Archive for the ‘Online Ethnography’ Category

‘Coming of Age in Second Life’: A review essay

Thursday, February 19th, 2009

‘Coming of Age in Second Life’: A review essay

Paul Manning

(This is very much a work in progress: comment appreciated)

I will admit that I am daunted by the prospect of writing this review, few books in recent anthropology are likely to have been written about so much as this one. What can one say about something about which so much has already been said? For this reason, I will confine this review article to drawing attention to the specific things this book has to say to linguistic anthropologists, and draw sometimes idiosyncratic connections to literatures and problems that do not specifically deal with issues of virtual communities or the technologies that potentiate them (since those are well-covered elsewhere). I also will draw comparisons between online communities like Second Life and online games like Everquest (Taylor 2006), partially because I think that aside from some differences in the over-all world internal teleologies, where SL focuses perhaps on Simslike creation of places and objects (building), online games like Everquest focus on character skills development (bildung), the quality of sociality of both kinds of worlds are strikingly similar in other respects. But before I do that, one might well ask, what is it about this book and its topic that has caused so much to be said about it?

As the explicit intertexts of the book’s title and various chapters suggest, the book’s central conceit is to explore the use of ordinary empirical methods of ethnography and participant-observation to an extraordinary topic: to use an absolutely ordinary, conventional genre (ethnography), in book form, no less, to study an online community (Second Life) (29-30). The author makes no secret of this, and, indeed, makes intertexts with many of the central canonical ethnographies (30-31), as well as with his own earlier work in Indonesia, to establish a continuity of method from actual to virtual life.

This alone has attracted a certain amount of controversy, as well as an almost endless series of (to my mind) pedantic quibbles (almost all of which the author attempts to anticipate, but as the published reviews and informal reception I have seen show, to little avail). I think much of this controversy relates little to the specific contents of this book, whose method and manner of exposition I find to be accessible and informative, representing a virtual world and its community in a way that I, as a gaming ‘addict’ and confessed geek, find to be both sympathetic and illuminating. Rather, as the author suggests (27, 32, 240), it is illuminating about our own ideologies about ‘technology’, a shifting and ill-defined category that represents, so to speak, a kind of irruption of the future into the present, whose specific objects are always shifting. As Leo Marx argues in a seminal article (Marx 1995), technology is a hazy, ill-defined, shifting category from which we demand, perhaps, too much, in terms of transformative potential: a recently invented category, technology has become essentialized and elevated to a universal, a transhistoric mainspring of human civilization, every new technology is instantly interrogated for its utopian or dystopian meaning, all new technologies are instantly accorded a wondrous, self-actuating, potential to transform social reality for good or for ill (for example, one is minded of George Orwell’s summations of dystopian and utopian modernist models of technology in the 1930s in chapter 12 of The Road to Wigan Pier (1937) the specific technologies have changed, but the basic antinomy of the narrative of technology has not).

Because the novel category of technology defines its essence in terms of novelty (32), what is true of technology in general is even more true of the latest wave, captured under terms like ‘virtual’ and its quasi synonyms (Boellstorff 16-19, 25, chapter 2). As a result of such unexamined assumptions, the very novelty of virtual communities, themselves founded on a number of such transformative technologies, appear to demand new methodologies, even new categories of humanity (such as the ‘posthuman’, which the author usefully engages in polemics against (27-9), and ‘cyborg’, which the author usefully points out refers to a completely different kind of nexus emphasizing the continuity of human and non-human (138)). Boellstorff helpfully reminds us throughout that if we assume (following this narrative about technology as novelty) everything about virtual worlds is novel, we will never actually find out which things about them actually are novel, and which are not (25). Therefore this book, instead of assuming a novel methodology for a novel technologically mediated form of community, instead seeks to explore the ‘gap’ between actual and virtual life precisely by approaching these novel communities using the most traditional, time-honored method of anthropology, participant-observation. Interestingly, this very method is precisely what most convincingly reveals those gaps that there are between the two kinds of community, actual and virtual, and interestingly, shows that a sense of virtualness is constituted precisely by the meta-awareness of this gap between the actual and the virtual (and not, for example, by immersion in the virtual world as if it were another actual world (what Boellstorff calls the ‘immersive fallacy’ (112-3))). I will return to this below, as it seems to me that this aspect of the book (‘the gap’) is one which has a number of homologies in linguistic anthropological theory and which provides many interesting avenues for research for linguistic anthropological research.

The book, then, takes on, and to my mind, banishes, several pervasive dystopian assumptions about virtual communities (25-7) (a set of assumptions that another recent ethnography of the online community of EverQuest (Taylor 2006) also discusses, rather than engage in an extensive discussion of both books, I will note here that the books can be read together and complement each other nicely). One, of course, is the ‘moral panic’ about the very idea of a virtual life, that life spent online is dangerous, addictive, escapist, diagnostic of social pathology or individual social death. Such a view is that found in Western media representations like the infamous South Park episode on the online game World of Warcraft.[1] Another is the argument that such online worlds, inasmuch as they are wholly owned by corporations, are pervaded and formed at their very roots by capitalism. Fair enough. But then, as Boellstorff points out, so is everything else in the actual world (ibid.): if anything, this suggests that such online worlds, far from being irrelevant sideshows to the ‘real’ object of ethnographic description (whatever that is), might provide an interesting window on capitalism (the matter of ‘creationist capitalism’ that Boellstorff discusses in chapter 8).[2] Lastly, the argument that such communities, not being based on face to face interaction, are simply unreal. Many of the forms of moral panic associated with such communities seem to be related to the way that ‘unreal’ online interaction can ‘leak’ consequentially into ‘real life’. Any situation where virtual worlds have actual world consequences, whether virtual goods sold for real money, virtual friendships, flirtations or even sexual dalliances that are subsequently consummated in the real world, generate states of moral panic or analytic interest seemingly disproportionate to the stimulus. This fear of ‘avatars out of place’, the mixture of virtual and actual lives, is perhaps the most tenacious set of fears about the virtual life. But all of these caveats seem linked to a primordialism which seeks to locate social reality only in actual life, a domain of an authentic world of ‘face to face’ community or interaction constitutive of a nebulous sphere of ‘everyday life’ or a Habermasian ‘lifeworld’, from which mediated forms of community and interaction appear to be excluded as being either pathological or inconsequental (Crook 1998). In addition, these dystopian narratives of escapism, addiction, social or individual pathology are linked to a methodological codicil, that such virtual communities can only be studied in the ‘real world’, if they can be studied at all. From this perspective, the book is faulted for not doing, at minimum, ‘multi-sited ethnography’, talking to the technicians, the code-writers, at Linden Labs as well as online in Second Life. As the author points out, there is nothing wrong with that particular methodology, but it asks different questions than this one (60-65, see Taylor 2006 for an ethnography of a different online universe that explores both actual and virtual world activities). The book also complicates several utopian narratives of virtual life, showing that the virtual life is just as capable, for example, of replicating essentialism of actual world status attributes in virtual life as it is of erasing them, and that precisely here we find some of the most striking normative conflicts between members of virtual communities (for example, the discussion the in world polemics about the introduction of voice chat, particularly interesting for linguistic anthropologists (112-116)).

As with any classic ethnography, particularly ones like the one from which it takes its title, this ethnography first situates us with respect to the virtual world, set us down on an unfamiliar beach in an exotic locale. Of course, the exoticism here has to do with the fact that this is a virtual locale, and it is precisely defining this concept in an operationalizable way (in relation to the actual world), both qualifying in what ways this new form of sociality is continuous with those based on the actual world, and in what ways not, and also, in what ways virtuality has always been part of actual life, that occupies much of the first chapter. From there, the author gives a kind of orienting introduction to the ‘imponderabilia of everyday life’ of this virtual world, much as one might give such an introduction to any fieldsite. While virtuality is often linked to concepts like ‘game’ which imply a circumscribed sphere of non-consequential interaction (pp. 21-24), and consequently defined in opposition to categories of consequential sociality like ‘everyday life’ (a term which, like ‘technology’, has a kind of false sense of concreteness and obviousness that hangs about it and yet defies definition (see Crook 1998)), the author shows us that SL too, has its own banalities (239), its own consequentiality, representing in its way a kind of ‘everydayness’ (8-16). After a ‘walk through’ of the virtual universe, the author returns to definitional issues, locating the opposition between virtual/actual in relation to a series of other competing candidate characterizations and terms, and centering the discussion on precisely how the awareness of the ‘gap’ between actual and virtual life, both the separation of these categories and the constant hybrid traffic or leakage across these boundaries, is a constitutive feature of the virtual (23).

Perhaps one of the most striking features of this chapter is not, for example, the careful separation of virtuality (which is predicated on a techne-mediated ‘gap’ between actual and virtual life) from other competing concepts like ‘cyborg’ (predicated on a continuity between human and non-human) (138), or the refreshing polemic with the all-too-rapid assumption of that virtuality involves a radical discontinuity with existing actual forms of human sociality, captured in the concept of the ‘post-human’ (27-9), but rather, a brief but telling linking of virtual life to Benedict Anderson’s notion of a ‘Imagined Community’ (24, also 36). Certainly the parallelisms with other kinds of non-face-to-face, imagined but consequential relationships such as are constitutive of nations, publics, and other social imaginaries (Warner 2002, Taylor 2002, Lee and LiPuma 2002), reminds us that the virtual has long been consequentially part of the actual. The author also explores the concepts of ‘game’ and ‘play’ (drawing on Huizinga, 22-3), to show that while the concepts of game and play are relevant to online worlds (many of which explicitly or implicit define themselves as gamelike), cybersociality is not inconsequential or trivial as the game or play analogy would suggest. This should not be a novel observation to anthropologists or sociologists, after all, the ‘play’ analogy is embedded in George Simmel’s discussion of ‘sociability’ as the ‘play-form of association’ (Simmel 1949: 255), for Simmel an omnivorous re-keying of each and every imaginable purposive form of social behavior as play. No one would suggest by now that the study of sociability, the ‘sociological play-form’, whose prototypical exemplar is idle conversation between peers (Simmel 1949: 259), is an idle or inconsequential topic, particularly in linguistic anthropology.

