« Gary Gygax, 1938 - 2008 | Main | Localization meets Culturalization »

Mar 05, 2008



Well, I think there's a deeper question here. "Is it really human behavior?" Yes, of course it is - it's the behavior people engage in when they create/play avatars in virtual worlds, with a specific set of motivations, stimuli and rewards (I actually dislike behaviorism so I don't know if I'm getting the terms right).

But I suppose the deeper question is "does this 'avatar behavior' tell us anything about human behavior in general?", or, to use a term I dislike even more, "does it tell us anything about human nature?" This is a very loaded question whose answer will depend on how we define "behavior in general", "human nature", and, indeed, "human". And the best reply is "possibly", I guess.

I think there are valuable things to be learned both from "avatar-only" studies and from studies relating avatars to users - as long as you're conscious of the pitfalls in both kinds, there's no reason to choose one over another.


Unless it's an NPC, it always comes back to the user. It may be a male user who chooses a female avatar, but it's a user nonetheless. Does the user's avatar behave differently than the user might in real life? Perhaps, and that's a large draw of virtual worlds. For many, it's a fantasy escape...a chance to be someone different, act different, live different. But those are all still reflections on the user's desires.


I think you do have to tie it back, if just to capture the phenomenon of people expressing themselves in different ways through different avatars.

eg EVE online players may have 'pirate alts', who participate in activities not condoned by their corporation. Or WoW non-guilded alts who gank newbies, scam, or whatever else - it's not my intention to limit this to 'bad' behavior.

If you studied all avatars you would capture all behavior, but you might conclude that there's several different kinds of players, when in fact it's the same person. Drawing conclusions about players from studying avatars would then be wrong.

I guess another way of putting it is asking what are you intending to learn from your study? If it's how residents act, study residents. But if you want to know how users act, you need to study users.


Are those 1995's in the brackets a year of the stats used? That was 13 years ago! That is a hell lot of time in this case.

It is not clear what the study wants to find out?
If the question is if a virtual world can be used as a society in vitro, so one can assume that avatars are behaving like humans, no that is a very bad presumption. But one can find out a lot about humans if kjeep in mind that those humans are using computers to access a world with very different characteristics (start with the opt-in gravity...).


1. The same avatar might be controlled by different players at different times (though I think this is fairly rare, and frowned upon in some VW's).

The same player might control several avatars

So, statistics like "X% of avatars do Y" really are about the percentage of avatars, not users. (What if the players whose avatars do Y have a greater than average number of alts? etc.)

2. You also need to be cautious about generalizing from VW behaviour to RL.

Two obvious examples:
- In RL, I don't drive my car the way I do in racing games
- RL reaction to seeing someone shot dead is likely different from in a first-person shooter.

There are similar doubts about how "griefing" in virtual worlds relates to real-life behaviour.

3. Information about an avatar's gender, age, race, and species may not correspond to the user. It may be interesting to know that "X% of furry avatars do Y", but this is clearly not the same as "X% of the users who are anthropomorphic animals do Y". The distinction is obvious with furries, but it's easier to fall into the trap of equating the avatar with the person with race, gender etc.

You could get really wrong answers by (for example) interviewing a bunch of white guys who role-play an Ali G impersonation, and assuming that this corresponds to the behaviour of players who are actually black.


Psychometric tests like the MMPI-2 have a similar issue. If you ask your test subjects whether they agree or disagree with

1. I like mechanics magazines.

what you are measuring is not whether they in fact like mechanics magazines, but whether they say they do when asked in the test. None the less, this can still be statistically correlated with something you're interested in. (It turns out that the mechanics magazine question isn't all that useful diagnostically, but the same principle applies to the other questions).

So, similarly, you can measure whether or not your test subjects present as a squirrel in Second Life, and this might be statistically correlated with something interesting. Even if the player is not actually a squirrel.


I think it's an interesting question, sure. But it has to be made in the context of all the different environments in which we find ourselves. Substitute "football fan" for "avatar" and "churchgoer" for "user" and you'll be testing whether people behave differently when they're watching their team get creamed vs. when they're going to Sunday School.

Of course we behave differently in different situations. It is, to a certain degree, an illusion that VWs are "differently different" than other mediated and variously social/cultural experiences.

