Gender Studies

I learned through experience as an undergraduate at Essex University that as a male I have been irredeemably socialized to hold oppressive, androcentric views about all things gender and that therefore my opinions on the subject are worth less than nothing. Even seemingly neutral actions on my part are inherently tainted by my early conditioning and testosterone-charged world-view, so can therefore only be regarded as manifestations of my poisoned, male psyche. Consequently, if you don’t want the purity of your beliefs about gender issues adulterated, you should skip this section.

For those of you who know sarcasm when you see it, read on….

The general point here is correct: Anything people say about gender is inevitably moderated by their own experiences. Whatever I write on the subject—even my selection of research paper references—really is influenced by my upbringing and physiology. However, the same applies to every individual on the planet, so I’m not alone in that; furthermore, it isn’t only true for gender studies, but pretty well “anything else,” too.

Much work in gender studies is from a feminist perspective ¹, which makes my position slightly more dubious. There isn’t a lot I can do about this, though. I shall strive to be impartial (although not necessarily unopinionated where virtual worlds are concerned); if I slip up you can blame me, not my gender.

So here we go….

Gender Studies is not Sex Studies; there is a difference between sex, and gender, essentially anthropological. Sex is a biological condition; gender refers to the behaviors associated with the sexes. Sex is uniform across all cultures; gender varies across cultures.

Gender influences people in two main ways: Through their gender identity (that is, their feeling of being male or female) and through their gender roles (that is, what their society establishes as appropriate behavior for females and males). Although there are only two sexes, some societies have more than two genders (usually to account for individuals whose biological sex does not match their gender identity). Gender roles may have a biological origin (the “boys will be boys” argument) or a cultural evolutionary one (tribes whose womenfolk fought and died alongside their menfolk had fewer babies than ones where only the menfolk fought, so the former were eventually outnumbered and defeated by the latter). Alternatively, gender roles may arise as the result of pressures from other cultural activities.

Although this is the classic, anthropological view of gender, it is not one that is universally held among gender theorists. They regard sex as an essentialist concept (that is, something that has intrinsic meaning beyond what perception reveals), whereas their preference is for a more constructivist ² approach (that is, that things only have the meanings that individuals construct for them based on the sensory data). This would suggest that sex/gender were part of the same system, but as Tomasz Mazur describes in an early review ³ of how gender studies theory applies to virtual worlds, this doesn’t work either—it flounders on the concept of identity. Mazur argues that virtual worlds themselves have a similar problem, with “real” being essentialist and “virtual” being constructivist, and identity existing in both and neither at the same time. He adopts a post-modernist ⁴ stance to address these issues.

Problematical though the fundamentals may be at a philosophical level, this is not something that normally concerns students of gender with regard to virtual worlds. Rather, they are primarily interested in two specific aspects of player behavior that introduce new variables into their own equations: cross-gender play and virtual sex. I’ll take these in that order.

If there were a pill that changed your real-life sex for one day only, would you take it? The science isn’t important—a mass of nanotechnology DNA-tampering machines could perhaps do it—but would you take it? Why would you want to? Why wouldn’t you want to? What would the effect be on gender issues in a world where anyone could take such a pill?

In virtual worlds, that pill is already available, without prescription.

The paper that brought the practice of cross-gender play to the attention of nonplayers was Amy Bruckman’s Gender Swapping on the Internet ⁵. This followed up on some of the activity she had noted in her earlier Identity Workshop paper, and includes interviews and other set pieces that have been heavily quoted by later researchers. She observed two behavioral patterns in particular that were to resonate with many people: That female characters are subjected to unwanted sexual advances; that female characters are assumed to be in need of help.

Bruckman indirectly suggests that the two may be related. Offers of help can be viewed as tokens of interest; by accepting the token, the recipient is signifying their acceptance of the other person’s interest; they are then obligated to respond favorably. Although some may see this as buying favor—and it may well be just that in some cases—it isn’t always. Sometimes, it’s just social interaction. Similarly, whether an offer of help is an esteem-damaging, patronizing assumption of vulnerability or the first move in a mating game depends largely on the attitudes of the players involved.

Unusually, Bruckman views gender as just one aspect of identity, on a par with (say) nationality. I share this view, but it’s not one that cuts much ice with those who politicize gender; Bruckman has been criticized for taking this position. I personally find it difficult to see why gender should be preeminent; race and sexuality are perhaps as much a part of a person’s identity as their gender, but these don’t get anywhere near the coverage that gender does. Perhaps it’s more to do with the way that most languages (through pronouns) force gender to be an issue; race and sexuality are not denoted in this way, and therefore have a much lower profile ⁶.

Bruckman characterizes gender as being malleable in virtual worlds. She means in the self, not in the character, but other researchers don’t always realize this and have occasionally questioned it. Bruckman, as (by then) a player of virtual worlds, could understand the difference between real and virtual self from an experiential point of view; researchers who have only an intellectual understanding are more prone to miss the point.

Sherry Turkle also writes about cross-gender play in Life on the Screen (in her Chapter 8); she even interviews some of the same people as Bruckman. She sees cross-gender play as challenging and psychologically complicated; characters that are neither male nor female she finds disturbing, yet evocative. As a result of her studies, she puts forward the hypothesis that gender exploration is a valid part of identity exploration in general; it can lead to the development of a person as an individual, due (at least) to the extension of their emotional range that it stimulates.

The existence of genders other than male and female in some virtual worlds is discussed ⁷ by Brenda Danet. Danet regards gender as a tyranny, and is therefore interested in the opportunities presented by the Internet—where all conventional gender cues are absent—for gender-free existence. She looked at Internet Relay Chat (IRC) and MOOs, and found that in both, individuals can successfully camouflage their identities for very long periods. Cross-gender play in MOOs, though, was much more elaborate than in IRC ⁸.

