31. "Face-to-Face and Computer-Mediated Communities, A Comparative Analysis" with Oren Etzioni. The Information Society Vol. 15, No. 4, (October-December 1999), p. 241-248.


Paper presented at Virtual Communities:
Eighth Annual Conference on Computers, Freedom, and Privacy
at The University of Texas

FACE-TO-FACE AND COMPUTER-MEDIATED COMMUNITIES, A COMPARATIVE ANALYSIS(1)

In this article we combine the perspectives of sociology and computer science to compare face-to-face (f2f) and computer-mediated communications (CMC) from the viewpoint of their respective abilities to form and sustain communities. We also identify a third type of community--a hybrid--that is based on a combination of face-to-face (f2f) and CMC (or off- and on-line) communications. The article thus in effect addresses an oft-asked question: Can virtual communities be "real," have the same basic qualities as f2f communities? The article is exploratory, because much of the necessary evidence has not yet been generated, and the relevant technologies are rapidly changing.

I. Community Defined

The comparison proceeds by pointing out that to form and sustain communities, certain conditions must be met. We examine whether f2f and CMC satisfy these conditions and whether one of the two kinds of modes of communication is more effective in this regard.

One difficulty confronted in studying the questions at hand is that various researchers define the dependent variable -- community -- differently. Hence, when some show that computer-mediated communications can provide for the formation and functioning of communities, and others demonstrate the opposite, they are not necessarily referring to the same phenomena. For instance, in writings about CMC systems, the term "community" is sometimes used to refer to tightly knit social groups, and in other occasions to signify aggregates of people who hardly know one another; sometimes it even is used to describe mere geographic places. For instance, Geocities (http://www.geocities.com) provides for "neighborhoods" which are purportedly designed to emulate real-world communities. The domain is divided into such areas as Hollywood for entertainment information, Rodeo Drive for shopping, or Wall Street for business. However, while Hollywood, Rodeo Drive, and Wall Street are places in which virtual communities may be formed, they cannot be assumed prima facie to exist.

This problem is not limited to the literature on CMC systems; it is also widely acknowledged in the writings about f2f communities. Robert Fowler dedicated an entire book to differing definitions thereof, and the confusion arising from the disparities.(2) Margaret Stacey even suggested that the term is so poorly defined that it should be abandoned.(3)

For the purposes of this article, we define "community" as having two attributes: first, a web of affect-laden relationships that encompasses a group of individuals--relationships that crisscross and reinforce one another, rather than simply a chain of one-on-one relationships. To save breath, this attribute will be referred to as bonding. Second, communities require a measure of commitment to a set of shared values, mores, meanings, and a shared historical identity--in short, a culture.

Community is defined as non-residential because many social entities that are not residential have many of the attributes of residential ones. A case in point many Jewish communities, whose members meet at a synagogue, Sunday school, festivities etc--even though they live in different residential areas. The same can be seen in gay communities and even on those campuses that do form communities although the faculty do not live next to one another.

If one follows the definition used here, one notes that there are many aggregations of people that do not qualify as communities. People who pass each other as strangers in bus terminals or railroad stations, or are all subject to an electronic broadcast, or "meet" for the first time in a chat room, are best referred to as aggregates or groups but not as communities, because they do not share social bonds and a culture. In any case, one should not conclude that because on-line aggregates lack attributes of off-line communities, that virtual communities are not--let alone cannot--be "real," that is, meet the prerequisites that are needed to form full-fledged communities. In the following analysis we compare off- and on-line aggregates to one another, and off- and on-line communities to one another.

It should also be noted that the two key attributes used here to define community are continuous and not dichotomous variables. Thus, both on- and off-line, a community may have either stronger or somewhat weaker bonds, and either an extensively shared culture or a less elaborate one. However, if both are low, there will exist no community but rather some other form of social entity. Some observers wonder if there are any examples of "real" communities on the Internet. There are reports of several such communities.(4) The extent to which they are full fledged communities, and which steps might be taken to make them stronger communities, will become clear as our argument unfolds.

We turn now to spell out the requirements of community building and inquire to what extent f2f or CMC can meet these specifications, and whether one can meet them more effectively.

