230 HUMAN ORGANIzATION Human Organization, Vol. 68, No. 2, 2009 Copyright �� 2009 by the Society for Applied Anthropology 0018-7259/09/020230-11$1.60/1 Introduction Sphysical ince the emergence of the Internet, ethnographers have adapted ways of studying communities in the world to the study of their online counterparts (Boellstorff 2008 Churchill and Bly 2000 Hine 2000, 2005 Mason 2003 Rutter 2005 Schiano and White1998). Online communities take many forms including newsgroups, chat rooms, online forums, blogs, social networking sites, online games, and more. Members of online communities, although they rarely meet face-to-face, nonetheless interact on a regular basis, form relationships, share knowledge and practices, and share identities as members of a community. Massively multiplayer online (MMO) environments are a recent form of computer-mediated communication. What distinguishes them from other Internet media is that they take face-to-face conversation as their primary metaphor for user interaction, rather than, say, the page or the bulletin board. MMOs are ���massively��� multiplayer in that thousands of concurrent users can inhabit the same virtual space. Millions of people worldwide are spending an average of 20 hours From 3D Space to third Place: the Social Life of Small Virtual Spaces Robert J. Moore, E. Cabell Hankinson Gathman, and Nicolas Ducheneaut Massively multiplayer online (MMO) environments are an emerging computer technology that makes possible new kinds of distributed communities and online sociability. What distinguishes MMOs from other Internet media is that they take face-to- face conversation as their primary metaphor for user interaction, rather than, say, the page or the bulletin board. Because they simulate 3D spaces and contain thousands of people who do not know each other, MMOs constitute public spaces, although virtual ones. As such, they can be studied in ways analogous to those of public places in the physical world. Inspired by the work of William H. Whyte and Ray Oldenburg on sociability in real-life public places, we take a similar approach toward the study of MMOs. We ask the question: what makes some virtual public spaces in MMOs successful ���third places��� while other similar places fail? Through our virtual ethnography of dance clubs and corner bars in three MMO environments, we find four features of virtual public spaces that appear critical for their success: accessibility, social density, activity resources, and hosts. We further argue that MMO sociability is just as authentic as that in ���real-life��� contexts while highlighting ways in which it is distinctly different. Key words: sociability, third place, virtual world Authors��� Statement and Acknowledgements: Robert Moore is with the Multiverse Network E. Cabell Hankinson Gatham is with the University of Wisconsin, Madison and Nicolas Ducheneaut is with the Palo Alto Research Center. The research reported in this article was conducted at PARC from 2003-2007 as part of the PlayOn project. a week (Yee 2008) playing together in MMO worlds. For example, World of Warcraft alone has over 10 million active users worldwide. As a result of this number of people interact- ing on a regular basis, communities and cultures are emerging both within and between virtual worlds. And not surprisingly, ethnographic research on MMO communities is beginning to flourish (Dibbell 2006 Ducheneaut, Moore, and Nickell 2004, 2007 Nardi and Harris 2006 Steinkuehler 2005 Taylor 1999). In terms of their basic demographics, we know that for MMO players in North America, the average age is 26 only 25 per- cent are teens 85 percent are male 50 percent are employed full-time 36 percent are married and 22 percent are parents (Yee 2008). If we consider second life in particular (analyzed below), the average age jumps to 30 and the proportion of males drops to 57 percent (Linden Lab 2007). MMOs are persistent 3D environments that are usually extensive in their geographic scale: they often simulate small continents that contain cities constructed from individual buildings and wilderness areas complete with mountains, trees, rivers, and even weather. Players experience these en- vironments from a near first-person perspective by moving a 3D character (known as an ���avatar���) through the landscape. But MMOs are not simply ���worlds��� in a three-dimensional sense, they are also social worlds inhabited by other real people. Players interact by bringing their avatars into close proximity, animating them using gesture commands and talking by typing text chat messages, which often appear in comic-strip-style bubbles over avatars��� heads.

