Working Draft -- Not for quotation
Final version to appear in:  Marc Smith and Peter Kollock (editors).   1999.  Communities in Cyberspace. London: Routledge


Communities in Cyberspace

Peter Kollock and Marc Smith

University of California, Los Angeles


The sociologists are going to love the next 100 years.

John C. Dvorak (1996)


Since 1993, computer networks have grabbed enormous public attention. The major news and entertainment media have been filled with stories about the "information superhighway" and of the financial and political fortunes to be made on it. Computer sales continue to rise and more and more people are getting connected to "the Net". Computer networks, once an obscure and arcane set of technologies used by a small elite, are now widely used and the subject of political debate, public interest, and popular culture. The "information superhighway" competes with a collection of metaphors that attempt to label and define these technologies. Others, like "cyberspace," "the Net," "online," and "the web," highlight different aspects of network technology and its meaning, role and impact. Whichever term is used, it is clear that computer networks allow people to create a range of new social spaces in which to meet and interact with one another.

Instead of people talking to machines, computer networks are being used to connect people to people (Wellman, et al. 1996). In cyberspace the economies of interaction, communication, and coordination are different than when people meet face-to-face. These shifts make the creation of thousands of spaces to house conversations and exchanges between far-flung groups of people practical and convenient. Using network interaction media like email, chat, and conferencing systems like the Usenet, people have formed thousands of groups to discuss a range of topics, play games, entertain one another, and even work on a range of complex collective projects. These are not only communication media – they are group media, sustaining and supporting many to many interactions (Licklider 1978; Harasim 1993).

What kinds of social spaces do people create with networks? Two opposing visions are popular. One highlights the positive effects of networks and their benefits for democracy and prosperity. A prominent proponent is Al Gore (1993) who captures this vision by saying, "Our new ways of communicating will entertain as well as inform. More importantly, they will educate, promote democracy, and save lives. And in the process they will also create a lot of new jobs. In fact, they're already doing it." The promise is that networks will create new places of assembly that will generate opportunities for employment, political participation, social contact and entertainment. At their best, networks are said to renew community by strengthening the bonds that connect us to the wider social world while simultaneously increasing our power in that world.

An alternative view notes that this glowing vision is in part driven by significant investments in public relations, advertising and political rhetoric. Critics see a darker outcome in which individuals are trapped and ensnared in a "net" that predominantly offers new opportunities for surveillance and social control. For Theodore Roszak, "information technology has the obvious capacity to concentrate political power, to create new forms of social obfuscation and domination" (1986, p. xii). While these critics do not rule out the idea that computers and networks increase the power of individuals, they believe that networks will disproportionately increase the strength of existing concentrations of power.

The chapters in this volume share a common understanding that the kinds of interactions and institutions that are emerging in cyberspace are more complicated than can be captured in one-sided utopian or dystopian terms. These papers do not ask whether online interaction is "good" or "bad." Our focus is on describing and analyzing patterns of online social interaction and organization as they exist.

The Internet is a strategic research site in which to study fundamental social processes. It provides a level of access to the details of social life and a durability of the traces of social interaction that is unprecedented. We use this research site to investigate how social action and organization change as they are refracted through online interaction. How do the economies (taking that term very broadly) of social life shift? What becomes easier to do? What becomes more difficult? And what are the aggregate consequences of these changes? The outcomes are not uniformly positive or negative. The new opportunities and constraints online interaction creates are doubled-edged, leading to results that can amplify both beneficial and noxious social processes.

Technology has its most profound effect when it alters the ways in which people come together and communicate. In this volume, we focus on computer network systems that directly support the interaction of people with other people. Before we turn to a discussion of the chapters in this volume, we review the types of systems discussed here and offer some technical background.



Each online communication system structures interaction in a particular way, in some cases with dramatic effect on the types of social organizations that emerge from people using them. We examine in turn email and discussion lists, Usenet and BBSs, text chat, MUDs, World Wide Web sites, and graphical worlds.


Email and Discussion Lists

Email and discussion lists are the oldest and most popular form of interaction on the Internet. Email allows an individual to send a message directly to another person. However, email is often used to go beyond a one-to-one interaction. In an email discussion list a message sent to a group address is then copied and sent to all the email addresses on a list. When people direct a series of messages and responses to the list, a group discussion can develop. As of 1997, there are tens of millions email users and thousands of public mailing lists as well as hundreds of thousands of less formal discussion lists in existence. These lists are maintained for the discussion and distribution of information on thousands of topics. This may be the most common form of group interaction on the Internet, and a number of lists contain thousands or tens of thousands of members.

Email discussion lists have some important qualities that distinguish them from other Internet communication tools. Email lists are typically owned by a single individual or small group. Since all messages sent to the list must pass through a single point, email lists offer their owners significant control over who can contribute to their group. List owners can personally review all requests to be added to a list, can forbid anyone from contributing to the list if they are not on the list themselves, and even censor specific messages that they do not want broadcast to the list as a whole. Because active review require significant time and effort, most email lists are run as open spaces, allowing anyone to join the list and anyone to contribute to it. Still, even open lists can be selectively closed or controlled by their owners when faced with disruption. Most email lists operate as benign dictatorships sustained by the monopoly power the list owner wields over the boundaries and content of their group. As a result, email lists are often distinguished by their relatively more ordered and focused activity.


Usenet and BBSs

Email discussion lists are asynchronous media. Interaction is structured into turns, but a reply may occur minutes or months after the prior turn. There are a number of benefits to asynchronous interaction. A group can interact without everyone gathering at a particular time. As a result people on very different schedules or in distant time zones can still exchange messages and sustain discussions.

BBSs (Bulletin Board Systems – also known as Conferencing systems) are another form of asynchronous communication that refine email discussion lists in a number of ways. Most BBSs allow participants to create topical groups in which a series of messages, similar to email messages, can be strung together one after another. There are a number of conferencing systems. Well known ones include the Usenet, the WELL (picospan), ECHO (caucus), and the bulletin board discussion groups run on the commercial online services such as America Online and the Microsoft Network. Each sustains a wide collection of topics of discussion and an on-going give-and-take between participants. BBSs differ from email discussion lists in another way. Email is a "push" media – messages are sent to people without them necessarily doing anything. In contrast, conferencing systems are "pull" media, people must select groups and messages they want to read and actively request them.