Here Boellstorff’s discussion invites comparison with Barker (2008), who invokes Simmel’s concept of sociability in a fascinating discussion of virtual relationships mediated by rhizomatic neighborhood Telekom networks in Indonesia. Like any virtual world like SL, the ‘hard-wired’ rhizomatic Telekom network shows a situation in which ‘online’ (virtual) relationships are segregated from ‘on land’ (actual) ones. The fact that online interaction in Telekom, like with SL, is interaction pursued for its own sake, a classic example of Simmel’s sociability, does not mean that such sociability is inconsequential or uninteresting. For example, I invite a comparison between Barker’s discussion of Interkom personae and Boellstorff’s characterization of virtual selves, in particular the use of pseudonyms and even the vocal equivalent of avatars, what Barker calls ‘voice personae’, achieved by using technical effects to achieve an ‘on air’ voice distinct from one’s ‘on land’ voice (Barker 2008: 137):

Users drew a sharp distinction between interkom life and regular life…. between an ‘on-air’ (di udara) world (sometimes ‘on-line’ or di jalur) and ‘on-land’ (di darat) world. Everyone who used interkom had both an on-land name (nama darat) and an on-air name (nama udara)…. Some people had different on-air names for different lines, while others kept the same name regardless of which line they were on. People who were frequently on-air together eventually learned each other’s land names. But when invitations were sent out for social events like picnics and anniversary celebrations, they were addressed using people’s on-air names. Even in person people referred to each other by these names. One thing these on-air names did was to provide a space for people to construct a sense of self that was different from the one they had on-land. Rather than being based on one’s familial ties, the place one lives, or one’s looks, this sense of self was established largely on the basis of one’s discursive style and sound on-air. The types of adjectives people used to describe the voices they liked were gentle (lembut), sweet and melodious (merdu), exquisite (bagus), attractive (menarik), and easy on the ears (enak didengar). People were always experimenting with their voice modulation by speaking in different tones and trying different bass, treble, and reverb settings. Since they could not hear the output of their own speech as it sounded on-air, they relied on others to help them find the settings that generate the most attractive sound.

But here, too, there arises a kind of ‘moral panic’ that results from the idea that these different personas, on land and on line, will ‘leak’ into one another in consequential ways: as with SL, telekom online relationships seem to generate moral panic when they become consequential for ‘on land’ relationships, when the network is used for other than pure ‘play’ forms of sociability. The hazards and disappointments here are the same as with avatars in an online community, the idea, for example (familiar, too, from personal ads, see for example Lemon 2008 and references there) that the projected ‘on line’ persona will not match up to the embodied one, or that on line relationships will compete with or complicate ‘real’ on land ones (Barker 2008: 138-9). While both SL and Interkom relations are technically mediated in very different ways, they both produce an analogous and problematic opposition between the ‘actual’ and the ‘virtual’, as well as generating a set of very similar moral panics that result from mixtures of the two categories. Boellstorff suggest other places that one might look for the idea that technically-mediated forms of social behavior might potentiate analogs of ‘virtual’ cybersociality (which could, in turn, lead to actual sociability), for example, ‘Pen Pals’ and other private relationships mediated entirely by the postal system as a form of virtual world (36). (Indeed, the first massively multiple virtual world I participated in, the Tribes of Crane, in the 1970s, was precisely a virtual world mediated by postal exchanges, a Play by Mail (PBM) virtual world).[3] More generally, Boellstorff claims, these novel phenomena should be situated within a ‘broader history of technologically mediated intimacy going back even to love letters’ (167). Here Boellstorff gives us the opportunity to see parallels not only with technologically mediated forms of public, such as print and internet publics (Warner 2002), but also with technologically mediated forms of private intimacy, which have captured rather less attention (for example Ahearn 2001, see also Conklin 1949 for an lovely example of a bamboo literacy used only for technologically mediated intimacy, and not, as predicted by utopian or dystopian technologically deterministic theories related to writing, for example, the liberation of human reason (Goody And Watt) nor, pace Levi-Strauss, ‘to facilitate slavery’ (1973: 299)).

If the first chapter locates us in a virtual ‘place’, the second chapter locates this place in a history of the virtual, in particular a discussion both of virtual technology and concurrent, but independent development of kinds of imagining of secondary worlds (particularly drawn from science fiction and fantasy literature and early paper-and-pencil gaming adaptations of these imagined worlds such as D&D). This observation recalls a major thesis of the book, that I find to be congenial as a participant as well as an observer, which is that the primary appeal of virtual worlds is that they involve alternate social worlds (and not, as Boellstoff notes is often assumed, alternate social selves) (17-8, 31, 91). It strikes me that this theme (which is developed with respect to SL in chapter 4, Place and Time), along with the co-constitution of actual and virtual, and the mediating role that techne has in constituting this opposition, forms one of the crucial observations of the book. While I can (and will) gleefully and geekily quibble about the whiggish history of assigning Tolkien and Lewis such a pivotal and unique role in the creation of imagined ‘secondary worlds’ to the neglect of earlier fantasy writers like E. R. Eddison, Lord Dunsany, H. P. Lovecraft and Clark Ashton Smith, I am forced to grudgingly admit that Tolkien’s specific imagination of a fantasy world contained in Lord of the Rings, and also his specific (obviously Christian) theory of ‘subcreation’ contained in his short essay ‘On Fairy Stories’ (1966), played, via their reception in role playing games like D&D, the source and continued inspiration of many of these online universes (37-8, see also Balfe 2004, Taylor 2006: 22). It is true that Tolkien’s conception of a secondary world as a place entirely separated from the actual world, and being in a sense ‘realistic’ according to its own terms (Balfe 2004: 78), makes it a bit different from these various fabulous predecessors, which remain in one way or another continuous with the actual world: Eddison’s fantastic worlds were located on other planets of the solar system, while H. P. Lovecraft, Clark Ashton Smith and Robert E. Howard focused much of their work on the ‘weird’ boundary between our world and other worlds (just as they were situated on the uncanny boundary of the fantastic, occultism and science, something else Tolkien (1966) would absolutely forbid in the constitution of modern fantasy): Lovecraft in Earth’s dreamland or on the outer margins of space and time, Smith and Howard are doomed, ‘lost worlds’ (Vandemeer 2006: vii), taking inspiration from Theosophically inflected versions of Lost continents like Atlantis and Lemuria (such as Scott-Elliot 1925). As Ramaswamy argues, lost continents like Atlantis and Lemuria are fabulous place-worlds in a special sense that differs from such purely fantastic ‘secondary worlds’ like Tolkien’s Middle Earth (on purely fantastic geographies see for example Balfe 2004). Inasmuch as they are lost worlds, they are real (unlike Middle Earth), but ‘utterly unavailable outside the imagination’ (Ramaswamy 2004: 6) (like Middle Earth). Therefore, much like any entirely imagined fantastic secondary world, these fabulous lost worlds could freely be appropriated as fantastic ‘secondary worlds’ (as they were in Lovecraft, Robert E. Howard and especially Clark Ashton Smith’s writings, for example). When we study virtual online worlds, then, one context to place them in is an ongoing study of imaginative geographies that not only potentiated them, but continue to be ongoing parallel practices of popular geographical imagination, virtual or otherwise.

I say this with hesitation. I cannot help but suspect (with Ramaswamy 2004: 2-4, Balfe 2004) that a call to study such imaginary place-worlds will be met with precisely the same snide dismissal that Boellstorff notes is associated with studying virtual sociality these worlds later come to be inhabited with. This uncritical relegation of fantastic and fabulous geographies to the academic ‘fringe’ is all the more inexplicable, given that these geographies are on a continuum with imaginative geographies such as Orientalism (Said 1978, Balfe 2004), which, by now, few would quibble about the importance of. Perhaps the academic alibi for the positive reception of studying orientalist representations of space is that imaginative dimension of orientalist imaginative geographies informed actual practices of political rule of the orient (Said 1978: 2-3), thus becoming ‘real in their consequences’, as social scientists used to say. But even where there are no such power effects, imaginary worlds have much to tell us about the social world that produces them as well. As has been argued recently (Ramaswamy 2004, Balfe 2004), the impetus to imagine such worlds, as well as the specific form that the imaginations of such worlds take, are not simply products of a transcendent imagination, but are socially embedded and informed by existing cultural discourses and imaginative geographies. However, like the virtual online worlds Boellstorff studies (27), the fact that much of the content of these worlds is derived from the actual world or preexisting (e.g. Orientalist) imaginative geographies does not in itself explain the desire or impetus to imagine them (Balfe 2004: 88, see also Ramaswamy 2004: chapter 1). Nor does it explain their subsequent circulation and consumption, the way that knowledge of these worlds, like knowledge of mythology, pointless knowledge that exists under the secular sign of culture because, unlike religious or scientific knowledge, there is no pretense that it has a referent, later circulates as an important adjunct of geek credentials, of a shared geek ‘culture’ which constitutes itself by agonistic sociable exchanges of esoteric, but worthless, knowledge. As Balfe notes ‘the ways in which imaginative geographies are promulgated and consumed remain largely mysterious’ (Balfe 2004: 88). Except, of course, in virtual worlds, where, as authors like Boellstorff and Taylor show us, we can actually watch the process ethnographically.

