Part II
As I said in Part I, one fundamental problem with “The Basics of Gaze Mechanics in Manga” is that Izumi Nobuyuki (hereafter “Izumi”) fails to integrate the various theories, models, and concepts he uses even at the most basic level. This also leaves him entirely unaware of the discontinuities within his own article. In its second half, when he begins to explain a “grammatical subject–object theory” directly, that discontinuity becomes completely exposed: is it the sequence in which narrative information is presented, or his so-called “gaze mechanics,” that determines left–right placement in the image? To take the question further, is it manga narratology or manga “gaze mechanics” that dominates the reader’s cognition when reading manga?
From beginning to end, Izumi never clearly distinguishes the two entirely different levels at issue: “the perceptual level” and “the narratological level.” He either evades or ignores this distinction, or connects the two in exceedingly vague terms. And this is precisely my point of entry for answering the central question, after so much destructive work undertaken with the aim of building a new edifice. Throughout Part II, our analysis and argument will revolve around this distinction.
Izumi Nobuyuki’s Fundamental Problems
Before that, though, I need to make a few things quite clear. First, although I have offered a comprehensive rebuttal of Izumi’s article, this does not mean the phenomena he observes do not exist. On the contrary, I have no doubt that a statistical analysis of the characters’ left–right placement in Negima! would very likely reveal a high correlation with character viewpoint and relative strength. Indeed, if the phenomena themselves could not even be observed, the theory would obviously have been questioned at once, and there would have been no need for me to make this argument.
Izumi’s real problem is that he wants to construct a theory of manga, yet falls far short of the level of scholarship needed to construct such a theory. As a result, he keeps exaggerating the influence of particular factors while ignoring others, wrongly attributing many impressions to entirely unrelated elements, failing even to recognize basic structures and elements, and going so far as to selectively omit and “re-edit” his material. Ultimately, he both reverses cause and effect and proves wholly incapable of untangling and integrating the knot of diverse opinions, models, and theories he has written. This is also one reason I must criticize him rigorously: he is not simply writing a manga review, or offering some straightforward analysis of form and style in a particular work. If that were all, I would have no reason to make such a battle of it. However many mistakes a review or analysis contains, they are relatively less important—and who has never made a mistake? The analyses of animation direction that I wrote online a few years ago were full of errors and omissions too; rereading one of them years later left even me deeply embarrassed. But Izumi is constructing a theory of manga cognition and reading, and that theory has had an influence. When it comes to theory-building, methods, analytical practices, and theoretical competence of this sort are unacceptable under any circumstances.
Second, I will now offer a straightforward analysis of the visual language of several sequences from volumes 8 and 9 of Negima!, using it to help Izumi find a genuinely constructive hypothesis about left–right placement, one that holds at least provisionally. Before I do, however, readers should note that I am by no means claiming the examples below are the best choices for analysis and theoretical exposition. I am using them because Izumi used sequences from these two volumes of Negima! as his examples. Naturally, using the same volumes makes my rebuttal more persuasive, whether I am addressing his selective omissions or his mistaken analyses of the manga.
This, of course, raises another question that must be answered: why do I think Negima! is not a good choice? Answering this both further clarifies my position on research methods and identifies another of Izumi’s fundamental problems. First, let us see why Izumi chose Negima!:
“The artwork in Negima! is a collaboration between Akamatsu Ken and the members of his studio, making it a manga somewhat resembling animation in style. To keep everyone’s expressive techniques in balance when working together, the work generally maintains middle-of-the-road, conservative compositions, rather than cutting loose (compared with other works with more distinctive styles). This is one of its characteristics. It is therefore relatively convenient as a sample for analyzing gaze mechanics.”
This explanation is barely acceptable in itself. Negima! is indeed, in certain respects, a manga whose style resembles animation, and its overall style rarely “cuts loose”—though it is by no means middle-of-the-road or conservative. But this most certainly does not make it “relatively convenient as a sample for analysis” in constructing a theory. Quite the opposite: as a sample for analysis, Negima! is extremely, extremely, extremely—add a few more exclamation marks if you like—bad.
