreement that a film’s text is coherent. At the level of a basic unit—the smallest passage that constitutes meaning—this coherence means that sound, perhaps commentary, explains the image or adds further information, while the image illustrates what the sound describes or sets up a charged contrast with it. Here, Oshii seems to break faith with us: a true image of the eight grifters never appears on screen. This naturally runs counter to ordinary viewing experience. Normally, the figure we see in a film is simply the character. To present a “beautiful woman,” for instance, filmmakers would probably cast a beautiful actor or produce a beautiful character design and model. Once that appearance is established, it does not change with the viewer’s subjective judgment. This film repeatedly emphasizes, however, that those we see are not the grifters themselves. They are figures handed down in writing or speech, as others remember them; the characters exist only in the commentary’s descriptions. It is rather like a novel describing someone merely as a “beautiful woman,” leaving each reader to supply a different beautiful image. Here, appearance, clothing, even names keep developing and changing through the parallel refrain “some say… others say…” shown in Figure 5. Viewers must therefore give up expecting sound and image to establish truth or correspondence, and enter a different viewing experience. These characters who “cannot be matched up” are indeed another version of “characters who keep appearing yet have no visible form.”

Now that we know how Oshii puts his character system into practice, the film’s structure becomes clearer. Let us approach it through genre—as a mockumentary—and examine how it constructs the real/fictional doublet.

Documentary reality derives from the physical properties of the machine. André Bazin praised the camera’s capacity for mechanical reproduction. This “integral realism,” produced by mechanical reproduction, had a profound influence on documentary, the form best placed to exploit it: for the first time, the apparatus had recorded facts without intervention or rewriting. Some filmmakers seem to overlook the presence of “fiction” even in classical documentary. Reenactments, though mocked as fakery akin to staged news photographs, have come into widespread use despite the controversy. Some regard them as a form of documentary creation; others as a way of emphasizing particular processes within real events. Some even advocate merging documentary with neorealist cinema. Amid this tendency, works beginning with The Sinking of the Lusitania, generally regarded as the earliest animated news film, also entered the debate over reality and fiction under the documentary banner. As an animated documentary itself, Tachigui undoubtedly enjoys an inherent generic advantage: the real and the fictional are already encompassed by the work as a whole.

To examine its narrative style more closely, we must naturally consider how Oshii’s distinctive character system produces the “mock.” For plot structure, I will use the passage illustrated in Figure 6 to approach the whole through a small example. Its dialogue runs:

In February of Showa 29 [1954], the Sharpe brothers fought the Rikidozan–Kimura team over three days, broadcast by NHK and Nippon Television, emblematic of professional wrestling’s golden age, then, on December 23 of Showa 33 [1958], the completion ceremony for Tokyo Tower, 333 meters high, then, on November 4, the first suicide by jumping from the tower, since its completion there had been rumors that “the tower is leaning,” instrumental measurements by the metropolitan building bureau established that this was untrue, incidentally Mothra would spin her cocoon on it a few years later, an unprecedented hula-hoop craze swept the streets, on November 13 of Showa 33 a mysterious thief calling himself “Satan’s Claw” was caught in the act of breaking into a house…

This sequence plainly lacks narrative connections altogether. To enter the vast debate over the nature of narrative connection would throw this essay completely off course, so I will consider only its function here. In an ordinary documentary-like narrative, fragments depicting “different events, in different orders, for different purposes” usually create the impression that they serve a common theme as a whole. Dispensing with clear temporal linearity, spatial unity, and logical relations produces dedramatization. There is no coherent, unified plotline or dramatic conflict because “that is how reality is.” In this passage, however, although the events are arranged chronologically, we really cannot supply narrative connections between them, much less identify a theme. The disorder within the passage of course works hand in hand with the noncorrespondence of sound and image discussed above. But it does more: by centering events rather than those who bring them about—being action-centered rather than agent-centered—it further develops the film’s context of doing away with characters. Technically, cutting continually between numerous events usually presents some difficulty. Here, the production method automatically resolves it, allowing scenes to change as freely as in a puppet show. The film makes fairly full use of its form’s distinctive capabilities.