MOOs typically have more than the usual two genders available to them ⁹, some of which may be considered a third gender, others a non-commitment to gender, and others completely genderless. Examples include:

• Neutral (“Sam puts its cards on the table”)

• Either (“Sam puts his/her cards on the table”)

• Plural (“Sam puts their cards on the table”)

• Spivak ¹⁰ (“Sam puts eir cards on the table”)

• Person (“Sam puts per cards on the table”)

From my point of view, this is interesting because in MUD1 I didn’t want gender to be a tangible property of characters. My own dialect of English—Yorkshire—uses near-plural by default for people whose gender is unknown or unspecified (as you will doubtless have noticed in this book, where I habitually refer to singular but generic players as “they” or “them”). As a student, I was once told off for using the word “themself” in a report, which up until then I hadn’t even considered not to be an English word ¹¹. For me, if I were informed that a new teacher had joined a school, it would be perfectly natural to ask “what’s their name?” rather than “what’s his or her name?” ¹². This approach was not suitable for MUD1, though, because for non-dialect speakers it looked bizarre (although it has since gained much wider acceptance, at least in England) and because sustaining it when gender is known is untenable.

My preferred alternative to implementing gender was to use a pronoun-free denotational approach instead (“Sam puts Sam’s cards on the table”), but I abandoned this on the grounds that it was too intrusive—I wanted to draw people into the virtual world, and the erection of constant billboards that said they were reading text was too much of an interference ¹³. It’s amusing to note that neither of the ways I considered for the linguistic avoidance of gender (near-plural ¹⁴ or denotational) have found any currency in MOOs.

Danet looked at gender choices for two large MOOs: LambdaMOO and (Amy Bruckman’s) MediaMOO ¹⁵. Both have non-binary gender classifications in place, and Danet examined the proportion of players who chose to take advantage of the alternatives to male/female that were available. She found that on LambdaMOO, roughly 50% of the characters were male, 30% were female and 20% were neither; on MediaMOO, 50% were male, 20% were female and 30% were neither.

There may be any number of reasons for this difference. Perhaps the media professionals of MediaMOO have a more progressive attitude to gender, or the balance of players’ real-life genders in the two games is significantly different. Danet’s observation that the real-life identity (and therefore gender) of MediaMOO’s players is known, whereas that of LambdaMOO’s isn’t, muddies the waters still further. Are MediaMOO players more likely to make gender-political statements, knowing that they may be called to account afterward if they don’t? Or is the fact that no one can accuse them of attempting to be deceitful by playing with their gender a liberating factor?

Danet’s paper ends with 11 questions for future research into cross-gender play, all of which have yet to be fully answered.

I should point out at this stage that there are two opposing views concerning whether cross-gender play is a good thing or a bad thing. They can be summarized as

• Cross-gender play is good, because it emancipates the individual.

• Cross-gender play is bad, because it subjugates the group.

At one level, its researchers’ views are inconsequential: Cross-gender play happens, and it will continue to happen no matter what people believe should happen and no matter what ideology they want it to fit. They can’t stop it, police it or exploit it, so incorporating it into their theories is purely an academic exercise. Virtual designers know implicitly, however, that things can be done to promote one side of the argument over the other, thereby encouraging or discouraging the practice. Creating mechanisms for characters’ gender to flip periodically through the use of magic or magiclike effects, for example, would encourage cross-gender play ¹⁶; making graphical avatars physically plain (if not ugly) would discourage it.

So, let’s look at some of the objections in more detail.

One of the hopes that feminists had about cross-gender play was that it would lead to a fluidity of gender. Lori Kendall points out ¹⁷ that in practice it does the opposite: People come to entertain a more rigid understanding of gender identities than they had before. They focus on the “true” identity behind the mask, rather than the one that is presented ¹⁸. Women came to virtual worlds later than men, therefore the cultures they had to join were already in place when they arrived, and thus the likelihood of changing this situation is not going to increase by much even if more women do participate in these virtual worlds—at least, not unless everyone makes a conscious effort.

This argument looks fine on the face of it, but it relies on regarding virtual worlds as extensions of Reality. In an absolute sense, they are part of Reality and therefore can indeed be considered physical extensions of it. They are not, however, extensions of the models of Reality that players have in their heads. Even social MOOs are separate from reality, as many players understand immediately and those that don’t will come to realize in the long-term. The hardware that supports each virtual world, and therefore the implementation of the virtual worlds, is in Reality; however, the virtual worlds themselves only exist in the imaginations of their players. They’re not worlds because they are worlds; they’re worlds because people think they’re worlds. What’s more, players think these worlds are not the real world.

The only things that exist in both the virtual world and the real world are the players’ identities. Kendall is correct in her observation that gender types in virtual worlds can be very stereotypical, tending to crystallize around some falsely perceived “essence” for each gender. Where I believe her to be mistaken is in concluding that these only serve to reinforce existing notions of gender and therefore add to the general woes of women (in particular) when they are acted on in real life. For players who can’t yet separate the real from the virtual, this may be true; for the rest, it isn’t ¹⁹.

Characters and players both move. The formal role that a character takes on may be some stylization of a false perception of Reality, but experience in playing that role will change both the character and the player. The character will change when behavioral expectations meet the reality of other players’ attitudes; the player will change when these expectations meet the player’s own attitudes. There is a fluidity of gender identity, and the stereotype drives it; in other words, where the stereotype exists it’s to show that it’s wrong.

Over time, the relevance of both Kendall’s and my positions may diminish. As players’ experience with virtual worlds grows, it could be that gender moves on and that stereotypes become irrelevant. While newbies still arrive with real-world gender issues to address, however, the likelihood of any such paradigm shift’s occurring is doubtful.

A second objection is raised by Jocelyne Voisin. She points out ²⁰ that in radical feminist theories, separatism sometimes plays a central role: The creation of a woman-only environment is seen the only way to escape patriarchal oppression. In the real world, such safe haven environments do exist—spaces from which men are banned—but they are relatively small in number. A separatist virtual world would be easier for many women to reach, and more conducive than real life to discovering whether the assumed benefits of such an environment (a supportive, egalitarian, nurturing community) would in fact arise and be sustainable over the long term. Gender masquerading, however, rules this out as a possibility: If there’s no way to tell that a female-presenting character has a biologically female player running it, the extent to which these spaces can be considered exclusively feminine is limited.