Part of the value of Boellstorff’s book, then, by thematizing precisely the way that much of the appeal of online worlds is precisely that they are worlds, is the way he indirectly draws attention to the way that such fantastic, fabulous, and imaginative geographies underlying these online worlds are in themselves worthy of more attention than they have been given. Certainly for linguistic anthropologists, there is a clear relevance in the related topic of the imagined languages, almost always with their own special arcane alphabet (an attribute of the exotic that dates back to John Mandeville), that, like maps, seem to be crucial semiotic technologies of the imaginary world (Rosenberg 2001 and references there). There is no question that these imaginary languages recapitulate the basic outlines of 19th century ideological preoccupations with language. Not only the basic romantic impulse to allocate to each imaginary ethnos its distinctive language, and perhaps territory or culture, inherited from romantic philology and continued into disciplinary linguistics. Such tendencies are well illustrated, after all, by Tolkien himself, the professional philologist and amateur constructer of languages, giving Elvish a respectable philological Stammbaum. But Tolkien was no innovator here, the tendency is found much earlier in Bulwer-Lytton’s invention (citing Max Muller) of an obviously Sanskritic or Indo-Aryan ‘inflectional’ language for the underground Vril-yas in The Coming Race, for example (1871: 12). But also one can see hazy outlines of different kinds of languages constructed for different kinds of imaginary worlds, which recapitulates in broad terms the 19th century opposition between a more logocentric Indo-European and a more textualizing Orientalist philology. Tolkienian languages follow Indo-European philological traditions grounded in folklore, supplying full grammars of slightly exotic (but not upsettingly oriental) languages, where ‘Celtic’ alterity (with respect to Germanic) figured in elaborate systems of consonant mutations and voiceless lateral fricatives borrowed from Welsh give Elves (themselves of Germanic origin) their not-quite-white vaguely ‘Celtic’ alterity. Meanwhile, the language fragments of ‘weird’ fantasy of orientalizing writers like Lovecraft and Ashton Smith revels in obscure antediluvian, perhaps prehuman, languages, written in unsettling prehuman hieroglyphs or by mad Arabs, preserved in mysterious inscriptions and texts with names like ‘Necronomicon’ or ‘the Pnakotic manuscripts’, and handed down through complex scribal chains of transcription and translation, the very image of Orientalist textuality (cf. ‘the Sacred Texts of the East)! While there has been considerably interest in imagined languages that figure in systems of religious belief (glossolalia, for example, has a large literature), there has been relatively little serious academic interest in fantastic and fabulous imaginary languages, presumably because there is no respectable alibi for their study in the idea that ‘someone believes that they are real’, as there is with glossolalia, or more generally, because occultism, unlike respectable religions, seems to be somehow intrinsically a ‘fringe’ topic (on which tendencies see Ramaswamy 2004, Owen 2004). And yet, these fantastic languages that are part and parcel of the imagining of fantastic worlds, fantastic languages like Klingon and the various dialects of Tolkienian Elvish, are all around us, they form part of the pedagogical materials of many linguistic departments.[4] Thus, they can be studied, but only as pedagogical devices to illustrate methods of phonemic analysis or historical reconstruction. They can even be learned and spoken, by both fans and professional linguists, but somehow, the fact that some of these languages are spoken more frequently by more people than ‘natural’ languages does not yet make them legitimate topics of study, even for the linguists who speak them! But surely, given all the interest in linguistic anthropology on language ideologies that involve imagined mappings of language to territory, one wonders why it would not be of interest to explore why not only imagined communities like the nation are often attended by imaginings of standard languages, but the same is true of imaginary (fantastic or fabulous) worlds as well in the same time periods? Constructed worlds (‘conworlds’) seem to demand constructed languages (‘conlangs’).[5] The practice extends to virtually every kind of imaginary world: recently I was trolling through a web site of a guild on the game Saga of Ryzom whose members are typically members of the ‘Tryker’ race (they look a bit like anime-influenced Kewpie dolls), and discovered that someone had decided to give the ‘Trykers’ their own language, replete with typical phrases used in the game! (cites). Given that many proto-nationalist language revival movements involve wholesale re-invention of languages to be spoken by imagined aspirational national communities (e.g. the revived Celtic language Cornish for a revived Cornish Nation (see for example Deacon 2006)), why does the no more nor less eccentric activity of inventing imaginary languages for completely imaginary worlds (e.g. the dialects of Elvish which seem to be, structurally, if not lexically, every bit as good members of the Celtic family of languages as re-invented Cornish) seem beyond the pale of serious academic study? When asked ‘Cornish or Klingon?’ (Deacon 2006), why is the answer so obviously ‘Cornish’? Is Cornish really any less ‘invented’ or more ‘real’ than Klingon?

But I digress. The central point is that the idea of the ‘secondary world’ (37-8) develops independently of the technology that will later allow these imaginary worlds to come to virtual life (which Boellstorff also discusses at length in chapter 2). That is not to say that such imaginary worlds first come to life when they are harnessed to a virtual online engine. Online worlds develop as much from imaginary worlds animated by paper and pencil role playing game (RPG) technologies, with their varied assortments of colored polyhedral randomizers and metallic figures, as they do on ‘cyber’ technologies. This general process identified by Boellstorff here seems familiar from my own personal biography. Certainly, I am my friends shared a childhood habit of creating often quite elaborate and populous imaginary ‘secondary worlds’ on the model of Tolkien’s Middle Earth or Lovecraft’s Dreamlands, and our creation of these worlds predated our ability to animate them, to virtually inhabit them, to share them with others, using such paper and pencil role playing technologies. When we discovered we could live in these worlds we had imagined using paper and pencil RPGs, we did so, becoming Game Masters (GMs) of worlds we had created, and the existence of RPGs as frameworks in turn acted as an impetus to further imaginings, our imaginary worlds changed to adapt themselves to the games that animated them. The independent confirmation of the validity of Boellstorff’s discussion here is surely in how many literary ‘secondary worlds’ have their own paper and pencil RPG worlds associated with them (Balfe 2004, Taylor 2006: 21), and many of these are now worlds associated with their own MMOs as well (Taylor 2006: chapter 2). The dependence is not merely a matter of simulating an imaginary universe that had been the theme of an RPG in an entirely new virtual game system: many online versions of these worlds (not worlds like SL, but certainly many of the ‘game’ worlds like the ones discussed by Taylor (2006)) often have as their core ‘game system’ something which inherits the very same basic categories and rules as the paper and pencil versions, for example, physical damage is still doled out in hit points as it was in the very first RPG (just as the language used by players in MMOs, who sometimes peak of ‘rolling’ a new character, recalls the dice-based technologies of paper and pencil RPGs). In effect, they are cyber simulations of paper and pencil RPG worlds.

The third chapter, on method, is a central one for developing the idea that such virtual ‘secondary worlds’ can be studied using many of the same ethnographic techniques as we use for actual worlds, and that precisely the ‘gap’ as well as continuity between the virtual and the actual emerges most clearly when we use the same methods for both. The author strikingly develops as a methodological thesis that virtual worlds cannot be reduced to actual worlds, but have a kind of relative autonomy (as used to be said of ethnographies of work places): ‘Actual-world sociality cannot explain virtual-world sociality. The sociality of virtual worlds develops on its own terms, it references the actual world but is not simply derivative of it’ (63). The remainder of the chapter shows how one would implement such an epistemic thesis in methodological terms, incidentally producing an excellent separable treatment of ethnographic methodology suitable for use in a methods course (an experiment I have tried successfully).

If Part 1 adduces the relevant presuppositions for the study of the virtual life, situating the virtual ethnography in virtual place, virtual history, and virtual ethnographic method in relationship to the actual, Part 2 represents the centerpiece of the ethnography, investigating the ‘culture’ of SL, specifically organized around categories of place and time (chapter 4), personhood (chapter 5), individual social relationships (intimacy, chapter 6) and finally community (chapter 7). Part 3 in turn takes up the topic that has interested much of the existing literature, the virtual political economies in which it is possible to make actual ‘real world’ money (chapter 8), and finally a conclusion about the nature of the virtual (chapter 9). While there is much of interest in these other chapters, I assume the discussions of cyber-sex and cyber-economies (which the author manages to do without the usual sensationalism) will attract their own animated commentaries, but I will confine the remainder of my remarks to just parts of chapters 4 and 5 that seem to raise important questions for a linguistic or semiotic anthropology focusing on performance or interaction, the presentation of self in virtual life, so to speak.

