Problems of Conflict Management in Virtual Communities

Anna DuVal Smith

Prepublication draft 3.0
Prepared to appear in Peter Kollock and Marc Smith (eds),
Communities in Cyberspace
Routledge Press, 1998.
Copyright 1997 by Anna DuVal Smith. All rights reserved.
Comments to the author at adsmith@advs.net are welcome and invited.

Abstract

This paper explores the sources of conflict and techniques of social control in an open-access, text-based virtual community. It argues that such social systems have the same kinds of opportunities and problems brought by diversity that real communities do, but that unique features of cyberspace make effective conflict management both more important and more difficult. Cases of interpersonal disputes collected during more than two years of participant observation revealed that power strategies of social control were generally counterproductive in managing the conflict that resulted from the multiplicity of values, goals, interests and cultural norms brought by members of the community. As in real life, methods that reconcile divergent interests (mediation and factfinding) and adjudicate rights (factfinding and arbitration) appeared to manage issue-based conflicts more effectively. However, their utility and, therefore, the community's ability to adapt and thrive as an open, goal-directed system depends on member awareness of the program, human resource availability and administration willingness to share power.

Table of Contents

Introduction

bulletTheoretical Background
bulletThe Setting and Data

The Problem

bulletBackground
bulletUser-Administrator Disputes
bulletDisputes among Users
bulletDiscussion

The Reforms

bulletBoundary-Defining Reforms
bulletDispute Resolution Tools
bulletThe Intervention

The Results

bulletA Dispute Managed by Fact-Finding
bulletA Mediated Dispute
bulletDiscussion

Evaluation
Conclusion and Summary
Notes
References

Introduction

In September 1993, tiring of the harassment my female identity received on IRC1, I read of MicroMUSE, a MU*2 reputed to be "nonviolent, noncompetitive and collaborative" because of its mission as an educational tool for children (Leslie, 1993, p.34; Kelly & Rheingold, 1993). Fascinated by the possibilities of self-representation and drawn to a more decorous environment than IRC, I logged in as a guest and shortly thereafter became a registered citizen of MicroMUSE. Within weeks I observed social conflict in this allegedly tranquil community to rival any I had seen or studied in real life as a social scientist and practitioner of mainstream Western dispute resolution techniques. For the next two years, first as an ordinary citizen, then as a nonvoting member of the governing body, more recently as a voting and technically-powered director (in MUD parlance, a "wizard"), I observed and participated in the community's interpersonal conflicts, enacting roles of audience, aggressor, target, investigator, reporter, confidante, judge and conciliator. In this paper, I report what I learned during that period about the problems, strategies and techniques of maintaining social order in cybercommunities. In particular, I assess the application of well-established physical world tools of conflict management to virtual disputes.

Theoretical Background

Various social scientific perspectives have different views of the incidence of conflict and its value in social systems, and accordingly offer different approaches to conflict management. Unitary models (also referred to as "monolithic" and "integration" models) seek and often find system-wide consensus, wherein values, meanings, goals, interests, rules and the like are shared and internally consistent. Individuals have clear understandings of what things mean and how to behave. They pursue the aims of the system because in so doing they address their own interests (McGregor, 1957). This integrated world is one of "harmony and homogeneity" (Martin, 1992). Many researchers and managers with this perspective see conflict as rare, disruptive to social integration, and symptomatic of system pathology. It is often attributed to deviants and provocateurs who can be controlled or eliminated by appropriate system design and action (Morgan, 1986, p. 188). The aim of conflict management under this perspective is resolution and prevention for greater system unity (Deutsch, 1973; Filley, 1975).

This negative view of conflict is shared by others who value harmony--or perhaps power and control--but who view differences and conflict as widespread, natural properties of social systems. Under this frame of reference, conflict management has the same aim as it does for integrationists-minimization and elimination. However, because their conflict management tools act in opposition to individuals and groups, they dominate, suppress or coerce compliance rather than achieve consensus. As many critics of bureaucratic theories of organizations have noted, such managers paradoxically obtain what they hope to avoid by this course of action: less control as people comply in inappropriate ways or find ways to resist (Dalton, 1971, provides an overview of this literature).

The radical model sees differentiation and conflict, too, but takes a positive view of conflict. Trade union leaders, for instance, induce conflict with management as a way to strengthen their position vis--vis management (Tannenbaum, 1965). Such conflict develops consensus within the subsystem while increasing separation from the other, thus polarizing the system and creating disunity. Ultimately, of course, radicals see conflict as a means to structural change through a process of dialectics. Thus, the opposing parties, one struggling to maintain the existing order, the other to reverse it, destabilize the system and co-produce change (see Morgan, 1986, especially chapters 8 and 9, for a more complete discussion of this perspective). Whether such conflict is thought good or bad depends on one's moral opinion of the system being destroyed. Certainly conflict that undermines an existing system is dysfunctional to that system, but it is good if the system is evil.

Not all views depend on moral judgment to find positive aspects of conflict. Georg Simmel (1956) saw human relations characterized by ambiguity rather than clarity. The way this plays out in terms of conflict is that social systems are seen always to involve both harmony and dissonance. To him, a completely peaceful social system simply cannot exist because of its incapacity for change and development. Harmony and consensus are not, therefore, per se good nor is conflict per se bad. Rather, both are positive, being necessary for system survival. For example, conflict structures a system, strengthening existing bonds (as within the union in the previous example) and forming new ones as participants become polarized into blocs that persist beyond the conflict. Small conflicts also allow parties to ventilate, reducing the likelihood of larger, destructive conflict. In these and other ways, conflict promotes system endurance and viability (Coser, 1956). By this line of reasoning, conflict management aimed at eliminating discord is not only an impossible task, but off the mark. Indeed, even some writers outside the radical school argue for and describe techniques of conflict stimulation (Peck, 1987; Robbins, 1974). Brown (1983), for instance, explicitly recognizes and integrates both functional and dysfunctional conflict outcomes into a unified model, hypothesizing a curvilinear relationship between conflict intensity and its outcomes. Too little conflict is said to produce a false sense of well-being and low levels of energy, creativity and adaptation. Too much conflict produces antagonism, restricted information flow, low-quality decisions and diversion of effort away from other tasks to deal with the struggle. Either condition may cause system failure. Positive outcomes of conflict and, therefore, system effectiveness are realized in the middle range of conflict intensity, where there is neither too little nor too much. Effective conflict management, then, becomes a matter of either reduction or stimulation as called for by the situation (Brown, 1995).

Another property of social systems that is thought to affect their effectiveness is diversity of population in terms of perspectives, skills, knowledge, abilities and the like. One way unitary systems achieve integration is through homosocial reproduction, selecting, socializing and promoting to positions of power similar others. While social similarity produces what Durkheim calls "mechanical solidarity," the stability and inertia can prove fatal when the environment changes and the system lacks the knowledge, skills and abilities necessary to adapt. Pluralistic systems, on the other hand, because they contain a broad variety of perspectives, skills, etc., are more likely to contain the requisite qualities for the new demands of the environment. If granted expression, differences can also generate a multitude of ideas for improving system performance, and greater capacity for their evaluation. Potential benefits of diversity, then, are high quality decision-making, creativity and innovation, and adaptation (Peck, 1987; Cox, 1991).

However, diversity of values, perspectives, beliefs, interests, knowledge and expectations is also a source of conflict (Pfeffer, 1981; Macduff, 1994). Pluralistic systems, which are more complex, more elaborately structured than those of the radical model which focuses on dichotomy and opposition, have a vulnerability of their own. As loose networks of people with multiple goals and diverse interests, whose alliances shift over time and across issues, achieving consensus on anything may become insurmountable. A system of extreme heterogeneity, while it may contain the raw material for creativity, flexibility and adaptability, is vulnerable to anarchy and anomie. Social systems operating in environments demanding creativity and innovation thus face an apparent dilemma. If they opt for tranquility either by denying expression of their diversity or by choosing homogeneity, they deny themselves precisely what they need to survive. But if they admit diversity and allow it expression, they may experience too much conflict to be productive and thrive. And so we come to the Hobbesian question of order: how is society possible?

The unitarist answer is avoidance of conflict through homogeneity. As argued above, this contains the seed of system failure. So, too, does the radical answer, coercion, which is conflict-seeking in its effects if not intent. The pluralist response is the Aristotelian ideal of politics wherein diversity is neither silenced nor avoided, but given expression through constructive conflict management. In the political model of social systems, order is created from diversity by bargaining and consultation that reconciles divergent interests and creates shared understandings (Morgan, 1986; Pfeffer, 1981). In sum, a social system may be more or less effective depending on its degree of diversity and its strategy for and skill at managing conflict (Cox, 1991).

MU*s are social systems which can evolve into symbolic communities, consisting of persistently interacting members with common interests who are linked primarily by symbolic (in this case, electronic) exchange rather than by face-to-face encounters in physical space (Gergen, 1991, p. 214; see also Clodius, 1997; Scime, 1994). They are also organizations when they are designed to achieve limited objectives through the coordinated activities of their members (Presthus, 1978, p. 1). The manifest goals of these organizations are those designated officially, often set forth in a mission statement. They also have informal or latent goals, which are the aspirations of their members. As with any organization, the two sets of MU* goals are likely to be at least in some degree inconsistent with each other. This research is a case study of an open MU* with limited, formally-determined and stated goals. Evidence is found of significant latent goal incompatibility with the manifest. I argue that such virtual organizations have the same kinds of problems and opportunities brought by diversity that real organizations do. Indeed, to the degree that they are open and attractive to the Internet population, they will be particularly prone to conflict and therefore in special need of effective conflict management techniques. This research also examines the conflict management techniques employed by the organization's members and introduced by the researcher. It concludes that unique features of the virtual world contribute to the problems of conflict management. Unless these problems are addressed by adapting existing techniques or developing new ones, there are limits to the degree of openness and diversity that virtual communities and organizations can tolerate. This, in turn, will affect their ability to adapt to their changing environment.