231 VOL. 68, NO. 2, SUMMER 2009 Because they simulate 3D spaces and contain thousands of concurrent users who do not know each other, MMOs constitute public spaces. As such, we argue that they can be studied in ways analogous to those of public places in the physical world. For example, in his pioneering Street Life project, William H. Whyte (1988) uses ethnographic observa- tion and time-lapse photography to study how people inhabit public spaces in city centers, such as those in New York City. He examines how the design of the physical spaces enables and constrains interaction among city goers, and asks: why do some public spaces support lively sociability while oth- ers are dead? Georg Simmel is perhaps the first sociologist to seriously examine ���sociability��� in its own right. He writes, ���Sociability extracts the serious substance of life leaving only ���together- ness,��� the sheer pleasure of the company of others��� (Simmel 1949:255). While the social sciences tend to focus on the many instrumental reasons that people interact with each other, for example, because of work or family obligations, Simmel reminds us that humans are truly social animals who sometimes seek out social interaction purely for its own sake. Simmel gives examples of ���sociability��� in the many varieties of playful conversation (e.g., anecdote, humor, witticism, or flirting) that tend to emerge around play activities such as games, music, or sports. Ray Oldenburg extends Simmel���s work by analyzing the places in which sociability tends to emerge. He calls such settings ���third places,��� and defines them as ���a generic designation for a great variety of public places that host the regular, voluntary, informal, and happily anticipated gather- ings of individuals beyond the realms of home and work [the so-called first and second places]��� (Oldenburg 1991:16). He argues that such informal public gathering places (e.g., caf��s, coffee shops, bookstores, bars, etc.) are essential to the health of communities, claiming that all great civilizations and great cities have had their characteristic public gather- ing spaces, but that in post-World War II America, which is characterized by the ���automobile suburb,��� third places have largely disappeared (Oldenburg 1991). Like Oldenburg, Whyte (1988) critiques certain urban centers, such as those of Los Angeles, for their conspicuous absence of sociable public life and identifies elements of urban design that inhibit urban sociability. While Simmel, Oldenburg, and Whyte describe activities and public places in the physical world, their approach has also been applied to the virtual. About the WELL, one of the first text-based virtual communities, Rheingold (1993:26) writes, ���It might not be the same kind of place that Oldenburg had in mind, but so many of his descriptions of third places could also describe the WELL.��� The WELL is an imagined space created purely through text description, yet even with such a simple technology, users could nonetheless construct engaging social experiences. MMOs take this kind of commu- nity to the next level by introducing simulated 3D spaces and bodies. This virtual ���material��� world provides new resources for social interaction. Oldenburg���s (1991) concept of ���third places��� has been applied specifically to MMO environments in prior research. Steinkuehler and Williams (2006) ask whether entire MMOs fit Oldenburg���s definition of ���third place��� and conclude that they do indeed. However, they also acknowledge that some situations within MMOs do not fit the definition quite as well. For example, when players participate in ���guild raids,��� group combat in which dozens work as a team for two to five hours, the social experience ���becomes increasingly more entangl- ing, time-consuming, and work-like��� and ���the function of MMOs as ���third places��� begins to wane��� (Steinkuehler and Williams 2006:903). In contrast to Steinkuehler and Williams (2006), we take an alternative approach in this study, as well as in our prior work (Ducheneaut, Moore, and Nickell 2004, 2007). Rec- ognizing the fact that MMOs are multifaceted environments in which players engage in a range of different activities, we focus only on certain kinds of places within them, namely dance club and bar-like locations. Unlike Steinkuehler and Williams (2006), we specifically examine how the design of 3D spaces themselves shapes sociability, building on work by William H. Whyte (1988). Understanding how 3D spaces and 3D avatars1 impact social interaction is critical for under- standing the social dynamics of MMOs because it is virtual 3D space that makes them unique. In this paper, we explore the relationship between virtual space, social interaction, and sociability. Like their real-world counterparts, virtual cities in MMO environments tend to lack sociable public life. We ask: what makes some virtual public spaces successful ���third places��� while other similar places fail? To explore this question, we focus on social life in three MMO environments: City of Heroes (CoH), Star Wars Galaxies (SWG), and Second Life (SL). For each, we analyze four factors that impact their success: accessibility, social density, activity resources, and hosts. Data and Methods For this study, we employed methods of virtual ethnogra- phy (Boellstorff 2008 Hine 2000, 2005). We began by gain- ing a player���s perspective through participant observation in several virtual worlds including SWG, CoH, SL, EverQuest II, World of Warcraft, EverQuest Online Adventures, and There. We inhabited each world (during different phases of the study) on a regular basis for an average of 10 hours a week, learned the gamer lingo and practices, observed lively public spaces, joined player groups and associations, formed online friendships, and more. As is true with most virtual ethnographies, we studied the players through their online personae���as embodied in their 3D ���avatars��� and text chat messages���rather than their real- life identities (Mason 2003:63). Although participants��� offline identities and lives are a very important area of research, we did not attempt to collect such data for this study. The analytic stance we take in this paper, informed by ethnomethodology, is that in order to understand the endogenous organization