When I read The Elephant in the Brain: Hidden Motives in Everyday Life by Kevin Simler and Robin Hanson, I felt what Hanson would call a “viewquake,” an insight which dramatically changes my world view. I recommend reading it in its entirety for the full exposé of our hidden motives, but I’ll summarize the relevant information for this post. Simler and Hanson devote a chapter to the hidden motives behind art, why we do it, and why we like it. Anime is a form of art I enjoy, and I often see debate¹ about the quality of anime decreasing pointing to various factors like Japanese vs overseas sales, the exporting of animation work done by non-Japanese studios, or even the transition from cel animation to digital animation decreasing the barriers to entry of for creating an anime; the overall difference is sometimes talked about as anime losing its “soul”. I think these are matters of taste that can be viewed as status claims. I’m also a fan of buddhist philosophy and its soteriological methods like meditation and trying to be a scientist in observing your own thoughts² to see the interdependence of all things and that the self is simply a fiction. So, I think the explanations in The Elephant in the Brain can help steer one toward enlightenment. While I write about Buddhist meditation and mention anime history, I am an expert in neither, but my favorite kind of article articles that which ties together knowledge from different domains; this is an attempt to do that.

The Elephant in the Brain

The introduction begins:

elephant in the room , n. An important issue that people are reluctant to acknowledge or address; a social taboo.

elephant in the brain , n. An important but unacknowledged feature of how our minds work; an introspective taboo.

with their thesis:

We, human beings, are a species that’s not only capable of acting on hidden motives—we’re designed to do it. Our brains are built to act in our self-interest while at the same time trying hard not to appear selfish in front of other people. And in order to throw them off the trail, our brains often keep “us,” our conscious minds, in the dark. The less we know of our own ugly motives, the easier it is to hide them from others.

Signals

What we don’t know yet is what these hidden motives exactly are and how they influence our actions. Often, these motives are to win allies³. We, as political animals, want to to woo others to further our place in the societal ladder so that we can gain concrete things like physical security and mating prospects. We do that in a variety of ways known as signals (“anything used to communicate or convey information”). Honest signals are said to “reliably correspond to an underlying trait or fact about the sender” and are otherwise deceptive or dishonest. We use and rely on signals all the time, but the most apparent ones, words in language, are cheap and therefore not as useful. My telling you “I’m a reliable person” carries much less weight than if you lost your job and I gave you a room to sleep in and helped you find a new job. In short: actions speak louder than words. Signals are also helpful, as we will see, in indicating status. We always want high status.

Art and Guilty Pleasures

Simler and Hanson devote a chapter to exploring art, beginning with noting how costly it is. Creating art can take up an incredible amount of time and energy. Art is impressive (i.e. signals fitness) in part because it is so costly. The more time one spends doing art and not doing other things important to survival, the more fit one seems, not to mention other traits like creativity, fine motor control, etc. one signals by creating art. These are the extrinsic properties of the artwork. These extrinsic properties cannot be derived from the work itself e.g. originality, how long it took to make, etc. The intrinsic properties in the artwork are what you see in, for example, a painting: colors, textures, etc. In the fitness-display theory of art, we allocate a much larger portion of an artwork’s value to the extrinsic properties than the intrinsic properties⁴.

We spend a lot of time honing our faculties to appropriately judge art because it helps us figure out what is high status and what is low status. We care so much about the extrinsic qualities because when we pass judgment on art, we are (at least in part) also passing judgment on the artist. If we praise a low-skilled artist and discover that they were low-skilled, we can feel embarrassed and even lose status among our peers. Think of how our opinion of a piece of furniture changes when we are told it was hand-crafted by an artisan rather than mass-produced and assembled on a production line. Simler and Hanson note that our focus on being able to judge so well suggests that we evaluate each other not only for our first-order skills, but for our skills at evaluating the skills of others. Think of foods, movies, series, etc. that are “guilty pleasures.” We acknowledge their low status (hence guilty) but note how we actually enjoy them. We can talk about them in the context of being a guilty pleasure but hesitate otherwise as it might appear that we consume media and think highly of it and the artist.

The cost of art on the part of the artist is important because it indicates the fitness of the artist. Waste is a feature (not a bug) of art, so technological advancements that increase certain intrinsic properties of art but are “easier” can backfire when they are known to be easier. That is, artists, pursuing technological advances, may lose status.

