The Moe Image

In this article I take a look at moe from the perspective of images as opposed to strict character types. I propose that there are four categories of moe images:

  • Junai-kei moe: images in which a loving, but not explicitly sexual relationship is depicted or implied between the moe heroine and the male viewer or his narrative proxy. Found most commonly in renai games such as those produced by Aquaplus / Leaf and Key.
  • Otome-kei moe: images in which male presence is deemphasized in favor of a scene which implicates the heroine(s) as the centerpiece of an idealized past or present. Found most commonly in (nonsexual) yuri such as Marimite and in anime / manga such as Kamichu, Kokoro Library, and Yokohama Kaidashi Kikou.
  • Erokawaii-kei moe: images in which the moe heroine is sexualized, to an extent limited by a) her innocence and b) her consent, for the benifit of the male viewer or his narrative proxy. Found most commonly in erotic games, manga and anime, and as such is often conflated with lolicon.
  • Denpa-kei moe: images in which style, fetish symbols and costumes crowd out significant narrative meaning, trading on the value of pure exhuberant cuteness. Found most commonly in mascot characters, the work of artists such as POP, and nonsensical anime (Digi Charat killing two birds with one stone here).

A common thread that runs through all of these is the image in which the heroine is first made aware of the viewer’s gaze, resulting in some form of embarassment; I’ve pressed the term kaimami into service to represent this phenomenon, although its original meaning is somewhat different.
That pretty much sums up the conclusions I’ve reached in the past few months of musing on the topic. If you’re interested in the circuitous route that brought me there, examinations of Akamatsu’s view on moe, the Moetry for You position, or any of several corollary issues and problematic elements of moe, by all means read on.
Note: in the interest of respecting artist copyright I’ve decided against including example images here that I don’t have express permission to use. Should a more visual discussion arise I would recommend the forums for that purpose.

Moe (萌え) is a term which has risen to commonplace use in the otaku lexicon over the past several years, and while its colloquial meaning is widely understood, its formal properties have seldom been addressed. This article proposes to delve more deeply into the otaku mind in an attempt to discover and expose the intricacies of moe appeal.
Moe’s current usage can be roughly equated with that of the more broadly defined verb 好き (suki; to like or love), although grammatically its position is more ambiguous; the phrase さくらちゃん萌え (Sakura-chan moe) can be parsed in several ways if we shunt a couple of particles around:

  • As a verb: [boku wa] Sakura-chan [ga] moe [desu] translates as “I (moe/like) Sakura-chan”.
  • As an adjective: Sakura-chan [wa] moe [desu] translates as “Sakura-chan is (moe/likeable)”.
  • As an interjection: Sakura-chan moe!! translates as “Sakura-chan love!!”
  • As a noun: Sakura-chan [wa] moe [ga nai] translates as “Sakura-chan lacks (moe/likeability)”.

These direct translations result in a reduction of meaning almost to the point of incoherence; the substitution of “like” for “moe” in all cases is obviously insufficient. My goal in this piece is to understand, based on an array of usage instances, the specific ways in which moe is rendered as a subset of taste (as denoting taste is the function of “liking”).
To do this, I will be focusing primarily on the term’s use as an adjective as it applies to characters as depicted in images and narratives. I contend that its usage as a verb is contingent on a subconscious understanding of a common moe referent system, which I will proceed to outline. Note that the history and etymology of the term are not part of this article’s purview, being both largely speculative in nature and well covered elsewhere; instead I will be focusing on its current construction as developed over the past decade.
Groundwork of Moe: Usage
Before we proceed, it will be useful to know a bit of sociological lingo: when we look at an image, we are subconsciously indexing it against what is called a referent system: a background set of properties, attributes, and associations that enable us to infer meaning from everything we see and hear (or otherwise experience). It is the moe referent system, existing in the mind of the otaku, that it is our goal to penetrate and understand. What are the laws that govern his perception, understanding, and appreciation of moe?
In other words, when an otaku declares his undying love for our hypothetical Sakura with the statement “Sakura-chan moe” (I like Sakura), what does he mean? What are the properties she possesses that make her attractive to him – that when viewed, cause him to melt into a puddle of gibbering goo? The understanding of these properties and their interplay with his network of subconscious associations will lead us to a fuller understanding of the moe phenomenon.
Because we are talking about properties first and foremost, and properties aren’t verbs, we must formulate moe as an adjective or noun. “Like” transposes awkwardly into “likeable” and “likeability” respectively, so it will be expedient to choose a more fitting term to correlate with moe: appeal.
Assertion 1
Moe is a type of appeal. When used as a noun, the word “appeal” can replace the word “moe” in a phrase with its meaning unchanged aside from the level of specificity lost when moving from a narrow classification to a broader one. Likewise, moe as adjective can be replaced by “appealing” with a similar retention of general meaning. Thus, Sakura wa moe ga nai becomes “Sakura lacks appeal” and Sakura wa moe desu becomes “Sakura is appealing”.
Like all forms of appeal, moe exists in the eye (and mind) of the beholder; that is to say, there is no moe without a moerer (viewer). Understanding what the viewer sees in an image he describes as moe will help lead us to a definition of the term. We’ll come back to this concept. A lot.
Assertion 2
Moe is a set of generally accepted stylistic and thematic elements. While they may not be explicitly defined, for the purposes of this piece we must assume the term does have some definition beyond “I know it when I see it.” I think it could be argued convincingly that within the sphere of anime and manga moe is in fact an art movement, but I don’t have the time or art history cred to go there myself. Note that this assertion is challenged, to a certain extent, by the original verb formulation of moe (see the Issues section below).
Assertion 3
Moe is a property of characters as depicted in images. These images are in turn informed by narratives (what is happening in the image). There is no moe without a character, there is no character without an image, and there is no image without a narrative (however sparse or nonsensical it may be).
When we speak of a moe anime, manga, or game, what we’re talking about is a work in which moe characters are depicted, and these depictions are informed by a narrative. Other narrative and structural elements of the work inform the moe image as well, the details of which we’ll get into later.
When speaking of an original image as moe, unrelated to any further sequential work, the narrative is derived from the content of the image in “a picture is worth a thousand words” fashion. This is done subconsciously, of course, as part of the lightning-fast mental process of interpreting meaning in an image. Identifying the narrative that informs a moe image is one of the keys to understanding its appeal to the otaku mind, and the ability to infer an image’s narrative is a key part of the moe referent system.
These three assertions (that a. moe is a form of appeal, b. it has generally accepted traits and c. it is a property of characters as depicted in images) are central to the continuation of this analysis; if you disagree with any of the three, try to bear with me until the Issues section, where I’ve attempted to address some of the problems with this approach.
Now that we’ve agreed on the basic usage of the term, it’s time to dive into the murky depths of the otaku mind and look through his eyes at a moe character. What is her defining attribute?
The First Law of Moe

