Race and Ethnicity in Fandom: Praxis

Outside oneself in World of Warcraft: Gamers' perception of the racial self-other

Thomas D. Rowland and Amanda C. Barton

Saint Louis University, St. Louis, Missouri, United States

[0.1] Abstract—The popularity of massively multiplayer online role-playing games (MMORPGs) has created a unique, heavily populated virtual reality wherein player characters are explicitly differentiated by the physical characteristics of their avatars. To investigate the way real-life race perceptions influence these adopted player-character identities, we invited MMO players to participate in an online survey. In this study, we are particularly interested in overlap, or deviation, between real-life racial perceptions and the perception of fictional fantastic races (elves, dwarves). On the basis of the data collected, we found that whether players consciously associate themselves with their avatars or consciously dissociate themselves from their avatars, real-life racial tendencies unconsciously manifest through players' choices of their avatars and in their interactions with other players within the game environment.

[0.2] Keywords—Avatar; Fantasy; Identity; Massively multiplayer online role-playing game; MMORPG; Race; Survey

Rowland, Thomas D., and Amanda C. Barton. 2011. "Outside Oneself in World of Warcraft: Gamers' Perception of the Racial Self-Other." In "Race and Ethnicity in Fandom," edited by Robin Anne Reid and Sarah Gatson, special issue, Transformative Works and Cultures, no. 8. http://dx.doi.org/10.3983/twc.2011.0258.

1. Introduction

[1.1] The recent wealth of scholarship dedicated to the rising field of online ludology and Internet culture confirms that those in the profession of studying literature and communication can no longer afford to neglect the medium of the Internet for cultural expression and interaction, and as a vehicle for storytelling. While early projections of the Internet had supposed it would remain largely a niche medium for utopian and virtual self-expression in a limited sense (Nakamura 2002, 2008), the post-Internet age into which we have entered has erased any such expectation (Nakamura 2008). The once-minor worlds of online role-playing games have expanded into their own macrocosmic digital universes and are no less a global and significant presence than commercial Web sites, Internet commerce, worldwide telecommunications, and information exchange.

[1.2] The extensive world building and design essential to these games, mostly taking their underlying conceptual format from the pen-and-paper role-playing games of the 1970s, means we can begin to investigate questions of real-life phenomena as they may or may not appear in virtual form. We have chosen to examine identity formation, particularly as it manifests in the question of race, racism, and virtual race. The design and selection aspects of massively multiplayer online role-playing games (MMORPGs) mean that the question of identity is entirely intentional, negotiated, and discursive. The design of a player's avatar, or in-game representative figure, through which the player interacts with the virtual environment and all other players in the game, means the player may assume an identity completely distinct, or "Other" than his or her real-life race, socioeconomic status, physiological traits, and personality.

[1.3] As Taylor (2003) points out, the design of these virtual avatar bodies is the end result of a long process of intentional design, resulting in what he calls "intentional bodies," and the choices in the design are intended to credit "legitimacy" to these bodies as real. Nakamura (2008), in her work on racial bodies in what we might call the Internet age (alluding to her latest identification of the post-2000s digital culture as post-Internet age), described the phenomenon of Internet avatar bodies as a sort of virtual tourism; players, regardless of their real-life race, gender, or physiological traits, could experience a simulacrum of life through the perspective of an alternate entity. However, the economic and cultural conditions giving rise to the Internet at this time were such that white became the expectation for any Internet user, regardless of his or her profile or avatar characteristics (Nakamura 2002). This is in keeping with what Feagin (2010) calls the "white racial frame" of late 20th-century American race studies and racial perspectives.

[1.4] In Nakamura's later work on Internet racism (2008), she retracts the description of white default for the Internet, mostly because of the rapidly accelerated expansion of the Internet, although she replaces it with the notion of neoliberal color blindedness, which is, she says, in its own way a form of racism. Nakamura is drawing on Bonilla-Silva's term color-blind racism, which is, in his understanding, a late 20th-century American manifestation of racism that does not recognize itself as racist, at least for those within the dominant (or privileged) race. In light of Nakamura's work on race on the Internet, we pose the question: do these MMORPGs, in particular World of Warcraft (WoW), which is the largest, although not the oldest, reintroduce notions of racism into interavatar interactions, or, under the ideas of white default and color blindness, subvert and contain the idea of race altogether?

[1.5] Following this idea about race and racism in MMORPGs, we formulated a survey to invite responses from players concerning what we hoped might be both unconscious and conscious notions of identity and race in WoW. Originally our work was not limited to WoW; we also invited responses from other popular MMORPGs, including Everquest, Age of Camelot, and Lord of the Rings Online, and we furthermore assumed that the concepts of identifying with an avatar would be familiar to respondents, as this phenomenon has been used in the majority of digital games and constitutes a major component of best-selling games such as Mass Effect and DragonAge.

[1.6] Our work on avatar identity formation and the phenomenon of intentional bodies follows a long line of excellent work. Taylor (2003) shows how designers work to create realistic bodies that allow for adequate player immersion but within a particular vision of the world that results from designers' experiences in real life. Turkle (1984, 1995) has pioneered the field of decentered subjectivity in avatars through computers as vicarious thinkers and through psychoanalytic examinations of the many new dimensions through which information is consumed in cyberspace. Pearce (2009) explores the way in which intentional communities prefer to maintain their group identity despite the termination of its virtual setting. She begins with a thorough and excellent investigation through the history of avatar design and virtual environment studies. Boudreau (2007) approaches identity formation as a process of networked interactions, focusing on its involvement of multiple identities, which includes, but does not privilege, the player-avatar relationship. More recently, Gatson (2011) has problematized the notion that online identity is less authentic than off-line identity by insisting that the phenomenon of a named, virtual self is not distinct from off-line markers of identity.