The Usenet is the largest conferencing system and has a unique form of social organization. The Usenet is composed of a distributed database of messages that is passed through an informal global network of systems that agree to a standard message format. As of 1997, more than tens of thousands of "newsgroups" are carried over the Usenet, each containing from a few dozen to tens of thousands of messages. On an average day tens of thousands of different people contribute hundreds of thousands of messages to the Usenet. A new site "joins" the Usenet simply by finding any existing site that is willing to pass along a copy of the daily "feed" (the collection of messages it receives). As a result, the Usenet has no central authority, no single source of power that can enforce boundaries and police behavior. No one owns most Usenet newsgroups; most newsgroups are anarchic in the technical sense of the term – they have no central authority though they do have an order and structure. Almost anyone can read the contents of a Usenet newsgroup, create entirely new newsgroups, or contribute to one. This makes the Usenet a more interesting and challenging social space than systems that are ruled by central authorities. Whatever order exists in the Usenet is the product of a delicate balance between individual freedom and collective good. Many newsgroups are wild unordered places, but what is startling is how many are well organized and productive.


Text Chat

Text chat differs from email lists and BBSs in that it supports synchronous communication – a number of people can chat in real time by sending lines of text to one another. Chat is one of the most popular forms of interaction on the Internet, and accounts for a sizeable proportion of the revenue of the commercial online providers such as America Online. Text chat is often organized around the idea of channels on a text-based "CB-radio" system. Most chat systems support a great number of "channels" dedicated to a vast array of subjects and interests.

Text chat also uses a centralized server that grants the server owner a great deal of power over access to the system and to individual channels. In the commercial chat services, chat channels frequently are policed by the provider’s staff or by appointed volunteers. In the largest non-commercial system – IRC (Internet Relay Chat) – each channel has an owner who can eject people from the channel, control who enters the channel, and decide how many people can enter.



Email discussion lists and conferencing systems are based on the models of postal mail and bulletin boards. Text chat is based on the model of CB radio. In contrast, MUDs (Multiple User Domains/Dungeons) attempt to model physical places as well as face-to-face interaction. MUDs are text-based virtual realities that maintain a sense of space by linking different "rooms" together. MUDs grew out of interest in Adventure style games that presented a textual description of different rooms and the objects in them and allowed the player to move from room to room, take and drop objects, and do things such as fight dragons and solve puzzles. With the growing availability of networked computers on university campuses in the late 1970’s, MUDs were developed to allow people to play Adventure games with other people instead of against computers.

Over the past fifteen years MUDs have become increasing sophisticated and complex. Modern social MUDs allow users to build new spaces, create objects, and to use powerful programming languages to automate their behavior. While many MUDs continue to focus on combat role-playing, many "social" MUDs have become a means for widely dispersed groups to maintain personal contact. MUDs incorporate a range of other modes of communication like email and discussion groups to link users with other users. But like text chat their key quality is that they support synchronous communication – people interact with each other in real time. MUDs allow a number of people in the same "room" to meet and talk by sending lines of text to one another. MUDs often support simulations of the multi-channel quality and nuances of face-to-face interaction by framing the lines of text users send to one another as "say," "think," or "emote" messages. This allows users to provide meta-commentary on their turns of talk and to create "gestures" or make parenthetical comments.

Like email lists, MUDs are typically owned by the individual or group that provides the hardware and software and the technical skill needed to maintain the system. Because these skills and resources have until very recently been rare, owners of MUD servers have had nearly complete control over the system. MUD owners are often referred to as "Gods." Gods can delegate their power in whole or in part to selected participants, who commonly take on the status of "wizards." Other users can be granted more access to the computer’s memory and network capacity, allowing them to build larger and more elaborate virtual spaces and objects. Users can be granted or denied the right to enter the MUD, be given the power to build new objects or enter specific rooms, and can have limited abilities to communicate with other users. MUDs can contain sophisticated forms of social stratification and elaborate hierarchies.


World Wide Web Sites

While the World Wide Web has been hugely popular for some time, it is only more recently that it has become a site for interaction. In its original incarnation, the Web served as a powerful way of accessing and linking documents. Web sites can now support both asynchronous and synchronous communication. Through the use of various software tools, web sites can host asynchronous discussion groups as well as real-time text chat.

Because of its graphical user interface and the ability to integrate images and sounds, web sites can create a more intuitive and richer context for text chat. As navigating through a web site is a familiar experience for most online users, entering into a discussion can be easier than learning a propriety system on a BBS. In addition, web sites can increase the channels of communication, setting the mood or style of the interaction through layout design, images, and sounds.

Web sites can also serve as a separate supplement to text-based communities. For example, a number of MUDs have elaborate web sites that are used to collect images and documents related to the MUD, as well as links to the personal web pages of its members. While text communication can be a very powerful form of interaction, the fact that MUDs establish web sites suggests that the web’s interface and graphical design provides important benefits.


Graphical Worlds

As computing power and network bandwidth increase, the kinds of media people can use to interact with one another expands to include images, sound, and two or three dimensional models of spaces. Real-time video and audio interaction tools have developed, as have online interaction systems that integrate text chat with a visual representation of each participant (often called "avatars") and some representation of a place. Some of these graphical worlds allow people to engage in a real-time audio conversation, a high-tech return to the low-tech telephone party line.

As these systems become more sophisticated they have taken on some of the characteristics of other older media. For example, WorldsAway ( – a descendant of the earliest graphical interaction system, Habitat – has developed a social structure similar to that found in MUDs. In WorldsAway, some users become "acolytes" who serve as helpers in the community and have a higher status than regular members. For other users a token economy adds another form of social stratification, creating virtual millionaires as well as beggars.

