Animation and Reality, or Isao Takahata's Realism as a Point of Reference

Prefatory note: This article was a prizewinning submission to the pre-festival and post-festival events of Lab on Roof’s Rags Drum 2020.

Words are flowing out like endless rain into a paper cup.

The question is not whether nature ‘really looks’ like these pictorial devices but whether pictures with such features suggest a reading in terms of natural objects. —E. H. Gombrich¹

Introduction

In his discussion of Tamako Love Story (2014), Ryota Fujitsu describes a lineage in Japanese animation devoted to the depiction of everyday life. Isao Takahata’s Heidi, Girl of the Alps (Alps no Shojo Heidi, 1974; hereafter Heidi) initiated and opened up this path, followed by Magical Emi, the Magic Star: Semishigure (1986) and Tamako Love Story. He writes: “Deliberately excluding dramatic incident and accumulating depictions of the everyday, the style known as ‘animation of everyday life’ had found a fully realized form in the progression from Heidi, Girl of the Alps to 3000 Leagues in Search of Mother (Haha o Tazunete Sanzenri, 1976; hereafter 3000 Leagues).” (1)

We can see in Heidi and 3000 Leagues Takahata’s attempt to present life calmly and objectively, in its ordinary, complete form, as it is. He proved that this approach could be realized in animation, and this set him apart from his contemporaries. For these two works of “everyday-life animation” to sustain themselves through the accumulated depiction of daily life, without relying on dramatic incident, they must have a grounding in reality. Viewers need to gain a sense of reality from the animation, and an unspoken understanding must develop between the work and its audience: “these characters really do lead lives.”

Takahata therefore adopted a cultural-anthropological perspective, designing the space in which the action takes place around “actual human life” by establishing (1) a social setting—food, clothing, housing, transport, education, working conditions, politics and economics, for example—and (2) a natural and geographical environment, including climate and temperature, visible light, mountains, rivers and seas, and ecosystems. (2) This is a broad program for working toward reality. How to construct animated space, and how characters are to act within it, are not merely questions of practical execution: they also involve a deeper question of reality. The way Takahata addresses these questions forms a central part of his realism.

In this article, we will take Takahata’s everyday-life animation as our starting point for a modest inquiry into the question of reality in animation. There are many complex issues to address, but let us begin by identifying two important questions, one methodological and the other concerning animation’s nature: (1) How does animation represent reality? (2) What makes it possible for animation to represent reality? As we will see, these are inseparable from another pair of questions: “(1) How does animation represent space and time? (2) What is the nature of the space and time that animation represents?”

In live-action cinema, the question of reality has been discussed extensively, in a considerable body of writing. Here we must mention André Bazin. His ontological proposition concerning the cinematic image—that “the image is identical with the photographed object in objective reality”—is foundational: one might say that an entire theory of cinema and method of criticism were built around it. More importantly, we can also see connections between Takahata and Bazin, working in animation and live action respectively, in their approaches to the reality of their media.

While making 3000 Leagues, Takahata was influenced by Italian neorealism, including Vittorio De Sica’s Bicycle Thieves (Ladri di biciclette, 1948), and portrayed social reality in harsh terms. Bazin was, beyond question, an ardent advocate of Italian neorealism. Toshio Suzuki recalls that Takahata often talked about Bazin’s What Is Cinema? (3); Takahata himself also observed that reading “The Virtues and Limitations of Montage,” collected in that book, confirms the importance of composition in depth, which “places opposing characters whose confrontation harbors a crisis within the same frame.” (4) It is therefore hard to imagine that Bazin’s film theory had no significant influence on Takahata.

Bazin’s theory advances three guiding principles concerning reality: (1) the reality of the subject represented; (2) the reality of space and time; and (3) the reality of narrative structure. The methodological realization of these principles is his theory of deep-focus cinematography. (5)

In “Respecting Real Time, Insisting on Acting-Led Direction: A Technical Study of Isao Takahata’s Direction” (6), Seiji Kano writes: “Takahata’s direction rests on three basic features: ‘a viewpoint of stationary observation based throughout on fixed-camera shots (Fix)’; ‘the frequent use of long takes lasting more than ten seconds’; and ‘meticulous scene planning and animation (acting) planning, which are prerequisites for the first two.’” The fixed camera is one source of Takahata’s restrained, calm and objective style. The long take, meanwhile, is an important component of Bazin’s theory of deep focus: compared with cutting up and reassembling a scene, it respects spatial continuity, temporal duration, and the actual connections between things.

Following Fujitsu’s account of the history of everyday-life animation, we find later directors along this path who also favor fixed-camera long takes, including Takashi Anno—episode 27 of Maison Ikkoku (1986), Magical Emi, the Magic Star: Semishigure, and Yokohama Shopping Log (Yokohama Kaidashi Kiko, 1998)—and Tomomi Mochizuki. Anno has a particular kind of long take in which he boldly dispenses with a musical score in favor of ambient sound. Slow changes in the landscape give visible form to passing time, while the characters’ inner states remain subtly implicit within these changes; the poetry of life emerges naturally over time. Mochizuki’s first episode of Twilight Q (1987) consists entirely of fixed-camera shots; episode 2 of Seraphim Call (1999) goes so far as to use only a single surveillance-camera shot throughout. Kimagure Orange Road: I Want to Return to That Day (1988), meanwhile, deliberately uses protracted, oppressive shots to convey youthful uncertainty and agitation, and the suffocation of a love triangle.

Another director along this path is Naohito Takahashi (To Heart, 1999). His directorial lineage goes back to Osamu Dezaki, and although his methods may differ from those of the others, the thinking is consistent: to depict everyday life in which “nothing has happened, yet something seems to have changed” (1), and to attend to its minute details.

Outside this lineage, directors such as Mamoru Oshii and Yoshiyuki Tomino also say that Takahata had an enormous influence on them. Both attach considerable importance to location research and to conveying a sense of everyday life; Oshii, in particular, also frequently makes a stylistic device of the fixed-camera long take.

Exploring Takahata’s influence and comparing these directors’ styles would be another complex undertaking. In fact, Takahata’s realism is a mountain that no history of twentieth-century animation can get around. The evolution of his work might even be seen as a “shifting axis of reference” in animation history. Like Dezaki’s, his lineage has woven an immense net extending over the animation world.