This order of presentation, which privileges the categories of the virtual world (chapter 4) over the categories of the virtual self (chapter 5), and relations between virtual selves (chapters 6-8), reflects the author’s thesis that a big part of the distinctive appeal of virtual worlds like SL is precisely that they are virtual alternate worlds, and not merely, as much of the literature has assumed, that they allow virtual expression of alternate forms of self, for example (though this is obviously also important). It is possible to imagine online environments that allow the latter without really creating the former, for example alternate online identities like email and especially chat rooms and discussion boards, which share with online games both the use of pseudonyms and avatars, so this is an important point. At the same time, his argument is that the appeal of these worlds is not that they are alternate actual worlds (in which one loses awareness of the real world, what Boellstorff calls the ‘immersive fallacy’ (138)), but their appeal is partially in the metawareness of the ‘gap’, mediated by processes that he groups under the name of ‘techne’ (54-59), between virtual and actual worlds.

Chapter 4 establishes the core categories of place and time of the virtual world, showing how the ‘virtuality’ of the online world inheres in their properties as places (the visual field (visuality and land), the properties of objects within the visual field (builds and objects)), while the actual world intrudes into the virtual place especially in relation to the category of time (102): ‘virtual worlds create a gap between the actual and the virtual not juts in terms of place, but in terms of time as well…. The gap with regard to space constitutes the binarism between virtual and actual; the gap with regard to time threatens it’ (105). Two particular, and entirely banal, kinds of disruption of the sense of immersion and presence fundamental to the virtual world are discussed in some detail. Of these is lag (101-106), in which time delays relating to loss of synchronization between computer and server disrupt the sense of place (103), but also while the universe is a shared place, this shared place is distributed across many time zones, interfering with the ability to share this space synchronously (104-5). The other is his discussion of the category of AFK ‘away from keyboard’, a state unique to virtual worlds that ‘presuppose synchronic virtual sociability’ (107). AFK is unlike for example, a busy signal (107), or even when one ceases to be a member of a public by virtue of a lapse of the attention that makes one a member (Warner 2002: 60-2), presumably because, no one else notices you have stopped paying attention in such a case. AFK is a defining term for introducing the next two categories of the chapter, senses of immersion and presence (112-117), senses which confer a Schutzian ‘accent of reality’ on a virtual world. An AFK state can be defined as ‘presence without immersion’ (112). Because one is present in the form of an avatar, one’s absence from the keyboard becomes a notable absence, an avatar that does not respond to tells, or continues to walk in a straight line butting its head against a wall, are evident signs of presence (online) but absence (of immersion). Lights on, nobody home. This particular state is so diagnostic of the intrinsic defining place-time properties of the virtual world and interactional hazards there that Boellstorff proposes an ‘AFK test’ for virtualness: ‘If you can go ‘afk’ from something, that something is a virtual world’ (112).

If AFKness is presence without immersion (the avatar is there, but reduced to its status as a puppet without the puppeteer), straddling ‘the border between online and offline…a kind of ghostly absent presence’ (117), then the actual world can undermine the sense of immersion by too much presence of the actual world, as well. The debate about bringing voice into SL (that is, using third party technologies to allow voice chat alongside text chat) is, perhaps, for linguistic anthropologists, of great interest (112-116). The category of voice) is ideologically construed as a proxy for selfhood (on voice see, for example, Taylor this issue). As Mrazek puts it in discussing the ‘montage’ of voice with image and motion in Indonesian puppetry, human voice is ‘special’ within the montage of elements that are juxtaposed to construct a singular performance in puppet theatre too, because unlike the puppet and its motions, the voice is associated with an identity (that of the flesh and blood puppeteer) separate from that of the puppet (hence voice acting abilities are among the most important for a puppeteer): “Human voice is associated with its source, with humans, and thus in complex ways implicated in what humans are, in human presence, in human identity” (Mrazek 2005: 270). As with puppet theatre, the debate in SL about voice is deeply revealing about our own ideologies about voice as imparting distinctive, individual characteristics, thus potentially allowing greater intimacy by revealing an embodied aspect of one’s unique, actual self, but also being essentially revelatory of status features (gender, race, ethnicity, age, education, language ability, disabilities), thus negating the sense of immersion and undermining the ability of participants to represent their virtual selves in a manner unencumbered by their actual life embodied status features (112-116). The debate in SL about voice is thus worth comparing with 18-19th century debates on anonymity and signature in the press (see below), since text chat, even if it potentially indexes inadvertently or intentionally a number of features about the writer (education, knowledge of English, fluency in use of different online registers, including emoticons [ e.g. :) or J] and face characters [e.g. (^_^)], which are often taken as a proxy for gender, among other things), is as nothing compared to the flood of status features indexed in one’s embodied voice (cf. Taylor, this volume). The result can be, in my experience, an almost jarring disjuncture between the persona embodied by the avatar and that embodied in the voice, leading to a complete collapse of immersion as the actual world discredits the virtual one. The same contradiction in voice in puppet theatre was noted early on as ‘Zich’s contradiction’: on the one hand, the synthesis or organic blending of the independent figure and movements of the anthropomorphic puppet and the human voice produces a naturalizing animation or vivification of the puppet. On the other hand, the often radical mismatch between the unmodified (either technically or theatrically) voice of the adult human puppeteer and the tiny, childlike body, and clumsy mechanical movements, of the puppet could lead to a jarring contradiction, which increases as the human voice and the puppet body become ever more distant (Bogatyrev 1983: 60-1). As Bogatyrev notes, in comments about the aesthetic problems presented by modernist puppets that could easily apply to avatars in online worlds,

With contemporary puppets whose shape differs striking from humans, the contradiction between the puppet’s shape and movement and the puppeteer’s ordinary voice naturally increases. It is difficult to imagine how a puppet made of cutlery (knives, forms, spoons) or figures made from laboratory equipment (the Martian musical comedians) could talk and sing in ordinary human voices. Consequently, contemporary puppeteers look for various means to distort the human voice to eliminate (or at least to lessen) the contradiction between the puppeteer’s voice and the distorted puppet. (Bogatyrev 1983: 61).

As Bogatyrev notes, just as interesting for linguistic anthropologists as masking the original (the puppeteer’s voice) is the question of what new modified voice the puppet will be given, and how. As Miyako Inoue shows in a brief but fascinating example, the question of what kind of ventriloquized voices will be given to a doll is instructive, in her case, a Barbie-like doll called ‘Licca-Chan’ who has everything that her owner does not ‘a sofa, piano, stylish clothes, a sports car…’ and lastly a special register of speech associated with ‘women on TV’ (Inoue 2003: 6). What is striking about online worlds is that voice chat in such worlds shows none of the technologies or techniques for voice modification found in Telekom or Puppet theatre, perhaps because the voice chat technology is provided by third party programs designed for straightforward chat (Vent, Skype). On the other hand, the sense of greater intimacy conveyed by revealing ones ‘true self’ in ones voice leads to a greater sense of presence, and in combat games, like the space war game EVE online, of course, the preference for voice chat is purely instrumental, to control the deployment of large fleets of ships in battle formation. More generally, theatrical mediations like puppet theatre share with technical ones like Telekom and Virtual Worlds a capacity to destabilize essentialisms and separate out components of interaction found together in face-to-face interactions, a lesson familiar from Goffman (1959, 1974, but also for example Mrazek 2005), as Barker (2008) puts it: “by definition all communications technologies provoke meta-pragmatic awareness since their definition as ‘technologies’ implies a reflexivity about how communication takes place. The advent of a new technology reminds people of other forms of mediation which seem relatively ‘old’ and often more essential.”

Categories like AFK and voice are as much categories of the online self as properties of the online world, which brings us to the chapter that raises the most interesting questions, SL considered as a performance of self, the topic of chapter 5. The debate held in SL about voice is not only revelatory of our own complex and fraught actual world semiotic ideologies about the category of voice, so often a proxy for notions of authentic self (Keane 2003, Mrazek 2005: chapter 5, Taylor, this issue), but it also of the risky and contingent nature of virtual embodiment, centering in the avatar. The debate about voice in SL can be seen as a rehearsal of the basic antinomies of the avatar embodiment, between cultural logics stressing ‘augmentation’ (the view of an avatar as being in a sense a ‘cyborg’ prosthetic extension of the actual self on one of possibly many virtual platforms) and ‘immersion’ (stressing instead the ‘gap’ between, and relative autonomy of, the actual and virtual worlds) (115-6, 121). The avatar, like one’s voice and one’s pseudonym, is seen as revealing or concealing one’s authentic self, revealing it because the mode of embodiment is truly an example of conscious self-fashioning, concealing it because, like the pseudonym one chooses as a condition of participating in the online world, it also erases or replaces one’s real world embodiment and embodied status features of race, class, gender, even as it creates new asymmetries specific to the virtual world (as the avatars of new players or newbies are frequently identifiable by how they look or move).