[1.7] Following Nakamura's (2008), Bonilla-Silva's (2010), and Feagin's (2006, 2010) description of late 20th-century American racial perspectives off-line (in real life) and online, we define race as a classification of humans (the species is important, as it distinguishes from virtual gaming in which race in fact indicates speciation) defined by similar, heritable physical characteristics and shared history and traditions, and thus both biological and cultural elements are present. This corresponds closely with the definition given in the Oxford English Dictionary, which ties the concept of identifying common physical traits to common descent from the work of Blumenbach (De Generis Humani Varietati Nativi, 1775). We believe this conception of race has become common in the vernacular.

[1.8] The most important aspect of race by this definition is that it is visible. For human race, physical characteristics such as skin color and nose shape are the most prominent markers of race (Alcoff 2006). Nevertheless, race is a problematic term; its definition in both the vernacular and academic senses is constantly evolving. The term can refer to a national or ethnic group, which may or may not display physical variation. Contemporary racial theorists now view race as a fluid cultural construction, defining it as more performative than physical, a combination of visible characteristics and behaviors (Alcoff 2006; Omi and Winant 1994). Furthermore, in interpersonal interactions, race is formed by both automatic responses and controlled reactions; even if the controlled responses are suppressed, this does not prevent these automatic reactions from establishing certain prejudices (Devine 1989). Bonilla-Silva (2010) describes race as a phenomenon of social construction, but one that insists, rather than denies, on its social reality. In this, it exists, like class and gender, as categories, and is undergirded by "racial structures" that reinforce racial privilege.

[1.9] In American society, a color-blind, default white perspective—what Feagin (2010) calls a "white racial frame"—obscures automatic or perceived racially differentiated responses, and equates actions and attitudes associated with white culture as positive. Race is performative on both the conscious and unconscious levels. Awareness of these levels can determine one's ability to pass as another race, but also allows one to parody the representation of another race. Nakamura (2002) identifies this as virtual tourism, or "passing," as far as it occurs online, particularly when a person (usually white) tries on or passes as another race. The second race can only be experienced as the Other, and thus no real vicarious experience can be had. A similar phenomenon from history (though not to suggest that this has passed from our society) is blackface, which follows a comparable pattern of exoticism or mocking, but not a true vicarious experience. Like transgendering, passing is not well tolerated by most spectators, though the personal act of passing does not cross lines of taboo.

[1.10] In the virtual environment of the MMORPG, the term race is used for the speciation of characters. During character creation, the player is required to select one of several races, providing a fundamental physiological framework upon which to add further defining traits of gender, name, and class (the character's skill set). This act of self-identity construction produces the avatar, who acts in preprogrammed behaviors according to the player's control, and often within several distinguishing, racial behaviors. For instance, dwarves may utter a bowdlerized oath in a dialect roughly approximating the Scottish dialect. Thus, though the differentiation of avatars crosses the species boundary (they are not all human), the two major components of race are represented in the physiological and behavioral differentiation. In most games, race furthermore carries a cultural, or ethnic, context, as members of the various races share common cultural frames: dwarves, for instance, in mining; elves in a symbiotic relationship with nature; trolls in rude speech and primitive tribal social structures. In fact, most MMORPGs operate within an inherent institutional racism, since players are first exposed to racially homogenous societies (the homelands of various races) that do not encourage (or sometimes allow) interracial interaction. This is particularly prevalent in WoW, wherein new characters are introduced in racially segregated environments and only egress these environments to encounter multiracial societies.

Figure 1. Player choices of race in character creation in World of Warcraft. [View larger image.]

[1.11] These races include traditional fantasy races from literary traditions, such as elves, dwarves, and trolls. Some races are original to various games, but are mostly hybridizations of man and beast (the Tauren race) or variations of traditional races (the Blood Elves) (note 1). Humans are one race among many, and this race usually does not feature more specific racial identification, such as black or white. The classification of this speciation as race complicates the situation further, as the term includes entities that might be considered monstrous (e.g., Taurens). In some cases, these races indicate a type originally human, but exposed to degenerating or transformative conditions, as the undead vampires or zombielike animate, sentient corpses. Within the context of these games, however, these fantastic races constitute a distinguishable type from the normal (i.e., human), without necessarily implying an evolutionary connection. They also share similar heritable characteristics, originate from the same virtual location once the player starts playing the game, share an in-game narrative of a racial history, and form a culture the way humans do. Therefore, these types operate according to our conception and to vernacular understandings of race; even as human, all in-game races maintain alterity from real-life categories and act as Other.

[1.12] The negative side of these types of social identification is the phenomenon of racism—that is, discrimination based on perceived or self-identified race, and prejudicial judgment based on racial stereotypes. Anna Everett (2005) shows that many of the video games currently available tend to utilize stereotypical human characters: epic heroes tend to be white, while martial artists tend to be Asian. Meanwhile, there are few black characters. Thus, she posits, we interact with games through "an understanding of encrypted, or encoded, meanings that represent desirable gaming heroes naturally as predominantly white, and victims and antagonists naturally as nonwhite 'others'" (313). Through the use of such ancillary gaming supports as magazines, she identifies the problem in many games: "Standardization practices increasingly reify or naturalize nonwhite characters as objectified third-person Others whose alterity is so irremediably different that ideal players would have little to no incentive to adopt them as avatars or skins" (316). She further claims that games such as Imperialism (1995) elicit what the postcolonial theorist Abdul R. JanMohamed identifies as "a complete projection of [the gamer's] self on the Other: exercising his assumed superiority, he destroys without any significant qualms the effectiveness of indigenous economic, social, political, legal, and moral systems and imposes his own versions of these structures on the Other" (Everett 2005, 316). Everett's assertion agrees with Patricia Devine's (1989) identification of racial impulses as automatic: players introduce several of their racial impulses into the game environment in terms of their own character creation and in terms of player-player interaction. Given that these impulses are automatic and without awareness, and considering the alterity of all races in WoW, we question whether WoW's attention to race/speciation highlights racism or subverts all identity as passing. These encounters, mediated through avatars (not infrequently with voice, through secondary chat programs) and as heavily differentiated as they are, might prompt unintended racial reactions—that is, racism. At the most basic level, this occurs as a differentiation of white and nonwhite.