    A moe character cannot be aware of her own appeal.

Indeed a large portion of her appeal comes from the viewer’s knowledge of this fact. When an otaku looks at an image he describes as moe, first and foremost he sees a character who is not consciously trying to leverage her other appealing attributes to attract or seduce the viewer. This fresh, unspoiled innocence lies at the heart of moe appeal.
This lack of self-awareness, when inverted and framed as a positive attribute, can be called the moe character (hereafter referred to as the heroine)’s maidenly virtue, modesty, or propriety. An image has several ways of engaging this concept, ranging from affirmation to abandonment, and it is this engagement that provides one of the principal metrics for gauging the moe quality of an image.
The following five categories encompass the various ways a heroine’s propriety can be addressed by a moe image:

  1. All maidenly virtue, all the time. The image’s narrative affirms a conservative, almost mythically hypertraditional sense of correctness.
  2. The heroine’s virtue is placed in an everyday context; that is to say, it is generally preserved by the narrative, but not to a rarified extent. The narrative may depict a fantastic scenario, but it won’t place the heroine’s sense of propriety in particular danger.
  3. The heroine’s virtue is compromised by the narrative in some improbable set of circumstances. The heroine’s reacts with righteous indignation to revert the narrative to its proper status quo as soon as possible.
  4. The heroine’s sense of a proper self image is seriously challenged by the narrative depiction of a fantastically improbable or impossible event. She either reacts with outraged embarassment (not possibly commesurate with the insult to her dignity that she has recieved), or begins to accept her role as an icon of sex or cuteness. Either way, the role her virtue plays in the image’s narrative is reduced significantly.
  5. The heroine’s virtue is either surrendered as she embraces the reality of her sexualization, or it is rendered irrelevant by her reduction to nothing more than the sum of her visual attributes.

In addition to an image’s narrative, as discussed above, there are several other components of a moe image that work to describe the character’s position relative to her sense of virtue as depicted in the image and interpreted by the moe referent system. These components include:

    Ambience: The feeling or emotion conveyed by, and the physical properties depicted in the image’s mise-en-scène.
    Relationships: Ties between characters, based on kinship or emotion, that bear on the image’s narrative. By necessity relationships are established within the broader context of the work the image is derived from, or spelled out textually within the image itself.
    Roles: The heroine’s occupation, and that of other characters, as depicted in the image. Often fetishized, roles can either be meaningfully tied to the image’s narrative and relationships or reduced to their fetish status as little more than costumes (see below).
    Symbols: Costumes and props / accessories depicted in the image that have gained fetish status. They can be employed by the image either for their original use or their fetishized use.
    Style: Character design elements that set moe characters apart from other character types. The predominant stylistic trait of moe is found in the head and face; the ears drop to the base of the jaw, pulling cheekbones, lower eyelids, and the bridge of the nose with them. Eyes and forehead become larger, nose and mouth smaller, head rounder. These alterations can vary from a subtle, more realistic look to cartoonish “BESM” deformations.

When taken together, how do these various properties of the moe image impact the viewer’s perception of the character it depicts? We find the answer in
The Second Law of Moe

    The greater an image’s emphasis on style and fetish symbol at the expense of narrative, ambience and relationships, the less relevant propriety becomes.

Conversely, the greater the emphasis on narrative the further style and symbol are rendered contingent to it and the greater role the heroine’s purity or virtue plays in the image.
We have now developed nearly all of the tools we need for a complete understanding of what the moerer sees in a moe image, but one important factor remains: the extent to which the viewer himself, whether directly or through a proxy character, is interpolated into the image’s narrative.
There are five possible levels of male presence in an image:

  1. None. The viewer is ignored by the image.
  2. Accessory. The viewer (or his narrative proxy) is acknowledged by the image, but his presence has minimal impact.
  3. Voyeur. The viewer’s direct gaze is acknowledged by the image, but he is not able to participate actively in its narrative.
  4. Central. The viewer is the focal point of the image’s narrative. His active participation is allowed.
  5. Intimate. The viewer is implicated as the heroine’s lover.

The Third Law of Moe governs the relationship between the viewer’s distance from an image and its heroine’s “maidenly virtue”, or propriety. It should come as no surprise that

    The closer the viewer (or his narrative proxy) becomes to a moe character, the harder it is for her to maintain her sense of propriety.