The advent of photography wreaked similar havoc on the realist aesthetic in painting. Painters could no longer hope to impress viewers by depicting scenes as accurately as possible, as they had strived to do for millennia. “In response,” writes Miller, “painters invented new genres based on new, non-representational aesthetics: impressionism, cubism, expressionism, surrealism, abstraction. Signs of handmade authenticity became more important than representational skill. The brush-stroke became an end in itself.”

Anime’s Lost “Soul”

In typical discussions on the decrease of anime “quality” over time, people will point to various intrinsic properties of various anime like the color palettes, character designs, general animation “quality”⁵, etc, comparing scenes from typically older, cel-animated anime to modern, digital anime. When people make aesthetic judgments, they will often cite the extrinsic qualities like originality or how much time and effort was put into a particular scene, episode, or series as a whole. Thus, we will focus on these extrinsic qualities and how the intrinsic relate. Animation is a complicated process, but because the discussion focuses on how anime has lost “soul” over time, it is especially important to note how the extrinsic properties have changed over time.

Blame the Computer

For much of the 20th century, animation was laborious and time-consuming. It required all sorts of special equipment and labor from skilled artists and technicians. You not only had to have a wide variety of paints but also needed to properly mix them and mix enough so as not to run out of your mixed color. You needed to hand-paint cels with a particular technique. You needed You needed technicians who knew how to operate the complicated cameras. You also need to keep track of all those cels.

Nowadays, much of the work can be done digitally. There is no need to mix paints, mistakes can be more easily remedied, there is no complicated, heavy, mechanical camera to operate, and there are no physical cels to keep track of. Computers just make most things easier, and specialized software can further reduce the costs of creating anime. Note, creating an anime is still a very difficult and costly process, but the point is that relative to before, it is less costly now to create the same or similar works. So, when we see people attribute high status to cel anime and low status to digital anime (when they claim older anime is better than newer anime) they might just be making a point to show their status since cel-animation is more difficult to create and, particular to anime due to its rise in popularity in the west, may indicate an ability on the critic’s part to be “ahead of the curve” when it comes to appreciating art if they appreciated it when it came out, or it may indicate the critic’s commitment to the medium (they went through the trouble of watching what came out before they could have shown an interest and don’t watch just what is currently airing).

Brutal 3DCG

We encounter a situation much like that of painting and photography Simler and Hanson mention. However, the intrinsic differences introduced to anime production by the computer are not as visible as those of photography to art⁶. In order to maintain the status of cel-anime, new anime must revert to the costly physical techniques of cel-anime, make its intrinsic qualities such that it appears cel-animated, or fully embrace the computer and create an end-result that is so different or impressive that it clearly shows how costly it was to make. Indeed, often anime producers do take advantage of some benefits afforded by digital means of production: more precise control over the colors and interchangeable three-dimensional computer generated (3DCG) models to be reused with different camera angles. Much like the brutalist style of architecture Simler and Hanson mention, 3DCG has become quite disliked among some anime fans. And still like burtalism, as you learn more about the process, there’s more to appreciate: like the creative means by which certain effects are achieved. Now, this isn’t to wholly discredit intrinsic criticisms of using 3DCG in anime; sometimes things do just look bad for many reasons⁷, but it hopefully sheds light on how we may dislike or argue about intrinsic qualities when we may really disagree about how the extrinsic qualities of the work and how much effort the artists appear to exert. Just as we’ve seen these arguments before, we will likely see more future complaints about machine learning-enhanced artwork. I predict, if we see use of tools to increase the framerate of an anime series or movie, we’ll see criticisms of its use similar to those 3DCG faces.

Enlightenment

Simler and Hanson actually touch on the consequences of becoming aware of the elephant in the brain:

In fact, knowing about our own blind spots should make us even more careful when pointing fingers at others. After all, many of our perceptions are colored by self-interest, including our perceptions of what other people are up to. […] The next time you butt heads with a coworker or fight with your spouse, keep in mind that both sides are self-deceived, at least a little bit. […] Above all, what the elephant teaches us is humility. It’s a call for more thoughtful interactions with our fellow self-deceivers, a spur to step outside our own conniving minds. […] Another benefit to confronting our hidden motives is that, if we choose, we can take steps to mitigate or counteract them.