The Setting and the Data

MicroMUSE was founded in 1990 under another name as a recreational/gaming MU*, but by 1991 its manifest goals had become educational and a second generation of administrators had adopted its first formal charter (MicroMUSE Charter, 1991).3 The setting is that of Cyberion City and its environs on a cylindrical space station orbiting the Earth. In 1993, guests arrived in the Main Transporter Receiving Station, having been beamed up from Earth. The total population of registered characters was over 2000 in 1993 (Rheingold, 1993, p. 162), but grew rapidly to over 3000 by April 1994 (MacDuff, email to Advisory Board, April 18, 1994) after the community received favorable publicity in a number of books and magazine articles. It is impossible to know how many people these characters represented, as some would have been abandoned from lack of interest or loss of Net access and some people would have had more than one character or shared with friends or family. The character population shrank to 1000 by December 1995 (online statistics) as a result of events which are described below, but since multiple and shared characters are now prohibited, and long inactive ones automatically recycled, this is probably a more reliable estimate of the human population now than 1993's estimate was for that time. The distribution of social characteristics is also not known, but it might be assumed that the largest group of users was male and of college age, since conventional wisdom has it that this was the largest single category of Internet users at the time (Rheingold, 1993, p. 150). The concentration of young men may be declining, along with the population, because of recent concerted efforts to refocus the community on its manifest function and widening global access to the Net. Few informal and no formal inquiries are made of a user's race, but nationality and geographic location are frequent topics of conversation. Most users clearly are from the U.S. and Canada, although it is evident that Cyberion City is cosmopolitan, drawing users from Europe, Asia, Australia and Latin America. Among the countries known to be represented are Austria, Czechoslovakia, Finland, Ireland, Italy, the Netherlands, Poland, Russia, the United Kingdom, India, Japan, Korea, the People's Republic of China, Singapore, Taiwan, Thailand, Australia, Brazil, Colombia and Mexico. Two features of MicroMUSE that distinguish it from most other MU*s open to the Internet are its educational focus and deliberate attempt to serve users under the age of eighteen. As discussed below, these features in interaction with its open-door policy have affected the character of its members' relationships as well as the composition of its population.

The formal organizational structure was changed during the period of this study, but underlying role and class distinctions were constant. The two basic class distinctions are users (or players) and administrators, the latter of whom perform the tasks necessary to keep the community functioning in a technical sense and following the direction outlined in the mission statement. User characters are of two types. Temporary characters (called "visitors," formerly "guests") lack the ability to create objects or rooms and to change their name. No password is necessary to use one of these and they are recycled when the user disconnects. Permanent characters (called "students" and "members," formerly "citizens") are registered and require passwords to access. They have the technical ability to create structures and processes, and to communicate and move about the community. The administrative structure is too large and complex for description here, but knowledge of the finer distinctions is not necessary to understand the phenomena under investigation.4 Here, "director" will be used to refer to the class with the greatest authority and technical power, "staff" to other ranks, and "administrators" to all powered ranks. The decision-making bodies will be identified and described as needed.

The sources of data for this investigation are several. As a participant-observer, I spent thousands of hours logged into MicroMUSE and its sister communities, immersing myself in the life of the MUSE for a period of two years. During this time I had thousands of conversations on line with guests, registered users and administrators. In addition to my own notes and transcripts, other users supplied transcripts. Official documents were obtained online and by ftp, gopher and the web. I also subscribed to the open-subscription email lists and, later, as a member of administration, I was a party to the closed- subscription lists by which much of the governance of the community is conducted. Finally, there were many email and telephone contacts, and a number of physical world encounters (known as "face-to-face" or "in real life"), including organized weekend gatherings and impromptu or planned visits by or to fellow MUSERs.

My entry and subsequent involvement in the community as a participant-observer is broadly described in the introduction. As a user, I was initially cautious about revealing my true identity, but became open about it and later my purpose after I made the decision to use the community as a subject of research. I then sought and obtained the consent of the Advisory Board to reveal the identity of their MUSE and make use of the data, but, with a few exceptions, I have used pseudonyms or titles to conceal individual identities.

Home Author's Home Contents Table of Contents Replay Previous Section References References

The Problem

Background

Although many of their projects are about science, MicroMUSE users are free to pursue any topic of interest within the constraints of the community's culture and technical resources. The process of learning, and the learners and their relationships with each other, are more important than any particular subject. Learning is supposed to be fun, self-directed, collaborative, peer- and project-based (Kort, 1995). People of all ages and backgrounds are welcomed, but there is a deliberate effort to attract and involve children in kindergarten though grade 12. This effort appears to be successful as my impression is that a large number of the currently registered users are, in fact, children, parents and teachers. In support of its mission, MicroMUSE attempts to provide an environment its administrators deem suitable for children and supportive of learning. Indeed, its Charter states its theme to be "based on an optimistic vision of the 24th century" (MicroMUSE Charter, 1994). This vision is optimistic both technically (matter transmission is easily accomplished across the distances of the orbiting space station) and socially (life is supposed to be peaceful with most time spent on scholarly pursuits of discovery, creation and discussion). Violence, obscenity, racism, pornography, harassment, theft, invasion of privacy, and other kinds of conduct are prohibited (MicroMUSE Bylaws, 1994) but not, as the cases below demonstrate, absent.

User-Administrator Disputes

Nearly the first thing I learned on line about MU*s was the power of the directors. PeterPan, the 27-year-old unemployed American who introduced me to MicroMUSE, was regularly punished for making frivolous announcements, using foul language in public, and harassing administrators. For this he would be fined, lose the privilege of making announcements to the community, enslaved (a condition that substantially impairs a user's ability to communicate and move), or arrested (forcibly placed in a room with no exits). On occasion, he would be summoned (transported against his will) to a director's office and lectured. On other occasions, he would be punished without being told by whom or why, although he frequently surmised the reason. However, the power of a director is not absolute, as Swagger's case illustrates.

Swagger was a teenager who built an Orgasm Room filled with sex objects to which he apparently invited female users for online sex. When a director discovered this, Swagger was summarily nuked (the MicroMUSE term for complete removal of the character and all its virtual property from the database). Swagger returned as a guest and complained about his treatment on the public channel (a medium of mass communication similar to a citizens-band radio channel). He was successful in raising the support of various citizens and staff who believed that he should not have been terminated without an opportunity to defend himself. Two staff cooperated in the re-creation of Swagger's character. At this point, the director whose action had been countermanded objected. A vote of the Executive Board (a body composed of all directors) affirmed the director's original action. Swagger was nuked again, and informed that he had been exiled for a period of two months, after which he could petition for readmission.

Before Swagger's exile by the Executive Board, there was a town meeting presided over by yet a third staff member. About forty people, including Swagger, attended and discussed restrictions on citizen speech, governance and the justice system. Sentiment was expressed for the election (rather than the appointment) of administrators and/or other means of providing citizen input to the governing body. The attendees also complained about the lack of a meaningful system of justice, alleging power abuse by directors, no opportunity for the accused to defend himself, no trial by jury, and no code of laws (the latter was untrue as everyone was emailed a list of rules upon registration). Shortly thereafter, an unmoderated, open-subscription mailing list was initiated upon which citizens aired their complaints about MicroMUSE and posted various items of general interest, such as the transcript of the town meeting.

One month later, a new Charter was adopted, creating a complex system of governance which included a Citizens Council with a nonvoting representative on the new primary governing body (Advisory Board), and a role for appointed voting advisors who might or might not have technical powers. The 1993 Charter also elaborated and codified the procedure for dealing with rule violations, but did not contain all elements of due process systems sought by the dissidents, such as peer jury trials. It did, however, promulgate the limited right of appeal first granted to Swagger. At about the same time, some of the more disaffected citizens mounted their own MUSE, dubbing it "AntiMicroMuse," upon which virulent attitudes towards MicroMUSE administrators were expressed. This MUSE was enormously successful for a brief spell, but it was operated in violation of the policy of its system administrator's university, so was short-lived.

The Citizens Council was a complete failure in the opinion of both administrators and users. The election was essentially a popularity contest, and one of the community's most notoriously destructive hackers claims to have stuffed the ballot box to assure a win for his girlfriend. The first Council worked on a few projects, without success, and eventually became completely inactive.

Swagger's petition for readmission was one of the early cases I took on. He was readmitted after an investigation concluded there was no evidence to rebut his assertion, supported by witness statements, that he had reformed. Although he was never again accused of sexual misconduct on MicroMUSE, he did break several other rules, most notably exploiting bugs in the code and using others' characters without their consent. This led to his permanent banishment from MicroMUSE, but for some time he continued to connect illicitly, logging on as characters whose passwords he cracked, and again exploiting code bugs.

Disputes among Users

Users naturally have problems with each other as well as with administrators and MUSE policy. Harassment, sexual harassment, assault, spying, theft, and spamming (filling a user's screen with unwanted, meaningless text) are typical complaints. The following cases illustrate the usual course of events in interplayer disputes.