To return to our earlier discussion: although live action—Bazin’s film theory, for example, or the fixed-camera long take—can offer animation points of methodological reference, we can expect animation and live action to differ in a whole series of expressive means and methods of realization, since their natures are necessarily different. This bears on the questions posed earlier. In what follows, we will undertake a modest inquiry into reality in animation, principally by describing and analyzing these differences. We will divide the discussion roughly into three aspects: shots and space, movement and time, and reality and symbols. These are not wholly separate from one another. The third is the core of this article, and we may need to bring in a little knowledge of painting.

Shots and Space

We should first note that it was on Heidi that Takahata and Hayao Miyazaki first formally introduced layout into the animation production process. At the technical level, a layout—the plan for a shot’s space and staging—has the most basic purpose of matching the characters’ forms and movements to the space indicated by the background. Layouts in the Mushi Production lineage emphasize graphic design, concepts and impressions, as in Dezaki’s Aim for the Ace! theatrical film (1979) and Haguregumo (1982), whose layouts were by Yoshiaki Kawajiri. By comparison, layouts in the Toei and A Production lineages—those of Miyazaki and Tsutomu Shibayama—give greater weight to spatial representation, camera position and lens characteristics.

Miyazaki handled the layouts, or scene designs, for both Heidi and 3000 Leagues, providing the technical basis for spatial coherence. His use of telephoto and standard-lens perspectives also laid an extremely important foundation for the subsequent history of animation layout.

In live-action photography with a focal point, the longer the focal length, the shallower the depth of field. To present space on multiple planes, Bazin’s theory therefore relies principally on wide-angle, deep-focus shots to establish the integrity of space and time. In cel animation, however, perspective is fixed on a flat surface, which is then photographed through a lens, generally with everything in focus. Depth of field is thus not a consideration to begin with, making possible long-focal-length, telephoto perspectives in which foreground, middle ground and background are all sharply defined, without the problem of insufficient depth of field.

Takahata made extensive use of telephoto perspectives in Heidi and 3000 Leagues. Compared with wide-angle perspectives, they (1) more closely resemble the image generated by the brain’s processing of what a person sees, and so are more comfortable to look at; (2) convey a stronger sense of space and atmosphere through their pronounced compression of depth; and (3) are more objective and stable, as opposed to the immediacy and unease associated with wide-angle views.

Unlike real space, moreover, animated space can combine several different kinds of space. Spatial design in animation often follows the principle that “how it looks” takes precedence over “how it ought to work.” For instance, different lens perspectives can be used within a single image. Framing near and distant objects differently—a wide-angle view nearby, a telephoto view in the distance—allows judicious exaggeration to give the image greater force. Or, in a one-point perspective drawing, one need not be bound by the position of the vanishing point: one can give precedence to one’s own feeling rather than to the correct vanishing point. An image can also contain several viewpoints, guiding the viewer’s gaze.

Compared with live action’s “mechanical reproduction” of physical objects, animated space has greater plasticity, and can also better accord with the aesthetics of human vision. In his production notes for 3000 Leagues, Miyazaki wrote, “I design the layouts according to the feeling of the human eye.” Precisely because his layouts are constructed on the basis of visual cognition and aesthetics, rather than strict perspective drawings—which more closely resemble live-action shots—they may give viewers a more immediate sense of reality at the level of feeling.

In a certain sense, this capacity of animation to represent space, as distinct from live action, might be called “plasmaticity.” Sergei Eisenstein first proposed this concept, describing it as follows:

“The rejection of the constraint of form, fixed once and for all, freedom from ossification, an ability to take on any form dynamically. An ability which I would call ‘plasmaticity,’ for here a being, represented in a drawing, a being of a given form, a being that has achieved a particular appearance, behaves itself like primordial protoplasm, not yet having a stable form, but capable of taking on any and all forms of animal life on the ladder of evolution.” (7)²

“Plasmaticity” is generally understood here as the capacity of line drawings to move freely, something very close to the concept of plasticity. But animation scholar Nobuaki Doi argues that it refers not to changes at the visual level, but to changes that occur, abstractly, in the viewer’s consciousness: the capacity to set formless, abstract “metaphors”—in Yuri Norstein’s sense of the word—in flux. A “metaphor” is not the drawn figure itself, in its materiality, but the fluid images it evokes, in their abstraction. Freed from physical constraints, animation uses plasmaticity to invent an ambiguous, pluralized reality within consciousness, moving among multiple realities and making viewers aware that their own relationship with reality is only one of countless possibilities. Plasmaticity does not mean leaving reality behind; it unsettles reality itself at its foundations. When animation convincingly presents a reality whose nature has changed, live-action images, by contrast, seem like lies and fictions because they are too homogeneous. To take this further, animation becomes an “intimate” reality, live action a “distant” one. (8)(9)

In this sense, Takahata and Miyazaki effectively use “metaphorized” animated space as a method oriented toward reality. Here, “metaphor,” grounded in bodily sensation and in the way our bodies apprehend the world, intimates another “reality”: one closer than live-action images to our physiological instincts, closer to our bodies, and one we may even find more intimate.

We will discuss how this sense of reality works in greater detail in the fourth section, “Reality and Symbols.”

To add to the earlier discussion of animated space: painting, too, has a way of working with multiple viewpoints within a single image—what is called scattered-point perspective. It is, in fact, quite common. Édouard Manet’s famous The Luncheon on the Grass (The Picnic, 1863), for instance, contains several different perspective relationships within one painting.

But we perceive animation very differently from a still painting. In animation, characters act within the space, and the camera begins to move in and out, pan and track. This capacity for movement and narrative is precisely animation’s advantage over still images, and the reason we can sense “life” in it. As Doi puts it, “A still image has no need for the magic that makes life visible. The magic must happen in the dimension of the story, in the circumstances of the characters.” (8)

A representative example of movement in a “multiple-viewpoint” image occurs in the opening sequence Yoshinori Kanada animated for Galaxy Cyclone Braiger (Ginga Senpu Braiger, 1981). A single image contains three viewpoints—looking down, looking straight ahead and looking up—corresponding to three characters. Spatial perspective is distorted, and the image begins to move according to that distorted perspective. Ryusuke Hikawa comments: “Normally, a layout like this could never fit into a single drawing. It’s amazing that he could draw it so matter-of-factly. And he didn’t just draw it—he made it move according to that distorted perspective. That feeling of a distorted image rotating like this is something unique to Kanada, something no ordinary imagination could ever draw.” (10)

Like Miyazaki’s layouts, Kanada’s—especially his work on Invincible Super Man Zambot 3 (1977)—had a profound influence on animation history.

