1.1

Networked virtual reality (VR) is a rapidly growing area of research and development. And although the implications of this technology as a communications medium are still unclear, the potential for more widespread uses in the future raises a number of questions: how is VR becoming integrated with existing information and communication networks? What are the aims of developers, and how are they translated into everyday uses of the technology? How do users interact with each other inside virtual worlds? And - a question that has been raised about computer-mediated-communication generally - do virtual worlds create novel social identities or communities?

1.2

So far there has been little research on the social aspects of networked VR or related technologies for computer-mediated-communication (see the review by Garton and Wellman, 1995). Yet this should not prevent us from exploring the social phenomena surrounding the development and current uses of multi-user VR technology. Even if the technology is still in its infancy, research on other new technologies has shown that early stages of development are often crucial for the social implications that follow later.

1.3

Before launching into these areas, some definitions are required: 'Immersive VR' will be used to refer to a computer-generated display by means of which users have a sense of being present in an environment other than the one they are actually in and to interact with this environment (see Schroeder, 1996: p. 25). 'Desktop VR' systems only display the virtual environment on a 2-D desktop computer screen, and therefore do not allow the kind of immersion that is experienced with head-mounted displays and other input/output devices like gloves or hand-held 3-D joysticks. But like immersive VR, they provide a first-person perspective on a 3-D computer-generated world. Finally, 'second-person VR' systems represent the user as an avatar (or figure) on the screen without a first-person perspective. These systems therefore only give a feeling of being immersed in a 3-D (or sometimes 2 1/2-D) computer-generated environment at one remove, so to speak. Nevertheless, the boundaries between these systems are fluid and, as we shall see, there are good reasons for including all of these types of systems here - even though desktop and second-person VR stretch the limits of what is meant by 'VR'.

Forerunners: Networked Interactive Computer Graphics

2.1

Networked interactive computer graphics were first demonstrated in 1972 on ARPANET, the computer network developed by Advanced Research Projects Agency (Norberg and O'Neill, 1996: pp. 119-52).^[1] One of the main purposes of this project was to allow for time-sharing of what were then costly and scarce computer processing resources. These networks were thus mainly intended for co-operative work and for sharing information. It is not necessary here to go into detail about the history of the use of networked computers for graphics; the main point in this context is that these networks have subsequently come to be used for a variety of purposes, including shared scientific visualization, training, cooperative work, military battlefield simulation, and entertainment games - the last two being particularly important for the development of multi-user VR.

2.2

Networked battlefield simulations have been developed in the United States from the mid-1980s onwards, from SIMNET (SIMmulator NETwork) to its current incarnation in the NPSNET (Naval Postgraduate School Networked vehicle simulator) project. A number of features of the NPSNET system are noteworthy: first, this is still the most complex networked interactive virtual environment. It allows the largest number of simultaneous participants (several hundred) in an immersive virtual environment and consists of the largest and most complex virtual world (more than one hundred realistically-rendered square kilometres, see Pratt et al, 1995: p. 90). Secondly, the driving force behind the network has been the lowering of training costs. It is much cheaper to bring together personnel and blow up virtual vehicles on a computer network than in real-life training exercises (Stytz, 1994). On the other hand, it is, of course, very expensive to build this network in the first place, but this has not been an insurmountable obstacle for the American military establishment. And finally, on a more technical note, while the network initially only supported desktop VR via T1 lines (1.5 megabit/second), it has now been upgraded to link immersive VR systems via T3 links (45 Mbits/sec). Hence this is also the most 'capacious' interactive wide-area network to date.

2.3

And yet, this military network has so far exercised little influence on the world of civilian VR. It is possible that it may do so in the future, especially since the software standards for this network are being considered for application in multi-user entertainment games (Katz, 1994). But this depends on a number of social and technical factors and, as we shall see, these factors are inextricably intertwined in networked VR systems. In any case, military training simulations have the longest history in the use of large-scale computer networks to engage with computer-generated worlds on a day-to-day basis.

2.4

The second area that has been important for multi-user VR are networked computer games. The use of networked computer gaming can be traced to text-based Multi-User-Dungeons or MUD's, more recently also known as Multi-User-Dimensions (Bartle, 1990; Cawson et al, 1995: pp. 135 - 174). From the mid-1980s until very recently, the only networked game with interactive computer graphics was Habitat (Morningstar and Farmer, 1991, see figure 1), first piloted in the US (where it failed) and subsequently in Japan, where it has been operating continuously.^[2] With the exception of networked computer games that have been part of interactive cable TV trials (for an overview, see Baldwin, McVoy and Steinfield, 1996), these games have so far only been carried over standard internet connections (telephone lines) - unlike military simulations which are carried via bundled and dedicated telephone lines.