[1.13] With the increasingly diverse array of avatars, few or none of which are characteristically light-skinned, or indeed even human, it becomes difficult to differentiate white from nonwhite. This phenomenon is not rare: Nakamura (2002) describes the experience of trying to find diversified avatar characteristics even in minority-oriented online chat rooms. Instead, players perceive two different relationships: the real self and the avatar, and the avatar (virtual self) and the Other (Turkle 1995; Pearce 2009). The first begins with the player's first experiences in the game, and the act of initiating this relationship through character creation is accomplished partially with deliberation toward the second relationship. In other words, the player constructs a new identity with a mind toward how this character will be perceived as hostile or friendly, or helpful or unnecessary by other players. The negotiation of the first relationship to the second underscores the significance of physical differentiation among avatars, and in so doing emphasizes the importance of racial relationships in the game.

2. Using a survey to discover racial perspectives

[2.1] To test the potential presence of a default white racial frame in character creation and interavatar interactions within MMORPGs, we sought data directly from players within these online communities by the use of an online survey. Bonilla-Silva discusses the use of surveys to identify racial ideologies, which, he says, are "produced and reproduced in communicative interaction" (2010, 11). He identifies the potential limitations of surveys; many of these utilize antiquated questions that draw on theoretical foundations from the 1950s and 1960s. Most importantly, surveys as they normally appeared (with fixed answers) limited the capacity of the respondent to provide original or unprompted answers.

[2.2] The scope of our project was limited by our inability to sift through the seemingly endless bulletin boards, fan fiction, wikis, and elusive in-game comments. Rather than comb through all of these products of player-publisher/programmer collaborative media, we invited players to provide us directly with structured feedback through a survey designed to assess players' perceptions of race. We invited all gamers, regardless of their real race, ethnicity, or nationality. Respondents were filtered only by language, as the survey was presented only in English. The survey was available for anyone who wished to participate for the month of January in 2010. During that time, we received 446 responses, mostly from players of the well-known online game World of Warcraft. We posted invitations to participate in the research survey on prominent bulletin boards and online publications for MMORPG fan communities, including WoW Insider (http://wow.joystiq.com), Ultima Online Forums (http://www.uoforums.com), Dark Age of Camelot (http://www.darkageofcamelot.com), and MMORPG.com (http://www.mmorpg.com).

[2.3] This survey was created using the Forms feature of Google Docs and comprised 18 questions divided into three topical sections (Appendix A). The first prompted respondents to enter information on the MMORPG and the avatar he or she played most often. The second gathered information on a secondary character, an avatar unique from the first that the player considered just as crucial to his or her self-identity as the first. The third section gathered information on the respondent's real-life identifiers in terms of race and demographic descriptors (age, gender). Each section furthermore asked respondents to identify racial or cultural descriptors of his or her in-game avatar race and real-life race. These data were intended for comparison to highlight areas where in-game racial perspectives mirrored real life, so that we might discover correlation.

[2.4] We had hoped for responses from a wide cross section of real-life races, with which we could begin to identify correlations of real-life and in-game racial perspectives of the Other. By comparing the two, we could begin to see how certain real-life racial perspectives might be imposed in-game fantasy racial relations. For this purpose, we relied on key descriptors (table 1) culled from studies of racial prejudice (Devine 1989; Krueger 1996; Wilson 1996; Taylor and Stern 1997; Dhingra 2003). Some of these descriptors included aggressive, arrogant, ambitious, hard-working, intelligent, lazy, musical, violent, self-confident, selfish, borrowed from Krueger's study (1996) of racial group perceptions in black subjects, and common stereotypes of Asians from Taylor and Stern's study (1997) of commercial advertising in America. In addition, respondents were asked to identify gross generalizations of an openly racist perspective of both the real-life races and in-game races. While these questions were somewhat leading and suggestive of racial perspectives, respondents seemed to indicate more confusion in the application of these questions to in-game race; thus, we cannot conclude any reliable results from these data, and we draw no conclusions from questions 5 and 19. Unlike the traditional surveys that Bonilla-Silva (2010) criticizes, this survey included open-ended answers. We acknowledge that such questions still cue specific types of answers, but in this case, this strengthened the survey rather than limited it. A survey with only open-ended questions would, we assume, only have confused respondents. For this survey, these overtly racial questions, such as 5 and 19, led respondents to give original answers in question 20 that provided much more relevant reflections.

Altruistic, philanthropicObnoxious
Civilized, sophisticated, educatedPeaceful, assuaging
CriminalProud, arrogant, narcissistic
DemureSelf-sacrificing, team player
Dull, unintelligentSelf-serving, selfish
Effeminate, feminineSexually desirable
Eloquent, well-spokenSexually perverse
Guided, ambitiousSpiritual, heavily religious
Hard-workingSubmissive, weak
Intelligent, studiousUncivilized, barbaric, savage
LazyViolent, aggressive
Masculine, manlyPolite, cooperative

Table 1. Key descriptors used in survey.

[2.5] From the responses we received, we aggregated the numbers to indicate the racial breakdown of players and their avatars, and we speculated as to possible reasons for correlations of real-life and in-game race. Furthermore, from the hundreds of personal original comments provided by the survey respondents, we attempted to provide an image of online racial perspectival frames among players and the ways in which interracial relationships are complicated through the phenomenon of players electing specific racial identities distinct from real life. We analyzed the results of the 446 responses to determine the racial composition of the players and of their avatars, and we looked for indications of racial perspectives of in-game races that possibly mirrored real-life racial perspectives. Data analysis revealed several weaknesses in our study structure. After consulting with our institutional review board, we included a notice in our informed consent form that survey participants must be over 18 years of age, and we included only age categories over 18 years of age in our survey. Our choice to use an anonymous public survey made it difficult to filter respondents by age, so we could avoid collecting and publishing data on minors. Furthermore, our preliminary research focused specifically on American racial categories, but because of technical limitations of the Google Docs form environment, we were unable to filter the nationality of respondents. Therefore, we were unable to analyze the data within a strictly American cultural context, and our efforts to understand data possibly gathered from an international community were unavoidably complicated. Despite these limitations, we think that the survey provides important preliminary data about gamers' racial perceptions and any relationship to the character creation process, and reveals the importance of further studies that will utilize more rigorous sociological methodology to collect data from online gaming communities.