II. Access

Access is defined as the ability to communicate, not in the sense of articulating a message but in the sense of being able to reach others. (Access is a prerequisite for communication but not tantamount to it. Thus, phone lines provide access; they make this form of communication possible among those who are hooked up. They do not though, per se, determine who will call whom and to what extent they will be utilized.)

To form and sustain communities, members require access to one another. All other things being equal, people who have a higher amount of access, as well as modes of access that encompass more of the people in a given aggregate, are more likely to form communities than those who have a lower level of access.(5) Indirect evidence to this effect can be seen in areas in which the level of f2f communications is declining, or has been low for a while; in such cases, community members tend to seek to increase communication in order to sustain their communities. They use various means to this effect, among them enhancing the safety of public spaces, increasing walkways and front porches, and holding more "community" meetings (e.g., PTA or town hall meetings). (For the purposes at hand, we ignore the fact that at very high levels of access this proposition may not hold; furthermore, we recognize that the proposition holds only for populations that are not highly heterogeneous, as well as other qualifications irrelevant to the comparative task at hand).

F2f contact among people within a given area provides opportunities for access. However, merely being in the same space does not necessarily engender f2f communication or community building, as many have observed about contact avoidance on New York City subways and the anti-communication culture of many car pools. For f2f contacts to provide for communications requires mores that legitimate and value communications.

One further notes that f2f communications often are built into other social occasions or relationships, such as meeting at a country shop, strolling on a promenade, serving together in a voluntary association, or attending adult education classes. Hence, no special arrangements or investment of resources, energy, or time are needed to ensure that the needed communications will take place (as long as the culture legitimizes initiating contact).

Computer-mediated communications enable people to communicate regularly without significant economic or other costs and without being in close proximity either spatially or temporally (CMC can be asynchronous). These communications evolve across both geographic borders and time zones, and they encompass individuals who are home-bound because of illness, age, handicap, or lack of social skills. They provide safety for people who seek to communicate but fear leaving home, a major consideration in many cities. And they can encompass a very large number of individuals.

In short, both f2f and CMC can provide the needed access. On-line communications seem to be superior in that they can reach more people, even those dispersed over large areas. Off-line communications benefit from the fact that they are built-in other social activities (for instance, having a drink in a neighborhood pub), and hence require fewer specific initiatives than on-line communications.

III. Encompassing Inter-Personal Knowledge: Identification, Authentication, and Accountability

Bonding, one of the two core elements of community, requires a high level of encompassing (versus specific) knowledge of the others with whom one bonds.(6) Accordingly, one would expect a group of individuals who meet for the first time--for instance, at a scientific meeting--to share only technical communications (like their findings and methods), to bond much less than a similar group of scientists who also share personal information (a recent divorce, the serious illness of a child, etc.) and discuss their personal feelings, formative experiences and life histories.

To gain encompassing knowledge of others, one needs (a) to anchor various different items of knowledge about those involved with specific identities and thus be able to compose a broad and inclusive images of others; (b) to trust that the communications from others are crudely correct, which often entails finding some way to authenticate some of the messages; and (c) to develop a sense that one is able to hold others accountable, that the members of the group are reasonably responsible. Thus, for instance, if a person promises to send a draft resolution or a letter of recommendation and does not, one is in a position to know which party failed, avoid dealing with that person in the future, and as such have a deterrent effect on irresponsible conduct. For communities to evolve and be sustained, they must be able to either marginalize or de facto exclude people who are found to be frequently irresponsible or otherwise anti-social.

F2f communication can meet these requirements with relatively few difficulties. Such communication relies on various personal identification markers, such as names, addresses, faces, and records. Thus, if one person protests strongly about the direction in which a town meeting is moving, most people present will know who that person is, an identity that often includes several relevant facts about that person. For instance, whether he or she is a major figure or a marginal one; whether he or she has good reasons to object or instead blows hot and cold over time, and so on. These factors are all taken into account in assessing this (and all other) responses.

While most off-line communities have difficulties in maintaining close personal knowledge of all members, especially as they grow in size, there is usually not more than one degree of separation between any two members, allowing A to ask B about C, as well as preserving records one can draw upon. Also, because communities tend to be homogeneous, it is often easier to empathize with others even if one does not know them personally.

Novels, plays, and movies are based on the highly unusual circumstances that arise when a person is able to acquire membership in the community without disclosing themselves or presenting a false-self. The Return of Martin Guerre is a fine example of this phenomenon.