This is not to say that Negima! is a bad manga. On the contrary, I think its technique is excellent and accomplished; I even believe that many parts of it stand at the very pinnacle of Japanese manga. If I were to write a history of Japanese manga, I would certainly praise Negima! lavishly and put its school-festival arc in my “Top 10 Japanese Manga Stories You Must Read Before You Die.” But for developing and testing theories, Negima! is far too complex. Throughout the work, the logics of montage, layout, and writing are mixed together, viewpoints and modes of cognition shift in complex ways, and there is even a great deal of ensemble action. By “a great deal,” I mean at least twenty-odd relatively important characters, perhaps forty or fifty, and sequences in which five or six, or even more than ten, often appear at once. The school-festival arc I love most also contains many exquisitely intricate ensemble sequences.
If Negima! is unsuitable, what does a good “sample for analysis” look like?
Anyone who thinks scientifically should easily understand why this is the best kind of example for developing and testing theories—especially concerning the characters’ left–right placement, which we have discussed over and over: it is very simple.
Again, this does not mean Doraemon is artistically inferior. It means that its stories are simple and easy to understand, its very small cast appears repeatedly, its panels are mostly narrative shots of a very simple form, and it is a very long series. These features make Doraemon much easier to compare, control, and manipulate than Negima! in terms of scientific research methods. Indeed, existing scientific studies of manga do favor works of this kind as their starting points. This is partly because Japanese manga was nowhere near as complex when Doraemon began as it is today, and partly because Doraemon addresses a broader readership and was deliberately made more accessible. Consider other famous works of the same period, such as Ryūjin-numa (Swamp of the Dragon God), still much discussed in manga teaching today: they are far more complex.
Negima!, by contrast, began in 2003, by which time Tezuka Osamu, Tokiwa-sō, Tsuge Yoshiharu, gekiga, the shōjo manga movement, and countless other manga and manga artists had already made their historical contributions, leaving Japanese manga with an extraordinarily rich range of styles, genres, grammars, and forms of expression. With its unstinting use of these grammars and forms, Negima! is fundamentally a “sample for analysis” in which variables cannot be manipulated or controlled, and even a single variable cannot be identified. Even its formally simplest sequences are far more complex than Doraemon is most of the time. And Izumi, who does not understand scientific methodology, has not even chosen the relatively simple sequences.
The real reason Izumi chose Negima! as his “sample for analysis”—allow me here to speculate about his motives; I really cannot resist—is probably that it was still being serialized when he was writing. Analyzing it would make his theory look applicable and up to date. He also happened to notice in Negima! that “the protagonist is always on the right and the antagonist always on the left,” whereas in many other manga this pattern is less obvious or simply does not exist. So Negima! “became” a “middle-of-the-road,” “conservative” manga, and other manga “became” “more distinctive in style.” When Izumi calls Negima! a “convenient” sample, he certainly does not mean convenience in the scholarly sense: ease of precise definition and classification, ease of rigorous reasoning and analysis, ease of observation, measurement, and comparison. What he means is probably the kind of “convenience” we ought most to avoid in scholarship—at least in the study of form and style: easy to interpret in my favor, easy to make me look right.
With that said, the preliminaries are complete. I will write separately about the application of scientific methods to manga and animation scholarship at some point in the future, though I cannot say when. Nor will I further discuss my criticisms and proposals concerning general methods and methodology in the study of form and style here. I will devote my next essay, “The Study of Form and Style and Scientism 2: Metacriticism,” to that subject.
Before we begin the analysis proper, let us lighten the mood. Look at this manga page:
According to Izumi’s theory, can you feel the airship in the first panel getting just a tiny, tiny, tiny bit faster? Or should we use his other rhetorical device here: the airship flying from right to left moves slowly forward along the “invisible current”?