While these new graphical spaces present interesting research possibilities, the case studies in this volume all concentrate on text-based social systems. This was done because text-based systems have been in existence much longer than graphical systems. Hence, social interaction and groups in text-based systems have a longer history, are often more developed and elaborated, and there is a greater variety of textual online social spaces to compare and contrast. Most of the lessons drawn in these case studies of text-based interaction are also applicable to the emerging graphical worlds.



Each of the sections in this volume explore the implications of online interaction in terms of a key concept, and each section builds on the previous one. The chapters are organized into four major groups. We treat in turn issues of (1) identity, (2) social order and control, (3) community structure and dynamics, and (4) collective action.



We begin with a consideration of identity, the basic building block of social interaction. All of our interactions, even those with strangers, are shaped by our sense of with whom we are interacting. In face-to-face and telephone interactions there are a wealth of cues of varying reliability to indicate our identity and our intentions. Our clothes, voices, bodies, and gestures signal messages about status, power, and group membership. We rely on our ability to recognize fellow group members in order to know who we can turn to and what we can expect. Our ability to identify others also allows us to hold individuals accountable for their actions.

Online interaction strips away many of the cues and signs that are part of face-to-face interaction. This poverty of signals is both a limitation and a resource, making certain kinds of interaction more difficult but also providing room to play with one’s identity. The resulting ambiguity over identity has been a source of inspiration to many who believe that because people’s physical appearance is not manifest online (yet), individuals will be judged by the merit of their ideas, rather than by their gender, race, class, or age. But others (including authors in this volume) argue that traditional status hierarchies and inequalities are reproduced in online interaction and perhaps are even magnified. In this section we examine how identity is established online as well as the durability of the institutions of race and gender in online interaction.


Honesty and Deception

How is identity – true or counterfeit – established in online communities? This question is at the heart of Judith Donath’s chapter on Identity and Deception in the Virtual Community. Deception can bring great benefits, but deception is useful only if some members of a community honestly signal what sort of person they are. A signal that is used deceptively by an entire population carries no information. Donath thus asks what maintains the balance between honest and deceptive signaling, and makes the point that some signals are inherently difficult to mimic.

A signal is inherently reliable to the extent that it is more costly for mimics to emit than it is for those who actually possess the quality. A classic example is conspicuous consumption as a signal of wealth. A person of limited financial means will find it very difficult to display the chauffeur driven-limousine, mansion, and other signals of wealth. To take an example from the Internet, one’s proficiency as a programmer can be signaled by included an example of one’s work in a Usenet posting. Indeed, there is a convention in the Usenet newsgroup alt.hackers of including a clever hack in each post to establish one's identity as a skilled hacker and to keep out impostors. Signals that are to some degree inherently reliable are called assessment signals. This is in contrast to conventional signals which have no necessary connection to the trait they advertise (e.g., simply stating one is wealthy or a skilled programmer).

With these issues as a background, Donath then goes on to dissect the anatomy of a Usenet post in order to examine how identity is established or concealed. The first source of clues is the account name itself. Although a seemingly sparse source of information, Donath demonstrates how much can be squeezed out of this basic ID. Knowledgeable members of the Internet realize that each domain name has its own reputation and recognize the different implications of commercial versus institutional accounts. The content of the post contains its own set of signals about the identity of the author. The writing style, the facts that are brought forth, the proper use of abbreviations and argot that are specific to the group, all help establish or challenge the user’s identity. At the end of each post is a signature which has also become an important element in establishing an identity. Signatures are a combination of business cards and bumper stickers that members use to display their interests, opinions, and occupation. By providing the address of one’s company and position, the signature can help establish one’s credentials and create the possibility of accountability (one can check the company to verify the information). Donath also provides a number of fascinating examples of the use of codes or esoteric knowledge as a way of establishing one’s identity. Another common element that is now seen in signatures is the address to the author’s web page. The simple link to a web page is significant in a number of ways. By providing a link to a detailed document, the author is able to establish his or her identity in a very elaborate way. The fact that an elaborate set of web pages represent a substantial investment of time and effort may also have the effect of encouraging identity persistence – throwing away an alias used in previous postings may be a trivial act, but dismantling a set of web pages and constructing a new set may involve costs that the author is not willing to bear.

There is yet another source of identity information: the author’s history of previous posts on the Internet. Interaction on the Net leaves a trace of varying durability. Posts to Usenet discussion groups can usually be accessed for a few weeks and some groups maintain archives of past discussion. The search for the traces of someone’s past interactions has changed radically in recent years with the development of extremely powerful search engines and durable, extensive databases.

Donath raises two extremely important issues at the end of her chapter: To what extent can online interaction and communities be structured so as to encourage honest signaling and identity persistence? And to what extent is a known, stable identity a desirable thing for the person and for the community?


The Persistence of Race

An early promise of online interaction was that it would render irrelevant such markers as race, gender, status, and age. Because online interaction strips away physical markers, the assumption was that social categories assumed to rest on physical characteristics would wither away. The next two chapters argue against this view, taking as their focus race and gender respectively.

As Byron Burkhalter points out in Reading Race Online: Discovering Racial Identity in Usenet Discussions, the connection between racial identity and the physical body has been so strong that it is simply taken for granted in most settings. Yet racial identity can be an ambiguous thing even in face-to-face interaction. Indeed, Burkhalter argues provocatively that race does not disappear online and that "racial identity is no more ambiguous online than offline. … Certainty of racial identity offline or online is always contingent – absolute proof is not available and rarely necessary."

At the same time, Burkhalter believes that online interaction does change the dynamics of racial identity, shifting how racial identification is achieved and how stereotypes operate. In online interaction racial identity springs from a participant’s perspective on racial issues rather than from physical cues. The reliance on a participant’s written words as a source of racial identity also reverses the usual sequence in stereotyping. Burkhalter argues that in online interaction, an individual’s beliefs and attitudes are used to make inferences about the individual’s race, rather than the more familiar route of inferring attitudes based on physical racial cues.

Burkhalter goes on to examine how racial identity is achieved, maintained, questioned, and reestablished in Usenet discussion groups. Using the soc.culture.african.american discussion group as his case study, he examines how any topic can potentially be framed as race-relevant and how this link to racial identity can ebb and flow during the course of a discussion.

