(a month or so ago I posted “a longish review”:/2008/06/13/more-on-coming-of-age-in-second-life/ of Tom Boellstorff’s book Coming of Age in Second Life. Tom has now whipped up this occasional contribution to expand some of his thoughts on the topic — enjoy! -R)

The Setup
1. I have been thinking about writing this essay for some time (in connection with an edited book project in its early planning stages), but was inspired to write it in this form in response to the interesting conversations that have appeared on this blog and elsewhere in regard to my book Coming of Age in Second Life: An Anthropologist Explores the Virtually Human. This essay is an experiment in two ways. First, in it I discuss emerging genres of ethnographic research with regard to virtual worlds. Second, I use the blog format to post a draft, in the hope that the feedback I receive will help me in writing the final version of this essay. There have been several recent experiments using blog-based peer review to complement more traditional forms of peer review. The best example of this to my knowledge is the case of Noah Wardrip-Fruin’s forthcoming book Expressive Processing: Digital Fictions, Computer Games, and Software Studies. Part of the “experiment,” then, is to see how the comments I receive with regard to this essay help me revise it.

1. Here is how I have set things up with these goals in mind. I have numbered the twenty paragraphs of this essay to facilitate commenting upon specific passages, and have kept the essay as succinct as possible. I see this essay as sharing genre features with the “conference talk,” a time-tested way to present work in progress. I have thus omitted endnotes, a bibliography, and most references to Coming of Age in Second Life, despite the fact that this essay extends arguments already present in that book. I encourage any and all comments on the essay, and would appreciate it if you could provide your name with your comment. If your comment inspires me in any way during the revision process, I will thank you in the acknowledgments section that will appear in the final version of the essay. I envision that final version being about twice the length of this draft, and of course as including endnotes and a bibliography.

The Background
3. Coming of Age in Second Life is an ethnographic study of the virtual world Second Life, as well as an analysis of the place of virtual worlds in human sociality more broadly. Since its unofficial release on April 21, 2008 and particularly since its official release on June 18, 2008, I have spoken in graduate seminars, been interviewed on radio stations, and appeared at a number of meet-the-author events (many of them taking place within Second Life itself). It is hardly a coincidence that in the same month as the official release of my book, the inaugural issue of the Journal of Virtual Worlds Research was released. All this indicates that we are at a moment in which a new research community is coming into being, a community whose growth will only be stimulated by the continuing emergence of new virtual worlds, from Age of Conan to Google’s Lively. This research community includes a wide range of persons, from those who have researched virtual worlds for decades to students just beginning to think about new projects. It is an interdisciplinary research community, including persons from many academic disciplines, persons working in nonprofit and industry contexts, independent scholars, designers, journalists, and residents (these are, of course, not exclusive categories). Two key questions that emerge around this new research community (and in all likelihood, all new research communities) are: (1) what is the object of our study?, and (2) what do various methodologies bring to the table in researching this object of study?

1. As with the case of my previous publications on Indonesia, it has been interesting to see what aspects of my analysis attract attention from different readerships. In both my Indonesia work and my Second Life work, discussions have emerged around questions of methodology and particularly the constitution of “the fieldsite.” I find this gratifying, because I have intentionally designed all my research projects to push on the boundaries of what we mean by “the fieldsite.” In Indonesia I have conducted research on gay Indonesians on three islands (Java, Bali, and Sulawesi), but in my books The Gay Archipelago and A Coincidence of Desires I discuss how, in a powerful sense, this research is not “multi-sited”: the fieldsite is “Indonesia” itself. This is because gay Indonesians have historically seen themselves as gay “Indonesians,” not gay Javanese, Balinese, and so on. There are many different kinds of spatial scales operative in human life, including local, national, regional, and global, and it is crucial not to equate culture with locality. Sometimes that equation is valid, sometimes not: it depends. Translocal cultural logics exist with regard to everything from religion to gender. In the case of gay Indonesians, while they may think of themselves in local terms with regard to some aspects of their lives, with regard to homosexuality they typically think of themselves as Indonesians. This makes sense given that the concept of gay subjectivity is associated with modernity, not learned from one’s parents or tradition, but the linkages to the nation turn out to be much more complex. I have suggested that one reason so little has been written on gay Indonesians is that these persons fall outside one’s analytical horizon if that horizon is founded in the spatial scale of locality. Researchers who equate culture with locality in Indonesia can miss the forest for the trees, so to speak: they will see all kinds of cultural logics that are local, but those that are translocal in some fashion will appear to be inauthentic impositions.