Does my avatar make different choices than I do? That is a bizarre way to form the question, because avatars don't make choices; users do. Do I make different choices through my avatar than I would make for myself? Again, stop filtering it through the lens of, "Wow. That really looks like another person," and think of it as another communications tool. Of course you do different stuff with your avatar; it can do different stuff.

However, it can't eat, pee, itch, etc. One of the things that seems to break up more WoW/VW sessions is the potty break. My avatar can go forever without visiting the WC. I cannot. Therefore, by definition, he/she/it/they and I will be doing different things.

I ask the question of myself this way: what fictions do I enjoy engaging in? Sometimes that fiction involves me as a viewer only (book, movie, etc.). Sometimes it involves me as a player, sometimes as a creator.

I think there might be some utility to comparing how actors feel about playing different parts and how VW/MMO players feel about playing different characters. Having known many actors in my life, I know that there are some roles that make some of them uncomfortable, and some that they really enjoy playing. I haven't, yet, been able to come up with a standard that lets me make judgments such as, "Actors who are poor love playing rich characters," or "Moral actors hate playing villains."

It's not a who-you-are thing. It's a what-you-do thing. And while there are lots of sub levels of what-you-do in VWs, I take it as an absolute given that I am *not* "changing myself" when I play a VW any more than I am changing myself when I talk on the phone or am more polite to my minister than I am to my guut buddies. I am just doing different things.

It is, I think, less than helpful to equate the actions of people within VWs as somehow being fundamentally different than those elsewhere. You can do different things, yes. But you are still the same person. Even (especially?) if you are pretending to be an opposite gender, dark elf assassin.


Andy Havens wrote:

But you are still the same person. Even (especially?) if you are pretending to be an opposite gender, dark elf assassin.

Completely agreed. I've had this discussion many times with people who claim that when they're roleplaying they are "a different person." That is simply not true, of course, as all input into the game that is user-controlled flows from the user. It's not possible to "play a different person." All you can do is play an aspect of yourself.



C'mon, it is possible to play different person. One can go create a character, like for a book or comic and play it.
Though, most of the times, we are playing aspects of ourselves, I agree with that.


No, it's not possible to play a different person. Any input you send to the character is, by definition, part of you to begin with, else it could not have originated with you.



Yes, it is. Otherwise literature wouldn't be possible. I can play a role of fascist dictator. It will originate from me, but hardly that I can tell that is me.


dandellion: If I am a man, I can answer the question, "What is your gender?" by saying, "I am female." I can answer subsequent questions in ways that, I believe, are more in line with how a woman might answer than how I would answer were I not playing a part. I may begin to feel differently (let's assume, more feminine) than I do when I play a male character, either based on feedback or because I am observing my own choices as being different in a self-imposed, self-imagined female way.

That doesn't make me a woman. It makes me a man playing a role. And, if I feel like it, I can stop and play another role. And another.

One of my favorite truisms of pen-and-paper RPGs is that the most difficult player to GM is one who is running a character who is smarter than the player. How do you make someone play a character in a way that is more intelligent? We work around this by giving that character better clues (if he/she makes his/her roll) or by making some things more evident to them; we substitute observation by proxy or revelation for intelligence.

I can pretend to be a woman, a giant, a mouse, a table a rutabaga. And it can be fun, interesting and instructive to do so. But if you study the behavior of "male avatars," what you are really studying is the behavior of players pretending to be something else which is, in this case, male. Is it interesting and, possibly, informative to do research like Nick Yee's into how people behave in VWs based on the gender of avatar? Sure. But we might also observe that when people pretend to be pirates, there is a propensity to say, "Aaarggh" a whole lot.

There may be cases when any behavior is the result of neuroses or some other disorder. We might observe that a particular individual, no matter what, always plays short avatars or ones with purple hair or always plays the opposite gender. And that data may be helpful.

What is not helpful is to treat the behavior of "avatars" as distinct from player behavior. You cannot, I believe, take the behavior of avatars as indicative of how their players would do... well, anything necessarily.

I do think that it's possible to make careful measurements and observations of avatars, and to have research conducted inside VWs... but to do so with the clear understanding that the game is as much a part of the real world as my clothes are part of a larger social construct.

I am reminded of the classic Statistics 101 example cited often as a way to familiarize students with casual relations vs. causal ones: a graph of candy consumed clearly shows that those who eat the most candy weigh the least. Students start by positing reasons such as, "Candy eaters eat less of other foods, as it's a snack," or "Perhaps it's an athlete thing." The the prof asks, "When in your life did you eat the most candy?" And the "a-ha" goes around the room as people realize that it's kids; children eat more candy than adults and weigh less. Without age data layered in, the stat is useless.