The Trojan Horse. One evening, a stranger named Sinon entered Hecuba's office. After some casual conversation, he gave her a present. Touched, she accepted it and looked it over. It came with instructions which she followed, whereupon she was spammed by silly material for some minutes. Hecuba thanked the stranger in her return spam, thinking this was a bit of harmless fun. Sinon laughed, saying he could win any spam game with her because he had a high-speed direct connection to the Net and the edge in technical skill. He then proceeded to demonstrate that he could disable her computer for such a long time that she would be better off disconnecting than waiting for her screen to stop scrolling. When she got control of her computer again, Hecuba spammed him back. Sinon's next present was subjected to a closer scrutiny. Hecuba examined it every way she could and saw only the shadow of an object inside the gift. Curious, she followed instructions again, thinking that at worst, she would have to disconnect to avoid more spam, but this time the game turned into war as this object destroyed nearly everything she owned and then booted her from the MUSE. She reconnected only to see the extent of the destruction, then logged off for the night to lick her wounds. Defeated and depressed she returned the next day and began reconstruction, but was so ashamed at having been duped that she told no one about it for a long time, not even the director who noticed the shambles of her possessions and asked her about what had happened. Eventually the director pieced the story together and nuked Sinon's character. However, Sinon returned again and again, as each different character he used was discovered to be controlled by him and nuked.

The Hired Gun. A similar incident took place on a sister MUSE, one less controlled by its administration than MicroMUSE is. This time the aggressor was an anonymous guest using the public channel to harass other players with obscenities. Outraged players solicited the aid of the aforesaid PeterPan, who was an accomplished spammer. He invoked a single command that flooded the offender's screen for some time, eventually causing him to disconnect. The offender returned in an angry mood, but his retribution was less damaging than Hecuba's Trojan horse as he only dropped a final, huge, spam bomb on the channel before leaving for good.

The Scorpio Mutiny. A final example of interplayer conflict occurred within a previously collaborative task-oriented team. By way of background, one of the larger and more interesting educational projects on MicroMUSE is Space. Participating users are able to construct ships of various types and fly them in real time and three dimensions to celestial destinations, which may also be of their own creation. Individual space ships are often large, complex structures crewed by multiple players. To make it easy for users to cooperate in the construction and maintenance of such large-scale projects, characters may be created as holding or operating corporations, to which approved users have access. One such group character was Scorpio, which was created to hold the starship ICS Scorpio. Among those using it were two men, Ajax and Achilles, and Achilles' robot. Achilles and Ajax evidently constructed and operated the ship successfully together for some time. If there were any internal power struggles on Scorpio, they did not come to the attention of MUSE administration.

In the fall of 1993, computer memory and disk space limitations caused MicroMUSE administration to decide that the entire space project should be moved to a separate MUSE (MicroMUSE Charter, 1994). When this was announced, Ajax lost interest in MicroMUSE entirely and argued for the destruction of the ship. Achilles was opposed, hoping to persuade administration to keep the ship on MicroMUSE as a museum. Announcing he would not be back, Ajax left the MUSE. He returned from time to time, but kept away from the ship. During this period, Achilles mothballed the ICS Scorpio, which was still his home, but eventually decided to take one last voyage. Worried about the risk Ajax posed to the ship and needing a captain, Achilles promoted his former first officer, Nestor, to captain, and demoted Ajax. He also had an administrator remove Ajax from the list of characters who could modify the ship. When Ajax returned to the MUSE, he took a look at the ship's roster and discovered the changes about which he had not been consulted. Believing his ship had been pirated, he gained access to it through Achilles' robot (whose password Achilles had forgotten he had shared with Ajax). Within about an hour, Ajax used the robot to destroy the entire ship, with its 120 rooms and objects, and enslaved its other robots. He was open about what he was doing, and asked to be nuked when he was finished. Coincidentally and tragically, Achilles was repeatedly trying to connect to the MUSE during this period and failing. When he finally logged in and saw the destruction, he asked a director to rescue his ship from a backup database, but the director refused, saying Achilles should have been more careful about sharing passwords. In the aftermath, all users who lost property in the attack--Achilles, Ajax, Nestor and others--levied charges against administration for both suspected actions (complicity in certain facets of the affair allegedly motivated by a desire to get rid of these users) and known inactions (failure to stop the destruction once it became known). Ajax was nuked as he requested and the password of the robot changed. A backup copy of the database was eventually made available to Achilles and Nestor, and they began reconstruction, but never regained their former interest in the project or MicroMUSE.

Discussion

These cases do not depict usual user conduct, nor is any one typical of the conflicts observed in the period studied. Rather, they collectively include many elements of the broad range of disputes and have therefore been selected as illustrative of sources of MicroMUSE conflict and the ways in which they were being handled.

Diversity as a Source of Conflict. Much of the conflict observed arose because of the different meanings of MicroMUSE to its users and administrators, and their diverse values, goals, interests and norms. The following sketches in broad terms some of those different realities.

Beginning with administration, which is responsible for directing the organization towards the fulfillment of its mission, its vision of MicroMUSE was most graphically expressed early in 1994 as the Advisory Board was trying to come to grips with the amount of conflict the community had been experiencing. One director likened MicroMUSE to a science museum which is open to the public but intended for specific educational purposes (D. Albert, email to Inner Circle, February 19, 1994). Under this model, all behavior in the museum is judged in relation to the museum's needs, which are prior to all others. If teachers cannot bring their classes to the museum because a room is openly designated for copulation and teenagers are announcing obscenities on the public address system, or scholars cannot study because of excessive noise, or the building is being vandalized by hoodlums and exhibits destroyed by bombs, then the museum will fail. To survive, virtual communities must protect their primary resources. Being no exception, MicroMUSE administration protects against threats to its physical plant (machine and code) and citizens (Internet gateway and users). Because of its manifest organizational objectives, it must also protect its capital supply (grants and donations), its labor supply (educators and technicians), and its markets (students and their parents). These efforts will come up against users with different perspectives and agendas (often in direct conflict with the museum's) who will resist or try to circumvent them in their own effort to maintain their own autonomy and achieve their own goals. Hence, there is user-administrator conflict as illustrated by the Swagger rebellion.

Turning now to the users, many--particularly those whose expectations of MU*s are shaped by social MU*s such as FurryMUCK or adventure MU*s like BatMUD and Revenge of the End of the Line- -have primarily social or recreational interests. Although they do some building and exploring through which they learn some coding and matters of substance, this is not what they really log on for. They come to visit with friends, meet lovers, play games, make toys, have parties, experiment with identity and behavior, act out difficult feelings, feel powerful, even to heal.5 To them, MicroMUSE is a playground, singles bar, bedroom, coffee house, shopping mall, theater, front porch, therapist's or social worker's office and more. Some are astonished to find that what is tolerated in those real life places is not permitted on MicroMUSE. Swagger, for instance, claimed--possibly disingenuously--that he was unaware of the restrictions. Others, holding that MicroMUSE is not real, being only text from which one can disengage at the flip of a switch, or who subscribe to the hacker ethic,6 believe that anything technically possible can and should be allowed. To them, MicroMUSE is a game with the rules subject to renegotiation or to being broken for one's own advantage or the sake of discovery. The Trojan horse case and Scorpio's internecine war were fed by this view.

Another vision of MicroMUSE is that of a city-state, a metaphor ironically supported by administration in its decision to represent MicroMUSE as a city and to label its users "citizens." At least some of those expressing this view see the existing form of government as totalitarianism (often calling administration "fascists" or "communists") and hold democracy to be the ideal. A few, who are also extremely vocal, said democracy is an impossible dream because of administration's lust for power. This view reflects the structural pattern of MicroMUSE: division into two parts, with the two halves performing specialized (in this case, reciprocal) functions. Such societies, known as "moities" to anthropologists, work best when power is distributed equally. If one faction is more powerful than the other, there is a tendency to dominate the weaker faction, which then struggles to become stronger so as to avoid domination (Thomas, 1984). In MicroMUSE, as in other cybercommunities, ultimate power rests with administration, as the party controlling the essential resources of the community--the code and the machine upon which the program runs. Dissident users complaining of fascism were acting in recognition of this asymmetry and the resultant imposition of the stronger faction's vision of the community on the weaker. The rebellion that was sparked by the Swagger case can be understood as a struggle for the right to shape the culture of the entire community (Morgan, 1986, p. 127), and of the user visions described here, this is the one probably responsible for most of the user-administration disputes during the pre- and immediately post-reform period of this study. The large influx of new users attracted by recent publicity and a database swollen to machine capacity put MicroMUSE at a transition point in the fall of 1993. While Swagger's treatment was a cohesive influence in one respect, providing a common cause for diverse elements of the community and mobilizing them to action, the intense political activity during that fall and winter was less about Swagger than the future direction of the community.

The three major interests of those who see MicroMUSE as a city-state were touched on in the town meeting following Swagger's resurrection: freedom of expression and other civil rights, user voice in MUSE government, and justice both through the equitable distribution of resources and due process in administering discipline. Of the three rights Karnow (1994) sees as important for protection of electronic personalities, users with a city-state view most value two, freedom from discrimination and freedom of speech. In fact, most of the administration-user disagreements are, like PeterPan's, over the content of public communication or public representation of self such as character and object descriptions. They are also the ones most likely to engender open community support for the accused. When a popular user called "Debs" was arrested for a second violation of public obscenity and later exiled when she failed either to agree with administration that her language was inappropriate or to promise to refrain from using such language in public again, the case was discussed for weeks both on and off MUSE. Although much of the discussion was both ill-informed and more accurately characterized as "flaming"7 than discourse, there were also serious attempts to understand the issues. For example, I received a request to moderate a debate on civil rights in virtual communities and a law student sent me a brief on the First Amendment and cyberspace for discussion by administration. Her argument did not persuade the Advisory Board to end its censorship.

There are, of course, many players who use MicroMUSE for its intended purpose or whose enjoyment of social activities is neither disruptive of nor disrupted by the organization's pursuit of its manifest goals. These users, in fact, probably appreciate and may even rely on administration's attempts to maintain order. For them, Karnow's (1994) first right of electronic personalities-the right to privacy-is primary. When teachers and parents see the safe environment they want for their students and children compromised, and when builders or conversationalists have their activities disrupted by other users' free expression, they often become complainants in MUSE disputes. Debs was reported on both occasions by other users, one of whom was trying to demonstrate the MUSE to educators at a time Debs exercised her right of free speech to utter obscenities on the public channel.