Movement and Time

In “The Virtues and Limitations of Montage,” Bazin writes: “When the essential content of an event requires the simultaneous presence of two or more elements of the action, montage must be prohibited. The right to use montage is restored once the meaning of the action no longer depends on physical proximity, even an implied proximity.” (5) This sets the limits within which montage is prohibited, defining the area in which the theory of deep focus applies: different actions unfolding simultaneously on several planes of space within a shot.

In hand-drawn animation, however, and in television animation broadcast weekly, shots in which “different actions unfold simultaneously” become particularly difficult. Moreover, the longer a single shot lasts, the more readily viewers focus on the moving characters’ acting, and the greater the challenge to the animators. In the 1960s and 1970s especially, shortages of money, time and staff were acute. Animation thus naturally tends toward using montage to represent “dramatic time.”

Bazin writes: “Breaking a scene down into several shots and assembling them through montage amounts to a kind of temporal expressionism, rearranging events according to an artificial, abstract time: dramatic time. Neorealism challenged almost all of these general conventions of cinema as spectacle.” (5) In animation direction, dramatic time manifests a discontinuity, perhaps also implying a “heterogeneity of mental activity.”

By omitting, abstracting, simplifying, magnifying or cutting short certain movements, and by using a wide range of camera techniques, Mushi Production sought, “when an object moves, to reduce its movement as far as possible according to whether movement is necessary, while making it feel as dynamic as possible as animation.” (11) These theories of animation and direction are collectively known as limited-animation techniques. In Heidi and 3000 Leagues, by contrast, Takahata tended to exclude dramatic construction, showing movements through their full course and allowing the time they take to unfold without abbreviation. This secured temporal continuity, while acting conveyed the characters’ subtleties. It certainly did not mean excluding limited-animation techniques altogether; rather, he took a stance that combined opposition with compromise and integration.

Take the number of drawings used in television animation. An episode of Heidi averaged 5,000–6,000 drawings, two or three times the number in an ordinary television episode. (12) Mushi Production, meanwhile, experimented with low drawing counts in Sabu & Ichi’s Detective Tales (Sabu to Ichi Torimono Hikae, 1968): episode 9, handled by Moribi Murano, used roughly 1,200 drawings, and episode 14 only 800. (13)

Collaboration with top animators such as Miyazaki and Yoichi Kotabe was the key to maintaining consistent animation quality in a long series with high drawing counts. Later, the Takahata–Miyazaki–Kotabe trio broke up. For Anne of Green Gables (Akage no Anne, 1979; hereafter Anne), Yoshifumi Kondo, recommended by Miyazaki, served as animation director for the entire series. With astonishing energy, he maintained control over the animation’s quality and achieved exceptionally delicate character acting. (It is worth noting that Miyazaki resigned from scene composition after episode 15; Michiyo Sakurai took over at episode 18. Production ran into difficulties for a time, but miraculously regained its brilliance under conditions stretched to the limit.)

Let us take another specific scene. At the end of episode 1 of 3000 Leagues, Marco’s parting from his mother includes a shot lasting more than a minute. In its first half, his mother says goodbye, but Marco does not make a sound. After she leaves, he remains silent, head bowed, for almost half a minute. Once the crowd has moved away and left him alone, Takahata makes an exquisitely judged slow track-in (T.U) to enter the character’s inner world. Alone, Marco hears the voices of everyone saying goodbye. The shot’s duration and the track-in allow Marco’s emotions to build, giving extraordinary force to the tears in Marco’s eyes in the brief close-up (UP) that follows. His emotions bursting out, Marco runs. He falls and gets up, falls and gets up again. The character is drawn with simple lines and almost no shading, and his movements are not showy—indeed, they are quite plain—yet viewers can feel the sadness welling up within his body. Toshiyuki Inoue suspects that Miyazaki handled the entire passage himself, from layout to key animation. He regards it as direction that strikes straight at the heart in a way live action cannot; Marco’s falls made him feel pain too. (14) In this scene we see Takahata’s respect for real time, his insistence on movements being shown in full, and the need for formidable animators to support the layouts and action scenes.

On the other hand, “action” is not actually the core of Bazin’s phrase “two or more elements of the action.” What matters is the relationship, or potential relationship, between two or more elements. Their interrelation is conveyed through their unity within a shared space and emphasized through the representation of time.

Discussing the Eskimo’s wait for a seal in Robert Flaherty’s Nanook of the North (1922), Bazin writes in “The Evolution of the Language of Cinema”: “[T]he length of the hunt is the very substance of the image, its true object.” (5)³ The moving hunter and the motionless hole in the ice appear in the same shot, both securing a strong spatial unity and expressing the tension of the relationship during the hunter’s wait.

For animation, static compositions that emphasize actual or potential relationships are especially important. They not only save drawings but also increase expressive power. Yet without movement to support them, background-only shots (BG only) and held drawings (tome-e) have difficulty expressing passing time, because animation lacks the depiction of subtle changes in light, shadow and nature. Take the roughly one-minute held drawing near the end of episode 24 of Neon Genesis Evangelion (1995; hereafter Eva): even though the background music continues, viewers still feel that time is standing still rather than passing.

How to represent time when the characters are motionless thus naturally becomes a major problem for animation.

Before Eva, both episode 4 of Anne and Angel’s Egg (Tenshi no Tamago, 1985) had experimented with such shots. They share several features: a fixed camera; a duration longer than that of an ordinary shot—about twenty seconds for a single shot in Anne, and an astonishing two minutes or more in Angel’s Egg; a tense relationship of confrontation or potential confrontation between multiple characters; and little or no character movement.

Unlike Eva, however, Anne gives visible form to passing time through the swaying of grass in the foreground and the passage of dark clouds, while Angel’s Egg does so through flames that flicker and go out, and the resulting changes of light and darkness.

In the pairings Bazin–Takahata and Tarkovsky–Oshii, we may find the roots of the common ground between Takahata’s and Oshii’s images. Oshii has also acknowledged directly, “I wanted to try making a film like Stalker (1979).” (15) It is worth noting that Oshii made further experiments with animation’s representation of time in later works: the many precisely drawn in-betweens added to Batou in the grocery-store scene in Ghost in the Shell 2: Innocence (2004), for example, or Suito Kusanagi’s “motionless movement”—or, rather, involuntary micromovements—in The Sky Crawlers (2008). These methods all make time pass slowly and heavily.