3. Questions of race and identity

[3.1] While interacting with(in) these games, the player negotiates multiple self-identities simultaneously. For Filiciak (2003), avatars enable pseudonymous management of multiple selves: "On the Internet people perceive us as our avatar, and such a perception is highly appropriate because when I play I am more my own avatar than the person sitting by the console/computer. To specify which of these identities is more true or more false is probably impossible. In any case, it would appear that our virtual 'self' is closer to our images of ourselves than the one we present" (93). This approach to identity agrees with Boudreau (2007), especially when Filiciak attributes this convergence of self-identity and avatar identity to the fluidity of the postmodern hyperidentity, one in which we are more concerned with a divergence of many identities specific to certain social situations—identities through which we move freely. The liquid identity is apparent by the correlation of self-descriptions and avatar descriptions and in the comments provided by our respondents.

[3.2] As we see in the comments below, the negotiation of the real self to the Other and the avatar is vital, and in many cases, it is a deeply personal experience for players. The decisions in creating an avatar are viewed in some ways as crafting an extension of the player's self-identity. Jos de Mul (2005) proposes to reallocate Paul Ricoeur's theories of self-identity through the use of video games; players are continuously engaged in the resolution of the entanglement of the "dialectics between the self and the other" (255). In the case of MMORPGs, the Other is recognized primarily as the avatar and secondarily as other players as elements in the virtual environment, which exist alongside virtual terrain, questing elements, and monsters. The first exists in the manipulation of the virtual environment within the player's personal computer or console; the second exists entirely within the virtual environment and results in players congregating into communities, called guilds, for the purposes of mutual enabling and social interactions, including friendship. These communities tend to be long-lasting and take on identities or brands of their own; members will often retain close connections across various gaming platforms (Pearce 2007). However these guilds form, they highlight the interactive and participatory culture of the MMORPGs, defining the game experience as necessarily communal. These interactions are influenced primarily by visual encounters between avatars based on perceptions of character skill set (class) and physical speciation (race). Thus, game experience is discursive in the constant perception and observation of the Other as player in the game, a discourse of evaluating other characters on the basis of visual cues against their potential for support in accomplishing common goals (Neitzel 2005).

[3.3] By these three elements—the negotiation of the self and Other through character creation with an eye toward future encounters of self (avatar) and Other (virtual environment, players); the phenomenon of self-forming guilds; and the primarily visual interface of the game—we can conclude that the virtual MMORPG gaming environment is predicated on the consequences of racial interactions and represents a real struggle of an individual of self and Other (world) played out in a hyperbolic visual manner. Furthermore, we can establish that crises of self (avatar) and Other (players) must be viewed as sympathetic or adversarial. In evaluating the hostility of the Other in this world, the player must rely primarily on visual cues of physical characteristics. Thus, decisions and interpersonal, or interavatar, relationships can be made solely along the same mechanisms by which we as humans understand race. We can safely conclude, therefore, that these games are predicated thoroughly on issues of race.

[3.4] We understand such a topic can be exceedingly difficult to navigate, considering the highly emotional and sensitive reactions respondents may have to questions of race, racism, and prejudice. In this regard, we had mixed responses from the respondents: while respondents found some of the questions too overtly racist and thus hesitated or declined to answer, respondents had no hesitation answering other questions that were potentially just as racist, likely because they disassociated fantastic race as it appears in the game with real-life racial sensitivity. In other words, respondents recognized racially discriminatory questions when they dealt with real-life race, but failed to recognize equally discriminatory questions when they focused on in-game race. Furthermore, because of the structure of the survey, the respondents were unable to see the later questions until they had finished previous questions, and they were unable to revise their answers. At the end, when respondents were given the opportunity to provide original unguided comments, several mentioned they could now see that in-game race discrimination was in fact very similar to real-life racism.

[3.5] The majority of respondents (97 percent) indicated a primary avatar from WoW (note 2). We anticipated a larger response from WoW players, as our experience indicates that members of this community are often more willing to communicate their interest in the game through outside projects such as ours. We did try to encourage a more balanced response by posting invitations on forums catering specifically to games other than WoW. Furthermore, we received additional publicity from a generous notification published in WoW Insider (formerly WoW.com), run by our request, though no remuneration was made (note 3).

[3.6] The majority of respondents (89 percent) identified themselves as white (Caucasian). Of these respondents, 32 percent played as Elven (18 percent Night Elf and 14 percent Blood Elf), and 16 percent played as Human (the remaining 52 percent are spread over various other races). Only 1 percent of respondents identified themselves as black. These respondents played only three races, the majority being Human (40 percent) and Tauren (40 percent). Another 3 percent of the respondents identified themselves as East Asian, and of these, nearly half (42 percent) played as Elven. Hispanic (3 percent), Native American (1 percent), and Other (3 percent) made up the rest. Hispanics played mostly as Elven (50 percent); Native American respondents each chose a unique race, but of these, all but one are arguably monstrous (for example, Troll, Tauren, Undead, and Draenai each represent a hybridization of man and beast, or a perversion of man). Figure 1 illustrates the results of players' responses.

[3.7] Although only 31 percent of respondents indicated their real-life gender as female, the avatar genders were much more balanced: 44 percent of respondents indicated their primary avatar is female. With secondary characters, female avatars were dominant at 52 percent (see figure B.3).