This law documents more of a trend than an uncrossable line; scenes in which the heroine is focused on the viewer or his proxy can certainly be pure and virtuous, but they may also tend toward sexual innuendo as the law would imply.
Moe as Emotion
We now know what the otaku sees when he looks at a moe image, but before we continue we need to turn around inside his head and find out what he feels. This brings us to
The Fourth (and final) Law of Moe

    The viewer’s emotional response to a moe image is a function of the convergence of his position relative to the image with the heroine’s state of maidenly virtue as depicted therein.
  • In response to an image where male presence is not implicated and the heroine’s virtue is affirmed, the viewer reacts with sentimentality, nostalgia, empathy, and a sense of affirmation or alignment with a (probably fictional) status quo.
  • In response to an image where male presence is intimately depicted and the heroine’s virtue is preserved, the viewer reacts with an urge to comfort or support the heroine (if she is weaker) or be comforted or supported by her (if she is stronger).
  • In response to an image where the male is present as a voyeur and the heroine’s virtue is compromised, the viewer is titillated by the resulting sexual tension.
  • In response to an image where the male is central or intimate, and the heroine’s virtue is challenged or surrendered, sexual arousal is elicited in the viewer.
  • In response to an image where male presence is not implicated and the heroine’s virtue is irrelevant, the viewer is prompted to react with excitement and diffuse energy.

Broadly considered, these five responses combined with their triggers provide a comprehensive scheme for the catagorization of moe images and narratives, as depicted in the following diagram. This can be seen, in essence, as a map of the moe referent system:
[image removed]
Junai-kei moe (純愛系)
The harmonic depiction of innocent lovers is the province of junai-kei moe, an image type commonly found in popular renai games. Note that the male character need not be shown in the image for his presence to be evident to the viewer (and the heroine); such images are often more powerful when the heroine is seen to be addressing the viewer (or player, in this case) directly.
Otome-kei moe (乙女系)
The beauty and grace of everyday life in nearly utopian tranquility is captured by otome-kei moe. Unfettered for the most part by concerns of love or sex, the heroine goes about her business in perfect and idyllic innocence. Male characters are rarely present, though female interaction in the image can lend it a strong yuri tinge (but not beyond the bounds of propriety).
Erokawaii-kei moe (エロ可愛い系)
The heroine acknowledges the proximity of the male gaze, and likes it. The erotically cute moe image is just that, a girl who may be shy but is willing to shed most conventions of virtue for the benefit of the viewer or his proxy. The extent to which sexual activity is allowed here is open to debate, but force or coersion perpetrated by either party is a hard limit to this form of moe.
Denpa-kei moe (電波系)
Denpa is a term used to describe nauseatingly saccharine, hyper-cute bishoujo game theme songs, and denpa-kei moe images are just that: heavily stylized, almost Dadaist depictions of heroines often so decked out in props and costumes that few physical design traits remain. Not overtly sexualized and with little male presence to interfere, these mascot-like characters are the iconic bastions of moe.
Kaimami (垣間見)
Translated directly as “looking through gaps in a fence”, kaimami is a literary device that dates back to the Tale of Genji. This time-honored tradition persists today in moments of voyeuristic titillation that can crop up across the spectrum of moe narratives, but most comfortably occupy an unstable position at the center of the mix, representing the moment when the viewer’s gaze compromises the heroine’s virtue for the first time.
Issues in Moe
There are several corollary themes that have thus far been assumed in the course of this analysis. What follows is an attempt to unpack some of them:
Primacy of the Image

    Q. It’s assumed here that the image is the primary vehicle of moe; that is to say, the viewer can look at an image in isloation and describe it as moe with no further referential material. Is this really the case?
    A. Yes and no. To clarify, when a moerer views an image he references various libraries of meaning to determine its moe content: first the character, then the narrative, then the other various stylistic elements and symbols. If he lacks knowledge of a particular image’s narrative referent (he hasn’t seen the anime or read the manga it is derived from) its appeal may be less obvious.

Alternate Formulations

    Q. Doesn’t moe just boil down to personal preference? Why spend so much time trying to categorize it?
    A. What I try to show here is that moe isn’t just liking (or loving, or being turned on by) anything, it’s a more specific emotion or trait that can only be applied to certain entities (moe heroines) without lapsing into incoherence.
    Of course, this is at odds with a segment of its actual usage (“train moe”, etc.) but I would argue that these are loose slang references to the type of emotion instantiated most directly in the moe categories I have outlined above.
    Q. What’s wrong with Ken Akamatsu’s interpretation of moe? Or that of Moetry for You? They seem perfectly serviceable.
    A. Akamatsu’s “nurture” theory is interesting, but it ignores a good three-fourths of the moe spectrum (otome, erokawaii, and denpa). These are all well-established uses of the term (though not by my arbitrarily coined terminology, of course) and by not addressing their broad public acceptance he limits his theory to being only partially complete, at best. Of tangential concern is how Akamatsu would rate his own work given the standards he sets for moe; his perpetually henpecked protagonists and fanservicey females don’t seem well-suited to his formulation (with the exception of Earth Defense Force Mao-chan, perhaps).
    Moetry for You also ignores the baser and more frilly points of moe, but generally does a fine job of encapsulating its spirit. It’s certainly a more authentic take on the topic than this.


