Welcome to Raph Koster's personal website: MMOs, gaming, writing, art, music, books.
Welcome to Raph Koster's personal website: MMOs, gaming, writing, art, music, books.



This brings about the issue of identity. A player’s online identity is largely a result of the intangibles, not the tangibles. If the player is not a careful role-player, what will happen is that the intangibles that leak through (such as eloquence, prejudices, moods, analytical ability, leadership, etc) will be completely unrelated to and unaffected by the formal definition of his identity within the virtual space, which is concerned only with the tangible elements such as statistics. Thus we get the phenomenon of the character with supposedly low intelligence that is the leader of his peer group.

The way to think of it is that the mud server relates to a user based on the tangible statistics, because they are database entries. And other players relate to the user on the basis of the intangible attributes. But these intangible attributes are not exactly the intangible attributes of the actual human behind the mask—rather, they are filtered intangibles that are modified both consciously and unconsciously by the user based on the inputs he receives from the server and based on the nature of his online identity.

Text Box:  A selection of avatars from The PalaceThis is a heavy burden for an avatar to carry. Fortunately, avatars are iconic and not representational. And this means that they can carry a much heavier semantic burden. In Scott McCloud’s seminal book Understanding Comics,[1] he defines cartooning as “amplification through simplification.” What he is getting at is selectively picking elements of a picture to concentrate on and make more important—like, perhaps, the way in which we select intelligence, strength, dexterity, and other such attributes to serve as an abstraction of a person in an RPG. He also says,

When two people interact, they usually look directly at one another, seeing their partner’s features in vivid detail. Each one also sustains a constant awareness of his or her own face, but this mind-picture is not nearly so vivid; a sketchy arrangement... a sense of shape... a sense of general placement. Something as simple and basic as a cartoon.

He goes on to note that the only thing more abstracted is text itself.

The implications for the avatar are clear. In the virtual world, what we are seeing is the cartoon selves of different individuals interacting directly. McCloud says, “Icons demand our participation in order to make them work. There is no life here except that which you give to it.” The process of fleshing out the reality of an icon, McCloud terms “closure.” Media which depend on this sort of thing were termed “cool” media by Marshall McLuhan—comics is one such, and television is another. A virtual space, because of its interactivity, of course cannot fall into the same category, but it is interesting to see that one aspect of virtual worlds, at least, can be regarded that way.

 In order to identify with an avatar in virtual space, we must make a very large leap of “closure.” However, the amount of closure required has been reduced over time as the tangible qualities of avatars expanded. This is complicated by the fact that the filtering performed by avatars is still very imperfect: just as in the paper RPGs, the player’s personal abilities to communicate and lead are more important than the alleged abilities of the avatar that represents him.

An interesting possible extrapolation of McCloud’s commentary is that it may very well be the fact that the typical virtual world avatar is an abstraction that makes them so easy to identify with. They serve as a direct projection of our unconscious selves in that like that self, they are a “cartoon” in McCloud’s terminology. This makes them extremely easy to identify with. This argues in favor of reducing the realism of our avatar portrayals as we design these environments, if we wish to have tighter identification. Interestingly, social spaces are those with the least definition to their avatars. They have a much higher proportion of intangible avatar characteristics, and correspondingly less world physics to relate to. At the simplest end of this scale lies The Palace, the project with which Randy Farmer and Chip Morningstar of Electric Communities followed up Habitat. This mud uses a simple room-based model, with only a backdrop screenshot and little sense of spatiality to convey the “location.” And avatars themselves range from full cartoon bodies to essentially disembodied cartoon heads, floating against this backdrop, speaking in comic-book bubbles.

In a similar vein we see The Realm, originally developed by Sierra. Its side view, shared by BSX muds, Habitat, and The Palace, deliberately invokes a comic book in its imagery and artistic style. The player in the figures on this page went on to play Ultima Online, where the greater realism of the artwork no doubt prevented her from finding as good a representation of herself.

Of course, even this representation of Janey isn’t really the way she is generally perceived in the game: Ultima Online offered a higher-detail version of the character and the regular version, which is what most people could see. And this version looks much more like a cartoon again, as most of the detail is lost, including (most crucially) facial features.

In a more realistic environment, like Asheron’s Call, not only do facial features follow “heritage groups” (meaning, you can get Asiatic features, or Caucasian, and so on) but you can pick from an assortment of eyes, noses, mouths, and so on. This allows you to more carefully craft something that mirrors your mental image of your avatar, but also may result in removing some of that iconic quality in characters.

The flip side of the coin in all of this, of course, is whether the iconic nature of the other avatars in the space makes them easier to objectify, or easier to empathize with. In general, the more cartoony spaces tend to have less problems with antisocial behavior, but that may result from self-selection of the participants, rather than any effect of the artwork. This remains as a potentially fruitful area of research.

It is a major mistake to minimize the importance of identity in your virtual space. It is also a mistake to assume that it can serve to tie a given player to your space (say, perhaps, for the purposes of charging an ongoing fee). Your environment offers an opportunity to incarnate in bits and bytes a persona that the player normally has only in their mind. But it is almost certain to be an incomplete mapping. The primary persona is and will always be the one visualized in the player’s head. As a result, it is eminently portable from world to world. There are countless incidents of role-players carrying a given character, complete with name, behavior, and even similar tangible attributes, from game to game, over the course of as many as twenty years.[2] The characters will not ever map perfectly, but they map as well as they ever do, since any mapping whatsoever fails to match the vividness of the conception in the player’s head.

That said, it is a wise thing to offer as wide a range as you can of means to reinforce an avatar’s identity, because it increases the chances of a player finding a mapping for the persona they wish to present. In addition, playing dress-up with a miniature doll is just plain fun, for either gender (though men may choose not to admit it), and the more choices you offer for avatar identity, the more you can offer that fun. It is no accident that Ultima Online and other games have termed the avatar customization window the “paperdoll.”


[1] HarperPerennial, 1994.

[2] A poll by the RPG news site The VaultNetwork (http://rpgvault.ign.com) found that 57% of players had played the same character or persona in multiple games, campaigns, or settings.