[3.8] White players described their avatar as sophisticated (37 percent), clever (40 percent), hard-working (37 percent), intelligent (37 percent), self-sacrificing (40 percent), cooperative (40 percent). This corresponds closely with how these players chose to describe themselves, with sophisticated, educated (58 percent), clever (57 percent), intelligent (71 percent), polite and cooperative (63 percent). While some descriptors probably viewed as negative were less frequent (e.g., criminal, 1 percent of respondents), some had a higher percentage of respondents than expected, such as sexually perverse (given with no elaboration provided), submissive, and violent each gaining 7 percent, 7 percent, and 4 percent, respectively, by white players. These showed even higher frequency in minority players. Some apparently positive qualities were less frequent than expected: sexually desirable (12 percent), guided, ambitious (19 percent), and spiritual (7 percent). Interestingly, while white players described their avatar as hard-working (19 percent; only 3 percent identified their avatar as lazy), players tended to describe themselves as either hard-working (31 percent) or lazy (30 percent).

[3.9] White players chose descriptors for their race predictably along those traits that make successful gamers, though they are in no way intrinsic to avatar choices, according to game mechanics. In other words, these races are not particularly handicapped or enhanced to succeed in the game: better at physical activity (35 percent), better at teamwork (37 percent), and better at reaching goals (34 percent). In choosing descriptors of their real-life race, white players tended to characterize their race as better at earning money (27 percent) and reaching goals (28 percent). Figures B.1 and B.2 break down these results.

[3.10] Players tend to value characteristics in their avatar that are practical in the virtual environment of the game and choose these same descriptors for themselves. In doing so, the respondents (or players) ascribed an intentionality or agency to their avatars that is not intrinsic to the game since, in terms of game mechanics, no in-game race is more clever or intelligent than any other, nor can any avatar sacrifice him- or herself without the player's intent to do so. Thus, in identifying the characteristics of their own (and we may assume therefore the ideal) avatar, the respondents in fact were providing the description of an ideal player, and hence clever, sophisticated, intelligent, self-sacrificing, and cooperative. In ascribing this sort of agency to the avatar, the player is essentially projecting his or her own identity onto the avatar.

[3.11] Several of the respondents' original comments reinforce the avatar-as-self identification. One respondent writes (all respondent's comments appear as typed): "I am not a role player by any stretch of the imagination so trying to describe the characteristics of my toon [avatar] was not easy for me to do. I basically based off of myself for the most part because I do not change my behavior when I am online." Another respondent indicated "I don't Role Play in games, so generally what my character is like is dictated by the class and my personality" (emphasis added). Since the game's programming rarely attributes a certain set of behaviors to the avatars directly, based on initial creation, the implication of this respondent is that gamers also project certain behavioral obligations to certain classes: healers must heal, and thus must produce an empathetic personality. However, behaviors produced by any given class will itself vary from player to player, thus indicating that while the player perceives a certain behavior-per-class expectation, no such standard exists objectively.

[3.12] At times, the association of player-avatar goes deeper. One respondent mentions a friend "who will only play blood elves because he is admittedly vain—he only wants to be beautiful. I myself typically prefer taurens due to their large size, I'm quite above the average in height and weight IRL [in real life] and to me being represented by a character that isn't head and shoulders above the crowd just seems…odd."

[3.13] We also encountered variations to this pattern. Some players seemed to sense that we were anticipating this avatar-race perception and tried to offer comments to the contrary. In so doing, they used implied racial perceptions of the fantasy races. For instance, one commenter said, "Never judge a character based upon appearance. More often than not you will be wrong. I've known stupid Ogres that Indian doctors [played] in real life, and an elven druid that turned out to be a black football player." Another offered: "I don't know any undead [creature] player that doesn't play like an evil bastard." Often players failed to realize that when they were making judgments on characters because of extreme differences in physical appearance, they were in essence making racial differentiations. Characters that looked "especially awesome" or had "visual appeal" were as important in decision making as class opportunities.

[3.14] For white players of WoW, this meant focusing on elven or human characters because of their visual appeal. The physical appearance of elf and human (light skin; long, straight hair; tall and slender) vaguely correlates with white characteristics. We do notice that elves and humans tend to exhibit historical similarities with cultural narratives of white/Western races. Players adopt these races as vicarious agents in real-life racial prejudices (human and elf as privileged, and nonhuman/elf as monstrous, liminal, or disenfranchised). This accomplishes what Bonilla-Silva calls "softly otheriz[ing]" nondominant races (2010, 3). While the aggressive tendencies of humans tend to be viewed as righteous or honorable, similar tendencies in nonhuman races are described by one player as "genocidal." The stereotyping of various other races comes out in jokes or taunts: "elves being lithe and arrogant, perhaps, or dwarves being gruff and industrious," and "blood elf males are all pansies, gnomes should be punted." One respondent acknowledged race stereotyping by claiming to work against it, where her "troll characters [are] different from the expectations people have for their race, just to make a character more interesting and developed in my role play. So, even though many troll roleplayers…play their trolls violently, or their gnomes good at mechanics, I specifically don't." Another player accepted that there is "a fair amount of stereotyping based on the race a person chooses. Blood elves are associated with females or gay men. The same could be said for female undead. Orcs, male undead, and male tauren are considered the most masculine of characters and it seems to me they attract the type of people that are the most insecure about their physical masculinity or body image." This comment was offered by a white woman in the 18–24 age bracket, and it is interesting in conjunction with the relative rarity of white, black, or East Asian respondents selecting "sexually desirable" or "masculine" as self-descriptive terms. (The comment also invites speculation as to how the player may be aware of other players' lack of positive body image.)