While any racial identity might be claimed in this textual world, this does not guarantee that others in the group will acknowledge or support the claimed identity. Racial identity, Burkhalter argues, is interactionally negotiated – people do not create racial identities by themselves and the identity portrayed in a posting might be accepted or vigorously contested by other members in the group. Sometimes a group member will attempt to cast the poster in another identity as a way of challenging the poster’s arguments. Burkhalter reports that this is a common method of challenging a person, and that arguments often center around what identity a poster can reasonably claim rather than directly challenging a poster’s arguments and views.

The structure of the Usenet also makes it a useful research site for studying the dynamics of group boundaries. Because messages can easily be crossposted to other groups, the audience for a discussion can rapidly expand or contract, drawing in other groups with different identities at stake. Burkhalter points out that a discussion that remains within soc.culture.african.american can make use of fine-grained racial distinctions that are a meaningful framework for members within the group. However, once another newsgroup becomes involved in the discussion, such distinctions tend to collapse as each group orients itself in a monolithic manner toward the other group. Within-group variation is attenuated while across-group differences are accentuated.


The Conservatism of Cyberspace

The most optimistic proponents of the Internet have argued that gender, race, and age become unimportant in online interaction. At the very least, many assume that the absence of these markers will provide the opportunity to explore and invent alternate identities.

Jodi O’Brien sharply contests this view in her chapter on Writing in the Body:Gender (Re)Production in Online Interaction. O’Brien argues that gender is such a central feature for organizing interpersonal relations that persons go to great pains to reproduce gender in online interaction. "Are you male or female?" is such a commonly asked question that it was long ago abbreviated to "RUMorF?" Significantly, no such abbreviations are in widespread use for questions concerning age, height, weight, socio-economic status, etc. Gender is the one characteristic of our embodied lives that is a central feature in interaction throughout the Internet.

And it is not simply that gender is reintroduced in a world without physical markers. O’Brien makes the striking point that gender is reintroduced in a more limited and stereotypical manner than exists in embodied interaction. There are no limitations to how one might describe oneself in cyberspace. Yet the gender descriptions one encounters on the Internet show far less variation and imagination than occurs in face-to-face interaction. People recreate themselves as stereotypical ideals, and O’Brien points out that this "hyper-gendering" is especially prevalent among those who attempt to "cross-dress," (i.e. males presenting themselves as females). The implication is that a world without constraints has led to greater homogeneity rather than new forms of identity.

The attention given to the possibility that someone might be gender-switching points out the crucial importance of gender identity in cyberspace. The fear is that a close friend or the object of one’s seduction is not "really" the gender they present themselves as. The debate as to whether and when gender-switching is appropriate has led to a vocabulary of motives (Mills 1940) that makes crossing acceptable as a way of avoiding harassment or experimenting with a new perspective, but inappropriate if there is an intent to "deceive." This is hardly a clear line, and one person’s experimentation is another person’s deceit.

There is a deeper tension. O’Brien points out that there is strain between those who view online interaction as an opportunity to "perform" a variety of perhaps fabricated roles versus those who see cyberspace as a new communication medium between "real people." O’Brien argues that the distinction between the intent to "be" and the intent to "perform" may be much more useful than discussions about what is real versus non-real, or honest versus deceitful. She concludes that "the most contested issue at the moment in the history of online communication may be how to establish ways of underscoring real versus fictitious sites so that users can reliably distinguish ‘real authenticity’ from ‘authentic fantasy’." Thus, the issue is clearly signaling the frame of the interaction so that individuals know whether this is a site having more to do with theater (and therefore authentic and consistent fabrication, as is true in many role-playing MUDs), versus a site where identity is tied is some reliable way to one’s "real-world" embodied self (as would be necessary for legal and financial transactions).



Social organizations vary in terms of the amount of control, coordination, and ultimately the coercion they can exercise over themselves. Social control rests in large part on a group’s ability to identify individuals in order to hold them responsible. Building on identity, we look at the nascent institutions of self regulation and governance in online groups. We examine the ways in which cyberspace enhances or erodes the methods of social order and control that already exist in human societies.

In this section we examine a range of power structures that have emerged in cyberspace. It is widely believed and hoped that the ease of communicating and interacting online will lead to a flourishing of democratic institutions, heralding a new and vital arena of public discourse. But to date, most online groups have the structure of either an anarchy or a dictatorship. Some notable experiments with democratic electoral politics have failed dramatically, raising questions about what sort of governance is possible and what the prerequisites are for democratic institutions. These chapters highlight the fact that cyberspace is often a domain of vast power imbalances. Several different approaches to social order are examined, including public punishment, ostracism, and mediation.


The First Age of Social Control

Despite the utopian hopes of early settlers, it became clear very quickly that some form of monitoring and sanctioning would be necessary in online groups. "The failure of the ideal of complete freedom in cyberspace was an early phenomenon." So begins the chapter by Elizabeth Reid on Hierarchy and Power: Social Control in Cyberspace.

Reid examines the dynamics of power and methods of social control in two kinds of MUDs: Adventure MUDs and Social MUDs. In Adventure MUDs, users participate in role-playing games which involve scoring and strict hierarchies. The structure of the games requires a serious commitment to the community for the reason that one must play the game frequently in order to collect the resources necessary to keep one’s character "alive." Social MUDs, in contrast, involve more free-form interaction and the ability to build new objects in the world. It is possible to participate in Social MUDs much less frequently than in Adventure MUDs because one’s character continues to exist over time without the need to explicitly find food, shelter, etc..., as is the case in the role-playing Adventure MUDs.

In common with many other forms of online interaction, the relative anonymity and lack of physical contact that is a feature of MUDs encourages users to become less inhibited. This is not to say that members of MUDs exist in an anomic state. Reid rightly points out that there is "no moment on a MUD in which users are not enmeshed within a web of social rules and expectations."