In CMC, "handles" or "log-in IDs" are generally used in lieu of proper identification; even when names are given, disclosure is often very limited, and identity is rarely authenticated. As a result, there is ample documentation of presentations of false-selves,(7) role playing, gender swapping, and so on.(8) In many chat rooms "membership" is open-ended and uncontrolled. People join and drop out at will. Members of groups that have been chastised, have been reported to drop out and return with other log-in IDs.

These CMC features may well be valuable for several purposes, such as to vent feelings one is inclined to expose only to strangers one is never expecting to encounter again. Moreover, these features enable people to experiment with selves (and identities) other than their "true" one. But these features, on the face of it, do not promote interpersonal knowledge.

To the extent that CMC systems provide participants with features that enable them to readily activate authentication procedures, the more amenable these systems are to community building. Note that supplying the features routinely available, does not mean that participants will seek to use them and certainly not that they be required to do so, by some external authority.

In short, f2f communication systems are significantly superior to most CMC systems as far as identification, accountability, and authorization are concerned. However, it should be noted that there are no design difficulties in providing a much stronger basis for interpersonal knowledge in CMC systems. In fact, several forums already demand that participants use their true names, and others verify such claims. For example, MediaMOO, a MUD (Multi-User Dungeon, basically a CMC) for professionals, requires participants to provide their actual names, e-mail addresses, and background on their area of research.(9) MediaMOO's administrator verifies this information. Participation is limited to members, and controlled by passwords and once a person's identity is known, over time a CMC group will learn to know its members by associating various attributes (e.g., a tendency to "flame") and viewpoints (e.g., she/he is the most liberal/informed/responsible member of the group) with various individual members.

Additional features that could be added would further enhance the ability of on-line groups to associate validated personal information with various log-ins, to be activated if the members so choose. For instance, CMC mechanisms could ensure that when a member has the floor, a recent picture of that person will be flashed on the screen.

IV. Interactive Broadcasting

To form and sustain shared bonds and values, communities need to be able to (a) broadcast, i.e., send messages that reach many people simultaneously rather than point-cast to one person at a time, as is done in typical f2f conversation. Communities also need to (b) provide for feedback from those who are addressed by the said broadcast, including that from many recipients of the broadcast to many other recipients (communal feedback) and not merely from one participant to another (as in telephone conversations). We refer to this combination of needs as "interactive broadcasting."

One reason community building requires interactive broadcasting is that if there are only point-to-point communications, interpersonal bonding may develop (e.g. friendships, romantic involvement) but not the crisscross bonding that characterizes communities. The reason that broad-based feedback (that from many to many) is needed is to allow those who broadcast to take into account the reactions they engender in a summary and continuous manner. If those who broadcast are unable to take into account responses to their messages, even if not in real time, their ability to persuade their audiences will be hampered. And if the members of a group are addressed but unable to provide systematic feedback, they will tend to feel propagandized rather than persuaded. Mutual persuasion is needed for members of a community, who start from divergent viewpoints on a specific issue, to be able to develop a shared position.

F2f communication generally provides rather limited opportunities for interactive broadcasting, especially if the number of people involved is large, or they are dispersed over a large area. The needs of interactive broadcasting are met to a considerable extent, but far from fully, in town meetings. When these take place, a person addresses all those present (meeting the first prerequisite) and it is possible for those present to respond in a way that both the original broadcaster(s), and all those that have been initially addressed, hear the responses (meeting the second prerequisite).

Note that in such town meetings(10) responses are not necessarily monolithic. For instance, some parts of the audience may applaud while others hiss, some may chuckle appreciatively while the rest frown, and some people may nod their heads approvingly while others look pained or even stream out of the room. All these cues allow those who broadcast to adjust the message, based on their "reading" of various segments of the audience, in a community-building manner. They may better explain their point, moderate it, even retract, and so on, all in response to the reactions they are receiving.

While town meetings might well be the closest a community can come to meeting the need for interactive broadcasting, opportunities for interactive broadcasting are limited even in town meetings by the fact that often only a fraction of the members of a community attend. Time limits on such meetings further restrict the potential for interactive broadcasting. Such meetings rarely last more than a few hours and cannot be reconvened too often.