To return to the original subject: Hideaki Anno, Takahata and Oshii are all attempting to relate the image to a stretch of time detached from character movement. This is not the same as animation’s use of separate layers or compositing merely to save drawings by moving only a character’s mouth or making the eyes blink. The directorial aim is to have the same stretch of time elapse for viewers and characters, while intensifying viewers’ experience of time. The viewer’s state of “waiting” also allows both identification with, and a contrast to, the characters’ inner states.

In these shots, the time in Eva and Angel’s Egg also builds anticipation for the story’s subsequent development, preparing for the characters’ ensuing actions. They can be regarded as a kind of deliberately prolonged sensory-motor situation. The scene containing the shot in Anne, by contrast, comes closer to Gilles Deleuze’s time-image: movement is no longer the principal force driving the images forward.

Anne rushes out of the carriage and sits on a collapsed fence some distance away, followed by a long silence… Then she gets up and returns wordlessly to the carriage. Here, “the sensory-motor circuit breaks down and ceases to function.” Anne’s act of withdrawing becomes an element that produces a break: “this new element will prevent perception from extending into action, so as to bring perception into contact with thought.” The sensory-motor situation thus becomes a purely optical and sound situation. Anne sits still; Marilla sits still. Both are at a loss, unable to act: “No foreseeable, determinate action is about to occur; there is only the arrival of an event improvised in the present… Latent emotions and mental life become manifest through the powerlessness and lack of action.” (16)(17)

Let us stay with Anne to discuss the overall organization of time in Takahata’s work. Anne followed Heidi and 3000 Leagues, the works in which the style of everyday-life animation was said to have “found a fully realized form.” Astonishingly, however, its very first episode immediately began to explore further possibilities in directorial method. It could even be said to have been still more pioneering than those earlier works, and it became a celebrated episode in Japanese animation history.

Film researcher Seiji Kano wrote on Twitter that episode 1 of Anne “breaks the rules from beginning to end: its slow pace faithful to actual time, long takes and fixed-camera shots, excessive dialogue, live-commentary-style male narration, objectivism, and so on.” (18) Oshii likewise drew attention to Takahata’s use of time. He explained that the first episode had greatly influenced his own animation work. Put simply, something that takes thirty minutes in real life is depicted in thirty minutes of animation. Could you really do that on television? The thought gave him encouragement. (19)

“Something that takes thirty minutes in real life is depicted in thirty minutes of animation” is something of an exaggeration, but it captures Takahata’s pioneering direction with considerable economy. We see that the first episode consists of several very long passages, each essentially a “live record” of time, with no truncation or compression. The time that elapses for viewers thus becomes the same as the time that elapses for the characters.

Matthew arrives at the station to collect Anne eight minutes into the episode; at its end, she is still riding in his carriage on the way back to Green Gables. This is a documentary style pushed to its limit. Amid the tension built into the script, Anne’s torrent of childishly extravagant talk, and the beautiful scenery of the expansive countryside, viewers have an utterly distinctive experience of time: a single episode feels both long and brief.

In fact, the series takes its entire first five episodes to recount these two consecutive days in Anne’s life. Episode 4, from beginning to end, likewise consists largely of Anne and Marilla talking in the carriage. They set off at the end of episode 3, and by the end of episode 4 still have not reached Mrs. Spencer’s house. Takahata lavishes attention on small details, striving to realize the third of Bazin’s guiding principles concerning reality: the reality of narrative structure. In this way, he respects the completeness of events.

Oshii later spoke once more of Takahata’s influence on him:

“I can’t begin to say how much I learned from Anne of Green Gables, and Chie the Brat (Jarinko Chie, 1981) taught me a great deal too. I’ve watched it more times than I can count. As a director, I don’t think any other work has been as useful to me, and everyone in the industry says the same. I have to keep saying this: Takahata at that time was a director of the very highest order… He was one of the rare people who could use the temporal axis properly in his direction while thinking of the film as a whole. I really did study his work for all I was worth.” (15)

Analyzing Oshii’s work with Takahata as a point of reference would be another complex but fascinating subject.

Reality and Symbols

In “The Ontology of the Photographic Image,” Bazin writes of the invention of perspective: “The artist was now in a position to create the illusion of three-dimensional space within which things appeared to exist as our eyes in reality see them… The fact that a human hand intervened cast a shadow of doubt over the image. Again, the essential factor in the transition from the baroque to photography… lie[s] in a psychological fact, to wit, in completely satisfying our appetite for illusion by a mechanical reproduction in the making of which man plays no part.” (5)³

This is the key to Bazin’s ontology of the cinematic image, the proposition that “the image is identical with the photographed object in objective reality”: “a mechanical reproduction in the making of which man plays no part.” The accomplishment of photographic technology thus surpasses once and for all the likeness achieved by painting and realist painting’s pursuit of likeness; photography has “freed the plastic arts from their obsession with likeness.”³ Painting consequently turns from imitation to representation, toward what Nelson Goodman calls “denotation,” which requires no similarity to its real-world counterpart.

Then animation’s realism—its pursuit of a sense of reality—must never be a pursuit of likeness, if we approach animation as an art, even where its reality-oriented methods of direction resemble those of live action.

Painting is, first of all, a combination—an act of “addition”—of human-made visual elements: forms, including concrete figures and geometrical shapes; color effects, including tonality, shading and transparency; and spatial arrangements, whether flat or endowed with depth through perspective. We might call these visual elements symbols.

Animation adds a temporal dimension to the static image. Obake—distorted in-between drawings, or smears, used to express fast movement—and changing speed lines, for example, can all be considered kinds of symbols. So can timing, the speed and rhythm of an individual action; spacing, the change in an object’s position from frame to frame within that action; and even the exposure sheets written by key animators. Just as a perspective drawing creates an “illusion of three-dimensional space,” timing and spacing create an “illusion” of real time. By allocating a certain number of frames to each drawing, and through changes in an object’s position and the connections between frames, time is born and becomes freely manipulable. Norman McLaren even defined animation directly in terms of this characteristic: “What happens between each frame is much more important than what exists on each frame. Animation is therefore the art of manipulating the invisible interstices that lie between frames.” (20)⁴

Addressing these symbolic features of animation, Tatsuyuki Tanaka writes: “Isn’t an attachment to, a taste for, ‘symbolic’ expression itself—a fetish for ‘a particular mode of expression’ in its own right—the very essence of drawing in manga and animation?” (21) The wonder of animation lies not in imitating reality, but in the power of representation and the freedom of plasmatic transformation latent in the symbols themselves. Through the body’s apprehension, articulation and imagining of the real world—at which point symbols naturally possess a latent “bodily character”—another world of illusion is brought into being. Here we need a modest analysis of how viewers “see” symbols. Perhaps we could say that, within the illusion, viewers undergo what Bergson calls the two-way process of memory and perception. Below is his famous “memory cone”:

One key to this diagram is that “the present instant always lies at the intersection of two opposing movements: perception continually falling into memory, and memory continually becoming perception.” (16) Within animation’s illusion, viewers “remember” experiences of the real world. Yet in watching animation—fluttering cloth or hair, explosions, walking, running, and so forth—they also apprehend that world anew.