The disinhibition that is common in MUDs can encourage the formation of intimate relationships and deep feelings of attachment to other users. The other edge of this dynamic, however, is that members of MUDs can act out in a hostile and violent way without the usual fears of retribution. While it is true that threats of physical violence in online interaction are almost always empty threats, there is the possibility to do great symbolic violence. Tales of online harassment and "rape" are already well known, and Reid adds a particularly striking example of virtual violence involving a MUD dedicated to the support of persons who have experience sexual assault.

Reid uses the community’s response to this crisis as well as examples from other social and adventure MUDs to describe a variety of methods for social control. Some forms of social control require the intervention of the owner of the MUD or one of the high ranking "wizards" who can control and rewrite the software that underlies the MUD. Other methods come from the members themselves, acting either singly or in concert with others. Among the possible forms of social control that are brought up in Reid’s discussion (and in the subsequent chapter by A. Smith) are the following:

Eliminating commands in the system that chronically lead to objectionable behavior. An example is the decision by the owner of the MUD mentioned above to eliminate the "shout" command in that community.

Providing commands that allow members to filter out objectionable behavior. "Gag" commands, which are available in many communities, allow users to ignore what a particular person is saying and doing. However, as Reid discusses, there are important limitations to this technique. Some communities also allow its members to intentionally select what actions to which they wish to expose themselves. In some adventure MUDs, one can choose to open oneself up to the risk of having one’s character killed, or alternately to prohibit this kind of attack on oneself.

Temporarily restricting the rights of transgressors. A member might have their movements restricted or lose the right to use certain commands for a period of time.

Shaming and humiliating a transgressor in public rituals. The owner or wizard of a MUD can force a transgressor into a public space, strip away their abilities and perhaps allow other members to shame and ridicule the person.

Banishing a transgressor from the community. A wizard can also force a transgressor out of a community, although the ease with which one can acquire another account and identity on the Internet means that a determined person might find a way back into the community.

Instituting stricter criteria for admission to the community. To try to avoid problems to begin with, some communities restrict membership, imposing a limit on the total number of users, for example, or requiring that new members be sponsored by existing members. Other communities prohibit guest visitors or severely restrict their abilities.

Increasing accountability by registering identities. Some communities require that members have one, persistent identity as a way of increasing accountability. Other communities also require members to register their legal names and phone numbers with the head of the community.

Regulatory committees. Some MUDs have experimented with regulatory boards composed of members to oversee, for example, the construction projects of the community’s members

Moral Entrepreneurs and Vigilante Groups. Of course, individual members can take it upon themselves to express their dissatisfaction with the behavior of another. In some cases, groups of members can even band together in order to try to right some perceived wrong. In some adventure MUDs, for example, posses are formed to hunt down and punish users who have killed other users. There have also been incidents in which groups have brought in a "hired gun" to punish another member.

Reid makes the provocative point that the emphasis in many MUDs on public punishment, humiliation, and ostracism marks a return to the medieval in terms of the technology of punishment. The body of the users – albeit a virtual body – is once again the site for punishment.


Institutionalized Mediation

If Reid documents the return to medieval forms of punishment in online communities, Anna DuVal Smith discusses a very different source of social control. In her chapter on Problems of Conflict Management in Virtual Communities, she documents the emergence of institutionalized mediation in one community.

Online communities, Smith observes, are distinguished from many face-to-face communities by their open boundaries (users can enter and leave the MUD with much lower costs than physically moving to a new community), the relative anonymity of computer-mediated interaction, and the possibility of great social diversity (members may come from many different countries or ethnic groups, and may have very different expectations about what the goals of the community are). Each of these features can encourage conflict. Smith argues this means that methods of conflict resolution may be both more important and more of a challenge than in many face-to-face settings.

Research in dispute resolution has identified three basic routes for conflict resolution: exercising power, reconciling interests, and adjudicating rights. Many of the examples in the previous chapter by Reid fall into the first category – wizards in a MUD will use their control over the software that is the infrastructure of the community in order to force a member to be shamed, humiliated, or banished. But a resort to raw power has many limitations, both in terms of the technical capacity to carry out decrees and the threat that is posed to the legitimacy of the rulers because of their capricious use of power. Another response is adjudication, though this carries its own costs and requires the establishment of a system of rights and judicial processes.

Between the raw exercise of power and a formal judicial system lies mediation. Drawing on her experience in dispute resolution in face-to-face communities, Smith introduced mediation and fact-finding procedures into the MUD in which she was a member. The catalyst for these reforms was an episode in which the wizards of this MUD intervened to punish a member by using the technical powers at their command to banish the member and destroy the objects he had created in the MUD. Their response was widely seen as inappropriate and illegitimate by many members and led to a series of reforms aimed at trying to better manage conflict. The community decided to (1) severely limit the abilities of visitors to the community, (2) to require sponsorship by two mentors and (3) a period of socialization prior to full membership, and (4) to provide a dispute resolution process (which was carried out by Smith).

Among the important lessons she reports is the double-edged effects of online interaction on the conflict resolution process. On the one hand, the mediator’s role was made more difficult by the fact that users connecting from different time zones made it difficult to schedule meetings. Lag in the network, and the fact that some members had to deal with much more severe lag than others, also meant that it was very difficult for some users to participate. There was also the problem that evidence (such as email logs) could be falsified. On the other hand, the same disinhibition that may have provoked the conflict to begin with also encouraged members to be forthcoming with facts and details about an episode. The absence of physical cues also had some positive consequences, making it easier, for example, to overcome the barriers of socio-economic differences. Even network lag had a positive side in that it provided time for thought and reflection before reacting.

Smith concludes that the conflict resolution methods that were introduced to the community were useful. The Advisory Board of the community agreed and institutionalized the mediation procedures. Nevertheless, she cautions that there are significant challenges to establishing and maintaining a system of mediation. Members must be aware that a mediation procedure exists and believe it is a fair and legitimate process. A second challenge is that cases can take quite a long time to resolve. Finally, there is the simple fact that members of the community who are both skilled and willing to invest substantial time and effort must staff the process. A system for resolving disputes is a public good, and there is the temptation to free ride on the efforts of others.