2. My interest in challenging meanings of “the fieldsite” was strongly supported by my dissertation advisors at Stanford. One of these advisors, Akhil Gupta (now at UCLA) has co-written some key works about notions of the fieldsite with James Ferguson. James Ferguson was my colleague at Irvine for several years before moving to Stanford. I still have many colleagues at Irvine who support my interest in rethinking “the fieldsite,” including George Marcus, whose work on multi-sited ethnography is well known. When I began conducting research on virtual worlds, I thus came to the project with a prior interest in problematizing conceptions of the fieldsite. I soon realized that the conceptual tendencies with regard to virtual worlds were strikingly opposed to those I had encountered in Indonesia studies. Whereas in Indonesia studies, the presumption was in the direction of locality, in the study of virtual worlds the presumption has been in the direction of translocality. For instance, there were (and still are) persons claiming that all virtual-world research projects must include meeting persons in the actual world to be valid, or at least that such projects must always privilege the actual-world lives of virtual world residents.
   
   The “Four Confusions”

3. As I discuss in Coming of Age in Second Life, both virtual worlds and the study of virtual worlds have histories to them. However, as noted above, the study of virtual worlds as a research community is in a formative stage, paralleling the recent growth in virtual worlds themselves. In such contexts of emergent inquiry, debates over definitions and terminologies are common, as a glance at the Table of Contents for the first issue of the Journal of Virtual Worlds Research indicates. A difficulty in moving these debates forward is the remarkably negative attitude toward virtual worlds found not just in some quarters of anthropology, but even in science and technology studies. In part this may be due to the simultaneously utopian and dystopian narratives that often co-occur with new technologies. It may also be relevant that to date, the most prevalent popular-culture reference to virtual worlds is The Matrix movies, in which a virtual world is used to enslave humanity.

4. I define virtual worlds as places of human culture realized by computer programs through the Internet. Before turning to my typology of ethnographic methods with regard for virtual worlds, it will prove helpful to set out what I term the “four confusions” regarding contemporary discussions of virtual worlds. Most of these confusions originate in mistaking something that frequently co-occurs with virtual worlds for a necessary condition of their existence.

5. Games. Virtual worlds are not games. Historically they have been and continue to be shaped by video games; they may contain games within them; they may even be largely structured in a game-like manner; but there is no way to equate virtual worlds with games without defining “game” so vaguely as to include all social life under its purview. The confusion originates to some extent in the English-language distinction between “game” and “play,” a distinction not found in all languages and cultures. Because it is incorrect to assume, by fiat, that all virtual worlds are games, it follows that the use of theories from game studies to virtual worlds must be contextual: in some cases such theories will be highly effective, in other cases less so.

6. Visuality. Despite the fact that at present, phrases like “3D web” are frequently used as synonyms for “virtual world,” virtual worlds need not be graphical or even visual. This is seen most clearly in the fact that historically, virtual worlds were exclusively text-based (as in the case of “MUDs”). The fact that nearly all contemporary virtual worlds are built around three-dimensional graphics is fascinating and important to study, but this does not mean that such graphics are a definitional precondition for deeming something a virtual world. For instance, one could in theory have a virtual world composed entirely of soundscapes, in which persons who are blind in the actual world would be on equal footing with the seeing. One could imagine a purely haptic virtual world, in which an interface technology like a glove allowed participants to navigate and interact solely through touch. There is no indication such virtual worlds would involve more than comparatively small communities were they to come into existence. If anything, the trend toward visuality seems to be accelerating. However, it remains crucial that we avoid conflating virtual worlds and visuality. Since most contemporary virtual worlds are structured around visuality, theories from visual studies will be crucial to understanding them, but it would prove less effective to use such theories to make categorical claims about virtual worlds.

7. Mass Media. Because virtual worlds are places, they are not mass media, though they may contain mass media within them (everything from magazines, books, and embedded websites to streaming audio and video media). Virtual worlds need not mediate two or more places, since they are places in their own right. If anything, it is more accurate to think of a virtual world as a “medium,” in the sense of a material in which one crafts things. This has consequences for the use of mass media theory for understanding virtual worlds: we cannot assume ahead of time how such theories will need to be reworked for virtual-world contexts.

