On animetism : or, the importance of sakuga to theory

Thomas Lamarre’s The Anime Machine is undoubtedly one of the most important books dedicated to animation, and especially to anime. It manages to be at the same time a historical overview of anime and its techniques, a thorough analysis of some of its most prominent artists, and a compelling theory of animation and media in general. Most importantly, it is one of a few very precious works that try to focus on “the materiality of the moving image itself” and uses it as a starting point, rather than examining anime as a part of a larger media system, or just like another form of cinema.

Lamarre’s perspective centers around the idea of technology. Using as a starting point the provocative statement that “Much of anime is […] unabashedly low tech” [Lamarre, 2009, p.xiii], he focuses at the same time on the technology (or technologies) that make anime possible and how anime thinks about and presents technology : “I wish to indicate that animation at once works with technology and thinks about technology – and the two processes are inseparable. In anime, thinking about technology is inseparable from thinking through technology (not only using technology but also aligning thought with its operations).” [Lamarre, 2009, p.xxx] Whereas most anime histories start with individuals or studios, he begins with a technical apparatus, the multiplane camera : its historical genesis in Disney Studios, and its harnessing by Japanese artists. It is from these purely technical concerns that he the approaches practices of production : compositing, character design, plot, while always interrogating how these all relate to the theme of technology.

The multiplane camera, or the animation stand (two related, but different objects, the most fundamental being the second one), the decisive objects at the core of Lamarre’s reading, can be understood and presented in two ways : 1) in their nature, that is as purely mechanical and technological objects, or 2) in their function, that is to stack cels in order to create a sense both of depth and unity in the final image, the one shot by the static camera. This second aspect is also known as “compositing” : “Compositing is a matter of assuring that the gaps between different elements within the image are not noticeable” [p.31]. What distinguishes anime is then not a particular style or nationality, but a specific approach, or set of approaches, towards compositing. From this quick overview (1), one clearly understands that what is key in anime is not animation, but compositing : “In fact, as I aim to make clear in this book, in the analysis of animation, priority should fall on compositing (the space within images that becomes spread across frames) over character animation (movement across frames).” [p.xxv]

However, it is precisely this focus that I believe is problematic. Indeed, the two aspects of the animation stand that I just presented lead, in my view, to two of the trappings that Lamarre seeks to avoid : an overly deterministic “apparatus theory” on the one hand, and missing movement and animation on the other. I will therefore proceed to analyze these two points and offer an alternative reading of animation in anime. However, in the scope of a simple blog article, I will not be able to engage with the entirety of Lamarre’s theory ; I will therefore mostly focus on chapters 10 (“Structures of Depth) and 11 (“The Distributive Field”) which focus the most, in my opinion, on the stuff of animation, through the work of animator and director Hideaki Anno.

Apparatus theory is a branch of cinema theory that emerged in the 1970’s and 80’s ; its main argument was that cinema was an ideological tool resting on the passivity of the viewer before a technological apparatus (mostly the camera and the screen) which decisively influenced representation. In Lamarre’s account, which focuses a lot on notions like space and depth, the most important point is that of camera movement and perspective : “Baudry, for instance, insisted that the monocular lens of camera constrained it to reproduce one-point perspective, which in turn resulted in the imposition of a seemingly rational and scientifically accurate grid upon reality, enacting the ascendancy of technologized optics over human perception and generating a world in which human actions were necessarily reduced to cause-and-effect relations.” [p.xxvii]. At a first level, insistence on the alternative apparatus that is the animation stand allows Lamarre to explore another kind of representation : most notably, Hayao Miyazaki’s use of “open compositing” which plays and insists on the different layers of the image, is read as an alternative relation to technology and an opening of a new relationship to space and nature. This is the core of the distinction between “cinematism”, centered around one-point Cartesian perspective, and “animetism”, a flat image comprised of different layers. But more importantly, what Lamarre contests is the determinism running in apparatus theory : indeed, it contends that “the singular apparatus determines the whole of cinema” [p.xxvii].