    Q. Can the moe “heroine” be male? Can the moerer be female?
    A. Yes and yes. I think it is clear that a fair segment of shotacon can be counted as moe in the various senses outlined here. Regarding the gender of the viewer I can only speculate that women are capable of appreciating innocent virtue in ways similar to men; if I were to sketch out a reverse-gender diagram it would look something like this:

[image removed]

    Q. How do manga and anime originally targeted at girls fit into the moe spectrum?
    A. Magical girl properties, along with some yuri, exist at a peculiar confluence of the moe and shoujo genres. They’ve been increasingly co-opted into the moe mainstream, to the extent that very similar male-oriented titles have been created to cash in on their popularity (compare Burston’s Engage eroge to Marimite, for example, or compare Nanoha to Precure).
    While not all shoujo is well suited to migration, magical girl shows tend to rest comfortably in the denpa niche with their superabundance of costumes, accessories, and simplistic narratives. Shoujo yuri isn’t as common and subsequently sees less representation, but when it is accepted it finds a place in the conservative end of otome-kei moe.


    Q. At what age can a heroine be described as “moe”? When does she grow out of it?
    A. The first law of moe dictates that the heroine cannot be aware of her own appeal. Implicit in this is that she has the potential to be made aware of it; i.e. that she posesses enough self-awareness and social consciousness to engage in a propriety-based narrative. Thus she must be well beyond Lacan’s mirror stage and have had an opportunity to begin the realization of social mores: around the age of five or six.
    On the opposite end of the spectrum we find the heroine aging out of the period of youth in which innocence is still credible; notions of girlish virtue can only linger for so long before they become a wink-and-nod caricature. I would place the absolute limit here at around age thirty, though such characters are extremely rare; the standard moe age range is between late elementary and high school (or some kind of mythical post-secondary “gakuen” level of education invented so eroge protagonists can be both legal and still in school).


    Q. Can a non-human character be moe?
    A. I would contend that non-human characters cannot be considered moe. With that said, the moeverse has managed to stretch the definition of “humanity” to previously unknown breadth, including androids, catgirls, bunny girls, various forms of undead, dolls, and etc; basically anything with a vaguely human shape and humanlike intelligence. Non-human characters are thus limited to animal sidekicks (which may be cute, but not moe) and any other personifications that a virtue-based narrative cannot be applied to coherently.
    Q. Can a nonliving item be moe?
    A. According to my formulation, no. Trains, spaceships, power tools, cars, guns and the like can’t be moe because, while they may argueably be characters in some narratives, they’re not sentient. I feel ridiculous even trying to explain this, so I’ll stop here. Anyone who calls a train “moe” is either hallucinating or using the term in an irresponsibly broad context when other words would be better suited to the task.


    Q. What is the difference between kawaii and moe?
    A. “Kawaii” translates relatively cleanly as “cute”, and is applied in a similar range of situations. It could be argued that all moe characters are also kawaii, but the reverse is certainly not true: animals, clothing, cars, faces, babies, houses, old ladies, phrases, etc. etc. are all instances of objects that can be described as kawaii but not moe. As I think has been made clear, moe is a property of human or human-like characters only, within an age range that permits them to be both socially aware and innocent.


    Q. What is the relationship between fanservice and moe?
    A. Discounting more esoteric forms of service (references to other works, detailed mechanics, realistic physics / props and the like), there are four types of fanservice moe is concerned with:

  • Heroine-initiated service occurs when the heroine deliberately exposes herself to the viewer, and is rare except in erokawaii moe.
  • Accidental service occurs when an act of nature, another character, or simple bad timing serves to expose the heroine to a (typically unwanted) male gaze. This is the textbook example of what I’ve termed kaimami and usually results in severe physical damage to any male characters involved, though the heroine’s emotion is occasionally more self-centered embarassment or consternation than an immediate concern for retribution and the righting of the virtuous status quo.
  • Camera-based service occurs when, absent a male proxy, the camera takes “artistic liberties” in its placement vis a vis a heroine’s particular assets. Too much of this in an image or narrative and its virtue-based credibility is reduced, shading either toward denpa, erokawaii, or Agent Aika.
  • Prop / costume-based service is the exploitation of fetishized moe symbols, and is found mostly in denpa and erokawaii moe.
  • In certain respects moe as a genre could be considered an exercise in fanservice, as some form of it tends to be present in almost every moetic work.


    Q. Can moe only be conveyed by images? What about other media, such as text and sound?
    A. This is a poser. I’m tempted to say that raw text can’t adequately represent moe unless reinforced by at least an image or two, as is standard for Japanese serial novels, but I’m by no means certain of this. I’ve never read a block of text that I would consider “moe” on its face, but my experience with Japanese literature leaves much to be desired. Using a slightly looser definition it could possibly be argued that novels such as the Betsy Tacy series are moe, but I’m not prepared to go very far in that direction.
    The obvious candidate for “moe sound” would be denpa, and to the extent that it shares a name with one of the categories I’ve chosen I feel obligated to include it here. As my analysis has thus far been limited to moe images I haven’t developed a vocabulary yet with which to discuss sound, and whereof one cannot speak, thereof one must be silent.

Actual Usage

    Q. To what extent do authoritative print sources and conversations with individuals who use the term in a natural context bear out this definition?
    A. The ideas developed here are based on prolonged exposure to self-titled moe publications (Dengeki Moeoh, Moetan, Moegaku, Maritan, the Moe. eroge brand, etc), works in which the term is employed (ranging from Comic Party to Welcome to the NHK), and individuals who use the term in spoken and written Japanese to describe their personal preferences. I am confident that a similar investigation of readily available resources (including a Google image search for 萌え) would yield comparable results.