2.5

With the recent take-off in the number of domestic internet users, it is not surprising that a number of developers are now starting to produce shared virtual entertainment worlds carried over telephone lines (a selection of these, which includes virtual worlds for entertainment and non-entertainment purposes, can be found on the internet at <http://www.ccon.org>). The limited amount of bandwidth has meant, however, that in most systems, it is only possible to exchange text-messages and to represent the user as an avatar in a two- and latterly simple three-dimensional world (although a few systems permit voice communication and some non-entertainment systems feature highly sophisticated and realistic graphics). Bandwidth and telephone charges thus remain the major obstacles to the widespread adoption of this use of networked computing.

2.6

It is conceivable that the increasing popularity of dial-up internet services and lower telephone tariffs may alter this picture. Even so, the problem of bandwidth remains: existing internet services put severe constraints on the complexity of virtual worlds and the on interaction with other users. More complex worlds and forms of interaction require more capacious telephone lines like ISDN (Integrated Services Digital Network, 64 kilobit/second), and these are still expensive for domestic users. The future of networked computer gaming therefore depends, as in military simulations, on aligning a number of technical and social forces - except that in this case, with a large number of homes and offices already equipped with internet connections, the balance is tilted more towards demand 'pull' for multi-user entertainment virtual worlds - as opposed to the technological 'push' that has been required for creating networked military training simulations.

Networked Virtual Reality

3.1

What has just been said applies to large networks connecting many users at a distance, but the possibility of putting two or more users within the same virtual world in the form of local networks was a part of VR systems development from the start. When Jaron Lanier introduced the first commercial VR systems in the late 1980s and early 1990s, one configuration was a system for two users, VPL's RB2 or 'Reality Built for Two'. Similarly, early immersive VR games like those produced by the firm Virtuality were available in a networked configuration from the time when they were first introduced in the early 1990s, typically for between 2 and 4 players. And although immersive VR (apart from prototypes and arcade games) has mainly been a single-user or stand-alone technology, various projects for multi-user systems or shared virtual worlds have been ongoing ever since.

3.2

At this point a distinction must be made between VR operating on local area networks (LANs) and VR operating on wide area networks (WANs). LANs typically rely on ethernet links that operate at 10 Mbits/sec, although some links between supercomputers are capable of 100 Mbits/sec. Ethernet is fast enough to allow multi-participant interaction in virtual worlds generated by the most powerful VR systems that are now commercially available. In these systems several simultaneous users with head-mounted displays and 3-D joysticks can move around and manipulate the virtual world. WAN communication networks like the internet, on the other hand, are typically limited to the bandwidth of telephony, 28 kilobits/second. Internet-based VR is therefore currently restricted to desktop and second-person VR. It allows only limited interaction with the virtual world (the change of position of one point, ie. the user's avatar) and there is little detail and limited manipulability in the three-dimensional world.

3.3

Locally networked immersive VR systems are currently either used for specialized purposes like shared engineering visualizations in laboratories and firms, or for arcade games. And while there are a number of trial uses of WAN- or internet-based VR systems, including projects that are at the experimental stage in universities, the most widespread use of commercial desktop virtual worlds is for entertainment. These virtual worlds can often be downloaded for free from the internet and they are accessible to anyone with a personal computer (a Pentium is typically required for reasonable graphics) and an internet connection.

3.4

This difference corresponds roughly to the differentiated market that is emerging between high-end and low-end stand-alone VR systems (Schroeder, 1996: pp. 139 - 142). At the low end, there are many software packages for building and interacting with virtual worlds on a personal computer, some of which, again, are available for free on the internet. These low-end systems can be used in conjunction with 3-D head mounted displays and 3-D mice that are priced for the consumer market. The performance of these systems, with low-resolution displays and minimally interactive worlds, is different from high-end systems which cost tens of thousands of dollars and yield very complex virtual worlds which look more 'realistic' and which the user can manipulate in a more sophisticated way. Nevertheless, the distance between these two markets is shrinking both in price and performance, particularly as there is fierce competition in computer graphics hardware and software.