That administration shares these users' high esteem for privacy is shown not only in its rules (entering private space without permission is forbidden, for example), but in a relative absence of monitoring user behavior. Complete permanent logs of commands executed are not made, and only directors have access to the brief transient logs that are made for debugging after a crash. Even then, reading these logs without good reason is considered outside the bounds of propriety, although doing so would make it easier to identify and prosecute miscreants.8 On the other hand, while administration does not go looking for violations of the rules, either by reading logs or through other forms of surveillance, it does try to prevent them and correct those that come to its attention before users complain. Swagger was caught because an administrator saw the name of his room in the publicly-viewable directory, not because a citizen complained or administration was reading command logs and saw what he was doing in the room. Thus, in terms of the rights discussed by Karnow (1994) as being fundamental to managing the private/public border in cyberspace, the line on MicroMUSE seems rather clearly drawn. On the one hand are administration and the users for whom privacy is paramount; on the other, there are the users valuing free speech more and seeking greater influence for themselves that they might be discriminated against less and achieve greater freedom of expression. In late 1993 and early 1994, there were substantial numbers of the latter who were freely and loudly expressing their opposition to the former.

A related source of difficulty is the variety of cultural norms brought into the community by its users. A complex interplay of factors including generation, gender and geographic cross- cultural prejudice and misunderstanding were involved when an adolescent girl from one country and culture accused a man from another of sexual misconduct, a charge he vehemently denied. About the only thing on which there was consensus when the case broke was that the girl claimed to be his victim. Some users believed the charge true because "that's the way men are in [that culture]." Others thought the man probably was unaware of how conduct that seemed innocuous to him could be construed by others as injurious. Other views focused on the girl. Some believed she did not fully understand the word she used to characterize his behavior. In their view, his crime, if any, was probably less serious than alleged. Others felt she had merely misinterpreted the man's behavior, and still others questioned her motives, believing the man to be the victim of an immature, and possibly disturbed, teenager's retaliation for a perceived rejection. When it was discovered that the victim of the alleged crime was not the accuser, but someone who was not even a member of the community, this lent credence to the retaliation theory, but still left most of those who were aware of the case distrustful of the accused. As a consequence, he was marginalized. Feeling unwelcome, he redirected his virtual life to other communities, although he did maintain a presence on MicroMUSE and forgave the girl for her role in the affair.

Although it is uncertain whether either of the antagonists in the preceding case supported the notion of a family environment, even those who do so disagree about what that means. For example, a husband and wife from the southern United States once complained about a young man's description that contained a reference to an alcoholic beverage and about swearing in the presence of women users, while college students frequently question why expressions commonly heard in shopping malls, playgrounds and in their own childhood homes are not permitted. Thus, language and behavior disallowed by administration are explained and sometimes defended on several grounds: ignorance of the rules, virtual inequivalence to reality, freedom of speech, and real-world community standards.

In sum, the experience of MicroMUSE suggests several things about virtual communities and conflict. First, to the extent such communities are open to and draw from the Internet's increasingly heterogeneous population with a diversity of values, interests and expectations, they will experience conflict. In addition, when the virtual community has manifest goals, conflict will be greater to the extent that those goals are not understood or shared by members of the community. Third, the social structure of such communities, being one of asymmetrical dependence and power, is an underlying source of conflict and instability. Fourth, given power inequality and a lack of community-wide consensus, administration's strategy for dealing with dissent may have the unintended consequence of increasing and intensifying user conflict behavior. I turn now to this latter point.

Social Control. The MUDs studied by Reid (in press) had a medieval system of authority and correction, focusing on the body of the condemned as the locus of punishment. While MicroMUSE has the technical and social devices available to it that Reid's MUDs do, it does not employ them dramaturgically to publicly shame or torture its deviants. In 1993, social controls were more likely to consist of explanations, requests, warnings or lectures in a director's office than public chastisement and humiliation. Technical controls, while they did affect the player's virtual body such as by removing the power used to offend (ability to communicate or move, e.g.), did not alter the appearance of the offender in order to shame him (a practice known as "toading"). In fact, administration, in keeping with the privacy norm, was loathe to report or discuss disciplinary cases in public. The more extreme measures used to tend community borders by gods in Reid's MUDs were, however, frequently used by directors on MicroMUSE: destruction of the offender's character (exile or banishment) and disallowing all connections from the offender's host computer (sitelock) for an hour or two to give the user time to cool down. The sitelock was also sometimes kept in force until the offender's account was cancelled by his host system.

That these measures, more modern and humane though they may seem to us than medieval humiliations, have been incompletely successful as is evident in the cases described above. Because the Net has so many routes and so many host systems on which to obtain accounts, there are endless ways to defeat the sitelock. Other features of cyberspace and its communities assure that a determined user can return: availability of anonymous guest characters, ease of deception in obtaining email accounts and therefore registered characters, impersonation possibilities because of the casual disregard for security many users have with their passwords, and insecure software. The impotence of banishment and sitelocks is illustrated by Swagger and by Hecuba's assailant, who, as far as I know, may even yet be connecting.

Once an exiled user does return, s/he has both technical and social means to harass other individuals, to disrupt the entire community, and otherwise to challenge the power of administration. Swagger raised community support through the mass media, as did Debs. Users have executed commands or written programs to slow execution of commands ("lag") or crash the computer running MicroMUSE, exploited bugs to destroy database, fraudulently subscribed administrators to hundreds of mailing lists, and harassed and assaulted other users. In other words, much of what they can do to get into trouble in the first place, they can also do to retaliate. Some, of course, amend their behavior and others simply leave, but in the fall and winter of 1993-94, it was evident that a significant share of administrative time and energy was being spent in a ceaseless and escalating battle with users.

Players, too, employ both social and technical means to protect and assert themselves, most of which are equivalent to real life measures. Among these are locking doors and possessions, screening communications, sweeping for listening devices, shunning, ridicule, ostracism, negotiation, and complaining to a person in authority. As the above cases show, there was a frontier ethic of taking the law into one's own hands. Retaliation and counter-retaliation in cycles of escalation were common, and formation of posses and use of hired guns occurred. How many and by what means interplayer disputes were resolved before they escalated to the point where administration became involved is unknown, but cases like the Scorpio mutiny demonstrate the external costs of an informal system of frontier justice: innocent third parties lose intellectual property, administration spends hours investigating and deciding the case and helping with reconstruction, and users lose respect for an administration felt to be unresponsive to their needs.

The reason for this is not just the many conflicts arising from the diversity present and the lack of absolute technical power to expel forever unwanted users. It can also be explained by the way in which administration proceeds when confronted with disobedience to the rules. The Swagger case is informative. Although the director who nuked Swagger originally acted within his authority, his apparent failure to investigate fully and fairly or to give Swagger an opportunity to amend his behavior shocked members of the community and some staff, two of whom countermanded him while acting outside their own authority. Administration as a whole lost respect and the director was under attack in private and public forums, wherein other grievances were raised. Although this was an unusually dramatic case, there were many others in which administration was viewed as acting arbitrarily, capriciously or discriminatorily and consequently came under vicious, even malicious, attack. When subjected to harsh criticism, personal insults and other acts of revenge, one might be expected to withdraw from the field, defend oneself, or counterattack. All of these reactions were observed to at least some degree. Thus, the charges that administration was unresponsive, remote and unmerciful became self-fulfilling prophesies, fueling still more criticism. The conflict escalated further as the community polarized and each side struggled for control. Administration was on its way to becoming what many in the community all along thought it was--fascist police--rather than the helpful servants, gentle educators, and wise mentors it envisioned of itself. Clearly, MicroMUSE was not the homogeneous community of consensus its powerful administrators wanted. Neither did it resemble the pluralistic vision of society, a web or mosaic of different realities and interests sewn together by conflict (Simmel, 1956). Instead, a malignant conflict was becoming intractable (Deutsch, 1995, p. 127). The radical model of dichotomy and opposition in upheaval was a better fit. It was at this point, "after years of adjudicating all manner of interplayer disputes with little to show for it but their own weariness and the smoldering resentment of the general populace" that LambdaMOO's Pavel Curtis turned social control of that community entirely over to the players (Dibbell, 1993), effectively altering the balance of power and resolving the structural conflict. Since the Swagger case, MicroMUSE has made some reforms, but it has not taken the democratic path chosen by LambdaMOO.

Author's Home Contents Table of Contents Replay Previous Section References References

The Reforms

Boundary-Defining Reforms

The changes to the Charter passed in 1993 did not effectively address the sources of the conflict identified here, either the competitive relationships produced by plurality and power imbalance or the positive feedback system of conflict management. So, in 1994, several new reforms were introduced. These were designed to define the boundaries of the community more clearly, to provide better monitoring of user conduct and more rational sanctions for rule violations, and to improve conflict resolution, all of which are among the design elements Ostrom (1990) posits as being essential for robust common-pool resource institutions. One reform was technical and intended to control unacceptable behavior by anonymous visitors. It involved rewriting the server to block nearly all commands entered by visitors. This code modification also significantly curtailed visitor ability to communicate with other users except those designated as "mentors," who act as docents and help visitors register for resident characters. This reform thus protected the machine, code and users from disaffected players and outsiders with little or no investment in the community.

A second reform significantly altered the process by which new users are introduced to the community. The objective was to reduce conflict caused by diversity by forming a more monolithic society. Like the technical reform, it had a boundary-defining effect. Full membership privileges now required sponsorship by two mentors after a period of socialization. During this period, new users are exposed to the intended culture and complete several tasks, working with mentors to navigate the process and learn rudimentary coding. Theoretically, at least, mentors do not sponsor someone until they know the user is aware of MicroMUSE's mission, accepts the rules and can get around in the MUSE environment. The premise of this reform is that people uninterested in a learning community will not join, while those who stay will be both better informed and more committed by virtue of overcoming a modest hurdle to full status. Thus, self-selection and acculturation would close up the boundaries and produce a more homogeneous population with everyone having at least a little investment in the community.