Against this position, Lamarre defends the alternative concept of “machine”, which is not only the material, mechanical apparatus, but also the set of presuppositions and practices upon which it rests and that inform its use. Therefore, the major object of study must not be the animation stand itself, but its uses and transformations across time, sometimes in contradictory directions : this is what Lamarre calls “divergent series”.

However, even if we move from concrete objects to an abstract machine, the question of determinism still subsists. Lamarre does avoid falling into an entirely deterministic position ; but setting entirely the focus on the machine’s stance towards technology transforms it into a set of technological restrictions rather than a creative force. In other words, the “anime machine” works as a (powerful) hermeneutical presupposition : it is a reading tool which allows the critic to instantly situate works in a grid of possible positions. For example, in his analysis of Daicon IV, Lamarre points to a sort of “hyper-cinematism” or “hyper-Cartesianism”, that is an emphasis on movement in depth, most notably in Itano Circuses. But, invoking the force inherent in the machine, he immediately reverses it into an animetically critical form of cinematism and technological optimization : 

“The Daicon animations, however, approach optimization from a very different angle. They embrace technological optimization. The density of information, the dizzying rapidity of cuts, the explosion of projectiles across the screen, not to mention the attention to spaceships and powered suits, all are part of a technical optimization of the perceptual field. […] When it comes to thinking technology, then, a great deal depends on whether one thinks that optimization is always just optimization, or whether one thinks that there can be different modes of technical optimization, some of them better than others, some of the potentially opening a critical relation to the technological condition. Can there be such a thing as critical optimization, or does optimization invariably result in incessant crisis, in the destruction of the human life world ?” [pp.138-139]

Putting aside the philosophical argument, what Lamarre does is two things : first, he immediately reduces animation as moving images to a representation of technologically-powered movement ; then, he proceeds to a general conclusion about Daicon’s stance towards technology. In another instance, when confronted to the hand crafted stuff of animation, Lamarre again diverts the focus :

“In his memoirs, Takeda recalls that at that time Anno Hideaki had never worked with cel animation but only with paper animation. In fact, when he interviewed Anno, Anno pulled out a pad of paper and quickly produced a flip-book animation of a powered suit with great detail and complexity. […] It is significant that, to demonstrate his abilities to Takeda, Anno chose to draw a powered suit. This choice is significant not so much because it shows Anno’s allegiance to certain kinds of SF anime, but because the powered suit is truly an embodiment of the multiplanar image within a character form. […] Nonetheless, to repeat my argument in the Introduction, even though animation relies heavily on the art of hand (as with Anno’s sketches of a powered suit), such hand arts do not explain animation. Animation folds the art of the hand into a multiplanar machine, where their relation to the machinic force of the moving image thoroughly transforms them. Animation is not the art of sketching characters that will then be forced into movement. […] The art of character design anticipates its movement within the multiplanar machine, anticipating the dynamics of compositing.” [pp.129-130]

Even though this is quite a long excerpt, I believe it is one of the most important paragraphs of Lamarre’s book, because it reveals more clearly than any other his stance on animation. It shows that theoretical and technological concerns take precedence over esthetical ones ; more importantly, it questions what Lamarre precisely means by “the materiality” of animation, which he claims to study. While the anime machine is not a deterministic system or structure, as Lamarre repeatedly points out, it is difficult to construe it as anything other than an essence, therefore both a key determination (without animetism, there is no animation) and a prerequisite for any understanding. Most importantly, it appears as a tool to integrate, maybe forcefully, all kinds of images : this is what’s expressed by the metaphor of the “fold”. Then, even though Lamarre focuses on “divergent series”, that is on apparent contradictions in the machine, its force is more one of uniformization than diversification. Determinism is then not the word, and Lamarre does not fall into it ; but his understanding might be as limited, considering that it does not account for exceptions.