Relationship with Lolicon

    Q. Isn’t moe just lolicon in a fancy costume, and by extension an excuse for pedophilia?
    A. The only aspect of moe that could possibly be conflated with lolicon is what I term erokawaii, which composes roughly a fourth of total moe material. While referring to real human children as “moe” is certainly disingenuous (see below) referring to moe illustrations as pedophilia is even moreso; as both Akamatsu and Moetry for You go to great pains to point out, the bulk of moe is not sexual, and the parts that are must by definition remain benign and consensual. Above all, it must be stressed that moe is an attribute of two-dimensional characters (again, see below).
    Lest I come across as an apologist, it is true that “moe” has been coopted to a certain extent by advertisers eager to make a few quick yen off of an otaku community with seemingly infinite amounts of expendable cash. As such, its usage in seedier parts of the internet has been stretched to include images well beyond the pale of general consensus on the topic (what this piece is concerned with).

Narrative Stasis

    A concept central to moe that hasn’t been touched on yet here (though Akamatsu addresses it) is the notion of narrative stasis, or the “happily ever after” syndrome. The desire for moments of perfection to be frozen in time and memorialized is a common theme in moe narratives, especially renai games; the heroines often express the sentiment “I like you just as you are, and I want us to stay this way forever.” While the narrative may continue past that point of perfect bliss, narrative stasis is accomplished through the basic mechanism of static archival images.
    Thus, I think it can be argued that moe in its purest form is distilled in these carefully composed one-shot images designed to capture a moment of perfect symbolism in whatever idiom it is representing. From the “sad girl in snow” to a flash of panties in mid-fall as a waitress stumbles and drops her tray, stasis as opposed to evolution, progress or improvement reigns supreme in the moe narrative.
    This is part and parcel with the notion of moe as reinforcing a mythical status quo in which the moerer / viewer / player is not forced to acknowledge his shortcomings (much less the ones that may or may not have caused him to turn to this particular form of entertainment in the first place). Instead, he is accepted “as he is” by the heroine, and moments of happiness with her are preserved until time immortal with no fear of aging or falling out of love. This, I believe, is one of the central attractions of moe.


    Another important question that remains is the issue of dimensionality. Is moe solely the province of two-dimensional characters, or can flesh and blood humans also lay claim to the attribute?
    When considering three-dimensional moe we immediately run up against a couple of problems: first, the initial assertion that moe is a property of characters as depicted in images. Humans aren’t characters, so to lay claim to the moe attribute they must adopt a character identity. In turn, such an adoption violates in principle the first law of moe; namely that the moe heroine is unaware of her own innocent appeal.
    When a person adopts a character identity (often through cosplay, or the appropriation of costumes and accessories associated with the character or role in question), her existence as moe becomes contingent not on her lack of knowledge of her appeal, but on her ability to induce the viewer’s suspension of disbelief. In other words, her ability to play pretend is set against the viewer’s willingness to be seduced.
    The one group obviously immune from the precondition of adopting a character identity is that which has yet to develop knowledge of its own (sexual) appeal; i.e. prepubescent girls (or boys). I would contend, however, that applying the moe descriptor to them is disingenuous; “cute” is more straightforward, and not laden with the referent baggage of moe (or the attendant implications of pedophilia).
    Another aspect of dimensionality that comes into play when considering moe is the or distillment that occurs when three dimensions are reduced to two. While it can be argued that behind every two-dimensional heroine is a three-dimensional artist who may have motives far less pure than the characters he or she depicts, compared to an in-the-flesh idol (or even an author) the moe artist is transparent, letting abstract brush strokes form meaning in the mind of each individual viewer untainted by residual artistic baggage or preconcieved notions of the creator.

Consider the possibility that two-dimensional characters are moe precisely because they are depicted in two dimensions, and it is this reduction, simplification , lack of pretense – it is this lack that allows the heroine to preserve her virtue unquestioned by the viewer.
The incessant and agressive commodification of moe stands at ironic odds against this central tenet, but compared with other options it may seem superior: what the moerer expends on moe brings guaranteed dividends, perhaps not of love, but of simplicity, understandability, purity, earnestness, honesty.
However you read the omens of moe – crass commercialism, social decay, artistic revolution, the harbinger of world peace, or just a bunch of delusional geeks who deserve to be bred out of the gene pool, in the near term the future of Japanese visual culture is very clear:
It will be flat, and it will be moe.

    Special Thanks
    This article would not have been possible without brain-picking conversations courtesy of Albert Elegance and OGT, Matthew‘s stellar translation of Ken Akamatsu’s blog post, and the few, the proud, the contributors to the Heisei Democracy forums. You guys rock.