8. Anonymity and roleplaying. The vast majority of existing virtual worlds require that participants have accounts in which their identity differs from their actual-world identity. For instance, in Second Life I am known as “Tom Bukowski,” because while one is allowed to choose any first name one wishes, last names must be selected from a pre-defined list. However, it is not a definitional precondition of virtual worlds that they be built around anonymity. One could imagine a virtual world that encouraged or required participants to use their actual-world names inworld, along the lines of social networking websites like Facebook. As virtual worlds are used increasingly in contexts like education, nonprofit work, and the corporate sphere, virtual worlds disallowing anonymity, or at least not mandating anonymity, may become more common. Linked to this question of anonymity is that of roleplaying. Since many virtual worlds are structured partially or overwhelmingly as games, and given the historical linkages between virtual worlds and fantasy fiction like J. R. R. Tolkien’s Lord of the Rings, it is unsurprising that forms of role play are crucial to many virtual worlds. Roleplay, however, is not a necessity for defining something as a virtual world.

9. There are, of course, other possible confusions: this is to be expected given the degree to which the study of virtual worlds is an emerging field of inquiry. For instance, it is still unclear if virtual worlds must employ avatars. Text-based virtual worlds, for example, did not use avatars as typically understood: one could define the description of a resident as a textual avatar, but such descriptions seem more like a “profile” than an avatar. However, the four confusions described above seem to be the ones that most often sidetrack contemporary research agendas and discussions with regard to virtual worlds.

Research Questions and Ethnographic Design
13. With the preceding discussion in mind, I now turn to the issue of research design, setting out a three-part typology of methods for ethnographic research with regard to virtual worlds. I intend “typology” to be taken in a heuristic sense, not an exhaustive one. My undergraduate mentor in linguistics, the late Joseph Greenberg, talked about “splitters” and “clumpers” in linguistic typology: researchers who sought fine-grained categorizations versus those who worked to gather the world’s languages into the fewest possible groups. For the purposes of this essay I will be a clumper: I will heuristically group all ethnographic methods with regard to virtual worlds into only three categories. I could easily have set forth a typology with five or ten categories, but this more parsimonious typology has the benefit of brevity, and well as highlighting some key distinctions.

1. It is crucial to foreground the relationship between “research question” and “method.” Any claim that a particular method is the best (or the only valid) method for researching virtual worlds misses how research always involves a coming-together of research question and methodology. How one conducts research is not determined by some essential property “out there;” it is shaped by the research questions one wishes to investigate. In my work as Editor-in-Chief of American Anthropologist, I find that one of the most common reasons I end up rejecting a manuscript is that the research questions (while fascinating) and methodology (while rigorous) do not match up: the methods are not working to answer the questions the researcher has ostensibly chosen to examine. If I wish to study patterns of HIV infection in a certain social group, quantitative methods will prove invaluable. If I wish to understand how a certain population comes to think of itself as a “social group,” qualitative methods will in all likelihood be a better fit. Methodological partisanship is not helpful in moving these kinds of conversations forward: what ideally emerges is a research community, with researchers using different methods to answer differing research questions with regard to a shared field of interest. With this in mind, here is my “clumping,” preliminary typology of ethnographic methods with regard to virtual worlds:

2. Virtual/Actual Interfaces. One class of ethnographic methods with regard to virtual worlds explores interfaces between virtual worlds and the actual world. An excellent example of this kind of research is T. L. Taylor’s Play Between Worlds: Exploring Online Game Culture, which opens with the researcher attending a hotel convention for participants of Everquest. This class of methods builds off a history of examining such interfaces with regard to the Internet more generally. A well-known example of this is Daniel Miller and Don Slater’s The Internet: an Ethnographic Approach, which focuses upon how Trinidadians use the Internet to reconfigure Trinidadian identity and community. Since work in this genre tends to emphasize relationships between virtual-world and actual-world selfhood and sociality, a logical methodological outcome is that researchers strive to interview the same persons in the actual world as they encounter in a virtual world or worlds, and have as one research focus those instances where residents of a virtual world meet collectively in actual-world contexts.

3. Virtual/Virtual Interfaces. Another class of ethnographic methods with regard to virtual worlds examines interfaces between two or more virtual worlds. In some cases, this can be a comparative research design in which residents of the virtual worlds in question do not (or mostly do not) move between the virtual worlds being studied. This is analogous to, say, Clifford Geertz’s book Islam Observed: Religious Development in Morocco and Indonesia, in which the Moroccans and Indonesians studied do not travel between Morocco and Indonesia and are, indeed, largely unaware of each other’s existence. In other cases, this can be a research design that tracks a community or communities moving between virtual worlds. An excellent example of this is Celia Pearce’s forthcoming book Communities of Play: Emergent Cultures in Online Games and Virtual Worlds. In this work, Pearce examines the “Uru diaspora,” a community formed when the virtual world Uru shut down and residents worked to rebuild their lost virtual home in other virtual contexts like Second Life and There.com. One frequent topic addressed by research in this genre is how notions of selfhood and community are sustained and destabilized across differing virtual contexts.