Names. First of all, avatars in online worlds, like on line voice personae in Indonesian Telekom (see Barker, above), are associated with persistent names, so that an online persona consists of both an avatar and a name, both of which produce potential for creative self-figuration and self-fashioning. In part, then, the resources for online self-figuration in worlds like SL overlap contemporaneously with those found in Telekom ‘voice chat’, internet chat rooms and discussion threads (which also feature visual figures that are also called ‘avatars’). They also remind us of the fashion for pseudonymy characteristic of 18-19th century print culture in many places, which was like the contemporary internet, and unlike what was dubbed by Matthew Arnold the ‘New Journalism’ beginning in the late 19th century, in privileging anonymous collective authorship (‘the editorial we’) or pseudonymous individual authorship over individual signature (for an overview of debates over signature and anonymity in the British press see Brake 1994: 19-32 and chapter 4)). Therefore, this aspect of the avatar could be compared to Warner’s discussion of ‘the principle of negativity’ (Warner 1990: 42), the process by which one enters the public, by erasing one’s private self (and rendering ethos, the persona of the speaker/writer, irrelevant) and assuming a public mantle of anonymity, to evoke a purely public figuration by which ‘arguments, not persons, are judged’ (Remer 2000: 83). In this sense, the online environment of a world like SL can be compared (not at all facetiously) to the virtual public created by 18-19th century ‘anonymous journalism’, both an MMO and the press can be ‘participated in by any number of unknown and in principle unknowable others’ (Warner 1990: 40). But as Remer points out in his discussion of Warner’s ‘principle of negativity’ in the anonymous press, echoing Boellstorff’s point that pseudonymity, by definition, cannot be reduced to anonymity (122), while an assumed pseudonym produces anonymity by erasing one’s private identity, it also creates a public persona that is not an anonymous ‘voice from nowhere’ or voicing out of a collectivity (‘editorial we’) (on which see Brake 1994), but, indeed, a specific individuating figure that in many ways resembles an avatar (Remer 2000: 83-4). Like avatars, each pseudonym is part of what Boellstorff would call a techne that helps constitute the gap between private and public identity, and as a mediation points in two directions, it negates ones actual identity, and posits a new virtual persona at the same time. These ‘gaps’ generate similar antinomies. Boellstorff finds an antinomic opposition between cultural logics of ‘augmentation’ and ‘immersion’ regulating traffic between actual and virtual worlds (115-6, 121). So too, Remer argues that Warner’s ‘principle of negativity’ (analogous to ‘immersion’) privileges the Antifederalist ideology of the press, by treating pseudonymy as being essentially a form of anonymity, attempting to eliminate ethos, the rhetorical ‘presence’ of the embodied person in their public arguments. By contrast Remer reveals that the Federalist ideology, which seeks to use pseudonyms to provide disembodied texts with characterological forms of ethos (e.g. classical pseudonyms redolent of Republican virtue like Agrippa, Atticus, Cato…), or, alternatively, to find out the true embodied identities of the authors behind the disembodied texts (analogous to the cultural logic of augmentation), were just as prevalent a cultural logic constitutive of the formation of the 18th century press (Remer 2000: 78-84), and certainly were an important part of the 19-20th century reaction against anonymous journalism. Nor do the prima facie resemblances between the naming sets of MMOs like SL and the 18-19th century anonymous press end there. In some online worlds (like City of Heroes) many participants will have a ‘main’ or favorite avatar as well as many alternate avatars (‘alts’) (leading to an addictive disorder called ‘altaholism’, a refusal to stick with a main avatar and endlessly producing new alts). Since these alts are all on one account united by a ‘global chat’ name which is generally given out to establish a friendship across incarnations, this global chat name will provide sometimes provide a kind of stable ‘true’ name or ‘intimate’ cross-incarnational name that friends and acquaintances will use in address even when one is not playing that alt. Alternately, since one’s ‘main’ is often one’s most advanced character and therefore the oldest, other players may index the depth and age of the cross-alt friendship by using the name of the avatar in whose guise that person was first met. But here again, the new virtual world is much like the old, the distinction between a ‘main’ and an ‘alt’ also finds its homology in the pseudonymous adoption of print personas in print culture. For example, in 19th century Georgian print culture, like many others, the practice of using pseudonyms (often playful kennings of one’s actual name) not only allowed male, often aristocratic Georgian intelligentsia authors to slum or ventriloquize, to assume voices of relevant social others (women, peasants, Europeans or ‘Asiatics’ (Mikadze 1984)), and vice versa, during periods of repression to avoid state supervision or retribution, but such authors might well create obviously humorous pseudonymous ‘alts’ especially for humorous or satirical writings, displaying a pseudonymous ‘altaholism’, repertoires of jocular pseudonyms that could run into the dozens, alongside a pseudonym for serious writings that provided a stable ‘main’ persona (Grishashvili 1969, cited in Mikadze 1984: 6).

But the greatest set of resources for self-figuration, self-presentation, self fashioning is the avatar itself. The avatar inherits all of the dualisms that afflict all performing objects, ‘Zich’s contradiction’ (Bogatyrev 1983: 60), namely that avatars, like puppets and other performing objects, ‘may be perceived either as living people or as lifeless dolls’ (cited in Bogatyrev 1983: 48). Earlier I adduced the two examples that Boellstorff discusses that in one way or another illustrate Zich’s contradiction, AFK and voice. An avatar that is AFK is an avatar reduced to a lifeless doll. An avatar that is attended by a discordant human voice, too, cannot be a living person, and must become, by contrast with its all too present puppeteer, a lifeless puppet. But equally interesting is the question of when avatars do not become Zichian ‘lifeless dolls’, but ‘living people’. One of the most interesting, and suggestive, aspects of Boellstorff’s book, is a brief discussion of an instance of a player attributing agency and an autonomous persona to an avatar, linking it productively not only with issues of performance theory (including metaphoric comparisons with performing objects like masks) but also with treatment of nonhumans as agents in anthropology of technology (149). Avatars exist on a continuum with more tangible performance objects and nonhuman actors, a heterogeneous collection of masks, costumes, dolls, puppets, animations, automatons, machines, and robots. What avatars share with other performing objects, as a kind of techne, is their capacity to produce a ‘gap’, to divide the individual into ‘dividuals’ or to divide or distribute a unitary human agent across multiple roles of a single performance (for example, a flesh and blood actor becomes a puppeteer and a puppet, or sometimes multiple puppets, just as a single person might have several alts, etc.) (150). This ‘analytic’ aspect of performing objects as a techne or mediating technology, the process of ‘breaking down and building up’ (Veltrusky 1983: 98-9), which Mrazek usefully compares to montage (2005: 17-29), which decomposes the organic unity of the human actor into separable parts (voice, puppet, gesture, movement) and then recomposes them as a new aesthetic unity (a ‘montage’), also provides an analytical vocabulary for the human. In this sense performing objects (virtual or actual) work like Goffman’s dramaturgical metaphors for unstaged social life, especially his influential division of the unitary role of speaker in to principal (the one responsible for the text or performance or in whose interests it is performed), the author (the one who composes the actual form of the text or performance), the animator (the one who actually animates the text or performance), and lastly, the figure (1974: 496-559).

As Boellstorff notes, apropos here is Latour’s discussion of the various kinds of non-human actants which substitute for, or are delegates of, human agency: Avatars not only remind us of performing objects like puppets, but also remind of all those objects that serve as proxies for human agencies, actants who are placed on a continuum with the humans who might have been required to perform the same task: ‘characters, delegates, representatives, lieutenants…– some figurative [in the sense of having iconic anthropomorphic form], others non-figurative; some human, others nonhuman; some competent, others incompetent’ (Latour 1992: 162). The comparison reveals a larger class of ‘performing object’ or ‘nonhuman actor’, the historionic objects that fascinated Praguean semioticians, who ‘perform’ or ‘act’ in the dramaturgic sense, and the non-human actants that fascinate Latour, who ‘perform’ or ‘act’ in the sense of performing delegated human tasks or labor. Just as the puppet breaks down the unity of the human into fragmentary qualisigns of the human (voice, image, gesture, motion) and rebuilds it as a montage, Latour here defines actants, human or non-human, initially in terms of being substituted for, or having servile human activities of ‘labor’ delegated to them, but they are often also semiotically anthropomorphized (gaining iconic figural human properties, which Latour here subsumes under the feature of being ‘figural’). Latour’s example of such a ‘performing object’ which is a ‘delegate’ and ‘figural’ is Petit Bertrand, a small named iconic representation of a human of the kind that often attended and sometimes still attends mechanical contrivances, which unite a mechanical actant, a delegated human agency (a servile object, in this case mechanical meat roaster) with an anthropomorphic figure (a performing object, in this case, a small named representation of a man who is made to appear to be the actor whose labor turns the device) (Latour 1992: 163): ‘The irony of the ‘Petit Bertrand’ is that, although the delegation to mechanisms aims at rendering any human turnspit useless, the mechanism is ornamented with a constantly exploited character ‘working’ all day long’ (Latour 1992: 163). In English, too, there is actually a field of polysemy that unites the two kinds of ‘performing objects’, historionic performers and servile ones. As Shershow’s discussion reveals, the term jack, moves from denoting a typical servile human laborer in general, ‘Jack’ (the name of the folkloric Everyman), to denote a whole variety of (figurative and nonfigurative, histrionic and servile) devices,

which in some way take the place of a lad or man, or save human labors…. The literal clock jack, for example, is an icon of a ‘common’ man, a kind of ‘servant’, whose apparent labor, actually performed by hidden clockwork motion, is a histrionic illusion or trick. The turnspit or ‘jack’ had fastened to its handle a tiny figure of a man who appeared to turn it. On the one hand, the word’s semantic sense of servitude has outlasted its iconicity: even today, for example, one refers to an electric outlet or a device for raising an automobile off its axle as a ‘jack’. On the other hand, the etymology also seems to foreground the very process of iconic figuration…. The ‘jack-in-the-box’, once denoting a specific cultural practice, eventually became a children’s toy, a puppetlike figure in a literal box. ((Shershow 1995: 80-1)

Avatars, then, in a sense, are the newest, virtual representatives, of a wide community of performing objects in this broader sense, including both the histrionic objects (dolls, puppets) which fascinated Prague schools semioticians, as well as ‘servile’ objects (objects to whom human labor has been delegated), that have been of more interest in anthropology of technology. As Boellstorff’s suggestive comparisons reveal, it would appear that avatars give an interesting empirical domain in which semiotic anthropology, performance studies and anthropology of technology overlap, and also an interesting site to explore precisely what is new about virtual worlds and virtual humanity, and precisely how it is new. These various other existing forms of actual performing object each provide a ready analytic metalanguage into which we can ‘break down and build up’ the avatar, by introducing it pairwise to other members of its extended family.