To make the analysis more concrete, let us take a single element of drawing: internal and external contour lines. Another reason for choosing this example is that it will be central to our later discussion of Takahata’s work. Things in reality have no lines. A drawn contour merely outlines the edge of a “surface” in reality, yet precisely these lines capable of “outlining” reality become one of drawing’s principal means of stirring memory and imagination. Whether simple or complex, contours with a strong representational power can quickly make us “remember” corresponding experiences of reality. Summoned by the “perception of contours,” our memories descend from the virtual plane of memory $AB$ to the apex $S$, turning into present perception. Or, to put it another way, this is the psychological mechanism of projection: we call on our memories and project them onto the image before us.

At the same time, contour lines inherently possess multiple meanings. Drawings exist as symbols freed from the constraints of reality, and can in themselves freely stimulate viewers’ imagination and thought, or set “metaphors” in flux. The plasmaticity of symbols is implicit here. Take the famous duck–rabbit figure:

“The point of this figure is that it can be viewed under more than one aspect: the same drawing can be seen as a duck and as a rabbit. Though the aspect changes, what is being looked at does not; the same drawing is a duck one moment and a rabbit the next.” (23) This sort of illusion continually calls on our capacity for projection. We keep recalling experiences of reality and proposing different interpretations. In Ludwig Wittgenstein’s vocabulary, it activates our capacity for “aspect-seeing”: our ability to see something “as” something. This ability of “seeing-as” requires our imagination, but also our culture, knowledge and symbolic conventions—or the virtual plane $AB$. The duck–rabbit also reminds us that behind the convincing illusion we see are symbols—shapes, lines, shadows, colors—with multiple potential meanings. From these symbols, we recognize the various images the artist wishes to represent.

We should note that the representational power of a contour does not mean accuracy. As the “Shots and Space” section pointed out, concern for visual reality can lead one to depart from the rules of linear perspective and use scattered-point perspective according to the feeling of the human eye. This changes objects’ contours, stretching or distorting them. A contour that feels real need not be “accurate,” nor should there be a standard for it. Gombrich comments on Paul Cézanne’s painting: “In all his struggles and explorations to achieve a sense of depth without sacrificing the brilliance of color, and an ordered arrangement without sacrificing the sense of depth, there was one thing he was prepared to sacrifice if necessary: the conventional ‘correctness’ of contours.” (24)

In the painting above, we find that Cézanne has distorted perspective relationships—the rim of the vessel, for example. He also strengthens the objects’ clarity with contours that appear and disappear, establishing their forms amid unstable air and light. This is the pursuit of “nature’s solid and enduring forms” that the early Impressionists neglected. In Cézanne: A Study of His Development, Roger Fry writes:

“For the pure Impressionists this question of the contour was not so insistent. Preoccupied as they were by the continuity of the visual weft, contour had no special meaning for them… But for Cézanne, with his intellectual vigour, his passion for lucid articulation and solid construction, it became an obsession… The contour is continually being lost and then recovered again. The pertinacity and anxiety with which he thus seeks to conciliate the firmness of the contour and its recession from the eye is very remarkable. It naturally lends a certain heaviness, almost a clumsiness, to the effect; but it ends by giving to the forms that impressive solidity and weight which we have noticed… At first sight the volumes and contours declare themselves boldly to the eye. They are of a surprising simplicity, and are clearly apprehended. But the more one looks the more they elude any precise definition. The apparent continuity of the contour is illusory, for it changes in quality throughout each particle of its length. There is no uniformity in the tracing of the smallest curve… We thus get at once the notion of extreme simplicity in the general result and of infinite variety in every part… In spite of the austerity of the forms, all is vibration and movement.”⁵

We can thus discern the latent bodily character of symbols. Contour lines are nothing other than the manifestation of an inner bodily sensation, as in Cézanne’s own conception: “Nature speaks within the painter and fulfills itself there.” The image of the body concentrated at the apex $S$ is the starting point of all perception and the center of action. Cézanne sensed both the solidity and order of forms and nature’s infinite variety. Through contour lines, clarity and infinity are conveyed together. Deleuze’s understanding of the “body” in painting is more radical still. He writes:

“[A]t one and the same time I become in the sensation and something happens through the sensation… it is the same body which, being both subject and object, gives and receives the sensation. As a spectator, I experience the sensation only by entering the painting, by reaching the unity of the sensing and the sensed. This was Cezanne’s lesson against the Impressionists: sensation is not in the ‘free’ or disembodied play of light and color (impressions); on the contrary, it is in the body, even the body of an apple. Color is in the body, sensation is in the body, and not in the air. Sensation is what is painted. What is painted on the canvas is the body, not insofar as it is represented as an object, but insofar as it is experienced as sustaining this sensation.” (25)⁶

Sensation turned toward the subject—the nervous system, vital movement, “instinct,” “temperament”—cannot be separated from sensation turned toward the object—the “fact,” the place, the event. This is the phenomenological view of “Dasein” dwelling among things.⁷ What is painted is nothing other than “my body.”

Building on this discussion of symbols, let us now go further within the field of animation. In a conversation with Yuichiro Oguro, Tatsuyuki Tanaka introduced the concept of “re-symbolization.” For Tanaka, reality is a springboard for symbols. Re-symbolization means “re-examining and reconstructing manga’s—or animation’s—symbolic modes of expression on the basis of reality, and presenting them once again as simple manga symbols.” (26) This inevitably calls to mind Gombrich’s theory of “schema and correction.” Of course, re-symbolization here does not refer to the ordinary process of “matching,” in which painters adjust the schemas already in their minds to meet the needs of depiction. It refers instead to an experimental transformation of animation’s schemas with historical significance. Tanaka’s example is Satoru Utsunomiya’s work on Gosenzosama Banbanzai! (1989). Both its investigation of three-dimensional form and its way of representing light profoundly influenced those who followed; together with Mitsuo Iso, Utsunomiya launched the realist revolution in the history of animation drawing.