More interesting still, after noticing the airship on this page—page 98 of volume 9—I made a simple count as I read. From this page to the last page of volume 10, the airship appears in long shot a total of twenty-three times. Twenty-one times it flies from right to left across the image; only twice does it fly from left to right. And on one of those two occasions, it occupies such a tiny area in the upper-left corner that I did not even see it on my first count. In other words, of these twenty-three panels showing the airship in long shot, Izumi just happened to choose the left-to-right airship that appears only twice—or effectively only once. Such a coincidence of fate gives me much to ponder.
Let us return to the subject.
Narrative Sequence or “Gaze Mechanics”?
What actually determines the characters’ left–right placement in manga? To answer this, we must first establish what kind of question it is. The discontinuities in Izumi’s article give us a clue.
In the first section of “The Basics of Gaze Mechanics in Manga,” Izumi devotes a great deal of space to explaining a particular visual-perceptual effect produced by the combination of manga reading and composition: an “invisible current” flowing from right to left. This current forms the basis of his entire discussion of so-called “gaze mechanics.” Later in the article, however, he suddenly begins to discuss a “grammar” of left–right placement, which puts the grammatical subject on the right and the object on the left. A manga panel’s sentence structure consists of the subjects, objects, and modifiers in the image.
First, we must clearly recognize that these are two things of entirely different kinds. When Izumi first explains “gaze mechanics,” it is indeed still a “mechanics” of vision—or, more precisely, the perception of directed tension or forces in a static image. Even here, we could already criticize him severely: with his extremely limited understanding of visual-perception theory, he departs from Natsume Fusanosuke’s originally quite apt judgment and rashly proposes an original view of the perceptual effects of manga reading. I also very much doubt whether readers can actually feel the directional tension in his examples. But in any case, at this point we are still talking about visual perception.
As Izumi gradually extends this to “viewpoint” and “strength,” and finally produces a visual “grammar” out of nowhere, how exactly does “visual mechanics” connect with these concepts? Is he really still discussing “visual mechanics” at this point? No. The concept has become meaningless. Through constant substitutions and shifts in meaning, it has become a theoretical edifice built in defiance of the rules. What connects directly with visual “grammar” is not visual perception but a narrative sequence: when we read manga from right to left, we follow a chain from A→B→C, reading deliberately arranged narrative information in a definite order.
Only after clearly distinguishing narrative sequence from the perception of visual dynamics do we have the conceptual basis for understanding the mechanisms behind “viewpoint,” “strength,” and even “grammar.” However much Izumi muddies the waters again, we can quickly clear them.
Let us start by sorting out the question of viewpoint. In “Positions Divided Between Left and Right,” Izumi attempts to develop the “speed theory” of “gaze mechanics” into a “viewpoint theory”:
“Merely differentiating the characters’ left–right positions or the directions they face (within the panel) can express the panel’s ‘viewpoint (subjectivity),’ or ‘the character in control (the main character in this panel).’ (Chinese translator’s note: Or, in other words, the character the reader identifies with—who ‘I’ imagine myself to be.)”
“A character drawn standing on the right of the panel, facing left, along the vector running with the current, is aligned with ‘the reader’s gaze (its direction of flow)’, so the reader can easily feel at one with the character inwardly (identify with that character).
Conversely, a character standing on the left of the panel, facing right, along a vector running against the current, seems to confront ‘the reader’s gaze (its direction of flow).’ Viewed from outside (from an objective standpoint? From the reader’s position outside the page?), it is as though, for the reader, (the character) suddenly appears from a part that has not yet been read, making the character seem hard to understand and therefore hard to identify with.”
“Furthermore, as readers enter the panel’s field of view in reading order, the first thing they look at must be the character on the right; the character on the left is seen second or later. (There may be exceptions depending on the particular composition, which I will discuss below.)