The real/fictional doublet constructed by the film’s narration has its own distinctive flavor. As noted above, we cannot seek truth or correspondence between sound and image here. Consider the sequence about O-Gin in the second section, outlined in Table 1. C1–C4 follow a typical narrative pattern, forming question-and-answer beats that establish basic narrative connections. At C5, narrator A enters, introducing O-Gin and explaining her customary behavior. Viewers can readily understand this in terms of a conventional documentary pattern; this narrator has also been present before the scene. The question-and-answer pattern ends at C5: C6 does not answer C5, but jumps straight to the proprietor berating O-Gin for eating twelve bowls without paying. C7 still contains O-Gin’s diegetic dialogue. But the “chignon” in C9 provides a cue for narrator B to pick up the thread at C10. This is only a change of delivery; I call it narrator B for convenience. By C11, narrator B is describing the action entirely, even reporting the characters’ dialogue. Viewers may feel a slight incongruity. Sure enough, at C12 narrator A identifies the preceding account as an “anecdote” (everything just presented is fictional). C13 shows a fictitious book, The Demonic Vixen. Viewers may infer that the “anecdote” comes from it, though this is not disclosed. Three more fictitious works follow, indicating that this anecdote is merely one of several circulating stories about O-Gin. Incidentally, Oshii himself has created several O-Gin stories; readers may pursue that thought for themselves. At C17 the subject turns to “an account that has been criticized”—the fox hand game, kitsune-ken—telling viewers in advance that what follows is very likely untrue.⁶ Oshii nevertheless presents it visually, then supplies evidence against it at C25. We can readily agree that visible images generally make a deeper impression than verbal descriptions and feel more definite. Yet in this passage, the vivid images viewers have seen are turned into vague hearsay, then declared fictional. As though teasing the audience, the second account announces its fictionality in advance, only to supply images anyway.

English transcription of Table 1:

The untruth and noncorrespondence of sound and image take other forms as well. The Mothra passage mentioned earlier, for instance, uses deliberately fake shots (…no model of the atomic heat-ray cannon could be found, so a Markalite FAHP was substituted…).⁷ Sometimes the images and the commentary simply describe two different things. Reality and fiction fill the images, sounds, and events, perhaps exhausting a whole matrix of combinations of true and false. If Oshii “spares” his audience anything within this stylized narration, it is the ambiguity that so often makes viewers of his other works wonder where “they themselves” are situated. Here, he abandons that ambiguity and becomes a narrator following Kiichi Inukai, conducting a dialogue with the audience. Not only are personified characters absent; the personified camera—the viewpoint—is absent too. Viewers can therefore occupy something like the position of readers of an essay, “clearly” observing the “clear” real/fictional doublet articulated through Oshii’s distinctive character system.

I am therefore prepared to argue that Oshii’s character system and his ambiguity of the real and the fictional illuminate one another. Toshio Suzuki says of him: “Miyazaki moves audiences through emotionally rich protagonists. Oshii creates by observing his characters objectively, carrying expressionless performance through to the end.” Oshii’s position has probably moved far beyond observation. Once he has mastered the construction of characters with indeterminate bodies that are not constituted as persons, it is only natural for his world to lose its “expression” and approach inanimate objects or landscape. Having broken through this threshold, Oshii maintains that cinema permits unlimited “addition and subtraction,” whether in live action, still “distinguished from animation,” or in the animation that “everything will eventually become.” In Tachigui, even what he sees as the last “barrier”—the actor—has been transformed into material. Quite literally, if a character is supplied by an actor participating as a human member of the production team, a director who is himself a person cannot completely control that other person. But if the actor is material, even an unconscious tilt of the head or blink becomes impossible: everything lies within the director’s control. Everything in the film thus becomes truly and entirely the creator’s creation, while the result truly exists as mimetic illusion—flowers in a mirror, the moon on water. As this analysis has shown, most ambiguity asks viewers to interpret ambiguously; Oshii’s asks them to read the ambiguity itself with certainty. Breaking through the threshold, Oshii walks straight from the eastern end of the street to the western end, then, almost as a show of defiance, walks back.

After narrative style, we must naturally turn to story content. The grifters’ invented history clearly overlaps with real history, adding another dimension to the real/fictional pairing. Generally, as the introduction suggested, content-centered works tend to be far more bound to their region than form-centered works. Beyond the difficult questions of culture and history, we would enter the aesthetics of translation, which is not my concern here. The movement from intellectuals’ theoretical enlightenment to propaganda directed at the masses; sōkatsu and “self-rejection”; “beef,” “hamburgers,” and “Disneyland” as cultural invasion; even wandering in India—I do not intend to approach the film from these angles.⁸ To avoid lengthy discussions of minor points, I also ask to be excused from analyzing each story’s local details, such as the amusement park as an “illusory” product outside the “real” city. I want to take a broader view of the highly stylized symmetry constructed by the film’s story.

Let us begin with the many textual correspondences, such as the parallel sentences closing the accounts of Ginji and Sabu.