[3.15] Players also seemed aware of their associative identity of avatar-player. One respondent said, "I consider my characters symbolizing a certain part of me." Several respondents imply a metaphorical relevance to the events of online game play with that of real life by associating in-game actions with real-life decisions: "I do not roleplay my characters. What I do and say online is what I would do or say normally"; "I play my character mostly how I would act in real life"; "I think I play my character a lot like I am in real life." The imposition of a person's real-life personality into a fantasy environment via the avatar is suggested by another's comment that "I suppose I created my character as an ideal version of what I would like to be, if I were in that environment." This avatar-as-self identity persists despite opportunities to reinvent or renew the process to more advantageous results: the commenter continues, "If I were to start over [I would create a different character], but with the time and effort I have invested in this character, I'm sticking with her." This persistent avatar-as-self identity is echoed by other players who indicate that having few avatars is preferable to experimenting with many different avatar-identities.

[3.16] Sometimes players authenticate this avatar-player association by supplying a narrative backstory. In one sense, this echoes de Mul's (2005) appropriation of Ricoeur's identity theory via narratives, in that we construct a sense of self-identity through identifying with characters of a narrative (mimesis3). One respondent offered, "I like my ingame characters to have a background story. I have even written some pages about the story of my characters, tho it isn't finished yet and I have much more that isnt written down on paper. I tend to give my characters names that somewhat fit their race. I also enjoy collecting items and such that increase the characteristics of the ingame character." (The perception of appropriate names-to-race suggestion and how players perceive this correlation are also of some interest.) This sort of identity construction is particularly practical within online role-playing games: while the nature of game represents playable narratives (Neitzel 2005) through its nature (mimesis1), the plot arc in the game through questing provides the mythos of coherency for the player-avatar, that while the avatar actually experiences the virtual challenge, the player through the avatar can appropriate that mythos into their own experiences (hence players expressing the actions of the avatar through the first-person pronoun) (mimesis2). Through the creation of a narrative onto the avatar, the player is in fact appropriating a narrative for the self (mimesis3) (Ricoeur 1984; de Mul 2005).

[3.17] Yet in the dual nature of the avatar, it is still a puppet, or a mask (à la Jung) (Filiciak 2003). Some players are aware of this identity association and specifically spoke against it. In several cases, respondents spoke of the attempt to separate player and avatar: "Most of my characters are not direct extensions or representations of myself, though they may embody a few traits that I don't or can't have but would like to experience"; "For the most part I try (emphasis on try) to keep myself separate from the characters I create"; "I see my characters as characters, not as avatars of myself. I usually make up a little backstory for them (that I never use, I don't role-play) just because that feels natural to using a character." In one case, the respondent chose not to choose descriptors for his character, saying "my character is just a puppet."

[3.18] While female players were more aware of and vocal about the fetishization of the female avatar, male players were more willing to explore "transgendered" (strictly defined as men playing female characters and women playing male characters) avatar-player relationships by playing as female characters, while seemingly unaware of a reason for this. Male respondents commented on the trend; for example, "Also i think girls tend to play female characters, and men tend to play both." One might speculate that male players unconsciously fetishize the avatar, adding another layer to the avatar-as-self identity, viewing the avatar not only as a projection of their narrative self-identity but also as a sexual object, a relationship that female players are aware of but do not engage in (figure B.4).

[3.19] Finally, several comments addressed one mitigating factor in the selection of race through character/avatar creation. In many online role-playing games, several classes (character skill sets) are limited to certain races. Any discussion of interpersonal (or interavatar) interactions and perceptions of race in virtual realities must take account of players' immediate perceptions of class and the restrictions on race based on class preference. While class does heavily influence the player's experience with the game, its influence matches at least the importance of race. Class tends to influence a character's utility, especially regarding group abilities and mechanics of group questing, while race tends to evoke an emotional or less rational reaction: dislike of the undead, hazing of short races, avoidance of orcs and trolls.

[3.20] It is gratifying to see one's research have an impact on the participants. Several respondents indicated that the survey made them aware of how real-life race relations can parallel fantasy race relations. One respondent offered: "I like that your survey has brought up the fact that we consider human ethnicities to be 'races' in the same way MMO games consider dwarves, night elves and humans to be different races. Poignant!" Another said, "I wouldn't want to stereotype about a human race in this way—which is now making me wonder about doing it for a fictional race"; "I like to consider my character a projection of the personality traits I like best about myself. However, I have never consciously considered a relation between the appearance and perceptions of my character's race and that of my own."

4. Conclusions

[4.1] In this project, we elected to gather data from gamers themselves to determine the amount that real-life racial perspectival frames have influenced virtual racial interactions. Following this line of approach, we ask whether WoW maintains phenomena of late 20th-century and 21st-century racism, including white default racism, and whether WoW attempts to subvert questions of race through means of neoliberal color blindness and passing, and in so doing, contains the undesirable aspects of racism.

[4.2] To answer this query, we propose, in looking at the comments and data aggregated from our respondents, that the data confirm how scholars have formulated avatar-player hyperidentification, and that by this phenomenon of players projecting their self into the avatar, players inadvertently reintroduce a visible notion of race into the game. However, because players recognize a differentiation of the avatar as a "self-Other"—a mask or puppet—the construction of interavatar relationships as inherently racist goes largely unrecognized by players. This is in keeping with Bonilla-Silva's (2010) description of American racism as being self-denying. Hence, our conclusion is to find aspects of both possibilities operating within online games: players softly otherize fantasy races but are unaware they are doing so. Concurrently, they own the behavior of their avatar but deny the consequences of their avatar's actions. Some of this may be attributable to the inherited tradition of many of these fantasy races from other media, including fantasy literature, online fan fiction, and cinema. However, we maintain that this racial perspectival frame is mainly linked to real-life racial ideologies, because even these inherited traditions are linked to the default white racial frame of contemporary American culture.

5. Appendix A: Survey

We have reproduced the questions from the survey. Respondents took the original survey in a Google Docs form, which prompted them to answer one section of questions before they were able to see the next section of questions, and so on. The page breaks below indicate the original section breaks.