Thus, readers preferentially adopt the viewpoint of the character on the right, and understand the left-hand character as an ‘object seen objectively’ in relation to the right-hand character’s inner life. This is a basic model.”
On a first reading, if you are not paying close attention—and with an illustration as misleading as Figure 15 to help him—Izumi may well take you in. Readers with some theoretical grounding will, of course, immediately doubt such an explanation. Japanese manga has adopted and adapted a great deal of visual vocabulary from film, including all kinds of ways of constructing viewpoint. If “gaze mechanics” really had the effect Izumi claims, how could this audiovisual vocabulary still be used in manga?
Once we have clearly distinguished narrative sequence from the perception of visual dynamics, the first question is: how much of the effect here comes from narrative sequence, and how much from the perception of visual dynamics? We then immediately notice that Izumi constantly conflates the two. He attributes viewpoint effects to his supposed “current” and “vector direction,” while simultaneously incorporating the mechanism of narrative sequence as a supplement. This confusion directly prevents him from understanding a composition like this:
This sequence shows Asuna watching several students playing with magic wands. Clearly, the construction of viewpoint here does not fit Izumi’s explanation at all. As we read the second row, our viewpoint belongs to Asuna on the left; on the right are the objects of her observation. So-called “gaze mechanics” plays no part whatsoever. What operates here is narrative sequence and narratology. After the narration of the first row, our viewpoint moves away from the students playing with their wands, back along the line of sight to their observer, Asuna. This transitional visual language means that the objects of observation must appear before Asuna, and therefore the observed students are placed on the right of the image, while the observer—the character to whom our viewpoint is also attached—is placed on the left.
Look more closely at the structure. We first see several people on the right practicing magic. At this point, our viewpoint belongs to no character; it belongs only to ourselves. As the narration moves leftward, Asuna is watching them. We notice her observation, and at the same time retrospectively recognize that our earlier viewpoint was hers, so that our viewpoint coincides with Asuna’s. Readers familiar with film language will immediately understand that this manga panel is equivalent to an over-the-shoulder shot pulling back from frame A to frame B. This, too, is a very common construction of viewpoint. Clearly, it does not fit Izumi’s “basic model” of viewpoint; indeed, it is its exact opposite.
Is this, then, one of the exceptions Izumi says “may occur depending on the particular composition”? Let us look at some more examples.
This sequence comes immediately after the chase in the first Negima! example Izumi uses in his article. Once the chase ends and the two sides establish one another’s identities, the man in the white suit asks Negi why he is protecting Chao Lingshen. According to Izumi’s so-called “viewpoint model,” the man in the white suit stands on the right, facing left, so we adopt his viewpoint; Negi, who appears after him facing right, is our object of observation. But this clearly does not fit the story. In the narration of this sequence, we learn at the same time as Negi that Chao Lingshen is being pursued by unidentified people; like Negi, we do not know the real reason for the pursuit. Our viewpoint coincides with Negi’s, and this shared viewpoint is also given visual form in the subsequent panels. Why, then, is the man in the white suit on the right and Negi on the left? Because in this sequence the man in the white suit asks Negi a question first, then explains to him their legitimate reasons for pursuing Chao Lingshen. In the narrative sequence as a whole, the man in the white suit’s actions precede Negi’s subsequent reactions, so he is presented first—and thus naturally appears on the right.
Later, we see Negi return to the right of the panel composition, with the man in the white suit on the left. The compositional viewpoint in the first panel is now closer to the man in the white suit’s, while the panel in the lower-right corner takes the viewpoint of the young woman with white hair. In fact, viewpoint is no longer at issue here at all, because these panels are all components of shot/reverse-shot logic. Which character occupies the right of the panel has nothing to do with any of Izumi’s theories—the “speed theory,” the “viewpoint theory,” or the “strength theory.” The real reason lies in who speaks first and who acts first. Shot/reverse-shot logic is one of the most common approaches to panel organization in Negima!. This is one reason I can just about agree with the claim that it resembles animation.