Voisin is right: Cross-gender play does undermine such feminist virtual utopias (as transsexuals pose problems for real-life “female-only” spaces). Deliberate infiltration of a female space is not likely to happen often, but even so that may be more often than is acceptable. Unfortunately for the radicals, there’s very little that can be done about this; even proof of real-life identity when an account was created doesn’t guarantee that the person later using the account is the one whose identity was proven ²¹. Creating a virtual world with all-female characters doesn’t mean you’ll get all-female players; the only way to guarantee all-female players is to have them under constant real-world surveillance as they play—hardly a situation that could be described as supportive or egalitarian. So although cross-gender play validates feminist conspiracy theorists’ worst fears, it’s unstoppable and unavoidable; taking an uncompromising view of it isn’t going to change a thing.

Countering that view is the suggestion made by Michèle White that perhaps virtual worlds are intrinsically not female-friendly places anyway, at least as they are currently designed. White’s complaint ²² is that the commands used in MOOs to obtain information about other players suggest a surveillant attitude and use voyeuristic terminology (“watch,” “scope,” “glance,” “peep,” “gawk,” “examine,” and so on). The “look” command itself, with the third-person messages it triggers, is worst of all; frequently, new players are distressed when they are continually looked at by other players. The concept of “gaze” is important in feminist theories, because implicit in it is a communication between the gazer and the gazee. MOOs unequivocally empower the gazer: Gazing is active; self-image is passive. In other words, you describe yourself for the benefit of the person who looks, not for the benefit of yourself. This limits the identity constructs that people can create, which is bad for everyone no matter what gender they are playing.

I have some sympathy with White’s position. In my own virtual worlds, players don’t get to describe their characters—I believe that allowing it anchors character development too much. On the other hand, looking at a character is permitted (it concocts a brief description based on their tangible properties), along with large numbers of other commands that have characters as their objects (including “hit”). Although White herself doesn’t argue this, anyone willing to construe such commands as empowering those who issue them at the expense of those against whom they are issued would surmise that virtual worlds (even graphical ones) were less than ideal. Quite how to implement these kinds of commands some other way is non-obvious, too. That said, virtual worlds are still much better than the real one in this respect, because at least the “empowering” actions are open to anyone.

Returning to the issue of how cross-gender play undermines attempts to set up female-only virtual communities, it could be that the problem might be worse in theory than in practice; it depends on the actual incidence of cross-gender play. If far fewer men play as female characters than anecdotal evidence suggests, the effects of their presence might amount to mere background noise. It would therefore be very useful if empirical evidence existed to show how widespread cross-gender playing really is. At a more general level, it would also be very nice to know if the predictions as to why people play across gender are in fact accurate.

I know of two studies that have attempted to answer these questions ²³.

The first study is by Kathryn Wright, who is the resident consulting psychologist at womengamers.com ²⁴. Her perspective is that of female computer game players (among whom she numbers), rather than of gender studies; nevertheless, the results of her study ²⁵ are relevant in this context.

Wright surveyed men who played female characters across a variety of game types, but mainly those of first-person shooters and (gamelike) graphical virtual worlds. 57% of her respondents (all of whom claimed to be male in real life) played as female characters for over half the time. Respondents were asked to describe in their own words why they played as female characters; when Wright analyzed the results, she found they fell into the following categories ²⁶:

• 25% said it added to the role-playing experience (of which 19% cited gender exploration as a motivation).

• 23% did it for the visual stimulation (that is, sexual aesthetics ²⁷).

• 19% did it because the avatars were smaller and harder to see/hit.

• 16% did it to be given gifts.

• 14% did it because female characters had better skills.

• 14% said they preferred the modeling (that is, non-sexual aesthetics).

• 13% did it for variety.

• 11% did it to get a psychological edge over the other players.

• 6% preferred the sounds that female avatars made.

• 6% did it to amuse themselves at the reactions of male players.

• 6% did it from habit: They started, and hadn’t found a reason to stop.

Although role-playing comes out as the top reason, many others are for gameplay purposes, that is, female characters have an edge for whatever reason, and so they are played as a means to an end. Wright anticipated this and asked a series of follow-up questions; read her papers if you want to know what she found (it’s interesting, but not relevant to the current discussion).

As a final question, Wright asked whether cross-gender players thought there was a correlation between men who played as women and men who were homosexual or confused about their gender; this was in response to something that happened when she advertised for respondents to her survey (she was accused of doing the survey in order to meet homosexual EverQuest players). She found that among the respondents themselves

• 60% said there was no correlation.

• 15% said there may be for some people, but not for everyone.

• 12% said there was for them personally, but not for everyone.

• 12% didn’t answer.

Wright notes that no cross-gender players reported a belief that men who play as female characters are homosexual by default.

Wright’s survey was limited, in that she had 64 respondents of which only 33 actually completed her survey (the rest gave their opinions via email). She concluded that lumping together role-playing, fighting and action gamers was inadvisable, because too many answers were dependent on the type of game that the players played. Nevertheless, her results do suggest that some cross-gender players really do perform it for reasons of exploring their identity.

A much more formal and detailed survey ²⁸ was undertaken by Lynne Roberts and Malcolm Parks in 2000. Their methods were very thorough: They selected the largest two social MOOs plus five others chosen at random, and from these selected (also at random) 1,200 candidate characters ²⁹ that had been played during the previous 14 days. A second study included players of role-playing MOOs. The candidates were contacted individually, and around 30% agreed to take part.

Unlike Wright’s survey, which looked only at the cross-gender players themselves, Roberts’s and Parks’s examined the views of all players on the subjects. Their findings make very interesting reading. For example, they discovered that the male/female player breakdown ³⁰ in the MOOs they surveyed approached parity (48.1% female in study 1, 47.0% in study 2); this is a much higher ratio than is common for regular computer games ³¹.

Roberts and Parks discovered that around 60% of the players of social MOOs had never played cross-gender, but that around 20% were engaged in it at some time during the 14 days the survey covered. In role-playing MOOs, 43.3% had never tried it but 40.0% were currently cross-gender playing. Including all the personality and sexuality tests that the studies incorporated, playing in a role-playing MOO was found to be the greatest predictor that an individual selected at random from the respondents was or had tried cross-gender playing. Perhaps surprisingly, real-life gender was not found to be a predictor of cross-gender play: Female players were as likely to play another gender as were male players.