The survey consisted of 5 pages after the informed consent form, which partially outlined the scope of the project and asked for respondents' permission to publish their responses. Respondents were allowed to withdraw at any time from the survey, according to specifications set out by the institutional review board of Saint Louis University.

Page 1: Survey of Race in Massive Multiplayer Online Role-Playing Games

Thank you for helping us gather information on fantasy races in games. Please pass this link on to your gaming friends and help us get a lot of feedback!

1. Characters: We would like you to answer some questions on what kind of characters you play. For this section, please answer according to your primary character (this can be whichever character you consider foremost: perhaps the highest level, the most recent, your favorite, the oldest, one for which you've written stories, or the one with the most game-hours logged). You will have the opportunity to comment on other characters later.

This character was created for: (please select one game)

World of Warcraft Star Wars Galaxy
Lord of the Rings Online Dark Age of Camelot
Everquest Ultima Online
Everquest II Other

---------------------------------------- End of page 1 --------------------------------------------------

2. Races: Please choose the appropriate race for your character. Question 2 asks the survey participant to choose the race of his/her character from a drop-down menu based upon the game chosen in question 1. See pages 4 and 5 for a breakdown of races by video game.

3. Character gender: What gender is your character?



Other (gender is not relevant to this race)

4. Character-player association: How long have you played this character:

Less than a weekLonger than 3 years
Approximately 10-100 hours Longer than 5 years
Approximately 100-500 hours Longer than 10 years

5. Race strengths: How would you describe your character's race?

Better at physical activity Better at making things
Better at earning money Better at individual confrontations (pvp)
Better at doing harm Better at talking trash, intimidating
Better at sneaking about Better at leading a group
Better at deceiving Better at communicating
Better at asserting itself Better at teamwork
Better at producing popular music, dancing Better at reaching goals

6. Race characteristics: How would you describe your character?

Altruistic, philanthropicObnoxious
Civilized, sophisticated, educatedPeaceful, assuaging
CriminalProud, arrogant, narcissistic
DemureSelf-sacrificing, team player
Dull, unintelligentSelf-serving, selfish
Effeminate, feminineSexually desirable
Eloquent, well-spokenSexually perverse
Guided, ambitiousSpiritual, heavily religious
Hard-workingSubmissive, weak
Intelligent, studiousUncivilized, barbaric, savage
LazyViolent, aggressive
Masculine, manlyPolite, cooperative

7. Secondary characters: would you like to repeat these questions for a secondary character?

Yes—continue to second section of survey

No—continue to last section of survey

---------------------------------------End of Page 2 -------------------------------------------------------

[Pages 3 and 4 were used to allow respondents to answer the above questions for a secondary avatar.]

14. Secondary character: Please indicate your reason for creating a secondary character. If none of these apply, please select "Other."

I wanted to try out a different race.

I wanted to try out a different gender.

I wanted to play the game from the beginning again.

I wanted to have a different experience with the game.

I lost, killed, or maxed out my primary character.

I wanted to finish off the subscription I had paid for, but didn't want to use my primary character.

I created it to appeal to someone else (a guild, a friend, a significant other, etc.).


--------------------------------------- End of Page 4 ------------------------------------------------------

15. Player gender: Please indicate your gender.



16. Player age: Please indicate your age.





17. Player race: I would describe myself as:

White, Caucasian East Asian
Black, African American Middle Eastern/Arabic
Hispanic Other, please specify
Native American

18. Player descriptors: I would describe myself in real life as:

Altruistic, philanthropicObnoxious
Civilized, sophisticated, educatedPeaceful, assuaging
CriminalProud, arrogant, narcissistic
DemureSelf-sacrificing, team player
Dull, unintelligentSelf-serving, selfish
Effeminate, feminineSexually desirable
Eloquent, well-spokenSexually perverse
Guided, ambitiousSpiritual, heavily religious
Hard-workingSubmissive, weak
Intelligent, studiousUncivilized, barbaric, savage
LazyViolent, aggressive
Masculine, manlyPolite, cooperative

18b. Additional answer: Would you like to offer more comments?

19. Race descriptors: I would characterize people of my race as:

Better at physical activityBetter at making things
Better at earning money Better at individual confrontations (pvp)
Better at doing harm Better at talking trash, intimidating
Better at sneaking about Better at leading a group
Better at deceiving Better at communicating
Better at asserting itselfBetter at teamwork
Better at producing popular music Better at reaching goals

19b. Additional answer: Would you like to offer more comments?

20. Additional comments: Would you like to share any personal anecdotes or comments on gaming experience that is pertinent to this topic?

Thank you for taking our survey. Your input has helped us to get a good glimpse into the online community of gamers!

6. Appendix B: Aggregated results

Figure B.1. Responses of all respondents when asked to identify traits of themselves, broken down by race. [View larger image.]

Figure B.2. Responses of all respondents when asked to identify traits of their individual avatar, broken down by race. [View larger image.]

Figure B.3. Real-life gender of respondents, broken down by race. [View larger image.]

Figure B.4. Percentage of respondents indicating their experience playing transgendered, broken down by race. [View larger image.]

7. Acknowledgments

[7.1] An early version of this work was presented at the 31st International Conference on the Fantastic in the Arts in 2010. Our thanks to Ruth Evans for her insightful comments and guidance, and to Kara McBride, who helped navigate the intricacy that is the institutional review board. We also thank both friends and the many survey respondents who helped by providing us with personal anecdotes, comments, and explanations of game mechanics and the experience of role-playing online. All data collection processes and procedures were approved by the institutional review board at Saint Louis University.

8. Notes

1. See Blizzard's official World of Warcraft Web site for images and a description of Taurens: http://us.battle.net/wow/en/game/race/tauren.

2. Several players noted more than one game as their primary interest; our software did not prevent respondents from selecting multiple games. These data are reparable, however, as race selection of the subsequent questions was game specific.

