Anthropology

The discipline of science concerned with the study of other cultures is anthropology. With ever-dwindling numbers of unstudied cultures, it is unsurprising that anthropologists should take an interest whenever new cultures develop on their doorstep (as they have in virtual worlds). That virtual worlds have not been flooded with anthropologists, however, can perhaps be ascribed to three factors:

• The traditional view that virtual worlds are games and therefore not worthy of serious study ¹.

• The fieldwork for virtual worlds isn’t as exotic and exciting as that promised by the real world.

• These cultures are so new, they could well be considered to be a subculture of that of the anthropologist, which makes their study sociology rather than anthropology.

Nevertheless, formal anthropological work has been undertaken for virtual worlds, some of it producing results of interest to virtual world designers.

There are several subfields of anthropology, not all of which are applicable to virtual worlds. In his excellent modern anthropology textbook ², Daniel Bates lists these as

• Biological anthropology. Also known as physical anthropology, this concerns the study of biological variations among populations. It has virtually no applicability in the context of virtual worlds ³.

• Archaeology. Archaeologists investigate the relationship between the artifacts people create and their behavior. This is unlikely to be applied to virtual worlds for many years, as the people who created the artifacts still live and their behavior can be assessed directly rather than deduced from what survives them.

• Linguistic anthropology. Linguistic anthropologists study the way language is constructed and used, and from this come to understand the nature of the culture it supports ⁴.

• Cultural anthropology. This is what everyone generally thinks of by “anthropology”: going out and living among a population for months on end, observing and recording aspects of their everyday life.

There are, furthermore, two different results of anthropological study. The first is the ethnography: This is a detailed examination of an individual culture that is purely descriptive; it paints a detailed picture of a culture, but offers no interpretation of the results. An ethnology, on the other hand, attempts to discern cultural patterns of behavior from the ethnographical data. Ethnographies describe the way things are; ethnologies explain why things are the way things are.

Ethnographic data usually begins with a demographic study of a population to determine where the cultural boundaries lie. In 1996, for example, Gamespot undertook an influential survey ⁵ of players of multi-player online games of all kinds; it uncovered different patterns of play for players of different types of games. Although useful for identifying possible areas that may be worthwhile as subjects for individual study, this kind of data snapshot has only a limited shelf life if the rate of change of player behavior is unsettled ⁶.

Virtual worlds, as (in this context) a subfield of multi-player computer games, do have a long enough history to have achieved some measure of stability. The earliest serious attempt by a non-player to collect ethnographic data about virtual worlds only was Herbert Emert’s 1993 survey ⁷, which was formally an investigation into the social effects that playing in virtual worlds had on the players themselves. It succeeded in discerning the different attitudes players had, particularly in why they played (that is, the presence of different player types).

Unfortunately, early surveys (and there were several that followed—Usenet requests for people to respond to questionnaires were a seasonal event at term paper time for several years) regarded the players of virtual worlds as a single community—something that hadn’t been true since the days of MUD1 (and even that had a distinction between its “internal” and “external” players, who played at different times of the day and rarely overlapped). Later studies ⁸ by non-players made a distinction between “adventure” and “social” virtual worlds, but still treated them as if they were communities as a whole.

Some aspects of virtual world culture are universal, of course, and others aren’t but are prevalent in many related worlds. This degree of homogeneity is caused by inheritance from parent worlds and the occasional arrival of immigrant players bringing ideas from more distantly related worlds. Nevertheless, individual virtual worlds must be regarded as the entities in which values are passed from one generation of players to the next, and it is therefore individual virtual worlds that anthropologists must study if they are to learn about these cultures.

Anthropologists studying virtual worlds will stay in a single one over time in the same way that in the real world they would stay in a single village. Expeditions to neighboring “villages” to examine differences and similarities of behavior are possible, but on the whole the research is based in one place.

It is interesting to consider whether large-scale virtual worlds that exist in multiple simultaneous incarnations have different cultures across those incarnations. The evidence seems to suggest that although there may be differences when particular servers have been set up to attract a particular playing style (for example, PvP is allowed or not), in general the culture seems to be fairly uniform. In The Norrathian Scrolls (Nick Yee’s exhaustive demographic study of EverQuest), no strange age/gender combinations were discerned for individual shards; the composition of player populations was pretty much the same from one to the next. This doesn’t mean that their players would necessarily behave the same way, of course, but it suggests that if they don’t conform, then their behavior is not sufficiently outré to affect overall population balance.