I said “another gender” there, because of course MOOs have multiple genders. 78.7% of cross-gender play was found to be to the opposite gender (that is, male or female) of the real-life player, whereas the remaining cross-gender ³² players opted for something else; this is in line with Danet’s figures from LambdaMOO four years earlier.

As for why they played across genders, the reasons players gave were consistent but the proportions varied between those who were currently doing it and those who had tried it but stopped. Among current cross-gender players, 31.3% said it was part of role-playing or a challenge to their role-playing skills, whereas 13.4% of former cross-gender players did. Conversely, 13.5% of current cross-gender players said they had started cross-gender playing to experiment with gender, whereas 34.3% of former cross-gender players did. Why there should be this reversal clearly merits further investigation.

The surveys produced many other results about who cross-gender plays and why; read the paper itself for the full details. From the point of the discussion here, there are two issues the results raise that are of particular importance.

Firstly, the studies do seem to show that cross-gender play is a widespread activity. Roberts and Parks themselves do not conclude that it is prevalent, pointing out that although many players had played cross-gender in the two-week period the study covered, only 22.6% of those Study 1 respondents who had done so had spent more than 60% of their playing time pursuing the activity. For Study 2, the figures were better: 31.5% of cross-gender players spent more than 70% of their time doing it, of which 20% said they played the whole time during the previous two weeks as a gender other than their real-life one. Roberts and Parks are correct that cross-gender play isn’t prevalent (if “prevalent” means “more than 50%,” which from a strictly statistical point of view it must); their results would seem to indicate that it’s quite a common activity, though, even in purely social (as opposed to role-playing) MOOs. I suspect that the figures for goal-oriented virtual worlds would be higher, since players of these are more formally embarked on an exercise of identity exploration ³³; my only caveat is that some people object to cross-gender play in principle, and these may be more common in such gamelike worlds.

This brings me to the second point of special significance in Roberts’s and Park’s results: Why people don’t play cross-gender. The studies found that many of those players who had never played cross-gender believed that the activity was manipulative, was used to lure people into having virtual sex, and was a way to cause trouble for people who didn’t do it. They were upset and uncomfortable when confronted by someone not playing as their biological sex. Personally, I’m very disappointed by this; the cited activities are ones that cross-gender play should weaken, not strengthen! I can only hope that the people who hold these points of view do so out of ignorance, and that as they come to understand the difference between the real and the virtual (that is, that there is one) they may revise their opinions ³⁴.

As I mentioned earlier, Roberts’s and Parks’s survey showed that cross-gender playing by female players was far more common than is usually supposed; their studies showed that it happened as often as it did for male players. The reasons that women play cross-gender are different, however. A short article ³⁵ by (female) player Natalia, who plays as a male character 90% of the time, explains that she does it mainly to avoid the nuisance of unwanted attention that female characters get (gifts and pickup lines). This treatment of female characters has been noted many times before, however a point that Natalia makes which is often not reported is that it’s only a very small minority of players who make playing as a female character unpleasant—sometimes only one in the whole virtual world. After all, when she plays as a male character and doesn’t hit on female characters, no one comments at all—it isn’t regarded as unusual. This would seem to suggest that intervention by the community management team could eliminate the problem, at least in relatively small virtual worlds.

Daphne Desser undertook a small study ³⁶ of women who mask their gender in virtual worlds, to judge the effectiveness of female rhetorical practices in an environment where gender was not a factor. The theory is that in real life, women use social connections through language, gestures (nodding, smiling, and so on) and linguistic tags that show agreement; this is because they have diminished power in patriarchal real-life societies. In virtual worlds, where no one need know a player’s real-life gender, such defensive insertions should not be necessary; women have the same social and physical presence as men. Previous studies of chat groups had suggested that even when women are gender-masked by pseudonyms, they are still more vulnerable than gender-masked men to harassment. This would mean that somehow people were picking up on the fact that they were “harassable” (if not that they were female); if women did not change their discursive style of interaction even when playing cross-gender, that could perhaps explain it.

Desser analyzed transcripts of gender-masked chat sessions from LinguaMOO ³⁷, and discovered that real-life females do indeed still use social connections much more than real-life males; their behavior in virtual worlds in this respect was not substantially different to their behavior in the real world. Desser sees this as an indictment of the power that real-life expectation wields to cause people to conform to gender roles ³⁸.

She also, however, looked at the behavior of male characters playing as female characters, to see whether their behavior was equally unchanged. Perhaps surprisingly, she found that it wasn’t: The vocabulary of males playing as females alters in a way that the vocabulary of females playing as males doesn’t. This may well be because men have a stronger stereotype to work to than women; on the other hand, it could be that men are more willing to behave differently than women. It would take a much wider study to determine whether men were, in practice, any better at cross-gender play than women, however ³⁹.

The differences between male and female characters has been the subject of speculation for many years. Back in 1994, Lynn Cherny examined ⁴⁰ behavioral differences in JaysHouseMOO, a small virtual world where many of the players knew each other in real life and didn’t switch gender. She discovered that in this context, men and women treat their virtual bodies differently: Women engage in hugs and other cuddlesome contacts with one another, but otherwise don’t interact “physically” much beyond that; men, conversely, have a much wider range of emotes, some of which evoke quite violent imagery. When women use such emotes, it’s almost invariably when in the presence of men who are using them; Cherny suggests they’re adapting to male behavior so as to fit in.

Cherny’s work pointed to differences between male and female behavior in non-anonymous virtual worlds. Whether these differences could be used as “gender fingerprints” to spot cross-gender playing in a virtual world with anonymity is another matter, of course. Spotting cross-gender playing isn’t easy even when you’re looking (which, most of the time, people aren’t).

In Kathryn Wright’s survey, players were asked whether they felt they could tell the real-life gender of characters accurately. She found:

• 50% believed they couldn’t.

• 31% believed they could most of the time.

• 19% believed they could sometimes, but couldn’t at other times, depending on the skills of the person doing it.