0 Responses to The Moe Image

  1. Rocket December 7, 2005 at 3:39 pm #

    Bravo sir, your work is a milestone! Finally – a treatment of the subject that does justice to it’s breadth (lack of depth) and also does not ignore it’s seedier side.
    So much to muse over.. however I must take exteremly minor issue with your contention that moe characters must be A) flat and B) not human.
    “Flatness”, or graphical depection certainly help in the reduction and purififcation process. Yet it seems obvious that a character depiction in felt cut out collage, or 3D Studio is just as capable of prducing the same viewer response. Let’s not forget the “figure side” shall we?
    As for “Not human”, if a character image can be presented to the viewer by pencil and paper, computer pixel, or injection molded plastic, its seems plausible that the character could be presented by a human in costume, or the image thereof. The character adopted is no more conscious of the self aware motives of the human player, than the character.jpg is aware that it is composed of pixels. I think the human may or may not be moe, but the character can be.
    Furthermore i find it easy to imagine images of actual persons that occupy the Juani-kei section. Historically we can consider many western Victorian era images of girl children rendered as idealized femininity to be equally valid (and more dimensional) examples of moe. At least if we take moe out of it’s Japanese context and rely only upon the definitions explicitly stated in this article.
    Consider this description from David M. Lumin of Seymour Guy’s 1867 painting, “Making a Train”. “That title refers to the imaginary skirt that a little girl, probably Guy’s nine year old daughter Anna, sweeps behind her. Alone in an attic bedroom she pretends to be a grown lady, a bride or queen, or simply her mother. Given the theatrical role playing of the girl as well as the stage set, stage lit quality of the, the painting enjoins its viewers to think of themselves as invisible spectators, positioned in the dark…the painting is buoyantly innocent, but also teasingly erotic.” Not mentioned is that the child is naked from the waist up. Depending on the viewer, the reaction can range anywhere from Juna-kei, through Kaimami, to Ero-kawaii. At this point in American history, as Lumin goes onto dissect, this range of responses was part of the national visual discourse, though of course the latter examples was repressed and not spoken of. Denpa and Otome are also possible responses, as we have the following quote from Mark Twain decrying the combined inspidness and predictability of both in an review he writes on the 1867 National Academy of Design ‘s annual exhibition. “…more than half of the paintings in the Academy [were] devoted to the usal harmless subjects…same old detachment of cows wading across a branch at sunset.. the everlasting farmers gathering their eternal squashes; and ‘Girl Swinging on a Gate’; and a ‘Girl Reading’; and girls performing all sorts of similar prodigies.”
    Moe is perhaps not an artistic movement, but is does seem to be a cultural movement. Perhaps we could say that moe is a modern Japanese expression of cultural pattern that occurs when an attempt is made to “reconfirm gender stereotypes…[and] ward off from a white (sic) middle class male dominated social order, threats arising from militant feminists and other agents for socal change.” At least that’s how this type of image and viewer response in the 1800’s America and Europe is viewed by some art historians.
    Indeed, if one considers the aburpt swing from militant, messianic, and apocalyptic Otaku visions of spaceships, robots, and noble heroes to a Moe dominated small town, is it any surprise that the underlying premises would be just as conservative? Cute, clumsy, airheaded, waitress turned popstar – samurai like hero piloting his transforming airplane against gigantic, cultureless, imperialist invaders – now they seem like they were two sides of the same coin all along. Just a thought.
    Did I mention your work was pure genius?

  2. Shingo December 7, 2005 at 4:25 pm #

    Thanks for taking the trouble to respond to this extent. Let me address one point first off: you’re quite right to point out the discrepancy of the “Figure Side” as well as three-dimensional computer rendered images; I agree that they belong firmly in the moe camp, and your pointing to the slippery slope of canvas change to argue that the human form could be used as a canvas to depict moe art, well… that blew me away. I’m tempted to agree with you right off the bat, but it’ll take a bit of thinking to process this concept (this being the three-dimensional commodification of moe that I’m primarily objecting to here).
    I’m guessing from your close reading of that 19th century painting that you have an art history background, something which I sorely lack. From your description of the painting I wouldn’t hesitate to brand it “moe”, at least by the formulation I have here, as you say; my brief mention of the Betsy Tacy books is the only leftover indication of that line of thought I was travelling down for awhile that included the written works of Louisa May Alcott and the like. I’m not well versed (not versed at all, really) in premodern “character art”, which I’m sure lends an entirely different perspective to this issue.
    I agree with your interpretation that moe may be a fundamentally conservative movement. The point I was trying to make about narrative stasis (listening to a denpa track earlier called “futsuu no anata ga daisuki”, or “I love the average you”) is certainly a backlash against modernist pressures of perpetual evolution and progress.
    The analysis of moe and apocalyptica as two sides of the same coin is attractive, but taken out of historical context may be oversimplified. To a certain extent I think it could be argued that all of Japanese popular culture, especially visual culture, is relatively conservative, and traditional warrior images and the femininity of moe are less related than tied to a common source of “what comforts, sells, what brings discomfort does not” (Evangelion and the works of ABe being exceptions that prove the rule).
    Definitely a whole lot of food for thought here… if you feel like it, it might be better to continue this on the forums to avoid cluttering up the comment area. ;;

  3. 天魁☆ December 8, 2005 at 3:30 am #

    Hi. I am the auther of “the poetry for you”
    Let me say several things.
    but it seems me you miss one important point. That is, moe cannot work out without moerer.
    Several people have pointed out that moe is just like quantam physics. The narrative that moe characters have can change according to observers, or in this case, moerers. This is because these fictitious characters inevitably have contingency, since no author can give his/her characters complete information. (i.e. What moe characters are doing or thinking while they are not appear on the page or screen?) Moerers supplement this lack by their imagination to make up a ideal character for them. Therefore, conflicts sometimes occur between Otakus, reagarding their moe characters when they supplement that lack in different way.
    I would like to call this “uncertainty principle of moe”
    One more thing. 3 years ago, I wrote “Moetry for you” as a kind of parody of “Poetry for you” (though I did not mention this in the article) You may or may not know, “Poetry for you” is written by a English poet during the WW2, when poem was looked down on. “Now, moe is also looked down on by the society. So what if I praised moe in just the same way “Poetry for you” lauds poem? Can it sound plausible?” That was why I started writing this, partly as a practice of English. Yes, “Moetry” is coined from “poetry”.
    Three years later, my view of moe has changed. I now think it is nonsense to anguish over moe: This is nothing more than an entertainment, right? But its content seems reasonable as it is still read by English people, which I did not really anticipate.
    By the way, shouldn’t “Don’t take it so serious” be ” Don’t take it so seriously”?
    I’m enbarrassed. orz
    Should I have time and inclination, I will re-write this.
    I would be very happy if you’d say “your English have much imporved.”