I think re-symbolization has two essential points. (1) In the process of re-symbolization, reality is only a springboard. The aim is not to approach reality, but to reconstruct, on the basis of our apprehension and articulation of reality, a model of relations that yields equivalent responses. Changes in schemas arise not from “the discovery of likeness, but from the discovery of equivalences; these equivalences allow us to see reality in terms of an image, and an image in terms of reality. Their basis lies less in the similarity of constituent elements than in the identity of the responses produced by certain relationships.” (27)

(2) A simplified treatment may often be more effective than a complex one. Re-symbolization can also be understood as omitting redundant and roundabout forms of symbolic expression, proposing a simplified treatment in which “one visual device does the work of several.” In Gosenzosama Banbanzai!, for example, Utsunomiya does not represent the head’s shadow as layers of attached shadow where the head meets the neck, as most previous animation did. Instead, he treats the head as an obstacle to the passage of light, casting its near-circular shadow directly onto the body. Compared with the earlier use of shadows merely to emphasize character detail, this seemingly crude approach both (1) brings out the head’s three-dimensional form and (2) directly suggests the location of the light source, actually increasing its effect. It is an extreme simplification of the complex behavior of light and shadow in reality, yet it feels quite real. Whether the shadow conforms to reality is of no importance at all.

The reality-oriented work of Takahata, Miyazaki and their collaborators on Heidi and 3000 Leagues likewise uses reality as a springboard, reconstructing animated symbols in which “great skill seems clumsy” to give viewers a sense of reality. Take the scene in episode 2 of Heidi in which Grandfather melts cheese over the fire. The construction of the cheese as a symbol is unprecedented. In the illusion of melting cheese, viewers see its smoothness, viscosity, softness and brightness. Through their “recollection” of the real world, they imagine its silky, rich flavor and texture, gaining an experience on several levels. Inoue says, “You can’t find a single unnecessary line in the image. There are only simple contours, highlights and shadows, yet somehow that cheese looks more delicious than real cheese. Even in an image whose information has been reduced to the absolute minimum, viewers can still extract from animation the impressions we get from seeing actual things. That is animation’s strength.” (14) We discover that our response to a white dot on a golden-yellow block is astonishingly complex: we seem to perceive in it the qualities of the whole.

In fact, cheese would never melt this way in the real world; the difference might even be considerable. This is certainly not a copy of reality. But as we watch, we understand the phenomenon of “melting” and apprehend the cheese’s qualities. Reality—the cheese and the phenomenon of melting—has been successfully represented, and it even “looks more delicious than real cheese.” Including its movement, this is an entire system of equivalent relations conveying material qualities. On the other hand, the ability to set our sensations and impressions in flux so freely also depends on animation’s plasmaticity. It invents another reality—or, rather, the body’s reality. Re-symbolization is a rediscovery of our inner bodily sensations.

Plasmatic figures return human consciousness to “prelogical, sensuous thought”—the thought of young children, primitive thought—freeing people from reason and allowing them temporarily to “forget” logic and plausibility. (28) Our projection is, in fact, a selective recollection: sometimes we remember more, sometimes less. In this complex process of memory and perception, we experience differences in the sense of reality—for example, between Miyazaki’s world and Hiroyuki Okiura’s, even though both are described as “real” and “convincing.”

Here we once more emphasize the position of the apex $S$. A sense of reality is a product of the body, not an inherent property of reality: a perspective drawing creates the illusion of depth on a flat surface; sliding layers held in contact at different speeds creates a multiplane effect—an apparent perspective; the brain fills out animation on threes—one drawing held for three frames—into fluid action. Compared with the camera’s mechanical recording and preservation of space and time, and the body’s outward presentation of itself, our apprehension of animation’s symbolic expression takes place at a more inward bodily level. More boldly, you might call this what Francis Bacon calls the “nervous system.” We feel a physiological pleasure in Kanada’s animation, akin to a fetish for a particular mode of expression. More fundamentally, the forms of animated symbols are themselves constructed and reconstructed through the body’s apprehension of real space and time, and of itself. The production of animated symbols is an “addition” that works through internalized experience: the existence of space, time and the body receives a certain intensification in symbolic form. (29) Just look at Yoshinobu Inano’s depiction of the flesh in the ending sequence of Aura Battler Dunbine (1983), or at his drawings of Lynn Minmay!

Finally, re-symbolization is a long road of development, and a road of innovation that entails pain. For various reasons, especially animation’s potential status as a commodity, a particular mode of expression often hardens into a fixed pattern.

In 2010, Tanaka observed on Twitter that in recent Ghibli productions, “the expressive power of the drawing and animation has reached another dimension, but the patterns of character acting haven’t changed or gained anything since Heidi and 3000 Leagues.” (30) This is a sharp criticism. Yet we can see that, even with the participation of animators as innovative as Kanada and Shinya Ohira, the overall mode of Miyazaki’s work has not changed qualitatively. Its animation as a whole continues the styles of Yasuji Mori, Yasuo Otsuka and Miyazaki from the 1960s and 1970s. In another sense, of course, this also demonstrates the strength of Miyazaki’s control. (For those unconcerned with changes in modes of expression, this criticism is of little consequence.)

In The Tale of the Princess Kaguya (2013), however, Takahata developed what he had done in My Neighbors the Yamadas (1999), radically reconsidering his work on Only Yesterday (1991) and before, and turning completely toward the expressive possibilities of drawing and painting themselves. In “Why The Tale of the Princess Kaguya Is So Astonishing” (31), Fujitsu argues that Takahata’s work on Heidi and other productions had been devoted to “verisimilitude.” More precisely, the verisimilitude animation had pursued for over half a century was “the verisimilitude of live-action cinema.” The Tale of the Princess Kaguya “graduated” from this tendency and turned toward “the verisimilitude of painting,” offering animation a possibility other than development toward “cinematic verisimilitude.”

Fujitsu’s account lacks sufficient detail and is, in several respects, inadequately formulated. First, as Eiji Otsuka says, Takahata’s realism is not a matter of faithfully copying appearances (32): it does not transcribe and replicate reality and nature. The work of Heidi is better described as “how to let viewers experience a sense of reality through childlike, awkward-looking drawings.” (30) As in the earlier example of the cheese, the focus is the problem of symbolic representation, not live action or likeness.