As I indicated in Chapter 3, it’s actually not all that hard to pass as someone you are not (irrespective of their gender) if you know what you’re doing. I’ve played characters incognito for years at a time. Two of those that lasted for over 24 months of pretty well daily play were female, and neither was detected. I even had people asking me if I were female in real life when my female alter ego was playing a male character, their having “seen through” her charade. My experiences are not uncommon. Someone who made a real study of it and read all the literature could perhaps keep something like this up indefinitely (although quite why they’d want to do so isn’t clear).

For instance, it’s often observed that no men play unsexy female characters ⁴¹. Whether this is any different from what women do is immaterial—it’s the perception that counts. So, knowing that people commonly believe that men don’t play unsexy females, the crafty male cross-gender player will reduce his chance of being outed by masquerading as an unsexy female. The same applies to all other metrics for detecting differences between the genders, once the diligent role-player has read up on them.

Gender is an important part of the identity of many individuals in the real world. Crossing gender therefore raises important questions: Does it give people a wider appreciation of the opposite sex? Or does it give them a dangerously false impression?

In real life, gender has a biological basis. Male and female exist because that’s how human beings reproduce. Virtual worlds enable people to escape their bodies, and be whatever gender they wish.

Well, that’s the conceit. Unfortunately, although you can take the person out of the body, you can’t take the body out of the person. If the virtual person is told a joke, their real-world body may laugh; if something very upsetting happens, their real-world body may cry; if they get excited, their real-world heart will definitely beat faster and they may get an adrenaline rush. Real-world bodies can react to virtual-world events, because the player’s mind is in both places at once.

This brings us to the second activity of players of virtual worlds that interests gender theorists: virtual sex.

The practice goes by many names, including netsex, cybersex, tinysex, and worse. I’ll be following Sherry Turkle’s line and calling it virtual sex, because that’s its most accurate designation. It occurs only in textual worlds and the text channels of graphical worlds ⁴².

Virtual sex was not common in early, adventure-style virtual worlds, only really taking off with TinyMUD and its immediate descendants ⁴³. Whether this was because the worlds or the players were more social, or because there wasn’t a great deal else to do in them, is a matter for conjecture.

As for what virtual sex is, well in practical terms it’s a series of emoted actions that simulate a sexual encounter. The command

:bats the ball to Jill.

when issued by a player called Jack will create the message Jack bats the ball to Jill.

Jill can then enter the command

:returns with a sliced backhand.

to generate the message

Jill returns with a sliced backhand.

Okay, so I’m showing how you’d use emotes to play virtual tennis here, but virtual sex is the same except that it entails much ritual removal of virtual clothes and inserting virtual objects into virtual containers. If the characters involved are not binary-gendered, they make up their own ways to exchange virtual bodily fluids, usually involving intermingling super-sensitive tentacle organs.

Transcripts of virtual sex taking place are presented in Laurel Gilber’s early investigation ⁴⁴ into the phenomenon. As is usual with logs, these appear dead on the page; it looks like you really had to be there to appreciate the emotion of the moment.

Gilber’s analysis takes a post-modern approach that manages (as is almost inevitable with post-modernism ⁴⁵) to explain absolutely everything without actually telling you any more than you knew anyway. However, she does make some interesting points, most notably that because seduction ⁴⁶ is involved in these encounters they are not pornography. She also points out that in a world where AIDS is a dangerous reality, virtual sex is a much safer prospect.

But why engage in virtual sex? It’s perfectly possible to fall in love with someone remotely without having to enact a romp in the hay. In the days before email, people occasionally fell in love by exchanging correspondence with individuals whom they had never met. Even in virtual worlds, it happens through mere conversation: At one stage in British Legends I was able to trace a chain of 14 people who had fallen in love with someone who professed to love someone else. What does emoting the sex act give that whispering sweet nothings doesn’t?

In her Life on the Screen examination of the issues, Sherry Turkle focuses on the effect of virtual sex on real-life relationships. She asks whether to have virtual sex is to be unfaithful, and whether the violation of trust amounts to infidelity. Again, though, the same could be said about explicit email exchanges; it’s not something that virtual sex delivers which is unavailable by other means, and neither is it of practical use in real life ⁴⁷.

Jennifer Mulcahy examines ⁴⁸ virtual sex as an aspect of virtual romance. When people form relationships, either the relationship is real (in the minds of the participants) or it is simply role-play. In the former context, virtual sex can be used as a means to seek out potential partners for relationships, or to consummate/strengthen an existing relationship; in the latter context, it can be a personal exploration of sexual identity or a way to have some relaxing fun. Clearly, there are problems if one virtual lover thinks they’re in a relationship and the other doesn’t.

Mulcahy also looks at FurryMUCKs, some of which have a very sexually liberal culture ⁴⁹. In these, players take on the role of humanoid animals, reacting as animals to various degrees. Citing a player who genuinely ⁵⁰ believes he’s a porcupine trapped in a human’s body, Mulcahy hypothesizes that players often don’t want to meet their partners in real life because it would collapse the images that they have built in their heads concerning their virtual lovers. In this situation, virtual sex takes place because real sex can’t. She makes the important point that the actuality of physical attraction can stop a virtual relationship—no matter how intense—in its tracks. Players enjoying the experience of a virtual relationship implicitly understand this, and may therefore be apprehensive about meeting their partners in real life.

So, is it that virtual sex is the lust that accompanies virtual love?

Shannon McRae also looks at the use of anthropomorphic characters in her discussion ⁵¹ of virtual sex, in particular the symbolism involved. Players’ choice of what animal to play reflects their attitudes to sex: A wolf, for example, is likely to be very dominant and predatory, whereas a rabbit would be timid and submissive. Virtual sex in this context allows players to explore aspects of their own sexuality; it’s practical role-playing.

The bulk of McRae’s article examines the attraction of virtual sex. She finally pins it down to eroticism, in which regard I think she’s correct. She also tries to ascertain why players find the experience so intense, but is less successful there: She applies three theories, none of which convince her (and none of which convince me, either). Personally, I see it as a consequence of immersion; then again, given my arguments earlier in this book, I would, wouldn’t I? (Virtual sex Zen moment: What is the sound of one hand typing?)