Data gathering for individual virtual worlds has been conducted either for administrative use only or as a means to some non-anthropological end. Indeed, researchers who want to look at virtual worlds for one main purpose will often pick a particular one and study it in great depth. This can be viewed as a form of anthropological study. Dieberger’s work on navigation was focused entirely on Igor MUD ⁹, for example, and Pargman’s examination of community management arose from a study of SvenskMud ¹⁰. There have, however, been anthropological studies by anthropologists, for anthropologists, of virtual worlds.

The first of these was Michael Rosenberg’s 1992 ethnography ¹¹ of WolfMOO. Rosenberg joined this virtual world as a participant observer, and spent several weeks of regular evening play getting to know and understand it. Although his report only scratches the surface of what is possible, it nevertheless sets a high standard. Reading it, you get a good sense of what WolfMOO was about and how its players behaved; it’s like a first-person travel guide to a distant city that gets the facts over without the hyperbole. The value of this kind of work is perhaps made all the more clear by the fact WolfMOO is no more; Rosenberg’s ethnography is all that remains to reflect the culture that once was.

John Masterson’s 1994 ethnography ¹² of Ancient Anguish ¹³ draws directly on Rosenberg’s work, although it’s less formal and more speculative in places. He played Ancient Anguish for two years before writing his (unpublished, but widely read) paper. Many of his observations would be regarded as unexceptional in today’s virtual worlds, but of course part of the reason for this is because pioneers like he and Rosenberg made those observations in the first place.

Masterson put his field experience of Ancient Anguish to further good use, writing a 1996 master’s thesis ¹⁴ on the sociolinguistics of virtual worlds. In addition to Ancient Anguish, he also spent extended periods in two other adventure-oriented virtual worlds ¹⁵ for comparison: Paradox II and TrekMUSE ¹⁶. For linguistic anthropologists, virtual worlds (especially textual ones) are unusual because people write in them as if they were speaking. Formally, linguistic anthropology concerns language that isn’t written down, so on the face of it textual communication via a keyboard ought to be disqualified. It’s the fact the players consider themselves to be speaking that validates it; their words are as ephemeral as spoken words are in real life ¹⁷. Logging these communications is more like recording conversations than it is reading a book.

Linguistic anthropologists have looked at non-verbal communication in some detail. In real life, people use many physical gestures other than mere speech while communicating. Bowing is so ingrained in Japanese culture, for example, that people will often do so to individuals with whom they are communicating by telephone. Picking up on the nuances of this kind can be very difficult for anthropologists, therefore it’s very useful for them to be able to look at virtual worlds where all communication is by text only—there’s only one channel to examine. The first person to examine this in detail was Lynn Cherny, who in an 11-month study ¹⁸ of ElseMOO ¹⁹ categorized five different kinds of emote command as mechanisms for conveying non-verbal communication:

• Conventional action. These are actions that have a practical effect, much like any regular action. Example: “Name waves.”

• Back channels. These occur during conversations, indicating continued attention/ understanding. Example: “Name nods.”

• Byplay. Jocular meta-comments on conversation. Example: “Name dusts off some adjectives.”

• Narration. Comments usually related to real-life events, often indicating boredom. Example: “Name munches on a ketchup sandwich.”

• Exposition. Statements describing internal or factual background. Unlike the other four categories, these can be in any tense. Example: “Name hated the movie.”

Influential though this was, Cherny was later to surpass herself with a full-blown linguistic ethnography ²⁰ of ElseMOO, exhibiting the same kind of depth and scholarship seen in any other major sociolinguistic work. It’s very impressive; the discussion of the ethics of studying virtual communities is worth the cover price alone ²¹.

More pragmatic virtual world designers than I may wonder what there is in all this for them. Clearly, the more you know about virtual worlds, the better informed you are when designing one; however, why should sociolinguistic studies of particular virtual worlds be compelling reading? The potential of virtual worlds for sociolinguists has been apparent since Eva-Lise Carlstrom’s 1992 linguistic ethnography ²² of LambdaMOO; is there any traffic of useful information the other way, though?

There is, yes; or rather, there would be if designers were to take the time to look. Anthropologists offer a unique opportunity for designers to see themselves through the trained eyes of others; it is foolhardy to ignore the objective portrait of themselves that results. Anthropologists may not like it (they try to avoid altering any culture they are studying, so for members of that culture to act on an ethnography is anathema), but that’s their problem.