3.5

This range of systems and of networks will shape the nature of networked VR in the years to come. Trade-offs will have to be made between the cost of VR systems and the complexity and interactivity of the virtual world one side - and the number of users on the other. Hughes' notion of 'large technological systems'(Hughes, 1987) is useful here. Hughes refers to 'reverse salients'(instead of trade-offs or bottlenecks) to describe the places in large technological systems where key obstacles need to be overcome that are simultaneously technical and social. In the case of networked VR, it is clear that these obstacles include several technologies (modems, computer graphics capabilities, telephone lines) as well as social institutions (for example, the cartels created by computer manufacturers and telephone companies, standards bodies like the Virtual Reality Modelling Language (VRML) architecture group, and government regulators). How these social and technical constituents intertwine, or how the reverse salients where they are conjoined are surmounted, thus depends on how high-end and low-end systems converge and what kind of networks emerge for computer-mediated communication and for what purpose. For the moment, we merely need to register that high-end and immersive VR systems have thus far only been used on a trial basis. They have therefore been used for short periods and for specific experimental purposes such as co-operative visualization. Low-end internet-based VR, on the other hand, has been widely used for long periods in the forms of desktop and second-person VR. The latter are therefore the most interesting from the point of view of the social interactions between users.

Social Interaction In Multi-User Desktop VR

4.1

There are now more than a dozen internet-based desktop and second-person VR systems (for a survey, see Roehl, 1996; 1997). Here I will concentrate on two examples, Circle of Fire's Alphaworld (also known as Activeworld) and Blacksun's Cybergate (see figures 2 and 3), both of which have been popular virtual worlds. They have both been used primarily for entertainment, but in addition they have been used for educational purposes, for holding meetings, and for advertising promotions. The account that follows is based on participant observation in these multi-user worlds between September 1996 and September 1997. During this period, I spent more than fifty hours in Alphaworlds, Cybergate, Worldsaway and Onlive Technologies. More than half of this time was spent in Alphaworlds, which is the most highly populated of these worlds.

4.2

In these two systems, users are represented by avatars which they choose from a selection - the shape of a human body in the case of Alphaworld, and a cartoon-like imaginary creature in Cybergate^[3] - but they have first-person perspective on the world. The avatars communicate by means of text which is displayed in a window underneath the world. In Alphaworld, the text is also displayed on top of the 'speaker's' avatar. Both worlds allow many users to 'talk' to each other simultaneously or choose to switch to one-to-one conversation. Users move around in a series of interconnected rooms or worlds and they can build objects to add to the world. Navigation is by mouse, but there are other options such as double-clicking on where you want to go and teleporting by typing in coordinates. In Alphaworld, it is also possible to adopt a birds-eye perspective on one's own avatar.

4.3

What are the sociologically interesting features of these two multi-user VR systems? First, both systems borrow extensively from other technologies and formats, particularly text-based MUD's and computer games. The same applies to the behaviours of users, which display some of traits that can be found in other electronically-mediated-communication, such as aggressive verbal exchanges. One way in which they depart from these technologies, however, is in terms of the degree of engagement with other participants and with the virtual world. Users are much more involved in these worlds than in text-based MUD's or in 2-D worlds with limited interaction.^[4]

4.4

One way to get a sense of this greater degree of involvement in the virtual world is by giving a brief account of the typical experience of users. On entering the virtual world, the convention is to move around to get a sense of the place. When this is done, users will start a 'conversation' with others, beginning with 'where-are-you-from?' or 'what do you do in the real world?' questions to try to establish a rapport. They also tend to talk about their initial experiences with the virtual world and its features. In short, these are the smalltalk introductory conversations of strangers - as in the real world.

4.5

At this stage, it becomes necessary to make a distinction between novices or 'outsiders' and the regulars or 'insiders'.^[5] 'Insiders' will talk and behave with each other in a knowing way, building on their familiarity with the conventions of talking in and interacting with the virtual world and their adeptness at navigating through it and manipulating it. They will entertain each other, for example, by inviting each other to go places together or to have parties. They will also try out special features of the virtual world such as flying or building their own houses. Moreover, an insider will often develop a distinctive 'persona', such as the 'prankster', the 'bully', the 'helpful guide', or the 'seen-it-all cosmopolitan'. In other words, advanced users will be able to present or 'stage' themselves. Here it is important to remember that computer-mediated-communication allows users to misrepresent their real 'personality', a point that has often been made in relation to the internet.^[6]