We've already discussed here many reasons why, for example, men play women toons. Unsurprisingly, to me, there is a wide variety of answers. Why? Because playing a cross-gender toon isn't a major lifestyle choice or a "choice to be." Very few men who cross-play would, I think, choose to go transgender in RL. On the other hand, I know many men who, if you asked them, "Would you like to be a woman for weekend, just to see what it's like?" would say, "Yeah!"

What we do in our various fictions is instructive. But they are all things that *we* do. When I read sci-fi, I am not actually transported to a future time. I am me. Reading sci-fi. My imagination allows me to enjoy that in an immersive way, same for when I play RPGs. But immersion in fiction else does not change me, except inasmuch as everything I do does so.

I am going to go pretend to be asleep now. I will let you know when I pretend to wake up. Oh wait. I'm already awake. Never mind.


So, basically having a drug induced gender-change is impossible... That's odd isn't it? Because, I think it is possible. So, I am not sure if there is any "true me", although we like to believe so.

I certainly have roleplayed or enacted people who doesn't have my qualities. However, for it to remain interesting I have to play on emotional aspects of myself, but it is certainly not a necessity. It's just very hard not to do so.

This discussion reminds me of all those MUD-papers from the 1990s...


@Andy: Bravo!

@Ola: Lol. And it reminds *me* of social theory from the 50s (that's still just as good today as it was then). Nothing new under the sun and all that. ;-)


@Thomas: funny you should say that. I was just sitting in the kitchen and thinking how much rhetorics matters in academia today and then I thought to myself "but Goffman got away with saying interesting things in a rather consise manner". I guess the S/N noise ratio was better when people were handwriting their manuscripts. Kudos to old theorists (even if the results perhaps aren't as solid in terms of data) that were concrete enough to provide people with frameworks to think about their own lives. My personal favourite is Pragmatics of human communication (Watzlawick et al., 1967). Not sure if they are right, but it is a good read anyway. Very good S/N ratio. ;-)

Another thing I came to remember was IIRC Turkle's assesment that even if men roleplayed women online they probably never would experience what it would be like to have menstruation. Which probably is true. One could try I suppose, but would probably never get close to what it truly feels like for some women. We can have a taste, but we can't have it all (thank God).


@Ola: Ah, Bateson. I've just been reading Fred Turner's very nice account of the powerful connections between post-WWII systems theory and the distinctive segment of 60s US counterculture that included Stewart Brand. Turner focuses on Brand, and there's a section in there on Bateson's influence on his thought through the 70s. All of that stuff that built on the assertion that there was a "system" (whether informational, as Bateson had it, or otherwise) which exhibited the capacity to adapt and self-regulate has had an enormous influence on current ideas about technology, emergence, and governance.

The echoes of that stuff linger throughout the virtual worlds we see today. But now we're off the path of the OP. I'll pick up this thread in a post here sometime soon, since it bears directly on some of the ideas I'm working through in my book about Linden Lab. Hmmm. Maybe I should handwrite it. ;-)


Ola wrote:

So, basically having a drug induced gender-change is impossible... That's odd isn't it? Because, I think it is possible. So, I am not sure if there is any "true me", although we like to believe so.

A drug-induced gender change simply highlights the fact that gender is a falsely binary category which we force everyone into as a convenient shorthand (though clearly everyone doesn't fit, with or without drugs - hermaphroditism for instance). It's still 'you' not someone else.



For some reason, this all reminds me of the movie "Brainstorm" [my god, Chris Walken has a huge body of work]. The machine that tapes my experiences, and you can play them back across all 5 senses.

So... if a woman records herself doing something... womanly... and I play it back, essentially "playing a woman," can my thoughts regarding what happens during that time be comparable to those of a woman? EG, if experienced the tape of a woman doing several things, and then being offered the choice of doing one of them again... am I more likely -- outside the tape -- to choose the one the woman would have, or the one I would have?

My answer: I am more likely to choose the one that men choose after spending time in the woman sim.

I do love to quote Yeats; "How can we tell the dancer from the dance?" In this case, it is, "How can we tell the player from the character?" The point is, we don't have to. It's me.