A third reform institutionalized conflict by providing third-party dispute resolution tools that are used to manage real life conflicts. This is the subject of the balance of this essay.

Dispute Resolution Tools

Ury, Brett and Goldberg (1988) identify three basic elements of all disputes: interests (the things one really cares about and wants), rights (standards with perceived legitimacy or fairness), and power (the ability to coerce someone to do something against his will). These elements provide three basic strategies for resolving disputes: reconciling interests, adjudicating rights, and exercising power. These strategies have different costs and benefits: the costs of disputing (transaction costs), satisfaction with outcomes, long-term effect on the relationship, and the probability of recurrence. Ury et al. argue that reconciling interests is less costly than adjudicating rights which in turn is less costly than exercising power. This is because reconciling interests tends to result in lower transaction costs, greater satisfaction with outcomes, less strain on the relationship, and less recurrence of disputes than the other approaches. The strategy employed in MicroMUSE in 1993 as described above was counterproductive because it depended on power, either acts of aggression (spamming, destruction of property, etc.) or withholding of benefits (the sitelock and loss of powers, e.g.). It accordingly had high transaction costs, produced unsatisfactory outcomes for one or more parties, strained relationships and led to recurrence of conflict since new injuries and disputes were created in the process. The third-party procedures introduced in 1994 were designed to reduce the dependence on power contests by attempting to reconcile interests or adjudicate rights, thus yielding a more effective system of conflict management. In the order in which they are described below, the amount of authority (and, therefore, intrusiveness) of the third-party neutral increases, as does the focus on rights compared to interests.

Mediation is negotiation assisted by a third party who facilitates the disputants' agreement on a solution to their conflict. The success of this approach depends more on the neutral's ability to communicate and influence than to decide, but on-the-spot analysis of rapidly evolving situations and deciding on appropriate ways to proceed are critical. Research into labor mediation has identified two strategies. In deal-making, the mediator figures out what settlement is mutually acceptable to the parties and then manipulates them into adopting it. In orchestration, the mediator enables the parties themselves to discover a settlement that meets their mutual needs (Kolb, 1983). The latter approach is more likely to result in settlements that are in the best interests of both parties and is thought to make them less dependent on future third-party intervention than the former. Mediation is commonly used in labor, environmental, community, school, family and divorce, and international disputes, often to keep them from escalating to the more costly rights or power procedures.

Factfinding is a quasi-judicial process in which the neutral conducts an evidentiary hearing and issues a report. Factfinders have the authority to decide the facts of a dispute and may also have the authority to make recommendations for resolution, in which case, the procedure may be termed "advisory arbitration." Typically, if the parties do not accept the factfinder's recommendation, the report is made public. Resolution is theoretically facilitated by the rationalization of the dispute in the first place (determination of the facts and rejection of meritless argument), and by the pressure of public opinion in the second place.

Arbitration is also a quasi-judicial process. It differs from factfinding in that arbitrators issue decisions which, by prior mutual agreement of the parties, are final and binding. It differs from court proceedings in its simplicity and informality which produces results faster and at less cost. The arbitrator's authority flows from and is therefore constrained by the arbitration agreement, which may be a simple ad hoc verbal agreement to let a third party make the decision or a more complex one contained in a written contract that also sets forth the rights that are in dispute. Arbitration is widely used in labor and commercial disputes in the United States, but also in less formal settings, such as the family when children take a quarrel to a parent for resolution.

Arbitration was not the procedure of first choice for a number of reasons. First, it requires both parties to give up power procedures entirely. It is therefore more difficult for a party believing itself to hold the balance of power to accept than the other procedures. Second, it is unlikely that the weaker party will accept and use arbitration unless it trusts the stronger one to comply with the arbitrator's directive. In real life, the courts are available to enforce arbitrators decisions, but litigation is impractical at this time for virtual communities such as MicroMUSE because of its high formality and transaction costs. In view of the superior, though not absolute, power position of administration and the lack of an enforcement agency, I decided not to recommend this procedure for the resolution of user-administration disputes. On the other hand, it offered promise for interplayer disputes if administration would agree to enforce the arbitrator's decision when necessary.

Another reason for favoring factfinding and mediation was that there were few civil rights set forth in the 1993 Charter. Rights would therefore have to be more fully elaborated by administration, jointly with the citizens, or by submission to an arbitrator before they could be adjudicated. Interests, on the other hand, can be discovered and reconciled through the other procedures.

Finally, the preliminary diagnosis of social disorder on MicroMUSE suggested that the most important initial objective of intervention was to interrupt the cycle of escalating conflict and improve relations between citizens and administration. Because mediation and, to some degree, factfinding reconcile interests, they are more effective at improving strained relations and preventing recurrence of conflict than is the rights-adjudication arbitration procedure. Thus, the distribution of power, the lack of a mutual understanding of citizen rights and the objectives of the intervention supported the introduction of these two procedures.

The Intervention

The essential qualification for dispute resolution practitioners, especially for mediators who depend upon personal power rather than authority, is acceptability to all parties of the dispute. Once a party loses confidence in the mediator, the neutral's usefulness is over. Arbitrators and factfinders also need acceptance because this enhances the likelihood that the parties will accept and abide by their findings and recommendations or decisions. Knowledge of the issues, skill with the process, impartiality and integrity contribute to acceptability, all of which the parties can best discover about the neutral from their own experience with him or her. Using an unknown neutral is risky business in real life, so much so that gaining entry to a relationship is often the most difficult task. In text-based virtual reality, where anonymity and the absence of nonverbal forms of communication impede development of the neutral's personal credibility, where there are no referring organizations and little, if any, experience with these tools, the entry stage is a significant challenge. This proved to be the case in MicroMUSE, where there were a number of false starts. Credibility of both the neutral and the process had to be built over a period of months. A turning point occurred in mid-December 1993 when a sister MUSE was broken into and laid waste in a particularly offensive manner. By this time I had established positive relations both with administrators and those suspected of the attack, so I was invited to observe the investigatory interviews. At the close of those interviews I was asked to write a factfinder's report of findings, opinion and recommendations, the latter of which were accepted by all parties. This case provided citizens and administration with experience by which they could evaluate the usefulness of third-party techniques in general and the neutral in particular. Submission of disputes to the newly-created Citizen Mediator Office soon became routine. The following cases illustrate the application of the two procedures that were introduced in 1994.

Home Author's Home Contents Table of Contents Replay Previous Section References

The Results

A Dispute Managed by Factfinding

In an early case, a user who was an elected member of the Citizens Council was amusing herself by stealing unlocked property. On one of her scavenging trips, she took an object from its owner who was connected at the time and witnessed the theft. The victim followed the thief to her lair, located her own property and uncovered a cache of numerous other objects belonging to others. The mediator was called in by an administrator to whom the crime had been reported. Two mediation sessions were held online, both of which were logged with the knowledge of the participants. The thief did not deny what she had done, but was otherwise uncooperative. At first she scoffed at the notion that it was a crime and blamed her victims for not locking their property up, but later she admitted her actions might have hurt some people. For her part, the victim said she felt both violated by the crime and foolish for not having taken better care of her property. She said she felt the thief not only owed her victims an apology, but should resign from Citizens Council and also help prevent future crimes of this sort by teaching a MUSE course about locks. The thief rejected this proposal, saying that what she did had no impact on her ability to perform her duties of office and that it was administration's responsibility to teach MUSE security. The mediator asked the thief if she had any ideas of her own for acts of restitution, but the thief said she preferred to submit to the will of the Advisory Board, a number of whom she claimed had been members of a previous gang of thieves. Mediation having failed, a report was written to the Advisory Board summarizing the facts and positions of the parties, and making recommendations. A committee was formed which reviewed the evidence and recommendations, and ultimately decided that the thief should be reduced in rank, which would make her ineligible to remain on the Citizens Council, and that she should make amends to the community by an act of her own choosing, subject to the approval of the committee. Eighteen months after this decision, no such act had been made and the thief remained reduced in rank. Although she was implicated in another case at about the same time as this one, there were no further charges filed against her. She continued to connect for some time, but eventually found other MUSEs more to her interest and abandoned her MicroMUSE character. The victim, however, reported that she felt she had regained self-respect when she confronted the thief, and was satisfied with the decision of the Advisory Board committee. Additionally, although the thief made no act of restitution such as would help protect the community from further burglaries, administration did adopt one of the factfinder's recommendations that an article be placed in the online news to remind users to lock their objects and report suspected thefts. The ad hoc committee and the process by which the case was handled became the model for the Disciplinary Committee, which was established in May 1994.

A Mediated Dispute

A later case involved one of the most charming characters on MicroMUSE, a Maas-Neotek robot named Caspian, who is programmed to wander the MUSE, making maps, giving directions, delivering messages, answering questions, and other similar things. One evening a bored user called "Pygmalion" spammed Caspian so badly he crashed. When Faceless, a staff member who maintains Caspian and other MUSE robots discovered this, he checked Pygmalion's record and found that he had been previously warned. He therefore knew that Pygmalion was aware of the prohibition against spamming and had probably agreed not to do it again. Faceless decided he needed to take measures to prevent Pygmalion's willful disregard of the rule again. He therefore arrested Pygmalion, retrieved a log of the spam as evidence, and sent the mediator a report of these actions.