This is why I believe we need to expand our understanding of anime production beyond a technological account. Technology is, indeed, a necessary component, but it needs to be understood in relation to social contexts which are the driving force behind the harnessing of the machine. In other words, the machine does not stand by itself, nor is it the beginning and end of production ; it is but a necessary step, a means for an end which is, in the case of animation, the production of movement. Animetism is not, therefore, the movement itself.

This alternative look has already been developed in fan discourse, and has taken the form of the sakuga community ; here, “sakuga”, does not mean just “animation” or particularly good animation, but a certain kind of gaze : the one that realizes “a conceptual shift from anime as something to be enjoyed (read, interpreted, consumed) to anime as a cool process of production” [Condry, 2013, p.44]. As this quote makes clear, this perspective has been adopted by some academics who do not look at anime from the vantage point of philosophy or media studies, but from that of social sciences and most notably anthropology. This marks an emphasis on human practices first, and how they shape institutional systems, rather than on abstract structures. Taken as a method of analysis, sakuga must therefore mean not just an appreciation for animation taken for itself, but as the result of human and economic investment : it has to be a holistic look on animation production, from a bottom-up perspective, rather than a top-down one. If it does not, and confine itself to simple appreciation and catalogue of outstanding animation, it is but another (however respectable) form of amateur connoisseurship. The other thing sakuga must not become, however, is some kind of simplistic auteur theory which would put all creative responsibility and praise on the individual animator : it would be a failure to understand anime in its complexity and an absolutely uncritical stance on it.

Sakuga must therefore be a study of actual movement : the movement that’s on the screen, but also the movement of individuals and objects in production processes (2) and the movement of value, both economical and social. The word “sakuga” is therefore not just a nod to the fandom : among the three terms used in Japanese to refer to animation, sakuga literally means “crafted” or “made” image : even in its meaning, it points to the importance of man-made production (3). It is in light of this that I can now argue that Lamarre in fact misses movement, or rather gives a very limited account of it.

The Anime Machine claims to be a theory of the moving image ; however, I would say that from its account only, the nature of movement itself is very difficult to construe. The way I understand it, it is either something too restricted and concrete, the simple practice of compositing, or on the contrary something too abstract : movement in general, a pure force. This indetermination of meaning is what allows Lamarre to shift meanings in his definition of animetic movement : for example, when discussing differences and parallels between animetic compositing and live-action cinema, he first argues that “compositing in cel animation is analogous to camera movement in cinema” [Lamarre, 2009, p.124] and then changes its role, describing it a page later as “internal montage” or “editing within the image” [p.125]. Later, the same problem of consistency arises : “Giving ontological priority to movement has led to an emphasis on compositing or “editing within the image”. In effect, compositing is analogous to camera mobility in cinema, and character action is analogous to montage.” [p.191] But more importantly, I believe that Lamarre’s view on movement rises two more fundamental problems :

1. It reduces movement to a purely mechanical force without intentionality. A passage I have already quoted makes it evident : “Animation folds the art of the hand into a multiplanar machine, where their relation to the machinic force of the moving image thoroughly transforms them. Animation is not the art of sketching characters that will then be forced into movement.” [p.130] The specific expressivity aimed at by the animator or character designer is put aside in favor of the more general concerns of technologically supported movement. The origin of this position might be found in Lamarre’s ambition to mimic “the attitude of experimental science and technology studies” [p.xxvii] : just like in modern science, movement is but data which can be analyzed and calculated to give birth to prediction – or, in the case of criticism, predetermined interpretations. However, on a general level, one could argue that this misses the core nature of movement, that is its dynamics, the ways movement is not a purely quantitative process, but a factor of change both in the object that is moving and the space and time it is moving in. Moreover, this mechanical point of view entirely rules out the possibility of final causes – that is, of intentionality. But considering that we are, after all, not just studying technology or media but art or something like it, creative intent (even if it is something as vague as creating dynamic movement) cannot so readily be cast aside.