But the pursuit of likeness can indeed produce a sense of reality. For animation to construct a fully realized world with depth, as live action does, “addition”—increasing the amount of information—and more likeness seem inevitable. Even Takahata’s own subsequent work, in the character design, acting design and background art of Grave of the Fireflies (1988) and Only Yesterday, seems to draw ever closer to resemblance and complexity. If I had to name a shortcoming in Grave of the Fireflies, it would be that the density of detail multiplies the intensifying effect of symbols. The resulting sense of reality is so chillingly strong that it interferes with the calm lucidity of Takahata’s direction.

Animation can certainly achieve some directorial aims through likeness. But where does that path end? Or, put another way, where does animation’s own advantage over live action lie on such a path—how is animation not to be regarded as something inferior to live action? Presumably, Takahata recognized the danger in his own work and therefore changed direction with My Neighbors the Yamadas. He completely rejected “external,” “false” realism: specifically, likeness within individual frames. His character designs became closer to manga, boldly adopting figures just four heads tall. He also actively absorbed the achievements of Utsunomiya’s realist revolution, employing animators such as Ohira, Shinji Hashimoto and Masaaki Yuasa on My Neighbors the Yamadas, and Osamu Tanabe, Hashimoto and Norio Matsumoto on The Tale of the Princess Kaguya. He pursued reality between frames, emphasizing the convincing movement produced by the rich, subtle trembling of the lines themselves. This power to convince comes from the key animator’s apprehension and articulation of rhythm, force, weight and other qualities of real movement. Making the fullest use of animation’s plasmaticity and bringing a distinctive imagination to bear, the animator conveys this understanding to viewers through the creation of symbols such as timing and spacing, with their pronounced bodily character. This is “inner” reality. Takahata’s thinking finds full realization in Hashimoto’s work on My Neighbors the Yamadas and The Tale of the Princess Kaguya: respectively, the father peeling and eating a banana, and Kaguya running wildly beneath the moon.

Take Kaguya’s moonlit flight. How can we recognize her running in those tangled, interwoven lines? And how is it that, as the lines tremble and tumble, we find our own heartstrings trembling violently too? If we isolate a single line from a single image here, we may be unable to recognize its meaning or function. The lines—and even their movements—are ambiguous and multivalent. (33) This is not the “spatialization” of time performed by a camera, clearly cutting up the world into a sequence of frames. Though it may formally resemble a sequence, each individual image has abandoned most “external” likeness and loses most of its meaning when isolated.

The meaning of an individual image lies in the images’ mutual “interpenetration.” This is by no means a precise, rigorous causal relationship. In a certain sense, this ambiguity and interpenetration might even be said to go beyond “articulation.” Yet it truly conveys force and rhythm, and lets us experience complex, chaotic emotions—anger, sorrow, madness—in the trembling of the lines, making the sequence an organic whole. This is very close to what Bergson calls duration (durée): “A violent love or a deep melancholy takes possession of our soul: here we feel a thousand different elements which dissolve into and permeate one another without any precise outlines, without the least tendency to externalize themselves in relation to one another.” (34)⁸ Through watching animated symbols—the movement of lines—viewers undergo an experience of continuous, freely unfolding change: plasmaticity. Could we say that the animator’s experience of time’s flow is conveyed to the viewer—or that the body’s intuition and apprehension of duration and the flow of life are given back to the body?

A further point needs to be made here: Hashimoto’s work demonstrates that there is no clear division between an orientation toward reality and expressionism. In his conversation with Hashimoto (35), Oguro is determined to establish when Hashimoto’s orientation toward reality began, but the two actually understand “reality” differently. Oguro also keeps pressing him on details: how long the video recordings were, whether he watched them frame by frame, whether he referred to them frame by frame, and so on. I suspect there may be a tendency to place such emphasis on animators’ ability to work unaided that animation using live-action or 3D reference is indiscriminately looked down on. But this only shows that such viewers can analyze the issue solely in terms of the complexity of forms and movements, without realizing that a key animator’s individual sensibility may not disappear through the use of “reference.” Hashimoto used extensive live-action reference for My Neighbors the Yamadas, but this in no way detracts from the excellence of his work; even Miyazaki praised it highly. More importantly, Hashimoto mentions the importance he places on the bodily feeling of performing actions while filming himself for reference. And we also know that, while animating Asuka’s battle, Iso even knocked holes in his desk and wall.

To return to The Tale of the Princess Kaguya: after the long development in animation history of “addition” within individual frames, this is Takahata’s return to the expressive possibilities of simple animated symbols, to movement between images that has discarded “outward appearance.” The quality of information becomes more important than the quantity that inevitably comes up whenever animation is discussed. Nor is it only a matter of movement. By maintaining the look of watercolor on paper throughout, emphasizing the plain brushwork and texture of the line drawings, and boldly leaving areas of the image blank, Takahata offers a directorial strategy for “subtraction” in animation. In other words, “the idea reaches beyond what the brush depicts.” As Secrets of Landscape Painting, attributed to Wang Wei, says: “A pagoda’s top reaches to the heavens; there is no need to show the halls. Now seeming present, now absent, now above, now below. Amid grassy mounds and earthen banks, let the eaves of storehouses show only halfway; of thatched dwellings and reed pavilions, give only a glimpse of the framing.” Such blank spaces can instead bring the viewer’s mechanism of projection into the fullest play, calling on memory and experience.

Another essential aspect of this method is the rejection of the cel-animation model. Takahata understood early on that this model had a weakness: the irreconcilable difference between layers that Tanaka describes as “the characters are line drawings, while the backgrounds are painted in an oil-painting style.” (26) This inevitably affects the sense of reality. In Chie the Brat (Jarinko Chie, 1981), Takahata therefore used backgrounds in a pen-drawing style close to original color manga artwork (36), giving both characters and backgrounds contours. But this still could not bridge the difference in material qualities between the character and background layers. By adopting a watercolor style throughout My Neighbors the Yamadas and The Tale of the Princess Kaguya, however, he achieved a unity of texture between characters and backgrounds.