4.6

'Outsiders', in contrast, act like tourists. They only engage in novice or 'superficial' conversations and remain on the periphery of conversations or of 'the action'. If, among 'insiders', there is a distinctive sense of belonging that is expressed in certain types of behaviours, it seems that the majority of users stay at the level of curious tourists whose fleeting acquaintance with the virtual world does not allow or enable them to participate in the more close-knit social networks of 'insiders'. The implication is that virtual worlds produce 'stratification' - in the sense that different groups have distinctive behaviours and roles that set them apart from other groups with a different status or with a different sense of cultural cohesion.^[7]

4.7

This stratification within multi-user virtual worlds relates closely to a second feature; spatial concentration. Alphaworld and Cybergate both contain many separate worlds, but the highest concentration of users is almost invariably in a central place - 'Alphaworld' itself inside Alphaworld and 'Pointworld' in Cybergate. After entering through 'The Gate' world in Alphaworld and choosing an avatar in 'Avatar Room' in Cybergate, these are the two spaces in which users congregate, typically in a small cluster at the center of the space.

4.8

'Insiders', however, tend to explore the further reaches of these spaces and other worlds. They also build new objects in these further reaches, create new worlds (Alphaworld currently contains more than 150 worlds), take people (usually other 'insiders') with them to see their creations, and the like. Often the reason for going to remote worlds is to gain a sense of privacy with other users or to develop closer ties with small groups. Center and periphery thus map closely onto the insider/outsider split, and this point is reinforced by the fact that even insiders must nevertheless spend time in the center in order to socialize.

4.9

Stratification and spatial concentration thus go against the notion that 'virtual communities' are completely egalitarian or boundaryless. Virtual worlds tend to become socially differentiated and spatially segregated - even if they have other features that could be construed as displaying egalitarianism or boundarylessness (the absence of immediate age and gender identification, for example, or the possibility of building one's own space in a relatively unrestricted manner).

4.10

A full sociological account of social relations inside these virtual worlds would need to put these features into the wider context of how this technology is coming to be used. That is, it would require setting these online relationships in the context of how the use of this communication tool fits into the everyday life of users, surveys of its popularity ^[8], and comparison with other similar technologies. It is impossible to cover all the aspects in a short paper^[9]. If, however, we regard the sociological importance of new technologies as lying in their intensification or extension of the novel opportunities and constraints for refining and manipulating the natural and social worlds^[10], then the significance of multi-user virtual worlds can be seen as lying in the fact that they take existing technologies for text-based computer-mediated-communication a step further, and thus intensifying or extending the features which were novel in these technologies. Hence verbal aggression can become aggression directed at avatars, spending time listening in or 'lurking' on other conversations becomes hanging around other avatars, sexually explicit conversations become simulated sexual encounters between avatars, exploring different milieus in imaginary text-based rooms becomes wandering about in landscapes and rooms with different visual atmospheres, relations that rely on keeping the other's attention by saying interesting or entertaining things to each other turns into the requirement to be able to 'do' things together, a sense of belonging to an established group created through text-based exchanges becomes a sense of belonging created through physical proximity or territoriality, and so on.

4.11

This intensification or extension of the features of text-based communication can be elaborated by reference to the notions of communication 'channels' and social 'cues' that have been used in the study of other media. Both refer to the modalities by which communication takes place. A communication channel relates to the senses (hearing, seeing, etc.) and social cues to the various types of verbal and non-verbal expressive behaviour that are used in communicating. Different communication technologies rely on different channels and allow different social cues to be conveyed. Email, for example, makes use of emoticons to convey mood, but these have a different 'channel bandwidth' and are governed by different social conventions from emotions that are expressed via videophones or in face-to-face encounters. A shift away from verbal communication in email and MUD's and towards the visual representation of users and of the environment therefore also entails a shift in the channels of communication and the cues that come with it.

4.12

It would be useful at this point to undertake a systematic comparison between the channels and cues in networked VR and other communication technologies. Instead, I will restrict myself to making the point that it is impossible to regard desktop VR as (text-based) MUD's plus pictures. The combined effect of using text, navigating in a 3-D environment and engaging with other users via avatars - in terms of the behaviours described above - is greater than the sum of its parts. One example, as I mentioned above, is how being able to seek a more private conversation with another user or group of users is not just a matter of a one-to-one communications channel, but of seeking spaces that allow privacy for a group of avatars.