2. It reinstates the duality between Wester full and character-focused animation, and Japanese limited, non-animation. Lamarre himself aims to counter this “tendency to think the distinction between full and limited animation in terms of movement versus stasis” [p.184], and I fully share his analysis of the problems arising within it. However, I wonder if compositing can really be called animation ; and most importantly, I believe that Lamarre’s emphasis on it may very well lead to a complete absence of care for other kinds of movement, most notably character animation : “I see a priority of (a) compositing over character animation, and of (b) compositing and character animation over camera movement and montage” [p.191]. In other words, Lamarre accepts too readily the idea that limited animation is limited towards acting and expression, and therefore seeks movement elsewhere ; but this preconception might very well be misinformed.

The origin of this problem is in fact fundamental, because it lies in Lamarre’s definition of animation : “animation as moving images” [p.ix]. This definition suffers from two fundamental problems. First, it is far too general : “moving images” could very well be the definition of cinema, and there would then be no way of differentiating between the two mediums. Second, and most importantly, it is too ambiguous : moving images could at the same time mean the act of changing the spatial position of images, or images that move by themselves. To clarify the difference, I think it necessary to introduce a distinction between the mobile image, that is the image that is moved by an objective and external force or agent, and the moving image, that is the image that is compelled to move by its own internal subjective energy. Whereas Lamarre’s insistence on the internal force of the animetic machine might make us believe that he is talking about moving images, his focus on compositing as the key component of animation in fact reveals that his object is mobile images. If instead we define animation as “a media form that is created one frame at a time” [Condry, 2013, p.9], or even as movement created frame by frame, the movement we are looking for becomes a factor of continuity between stases of immobility : what we are looking for is then concrete dynamism (the “art of the hand”) and not a soulless technological force (to borrow the title of Ian Condry’s book, The soul of anime).

To illustrate my points, I believe it necessary to reconsider and reevaluate the legacy of Japanese character animation ; this is where sakuga becomes an essential study tool, as it gives us the much-needed care and appreciation for animators and their work. I will therefore proceed to a study of animator Yoshinori Kanada, who is both a paragon of limited animation techniques (4) and arguably one of the most important character animators in anime history.

The most common reading of Kanada’s animation and style is inspired by Japanese contemporary artist Takashi Murakami. Starting an art movement called “superflat”, he sought to root it both in traditional Japanese art and animation.  

The idea is quite simple, and was reinterpreted by Lamarre : according to Murakami, the main visual characteristic of anime is the flatness of the images ; a flatness present in the Edo-era ukiyo-e style of wood print which also rarely used “Western” Cartesian perspective. It is mostly in Kanada’s work that Murakami sees a prolongation of such a tradition ; and, most importantly, it is in Kanada’s effects animation. One of Kanada’s most well-known technique is that of the “Kanada dragon” : smoke, fire or lightning taking the shape of flying dragons. The important point here, for both Murakami and Lamarre, is that feelings of movement and depth are not created through geometric perspective or space, but only line and color : in this masterpiece of Kanada’s portfolio, the dragons’ movement and shapes are only given form thanks to the play between colors – black, red and yellow. This technique introduces in animation two key ideas : those of flatness (opposed to depth) and dehierarchization (opposed to the Western/Cartesian vanishing point). Kanada’s animation therefore opens a new relationship to space, representation and, if we follow Lamarre, technology.

However, Lamarre strongly critics Murakami and the two theses must not be mixed up. And I have to note that I completely follow Lamarre’s account here : “Murakami thinks entirely in terms of the structural composition of the image. He has little to say about animation as movement. […] The emphasis falls on techniques of image composition rather than animation and the force implicit in the moving image as a mechanical succession.” [Lamarre, 2009, p.112] But then, in what follows, Lamarre mostly discusses Murakami’s uncritical stance toward technology and militarism ; in other words, he never directly engages with Kanada’s animation, therefore implying that what is problematic in superflat is its general theory, and not its local analyses. While he critiques Murakami’s lack of discussion of animation, I do not believe that he provides an alternative look at it.