Cyberspace is already the home of thousands of groups of people who meet to share information, discuss mutual interests, play games, and carry out business. Some of these groups are both large and well developed, but critics argue that these groups do not constitute real communities. Something is missing, they argue, that makes these online groups pale substitutes for more traditional face-to-face communities. Others respond that not only are online communities real communities, they also have the potential to support face-to-face communities and help hold local communities together. Opinion rather than analysis and evidence characterizes much of this debate, and surprising little is know about the actual structure and dynamics of online groups.

In this section we consider questions of community structure and processes, beginning with a detailed comparison of online and "real-life" communities. We then examine the structure of online groups, and the surprising amount of cooperation in cyberspace.


Online Communities as "Real" Communities

Online communities surely differ in important ways from face-to-face communities – that is a premise of this volume. But when contrasts between online and face-to-face communities are made, especially if these contrast have a moral character, it is imperative that we be clear what the comparison point is. When, for example, critics describe online communities as more isolated than "real-life" groups, their comparison seems to be to an ideal of community rather than to face-to-face communities as they are actually lived. There is a great deal of loneliness in the lives of many city dwellers.

What is needed is a detailed comparison of online and "real-life" communities by someone who has studied both. Barry Wellman and Milena Gulia provide this in their chapter, Virtual Communities as Communities: Net Surfers Don’t Ride Alone. In asking what kinds of communities exist online, they chastise many participants of this debate on several grounds. Most commentators have fallen on one extreme end of opinion or the other, painting pictures of unrelenting hope or despair. Critics and enthusiasts alike also speak as if they were unaware that there is a long history of thinking and research on the topic of community. Finally, Wellman and Gulia make the point that most analysts of online interaction and community treats the subject as if it had no connection to the other facets of a person’s life.

The authors begin their analysis by pointing out that even before research on online groups had begun, researchers on community had gone through a very important shift. Community is now conceptualized not in terms of physical proximity but in terms of social networks. Telephones, automobiles, and airplanes have long meant that it was possible to establish and sustain important social relationships outside of one’s immediate physical neighborhood.

Wellman and Gulia frame their discussion around several key questions: (1) Are online relationships narrowly specialized or broadly supportive? (2) In what ways are the many weak, less intimate relationships on the Net useful? (3) Is there reciprocity online and attachment to online communities? (4) Are strong, intimate relationships possible online? (5) How does online community affect "real-life" community? (6) Does the Net increase community diversity? (7) Are online communities "real" communities? Their conclusion is that online communities meet any reasonable definition of community, that they are not a pale, artificial substitute for more traditional forms of community. At the same time, there are many distinguishing features of online communities that change the economies of social action and organization in important ways.


Visualizing Social Cyberspaces

In William Gibson's now famous vision, cyberspace is "A consensual hallucination experienced daily by billions of legitimate operators, in every nation … A graphic representation of data abstracted from the banks of every computer in the human system. Unthinkable complexity. Lines of light ranged in the nonspace of the mind, clusters and constellations of data. Like city lights, receding." (Gibson 1984, p. 51)

In practice, however, the landscape of the Net has been difficult to see. There is no single unified representation of cyberspace or even of its major components. The tools used to connect to social cyberspaces leave us blind to a range of information that is otherwise visible in face-to-face interaction – information such as who is in the room, where they are located and clustered together, what the room is like and what expectations for behavior it signals. In online interaction the experience is something like attending a cocktail party and only being able to see people who are actively speaking, while the room and all the listeners are invisible.

But what if we could see the lanscape of specific parts of the Net? In Invisible Crowds in Cyberspace: Mapping the Social Structure of the Usenet, Marc Smith explores a method for generating maps of the social networks created in the Usenet. Smith proposes a research agenda in which long term historical studies of large-scale social network can be carried out. His goal is to create a series of maps showing what different regions of social activity look like.

Smith discusses a range of measurements that can be generated from a collection of data from the Usenet news feed. Relationships between groups – in particular through a practice called crossposting that links groups together through shared messages – can be visualized as networks and then contrasted with other groups. The Usenet is examined in terms of the distribution of activity and diversity of participants. Smith proposes to use these measures to develop a classification system for online groups. What his research makes clear is that cyberspace is not homogeneous. The Usenet newsgroups he studies are as varied as any set of face-to-face social groups and spaces.

Smith highlights the range of methodological and ethical issues raised by research conducted in this manner. The maps created by this project reveal in great detail a social space that had previously been cloaked by the unwieldy nature of the data. Many people have a high expectation of privacy in computer-mediated interaction despite the fact that they interact in what can be seen as a public space. But new technical tools change and shift the nature of this space. Smith explores the implications of these tools and data for researchers interested in developing further studies based on records of interactions in cyberspace. He concludes that while online records are remarkable sources of data, they are also ambiguous and limited and require that researchers to handle them with care.


The Possibilities and Limits of Online Cooperation

Anyone who has spent time online has probably experienced aggressive, flaming responses from someone. The hostility that exists on the Internet is a common and oft-repeated complaint. But Peter Kollock, in his chapter on The Economies of Online Cooperation: Gifts and Public Goods in Cyberspace, argues that the puzzle to be explained is not the prevalence of hostility in online interaction, but rather how it is that there is any significant cooperation at all.

In fact, it is common to find individuals who are remarkably generous with their time and expertise. Howard Rheingold has even described the interactions within one online community as a gift economy. Kollock takes this metaphor as his starting point and examines the structure of gift-giving and whether that is an appropriate model for online interaction. He argues that much online interaction is characterized by a form of exchange that is both more generous and riskier than gift-giving. Many of the favors and benefits provided in online communities are public goods, i.e., a good from which all may benefit, regardless of whether one has helped create the good. In such a situation there is a temptation to free-ride on the efforts of others, but if all do so, everyone is worse off.

Kollock argues that the economies involved in producing many public goods change radically as one moves to an online environment. The costs of communicating and coordinating the actions of a group, for example, are often much lower than in face-to-face interaction. And the value of a piece of information or advice that is offered to a group can be amplified because of the fact that this is a realm of digital information in which an unlimited number of people might use or make copies of the information provided. It is also the case that the size of the group necessary to produce many public goods is often reduced to one. This is to say that single individuals are capable of producing and distributing valuable goods or services to a huge audience on the Internet. As he states: "Shifts in the economies of production mean that individuals are able to produce many public goods on their own. And the decrease in contribution and coordination costs as well as the potential amplification in the value of the contribution (because of the huge audience) makes it more likely that an individual will experience a net benefit from providing the good."