4.13

This distinguishing feature of multi-user VR systems allows us to make some brief comparisons between the social relations inside virtual worlds and the social relations attendant upon the uses of related technologies. Low-end desktop VR systems for entertainment have some similarities with internet-based computer games, but the social interaction inside multi-user virtual worlds is quite different: in popular competitive networked games like Doom or Quake, the ties between users are structured by specific game tasks. The behaviour and the ties between users in worlds like Alphaworld in Cybergate, in contrast, are relatively unstructured. In a similar vein, the ties between users in these two systems are also not instrumental as they are, for example, in the context of computer-supported-cooperative work. This does not mean, however, that these relations are purely spontaneous or that a non-instrumental 'virtual society' or 'community' is created: social relations inside these worlds are social, as Homans puts it, 'in the special sense of "social", meaning occasions where there is no compelling job to be done' (Homans, 1961: p. 317). They can be seen as a combination of tourism and fraternizing pursued for its own sake. Virtual worlds have this quality of 'tourism' in common with 'surfing' the World-Wide-Web, but unlike surfing the web, in this case the user explores these spaces together with other users. And again, virtual worlds have some similarities with certain types of text-based MUD's, but multi-user VR differs in so far as it provides a space for 'face-to-face' interaction between avatars.

4.14

At this stage we can relate the features of social interaction in desktop virtual worlds back to the discussion of high-end vs. low-end VR systems and of wide-area vs. local-area networks. Again, high-end fully immersive VR systems have not been used extensively enough to yield information about social interaction in virtual worlds. They have mainly been used for short periods, over local area networks, and for demonstration or proof-of-concept purposes. Desktop and second-person systems, on the other hand, have been widely used via the internet, mainly for entertainment. The development and the applications of the two types of systems are likely to have different trajectories, but the pairing of high-end systems with local networks and low-end systems with the internet is one way that VR has already congealed around a particular combination of technology and its uses. Social phenomena inside virtual worlds therefore need to be placed in the context that their possibilities and constraints are yoked to this state of socio-technical development.

Conclusion

5.1

This paper has traced the emergence of networked VR as a technological system and explored some of the social interactions between users inside virtual worlds. It remains to draw out some of the links between the two areas. As we have seen, the direction of networked VR as a large technological system is determined by several intertwined social and technological factors. On the side of the technology, again, examples include the graphics capabilities of the machines, compression techniques for transmitting data between computers, and computer processing power. On the social side, they include the number of adopters, standards bodies, and the competition between internet service providers. The key point here is that the technological system shapes the micro-social interactions between users - how they are represented in the virtual world, interact with each other, and the complexity of the virtual world. The shaping that takes place in the other direction, from the social to the technology, arises primarily from the fact that the popularity of different systems, worlds and features is leading to the development of new types of VR systems and more capacious and extensive network connections. In this way users adapt their needs for interacting with other users and for manipulating the virtual world to the constraints and possibilities of the systems - and these systems, in turn, adapt to the needs of its users.

5.2

To mention just two examples of how users and uses of the system exercise an influence on the technology: some features of the Cybergate system, such as the replay function (for recording scenes and reviewing them) are not being used at all, and the whole Cybergate system is being abandoned because it has become less popular (only a few occasional visitors pass through Cybergate worlds nowadays, which gives them the appearance of ghost towns). Alphaworld, on the other hand, is becoming ever more popular and some features, such as being able to locate other users, have been added in response to popular request. Moreover, these user-driven features are being further modified by users: the locating system, for example, is to be changed in the future so that users can remain 'invisible' if they wish.

5.3

Social relations within virtual worlds therefore depend on how networked VR technology has congealed into a particular combination of tools that are used on an everyday basis. Thus there are two separate levels to the development of networked VR, with the level of the technology as a system shaping how virtual worlds are used - and how users, in turn, are shaping the system. Much of the writing about using computers as tools for communication concentrates exclusively on the first level, and this is justified insofar as the technology sets the parameters for what goes on inside virtual worlds. At the same time, it would be ill-advised to wait with the study of the macro- or the micro-level of the social aspects of this technological system until it has taken on a fixed or permanent shape - since there are already many interesting social phenomena and interrelations between the two levels emerging in the technology's initial uses.