They even introduce a philosophy of “enlightened self-interest”:

the notion that we can do well for ourselves by doing good for others.

Meditation

Typical Buddhist meditation often focuses on becoming ‘aware’ of not just what is happening around us but especially on what is happening within our minds. Meditation is not just sitting cross-legged with your eyes closed and focusing your mind on your breath. That is useful, but we forget other forms of meditation:

[…] be aware and mindful of whatever you do, physically or verbally, during the daily routine of work in your life, private, public or professional. Whether you walk, stand, sit, lie down, or sleep, whether you stretch or bend your limbs, whether you look around, whether you put on your clothes, whether you talk or keep silence, whether you eat or drink, even whether you answer the calls of nature — in these and other activities, you should be fully aware and mindful of the act you perform at the moment. […] Then there is a form of ‘meditation on ethical, spiritual and intellectual subjects. All our studies, reading, discussions, conversation and deliberations on such subjects are included in this ‘meditation’. To read this book, and to think deeply about the subjects discussed in it, is a form of meditation.⁸

This is akin to a kind of self-help meditation we can perform every day. In noting our own emotions when they flare up e.g. when you get angry, simply saying or thinking “anger is happening” can be quite useful in becoming aware of that anger and realizing the futility of that anger. Doing this distances the anger to a point where we can more objectively understand it. When we encounter frustration towards people not appearing to behave as we expect them to, remembering the elephant may help. That is, we can look to a hidden motive of which they are likely unaware and see how their actions may seem more reasonable in this new light. We should therefore have compassion and patience with others.

Non-Self

It follows we should be aware of especially our own motives to properly see the world, particularly the mind, as it really is. Becoming aware of the possibility of these biases and knowing their form helps us to begin looking for them in our daily lives. If we better understand our motivations and the actions that result from them, then we can more objectively understand how to change our behavior to better orient ourselves towards enlightenment. As motivations are part of our mind, knowing them and how they arise and cease can helps us see the arising and cessation of consciousness itself. This in turn, helps to detach one’s mind from itself, that is, understanding how the self is a useful fiction.

This is a common way of talking about the self in Buddhist philosophy. Briefly, the main idea is that what we understand as the self or ‘I’ isn’t real (hence ‘fiction’); rather, what we call the self is really a collection of all sorts of ever-changing things (the precise nature of these varies depending on who you ask). It’s useful because it’d be cumbersome to refer to each other as “causally linked collection things” rather than, say, “Steve.” It is useful to realize no-self as it is one of the falsehoods we cling to most vigorously. It prevents us from seeing clearly the world as it really is. This is simultaneously a metaphysical claim (the exact nature of which varies for different schools of Buddhism) and soteriological method for achieving enlightenment. Many teachings in Buddhism function in this way, especially interdependence.

Interdependence

Seeing the causes of our actions (often, our motives) and then discovering the causes of those motives is an exercise in seeing the interdependence of all things. The basic idea of interdependence is rather simple: all things are dependant and all things influence each other. The metaphysics of this idea get complicated and, again, vary according to the school of Buddhism. Soteriologically, we should understand that we are dependent on all sorts of things. From the food we eat causing changes in our physical appearance and even mental states, to the friends we have influencing our tastes in music, movies, and more, we are inescapably dependent on other things. We should also realize that we influence others as well, therefore, we should be more mindful of our actions as they have effects we often don’t notice. In true Buddhist fashion, we should also realize how interdependence plays a role in realizing how the self is a useful fiction. Interdependence clearly rejects the self as some uncaused causer. Even if we reject that the self has any physical elements, we have to see how our mental states are dependent on our physical conditions as well.

A really simple way to think about independence is as karma. The Mahayana Buddhists used karma to stress the interdependence of all things and the importance of our own actions in shaping the world⁹.

Conclusion

We have motives and hide them even from ourselves. We can find them if we look, and they often hide behind signals. They can even hide behind our opinions about art. When we praise or dismiss a work of art, we might really be passing judgment on the artist and signaling things about our own status to others. When it comes to anime, I think we see elements of this kind of behavior from those who praise cel-animation and dismiss contemporary, digital anime. Rather than be dismayed by these hidden motives, we see that better understanding these motives can help bring us closer to enlightenment. Now, as Simler and Hanson note, we also shouldn’t delude ourselves into thinking the elephant can convincingly explain all individual behavior (beware seeing only nails), but it’s still a useful tool to have in our back pockets.