@Staarkhand eg EVE online players may have 'pirate alts',

@andy "It is, I think, less than helpful to equate the actions of people within VWs as somehow being fundamentally different than those elsewhere. You can do different things, yes. But you are still the same person. Even (especially?) if you are pretending to be an opposite gender, dark elf assassin."


I agree. I think, however, the difficulty is in contextualization. I like Staarkhands reference to pirates in EVE as an example here. Pirates are a fascinating and miniscule sub-culture in a tiny sub-culture (EVE) in a small sub-culture (MMORPGs) in a sub-culture (online gamers).

First, as background, there is a gradient of stigmatism at work with EVE pirates: 1.) Pirates are generally disliked by most EVE players (though I think most understand their role). 2.) PVE MMORPG players think EVE PVP is a fringe gaming culture. 3.) many online gamers view MMORPGs suspiciously.

(1.) and (2.) is likely transitive: PVE MMORPG player think EVE Pirates are a way-out cultural fringe (uber gankers).

But here is the thing. Based on in-game experiences, I actually found some EVE pirate groups surprsingly sophisticated, nuanced, and dare I say it - likeable. What struck me was how tightly knit, e-honor bound and surprisingly lawful (in the D&D sense) some were. They had a philosophical viewpoint about their role in the game and the executed it with precision and with a code of conduct. Yet likely many looking in from afar would see quite a different entity.

So yes, (viz Staarkhand's example) the true-blue loyal alliance member who moonlights as a pirate in EVE is the same person, however, to recognize what it says about that person would understanding the niches and contexts they operate in.


Andy, I agree that one can get too far out of own experience. But, unlike what you said about sci-fi, our experiences (including reading, movies, immersive RP-ing) are changing us and expands us.
Your example with gender switching is a heavy one. On the other side stands a survey published in New World Notes about a year ago about how different are players from their avatars. After all gender/age, hair colour things, there were 11% of players that play opposite political views from their own. There is no reason for somebody not to play different political views as good as s/he plays their own.
So, it gets us back to the start. For a good survey, one have to follow both humans and avatars. Differences vary depending on the category that is researched.


@dandellion: But Andy and Matt and others here are exactly right. No matter how *differently* any of us may act from one context to another, perhaps supplemented by the affordances of the trappings of our role (whether that be a uniform, a gavel, a machine gun, or an avatar), it is always a human being acting, as it were. To elevate the avatar into some other category ontologically (as the philosophers would say) on a par with human beings, is to commit a serious error.

This is not to say that the nature of the avatar is not of interest for human social performance. Nor to say that using an avatar cannot transform us in certain ways (as any role - for lack of a better word - can). Furthermore, the avatar is a distinctive and sophisticated affordance for human beings in many ways, all worthy of research. But it is not another human being, and should not be researched as such.


Don't get me wrong. I am not opposing to them nor saying that avatars are independent on humans. Avatar is a representation of a human, but not exact one on one copy. And humans can be very diverse in their behaviour.


I also think that this ignores a critical function: game mechanics. For example, in one MMO I play, female armor parts are much easier to find than male, and less demand means much lower cost. Thus, there is an economic incentive within the game itself to play a female character.


I’d agree with Andy that the avatars is ultimately expressing the self of the player. But any normal human plays a wide range of roles even in everyday life. Their behavior isn’t the same in the office as it is out drinking with the gang, as it is back with the parents for Sunday dinner.

The avatar effect just amplifies that natural human ability. As does dressing up in costume. Or wearing a mask. On online avatar is just a better quality costume and mask.

I do think I’ve learnt something from the experiences of my avatar. Dressing up for a pantomime made me wonder how an earth women deal with walking round in high heel shoes. But being in a show, it didn’t tell me much about how men interact with women. On the other hand, in my early EQ experiences, I was quite shocked at how differently male characters treated my female and male characters. I’d read about such differences, but it’s a bit different from experiencing them, even at the one remove of an avatar.

On the subject of studying avatars, I think there is something to be learnt from that. In early EQ, on my server, a pretty distinct role-play style evoloved for how to play an Ogre. It defined a socially acceptable way to play an “evil” character while still being nice to other players. It might have been an interesting study to monitor how that little culture absorbed and educated people. While you could tie it back to the guy behind the screen, there is something to be said for just studying the avatar culture as its own world.


肺癌 肾癌 膀胱癌

The comments to this entry are closed.