While I aim to do such a thing, I, as well, do not believe that Murakami’s (and Lamarre’s) understanding of Kanada’s effects is wrong : there is, indeed, an unavoidable play on depth and space here. However, focusing only on the “superflatness” of Kanada’s effects appears as a very poor appreciation both of Kanada’s work and anime in general. Indeed, Kanada was also a prominent character animator, who engaged with notions like space and depth in other manners that the superflat. 

First, in the realm of effects, it must be noted that Kanada dragons are not systematically associated to superflat techniques. Most notably, in this cut from Laputa : the Castle in the Sky where the dragons are made of lightning instead of fire, they are precisely made to emphasize depth rather than suppress it, and are not antithetical with (albeit somewhat distorted) vanishing lines. It is by the way most significant that Kanada, who frequently worked with Miyazaki, almost only worked on scenes of flight or dynamic 3D movement. In the same vein, one must recognize that one of Kanada’s central techniques is background animation, something he explored more than ever in his passion project, the OVA Birth. Here, what Lamarre would call an “optimization” of Cartesian perspective is pushed to its limits in an exhilarating chase. And it is important to note that, against readings that would put anime’s flatness at the forefront, and therefore insist on the proliferation of Kanada dragons, even in Kanada’s so-called “school”, that is animators who took after him, background animation remained a prized technique. What instantly becomes visible from this simple catalogue of cuts is that there is no univoqual reading of animation’s and animator’s space : the space of the image is first and foremost built and dictated by the animator according to the object represented, the effect sought, and probably something like individual inspiration and adaptation to storyboards or layouts – even though, in the case of Birth, the layout was entirely Kanada’s.

Character animation properly speaking is entirely missing from both Murakami’s and Lamarre’s account ; and this is probably what makes them miss actual movement, because movement is not just about composition or space : it is those two parameters considered dynamically, in regard with time. And time, or more precisely timing, was one of Kanada’s most distinctive traits. Indeed, in the anime production process, it is the individual animator which decides the timing of his cuts, and most notably, the number of in-between frames between each key frame ; and Kanada is famous for the highly irregular or unusually low number of in-betweens, something that then emphasize key frames. In other words, movement is created by a paradoxical emphasis on stasis. This technique has been overemphasized in later Kanada-style animation and became a staple of Hiroyuki Imaishi’s style (among others) : in this cut, for example, movement is not conveyed through the addition of in-betweens, but by the striking pose adopted in each new frame – as if each frame was a key frame (5). This creates a movement that is far more dynamic, while staying consistent with limited animation, superflat and open compositing techniques (especially considering that pull-cells are another staple of Imaishi’s directing style). Considering timing, which is not just a mechanical part of movement, but relies heavily on the animator’s own eye and artistry, therefore shows that limited animation is not antithetical to movement, character animation, or expression.

Moreover, this apparent detour by time allows us to understand space and depth in another, more dynamic manner. Indeed, it makes us understand how the famed “Kanada poses” create their own sense of spatiality, which I would argue is close to, but not reducible to, the superflat space. My main reference here would be this cut from Galaxy Cyclone Braiger’s OP, another one of Kanada’s masterpieces in both effects and character animation. What’s striking here is the centrality of the characters, who drive all of the first cut with their movement in depth (from the front to the back of the image) and their striking poses. Moreover, this movement in depth is created by the characters themselves : they are the ones establishing space, which is not a preexisting dimension. For example, if you look at it frame by frame, you can see that it is the first character shooting which apparently materializes the two other characters ; in the same vein, the square monochromes that appear behind the characters to emphasize them are in fact their shadows which first take an abstract, triangular shape, then take the form of a square. In other words, it is character poses which generate spatial composition and movement, and not compositing or cinematography. This is both an animator and character-driven animation, and not the result of pure, abstract, force.