Upon receiving the report, the mediator got in touch with the Pygmalion through email, told him he was free to represent himself before the Advisory Board's Disciplinary Committee if he wished, but offered the alternative of mediation, pointing out that the committee usually endorsed mutually-agreed to settlements. She also encouraged him to tell her his side of the story and to suggest a fair outcome. Pygmalion had several responses. First, he complained about the absence of the accused's rights. The mediator explained the reasons for the restrictions and arranged to have the unnecessary ones lifted. Pygmalion also admitted his guilt and told what had happened from his point of view, corroborating Faceless's version of events. He explained he had forgotten he was disobeying an order of an administrator and suggested that a just outcome would be the removal of his ability to use the command he used for spamming. His explanation led the mediator to question whether he truly understood why, from administration's point of view, spamming is an unfriendly act. She knew from past experience that administration would not be satisfied with another promise unless Pygmalion knew the reason for the rule as well as the rule itself, so she discussed administration's interests with him. She then asked him if he was willing to restate what he had learned by writing a brief explanation of why spamming is prohibited, especially of robots. Pygmalion agreed to do so. The mediator also informed him that it was not technically possible to remove his power to use the offending command without removing all his membership powers. The matter was left open until Faceless reacted to Pygmalion's explanation.

The mediator forwarded Pygmalion's written explanation to Faceless. He replied that the explanation was satisfactory and that if Pygmalion would make an act of restitution to Caspian and agree this was a last chance, the matter would be closed. Pygmalion accepted these terms, agreeing to write an apology to the robot and a statement to the effect that a third spam offense would subject him to exile. In one of MicroMUSE's more touching moments, Caspian visited Pygmalion in the jail where he listened to the letter of apology and accepted it, saying, "Never mind, Pygmalion," which is his programmed response to expressions of regret. On the strength of the apology and last-chance agreement, Faceless released Pygmalion. The Disciplinary Committee shortly thereafter accepted this mutually-agreed-to resolution. There were no further complaints against Pygmalion after this until he, too, abandoned his character.

Discussion

Outcomes. These cases illustrate the processes used to attempt reconciliation of diverse substantive and procedural interests with the primary objectives of improving relations and preventing the recurrence of conflict.

In the robot case, Pygmalion's substantive interest was to remain on MUSE (i.e., not have his character recycled) free to pursue his own objectives without undue restriction. He had due process issues of having his case expeditiously and fairly handled, and of having some of the arrest conditions relaxed so he could communicate with other users while his case was pending. For administration's part, its substantive interest was in preventing further abuse of the command queue and other users. It also wanted to prevent social and technical retaliation for perceived injustice and to resolve the case expeditiously. While it may have been faster and taken less effort in the short run simply to recycle the character or to let him languish in jail, none of Pygmalion's interests would have been addressed. Had it taken this power approach, administration would have risked acts of retaliation requiring even more disciplinary actions. In the long run, then, the more labor-intensive process of mediation, because it successfully reconciled all interests of the participants, furthered mutual understanding and respect, and changed the conduct of the miscreant, resulted in a lasting peace. Additionally, it modeled to one of its young users an effective means of managing social conflict and, therefore, was consistent with the educational mission of the MUSE.

In the theft case, once her property was returned, the victim's primary interest was to regain a sense of control and self-esteem. This was achieved in mediation when she confronted the thief onMUSE in real time and later when she was told the results of the Advisory Board's deliberations. Although her specific proposals for resolving the substantive issue of restitution were neither accepted by the thief nor adopted by the Advisory Board, the outcome was not dissimilar: rather than being taught by the thief, citizens received information about locks in the news item; rather than resigning, the thief was de facto removed from the Council by virtue of her reduction in rank; and the citizens received some protection from future thefts from the news article that contained the self-protection information and may have dissuaded potential burglars by reinforcing a behavioral expectation. Administration experienced the transaction costs of considering and acting on the factfinder's report, but these were mitigated by referring the case to a committee and relying on the neutral's findings, if not her entire set of recommendations. As with Pygmalion, there were no acts of retaliation or repetition, so administration's interest in conflict prevention was addressed. About the only goal of the thief that was achieved was her continued presence on MUSE, albeit at a reduced rank. Procedurally, she had requested a decision by the Advisory Board, perhaps hoping that her friends there would protect her. The integrity of the factfinding process was guarded, however, by appointing a committee composed exclusively of administrators without personal relations with the accused. Thus, the citizen interest in having a fair tribunal was addressed to some degree, although not in the peer-jury form discussed at the October 1993 town meeting. Finally, there was the matter of the thief's position on the Citizens Council. The factfinder recommended that her tenure on the Council be placed before that body, but this was rejected by the Advisory Board, which seemed concerned that this would set a bad precedent and create the expectation that the Council would have authority in future disciplinary matters. In rejecting the factfinder's recommendation, the Advisory Board signaled its unwillingness to cede decision-making authority either to a democratically-elected body or to an appointed professional. In conclusion, this case demonstrates the peace-making utility of third-party dispute resolution techniques even when the process does not completely reconcile all the interests of all the parties.

Role of Technology. It is difficult to capture in case descriptions the role played by the technology of cyberspace. Even transcripts of online sessions do not fully depict its influence. On the one hand, the frustrations experienced and time consumed can be significant. On the other hand, technology of the community and its environment also makes some things easier to accomplish than in real life. The following are some of the normal consequences of existing technology.

First, mediator/factfinder control of meetings and hearings is significantly impaired by parties connecting from different time zones with different real life schedules, and by the neutral's inability to isolate the meeting from real life interruptions, disrupted connections, the need of some users to relog periodically, and lag from various sources. Administrators can forcibly restrict a user's communications with others as well as inhibit player movement about the MUSE, but even directors are socially and technically powerless over extra-MUSE conditions. The effect of this is to delay the start of meetings, and to interrupt and prolong them, as participants fail to arrive at the appointed hour, leave and return, or leave and fail to return to finish the meeting. Comings and goings disrupt concentration and create the need to repeat or simply to wait until key players return, possibly hours or days later. Mediation is a labor- intensive activity requiring patience anyway, and these conditions make it more so.

People at different places in the Net experience different degrees of lag. Those who are lagging the most may be effectively shut out of a group meeting or even a dialogue because of their inability to contribute to the discussion in a timely fashion. Their interests, then, may never be addressed unless the mediator takes care to structure and pace the discussion or to supplement onMUSE synchronous communication with asynchronous email. Likewise, it is easy for one or two individuals to dominate a meeting by virtue of typing skill, a fast, reliable connection and much to say.9

The fact that messages are typed and read, rather than spoken and heard, has several effects. Different people may have the communication edge in this environment than those who do in real life. For example, those who have difficulty expressing themselves or understanding others through speech may be more communicative online while those who have difficulty reading, composing and typing may be less so.

Typing and lag interact to affect the sequencing of messages, such that one frequently tends to be one or more turns behind in a dialogue, thus one makes statements that appear to have ignored the other party's most recent message, which was itself typed and entered prematurely in an attempt to compensate for lag. Experienced users are accustomed to this and other vagaries of the Net. Most learn not to take offense, but it can be a source of substantive misunderstanding and hard feelings.

Another problem is the quality of the evidence available. Email addresses and logs can be falsified, objects easily destroyed, and passwords cracked. Frequently there are no witnesses to an incident, the complainant has been unable to log what occurred, and no administrator is available to retrieve the transient command log before the evidence disappears. The case then turns on credibility, which is extraordinarily problematic in text-based, multicultural virtual reality where there is an absence of nonverbal communication (except those described textually, either in words or emoticons, which are necessarily consciously chosen) and where meaning is so easily inaccurately conveyed and misconstrued, both unintentionally and deliberately. Nowhere is the problem of finding facts more evident than in the determination of who is responsible for a given act. In an environment where experimentation with identity and role-playing is the norm, deception and detection become a game for some. Under such circumstances it is easy to lose faith in one's evaluative ability and find deception everywhere. Shell (1995) argues that the computer medium itself eliminates much of the warmth of interpersonal contact, making it harder to gauge reactions to ideas under discussion, and that this could lead to misjudging commitments. While there is certainly that risk, my own experience, and that reported by Rheingold (1993), is that interpersonal contacts run the gamut from the detached and impersonal to the intensely passionate. Moreover, users tend to assume distinctive, individual virtual personalities, styles of discourse, and modes of expression. Just as in real life, the more familiar a person is with another, the easier it is to avoid errors due to individual differences (Ekman, 1983). Members of the virtual community who know their fellow users' personae should be less vulnerable than outsiders to misjudging their messages. Real life, face-to-face meetings with online acquaintances are helpful, too, in assessing one's ability to judge virtual human character, although the real character and personality may be quite different from the virtual.

The anonymity and physical separation of cyberspace supports social experimentation as well as explorations of identity and self. Reid (1994), Rheingold (1993) and others have noted the disinhibition effects on behavior in cyberspace. Being free from experiencing the effects of their behavior on others and free from fear of punishment, users find it easier to speak and act aggressively and dishonestly than they do in real life. This complicates discovery of the truth as well as contributing to the incidence of online conflict. But there is another side to disinhibition, and that is in its encouragement of truth-telling. The protection that physical distance and anonymity gives users supports intimacy and confidences, as well as aggression and deception. It is easier in virtual reality than in the real world to persuade people to confide their needs, desires and secrets. This facilitates reconciliation of differences and discovery of the facts of a particular case.

