Transcription
For a change of pace, how about we try a little experiment? It’ll be easy, I’ll just show you a couple of characters and we’ll see if we can guess what they’re like purely based off of their character design. Sound good?
Okay, first up. She’s cute! Her eyes are large and have a focus on this sweeping downward curve as opposed to being angular—making her come off as gentle and sweet to me. She’s also got a crazy amount of hair, but it doesn’t seem in any way unkempt. In fact, this braid and these accessories show someone who cares about their appearance, or at least has a parent who does. Finally, her clothes are relatively plain and simple. This pink frilly top almost looks like it could be homemade or sewn together.
Alright, how’d we do? Well this is Kokona from Yama no Susume! She’s a shy kind girl that eventually joins the two energetic mains, Aoi and Hinata, as they explore the world of hiking. Unfortunately her family is struggling financially, causing her not to see enough of her constantly working mother. However, they go out of their way to show how much they care for each other when they can. So, what do you think about my super character design analyzation skills, impressive? Wait, what do you mean “not really?” Anyone can retrofit known traits onto a character whose show they’ve seen? Why I never.
Alright alright, how about this. I’ll choose a random character from a show I’ve never heard of, and we can look it up later to see if I’m right. Sound good? I promise I won’t cheat. Oh, she looks nice. Overall she has a gentle sort of look to her. The color palette leans very soft and less saturated than usual. She’s wearing a vest and an ascot looking thing, but she doesn’t seem too concerned with being prim and proper. It looks like she might work in a restaurant, or be in a band. In the context of a show, I don’t think she’d be the main character, but a friend or supporting one with a side arc maybe.
Okay, how’d we do? Just give me a second to stick this into Google Reverse Image Search and… no results? Alright, I’m sorry, I stuttered. This character actually isn’t a character at all. She was generated using a deep learning StyleGAN generator, or if that confused you much as it did me, a fancy artificial intelligence. Basically, this waifu does not exist.
I came across this website a while back, and I was just as fascinated as I was with its human faces counterpart. And the thing that most interested me was that, even though these faces were generated by a program, I could *see* the characters in my head. They all had their own vibrant personalities, affects, and airs about them. Okay, maybe not all of them, some of the products are pretty gnarly. That got me thinking… why was this? How come these characters that were randomly assembled by a mindless machine nonetheless seem to have minds of their own? Essentially, how do these traits get created, adopted, and incorporated into these characters and constitute meaning? What makes a waifu, a waifu?
Well, we can try to take some inspiration from biology. Evolution posits that organisms adapt to their environments over time, gaining new traits that help them survive while the less fortunate are lost to the sands of time. After some drastic beginning changes, there tend to be more fine improvements that eventually leads to harmony with their environment. Birds get new beaks, amphibians grow strong limbs, and so on and so on. What of the waifu? Well, it *sort of* fits? However, unlike a lot of habitats, the anime climate is constantly changing at a rapid pace. Trends come and go, top waifus are crowned and dethroned in the blink of an eye. This year’s Lum becomes next year’s Rem becomes next year’s Nezuko (which are all fine choices by the way). I believe there’s some hint of adaptation, but evolution is inherently linear. You can trace a clear througline from common ancestor to common ancestor. But can you say that’s really true of waifus? Who is the common ancestor of Zero Two, or Asuna?
Hiroki Azuma, notable Japanese media theorist, noticed this incongruency. He is the author of Database Animals, a work discussing how we consume anime, and in it he questions the idea that waifus have clear spawn points (more elegantly phrased of course). Unlike evolution, where in theory you could easily trace back the origin of a certain trait, waifus obey by different rules. Instead, he puts forth the idea of the database. Think back to our little experiment that started off the video. We used (my very pitiful) knowledge of character design to try and guess at how certain characters would act. A lot of people don’t realize that a character is a story in and of themselves. But rather than being told through writing, it’s the little things. The way their clothes move in the wind, the way they adjust their hair—down to the size and shape of their eyes. Azuma took these various tenets of character design and deemed them “moe-elements.” Instead of traits evolving and appearing over linear time, there’s a bit of a “build-a-waifu” system in place. As characters are created and defined, they are broken into their base elements and incorporated into the mysterious “database.” Then, in the future, characters are born out of this primordial stew of moe-elements, often subconsciously. Azuma argues that in essence, there is no longer a need to have a character be from a series with a strong narrative, or be compelling in and of themselves. Best Girl and Best Character have been separated—the “idea” of them and ability to evoke “moe” is enough to propel a character into waifu status.