Another example. Figure 20 has already appeared in Izumi’s article. According to him, what operates here is the “strength theory”: the boy on the left is dazed and has lost his will to fight, so the man in black controls the fight, which is why he is on the right. Izumi does not explain why it is reasonable to invoke the “strength theory” here rather than the “viewpoint theory.” Of course, we can easily infer the real reason: his so-called “viewpoint theory” cannot explain this case. But before we laugh, let us help him out with a supplementary explanation. Why is there no subjective viewpoint here? Because this is an establishing panel—an establishing shot. As we can see in Figure 19, the opening panels of the sequence offer partial views: the boy knocked to the ground, the man in black taking his stance, and the different reactions—in facial expressions—of several people to what has happened. Only in Figure 20, on the next page, are readers clearly shown the characters’ spatial relationships and positions, acquiring an understanding of the scene as a whole. This overview of the scene belongs to no character’s viewpoint. It is an objective narrative panel. What, then, of the “strength theory”? Does the relationship of strength determine that the man in black is on the right and the boy on the left? No. The real reason the man in black is on the right is that he acts first in the scene, and also acts first in the narration—note the crucial difference—so he is on the right of the image.
We have now listed three types of panel: panels with the viewpoint character on the left, shot/reverse-shot panels, and establishing panels. All three are extremely basic components of manga’s visual language. They are also extremely common in Negima!—so common that they are used in most of its sequences. After the preceding examination, readers will presumably have noticed that what determines who is on the left and who is on the right in these panels always has to do with the mechanism of narrative sequence, and nothing to do with Izumi’s so-called “speed theory” or “viewpoint theory.” If anyone thinks this is still insufficient, I can find more types of panel in the same two volumes Izumi uses, all unrelated to or even at odds with his theories. Now let us look again at his statement:
“Thus, readers preferentially adopt the viewpoint of the character on the right, and understand the left-hand character as an ‘object seen objectively’ in relation to the right-hand character’s inner life. This is a basic model.”
His so-called “basic model” amounts to nothing more than a composition like this:
This is why I call Figure 15—the illustration Izumi supplies when explaining the “viewpoint theory”—extremely misleading. What about the three types of panel I described above, three constructions used extensively in Negima!? Are they “basic models”? Or are they, along with still more visual constructions that do not fit his account, all his so-called “special cases”? The situation could not be clearer: what Izumi calls the “basic model” is merely one of the principal constructions in manga’s visual language. Beyond it lies an immensely rich variety of other visual constructions, used not only in Negima! but throughout Japanese manga and even animation.
At this point, I hope readers really understand why, at the end of Part I of “A Critique of Izumi Nobuyuki,” I asserted that Izumi “fundamentally lacks the ability to analyze an image’s elements and structure,” and that at his level he may not even be able to “identify the details of manga correctly”; why, as I write this essay, I cannot help slipping into mockery from time to time. What Izumi actually does in “The Basics of Gaze Mechanics in Manga” is like a toddler just learning to talk: he has learned a few words of manga’s visual language and excitedly starts misusing them everywhere. The result is that every unfamiliar grown man Izumi sees gets called “Dad.”
Once everyone clearly understands what Izumi really amounts to, it becomes easy to see through his other claims, his so-called models and theories. Take the “strength theory” again. In his account, right-to-left movement is stronger because it goes “with the current,” while left-to-right movement is weaker because it goes “against the current.” This effect extends to a character on the right, facing left, having greater control over the scene than a character on the left, facing right. As we examined in Part I, Izumi even claims that in Negima! Akamatsu Ken puts the “strength theory” to work in concert with specific turns in the plot: when the characters’ left–right positions reverse, control of the fight changes hands too.