3. Our research invitation was published on WoW.com in a January 13, 2010, entry, "The Classifieds: News Briefs on Guilds and Players": http://www.wow.com/2010/01/13/the-classifieds-news-briefs-on-guilds-and-players/.

9. Works cited

Alcoff, Linda Martin. 2006. Visible Identities. New York: Oxford University Press. doi:10.1093/0195137345.001.0001.

Bonilla-Silva, Eduardo. 2010. Racism without Racists: Color-Blind Racism and the Persistence of Racial Inequality in the United States. Lanham, MD: Rowman and Littlefield.

Boudreau, Kelly. 2007. Pixels, Parts, and Pieces: Constructing Digital Identity. Saarbrücken, Germany: VDM Verlag Dr. Mueller.

Dark Age of Camelot. 2001. MMORPG. Fairfax, VA: Mythic Entertainment.

de Mul, Jos. 2005. "The Game of Life: Narrative and Ludic Identity Formation in Computer Games." In Handbook of Computer Game Studies, edited by Joost Raessens and Jeffrey Goldstein, 251–66. Cambridge, MA: MIT Press.

Devine, Patricia. 1989. "Stereotypes and Prejudice: Their Automatic and Controlled Components." Journal of Personality and Social Psychology 56 (1): 5–18. doi:10.1037/0022-3514.56.1.5.

Dhingra, Pawan H. 2003. "Being American between Black and White: Second-Generation Asian American Professionals' Racial Identity." Journal of Asian American Studies 6 (2): 117–47. doi:10.1353/jaas.2004.0004.

DragonAge: Origins. 2009. Computer game. Edmonton, AB: BioWare.

Everett, Anna. 2005. "Serious Play: Playing with Race in Contemporary Gaming Culture." In Handbook of Computer Game Studies, edited by Joost Raessens and Jeffrey Goldstein, 311–26. Cambridge, MA: MIT Press.

EverQuest. 1999. MMORPG. San Diego, CA: Sony Online Entertainment.

EverQuest II. 2004. MMORPG. San Diego, CA: Sony Online Entertainment.

Feagin, Joe. 2006. Systemic Racism: A Theory of Repression. New York: Routledge.

Feagin, Joe. 2010. The White Racial Frame: Centuries of Racial Framing and Counterframing. New York: Routledge.

Filiciak, Miroslaw. 2003. "Hyperidentities: Postmodern Identity in Massively Multiplayer Online Role-Playing Games." In The Video Game Theory Reader, edited by J. P. Mark Woolf and Bernard Perron, 87–102. London: Routledge.

Gatson, Sarah. 2011. "Self-Naming Practices on the Internet: Identity, Authenticity, and Community." Culture Studies, Critical Methodologies 20 (10): 1–12.

Guild Wars. 2005. MMORPG. Bellevue, WA: NCSoft.

Krueger, Joachim. 1996. "Personal Beliefs and Cultural Stereotypes about Racial Characteristics." Journal of Personality and Social Psychology 71 (3): 536–48. doi:10.1037/0022-3514.71.3.536.

The Lord of the Rings Online: Shadows of Angmar. 2007. MMORPG. Westwood, MA: Turbine, Inc./Midway Games.

Nakamura, Lisa. 2002. Cybertypes: Race, Ethnicity, and Identity on the Internet. New York: Routledge.

Nakamura, Lisa. 2008. Digitizing Race: Visual Cultures of the Internet. Minneapolis: University of Minnesota Press.

Neitzel, Britta. 2005. "Narrativity in Computer Games." In Handbook of Computer Game Studies, edited by Joost Raessens and Jeffrey Goldstein, 227–45. Cambridge, MA: MIT Press.

Omi, Michael, and Howard Winant. 1994. Racial Formation in the US: From the 1960s to the 1990s. 2nd ed. New York: Routledge.

Pearce, Celia. 2009. Communities of Play: Emergent Cultures in Multiplayer Games and Virtual Worlds. Cambridge, MA: MIT Press.

Pearce, Celia, and Artemesia. 2007. "Communities of Play: The Social Construction of Identity in Persistent Online Game Worlds." In Second Person: Role-Playing and Story in Games and Playable Media, edited by Pat Harrigan and Noah Wardrip-Fruin, 311–18. Cambridge, MA: MIT Press.

"Race, n.6." OED Online, June 2011. Oxford: Oxford Univ. Press. http://www.oed.com/view/Entry/157031.

Ricoeur, Paul. 1984. Time and Narrative. Vol. 1. Translated by Kathleen McLaughlin and David Pellauer. Chicago: Univeristy of Chicago Press.

Star Wars Galaxies. 2003. MMORPG. San Diego, CA: Sony Online Entertainment.

Taylor, Charles, and Barbara Stern. 1997. "Asian-Americans: Television Advertising and the 'Model Minority' Stereotype." Journal of Advertising 26 (2): 47–61.

Taylor, T. L. 2003. "Intentional Bodies: Virtual Environments and the Designers Who Shaped Them." International Journal of Engineering Education. 19 (1): 25–34.

Turkle, Sherry. 1984. The Second Self: Computers and the Human Spirit. New York: Simon and Schuster.

Turkle, Sherry. 1995. Life on the Screen: Identity in the Age of the Internet. New York: Simon and Schuster.

Ultima Online. 1997. MMORPG. Austin, TX: Origin Systems/Electronic Arts.

Wilson, Thomas. 1996. "Compliments Will Get You Nowhere: Benign Stereotypes, Prejudice, and Anti-Semitism." Sociological Quarterly 37 (3): 465–79. doi:10.1111/j.1533-8525.1996.tb00749.x.

World of Warcraft. 2004. MMORPG. Irvine, CA: Blizzard Entertainment.

Transformative Works and Cultures (TWC), ISSN 1941-2258, is an online-only Gold Open Access publication of the nonprofit Organization for Transformative Works. TWC is a member of DOAJ. Contact the Editor with questions.