Appendix

Contemporary Examples

Noodle’s April 16, 2021 Video

Youtuber, animator, and essayist Noodle made a video called Smoother animation ≠ Better animation [4K 60FPS] making some criticisms of technologically-assisted animation. This also gives me an excuse to write on signaling again.

Noodle opens against a recent trend of naively applying digital effects to art (e.g. using interpolation to increase the framerate of animation) by showing that the framerate is often used as a deliberate choice by the artist for a number of reasons including conveying a specific meaning or matching a certain aesthetic. He then claims that the act of modifying another’s art in such simple ways in general is bad because it goes against the original artist’s intent and violates some ‘rules’ of animation.

The former, specific, descriptive statement of framerate being a stylistic choice and then claiming that a higher framerate does not always indicate a higher quality animation makes sense and works. Framerate, ceteris paribus, is a metric that directly corresponds to how much effort was put into an animation by an animator or a team, so it makes sense why many see increased framerates as increased quality. Most people are unfamiliar with the animation process and interpolation, so even if it is done through simple methods requiring little effort, an interpolated animation may seem impressive.

With regard to the latter statement on how using these methods to modify existing work is bad: this looks a lot like when people accuse artists of “tracing” another artist’s work and then adding their own coloring, shading, etc. This is usually done without explicit reference to the original artist’s work¹⁰. Accusations like these can be damning for artists. Typically, the main accusation is that of low originality with low effort playing an important but secondary role. As shown above, lower effort means lower status. This video, like accusations of tracing, can be seen as trying to lower the status of those who create or modify animation through the use of interpolation but instead of focusing on lack of originality, it focuses on the low effort of interpolation. Noodle’s pointing to the fact that currently humans get paid to draw inbetweens and saying that AI draws “boring, lifeless” frames with his hesitant endorsement of specialized inbetween creation software (it seems he only endorses it as long as animators “remain in control”) is more evidence that this is mainly a bias a la more effort means ‘better’ art.

Later in the video, Noodle references the fundamental principles of animation from Disney Animation: The Illusion of Life claiming that AI that interpolates animations ignores most of these. While AI may ignore them¹¹, human animators may stretch and squash the principles as well. He then shows some examples of normal animations vs their interpolated versions. His critiques are mostly on the inbetweens (the frames between the key poses of what is moving), showing that they do not look as clear as a typical animation’s inbetweens would look, even if the inbetweens are “smear frames” (inbetweens that greatly distort or “smear” the animated object). He pauses or slows down animations to make this point. He even mentions that a viewer might not “see” this but may instead “feel” the difference. This is what I don’t buy. If someone can’t tell the difference in the animation displayed at the appropriate speed (without slowing or pausing to emphasize AI-generated inbetweens), just as if someone can’t tell the difference between an imitation of the Mona Lisa, then there aren’t any intrinsic differences in the person’s experience of the art, so because they don’t see and don’t feel any difference, there simply isn’t a relevant one!

With regard to his point about violating the original artist’s intent: that’s true! The original artists often have a vision for their work and wish to convey something with it. By increasing the framerate with interpolation, one creates a new work conveying a new meaning. Think of the new work as more like fanart with a new style, trying to appeal to people in a different way (thinking this way leads to less frustration). The main problem (from artists’ perspectives) seems to be when the modified work gains a higher status than the original work, but that seems unlikely to happen since if tastes change enough to prefer and reward interpolation, original creators can release their work already interpolated (How long can aritsts and consumers’ tastes differ? Is the market for interpolated videos only big enough for hobbyists on youtube?). Also, as Noodle mentions, even 3DCG movies are adding smear frames to their animations (rather than cheaply use interpolation) and newer anime don’t seem to take advantage of interpolation either despite taking advantage of plenty of other computer-assisted techniques to create animation.

The Intrinsic Problems of 3DCG in Animation

While I point to criticisms of extrinsic qualities of 3DCG in animation that pose as criticisms of their intrinsic qualities, there are intrinsic qualities that make it noticeable and make it sometimes worse at fulfilling some reasonable goals of art than traditional animation. My initial attempts to answer here are provided without much research and are purely from the perspective of a fan of animation.