Now look at Figures 22 and 23. Can you really feel how feeble the man in black’s attacks are, or that the two protagonists are about to win? Or should we dodge the main issue and say instead that this fits Izumi’s “viewpoint theory,” which is why the protagonists are on the right? Or should we use another of his shameless, unfalsifiable lines of reasoning—that this is “deliberately breaking the rule of acceleration with the current, drawing an otherwise powerfully accelerating punch against the current to convey a sense of counterattack”—and turn a direct counterexample into an indirect supporting example? This is still the problem we discussed at length in Part I. Within a composition or a passage of visual narration, so many elements and techniques can guide our cognition. Left–right placement is not only just one of them; it is also one of the less effective ones.
A Narrative-Sequence Theory of Left–Right Placement
One last question remains. If Izumi’s “gaze mechanics” is completely wrong, what actually determines the characters’ left–right placement in manga panels? In fact, when we open the manga, we find that although there are many cases like my examples above that do not fit his theories, the characters’ left–right placement in Negima! does show a regularity, and that regularity largely fits Izumi’s observations. Moreover, when we turn to the range of Japanese manga on the market, we find that, though the pattern is not so widespread, regularities in left–right placement exist in at least a substantial proportion of manga.
How can we explain this regularity once we have rejected “gaze mechanics”? I have effectively spelled it out already. In every counterexample to “gaze mechanics” we have examined, the mechanism of narrative sequence is at work. Its role is certainly not confined to these examples: it is a fundamental mechanism of manga construction. Readers read manga from right to left, and artists arrange the panel sequence from right to left. A person who speaks first or acts first always comes earlier in the narrative sequence; a person who speaks later or responds to the first person’s words and actions comes later. Naturally, the result is that the former is always on the right and the latter on the left.
What is the relationship between this simple mechanism and regularity in left–right placement? Let us think. Whoever comes first in the narrative sequence is on the right—but who always comes first? A manga’s storytelling always focuses on a protagonist: we watch what the protagonist does, share the protagonist’s field of observation, and listen to the protagonist’s thoughts. In the most common form, internal focalization, the protagonist constitutes our viewpoint; even with external focalization, our viewpoint often follows the protagonist. In narration centered on the protagonist, who always acts first, and who always observes other people? It must be the protagonist. In other words, because the characters who supply our viewpoint are always the first within our field of view to act, speak, and observe others, they always come first in the narrative sequence and ultimately appear on the right of the image, looking left. Think again: in a fight, who comes first in the narrative sequence? Naturally, the attacker. Is a character who is always actively attacking, who can always choose to attack rather than being forced to defend, often the stronger fighter? Or at least the one temporarily in control of the scene? Of course, this is a comparatively less probable association, so the regularity in manga is weaker too. In this way, we have “recruited” and reshaped Izumi’s “viewpoint theory” and “strength theory,” integrating them with a sound, solid mechanism of manga reading that is clearly more fundamental in theoretical terms: narrative sequence. We can call this the “narrative-sequence theory” of left–right placement.
In comparing it with “gaze mechanics,” we can easily see why “sequence theory” is a more reliable theory with considerable explanatory power, while “gaze mechanics” is riddled with serious problems. First, narrative sequence is plainly a much firmer foundation than Izumi’s “current.” This is not to say that narratology is more useful or more fundamental than visual expression in manga reading. Let me remind you once again: the cognitive basis of the supposed “current” and its various applied effects are entirely Izumi’s own original thinking. I have done my utmost to show that he understands virtually nothing about either visual-perception theory or manga’s visual language. His original thinking is, for the most part, speculation devoid of rigor, and naturally cannot provide a good theoretical starting point.
Second, within Izumi’s “chaotic system,” we cannot actually explain how the cognitive effects he describes arise. There is a fundamental discontinuity here. How are the cognitive modes of the “viewpoint theory” and the “strength theory” recognized and triggered, and how do they disappear? How do these two modes—or more—switch between one another or coexist? As we read manga, does our cognition of it keep jumping back and forth between viewpoint mode and strength mode for some reason Izumi has never explained? Once we recruit both into “sequence theory,” we can explain their emergence with ease: the fundamental cause of the pattern of a perceiving subject on the right and an observed object on the left lies in narratology, while the pattern of the strong on the right and the weak on the left is merely a particular manifestation of action–reaction narration in fight sequences.