Other effects of the technology of cyberspace assist the dispute resolution process as well. For instance, although lag can be frustrating and lengthen the time it takes to complete an interview, it also gives one time to think, decide on courses of action and compose one's words. Shell (1995) reports that people may pay more attention to the substantive content of messages on computer screens than to the same content delivered verbally. One would expect such reflection and concentration to improve communication, relationships and problem-solving. At the least it affords some protection against faux pas. In addition, the absence of physical cues and nonverbals can be used by the mediator as well as by the principals, as one does not have to work so hard to overcome barriers of socio-economic differences or to control physical expression of one's own feelings. Another example is the technical ability to make a verbatim record of any interview without the distraction of recording devices, and after an event has occurred to recall and record what is in the buffer. This provides instant and accurate transcripts (logs) for later study or evidence, such as in the Pygmalion case. Finally, the ability to multiplex allows one to support, consult and educate others on private channels, by whispering, paging and through remote control devices called "puppets,"10 while simultaneously monitoring or participating in a joint meeting. Thus, caucuses and joint meetings can occur simultaneously as well as sequentially. Multiplexing makes dispute resolution more effective as well as more efficient, because one is more likely to have access to key people, important information and creative ideas during critical moments of the process.

A final case illustrates some of the mediation-enhancing effects of the technology. This case involved a member, whom I shall call "Oedipus," who wrote some code that was unfriendly to the community because it used more than his fair share of bandwidth. I did not observe the incident as it happened, but was able to observe both it and the ensuing confrontation between the member and staff complainants after the fact by virtue of a program that allows me to stay connected to the MUSE even though my own computer is turned off and I am away from my keyboard. The director who later stepped in to back up the complaining staff made a log of his meeting with the two and emailed it to me as evidence that Oedipus had promised to rewrite the offending code. I was thus able to analyze the situation directly, rather than indirectly through another's interpretation, before attempting to mediate the dispute when Oedipus failed to keep his promise. My working hypothesis was that Oedipus was unable to accept his own culpability, and that this made him hostile towards his accusers and resistant to examining and understanding his code. I also concluded that administration's long-term interests would not be satisfied unless Oedipus understood the particular coding problem well enough not to repeat his mistake in the future. These hypotheses were supported when I met separately with each side simultaneously by multiplexing (I spoke with Oedipus, who was detained in the Cyberion City Jail, while discussing the situation with administration on an official-only channel). My objective then became to help Oedipus achieve a degree of understanding satisfactory to administration and sufficient to make an informed commitment, should he choose to make a new one. I knew if I could do this, he would be released as rehabilitated and there would be a good probability he would honor his commitment.

Because of his lingering hostility towards NetRunner (the staff who originally had confronted him), Oedipus refused a joint meeting. However, it was important that NetRunner be well-satisfied with Oedipus's knowledge, so I decided not to rely on shuttle diplomacy, but to exploit the technology to create a pseudo face-to-face meeting. I had NetRunner himself explain the problem, but use me as his medium, paging me lines which I then repeated almost verbatim to Oedipus by pasting them unedited into my own "say" commands. NetRunner heard for himself what was said through a puppet I placed in the room where I met with Oedipus. He thus could verify the accuracy of what I said and could hear for himself Oedipus's questions and comments, adapting as he, himself, felt necessary to develop Oedipus's understanding.

The strategy was effective. Oedipus was more open to direction emanating from my character, even though the words were not mine, than directly from NetRunner because he could concentrate on the substantive message rather than on his feelings about the messenger. Additionally, because we were in the same virtual space at the same time, the transaction was conversationally interactive. Misunderstandings were immediately corrected and understandings instantly reinforced. This made the effort both effective and efficient. There were external benefits of this session as well, when its log was edited to remove identifying features and later circulated to the mentor email discussion group and used in a help file to teach others better coding and, by modeling, coaching techniques.

Home Author's Home Contents Table of Contents Replay Previous Section References References

Evaluation

This experiment was undertaken to assess the usefulness of real life peace-making tools for improving the adaptability of virtual communities. After the 1993 Swagger case and resultant rebellion, the Advisory Board established a Mediation Office to assist users with interpersonal conflicts and a Disciplinary Committee to review and act on cases of user misconduct. All accused users have the right to controvert and be heard and are not punished for so doing. While they are encouraged to use the mediator, who is familiar with and informs them about the process, they need not do so. If a case cannot be resolved through mediation to the mutual satisfaction of the disputants, it receives careful, systematic and thorough consideration by a body composed of administrators, none of whom are an immediate party to the complaint. Corrective actions are directed towards rehabilitation and restitution, not punishment. Information about cases is made available on a need-to-know basis onMUSE and off, and the community is kept informed about procedures and Disciplinary Committee actions through news articles circulated on the email list and by ftp, gopher and the web. Thus, conflict has become institutionalized through structures and procedures chosen to (1) reconcile the diverse interests of a pluralistic community, both among the community's users and between its users and its administrators, and (2) provide due process for users and administrators accused of misconduct. All of this serves to give expression to conflict, but in a way that is not destructive to the community. Nevertheless, there are indications of problems with the system.

Criteria for evaluating the effectiveness of due process systems include the fairness and timeliness of the settlement and the ease with which the process is utilized (Aram & Salipante, 1981). Taking the procedural criteria first, the Mediation Office acts in the role of ombudsman to educate disputants on and assist them with the process, which itself is not complicated. But for disputants to use the process, they must be aware of it. A case that came to the Office fully 18 months after mediation was introduced clearly demonstrates that the community is inadequately aware of its existence. In this case, a two-year user named "Cassio" felt himself to be slandered in an administrative decision and took his complaint to the director he held responsible. Obtaining no satisfaction and unaware of any other option save leaving the community, he next took his case to the public channel, as Swagger had done 21 months before, and complained bitterly about the director. Eventually the director tired of this and filed a formal complaint of harassment with the Mediation Office. When contacted by a mediator, the member expressed surprise that he had any recourse save the political one of raising public opinion or the market one of withdrawing his membership (Hirschman, 1970). He subsequently ceased his public criticism of the director, wrote a petition to appeal the administrative decision on the grounds that it was based on hearsay evidence and offered suggestions about how to make the community better aware of its conflict management options.

A second problem with the system is that cases seem to the users to take a very long time to be resolved. Some of this perception is a function of user expectations in a cyberworld of nearly instantaneous global communications and the inability of some to delay gratification on account of their immaturity. The criticism also has some basis in fact. As described above, existing technology and global usage prolong the process. Little can be done about this except to make users aware of their impact. Another source of the problem is the labor intensity of the process, the community's reliance on volunteer administrators who have real life commitments taking them offMUSE at times inconvenient for the community, and few users or administrators with the requisite skills, acceptability or desire to act as conciliators. Increasing community awareness of the program and its needs could help relieve the labor shortage, make community expectations more realistic, and improve actual and perceived timeliness.

Turning now to the substantive criterion, two factors appear to have an impact on the perceived fairness of the settlements. One of these, the evidentiary problem affecting the rationality of outcomes, is discussed above. The other factor is the perceived impartiality of the decision-maker. For the first nine months of my engagement as mediator, I was an unempowered citizen who nevertheless had influence with administration. This gave me credibility with other citizens who perceived me as one of them, but I had to deal with their skepticism about whether I was influential enough to get results. I later accepted a voting and technically powered position on the Board of Directors reasoning that this would both give citizens a powerful voice in MUSE governance (I had veto power) and afford them visible protection when they have grievances against administrators. It also provided me with a degree of independence because, under the ByLaws (1994), directors are not easily removed from office. This solution avoided the problems inherent with employee ombudsman systems wherein the employee advocate is viewed as being under the control of management by virtue of serving at its pleasure. However, it also raised the question of insidious co-optation. It is clear that many members of the community, especially new users, believe I am an enforcer and administrator, not an impartial third party. Because of this, I am not always trusted to mediate interplayer disputes where, as in the Scorpio and Hecuba cases, administration is thought to have an interest counter to the disputants.' Acceptability has again become a significant issue.

Additionally, unless users believe they will get a fair hearing when, like Cassio, their grievance is with administration, they will either leave (depriving the community of potentially positive contributions), suffer in silence (depriving the community of indicators of potential problems), or take matters into their own hands, as did Cassio, Swagger, Ajax and others (Hirschman, 1970). In other words, as in real life, unless the third party is viewed as being unaligned with and independent of the good will of one side or the other, the conflict management system will be ineffective. The techniques presently employed can continue to resolve many issue-oriented conflicts, particularly if the community is made better aware of the program, the number of mediator/factfinders is increased, and the Office as a whole listens to all sides, respects differences and generally attends to its reputation. But because the asymmetrical power structure of MicroMUSE is itself the underlying source of much of the conflict, there are some disputes that cannot be resolved without first changing that structure to provide members with more influence (Rapoport, 1974). Employing a tri-partite arbitration board (consisting of an administrator, an elected member representative, and a third party chosen by the other two) authorized to develop a bill of member rights and resolve interplayer disputes would be a step in that direction and one administration may be ready for now that it has had some positive experience with third-party dispute resolution. Such a step would also move the community toward a more adjudicatory society and less mediative one, an evolutionary step that has been observed in societies elsewhere (Thomas, 1984).

Home Author's Home Contents Table of Contents Replay Previous Section References References

Conclusion and Summary

This essay began with the observation of substantial social conflict in the subject virtual community that was, and still remains to a large degree, open to the Internet. It noted a number of features of this world that make conflict more likely and more difficult to manage than in real communities: wide cultural diversity; disparate interests, needs and expectations; the nature of electronic participation (anonymity, multiple avenues of entry, poor reliability of connections, and so forth); text-based communication; and power asymmetry among them. These features and their results are not unique to this community or even to MU*s, as Macduff (1994) made similar observations about email networks. Some features may be limited to synchronous- communication communities such as IRC and MU*s, to MUSEs, or to MicroMUSE in particular: the publicity attendant on MicroMUSE's success, for example, its choice of a mission at odds with perhaps the majority of the Internet's population attracted to such environments, and its decision to be an open-access community. Clearly, though, what MicroMUSE shares with so many other virtual communities and what is at the heart of most, if not all, of its internal disputes, are open boundaries and substantial social diversity, both in degrees uncommonly found in real life. The MicroMUSE experience would thus seem to have lessons for other communities in cyberspace.