ある者は銀二は長身痩躯の老人にして常に漆黒の二重回しを纏うと伝え。そしてある者は軍帽の庇を深くして双眸を窺わせず、憲兵の外套を羽織る剛直のものなりと伝う、その素性についても神農を奉ずるもの世間士なりと言いあるものは山人の流れにして西行を為し境地に達するものなり言うものがあれば旧満鉄調査部の諜者なりとしたり顔にて答えるものあり。

Some say Ginji was a tall, slender old man, always clad in a jet-black Inverness coat. Others say he was an unbending man in a military policeman’s overcoat, his cap pulled so low that neither eye could be seen. As for his origins, some call him a worldly wanderer who worshipped Shennong; others say he was descended from mountain folk, wandered west, and attained spiritual understanding; still others knowingly reply that he was a spy from the former South Manchuria Railway Research Department.

ある者はサブは再びアジアを放浪し、某国で肝炎を患ってその地に没したと語り。またある者はカルカッタの路上で喰い歩きをしているサブを見たとも言う。そもそも中辛のサブなど存在しなかったあれは本物のインド人の亡霊だったのだと真面目な顔で語る者もいる。

Some say Sabu wandered through Asia again, contracted hepatitis in an unnamed country, and died there. Others say they saw him eating his way through the streets of Calcutta. Still others solemnly insist that Medium Hot Sabu never existed at all: he was the ghost of an actual Indian.

The symmetry is not confined to the text. First, it is easy to see that the film’s two halves have completely opposite styles. To put it plainly, the first half has that familiar “Duck-style” slog, the second those equally familiar “Duck-style” comic routines. Looking more closely, either at the content or at the commentary, we can readily identify four sections, each comprising two grifters: their legend, their self-destruction, their self-establishment, and their end. These four processes follow the classic “journey plot” pattern. Words such as “enlightenment,” “myth,” and “legend” all suggest that Tachigui tells the story of a journey undertaken by “postwar currents of thought.” This journey plot naturally entails symmetry in story form and descent in story content. Figure 7 summarizes these features.

English key to Figure 7:

To construct this symmetrical journey, Oshii could actually have taken greater liberties with Sabu, who by the end has already “lost his significance.” Instead, he deliberately constructs the textual oppositions described above. Jobs that “abandon language,” personal histories, wanderers returning home: these story elements indicate that the whole has been completed according to a more rigorous symmetry.

Most readers may expect me to use Figure 7 to compare the film with Herbert Quain’s April March and Statements.⁹ Space, however, makes this a good place to stop. I only want to point out that this myth-tinged, canonical journey plot, with form functioning as content, both adds a touch of mystery to the grifters and makes this “self-invented mythology” more persuasive.

With a work of such style and content, surely this must have been the animation Oshii most enjoyed making. Of course, it cannot have existed fully formed in his head from the start, nor should it be understood merely as the objective expression of a completed analysis. As far back as Urusei Yatsura, he already wanted to organize the grifters into a system, though the conception was still simple. Figure 8 shows the grifters who appeared then: Sabu, Ushigoro, Tetsu, and O-Gin. At the time they belonged to schools such as the die-hard stay-putters (inokori), the crushers (kurasshā), the imperials (teikoku), the wanderers (sasurai), and the actors (engi). After more than twenty years fermenting in Oshii’s mind, they have assumed an entirely different form.

The story forms an aesthetically pleasing symmetry. Yet, with the inevitability already foretold by its “journey” form, the grifters’ raison d’être steadily dwindles as the postwar years unfold. Sabu, whose significance has been lost, is an example. As an “urban legend,” he could presumably exist just as well in Heisei—even in the present Reiwa era—without anything changing. This accords with Oshii’s view that “the age from Heisei onward has lost its theme.” Thus, although the Economic White Paper had already appeared in 1956, Oshii’s—or Japan’s—“postwar” has still not ended.¹⁰ Strictly speaking, his ideal grifters are Ginji and O-Gin, whose work is “enlightenment,” or even “propaganda that falls short of enlightenment.” They are the grifters whose “jobs truly count.” As the descent continues, the tragic experiences of the earlier figures may give us pause; the later ones may only make us laugh.

“Inevitable descent,” “a sense of loss determined in advance,” or “a struggle—a war—undertaken despite knowing it is futile”: these are familiar themes in “postwar Japan,” and they constrain the subject matter of the great majority of Japanese creators. “Ginji’s Counterattack,” the film’s open ending without a definite answer, cannot believe in an unlimited future as Char’s Counterattack, Oshii’s favorite Gundam film, does. It can only call to mind the equally unresolved ending of The Sky Crawlers. Is Tachigui, then, really another work of ambiguous expression and an uncertain road ahead?