  4. Shingo December 8, 2005 at 6:07 am #

    Your English is far better than my Japanese, so please pardon me for replying in English.
    The first point you mention is true. While I try to suggest at times that the moerer plays an active role in taking his own meaning from a moe image (with such lines as “the moe artist is transparent, letting abstract brush strokes form meaning in the mind of each individual viewer”), I think I may push too hard on the idea that there is only one moe referent shared by all moerers.
    My point in stressing this is not that each moerer percieves moe in exactly the same way; but rather that there are limits to the “uncertainty principle” that you suggest. I think that when one goes too far afield of certain accepted ideas of moe, it ceases to be moe and becomes something else (like “cute”, for example). However, by my suggestion that a moerer MUST react with a certain emotion to a certain type of moe I think I am going too far; those are just general observations on my part and should not have been stated as fact.
    I was not aware of the origins of “Poetry for you”, but now that you mention it I have even more respect for your work. The parallel you draw between post-WWII poetry and current moe could probably be the subject of an article even longer than this one (;;). Despite your original intentions I think your piece has become one of the founding works on moe in the English language, and if you ever have the time to update or revise it I think many people would be interested to hear your new ideas. I know I would.
    Regarding “anguish” over moe, I think there is a fine line between being apologetic or defensive and actually trying to understand something by lending it serious thought (that it may or may not deserve). I think popular culture is a fascinating subject that is often ignored, and while it may not be deeply intellectual I think the study of entertainment can tell us a lot about who we are as people. That’s my justification for my academic interest in moe, at least.
    On your last point: you’re right, “don’t take it so seriously” is the proper form of the sentence. I can’t criticize your English at all, though – aside from a few spelling mistakes it could nearly pass as native. You remind me that I still have a long, long way to go with my Japanese.
    Thanks again for replying! I’m very happy you took the time to read my piece (although it is way too long;;). I would welcome the chance to continue discussion of moe in the future.

  5. Rocket December 11, 2005 at 9:41 am #

    I’m so glad my somewhat overzealous response was met with the proper spirit. I’ve long been musing on related topics – after reading your piece I could barely contain myself! orz
    I’d love the chance to answer some of your points and clarify a few of my thoughts. I’m going to link to your article directly from my site after I’ve gathered a few of my older contextual thoughts on moe from the past. Regardless, if you’d point me at the appropriate forum and or thread to continue in I’d love to jump right in, especially since I’ve also had some thoughts regarding the self-unaware aspects of moe which I’d like to hear your thoughts on.
    天魁☆, let me thank you for your original article. Stories are the way in which our minds understand reality. What could be more worthwhile than trying to understand stories and images which result in such strong emotion. Shakespere and Dickens were pop entertainment in their day, and their less acomplished contemporaries are also worth serious study!

  6. Shingo December 11, 2005 at 2:47 pm #

    I’m not sure what I need to do to make the forum more visible… ;; Feel free to link and comment at will, of course. I look forward to seeing more of what you have to say on the subject.
    By the way, is that your art on your blog? Sweet stuff, the perspective is devilishly tricky.

  7. Brendan December 17, 2005 at 5:01 am #

    I would like to see more of an explanation of the relationship between shounen’ai and the moe image. I think there is an interesting conundrum present with the insertion of characters such as Guilty Gear’s Bridget into the moekko vernacular.
    Also, I suggest you read Baudrillard’s “Simulacra and Simulation” as a way to re-contextualize the precession of moe images. All in all, though, fine work!

  8. LiK February 2, 2006 at 5:09 am #

    very good article with lots of research and thought. my head spun a little from all the explanations but i think this was omething very informastive to fans who wonder what moe meant.

  9. s April 9, 2006 at 2:48 pm #

    A very nice and comprehensive read. Although most of the context I’ve come to understand (in the form of the terms “platonic love/relationships), and somewhat sentimental/undilluted sappiness at times, I still have a question: Based on what you’ve described about moe, being a degree from virtue being preserved to surrendered, does that make Hayao Miyazaki’s statement valid about moe culture being the source of the objectification characters to their appeals (which when decronstructed can be attributed to fetishes, either on accessory, personality, etc.), which would make nearly any heroine to be taken seriously as a (non-harem) male lead would? I only ask because you seem to have know clearer context of its usage in shows.

  10. megan May 5, 2006 at 9:36 am #

    Thank you, Shingo, for your work on this subject. As a professor with a graduate student working on Aya Takano and Chiho Aoshima, your observations about “moe” have been really helpful.
    However, I note that you assume a universally male viewer and point of view. Have you thought about the female consumer of these images? And conversely, about female producers (and possibly parodic producers) of these images such as Takano and Aoshima?
    I’d love to discuss this topic.