He goes on to discuss the various motivations that might encourage individuals to produce public goods. Kollock then examines two provocative cases studies: the production of a new computer operating system, know as Linux, through the use of voluntary labor, and the 1996 effort to wire California’s elementary schools for Internet access. The success of these two projects illustrates the potential of online interaction in facilitating the production of public goods that would otherwise be much more difficult or even impossible to produce.

However, this is not to say that online cooperation is inevitable or expanding. Kollock ends his chapter with a caution on the limits of online cooperation, arguing that while shifting costs make the provision of some public goods much more likely, there are also a number of requisites for online cooperation that limit the extent to which many other kinds of goods will be produced. Kollock also make the point that successful collective action will not always be in the interest of the larger society – the changing economies of online cooperation benefit all groups regardless of their aims and goals.



Communities rarely exist exclusively in cyberspace. It is important to investigate the ways in which social groups in cyberspace spill out into the "real" world and vice versa. Can the social relations created or supported in cyberspace alter the fabric of our physical communities? We look at the concrete uses of networks to build or enhance communities and the ways online networks can be used for collective action. Social protest, the linking together of rural communities, and collective action by a disadvantaged community are examined.


The Dynamics of Online Social Protest

The effectiveness of the Internet as a tool for coordination and communication can be seen in the success of social protests that were carried out online. Particularly striking was the critical response to MarketPlace – a CD-ROM that was to contain information about 120 million American consumers – and the Clipper chip, an encryption technology that provided a "backdoor" to be used by the government as a form of wiretapping.

These are the two case studies that Laura Gurak analyzes in her chapter on The Promise and Peril of Social Action in Cyberspace: Ethos, Delivery, and the Protests over MarketPlace and the Clipper Chip. Her interest is in examining how online interaction either enhances or complicates traditional rhetorical activities such as speeches and debates. In particular, she focuses on two rhetorical features: ethos and delivery.

Ethos is classically used to refer to the character or tone of a speaker. Gurak makes the case that ethos can also be a group quality, referring to the cultural and moral tone of a community. This is the sense in which she uses the term in her examination of social protest on the Internet. The second rhetorical feature she examines – delivery – traditionally refers to a speaker’s gestures and expressions. However, the delivery of messages online differs in many significant ways from face-to-face communication. Gurak discusses a number of features of online communication, including the speed, reach, and durability of messages.

Picking up on a theme woven throughout this volume, the effects of online interaction on rhetorical activity are neither simple nor unitary; rather the outcome is double-edged. On the one hand, online interaction permits individuals to mobilize far more quickly, cheaply, and efficiently than is usually possible. Relevant information can be posted publicly or forwarded to particular individuals or groups at little or no cost. Significantly, it is also relatively easy for individuals to find other individuals with similar concerns and interests. In both the cases Gurak examined, specialized discussion groups quickly formed and served as a locus for debate and the exchange of information. Gurak argues that the protests were effective not simply because of these reductions in time and costs, but also because of a strong community ethos that focused participants. The fact that specialized discussion groups brought people of similar interests together meant that participants were likely to share many assumptions and concerns, which would facilitate communication within the group.

Yet these same features can also be problematic. The speed and range of communication can encourage the spread of inaccurate information, and the common ethos shared by many of the participants can discourage challenges to the information and conclusions of the group. The specialized nature of the discussion groups can become too insular and force out dissenting voices. Gurak documents a number of examples in which inaccurate information was widely spread without being seriously challenged or questioned by the participating protesters.

Thus, her conclusion sounds more than one note. While Gurak applauds the ability of online interaction to "provide space for many more voices," she also cautions that "speed may supersede accuracy and that the beliefs of the community may preside over the responsibility of citizens to make informed decisions."


Bringing Rural Communities Online

While space can be rendered nearly irrelevant by online interaction, there have been many online networks that have been organized around specific physical communities. Willard Uncapher discusses one of the earliest community networks in his chapter on Electronic Homesteading on the Rural Frontier. In the late 1980’s a computer network was set up to serve the needs of an entire physical community. The location was Montana, which is a huge and sparsely populated state. Social and business communication is often difficult and expensive across such a vast place and the hope was that a computer network would make it easier for people to share resources and socialize. This case study is valuable both for its historical significance and because it illuminates many of the issues surrounding the adoption of new technology, as well as how physical and online communities mutually affect each other.

The community network, called Big Sky Telegraph, was the idea of Frank Odasz. While his hope was that the network would serve many needs and constituencies in Montana, he initially started with the idea of linking teachers together online. Montana still has a number of "one-room" schools run by single teachers, and it was felt there was a real need and desire for isolated teachers across the state to come together and help each other. Odasz also believed strongly that technology should be deployed to serve clear, demonstrated needs, rather than being implemented for its own sake.

Odasz approached Dave Hughes, a well-know computer network activist, to help set up the network, and in January 1988 Big Sky Telegraph went online. There was immediate interest in the network and teachers learned about Big Sky Telegraph through the mail and through the face-to-face conferences teachers had twice a year in Montana. The network made use of outdated but still serviceable Apple II/e computers that many schools had. Once teachers connected to Big Sky Telegraph, they could take a variety of classes through the network to improve their online skills.

The network had a number of important benefits for teachers For example, people made themselves available online to answer questions, lesson plans were shared, and the circulation of library materials was organized online. New areas on Big Sky Telegraph were also set up for other groups and activities including business, Internet access, tourist information, and community services.

As Uncapher reports, despite the many benefits of the network, the adoption rate stalled after a time. Only 30 out of 114 one-room schools were active on the network two years into its existence. This leads to the crucial question of what was keeping teachers offline. While it was true that money, modems, and the information to make it all work were sometimes in short supply, Uncapher asserts that the factors driving adoption rates were more complex than simple structural constraints.