5.4

But the question of the interrelation between technological and social shaping in this case is more complex than this. This shaping should not just be regarded as being simultaneous from both sides, but of the two sides shaping each other independently: in view of the fact that desktop VR (and, in a different way, immersive VR, see Schroeder, 1996: pp. 99 - 122) lacks many of the subtle cues and the multi-channel nature of face-to-face communication, the diffusion of VR and VR-like technologies means that people will acquire a mode of communication to fit these technologies regardless of their (micro-social) possibilities and limitations. Hence new (micro-) social relations will emerge that are a product of the novel characteristics of this technology. An example, as we saw earlier, is the extent to which users differentiate themselves into groups with distinct characteristics and how they congregate and disperse spatially. Another example that has not been discussed (for reasons of space) is the way in which users take the opportunity to modify the virtual world, to build new objects and spaces, which, in turn, creates new possibilities and constraints for sociability (one possibility in relation to building in the virtual world that needs to mentioned here, however, is that users may mainly avail themselves of the 'pre-manufactured' spaces and objects, as the vast majority of users do now).

5.5

A further example of the interplay between the technological system and the social relations inside virtual worlds - in this case an example which is not just to do with the relations between individual users, but with groups of users - is the fee-paying structure of these systems. Initially, Alphaworld was free to users (Cybergate and some other systems still are). Recently, however, Alphaworld has begun to charge a fee ($19.95 for two years). Users who make full use of the virtual world (choosing avatars, messaging, building in the world) must pay this fee. 'Tourists', on the other hand, can still write messages to each other and wander around, but they do not have these privileges. Also, tourists are immediately identifiable; they have single 'standard issue' avatar - with a camera hanging from their shoulders! Another example of a system that began by being free to users and that has subsequently become available only to fee paying subscribers is the Worldsaway system which was mentioned earlier. In this case however, no free 'tourist' status is available.

5.6

The sociological interest of this change is not only that, for example, a two-tier structure of citizenship has been created in Alphaworlds ('tourists' versus full users). It is also that the providers of these systems have deployed a tactic of 'bait and switch' - or hooking the customer and then changing the rules of the game.^[11] What is of interest here, in turn, is that this tactic can be seen as not only being based on the popularity of the virtual world, but also on the strength of the social bonds between users: the social ties within the virtual world, and particularly the bonds between 'insiders', have been exploited for commercial gain because users who have made friends with other avatars do not want to be locked out. If we think about the implications of this kind of segregation, then it is possible to conceive of further differentiations in addition to that between first and second class citizens: perhaps a more distinctions will develop in networked virtual worlds between inhabitants with finer gradations of stronger and weaker social ties, or between various levels of active and passive 'citizens'.

5.7

It should be added immediately that this differentiation of groups and roles is not simply a matter of the fee paying structure of the system. Similar observations can be made, for example, about ExploreNet, which is a free multi-user second-person VR system that has been used in education. ExploreNet (see figure 4), a system based on Habitat, has been designed so that users have different roles, such as experienced 'world builders' who modify the world, 'cast members' who help pupils, and 'guests' who are new to the learning task and to the world. In this case, too, avatars have different capabilities in terms of how they are able to interact with each other and with the world.^[12] But here the differentiation of roles is to do with facilitating the learning process - mainly so that users who are familiar with the system can give guidance to novice pupils - and not with fee-paying.

5.8

I have mentioned this example only because role- or group differentiation inside multi-user virtual worlds can become part of the system for a variety of reasons. The important point about the systems that have been discussed here is that segregation into groups and differentiation between roles is based both on the system - the purpose to which the system is put and how this purpose is technically achieved - and on the changing patterns of use among participants within the virtual world. This link between the two levels points to the fact that differentiation is emerging not only within - but also between - different virtual worlds. What, for example, is the relation between fee-paying virtual worlds and free ones that are otherwise broadly the same? And what is to prevent users from switching from one to the other, apart from the attachments to people and places that they have formed in particular virtual worlds? It is easy to see that these are issues which, again, link the level of the 'large technological system' to the level of relationships within the virtual world, and that social relationships such as 'stratification' will be reproduced here that can be found in the social world at large.