Dolls. Avatars are first encountered on a screen in which their distinctive physical appearance is created by the maker (128-30). At this point, avatars can be looked at as static images or decorative objects, in effect, a kind of doll or costume (Prochan 1983: 7-8). Accordingly, part of the appeal of some worlds (SL, City of Heroes [COH]) is the way that avatars appearance can be customized and is independent of avatar abilities. One player of the superhero world COH confided in me ‘The reason I like COH is because it’s like a Barbie League of Justice’, that is, a cross between a dressable doll (Barbie) and a superhero comic book figure (League of Justice).

Toons. But like some doll that has programmed movements, the avatar can involve not only static iconography, but also preprogrammed ‘scripts’ or ‘animations’ as well for certain kinds of ‘gestural’ actions, including both bodily movements like ‘emotes’ and ‘powers’, a kind of dynamic iconography (Prochan 1983: 7-8, Tillis 2001: 175-6). The superficial similarity between this aspect of avatar motion, half-way between motionless doll and real-time animation of the puppet, and cel animation, is captured by use of terms like ‘scripts’ () or ‘animations’ (Tillis 2001: 176), as well as by the in-world term in COH for avatars as ‘toons’. Like scripts in SL, animations are often considered to be part of the over-all ‘look’ of the ‘toon’, a dynamic extension of the ‘doll’ aspect of the performance object. Thus, in COH superpowers themselves are sometimes chosen not because of their efficacy, but much in the same way animated emotes are, on the basis of the aesthetic properties of the animation, becoming part of an an over-all costume (including a biographical ‘origin story’ explaining the costume and powers). One example: some COH players will pick an anotherwise nearly useless ‘whirlwind’ power as part of a macro or key-bind that is linked to a costume change, so that they can change into their ‘superhero’ costume with appropriate flare. The similarity of terminology (‘animation’, ‘toon’), whether or not it captures actual similarities in the way these images are produced, still places avatars on a continuum as images with animated cartoon figures (see Silvio 2008 for a interesting parallel continuum between religious images through puppets to cartoon characters in Taiwan). I note here in passing that the term ‘motion’ (etc.)

Costume. The outward appearance of an avatar is often called a ‘costume’, reminding us that it has the property of being alienably associated with bodies, just as avatars can have multiple costumes, the costumes of avatars can be recreated in the real world. Worlds like CoH and SL that allow considerable customization of avatar appearance potentiate practices like virtual world and actual world costume play (cosplay), placing avatars on a continuum with other cosplay practices, in which fans dress up as favorite anime or cartoon characters (Winge 2006, Kotani 2007, Silvio 2008, Gagne 2008). For example, player driven costume contests are a major in world event in CoH which is complemented by actual world costume contests in which players dress up as their favorite online character, and at the same time avatars are often designed to cosplay favorite comic book or anime characters in world, or even dolls (a Barbie based ‘supergroup’ for example). Avatars, then, can be placed in a ‘cosplay continuum’ with actual personal adornment, one that often involves intertexts between virtual and actual world performances.

Puppets. When one enters the world itself, the appearance, the repertoire of animations, and the actual executed movements (both gestural and proxemic (Tillis 2001: 175-6) of the avatar resemble those of tangible performing objects like puppets. In particular, the window that opens up onto the world not only shows the world and the avatar, but also, a set of ‘handles’ (Mrazek) or ‘strings’ (Tillis 2001: 176) in the form of the on screen interface which continuously reminds one that the avatar without a controller is but a lifeless, AFK, doll. The clumsy motions of other avatars of others who are, perhaps, unused to the interface, reminds one that the avatar is not entirely contained within itself, but has strings attaching it to an unskilled puppeteer. Here not only their formal potentialities as performing objects (for example, avatars behave like real time or stop motion puppets in having specific points of potential articulation, or avars, which are part of their build), but also the manner in which they are controlled in real time through some kind of interface (both of which are unlike cel animation figures which they may otherwise resemble as animated images), invites comparison with performing objects controlled in real time like puppets. The fact that some of the motions of avatars are scripted with specific animations (for example, emotes as well as powers or skills), or combinations of existing animations by means of macros or key binds, place makes their motion somewhere between real time puppetry and animatronic or even stop motion puppetry (citations).

Bots. Automatons. If a badly controlled avatar reminds that avatars are puppets, an avatar that does not respond as a person would, that does not pass a ‘turing test’, is an automaton of some kind. Sometimes called ‘bots’ (), this list includes not only the limited AIs that control ‘non-player characters’ (NPCs, another category inherited from paper and pencil roleplaying) as well as player designed automatons that are run entirely by macros (a major problem in a game like Eve are ‘gold-farmers’ who strip mine asteroid belts using ship pilots run more or less entirely by macros).

Builds. Avatars are often evaluated in terms of their intrinsic capacities to interact with the world, either as nearly cyborg-like extensions of the player, or stripped down to the raw numbers of the acquired skills, levels, powers, and other attributes (add discussion of Taylor). This dimension of an avatar, and sometimes the avatar itself, in this sense is often called a ‘build’, almost always designated with some specific purpose (e.g. a ‘PVP build’ for player versus player playing, a ‘PVE build’ for playing against non player characters). In some sense, perhaps, they can be treated as proxies or prosthetic extensions of the actual self, but primarily here the avatar is being considered as a bundle of aquired in world abilities, treated instrumentally in a means-end maximizing calculus of efficiency, which allow the player to interact with the world and others in the world in new ways, characteristic of a specific kind of ‘instrumental play’ that Taylor calls ‘the power gamer’ (Taylor 2006: chapter 3). Here the avatar presents itself as divided between two selves, the decorative, qualitative ‘doll’ and the numerical, quantifiable ‘instrument’, the avatar seen in terms of the numerical values of its acquired abilities. In some worlds the decorative aspect of the avatar (doll) has an instrumental referent, for example rep resenting ‘armor’, which can only be individuated by its color, or abilities in other ways affect the costume, with the net result that the best, most efficient, ‘builds’ tend to all look exactly alike except for their color. This division is less characteristic of ‘building’ worlds like SL and more typical of ‘bildung’ games that emphasize development of the avatar through experience. Taylor distinguishes between broadly ‘social’ and ‘instrumental’ modes of play and players, and while these are not determined by the gender of the players, especially inasmuch as most players will oscillate between the two, they do seem to have a gender association, here we see the divide between two gendered kinds of play, dolls and toy cars (reference) writ large, but also between the general way in which technical play is masculine (finish this, eckert too).

Finally, as persons, avatars seem to be on a continuum between humans and non-humans, subjects and objects, actors and non-actors, selves and not-selves. First of all, as performing objects, avatars are located squarely in the antinomic world of Zich’s contradiction between human and thing, as dramaturgic proxies for the self, they re-enact all the central oppositions in the performance of self (Goffman 1959) that are carried over and developed in the typology of ‘figures’ in Goffman’s Frame Analysis (1974: 523-534; Hastings and Manning 2004: 302-304 for a discussion). The avatar embodiment by itself, as a kind of performing object like a costume or mask (to which they are sometimes compared explicitly, e.g. 130, 133), or a puppet (see above), automatically produces a sense of a gap between the actual and the virtual self, an intrinsically dramaturgic opposition, and players approach this gap dramaturgically in a number of ways, which one can capture in the Goffmanian category of ‘figure’ (here I am also influenced by Gagne’s (2008) insightful discussion of a ‘cosplay continuum’ of kinds of performance). On the one hand (taking my inspiration from Bauman’s discussion of intertextal gaps in meditational performances above), they may seek to minimize the gap, trying as much as possible to overcome the gap between the principal and the animator, the person and the avatar, turning the avatar, as much as is possible, into Goffman’s ‘natural figures’ ‘‘live, physical, flesh and blood bodies – animal or human– each with an ongoing personal identity’’ (Goffman, 1974, p. 524), a kind of figure typified by a situation when a performance is felt to be a performance of one’s authentic self, as no performance at all. This can be done, in an online environment, either by subordinating the avatar iconically to one’s physical embodiment (making the avatar mirror one’s actual appearance), or by claiming that the avatar mirrors one’s true, inner self, for example. Alternately, given how many of these online worlds originate in paper and pencil roleplaying games, one might expect a great deal of role-playing, of maximizing the gap in performance, to produce a ‘staged figure’, one in which there is no sense that the actual and the virtual meet. The opposition between ‘natural figure’ and ‘staged figure’ finds a homology in the idea that some avatars are ‘main’ avatars, with ‘ongoing personal identities’ with which one identifies as versions or ‘reflections’ of one’s authentic self, as opposed to secondary ‘alts’ who are consciously given personas distinct from that self. As Boellstorff notes, the use of explicitly dramaturgical metalanguages comparing avatars to performing objects (‘costumes’,’masks’) seems to occur especially when the character is felt to be an ‘alt’ (staged figure) rather than a ‘main’ (natural figure): “This frequent imbrication between the primary avatar and what was understood to be actual-world selfhood explains why it was that when residents referred to alts as ‘costumes’ or ‘masks’ or emphasized that they created alts that did not sound or look ‘like themselves’, the implication was that the avatar in question was ‘alternative’ to their primary avatar as much as their actual-world self” (133). Beyond this, we find some alts are more like Goffman’s ‘cited figures’ (Goffman 1974: 529-534), as when players of COH tread create a ‘tribute’ toon with a name or costume resembling a ‘real world’ superhero or celebrity. Here we might speak instead of a ‘cosplay’ figure, enacting the identification of a fan with their object of adulation. In some cases, like ‘joking’ alts, we find ourselves dealing with Goffman’s ‘mockeries’, that is, figures that are the diametric opposite of the figured self of the ‘natural figure’, these are a joking ‘not-self, in which there is no pretense of identification: “At the center is the process of projecting an image of someone not oneself while preventing viewers from forgetting even for a moment that an alien animator is at work” (Goffman, 1974, p. 534). In super-hero games like COH, where many or most of the avatars are wearing actual superhero costumes or masks in any case, the term same opposition is used (often the term ‘toon’ used synonymously with ‘alt’), one player friend, reminiscing with me about the avatars we were using as mains when we first met, informed me that the avatar name I met them under and still call them by was really a ‘joke toon’ and it was not under the second or third alt that they discovered a toon they identified enough with that they actually wanted to ‘take it to 50’ (that is, play it enough to reach the maximum experience level).