This essay has also argued that cybercommunities, like any social system, must include diversity and find some way to integrate it. Because open cybercommunities are likely to be extremely diverse, managing the resultant inevitable conflict is an especially important task. The computer interface, the anarchy of the Net structure, and the power asymmetry of most virtual communities, though, make the task of conflict management especially difficult. If, in their attempts to control behavior, such communities drive out ideas by suppression or exclusion, or escalate into chaos as a consequence of power struggles, their life and purpose will be threatened. To avoid this they must not still the voices of their members, but give them expression. As Scott Peck (1987, p. 71) puts it, communities must not give up fighting, but learn to "fight gracefully." Finally, this essay has demonstrated that several tools from the real world can be adapted by virtual communities to promote the environment of respect and safety necessary for their vitality.

Home Author's Home Contents Table of Contents Replay Previous Section References References

Notes

Return to body.1Internet Relay Chat. A 'party line' network for chatting in real time.

Return to body.2A MUD (Multi-User Dungeon or Multi-User Dimension) is a real time chat forum with structure (Raymond, 1993, p. 287). MU* is a general class of these virtual worlds including MUDs and their progeny such as MUSHes, MUCKs and MUSEs. 

Return to body.3"MicroMUSE is chartered as an educational multi-user simulation environment (MUSE) and virtual community with preference toward educational content of a scientific and cultural nature. The MicroMUSE administration works towards the development of MUSE technology to enhance the exchange of ideas, the learning process, and the expression of creative writing for individuals of all ages and backgrounds." (MicroMUSE Charter, 1994)

Return to body.4Interested readers should consult the 1994 MicroMUSE Charter and, especially, the 1994 MicroMUSE ByLaws for details of the administrative structure in place following the reforms implemented during the term of this study.

Return to body.5A psychological analysis is beyond the scope of this paper, but it is worth noting that MicroMUSE's population includes adolescents with developmental issues, adults wrestling with unresolved personal issues, players diagnosed with more serious psychological disorders such as depression, attention deficit-hyperactivity disorder, and substance abuse, and others simply escaping from real life's pressures. Turkle (1995) describes how MU*s, which encourage projection and transference, and offer a place to play and experiment, provide a forum for enacting and working through psychological issues. The MU*, then, is "a medium for working with the materials of...life" (p. 188).

Return to body.6Belief that information should be freely distributed. A corollary is that barriers to the free flow of information are evil, thus, central authority and bureaucracy are to be mistrusted and barriers are to be breached. Levy discussed the hacker culture at length in a popular book originally published in 1984.

Return to body.7Communicating with the intent to insult or provoke. Also communicating "incessantly and/or rabidly" on some uninteresting topic or with a ridiculous attitude. (Raymond, 1993, p. 181).

Return to body.8On some MUSEs (but not MicroMUSE), there is a 1984 Big Brother quality to behavior control as administration logs all commands executed by players suspected of misdeeds.

Return to body.9This can be such a significant problem in large meetings where many are vying for the floor that code has been written into the server to allow a meeting's moderator technical control over attendees' ability to speak.

Return to body.10Objects programmed to relay all they see and hear to its owner. Puppets may also be controlled by their owner to move, say, pose, or do anything their owner can do. Thus, the owner can be in two virtual places at once, the room of the owner and the room of the puppet.

References

Aram, J. D. & Salipante, P. F., Jr. (1981). An evaluation of organizational due process in the resolution of employee/employer conflict. Academy of Management Review, 6, 200.

Brown, L. D. (1983). Managing conflict at organizational interfaces. Reading, MA: Addison Wesley.

Brown, L. D. (1995). Managing conflict among groups. In D. A. Kolb, J. S. Osland & I. M. Rubin (Eds.), The organizational behavior reader (6th ed., pp. 317-328). Englewood Cliffs, NJ: Prentice Hall.

Clodius, J. (1997, January). Creating a community of interest: "Self" and "other" on DragonMud. Paper presented at the combined winter conference on educational uses of MUDs, Teton Village, Jackson, WY. [On-line]. Available http://dragonmud.org/people/jen/mudshopiii.html.

Coser, L. A. (1956). The functions of social conflict. New York: Free Press.

Cox, T., Jr. (1991). The multicultural organization. Academy of Management Executive, 5(2), 34-47.

Dalton, G.W. (1971). Motivation and control in organizations. In G.W. Dalton & P.R. Lawrence (Eds.), Motivation and control in organization (pp. 1-35). Homewood, IL: Richard D. Irwin & Dorsey Press.

Deutsch, M. (1973). The resolution of conflict: Constructive and destructive processes. New Haven: Yale University Press.

Deutsch, M. (1995). Commentary: The constructive management of conflict: Developing the knowledge and crafting the practice. In B.B. Bunker, J.Z. Rubin & Associates, Conflict, cooperation, and justice: Essays inspired by the work of Morton Deutsch (pp. 123-129). San Francisco: Jossey-Bass.

Dibbell, J. (1993, December 21). A rape in cyberspace or how an evil clown, a Haitian trickster spirit, two wizards, and a cast of dozens turned a database into a society. Village Voice, 38(51). [On-line]. Available gopher: gopher.well.com Directory: Community.

Ekman, P. (1985). Telling lies: Clues to deceit in the marketplace, politics and marriage. New York: Norton.

Filley, A. C. (1975). Interpersonal conflict resolution. Glenview, IL: Scott, Foresman.

Gergen, K.J. (1991). The saturated self: Dilemmas of identity in contemporary life. Basic Books.

Hirschman, A. O. (1970). Exit, voice and loyalty: Responses to decline in firms, organizations, and states. Cambridge, MA: Harvard University Press.

Karnow, C. (1994). The encrypted self: Fleshing out the rights of electronic personalities. John Marshall Journal of Computer and Information Law, 13(1),1-16.

Kelly, K. and Rheingold, H. (1993, July/August). The dragon ate my homework.... Wired, 68-73.

Kolb, D. M. (1983). The mediators. Cambridge, MA: MIT Press.

Kort, B. (1995). Project based learning and communitas. [On-line]. Available gopher: cyberion.musenet.org Directory: MicroMUSE/EdNet Articles File: ednet.5.

Leslie, J. (1993, September). MUDroom. Atlantic Monthly, 272(3), 28-34.

Levy, S. (1994). Hackers: Heroes of the computer revolution. New York: Dell Publishing.

Macduff, I. (1994). Flames on the wires: Mediating from an electronic cottage. Negotiation Journal, 10(1), 5-15.

Martin, J. (1992). Cultures in organizations: Three perspectives. New York: Oxford University Press.

McGregor, D.M. (1957). The human side of enterprise. Adventures in Thought and Action: Proceedings of the Fifth Anniversary Convocation of the School of Industrial Management, M.I.T., Cambridge, MA, April 9, 1957. Cambridge, MA: Technology Press.

MicroMUSE charter [On-line]. (1991). Available gopher: cyberion.musenet.org Directory: MicroMUSE/ImportantDocuments File: Old 1991 Charter.

MicroMUSE by-laws [On-line]. (1994). Available FTP: 192.1.100.39 Directory: micromuse File: bylaws.

MicroMUSE charter [On-line]. (1994). Available FTP: 192.1.100.39 Directory: micromuse File: charter.

Morgan, G. (1986). Images of organization. Newbury Park, CA: Sage Publications.

Ostrom, E. (1990). Governing the commons: The evolution of institutions for collective action. Cambridge: Cambridge University Press.

Peck, M.S. (1987). The different drum: Community making and peace. New York: Simon & Schuster.

Pfeffer, J. (1981). Power in organizations. Boston: Pitman.

Presthus, R. (1978). The organizational society (2nd ed.). New York: St. Martin's Press.

Rapoport, A. (1974). Conflict in man-made environment. Baltimore: Penguin Books.

Raymond, E.S. (Compiler). (1993). New hacker's dictionary (rev. ed.). Cambridge, MA: MIT Press.

Reid, E. (in press). Hierarchy and power: Social control in cyberspace. In P. Kollock & M. Smith (Eds.), Communities in cyberspace. London: Routledge.

Reid, E. (1994). Cultural formations in text-based virtual realities [On-line]. Unpublished master's thesis, University of Melbourne, Melbourne, Australia. Available http://www.ee.mu.oz.au/papers/emr/index.html.

Rheingold, H. (1993). The virtual community: Homesteading on the electronic frontier. Reading, MA: Addison-Wesley.

Robbins, S. P. (1974). Managing organizational conflict: A nontraditionalist approach. Englewood Cliffs, NJ: Prentice-Hall.

Scime, R. (1994). <cyberville> and the spirit of community: Howard Rheingold-meet Amitai Etzioni. [On-line]. Available gopher: gopher.well.com Directory: Community File: cyberville.

Simmel, G. (1956). Conflict and the web of group affiliations (K. H. Wolff, Trans.). Glencoe, IL: Free Press.

Shell, G.R. (1995). Computer-assisted negotiation and mediation: Where we are and where we are going. Negotiation Journal, 11(2), 117-121.

Tannenbaum, A. (1965). Unions. In J.G. March (Ed.), Handbook of organizations (pp. 710-763). Chicago: Rand McNally.

Thomas, Darlene K. (1984). Dispute resolution from an anthropological perspective (Occasional Paper Number 84-1). Washington, DC: Society of Professionals in Dispute Resolution.

Turkle, S. (1995). Life on the screen: Identity in the age of the Internet. New York: Simon & Schuster.

Ury, W. L., Brett, J. M., & Goldberg, S. B. (1988). Getting disputes resolved: Designing systems to cut the costs of conflict. San Francisco: Jossey-Bass.

Home Author's Home Contents Table of Contents Top of Document Top of Document

Send mail Send mail to Anna DuVal Smith