The ascension of the Internet has also created questions and tensions about the future of Big Sky Telegraph itself. Moving to a model based on the World Wide Web, for example, risked leaving behind those users with outdated equipment. Big Sky Telegraph is attempting to adopt to these changes, and whatever the outcome, it remains an important model for community based computer networks. Uncapher lists the key lessons Frank Odasz learned in running the system, and they serve as valuable principles for other community networks. Uncapher also makes the point that although technology continues to advance the simple, text-only system used for Big Sky Telegraph is a useful model for those areas of the world where the most current technology remains out of reach.


Online Interaction and Social Change

We end this volume with an optimistic story. In Cyberspace and Disadvantaged Communities: The Internet as a Tool for Collective Action, Christopher Mele illustrates how online networks can be used as a tool for collective action and empowerment. His case study is Robert S. Jervay Place, a low-income public housing development in Wilmington, North Carolina.

As part of a major downtown redevelopment program, the city of Wilmington wished to renovate Jervay Place. This brought the residents, all African-American women, into contact with the local housing authority. Residents were encouraged to take part in the planning process as decisions were made about how to modernize and rebuild the public housing project. Although the residents had some early success in working with the housing authority, the relationship soon fell apart.

The residents and the local housing authority jockeyed back and forth, and in the end a Jervay Place Task Force was created to plan and implement the reconstruction of Jervay Place. While the Task Force provided a formal means by which the residents could make their views known, they felt they could not rely on the housing authority to provide them with the information and technical knowledge they needed to participate effectively. If the residents’ organization could not bring to the Task Force detailed plans, the housing authority would have an easy time pressing forward with its own development ideas. To make matters worse, there was no local organization in Wilmington that could provide the residents the kind of specialized information and assistance they needed.

This encouraged the residents to strike out in another direction. Some Jervay residents had become familiar with personal computers through courses at a community college. With this background, the resident leaders began to search the Internet, at first through public terminals at a university library, and then through their own computer, purchased with a public service grant from a local university.

Their first action was to search for discussion groups devoted to urban planning and architecture. They sent out a plea for help to three such groups and within two weeks the residents received responses from 23 individuals and organizations, including architects and lawyers specializing in low-income housing. Working back and forth with three architectural firms, the residents were able to design a housing community that they felt would best meet their needs. As a result of this collaboration, the residents’ organization was able to dominate the discussion at the Task Force meetings, making use of their sophisticated and carefully prepared plans and graphics.

The residents also used the Internet in another important way by establishing a web site devoted to Jervay Place. The web site allowed the residents to publicly post status reports on the redevelopment project. They also included historical and cultural information about Jervay Place as well as providing a set of links to other relevant sites on the World Wide Web.

Thus, we have a story that seems to confirm the optimists’ hope that the Internet will bring new power and reach to traditionally disadvantaged communities. While these possibilities are exciting, Mele points out that the implications for social change are unclear. The success of the Jervay residents rested on their access to knowledge, technology, and financial resources. The ability to export the organizational lessons of Jervay Place will therefore be limited to the extent that other communities lack these resources.

Nevertheless, the Jervay residents have at least demonstrated the potential of online networks to decrease the costs of communicating and organizing. As one resident said, "On that machine in the center, there are people who are listening."



The lessons from these studies are many. Online interaction creates new forms of deceit and new ways to establish identities. And despite the new freedoms of online interaction, old institutions and stereotypes are reproduced, sometimes in exaggerated forms.

Social control is a necessary component of online communities. The forms examined here – the exercise of power and reconciling interests via mediation – are all still evolving, and the latter form is still in its infancy. As the number of people interacting online continues to increase, and as the interactions online become more important, developing forms of mediation, and eventually adjudication will be a necessary though difficult task.

One can find online groups that meet any reasonable definition of community, but this is not to say that online and face-to-face communities are identical. The economies of cooperation and collective action, as one example, shift significantly as one moves to online interaction. We also need better tools and models for studying online communities, and we see here new tools for mapping social structure and new models for exploring the ecology of online groups.

We also see the interplay between online communities and the "real" world. The Internet has been used as an extremely effective tool for carrying out social protest, but many of the same features that make the Net effective for coordination and communication also encourage the spread of inaccurate information and force out dissenting voices. Online networks are used to link dispersed rural communities, but the effort also brings up complex issues about the adoption of new technology. The Internet is used as a tool for change in a disadvantaged community, demonstrating a way to route around unhelpful government agencies to find people with information and expertise that are willing to share. But the ability to take these lessons to other disadvantaged communities may be limited by the lack of equipment and the information to make it work.

We return again to the double-edge of online interaction. These chapters avoid the extremes of utopian and dystopian visions to examine the details and sometimes conflicting processes within online communities. Assessing the meaning and impact of new technology is always a challenge. Many predictions about the ways new technologies will transform society fade quickly – the telegraph, radio, movies, and television did create revolutions, but not the ones that were expected. Hence, it is especially important that we turn from opinions and predictions to the serious analysis and description of online groups.



Blythe, James and Cathleen McGrath "Web Linkages as Social Networks" Paper presented at Sunbelt 1997 International Network for Social Network Analysis Annual Meeting.

Dvorak, John C. 1996. "Understanding Bizarre and Kinky Web Sites." PC Magazine Online 1 July:

Gibson, William, Neuromancer, 1984. New York: Ace.

Gore, Al. 1993. Speech at the Superhighway Summit Royce Hall, 11 January 1993, UCLA, Los Angeles, California (

Harrasim, Linda M. 1993. "Networlds: Networks as Social Spaces." In Global Networks: Computers and International Communication, ed. L. M. Harrasim. Cambridge, MA: MIT Press, pp. 15-34.

Licklider, J.C.R. Robert Taylor, and E. Herbert, "The Computer as a Communication Device," International Science and Technology, April 1978.

Mills, C. Wright. 1940. "Situated Actions and Vocabularies of Motives." American Sociological Review 5:904-913.

Roszak, Theodore. (1986) The Cult of Information: The Folklore of Computers and the True Art of Thinking. New York, Pantheon Books. (p. xii).