But even in such a comprehensive typology of figures, it seems that much of what happens in worlds like SL of COH falls between these categories. As Boellstorff notes for SL, and in my experience elsewhere, self conscious role-playing is rare compared to, for example, the paper and pencil roleplaying games that preceded these worlds. Even in superhero games like City Of Heroe, which would seem to demand over the top campy role-playing of avatars which are, after all, representations of comic book heroes, role-playing is done only sporadically at key moments (often reduced to shouting out heroic one-liners that have been turned into key binds or macros and tied to some figure animation, like “Eat Electric Death, Villain!” plus the execution of some suitably electrical attack—even these quickly ‘get old’ and cause complaints from teammates) and obsessive role-players form an actual subculture. Perhaps this is because, unlike a paper and pencil role playing game which is, after all, played face to face, in an online world there is no need to ‘bracket’ one’s actual self within a performance frame, since taking on an alternate not-self embodiment (the avatar) is a condition of participation. What seems most characteristic, then, of online environments is neither the essentializing creation of natural figures nor the dramaturgical creation of staged, cited or mocking figures, absolute identity or absolute alterity, but a kind of labile playfulness, a sociable playing with personas that seem somewhere in between, that Boellstorff calls (borrowing from the participants) ‘persona-play’, with varying degrees of interpenetration, autonomy and mutual constitution between the actual and virtual selves (119-122), just as ideologies of immersion (absolute autonomy of virtual and actual) and augmentation (prosthetic continuity between virtual and actual) represent extremes within which most SL residents locate themselves.

As agents, avatars represent a kind of continuum not only of performance of self, but also of agency, between the human actor and the non-human actant. Avatars have attributes of both, they are both controlled in real time by player agencies, but some of their motions are scripted like robots. As Boellstorff reminds us, avatars and other performing objects are not quite the same mediation of humanity as the cyborg. The cyborg is a creature of ‘augmentation’ (see above), a ‘prosthetic continuity between human and machine’, avatars are predicated instead on ‘immersion’ (see above), on a discontinuity between actual and virtual (138). Both produce a continuum between the human and the non-human: the cyborg is a metonymic continuity, a hybrid; the avatar is part of a metaphoric continuity, avatars are substitutes, delegates, representatives for the human they replace (cf. Latour 1992 pp.). At the same time, as Boellstorff reminds us, avatars can act on their actual world counterparts, assuming their own agency (similarly to the ways that fans attribute agency to cartoon characters, for example, see Silvio 2008). Sometimes, Boellstorff suggests, the roles can be reversed, not only do players come to take on the personas of their avatars (the ‘proteus effect’), but some ‘alts’ are given ‘real life’ personas in such a way that the alt could be said, in effect, to have a ‘real world’ avatar (133). As Boellstorff, citing Latour, suggests we should not assume that the agency is all with actual world human flesh and blood figures, that the avatar has no agency (see also Silvio 2008): “Avatars were not just placeholders for selfhood, but sites of self-making in their own right…. Yet the dynamic was more complicated that the residents controlling agentless avatars. Many felt the avatar appearance affected their behavior” (149). Through a kind of ‘proteus effect’, avatars can reciprocally affect the person who controls them, allowing that person to become more outgoing, for example (149). But this is only surprising from the point of view of a theory of performance that assumes that one’s self is a set of exteriorizations of a transcendent inner self. As we see in other genres, personhood is partially created by imputation (see Silvio on fans who impute personhood to cartoon characters placing this on a continuum with imputation of personhood to religious figures, also discussion of the bivalency of puppets in Prague school from Shershow 214. To paraphrase Latour: “When we say they [e.g. avatars, puppets, machines, figures] are “mere automatisms.” We project as much as when we say they are “loving creatures;” the only difference is that the latter is an anthropomorphism and the former a technomorphism or phusiomorphism” (Latour 1992: 241). In answer to Boellstorff’s question ‘Can the Avatar speak?’, beside Latour’s observations we might append the Russian director Meyerhold’s much earlier observations on the puppet theatre, which Meyerhold approaches both from the perspective of ‘what the puppeteer wants’, and the ‘what the puppet wants’. Meyerhold creates two imaginary puppeteers or puppet theatre directors, each of which embodies one aspect of the Zichian perspective on puppets, or alternatively, two approaches to the ‘immersive fallacy’ (), one based on the idea that the virtual world should be a naturalistic ‘other’ actual world, the other that asserts the autonomy of the virtual world. The first director embraces naturalism, and wants his puppets to imitate real people, until finally he realizes that the simplest solution is to replace his realistic puppets with real human actors.

….[Like the first director], the other director wanted to make his puppets imitate real people, too, but he quickly realized that as soon as he tried to improve the puppet’s mechanism it lost part of its charm. It was as though the puppet were resisting such barbarous improvements with all its being. The director came to his senses when he realized that there was a limit beyond which there is no alternative but to replace a puppet with a man. But how could he part with the puppet which had created a world of enchantment with its incomparable movements, its expressive gestures achieved by some magic known to it alone, its angularity which reaches the heights of true plasticity?

The second director encounters in the puppet, as the SL resident in their avatar, a Peircean secondness, an alterity, an otherness, a set of dispositions, needs, desires, alien to themselves. Like the avatar, the puppet is not just a realism manqué, and the world it creates is not just a slightly unrealistic version of the actual world (cf. Boellstorff’s discussion of the ‘immersive fallacy’). The gap between virtual and actual is not a flaw, a lack of realism that, following the impetus of Meyerhold’s ‘first director’, will be eventually erased by techne in the form of better CGI, graphics engine and interface (though certainly many computer game players and designers are like Meyerhold’s first director). The techne of the puppet or avatar is not trying to erase the gap between the actual and the virtual world, but to create it (150):

The puppet did not want to become an exact replica of man, because the world of the puppet is a wonderland of make-believe, and the man which it impersonates is a make-believe man. The stage of the puppet theatre is a sounding board for the strings of the puppet’s art. On this stage things are not as they are because nature is like that but because that is how the puppet wishes it—and wishes not to copy but to create (129)

All of which brings us back to the matter of worlds. As Boellstorff has shown us, virtual worlds can be profitably studied using some of the same categories we have used in the past to study actual worlds, and all the while, what we learn about technically mediated virtual worlds and performances shed new light on things in the actual world (the debate about voice, for example). In this sense, Boellstorff’s work and choice of data continue the analytic strategies chosen by Goffman, not only the use of dramaturgical models from the stage to study unstaged everyday interaction, but also, his eclectic choice of data (‘cartoons, comics, novels, the cinema, and especially, it turns out, the legitimate stage’) and Goffman’s refusal to countenance an analytical priority for ‘actual’ experience over ‘the other kinds’ (Goffman 1974: 15). But, virtual worlds are, indeed, worlds with a difference, too, because they are, indeed, often owned by corporations at whose whim they could be closed at any point, and the sense of reality they have is that conferred by the liveliness and commitment of the population of the avatars who frequent them and who may, without warning, log off and never return for whatever reason, including precisely because of the changes in the world made by the owning corporation, or sometimes, they simply die in the actual world (126-8). Boellstorff notes (58), citing Kelty (2005: 185), that online communities therefore sustain a strong sense of metawareness about their own contingency, and, perhaps, fatality. Each game is a prisoner’s dilemma, as one immerses oneself, commits to, gears into, this persistent virtual world with the understanding that others will, as well, and when a population or popularity of a game world seems to be declining, this sense of impending doom, of a dying world, can cause stampedes to avoid being the last avatar standing. As Boellstorff notes (182), citing Taylor (2006: 47), nothing is more lonely, more haunting, than an online place that is empty of other avatars, inhabited only by avatar automata (182). Such avatar-less virtual landscapes produce the same melancholy sense of loneliness as famous scenes of empty public urban spaces from apocalyptic films like On the Beach or 28 Days Later.