For some people, the attraction of virtual sex is pure hedonism. You can emphasize features of your sexuality without fear, you won’t catch any sexually transmitted diseases, and no one will become pregnant as a result. This view of virtual sex was popularized by Claire Benedikt in an up-beat, influential 1995 article ⁵² on the subject. With virtual sex, you get all the emotional pleasure and none of the dangers of real sex. If people try to mess you around by emoting an event you don’t like, you can quit, ignore them, or emote the inverse of the event just as easily. If someone tears off your shirt, you can snap your fingers and the shirt reappears; your virtual power is equal to that of any abuser. Role-playing is consensual; actions are only validated by the consent of all individuals concerned. If you don’t want to play along, all they can do is to go look for someone else who will.

Benedikt makes her point forcefully and with examples. Virtual sex is (in an implementational sense) intangible; anything that “happens” only happens if your mind accepts it or its consequences as real. Investing words with emotion doesn’t make them true; even if you feel violated, you haven’t been. If someone is harassing you, speak to a wizard and they’ll punish your tormentor. The only real threats are the ones that involve real life, and for those you can call on the real police.

But if you can’t prevent or undo the virtual abuse? If there are no administrators you can call on to stop it while it’s taking place? If it’s witnessed by passers-by who think you’re a willing participant?

Thus, we come to the most famous of all articles about virtual worlds: Julian Dibbell’s A Rape in Cyberspace ⁵³. Oft-reprinted not just because of its subject matter but also for the quality of its writing, its main concern is not the incident of virtual rape to which its title refers; rather it describes the way the community of LambdaMOO, where the event took place, reacted. Nevertheless, the virtual rape itself is a potent totem for gender theorists; a reagent rather than a catalyst.

Warning: If you thought cross-gender play was controversial, you’re almost guaranteed to be offended by some of the discussion that follows.

Virtual sex was quite prevalent on LambdaMOO at the time of the incident Dibbell reports, to the extent that he suggests that it was in their first act of virtual sex that many newbies experienced what would now be termed as immersion. There is a difference, however, between virtual sex in private between consenting players and that conducted in public against the will of at least one of the participants. This is nevertheless precisely what one of the LambdaMOO players did. His character’s name was Mr. Bungle.

Mr. Bungle used a virtual object called a voodoo doll to force other characters to behave in ways that their players had not instructed. He caused his victims to perform a number of debased acts, all to the accompaniment of his maniacal laughter. Because he had an audience in the form of the other players who were in the room when his assault began, his victim’s logging off would have made no difference—enough witnesses could see what was going on. For the same reason, the victim’s gagging of him would have had no effect ⁵⁴. Simply ignoring someone who is controlling your character is not an option, either. Mr. Bungle was only making intangible changes to the virtual world, but was doing so using a device with tangible functionality such that no one could prevent him. It wasn’t until a player called in an experienced old-timer who had some kind of virtual stasis gun at his disposal that the orgy of violence stopped.

Dibbell goes on to describe what happened afterward as the community struggled to come to terms with the events. Most players wanted Mr. Bungle expunged from the database (that is, the character would cease to exist), but unfortunately none of them had the power to do it. The administrators had handed over LambdaMOO to its players some four months earlier, and would only act on social issues when requested by the community as a whole. Unfortunately, the players had not formed a body to represent their views, so before they could get Mr. Bungle removed they had to create one. They disagreed over how to effect this, and in the end the wizards had to implement/impose a voting system themselves.

Many interesting points are raised here. The first is that the violation of a character was referred to almost universally as its “rape.” Although this word can technically be used in a non-sexual context (“the rape of the country”), that’s not how it was meant here: The players—those who experienced it and even those who merely had it described to them second-hand—all saw it as the non-consensual sexual desecration of an individual, for which the word “rape” was entirely appropriate. This is not to say that they felt it was the same as real-life rape, because there was no physical side to what had happened. It was a rape of a mind, but not of a body.

The community debated the possibility of responding to the virtual crime in Reality. The player behind Mr. Bungle was a student at a university with a strong antiharassment policy, and it was conceivable that had its authorities been presented with evidence of his actions (committed while using their computer) they might have removed him from his course. The players determined that a LambdaMOO crime best warranted a LambdaMOO response, but that’s not to say that players of other virtual worlds would necessarily feel the same way if something similar happened there. Indeed, as we shall see shortly, there is good reason to suppose that the crime was more real-world than virtual world.

Also, the proposed punishment amounted to character death at worst and permanent exile at best; in a social virtual world like LambdaMOO, this is about the most drastic thing that can happen to anyone. Heinous though rape is, it is second to murder in seriousness. By ordering the death of the alleged rapist, wouldn’t that mean the punishment was more severe than the crime? And what purpose was the punishment to serve—was it a warning to others, a step toward rehabilitation or merely an act of revenge?

As a developer, I can only shake my head that this situation ever arose. Non-admin players have a prevailing attitude that anything the virtual world lets them do, they’re allowed to do. The only exceptions they will countenance are those that involve freeform input (for example, speech) that the programmers can’t reasonably be expected to prevent. Over time, they come to appreciate other rules of behavior that exist, too, but the default is that if the world lets you do it, it’s allowed. If it weren’t allowed, the programmers would just stop it, right?

Consequently, it came as no surprise to me at all to learn that a virtual world in which players could control other players’ characters would, lo and behold, lead to a situation where one player did exactly that. Why would they do it? In the Mr. Bungle case, Dibbell suggests it was because the perpetrator was a sociopath: He looked on the virtual world as a laboratory for experimenting with other people’s emotions, a viewpoint that is only perhaps excusable when held by total newbies.

A Rape in Cyberspace succeeded in drawing the attention of researchers and journalists to virtual worlds (Dibbell himself is a journalist), its ultimately positive spin doing much to help people take virtual worlds seriously. Virtual sex was not presented as a sordid activity undertaken by emotionally crippled adolescent males; players were depicted as responsible individuals genuinely shocked by the actions of a madman. The critical differences between the real and the virtual were thrown into relief.