The database model suggests that over time, the priorities of otaku have changed. Azuma uses the examples of Mobile Suit Gundam and Neon Genesis Evangelion to illustrate this. When you watch any iteration of Gundam, you’re buying into a long reaching fictitious history, driven forward with the undercurrent of a grand narrative—the draw is figuring out its worldview and seeing everything through that lens. Over time however, fans began to lose interest in this kind of offering. With the importance of the grand narrative declining, characters instead became the focus. If we look at NGE, its plot is largely inconsequential. It exists, and “stuff” happens, but I’d be hard pressed to tell you what that “stuff” is, and it’s one of my favorite series. Rather, it serves as a means to an end. Its setting is simply a delivery vessel for the characters, and the characters themselves are delivery vessels for moe-elements. Evangelion is a show made by otaku for other otaku. Not to say the characters can’t be deep or well developed, the fact that they’re so obviously archetypes is part of the appeal—along with the show’s self-awareness of that fact. However, Asuka and Rei are almost hand-crafted to appeal to otaku, incorporating many traits recognizable to them. These traits are then broken down and recycled, destined to appear again. You can find evidence of this throughout the anime fandom. I can’t count how many times I’ve downloaded someone’s picture or put a show on my plan-to-watch simply because their design appealed to me. (I also can’t count how many of those were Fate/GO characters, but let’s not talk about that). We live in a climate where any of your particular niche interests are but a quick search away, an endless sea of different iterations on the concept. It is often the case that the characters are created first, the small narratives and settings to house them later. What does this have to do with waifus? Well, simply put, your waifu can only come to be because of the thousands of waifus that have come before. Your waifu is all waifus.
Now, this phenomenon is seldom described favorably. “Animalization,” after all, doesn’t have the best of connotations. One could interpret otaku as voraciously consuming whatever crawls out of the seasonal soup. Instead of searching for some underlying deep meaning or worldview, merely the act of consuming, cataloguing, and managing the database is enough. Database Animals says that we have lost the human desire for something more, and are settling for the simple animalistic fulfillment of needs. But, as I write these words, I can’t help but be more optimistic. Yes, like you I am also tired of getting the millionth reskin of an isekai (which are almost entirely born from the database), or shows with school settings, or the many battle harems I will never watch. But at the same time, you can see many productions that are aware of this database structure, and make use of this awareness. Neon Genesis Evangelion humanizes the traditional tsundere character while deriding the idea of the perfect submissive Japanese wife, Haruhi harbors clear irreverence towards faithfully recreating school romcom tropes, and of course Monogatari might be the ultimate evidence towards characters usurping narratives. These are all shows that rely on a certain feeling of being “in-the-know,” and looking back to my first few times watching anime, I can see how far my knowledge and enjoyment have come since first catching Inuyasha late at night. It’s a bit like seeing the matrix. There’s a feeling of community that comes with understanding the database and the elements that get pulled from it. You can watch a show and just “get it,” every little understood reference granting a hit of dopamine.
Hideaki Anno was once quoted by a bunch of clickbaity articles as saying the death of anime was imminent. But what he actually said is that he thinks *Japanese* animation is decline, and will soon be usurped by foreign animators, describing their passion and energy. One might wonder if the database model detailed by Azuma, reflecting anime’s more incestuous, insular, self-referential nature, might be the cause of this. On one hand, we have gone from the Tezukas and Miyazakis of the world expressing their worldviews which just so happened to be through the medium of anime—to otaku creating works aimed at otaku. But as I said, I am optimistic. Azuma wrote Database Animals in 2001, and while many of his fears are well-founded, the industry has definitely grown and innovated. Studio Orange has done the impossible and made beloved CG anime, expertly intertwining traditional 2D animation with 3D modeling. Creators like Masaaki Yuasa have sprung forth to lead the charge, bringing with them their own unique design and worldviews reminiscent of days past. I hope that ultimately, we can have the best of both worlds. The feeling of belonging to a special group, of saying “oh, this is their take on a tsundere,” or recognizing references and influences when you see them. But also innovating out of this rut of bubblegum pop or popcornification of anime described by Azuma as a result of postmodernism. However, regardless of how the industry evolves, I have a feeling that waifus in their many forms are here to stay. And frankly, I think that’s just fine with me.
Thanks for watching! If you enjoyed the video, please click the like button. It’s the best way to support the channel, tells YouTube that this is a good video and to recommend it to more people. This video is essentially an attempt to distill some base components of Hiroki Azuma’s Database Animals into digestible pieces, and I hope I did it justice. If you’re to the task, I do recommend reading it for yourself. I can’t thank Joe from Pause and Select enough for looking over this script, and highly recommend their channel for more concepts like these. Apparently I lean more Ito than Azuma here, there’ll be recommended reading in the description. If you want to support me further, feel free to donate to my Patreon. Special thanks to Less Shy Taluun, Frisky Tipsy Whisky, Khor-kheh Cobbler, GamerQuake236, Fed Ted, OniStuck, and Yo Ho Rollimor for their support. And of course, if anything I said was wrong, I’m sorry. I must’ve stuttered.