Third, when “gaze mechanics”—or rather, Izumi Nobuyuki—deals with counterexamples, we see him unhesitatingly choose to eliminate his propositions’ falsifiability and apply all kinds of patches without principle, even resorting to selective omission and “re-editing” of manga examples. Not only does this make the construction of “gaze mechanics” entirely unprincipled; it also makes Izumi himself unworthy of trust. In “sequence theory,” however, the patterns of perceiving subject on the right and observed object on the left, or strong on the right and weak on the left, do not stand in a direct deductive relationship to the cognitive process of reading manga. Rather, they are phenomena that the sequence mechanism produces with high probability. This means that counterexamples to “gaze mechanics” are not counterexamples to “sequence theory,” but supplementary cases: when these patterns do not occur, the theory has not gone wrong; the sequence mechanism has operated in another form. I have already devoted considerable analysis to this above, in listing the various constructions of visual language.
Finally, as I have repeatedly criticized above, in building “gaze mechanics” Izumi gives priority to accommodating more explanations above all else. This not only deprives him of any principle or rigor; it also leads him to conflate mechanisms of entirely different kinds within the reading process. The result is that he has fundamentally explained nothing, because “gaze mechanics” has become a boundless mire. Although “sequence theory,” too, relates directly to the construction of manga’s visual language, it is strictly confined to the explanatory domain of narratology. Thus, it does not need to explain things beyond what it has undertaken to explain, and it can readily accept other theories as supplements. Take the pattern of a perceiving subject on the right and an observed object on the left, or a protagonist on the right and an antagonist on the left. “Sequence theory” already provides a fairly good explanation, but clearly there are further causes. For example, the logic of manga layout means that, once characters have been placed in directional opposition, maintaining that left–right opposition in the subsequent action is not only better layout design but also a visual manifestation of the principle of cognitive economy. “Sequence theory” does not need to reject these explanations, still less pass them off as parts of itself in the manner of “gaze mechanics.” In explaining the phenomenon as a whole, it already acknowledges the need for them as supplements.
With that question answered, this essay could really end here. But I am still obliged to offer a rather tiresome response, because, as luck would have it, Izumi has already criticized “sequence theory” in advance in “The Basics of Gaze Mechanics in Manga”:
“Positioning characters does not follow a logic as simple as ‘put whoever speaks first on the right.’ Manga composition is not so simple. By manipulating a particular composition, the left–right arrangement of the directions characters face can work independently of ‘speaking order.’ Figure 20 shows an example.
“Here, too, we can observe a reversal of grammatical subject and object. The dialogue in the speech balloons has not been written out specifically; the order in which it is spoken is the same. Let us call the character who first says ‘A’ the ‘speaker,’ and the character who follows the other’s words with the reply ‘B’ the ‘listener.’
“Before the change, then, the manga expresses ‘the speaker asks the listener something’; after the change, it expresses ‘the listener asks the speaker something in return.’ Their respective viewpoints and relations of advantage and disadvantage are reversed. Before the change, readers feel, ‘Why is the listener still grinning after hearing me speak so seriously?’ Afterward, this instead becomes the subjective feeling, ‘Why is this speaker being so serious?’ This is the kind of effect the change may well produce.
“In this way, perhaps we can judge a manga artist’s intuition or taste by their ability to ‘compose consciously and in keeping with the intended staging.’ If the idea that ‘the first speaker goes on the right’ becomes a preconception, there is a danger of producing effects contrary to one’s intentions.”