Another form of f2f communication systems that communities rely upon is the cobbling together, over time, of various discussions of subgroups that may take place in pubs, country stores, PTA meetings and over the back fence. This cobbling usually takes place informally, without any particular institutionalized arrangements. This informal piecing together is discussed below. It suffices to say here that it provides a measure of interactive broadcasting as people report to one subgroup what other subgroups were told and what their initial and closing reactions were. It is though a rather slow, and somewhat cumbersome, process whose results are difficult to predict.

CMC systems readily meet one of the interactive broadcasting requirements: they readily enable broadcasting messages to all members of a group simultaneously. However, so far most CMC have no features that allow meeting the other requirements. In most cases, a person "addressing" a CMC-based community will have no information while he or she "speaks" about the reaction of the members of the community (especially in real time), and the members who receive the broadcast will have no knowledge about how others react to the same broadcast. However, this defect can be remedied and is not a congenital trait of CMC systems.

One of the authors of this essay conducted a lab experiment in which ten participants who were listening in ten isolated booths to a broadcast speaker could--while the person was speaking--tilt a lever to the left for agreement or the right for disagreement. The more they tilted to the right, the stronger an amber light was shining, and each of the participants had such a light, so the speaker as well as all others could determine how members of the group were reacting while the broadcast was taking place.(11) These devices so far are available mainly in social science labs and in theaters that test audience reaction to commercials and movies, but not yet on CMC. Moreover, as long as such devices provide only a summary feedback, they do not allow reading of sub-audiences and changes within them as well as changes in intensity of commitment. The light may shine equally brightly if some of the members are strongly won over, although the number of supporters dwindles. In short, this mode of feedback is typically highly condensed and insufficient.

The ability to record and transmit audio and video across the Internet in real time could form one basis for much richer feedback transmission. A person broadcasting to a group could benefit from serial scanning of individual faces or view several simulations on a split screen, or even allow for a summary of noise feedback, while others in the audience may receive the same feedback (that is, they will be able to see and hear how others in the audience react).

A more serious design challenge for CMC is to provide for communal feedback that will be segmented rather than monolithic and will enable the separation of various sub-audience responses. (For instance, a person who has the floor may seek to know how young people in the audience respond vs. the older ones, or minorities vs. whites, or those who have kids in school vs. others and so on.) This might be relatively easy to do for survey responses but more difficult for rendering audio and video attachments to capture sub-group and not merely aggregate responses.

In short, CMC are much superior broadcasting systems to f2f systems. However, neither system satisfies the need for feedback, communal sharing of responses, and segmented signals. While f2f communications meet this requirement relatively well in town meetings, these are rare. While CMC can be adapted to meet this requirement, this has not been done so far on more than an experimental basis.

V. Breakout and Reassemble (B&R)

Communities differ from intimate, close-knit groups of friends and families (including most extended families) in that they encompass a larger number of members. The larger the number, the more difficult it is for all members to participate actively in dialogues that maintain shared values or for members to bond with one another. For this reason, many Israeli kibbutzim strongly opposed increasing the membership beyond several hundred members. (They were horrified when, for the first time, one exceeded a thousand members.)

Large groups that seek to maintain a high level of dialogue provide structured opportunities for breakout groups, in which members meet in subgroups and then share their conclusions with the larger group by asking a representative of each breakout group to briefly address the reassembled plenary group. Such B&R systems are not considered as effective as scenarios in which all members participated in one small group (because then all those present could receive and respond to all communication), but are considered more effective than if all participate only in one large group and no breakout is provided.

F2f communication systems sometimes provide for such B&R. This is the case, for instance, during conferences of members of large voluntary associations. Here, all those seated around a table are asked to discuss the issues at hand with one another before the discussion is returned to those who address the whole hall. (Because of the resulting noise these sessions are sometimes referred to as "buzz" sessions.) In other conferences, the breakout groups retreat to separate meeting rooms before they reassemble. Much more informally, in many communities B&R occurs when subgroups of members meet in the country store, pubs, etc., and discuss subjects to be raised at a forthcoming town meeting or referendum, and then informally share with one another what they concluded in other such rump meetings.