But such contingencies produce new possibilities as well. Virtual worlds have been around long enough now that some have, indeed, disappeared and taken their communities with them. One such world, the virtual world of Atys of the game Saga of Ryzom has passed from one bankrupt owner to another, with progressively worse server support (so that the final period of the last owner was called ‘the Time of the Great Lag’). When I joined the game a couple of years back it was in extremis, and yet the few avatars who were present were unusually enthusiastic in greeting me and showing me around a world that seemed almost empty and very laggy. Later, I logged on again, and found the game no longer was online. Inquiring into the background, I found that the game had not only a lively community of refugees, but that this game had also been the focus of a concerted attempt by ex-designers and players to assert community control over a gaming world. The Free Ryzom Campaign (which solicited donations during the liquidation of the bankrupt previous owner’s assets) formed one important campaign of the Virtual Citizenship Association, whose goal was not only to restore Ryzom, but to make it a free MMORPG, in line with a virtual ‘social contract’ that would commit the Association, aong other things, to (1) turn Ryzom into free software, (2) ‘give back to the community’, (5) make the avatars into ‘the property of their respective players’ and finally (9) base their decision process on ‘participative democracy’ ( On this level, the vanishing world of Ryzom, became briefly an opportunity to turn a virtual world into a utopia enacting a new virtual social contract, with models of property and governance that deserve comparison with Boellstorff’s discussion of ‘creationist capitalism’ in SL (chapter 8). These actions were closely monitored by player refugee groups who reconstituted themselves at independent websites and forums. But for most of the players the end of Ryzom was not an opportunity to build a freeware utopia, a new kind of virtual world at the abstract level of property and code, but a tragedy of the loss of the concrete world of Atys. These communities constituted themselves as communities of exile, of mourning for the landscape of lost world, of ‘yearning… to dig with the threat of death by failure to careplan, yet with a relaxing serenity I’ve found nowhere else as all around the animals go about their business, fellow homins dart by, insects flit, fog and luminous spores drift from the ground, the trees and grasses sway, the weather changes….’ (Katriell, May 20 2008 Re: Yearning… citation).[6]

While the Virtual Citizenship Association was not able to realize their goal of resurrecting Ryzom both as Atys and as a freesoftware utopia, the game did, suddenly and without warning, reappear under new ownership, and the existing player base was invited this year to return to Atys and play for free. I don’t know whether this is a unique event in MMO history or not, but certainly it was a strange homecoming, a resurrection of a world, a sudden re-efflorescence of an exiled virtual community. This strikes me as the sort of thing we couldn’t possibly make up, and probably merits our attention. And this is the book that invites us to pay attention.

Citations (not nearly done)

Allison, Anne. 2006a. Millenial Monsters: Japanese Toys and the Global Imagination. Durham, NC: Duke.

Balfe, Myles (2004) ‘Incredible geographies? Orientalism and Genre Fantasy’, Social & Cultural Geography, 5:1, 75 — 90

Barker, Josh. 2008. Language & Communication

Bauman, Richard. 2004. A World of Others’ Words: Cross-cultural perspectives on intertextuality. Blackwell.

Stephen Crook. 1998. Minotaurs and other monsters: ‘everyday

Life’ in recent social theory sociology vol. 32 no. 3 august 1998 523–540

Ellis, M., 2005. The Coffee House: A Cultural History. Phoenix, London.

Ellis, Markman. 2008. Introduction to the Coffee House: a discursive model. Language and Communication 28: 156-164.

Keane, Webb. 2001. Voice. In “Key Terms in Language and Culture”, ed. Alessandro Duranti. Malden, MA & Oxford: Blackwell, pp. 268-271, 2001)

Keane, Webb. 2003 Semiotics and the social analysis of material things. Language and Communication 23 (3-4), 409-425.

Kotani, Mari. Doll beauties and cosplay. Frenchy Lunning (ed.), Mechademia 2: Networks of Desire, 49-63. Mineapolis: U Minnesota Press.

Marx, Leo. Technology: The Emergence of a Hazardous Concept

Owen, Alex. 2004. The Place of Enchantment: British occultism and the Culture of the Modern. University of Chicago press.

Laurier, Eric and Chris Philo. 2007. ‘A parcel of muddling muckworms’: revisiting Habermas and the English coffee-houses. Social & Cultural Geography, Volume 8, Issue 2 April 2007 , pages 259 – 281

Munn, Nancy. 1977 Spatiotemporal Transformations of Gawa Canoes. Journal de la Societe de Oceanistes 33: 39-55.

Munn, Nancy. 1986 The Fame of Gawa: a Symbolic Study of Value Transformations in a Massim (Papua, New Guineau) Society. Durham, NC: Duke.

Scott-Elliot, W. 1925. The Story of Atlantis and The Lost Lemuria. London: Theosophical Publishing House.

Lovecraft, H. P. 1927, 1933-5. The supernatural horror in literature. Http://

Bulwer-Lytton, Edward. 1871. The Coming Race. Electronic text

Seed, David. 2005 Introduction. In David Seed (ed.), Edward Bulwer-Lytton’ The Coming Race, xiii-liii. Middletown, CT: Wesleyan University Press.

Vernant, J. P. 1991 The figure and functions of Artemis in myth and cult. In J. P. Vernant, Mortals and Immortals: Collected Essays. Edited by Froma Zeitlin, pp. 195-206. Princeton: Princeton University Press.

Taylor, Charles. 2002. Modern social imaginaries. Public Culture 14.1, 91-124.

2004. Modern Social Imaginaries. Duke: Durham.

Warner, M., 2002. Publics and counterpublics. Public Culture 14 (1), 49-90.

Ramaswamy, Sumathi.2004. The Lost Land of Lemuria: Fabulous Geographies, Catastrophic Histories. Berkeley

Daniel Rosenberg 2000 Speaking Martian. Cabinet Issue 1 Winter 2000/01

Said, Edward. 1978. Orientalism.
Silvio, Terry. 2008. Pop culture icons: religious inflections of the character toy in Taiwan. Frenchy Lunning (ed.), Mechademia 3: The Limits of the Human, 200-221. Mineapolis: U Minnesota Press.

Smith, Clark Ashton. 2003. Selected letters of Clark Ashton Smith. Edited by David Schultz and Scott Connors. Sauk City, WI: Arkham House.

Williams, Rosalind. 1992. Notes on the Underground: An Essay on Technology, Society, and the Imagination. MIT

Winge, Theresa. 2006. Costuming the imagination: origins of anime and manga cosplay. Frenchy Lunning (ed.), Mechademia 1: Emerging Worlds of Anime and Manga, 65-77. Mineapolis: U Minnesota Press.

Latour, Bruno 1992. Where are the Missing Masses?

Brake, Laurel. 1994. Subjugated Knowledges: Journalism, Gender, and Literature in the Nineteenth Century. New York: New York University Press.

Akrich and Latour 1992

Levi Strauss, C. 1973. Tristes Tropiques. Atheneum.

Bernard Deacon, 2006 Cornish or Klingon? The standardisation of the Cornish language.

Cornish or Klingon? The standardisation of the Cornish language.

(published in Philip Payton, Cornish Studies Fourteen, University of Exeter Press, 2006)


Richard Cairney 2005. U of A linguist puts words in video-game characters’ mouths Feb 24, 2005.

[1] Episode 1008 ‘Make love, not Warcraft’ (originally aired Oct 4 2006). By contrast, the popular anime series Lucky*Star, whose main character is an otaku (geek) girl named Konata Izumi (Kona-chan), presents a much more sympathetic and convincing insider view of online gaming subcultures. In different episodes, Konata is shown teaming with her homeroom teacher (Nanako Kuroi) as a deadly tank and mage combo online, and then having the teacher break frame in ingame chat with ‘real life’ concerns about the negative effects of her gaming on her study habits. In other episodes we see the difficulty one of her real life friends (Tsukasa Hiiragi) has, as a newbie, in controlling her avatar, or the impromptu online celebration Konata throws with her online friends, including buying them online ‘gifts’, when she acquires an ultra-rare sword as a ‘drop’.

[2] I note that there is absolutely nothing novel about the vendibility for ‘real money’ of immaterial commodities, a commonplace since Say (Manning 2006), this is as true of actual as it is virtual capitalism, the specifica differentia for why these exchanges of real world money for virtual swords and armor seem strange has to be located elsewhere (on which Boellstorff devotes a whole chapter (chapter 8)).

[3] Indeed, many have made prec isely the argument that early PBM games like the Tribes of Crane were ‘the infancy of cyberspace’ (

[4] A widespread practice, e.g. Meha Verghese (2008) Lord of the Languages: Prof teaches with Elvish, Orkish. Brown Daily Herald 11/230/07, posted at .

[5] And increasingly not only in fantastic worlds in movies or television but also game worlds are replacing simple alien ‘gibberish’ with actual invented languages, often by the same Klingon-speaking linguists who use such languages in their course materials. One linguist hired to create languages for two games made by the game company Bioware comments on the difference between ‘fake’ ‘gibberish’ and ‘authentic sounding fantasy languages’: “The design team decided that it would be a great idea to have their non-English speaking characters or even non-human characters have authentic sounding fantasy languages. My feeling is that gibberish instantly compromises the entertainment experience because it is fake. I say it as a gamer and a film viewer. And movies especially – the Star Trek franchise has worked hard to give languages a sense of authenticity. The Lord of Rings movies did the same thing with Elven.” (Wolf Wikeley, cited in]

[6] Ryzom is also the only MMO I am aware of that has in world animal rights activists, specifically to prevent cruelty to Yubos, a small, cute herbivore that is the object of newbie hunting missions.