In contemporary animation, there is a tendency to paint backgrounds so that they resemble photographs of actual locations. This makes the textures of the layers increasingly disparate; attempts are then made to bridge the gap through photographic techniques resembling live action—filters, lighting and shadow effects, the introduction of depth of field, and so forth. For example, Kyoto Animation art director Mutsuo Shinohara says of Sound! Euphonium (2015):

“Although the instruction for the backgrounds was to ‘aim for reality,’ in the first season we went too far and made them look like live action. Too much is as bad as too little. Since this is animation, I think it ultimately has to come down to what drawing and painting can express. We have to make backgrounds that work with the characters’ actions, but the first season fell a little short in that respect… Reflecting on this, I paid particular attention to it in the second season and made sure to convey that thinking to everyone in the department.” (37)

Shichiro Kobayashi observes that ours is “an age that has lost tactility”: this is why images that look as though they were shot with a camera can be widely accepted even in a medium such as animation. (38) This, too, is a sharp criticism. “The verisimilitude of live-action cinema” certainly remains an important path and method. But we have to ask: as animation strives ever harder to approach the effects of live action, are we gradually losing our bodies?

In an essay on Kobayashi’s art, Dezaki says outright: “At any rate, I have always rejected an attitude that seeks merely to copy its subject accurately. A sense of reality lies not in realistic depiction but in the essence of things”; and “cel drawings and backgrounds cannot be considered wholly separately… Only when the two become one does a mode of expression begin.” (39) These direct statements are all the more profound for their directness, and they resonate with our earlier discussion. Of course, the “essence of things” Dezaki means here is, more precisely, “not the essence of the material world, but the essence of our responses to the material world.” (27) What matters is not the causes at work in reality, but “the mechanism that produces a particular response.” The production of The Adventures of Gamba (Gamba no Boken, 1975), for instance, began with the question of how mice see the world. Dezaki writes: “For mice, the uneven, pitted surface of a rock close at hand is more essential than the overall shape of rocks and other things. This means, for example, that even when they reach ‘Battleship Island,’ what first leaps into view is the rusted surface of steel.” “Essence” here means the mouse’s bodily sensation. Where “the sensing and the sensed” become one, we identify with the painter, and with the mouse.

It is worth noting that, in an echo of Takahata’s rejection of “false” realism in The Tale of the Princess Kaguya, Sunao Katabuchi, deeply influenced by Takahata, also drew his characters in a style close to their manga forms in In This Corner of the World (2016), likewise achieving a vividly lifelike sense of reality.

Recently, in a discussion of Satoshi Kon’s work for the special feature in the August 2020 issue of Eureka, Inoue perceptively remarked:

“Kon was always someone who really sought new things in image-making, and I think there are both achievements and failings in that. Detailed drawings aren’t necessarily good drawings. You can’t say the pursuit of detail is wrong, but it also has many negative aspects for animation… Once you’ve drawn enough, you realize that the appeal of a drawing can never come from pursuing detail and correctness alone… So now, looking again at the pursuit of detail and correctness that advanced from the time of Kon’s debut, I feel strongly that we need to go back to the starting point and reconsider what makes a drawing appealing. Isn’t Japanese animation at precisely such a turning point now?” (40)

Takahata, a leader and energetic driving force of realism in twentieth-century animation, underwent just such an important turn in his career. The radicalism of The Tale of the Princess Kaguya lies not in a return to animation’s dawn in the 1960s, but in something traceable back to the expressive forms of Heian-period picture scrolls (31)… The significance of this deserves repeated reflection.

Takahata’s achievements in animation history go far beyond what we have discussed here. I think the animation world still needs to digest the legacy of The Tale of the Princess Kaguya slowly, and again and again. In China, the ink-wash animation produced by the Shanghai Animation Film Studio in the last century is an art form comparable to Takahata’s work, and one of the few forms capable of expressing what is distinctive about our national culture within animation. Take The Cowherd’s Flute (1963). Yet this great legacy does not seem to have blossomed and borne fruit in today’s animation world. The development of modern technology has produced works like The Tale of the Princess Kaguya, which uses the latest technology to create a classical sensibility. What possibilities, then, are there for traditional ink-wash painting to revive and flourish anew in Chinese animation?

Conclusion

In the discussion above, we have tried to take Takahata’s realist animation as a starting point for exploring the question of reality in animation. We began by seeking answers in Bazin’s treatment of the question of reality in cinema. We found that Takahata’s direction can be compared with Bazin’s film theory: for the methodological question of “how animation represents reality,” live-action cinema offers examples from which animation can learn and which it can assimilate. But we also found that animation has many distinctive characteristics of its own, different from those of cinema. Thus, even where animation and live action resemble one another in directorial method, their expressive forms and inner natures remain very different. Animation’s “orientation toward reality” does not mean copying reality; animation is not an “asymptote of reality.” Its “sense of reality” has more complex sources, related to the ways people apprehend the world. In the case of “space,” for example, it is related to the images that human vision generates in the brain; in the case of “time,” to how people apprehend time through movement.

In “Shots and Space,” our focus was on a major foundation of Takahata’s method: scene design, or Miyazaki’s layouts. Miyazaki’s design approach, based on the feeling of the human eye, greatly influenced the animation that followed. It also gave us a glimpse of animation’s ability to invent another reality: the reality of our bodies. In “Movement and Time,” we principally discussed three things: (1) Takahata’s respect for real time and his insistence on movements being shown in full, in animation planning and acting design; (2) the time that passes when characters are motionless; and (3) the reality of narrative structure, Takahata’s respect for the completeness of events.

Finally, “Reality and Symbols” is the core of this article, occupying half its length. To explain the question concerning animation’s nature—“what makes it possible for animation to represent reality?”—we borrowed concepts such as Bergson’s “memory cone” and the psychological term “projection.” We also had to bring in a little knowledge of painting. I hope these concepts have not made you lose sight of the direction of my argument. The main points are, in fact, quite clear, though they connect with and pull on one another: (1) the power of representation and the freedom of plasmatic transformation latent in animated symbols themselves; (2) the latent bodily character that animated symbols naturally possess; and (3) animation’s reality as something directed not toward likeness or similarity, but toward the development of equivalences, the investigation of bodily responses and bodily sensations. These points run through almost the entire section, recurring and continually being extended and deepened.

But given the limits of length and the breadth of what is involved, I can offer only a modest inquiry. In fact, this article constitutes only the first half of what I had envisaged, merely the foundation for the second half—or, as the title has it, a “point of reference.” Yet this “first half,” which I initially thought I could handle in five thousand Chinese characters, kept multiplying and extending itself until it became a monster that had taken over the entire project. I therefore had to cut away the entire framework of the second half. But if this imperfect essay can be of some help, some use, to its readers, nothing would make me happier.