Creating a Visual Language

Traditional animation as a medium differentiates itself from other previous media by being animated. Obvious but important since people’s visual intuitions are accustomed to the physical world, especially when it comes to objects that appear to move, so when creating works in this new medium, animators must use people’s expectations of how real objects should move to inform how a drawn object must move. In The Illusion of Life: Disney Animation, one of the fundamental principles of animation is “Squash and Stretch.” It is noted as “by far the most important discovery” in animation:

When a fixed shape is moved about on paper from one drawing to the next, there is a marked rigidity that is emphasized by the movement. In real life, this occurs only with the most rigid shapes, such as chars and dishes and pans. Anything composed of living flesh, no matter how bony, will show considerable movement within its shape in progressing through an action.

The next fundamental principle is “Anticipation.”

People in the audience watching an animated scene will not be able to understand the events on the screen unless there is a planned sequence of actions that leads them clearly from one activity to the next. They must be prepared for the next moment and expect it before it actually occurs. This is achieved by preceding each major action with a specific move that anticipates for the audience what is about to happen.

This anticipation can be as small as a change of expression or as big as the broadest physical action. Before a man runs, he crouches low, gathering himself like a spring, or, the reverse, he draws back in the opposite direction, raising his shoulders and one leg, as he aims himself at the place of the next activity.

The opposite is a “surprise gag,” which only works when the audience is expecting one thing to happen, and suddenly, without warning, something entirely different happens. The surprise gag cannot work if a different action has not been anticipated by the audience. Similarly, no action on the stage can be anything but a series of meaningless surprises without anticipation.

Another fundamental principle is “Follow Through and Overlapping Action”:

When a character entering a scene reached the spot for his next action, he often came to a sudden and complete stop. This was stiff and did not look natural, but nobody knew what to do about it.

There are about nine more principles, but they are similar enough for my purposes. The point of most of them and the ones shown above is to instruct the animator on how to create a visual language that is understood by the audience, and to understand that visual language so they can accurately predict the audience’s reactions to and interpretations of their animation. In philosophical terms, the animator creates a new world, or ontology, for the audience, and thus must create a physics for them as well. Perhaps if someone had never seen an animation, that person might feel confused as to why Wile E. Coyote from his namesake show can run a tens of feet off a cliff without succumbing to gravity, be propelled through a rock and come out the other side unscathed, or be blown up by dynamite and suffer only light aesthetic damage; however, after probably just a few scenes, one understands the physics and can anticipate the coyote’s reactions to and physical damage incurred as a result of the road runner’s hijinks.

Nowadays, this happens in live-action media with CGI effects as well as in animation¹², but the process is the same. The first few scenes or even episodes of a show establish the visual language and set the expectations for the work going forward.

The audience’s interpretation of visual language in animation, like natural language in novels, is influenced by not just their real-life experiences and the current work they are consuming (the current animation they are watching) but also by all the works they have consumed in the past (other animation they have seen), so just as authors must write to make themselves properly understand by their audiences by using vocabulary relevant to their contemporary audiences so too must animators animate to accomodate the visual vocabulary of their contemporary audience.

Unwelcome Surprises

Here, I think, is the most important way in which 3DCG mixed with traditional animation becomes importantly distinct. Breaks from the visual language in traditional animation are typically done intentionally to surprise the audience (e.g., to create a visual gag that makes the audience laugh). With 3DCG, audiences are surprised without any gag or reason. I say this is simply because 3DCG is relatively new.

Audiences are less used to seeing 3DCG in animation compared to traditional animation, especially when there are 3DCG elements mixed with traditionally drawn elements. Because audiences are less used to it, it stands out more. If it stands out without specific intent (that is, without a visual gag), it simply looks how a foreign or esoteric word would in a novel: out of place.

Because 3DCG is new, especially in the context of trying to mix 3DCG elements with traditinally drawn elements, animators are also less skilled at properly utilizing 3DCG elements to create a consistent visual language from which the audience can anticipate.

As animators more often use 3DCG elements mixed with traditionally drawn elements, audiences will notice the 3DCG elements less. Animators will also become more skilled at properly animating with 3DCG elements and better create a consistent visual language audiences can understand.

Notes