4. Virtual Worlds In Their Own Terms. The third class of ethnographic methods making up my heuristic typology involves studying a single virtual world. This is the primary method I employ in my book Coming of Age in Second Life, where I refer to it as studying a virtual world “in its own terms.” If Geertz’s book Islam Observed can serve as an analogue for studying virtual/virtual interfaces, then several of his other books (for instance, his first ethnography, The Religion of Java), can serve a similar purpose in regard to studying a virtual world “in its own terms.” Geertz’s Religion of Java is primarily a study of Islam, and there are by how hundreds of insightful ethnographies of Islam, exploring Muslim life around the world. That such ethnographies usually focus on particular places and communities does not mean they ignore that Muslims are found worldwide, that many Muslims make the pilgrimage to Mecca, that persons migrate, and so on. Instead, it means that they examine how such translocal cultural logics and practices shape what is emicially understood to be a particular community or communities. It is absolutely crucial to recognize that an interest in intersectionality, translocality, and the co-constitution of cultural domains is typically common to all three classes of ethnographic methods I discuss in this essay. Some studies of virtual worlds “in their own terms” focus upon subcultures or specific topics (say, sexuality, or economics). Others strive for a more general portrait. In any case, researchers working in this genre are often interested in how shared practices and meanings emerge and are contested within a virtual world.
   
   Concluding Thoughts

5. This rough typology of ethnographic methods with regard to studying virtual worlds is meant to underscore how different genres of research design allow for exploring varied sets of research questions. Most researchers end up working in all of these genres over time, but at any point in time the best research is based upon the difficult choice of focusing one’s methods in line with a particular avenue of investigation. Arriving at a workable and compelling design is perhaps the most challenging and important step in conducting research. It is not possible to do everything at once. In the emerging research community around virtual worlds, I have sometimes encountered a misreading of my colleague George Marcus’s work on multi-sited ethnography, wherein it seems to be assumed that the more “multi,” the better. In this misunderstanding, research on a single fieldsite is, a priori, suspect or outdated, while research on multiple fieldsites is, a priori, valorized as cutting-edge. In reality, both single-sited and multi-sited methods go back to the earliest decades of ethnographic research: indeed, nineteenth-century anthropology was dominated by evolutionary approaches that presumed multiple sites of research and comparison. This is why in 1896, Franz Boas, a founding figure in American anthropology, felt compelled to write his influential article “The Limitations of the Comparative Method of Anthropology.”

6. It should be clear that all three of the methodological approaches discussed above (or the many additional methods that I could have set forth with a more “splitter” typology) are valid approaches to researching virtual worlds. All have strengths, and all involve sacrifices in terms of honing a doable research plan. In what I find to be the best research in virtual worlds or the actual world, we are moving toward forms of what I have elsewhere termed “postreflexive” modes of ethnographic engagement, focusing upon how “the fieldsite” of any research project emerges through that ethnographic engagement, rather than being set in stone “out there.” It is by now well-acknowledged that the single fieldsite is, in this sense, an ethnographic fiction. The irony is that in virtual worlds research, what sometimes appears to be less well-acknowledged is that multiple fieldsites are also so constructed.

7. What does the future hold? It appears that research on virtual worlds will continue to increase and diversify. A subset of that research will continue to be ethnographic in some sense, and this work bears every indication of representing an innovative set of contributions. Obviously, there is no need to choose between the various methods for ethnographic research with regard to virtual worlds that I have discussed above. All can be done well or badly, but none of them are by definition invalid. When properly keyed to appropriate research questions, each has something to offer. Each can contribute to building a body of ethnographic work that will help illuminate what virtual worlds are, as well as their changing place in human life. In some ways this body of work will be specific to virtual worlds, but it will continue to draw from a range of other fields as well. For instance, while (as noted above) virtual worlds are not necessarily visual and are not necessarily games, they do tend to be highly visual and often are games or emphasize play. As a result, theoretical perspectives from game studies and visual studies, among other disciplines, will continue to be crucial for understanding most virtual worlds. In turn, the growing body of research on virtual worlds, informed in part by various modes of ethnography, will have much to offer many other fields of inquiry. The conjunction of ethnography and virtual worlds will continue to stand as a vibrant field of research, contributing to central debates about human selfhood and sociality into the future.

—Tom Boellstorff is Associate Professor of Anthropology at the University of California, Irvine, and Editor-in-Chief of American Anthropologist.