5.9

As mentioned in the first part of the paper, the difference between high-end and low-end VR technology is shrinking. Advances in VR technology will therefore enhance the manipulability and flexibility of virtual worlds, even if the capacity of the network remains a stubborn 'reverse salient' within this 'technological system'. As we can see now, however, the development of these technological systems also depends on social interaction within virtual worlds: how popular and sociable they are, and whether they are used for minimally interactive entertainment or for creating dynamic social groups and building attractive worlds. Howsoever these networked worlds may develop in the future, worlds like Alphaworld and Cybergate already provide laboratories for the study of different modes of computer-mediated sociability which confound many of our expectations about the shape of virtual 'communities' - as well as providing an interesting example of social and technological shaping.

Acknowledgements

I would like to thank Ray Lee, two anonymous referees for this journal, the participants of seminars at the University of the West of England and at the American Sociological Association's 1997 conference for very helpful comments and Stuart Peters for his help with the illustrations for this article.

Notes

¹ Contrary to most accounts of the origins of the internet, O'Neill argues that ARPANET, the precursor to the Internet, was 'designed for other than explicitly military objectives'(O'Neill, 1995: p. 76).

² Habitat can be described as a second-person 2-D VR system, but again, this stretches the definition of VR since the sense of being 'present' in the virtual world is only via the avatar - the user has no first-person perspective on the virtual world. Habitat has recently been relaunched globally under the name Worldsaway, from which illustration 1 is taken. A key factor in the success of this system, as Dr.Kazutomo Fukuda (manager of the Cyberspace Systems Development department at Fujitsu, interview September 22, 1994) explained to me, are the telephone tariffs in different countries. In Japan, for example, the system was most popular late at night during the hours of 'off-peak' rates.

³ In Cybergate, it is also possible to design one's own avatar.

⁴ This greater sense of involvement of users can also be gathered from questionnaires: in pilot questionnaires for two separate studies (one in relation to Alphaworld and Cybergate, the other in relation to a networked VR system used in university research), for example, users were asked questions such as whether they feel 'immersed' in the virtual world, whether they feel the interaction with others is 'engaging', or whether they have sense 'of being' in the world or of 'being with others' in the virtual world. Their responses, both in scaled questions and in open questions which asked them to describe their experience, strongly support the idea that they have a sense of being inside the virtual world and with others. It needs to be added that these are only preliminary findings based on small samples.

⁵ The split between insiders and outsiders is, of course, not clear-cut; perhaps it would be better to envisage this divide as a continuum with two extremes. Some of the principles of micro-social or face-to-face interaction that are used here are summarized by Collins (1988: pp. 188 - 371). The discussion of the differentiation between communities in the following paragraph, and how this differentiation relates to status cultures and micro-social interaction, again, relies on Collins' theory (1975: pp. 49 - 160).

⁶ See Curtis (1992) and Turkle (1995). Common forms of misrepresenting oneself are to use several different names, to adopt a false gender, and the like. In this context it may also be useful to point to the studies carried out by Reeves and Nass, which found that people treat cartoon characters (1996: pp. 81 - 87) and other aspects of computer interfaces as more lifelike than they think - or than they are willing to admit.

⁷ A similar differentiation between 'newbies' and more experienced users can also be observed in MUD's, and it would be interesting to investigate the differences between MUD's and networked VR in this regard.

⁸ One indication of the popularity of these systems is that more than 250000 people have visited Alphaworld.

⁹ In relation to the internet and e-mail, some of these areas are discussed by Wellman (1996), McKenney et al (1992) and Nohria and Eccles (1992).

¹⁰ The idea that new technologies extend or intensify the means by which we manipulate the natural and social worlds is one that I have developed elsewhere (Schroeder, 1995, 1996; cf. Rogers, 1986.

¹¹ 'Bait and switch' has been mainly been used in the context of cable television in the United States, where, again, customers are hooked on certain programmes at a particular price which is then changed once the customers have come to rely on the service.

¹² One of the main lessons of the trials which have used ExploreNet is that this differentiation into roles has not been successful among users, who have tended to ignore it (see Hughes and Moshell, 1995).

References

BARTLE, Richard (1990) 'Interactive Multi-User Computer Games', MUSE Ltd. (Unpublished research report), December.

BALDWIN, Thomas; McVOY, D. Stevens and STEINFIELD, Charles (1996) Convergence: Integrating Media, Information and Communication. London: Sage.

CAWSON, Alan, HADDON, Leslie and MILES, Ian (1995) The Shape of Things to Consume: Delivering Information Technology into the Home. Aldershot: Avebury.

COLLINS, Randall (1975) Conflict Sociology - Towards an Explanatory Science. New York: Academic Press.