I do not know whether nearing the end of “The Basics of Gaze Mechanics in Manga” made Izumi a little impatient, but in this passage he is twisting things around without even bothering to hide it. Manga composition is indeed not simple, but its basic logic is extremely simple: “put whoever speaks first on the right.” The difficult question is not the order of narration, but how to decide who comes first and who comes later in the narrative sequence.
Look again at his illustration. After using a great many concrete examples from various manga throughout his article, this time Izumi has not even bothered to invent a specific example. Why? Very simply, because once he supplies actual dialogue, we can easily see that he is talking nonsense for paragraphs on end. Let me quickly make up an example that fits the panel design in the first row:
Panel 1
A: Why did you step on my foot?!
B: I didn’t step on it. I only brushed against it.
Panel 2
C: Nonsense! You clearly did step on it!
D: Really? Sorry, then.
Panel 3
E: Hmph. I’ll let you off this time.
Now let us look again at Izumi’s explanation. The first thing we notice is that if we put these lines, designed for the first row’s panel construction, into the second row, they cannot possibly fit without changes to their beats and meaning. Next, consider his analysis again:
“The manga expresses ‘the speaker asks the listener something’; after the change, it expresses ‘the listener asks the speaker something in return.’ Their respective viewpoints and relations of advantage and disadvantage are reversed. Before the change, readers feel, ‘Why is the listener still grinning after hearing me speak so seriously?’ Afterward, this instead becomes ‘Why is this speaker being so serious?’”
Looking back at the example I have supplied, can it really have the effect Izumi describes? Not only does it not; every sentence of Izumi’s explanation is nonsense. After the change, does this dialogue become the listener questioning the speaker in return? Absolutely not. To achieve that effect, the meaning would have to change. Have the viewpoints and relations of advantage and disadvantage reversed? Now that we understand “gaze mechanics” to be mistaken, the question of viewpoint here no longer has any significance. But the relation of advantage and disadvantage has clearly not reversed. In fact, I would like to put the question back to Izumi: how could it reverse? A character who can grin flippantly at someone standing before him in a fury clearly has the upper hand and is in control. However we change the panel construction, as long as the meaning remains unchanged, how could his advantage and control be reversed? The final sentence is more incredible still. Izumi is simply saying outright that changing the panel construction can express an entirely different meaning even with the same dialogue and the same events. But how is that possible? In the example above, does anyone really think it could turn into the question “Why is this speaker being so serious?” Of course not. His seriousness has a causal premise: “Someone stepped on my foot.” However we change the panels without changing the dialogue, the causal relationship between his foot being stepped on and his serious manner remains.
This essay ends here. I had originally intended to add a supplement on Izumi’s various other errors that I could not fit into the line of argument. But writing this has truly exhausted me, so let us stop here.
Afterword
As I said at the beginning, this essay is the first in a larger writing project. Having offered comprehensive, targeted criticism of so representative an example as “The Basics of Gaze Mechanics in Manga,” we can move on to the following questions. Where do practices of manga criticism and research such as Izumi Nobuyuki’s sit within the overall landscape of research into form and style in anime, comics, and games? Compared with other research into form and style, is Izumi’s level relatively high or low? From here, in “The Study of Form and Style and Scientism 2: Metacriticism,” I will set out my general view of the study of form and style. I will then propose my own approach to research, explaining what kind of research truly possesses “empirical grounding, systematic organization, and objectivity.”
There are also many subjects I have touched on without developing here, because they were difficult to bring into the essay’s line of argument. For example, in constructing his “gaze mechanics,” Izumi Nobuyuki undoubtedly drew on the paired concepts of kamite and shimote.1 I will devote a future essay to my views on this question, with brief responses to several related studies. I will also offer more specific proposals and comments on methods of straightforward manga analysis, especially eye-movement analysis. Scientific research into manga and animation is a much larger subject still. In “Metacriticism,” I will briefly explain it and situate it; more detailed introductions, translations, and so forth will have to wait until later.