While it seems that only very few CMC systems meet this requirement so far, they could readily provide for B&R sessions by adding this feature to the common chat room or news group format. Many chat rooms already enable any two participants to break out, hold a private discussion and return to the larger group; however, these amount to point-to-point dialogues that lead to interpersonal relationships or business deals, but not to community building (which could occur if the breakout consisted of sub-groups rather than pairs).

Group breakouts from chat groups do occur when sub-lists are created from a list and broadcasts are tailored to the sub-lists. Still, the prerequisites of a full-fledged B&R system are not met because those who engage in sub-list dialogue do not "reassemble" the full group for reports or final community-wide deliberations.

There seem to be no insurmountable difficulties that would prevent one from designing a full-fledged B&R system on-line. All such a design would require is a predesignated schedule for breakout and reassembly. For example, if 120 people participate in a given chat room, they must be able to determine at the outset, perhaps by a vote based on preprogrammed options, that after a given time period, the group will break into 20 sub-lists for a given time period, at the end of which the full 120-member forum will reopen. Each of the smaller (20-member) groups would select a representative who would report each sub-group's deliberations after which general chat would resume.(12)

Variations on the system could include enabling the participants to choose which sub-group to join, to vote on the length of the total meeting, and to decide the proportion of time spent in breakouts vs. the plenary.

In short, f2f allows B&R systems, needed to build consensus in a large group, to build this consensus into ongoing communal activities. They also do so in meetings of various voluntary associations and other social groups. There seem to be no great difficulties in allowing CMC systems to meet the same requirements, allowing large groups to communicate in sub-groups and reassembling in larger ones to share the deliberations of the constituent sub-groups. So far, however, B&R opportunities are not routinely provided.

VI. Cooling-off Mechanisms and Civility

Community dialogues, a major source of sharing a culture, seem to function most effectively when delay loops are built into the communication systems, and seem to function most poorly when they take place in real time. The designers of democracy, particularly the founding fathers of American democracy, were greatly concerned about the rule of the mob. They feared that demagogues would whip up irrational emotions and drive people to demand unreasonable policies and actions from their leaders. Democracy requires mechanisms for keeping passions in check and allowing reason to prevail; the House of Lords in the United Kingdom and the Senate in the United States are said to serve as such checks on the more populist lower houses. More recently, teledemocracy, in which the electorate is exposed to a message and asked to vote immediately, as suggested by Buckminster Fuller and Ross Perot, raises the same concerns regarding electronic voting and polling of public opinion.(13)

Most recently this issue has been faced when concerns have been expressed about the loss of civility and the rise in emotive tensions and polarization both in the public (especially under the influence of hosts of talk radio shows such as Rush Limbaugh and Howard Stern) and among elected officials. In short, constructive community dialogues and democratic government are believed to require cooling-off mechanisms. Effective cooling-off mechanisms provide delay loops, time intervals between receiving a message and sending a response, and use this lapsed time for dialogues that cross (and mute) previous divisions.

F2f communication systems have built-in informal delay loops due to the constant nature of informal communication and gossip; furthermore, decisions often are deferred from one town meeting to another, or resolutions are reconsidered in subsequent meetings. Also, as townspeople meet, groups are regrouped so that dialogues occur among different combinations of people rather than only within one subgroup (to the exclusion of others). Both systems allow passions to cool as one learns about the objections and concerns of others, of people he or she is bonded with and of those whose views he or she hence respects. The passions of the moment have a chance to cool off. While these informal mechanisms do not always suffice, especially when community bonds are weak or the community is in conflict with another, all f2f communities possess some cooling-off mechanisms.

Some CMC systems contain a modicum of cooling-off potential. When people compose a message on e-mail and order it to be sent, the program verifies with them if the command is definite. This allows for at least a short moment of reflection. America On-Line enables its users to compose mail and arrange for it to be mailed later, which allows one to reformulate messages composed in the heat of the moment.

Design changes could readily provide for much more effective cooling-off mechanisms. For instance, when one participates in certain community forums in which highly charged issues are deliberated, responses could be delayed automatically for 24 hours, during which one can pull them back before they can be retrieved by others, somewhat like being able to get a letter back from the post office after mailing it in a moment of impulse. (This design would differ from the America On-Line feature in that it would be an agreed communal feature rather than subject to individual impulse.) Most importantly, an agreed time interval could be set between the initial broadcasting of a message and the tallying of votes, rather than allowing instant responses.