The same could be said of Ichiro Itano’s “circuses” ; I believe that, in their analysis, Lamarre is closer to a real approach of movement, most notably in his reading of the distortion of classical Cartesian perspective and the “use of movement to generate a sense of depth rather than the reverse” [Lamarre, 2009, p.130], which is precisely what I just argued for. However, he immediately comes back to his technology-centric discourse and tries to reverse Itano’s apparent perception of movement : “Instead of movement into depth, you have movement that generates an exploded view of movement on the surface of the screen. […] Density of information, a sense of tightly packed elements with potential depth, begins to take precedence over movement within a world.” [pp.132-134] Here, Lamarre in fact focuses on just one of the two aspects of the Itano Circus, that is the impression that each missile is moving independently (6) ; but he misses the other key aspect, that is the feeling of being at the center of action. The genesis of the Itano Circus is almost legendary : it comes from animator Ichiro Itano having attached fireworks to his bike, an adventure he barely got out of alive. This is this highly subjective experience which then inspired the technique, whose core idea was for the animator to go “inside the action” and “think himself into the places where no camera has been before” [quoted in Clements, 2010].

Once again, then, the individual animator’s experience and art are key to understanding what is being seen and experienced : that is, movement, and not just any kind of it. Such an analysis should probably be pushed further, because sakuga, as I said, cannot just be an uncritical admiration of individual animators-auteurs ; however, this kind of poor understanding can only be avoided by looking further into the concrete aspects of production, instead of looking at it from afar, as an abstract process. This does not mean that we should relinquish Lamarre’s theory or methods : on the contrary, they are as necessary and insightful as ever ; however, we should admit that they do not totalize the experience of the animated image, and we should therefore endeavor to develop a complementary approach.

Notes

(1) I have given another quick summary that does not do any more justice to the complexity of Lamarre’s thought here ; you can also watch these two interviews of Lamarre himself : https://www.youtube.com/watch?v=6GpUQ42qtRA and https://www.youtube.com/watch?v=T6HSrR0MHvE

(2) The paradigmatic figure might then be not the animator, but the production assistant who moves from animator to animator to collect the cuts

(3) The same could be argued of genga, which means “original image”, and would refer more to the individual’s work and creativity ; but it is not the case of douga, which only means “moving image” without specification about the origin of this movement. This is the same difference I noted in an earlier essay between the “sakuga system” and “charisma animators” system

(4) Which is why I do not study the fathers – arguably – of Japanese character animation, that is Yasuji Mori, Akira Daikubara, and Yasuo Otsuka, considering that they all come from a full-animation tradition, that of Tôei Dôga

(5) This description is not far from Mitsuo Iso’s “full limited animation”, which aims as well to erase the distinction between key and in-between frames by having the key animator draw even the in-betweens ; the fact that it is common to Kanada-inspired stylized and Iso’s realistic animation would tend to indicate that this might be a common technique in anime, and one of its distinct stylistical traits. Lamarre also uses the expression “full limited animation”, but it has a completely different meaning and has more to do with character design than animation – I don’t even know if Lamarre knows about Iso’s terminology [edit : I have recently learned that the terminology “full limited animation” isn’t used in Japanese, but has been introduced by Ben Ettinger on his website Anipages. The question still remains if Lamarre knows about it or not]

(6) Which I would describe as each element of the image having intentionality, rather than as creating potential depth, even though both views are, here, not incompatible

Bibliography

Chung, P. (2007) “Japanese Animation Theory”. Retrieved from http://www.pelleas.net/forum/viewtopic.php?f=1&t=238&start=15&sid=e6d1e73d062eb1d1229ee4f121dc2d7

Clements, J. (2010) “Entering the Itano Circus”. Retrieved from https://schoolgirlmilkycrisis.com/2010/11/14/entering-the-itano-circus/

Condry, I. (2013) The Soul of Anime. Collaborative Creativity and Japan’s Media Success Story. Duke University Press.

Ettinger, B. (2011) “The anime production line”. Retrieved from http://www.pelleas.net/aniTOP/index.php/the-anime-production-line

Lamarre, T. (2009) The Anime Machine. A Media Theory of Animation. University of Minnesota Press.