Players are real, characters are virtual. Immersed players are their characters—they’re personae—and therefore any assault on a persona is an assault on the player. The assault can only be emotional or intellectual, because those are the only aspects of the real that can be affected by the virtual. A player who doesn’t understand the difference between a character and a persona (for example, someone who thinks “it’s just a game”) may assault other characters—and therefore players—without fully realizing the consequences. A player who does know the difference can perform an assault without censure if the context of the virtual world reasonably permits it (for example, it’s clear from the genre ⁵⁵ that characters can be raped). A player who knows the difference and isn’t in a virtual world with a culture that allows such assaults, yet who still goes ahead and does it without provocation anyway, is indeed a sociopath ⁵⁶.

As to whether these assaults violate real-world law, that’s a matter for the lawyers. In a real world where people can be successfully sued for inflicting emotional distress by failing to deliver on a promise to meet Jesus Christ in the flesh ⁵⁷, who knows what a gullible player might come to believe? Personally, I feel that anyone who enters a virtual world takes a certain emotional risk in doing so, and that they’ve no more right to sue than has a member of the crowd at a golf match who is hit by a stray ball. By the same token, though, any golfer who deliberately or irresponsibly hit a ball at someone knowing it would cause them physical damage could expect to be sued; anyone in a virtual world who deliberately or irresponsibly causes real-life emotional distress beyond that which a reasonable individual could expect to endure in the context of that world deserves all they get.

For those who think that freedom of speech rules would act as a defense, consider a different incident of virtual rape reported by Elizabeth Reid in her 1994 Cultural Formations thesis. This occasion did not involve emote commands: Rather, the “rapist” depicted the violations he ⁵⁸ was committing through use of the “shout” command. The effect on the other players was actually worse than in Dibbell’s example, because of the nature of the virtual world in which it took place. JennyMUSH ⁵⁹ was set up to help the victims of real-life sexual assault or abuse overcome their trauma, and developed into a strong, supportive community where people could talk through their problems openly with others who had suffered similar violations. Consequently, when the rogue player waited until there were no administrators playing before changing gender to male, name to Daddy, and then screaming graphical and violent messages of virtual assault, it was calculated to cause maximum psychological damage. If someone had used a public address system at a psychiatric hospital to manifest the worst fears of everyone attending a recovery group, they could expect to be punished; the same should be said when people choose to conduct similar real-life violence through actions undertaken in virtual worlds. New prospective players of JennyMUSH now have to be recommended by two other players and give their real-life contact details to the system administrator before they are admitted. They remain anonymous from the other players, but can be called to account if they violate trust.

Simply the best discussion I have read about virtual rape is by Richard MacKinnon ⁶⁰. He takes the view that the concept of “rape” is a social construct that varies from culture to culture; in plain English, what members of a society understand by “rape” defines what rape is in that society. The LambdaMOO players who witnessed Mr. Bungle’s actions or heard them described almost universally viewed it as rape, therefore in that community it was rape.

Virtual communities differ importantly from real-life ones, however, in that their members don’t have bodies. The wholesale importing of real-life concepts of rape into virtual worlds is therefore a bad idea: Instead, players should reconstruct the notion of rape within the context of the virtual world. Furthermore, because there are theories of rape, this can be done in a guided fashion—learning from real life, rather than repeating its mistakes.

MacKinnon describes the main prevailing theories that explain the causes of rape in real life. These are:

• Feminist. Rape is the result of deep-seated social traditions dominated almost totally by men.

• Social learning. Aggressive sexual behavior is acquired through repeated exposure to it.

• Evolutionary. Rape is a manifestation of the different reproductive priorities of the sexes.

Whichever of these is used in the context of virtual worlds, it must be broad; virtual worlds involve no body, therefore any definition of rape that requires one is inapplicable. In this regard, some of the wider feminist viewpoints are most appropriate, defining rape to be any “damage to the self” inflicted by physical, emotional, psychological, or material means. This being the case, rape (of the mind) is indeed possible in virtual worlds. The question is, therefore, how to prevent it?

In considering his solution, MacKinnon notes that Mr. Bungle’s actions were automatically accepted as events in the real world by the observing players. Had Mr. Bungle used his voodoo doll to make his victim appear to burst into flames, the players would not have accepted this as a possibility and would have mocked him for it. Putting this in terms I have used elsewhere in this book, the players would have thought spontaneous human combustion to be “unrealistic;” why, therefore, did they not look on the virtual rape this way? MacKinnon’s answer is that they were using a definition for rape taken from the real world that didn’t fit; it would have been better were they to create a new definition appropriate to their virtual world. Claire Benedikt’s seemingly blasé suggestion of deflecting the attacker’s weapon by deferring to the consensual nature of the virtual world’s reality is actually the solution. Rather than accepting rape to be an assault on the mind even when there is no body, players should merely point out the obvious: You can’t sexually assault someone when neither you nor they have a body. The act of attempting to do so is ridiculous, and therefore so is anyone making such an attempt.

MacKinnon accepts that during any period where new definitions of rape are coming into acceptance there will be problems, but the eventual reconstruction of the concept for virtual worlds will be far more helpful in the long term. I’m not convinced that it’s this simple: Immersion means that character and player are, in a sense, one persona, so an assault on such a character is an assault on a player. Unless they consented to it, it really could be damage to their self that amounts to rape under this broad definition ⁶¹. To respond as MacKinnon suggests would mean unimmersing, but this could be problematical for some players if they didn’t realize what was happening in time.

MacKinnon makes another interesting point concerning Mr. Bungle’s punishment. He is of the opinion that punishing a character is no punishment at all unless the player has a stake in that character. Mr. Bungle’s player didn’t care about the Mr. Bungle character, therefore any sanctions imposed on him were worthless. As Mr. Bungle’s player deliberately hurt his victim’s player ⁶², the offense was in real life, not in LambdaMOO. Perhaps it would have been better if the LambdaMOO denizens had asked for the offending player to suffer the consequences of his actions in real life? I subscribe to this same view, while also noting that deleting Mr. Bungle might nevertheless have served some small purpose in serving as a deterrent for other players contemplating similar rapes.

I also ought to state that personally I don’t feel that the violations delivered by Mr. Bungle amounted to rape. They were certainly cruel and certainly unpleasant, but to label this incident as an example of perhaps the second-worst ⁶³ crime a person can commit in Reality tends, in my opinion, to trivialize the latter rather than raise the former’s significance.