In short, both f2f and CMC systems already have some built-in delay loops that enhance civility and provide time for cooling the emotions aroused by the initial messages, deliberation, and occasions for dialogues among members of the community, before a final decision is rendered. However, the systems seem inadequate because currently they have at best rather meager delay loops. It seems that one could readily provide CMC systems with stronger civility options. (Measures that would enhance civility of f2f have not been discussed but are often explored when the need to promote a civil society is studied.)

VII. Memory

The communal sharing of culture never starts from a tabula rasa. The process of sharing value draws on a prior sharing of history, communal identity, experiences, and rituals. Hence, the need for a communal memory.

F2f communities use for such memory various devices including elder talk, the town library, archives and records, minutes, monuments, and grave stones. These devices are much more effective in recalling and reconstructing a community's history, past normative commitments, shared tragedies and celebrations than they are in recalling cognitive materials. These devices are often rather unreliable, which makes it easier to reconstruct the past in line with contemporary needs, but provides a poor foundation for logical and empirical endeavors.

CMC systems already provide for very powerful memory and retrieval systems. For instance, many representatives and senators maintain web sites in which constituents can access full archives of their speeches and voting records. CMC systems are much better at retrieving cognitive information, such as the text of resolutions previously passed by the town's council, past budgets, and earlier voting results, than at evoking the communal past.

In short, memories produced by f2f systems may be vague and unreliable, but they often are highly evocative and help forge a shared culture out of a varied past. Even without any additional increases in capacity and design modifications, CMC systems are vastly more effective than f2f systems in maintaining memory and retrieving it for communal purposes, as far as cognitive memories are concerned.

VIII. Mixed Communication Systems

So far we have seen that when it comes to satisfying the communication needs of communities, it is a mistake to assume that f2f communications are greatly more effective than CMC systems or that it is difficult to form and sustain communities on-line. Thus, while f2f systems are better at providing encompassing knowledge and B&R systems than CMC systems, the opposite is true for retrieval of cognitive information that has been generated earlier. Moreover, in some areas, the differences are of degree rather than substantial. For instance, both B&R systems are rather primitive, although those of f2f are somewhat more effective. When all is said and done, the specific profiles of CMC systems obviously differs from those of f2f ones, but this should not be taken to indicate that CMC are unable to meet the basic community criteria of bonding and sharing values. It follows that it is more productive to compare one type of community to another type and see their relative merits and demerits, than it is to ask what virtual communities lack compared to "real" ones.

Most importantly, we know from other comparative studies the faulty thinking that arises when one compares one type of system to another and overlooks the existence of hybrid systems. One case will stand for many other examples that could be given. Numerous studies originally compared traditional teaching to instruction on television (ITV). It was hypothesized that ITV would be superior to f2f teaching because ITV could use the best teachers, film lessons again and again until a very strong presentation was achieved, and allow the students to view the tape at their own pace (as many times as they liked, at the times most suitable to them, etc.) Others feared that the lack of personal involvement by the teachers on ITV would negatively effect the students' motivation to study, suppress their ability to absorb material, and prevent them from asking questions. It turned out that the best instructional systems are person-machine combinations in which teachers provide students with small discussion sessions that follow viewing of a teaching tape.

If one follows the same line of analysis, one would expect that communities that combine both f2f and CMC systems would be able to bond better and share values more effectively than communities that rely upon only one or the other mode of communication. While there is a very rich body of experience, experimentation and literature on person-machine relations when it comes to computers, there is not much to draw on when one seeks to analyze mixed f2f-CMC systems in the practice of community building. One hence draws on informal observations and reports.

One kind of hybrid system known by many in professional or public life or in business is when groups of people meet in person and form some measure of shared understanding, then maintain bonding through CMC. There are also ample informal reports about people who bonded on a CMC--and then met f2f, although these reports are largely about dyads rather than larger groups.(14)

Hybrids allow the special strengths of each system to make up for weaknesses of the other. Thus, hybrid systems--drawing on their CMC-have better memory capabilities than sheer f2f systems, and--drawing on their f2f links--a higher level of interpersonal knowledge than exclusive CMC systems.