COLLINS, Randall (1988) Theoretical Sociology. New York: Harcourt Brace Jovanovich.

CURTIS, Pavel (1992) 'Mudding: Social Phenomena in Text-Based Virtual Realities'. Palo Alto: Xerox PARC (Unpublished technical report).

GARTON, Laura and WELLMAN, Barry (1995) 'Social Impacts Electronic Mail in Organizations: A Review of the Research Literature' in Brant Burleson (ed.), Communication Yearbook, vol.18, Sage: London, pp.434 - 453.

HOMANS, George (1961) Social Behaviour: Its Elementary Forms. London: Routledge and Kegan Paul.

HUGHES, Thomas (1987) 'The Evolution of Large Technological Systems' in Wiebe Bijker, Thomas Hughes and Trevor Pinch (editors), The Social Construction of Technological Systems. Cambridge, Mass.: MIT Press, pp. 51 - 82.

HUGHES, Charles and MOSHELL, Michael (1995) 'Shared Virtual Worlds for Education: The ExploreNet Experiment', <http://www.cs.ucf.edu/~ExploreNet/papers/VA.Experiment1195.html>.

KATZ, Warren (1994) 'Military Networking Technology Applied to Location-Based, Theme Park and Home Entertainment Systems', Computer Graphics, vol. 28, no. 2, pp. 110 - 2.

McKENNEY, James, ZACK, Michael and DOHERTY, Victor (1992) 'Complementary Communication Media: A Comparison of Electronic Mail and Face-To-Face Communication in a Programming Team' in Nithin Noria and Robert C. Eccles (editors), Networks and Organizations: Structure, Form and Action. Boston: Harvard Business School Press, pp. 262 - 287.

MORNINGSTAR, Chip and FARMER, Randall (1991) 'The Lessons of Lucasfilm's Habitat' in Michael Benedikt (editor), Cyberspace: First Steps. Cambridge, Mass.: MIT Press, pp. 273 - 301.

NOHRIA, Nitin and ECCLES, Robert (1992) 'Face-to-Face: Making Network Organizations Work' in Nithin Noria and Robert C. Eccles (editors), Networks and Organizations: Structure, Form and Action. Boston: Harvard Business School Press, pp. 288 - 308.

NORBERG, Arthur and O'NEILL, Judy (1996) Transforming Computer Technology: Information Processing at the Pentagon, 1962 - 1986. Baltimore: Johns Hopkins University Press.

O'NEILL, Judy (1995) 'The Role of ARPA in the Development of the ARPANET, 1961-1972', Annals of the History of Computing vol. 17, no. 4, pp. 76 - 81.

PRATT, David et al. (1995) 'NPSNET: Four User Interface Paradigms for Entity Control in a Virtual World', Journal of Intelligent Systems, vol. 5, nos. 2 - 4, pp. 89 - 109.

REEVES, Byron and NASS, Clifford (1996) The Media Equation: How People Treat Computers, Television and New Media Like Real People and Places. Cambridge: Cambridge University Press.

ROEHL, Bernie (1996) 'Shared Worlds', VR News, vol. 5, no. 8, pp. 14 - 19.

ROEHL, Bernie (1997) 'Shared Worlds', VR News, vol. 6, no. 9, pp. 10 - 15.

ROGERS, Everett (1986) Communication Technology: The New Media in Society. New York: Free Press.

SCHROEDER, Ralph (1995) 'Disenchantment and Its Discontents: Weberian Perspectives on Science and Technology', Sociological Review, vol. 43, no. 3, pp. 227 - 50.

SCHROEDER, Ralph. (1996) Possible Worlds: The Social Dynamic of Virtual Reality Technology. Boulder, Co.: Westview Press.

STYTZ, Martin R. (1994) 'An Overview of Current Virtual Reality Research and Development Projects by the United States Department of Defense' in Sandra Helsel (editor), Proceedings of the Fourth Annual Conference on Virtual Reality London: Mecklermedia, pp.152-159.

TURKLE, Sherry (1995) Life on the Screen: Identity in the Age of the Internet. New York: Simon & Schuster.

WELLMAN, Barry et al (1996) 'Computer Networks as Social Networks: Collaborative Work, Telework, and Virtual Community' in Annual Review of Sociology, vol. 22, pp. 213 - 38.

Copyright Sociological Research Online, 1997