I mention all this for two reasons. Firstly, it’s the brief of this chapter to report on what researchers in other fields think about virtual worlds, and this is one of the areas in which there is particularly constructive activity. Secondly, it raises a general point concerning the responsibility of virtual world designers for what goes on in their worlds. Consider the “rape” command in a text-based virtual world.

Yes, you heard me.

Look, someone in your virtual world—probably a newbie, but it could be a more experienced player—is going to try the command “rape.” They can “kill Xyzzy” and get away with it, murder is a greater crime than rape, and so it should be okay to try “rape Xyzzy,” shouldn’t it? Or maybe they’re doing it in an educational virtual world because, hey, their watching friends in the real world will think they’re big. However they reason, someone is eventually going to do it. You, as the virtual world’s designer, have to decide how to respond.

Here are some typical solutions:

• Do nothing. The response to the command “rape” is the same as would be generated by any other non-command. This is the solution most designers take. They do so either because they didn’t consider rape to be an issue, or they did but it was too hard so they ducked it, or they have some rationale along the lines of “I won’t dignify it with a response.” None of these are acceptable. It is important, it shouldn’t be avoided, and you are giving it a response (specifically, that you don’t mind having in your world people who think rape is okay).

• Issue a response telling the player off. This has the advantage of showing that you don’t approve of this kind of behavior, but the disadvantage of being toothless. The would-be rapist can simply laugh it off.

• Issue a warning, and also alert the customer service staff. A 20-minute lecture by an unamused admin undertaken in a disconnected sin bin room is enough to persuade most of the “I just wanted to see what would happen” brigade that no, actually, you didn’t. This works, but administrators don’t like doing it, especially if they end up being hated for “not being able to take a joke” as a result. It also runs the risk that people will try it simply to gain attention.

• Give the warning tangible force. Don’t just say you think that this is inappropriate for your virtual world, but (temporarily) damage any character that tries it. Now they know you’re serious.

• Give the command tangible meaning. They want to assault another character, okay, so invoke whatever combat routines the virtual world uses. Couple it with one of the preceding responses (it works particularly well with an automatic hit point loss). Hey, if you attack someone while your pants are round your ankles, you’re bound to be at a disadvantage….

• Mete out instant PD on any character that issues the command. If the player wants it back, they can come cap in hand to the administrators and beg. Only after a suitable display of contrition will their character be resurrected. This zero-tolerance approach makes its point, but can be overkill. Furthermore, experienced players will persuade naïve newbies to try the command, then laugh their socks off at the result. The newly dead newbie, however, will suffer all the usual effects of PD and may be hurt as a result. Their punishment is beyond their crime.

• Do any of the above, and warn the intended victim. They can then take action to avoid the odious individual concerned in future. This sounds a good idea, but in practice it isn’t; it can be highly disturbing if, during the course of normal play, you suddenly receive a message telling you that your character has just successfully resisted an attempted rape by some other character. There is a strong possibility that a player would eventually exploit such a rape command solely to hurt their victim—even if it meant they’d eventually lose their own character as a result.

I’ve had numerous debates ⁶⁴ on this subject over the years concerning MUD2’s approach to the command, and have tried various different solutions. The one I have settled on is “tangible meaning,” without informing the intended victim that the attack came from a “rape” command. This means that the perpetrator suffers the consequences of their action, but they’re the only person who does. On the whole, people prefer not to see the word “rape” coming from their virtual world, even if this is at the expense of not being informed when someone tries it ⁶⁵.

Returning at last from the fringes of Gender Studies back to the mainstream, how can attitudes to virtual sex (and virtual rape) best be summed up? As with cross-gender play, there are basically two points of view:

• The radical view is that virtual sex victimizes women. There is a continuum of harassment, with virtual rape at the end-point.

• The liberal view is that virtual sex liberates women. Sex can be as frequent and different as you want, and is often better than in real life.

Neither of these is satisfactory, as Nicola Döring points out in her excellent critique ⁶⁶. Victimization sees women as always having the status of objects, and does not admit the possibility that virtual sexual experiences can exist and be enjoyed; liberalization idealizes women’s choice and control, and does not admit the possibility that there may be a power imbalance between the genders. Döring proposes a third, empowerment model that addresses the concerns of the other two. Central to this is the notion that sexual empowerment is an individual learning process, as well as a process of political emancipation. Women should question their needs, and explore them critically; in other words, through their actions in the virtual world, they should develop a better understanding of themselves.

Yes, that does sound strangely familiar. Isn’t it nice when all these theories come together?

Back in Chapter 4,“World Design,” I referred to J. C. Herz’s 1994 zeitgeist-capturing book, Surfing on the Internet. Her experiences with virtual worlds cover several chapters, as she moves from newbie to seasoned veteran ⁶⁷. When she begins, she has experience of IRC but not much idea of what’s possible in a virtual world. She starts with a look at the education-oriented MicroMUSE ⁶⁸, then moves on to the more socially oriented LambdaMOO. There, she meets a character called Thorin who gives her a tour of the place.

Herz leaves LambdaMOO and explores many other virtual worlds over the coming weeks, playing several quite extensively. Finally, she returns to LambdaMOO no longer a newbie, and seeks out the friendly Thorin. At first he doesn’t recall her, having showed around many guests in his time, but when she jogs his memory he remembers. He then furthers her education by demonstrating how to create her own virtual space within LambdaMOO. Then he hits on her.

As a guest, Herz was an “it.” Now she is experienced and has created her own character, she’s a “her.” When Thorin helped her as a guest, he didn’t know she was female (although he suspected it). Now she’s confirmed her real-life gender and accepted his further help, he thinks she’s obligated to him.

Well yes, maybe she is, but she’s not obligated that much. When she turns down his advances, Thorin’s mood changes drastically. He subjects her to a torrent of power-trip emotes, describing the mean things his character is doing to hers and how her character is reacting. Herz subverts this in exactly the manner described by Benedikt, undoing his every virtual act with an emote that returns to the status quo, all the while pointing out his foolishness with smiles and laughter. She ends with the comment, “Ooh. MUD harassment. Think I’ll have to see a shrink about this. Not.”

Now that’s the attitude.