One must expect that hybrid systems will not be all of one kind; a large number of different combinations can be readily conceived. One variation stands out: the extent to which a f2f community is also "wired." For instance, f2f town meetings could not be followed with additional deliberations in a CMC forum if only half of the members had access to it. And bonding at PTA meetings would not benefit (and might be even be set back) if only half of the parents (most likely the more affluent ones) had access to a CMC parent forum. Setting terminals in numerous public spaces, from the community library to shopping centers, may mitigate this problem, but will hardly overcome it.(15)

Another major consideration for a hybrid system is when its CMC component reaches people who are very geographically dispersed; it is much more difficult for these people to develop f2f communications than when people are concentrated in a relatively small and contiguous area.

The various deficiencies of CMC and f2f systems stand as challenges to designers to create the kind of system that has the highest potential for bonding and evolving a shared culture, that best catalyzes the building of genuine communities. Far from finding that CMC systems cannot meet the needs of "real" communities, we find that there are no conceptual reasons or technical ones, that CMC based communities, especially given additional technical development, could not become full fledged communities. Finally, we suggest that both f2f and CMC systems have strengths and weaknesses of their own, and that their proper combination promises to meet more of the prerequisites of community than either of them could separately.

For additional information visit the Communitarian Network's home page.

ENDNOTES

1. A very brief discussion of some of the points in this article has appeared in Science magazine, in an article entitled "Communities: Virtual vs. Real" (Science, 18 July 1997, 295). The authors are indebted to Peter Rubin for research assistance and comments on a previous draft of this article.

2. See Robert Booth Fowler, The Dance with Community: The Contemporary Debate in American Political Thought. Lawrence: University of Kansas Press, 1991.

3. Cited by Colin Bell and Howard Newby in Community Studies: An Introduction to the Sociology of the Local Community (New York: Praeger, 1973), 49.

4. See, e.g., Steve Jones, Cybersociety: Computer Mediated Communication and Community (Newbury Park, CA: Sage, 1995), especially p. 138. The book includes reports of people attending funerals, weddings, etc.--basically, the sort of things people do in f2f communities--in a CMC. The WELL (Whole Earth Electronic Links), founded in cyberspace in 1985, is also considered a vibrant community.

5. Bernard Berelson and Gary A. Steiner, Human Behavior: An Inventory of Scientific Findings (New York: Harcourt, Brace & World, Inc., 1964).

6. For more information on this variable (which is labeled "diffuse"), see Talcott Parsons, The Social System (Glencoe, Ill.: Free Press, 1951).

7. Sherry Turkle, Life On the Screen: Identity in the Age of the Internet (New York: Simon and Schuster Trade, 1995).

8. See, for example, Julian Dibbell, "A Rape in Cyberspace," The Village Voice, 21 December 1993, 36-42.

9. See http://lcs.www.media.mit.edu/people/asb/MediaMOO/.

10. Jane Mansbridge, Beyond Adversary Democracy (New York: BasicBooks, 1980), especially Chapter 5, "The Town Meeting." Also Amitai Etzioni, The New Golden Rule: Community and Morality in a Democratic Society (New York: Basic Books, 1996), 127.

11. Amitai Etzioni, "An Engineer-Social Science Team at Work," Technology Review January 1975, 26-31.

12. See Amitai Etzioni, "Minerva: An Electronic Town Hall," Policy Sciences 3 (December 1972), 457-474.

13. Amitai Etzioni, "Teledemocracy," Atlantic Monthly, October 1992, 36-39.

14. William J. Mitchell, author of City of Bits, is quoted as having stated "the more electronic communication expands and diversifies our circles of contacts, the more we are going to want to add the dimension of face to face" (New York Times, 25 February 1997, A12). This may well be true, but it is at least as valid to suggest that those who met in f2f groups are keen to stay in touch via the Internet.

15. For a preliminary discussion of these issues, see a report concerning a community that is 80 percent wired and maintains public access terminals, in Andrew Michael Cohill and Andrea Lee Kavanaugh, eds. Community Networks: Lessons from Blacksburg, Virginia (Boston: Artech House, 1997). Also see The Chronicle of Higher Education, January 17, 1997, p 24.

The Communitarian Network
2130 H Street, NW, Suite 703
Washington, DC 20052
202.994.6118
comnet@gwu.edu