  11. 5parrowhawk October 11, 2006 at 12:29 pm #

    Apologies for bumping this post so far after the fact. I’m really impressed at the amount of effort that has gone into this, but feel the need to braindump.
    In defense of Akamatsu, I must say that he has put his “moe theory” to the test in Negima ch. 122-125, the Ako story arc (in which it functions in a most magnificent manner), and that his definition at least covers otome-kei as well as junai-kei.
    The Ako arc embodies most of Akamatsu’s moe theory:
    – It is practically 100% nonsexual. (Notice how, even in the scene where her bare back is shown, it is done so in such a way as to arouse concern and sympathy rather than sexual desire)
    – During this arc, Negi (with whom, one assumes, the reader is expected to identify) is his “adult self” and so is in a dominant/authoritative position – this is no coincidence.
    – There is virtually no possibility of further development, due to Negi’s returning to child-form – hence Ako’s memory will remain untainted (in an idealistic sense) – hence affirming the present situation.
    – It is noteworthy that Akamatsu very deliberately divorces this from the loli/denpa archetype. Yes, Ako is underage, but her “state of being” is completely different from that of the standard moetic-loli character or the denpa-type character (despite minor trappings of nurseness which don’t show up in this arc). One could conceivably transplant her personality, attributes, &c. to a character 10 years older without an obvious disjunction.
    – It is also noteworthy that the character of Ako was given very little “screen time” before this arc. This supports the idea that the Ako arc was intended as an experiment in concretizing Akamatsu’s theory of moetics.
    As for the otome-kei aspect of the definition, one has to note that throughout Akamatsu’s definition there is no mention of male *characters*, only a male viewer. I think it is possible for a male viewer to identify with a dominant female character in the case of yuri (e.g. Marimite), or to conceptually place himself in the situation depicted, not as a predator but as a nurturer. The reverse would be true for female-to-male moe, which could potentially account for the volume of non-erotic BL material.
    In the final analysis, it might serve better to categorize moe by potential aspects of the moerer (can we find a more elegant word for that?)’s emotion than by situation. Nurture covers more than jun’ai; it extends into otome and denpa territory as well. Eros stands by itself. Admiration, despite protestations to the contrary, can also be part and parcel of moe, as with certain classes of tsundere or other “powerful” characters (Konpaku Youmu esp. in the Western Touhou fandom, Subaru Ryoko, Ruri-Ruri, any of Akamatsu’s Shinmeiryuu swordswomen, etc), and often is conflated with implicit contradiction (between innocent appearance/behaviour/personality and mythopoeic levels of power/intelligence/etc). Memetics (the denpa phenomenon – or should we call it “mimi-tics”?) or symbology is another aspect of moe which is difficult to nail down, although the canonical example would be “Mecha Musume”.


  1. » Shuffle #20 - December 7, 2005

    […] Touching on Shingo’s theory of moetic characters, Kaede has really taken the sort of naive loyalty found in Nerine to ridiculous sociopathic extremes, culminating with her sneaking into Rin’s bed completely naked. Undoubtedly the new Psycho Kaede strikes a chord with many a fan (this one included). It’s that crazy, blind, totally unrealistic world-be-damned dedication that’s so endearing. […]

  2. “あいとゆうきのおとぎばなし” » Blog Archive » the verb form of moe - March 18, 2006

    […] So I had started rereading NHK 4 so I could put up the review for that one when I got sidetracked. There’s a scene where Yamazaki and Sato are channel surfing and they come across a show about the inspirational story of a loli who’s terminally ill and dying. Yamazaki is dumbfounded and can only reply with “…moeru” (”…萌える”), the verb form of “moe”. Shingo’s excellent essay on moe covers the noun and adjective forms of moe (which are the moe aesthetic), but I’m never really sure how to treat the verb form of moe (”moeru”). It doesn’t help, of course that Japanese is often vague and just assumes that you know what the subject is. Going back to NHK, I was doubly confused here because I normally go back and dig up the scanslations of the mangas after reading the raw Japanese tankoubon in order to make sure that I didn’t miss anything. I noticed that the group (Evil-Genius, I think?) whose scans I was downloading would normally leave the word “moe” alone, but they translated “moeru” here as “…that’s SO HOT.” […]

  3. adorable rockets » Blog Archive » ロケット萌え - March 16, 2007

    […] I’ve added a new page with some thoughts on that very subject. As well as our favorite topic, the elusive moe. […]

  4. adorable rockets » Blog Archive » On Moé - March 19, 2007

    […] • Moetry for You – by 天魁☆ • Akamastu-sensei talks Moe by Ken Akamastu • Otaku Talk by Toshio Okada and Kaichiro Morikawa, moderated by Takashi Murakami, from Little Boy • Beyond the Pleasure Room to a Chaotic Street by Midori Matsui, from Little Boy • OTAKU : persona = space = city by Kaichiro Morikawa, from the catalogue of the 2004 Venice Biennial • Otaku Sexuality by Tamaki Saito, from the catalogue of the 2004 Venice Biennial • The Moe Image by Shingo • Nothing in Anime Makes Sense, Except in the Context of Fetishism by Love • Otaku History by Zer0 […]

  5. adorable rockets » Blog Archive » What is Moé? - August 4, 2007

    […] think with Shingo’s groundbreaking work, and my modest addition we have a pretty good understanding of how we can define moé, that is a […]

  6. Reverse Thieves » Moe! Good God ya’ll! What is it good for? - July 10, 2008

    […] the odd thing is, I don’t wholly hate the idea of moe I just hate the horrible machine it has become. Pretty princesses sometimes need saving. As I have […]

  7. My Delusions As A Fanboy: If Moe Is A Cancer… « Drastic My Anime Blog - July 7, 2009

    […] Here are three of my favorite takes on moe (the first is not safe for work, and hopefully SDS can update his post or repost it on […]

Leave a Reply

Powered by WordPress. Designed by WooThemes