I believe we can find an answer by completing a puzzle across Oshii’s works. Notice how, at the level of story, Tachigui builds against the grifters’ earlier appearances. In Gosenzo, Inumaru is indeed the victim of fraud; here, the Inumarus who rally under the banner of self-rejection share a propensity for fabricating stories. In Patlabor, Masa of Mach-ken reforms after Ota “sets him straight,” giving up petty theft entirely. This is the exact reverse of the Masa who dies through sōkatsu. The terrorism-inflected Ushigoro and Tetsu were merely helpers hired by Shutaro Mendo in Urusei Yatsura. In The Red Spectacles, standing-room soba shops are regulated, and Moongaze Ginji can only get plain noodles in broth. Calling himself “Ginji the Nameless,” he works as an information broker. With these deliberate inversions, if we take the earlier works as illusory images, then on this side of the reflection stands, without doubt, a true image of Oshii himself. At the end of his pursuit of illusion, Oshii records himself in documentary form. This is his “doppelgänger announcing the truth about himself.”

“This counts as an inference drawn with certainty”: the line recurs throughout the film until it becomes a motif. Indeed, Oshii cannot find evidence. As the line about the yellow airship in Gosenzo has it—the yellow airship also serves as a sign of illusion in Patlabor 2—“Such a silver tongue! If you have no evidence, get out of here!” He can reach his conclusion only through feeling, without rational evidence or proof. He therefore has no choice but to accept, however unwillingly, “a sense of loss determined in advance.” Yet inwardly, his struggle is not “another war next time” (kondo no sensō), but “war continuing from now on” (kongo mo sensō). “Standing on the moon” will come in the end. Perhaps his persistence as a creator is the same. Shinji Higuchi marvels at the film and shudders at the deception of Oshii’s earlier works: “This is not a film by Oshii. It is Oshii himself.”

Looking back over this essay and its scattered ramblings, I worry terribly that readers will feel I have been talking only to myself. I also fear that familiar, intractable Oshii problem: “If you don’t understand what Oshii is saying, then even if someone tells you what Oshii is saying, you still won’t understand what Oshii is saying.” That is why I have repeatedly broken off just as a subject gets going. Still, I have offered readers everything I wanted to say about Tachigui. If I have held anything back, it is the statements of creative belief, expressed as dramatic dialogue, that recur throughout Oshii’s work and appear here too. These passages themselves seem by turns true and false. I feel I remain at the stage of “inference drawn with certainty”; I really have no way to substantiate them. Readers have surely felt this themselves while watching. So I hope you will allow me to call this essay a “guide to watching,” full of an amateur enthusiast’s passion. Its greatest distinction is probably that it is a 000 essay written in the year of 00 (laughs). Finally, permit me a friendly warning: if you cannot take Korean cold noodles, perhaps you would be better off not actually trying soba. Given the confusion translation can cause, I have favored fluency in the Chinese, and ask readers to look to the foreign-language wording supplied for precision. There are doubtless many errors, and corrections are welcome. For substantial Oshii criticism that is logically clear and genuinely insightful, please turn to essays such as “A Comparative Analysis of the Direction in The Battle for the Capital and Patlabor 2” and “The Addition of Anxiety.”

Further Reading

• Terayama, Shuji. 書を捨てよ、町へ出よう [Throw Away Your Books, Go Out into the Streets]. KADOKAWA/Kadokawa Shoten, 1975.
• Oshii, Mamoru. 立喰師、かく語りき。 [Thus Spoke the Fast Food Grifters]. Tokuma Shoten, 2006.
• Bordwell, David. Narration in the Fiction Film. University of Wisconsin Press, 1985.
• Bordwell, David. Making Meaning: Inference and Rhetoric in the Interpretation of Cinema. 1991.
• Oshii, Mamoru. すべての映画はアニメになる [All Films Will Become Animation]. Tokuma Shoten, 2004.
• アニメ監督 押井守 妄想を形にする 密着ドキュメント [Animation Director Mamoru Oshii: Giving Form to Fantasy—An Inside Documentary]. NHK, 2008.
• Bazin, André. What Is Cinema? Vol. I. Translated by Hugh Gray. Berkeley: University of California Press, 1967, pp. 16–40.
• Winston, Brian. Claiming the Real II: Documentary—Grierson and Beyond. Documentary Film, 2008.
• Borges, Jorge Luis. Ficciones. 1944.