There are practical advantages, too. Take sociolinguistics: By looking at why people use the emotes they do, it is possible for anthropologists to draw up structured lists of commands that address these needs; virtual world designers have previously had to evolve such lists through requests from players (plus hours of thumbing through a thesaurus). Such emote lists are only of marginal use in textual worlds, where exhaustive sets of emotes have accrued over the years sufficient to cover all situations. For designers of graphical virtual worlds, though, these sets are unusable. It may only take seconds to add a new emote to a textual world, but for graphical worlds it’s a non-trivial exercise—it could mean a trip to the motion capture studio in some circumstances (expensive enough in itself, but ruinously so when it must be modified for all valid skeleton types). It’s important that players of graphical virtual worlds have available to them the same non-verbal forms of communication they can use in real life, but graphical worlds aren’t as expressive as textual ones in this regard. Players of graphical worlds can’t use free-form emotes like players of textual worlds do (except textually), and neither they can they expect every fixed-form emote to exist—it’s too expensive. Only some emotes can be implemented: The question is, which ones?

Sociolinguistics provides some of the answers. Rather than picking what players say they want, or what the designers think the players want, it’s possible to look at what players actually use and what they need to know is available (whether or not they use it). This means players can get on with the business of communicating, instead of having to attribute new meaning to bizarre combinations of whatever emotes the designers have actually provided.

This is just one example of where linguistic anthropology can be of possible use to virtual world designers; if you read Cherny’s or Masterson’s work, others will present themselves, too.

As for cultural anthropology, that also has things to say about virtual world design. Jen Clodius is a cultural anthropologist who immersed herself in DragonMUD ²³ (the oldest continuously running TinyMUD) so extensively it could almost be said she “went native.” In a series of short papers, some of which have already been cited here, she made a number of penetrating insights about why people in virtual worlds do what they do; in this sense, her work is ethnology rather than ethnography. For example, in a 1996 paper ²⁴ she describes how the players of DragonMUD developed and maintained a community of interest, touching on issues of identity construction to suggest why this may be. Developers of large-scale virtual worlds who rake in several million dollars in subscriptions monthly would do well to engage people such as Clodius to tell them how their community works, instead of operating in the dark (or worse, overreacting to the unrepresentative minorities who often dominate the public forums).

Ethnologies teach us about other cultures. By comparing behavioral similarities across many cultures, it is possible to discern universal patterns that can then be sought in more diverse cultures, including our own. This is, for example, how Joseph Campbell formalized the hero’s journey and why George Lucas was then able to use it for Star Wars (of which the virtual world Star Wars Galaxies is a direct result).

The first major thesis ²⁵ about virtual worlds, written by Liz Reid in 1994, was an attempt to identify patterns of cultural formation common to virtual worlds. She traced behavioral tropes back to the mechanisms that underlie virtual worlds, rather than the culture of the people who played them. For example: The reason people are more open in virtual worlds than they are in the real one is because of the anonymity virtual worlds afford them; it’s not because they come from a real-world society that emphasizes openness. Players have needs that virtual worlds can fulfill; however, if they are to satisfy those needs then they must live alongside other players. This implies they must develop a set of mutual understandings that support both their goals and their relationships; these feed back to provide a context to give meaning to their actions (and themselves). These are the cultures of virtual worlds.

Reid is right: Virtual world cultures are formed by the needs of the players, acting under environmental conditions both inherent to the access medium and provided by the virtual world’s designers. I would add, however, that these cultures rarely form spontaneously any more: It is possible to shape a culture in advance of its creation, because much of what new players learn about a virtual world they pick up from existing players. The original, seeding players therefore have a strong influence on the direction a culture takes; as most of these players will have played other virtual worlds beforehand, so some aspects of culture are passed from virtual world to virtual world to virtual world. As an illustration, consider whether the attitude to cross-gender play in virtual worlds would be the same today had they been invented in a right-wing, reactionary college in the Deep South rather than a British university notorious for its left-wing politics ²⁶.

Were virtual worlds to be created from nothing, quite different cultures could develop as per Reid; nowadays, though, much culture is passed on and therefore evolves, rather than forms. Furthermore, as increasing numbers of people become involved with the Internet, virtual communities are starting to have an impact on real-world ones. Given that (as always) the virtual is governed by the real, it therefore makes sense to suppose that to some degree at least, virtual society can be regarded as a subclass of society as a whole.

This brings us to our next topic: sociology.