Computer-Mediated Communication and The Public Sphere: A Critical Analysis


  • Lincoln Dahlberg

    1. Post-Doctoral Fellow in the Department of Sociology with School of Sociology, Social Work, and Social Policy at Massey University, New Zealand. His research involves a critical investigation of the practices and meanings surrounding Internet use, with particular focus on processes of democratization. His doctoral research explored the possibility of the Internet enhancing the public sphere and is being published as a series of conference and journal articles, one of which is the current paper. Dahlberg also teaches media studies at Victoria University, New Zealand, and is a member of the editorial board of New Zealand Sociology. He is currently co-editing a special issue of New Zealand Sociology on the topic of cultural studies in New Zealand.
    Search for more papers by this author


In recent times much has been said about the possibility that the two-way, decentralized communications of cyberspace can provide sites of rational-critical discourse autonomous from state and economic interests and thus extending the public sphere at large. In this paper the extent to which the Internet does in fact enhance the public sphere is evaluated. Online deliberative practices are compared with a normative model of the public sphere developed from the work of Jürgen Habermas. The evaluation proceeds at a general level, drawing upon more specific Internet research to provide a broad understanding of the democratic possibilities and limitations of the present Internet. The analysis shows that vibrant exchange of positions and rational critique does take place within many online fora. However, there are a number of factors limiting the expansion of the public sphere online. These factors include the increasing colonization of cyberspace by state and corporate interests, a deficit of reflexivity, a lack of respectful listening to others, the difficulty of verifying identity claims and information put forward, the exclusion of many from online political fora, and the domination of discourse by certain individuals and groups. The article concludes by calling for more focused Internet-democracy research to address these problems further, research for which the present paper provides a starting point.

Introduction: Approaching the Question of the Internet and the Public Sphere

Over the last decade many respected media researchers have pointed to the possibility that the Internet's decentralized communications can enhance the public sphere, the social sphere constituted by rational-critical discourse that enables the formation of public opinion through which official decision making can be held democratically accountable.1Douglas Kellner (1998), for instance, states that the Internet has “produced new public spheres and spaces for information, debate, and participation that contain the potential to invigorate democracy and to increase the dissemination of critical and progressive ideas.” Indeed, the thousands of conversations taking place on a myriad of diverse topics through Usenet groups, e-mail lists, chat rooms, or Web pages seem to attest to the expansion in cyberspace of the loose network of informal public discourse that has existed to various degrees in modern society. However, observations of cyber-discourse also show that “confrontation, misinformation, and insult … characterize many public forums on the Internet” (Katz, 1997, p. 190).

In this paper the question is raised as to whether online discourse is in fact extending the public sphere. To formulate this question, everyday online discourse will be compared to a normative conception of the public sphere. As well as enabling the evaluation of the extent to which the public sphere is being facilitated, the comparison between norm and practice will permit the identification of any factors inhibiting the expansion of rational-critical discourse through the Internet. A public sphere norm for this evaluation can be developed from Jürgen Habermas' theory of rational communication. Habermas is appropriate to the analysis because he offers the most systematic critical theory presently available of democratic communications. For Habermas the public sphere is constituted by moral-practical discourse, which is interaction oriented to resolving political problems.2 Habermas' analysis of communication reveals that every participant engaged in moral-practical discourse makes reference to a number of pragmatic presuppositions and thus to a set of normative conditions of the public sphere. This set of conditions can be summarized as follows:3

  • i. Autonomy from state and economic power. Discourse must be based on the concerns of citizens as a public rather than driven by the media of money and administrative power that facilitate the operations of the market and state.
  • ii. Exchange and critique of criticizable moral-practical validity claims. Rational-critical discourse involves engaging in reciprocal critique of normative positions that are provided with reasons and thus are criticizable, that is, open to critique rather than dogmatically asserted.
  • iii. Reflexivity. Participants must critically examine their cultural values, assumptions, and interests, as well as the larger social context.
  • iv. Ideal role-taking. Participants must attempt to understand the argument from the other's perspective. This requires a commitment to an ongoing dialogue with difference in which interlocutors respectfully listen to each other.
  • v. Sincerity. Each participant must make a sincere effort to make known all information, including their true intentions, interests, needs, and desires, as relevant to the particular problem under consideration.
  • vi. Discursive inclusion and equality. Every participant affected by the validity claims under consideration is equally entitled to introduce and question any assertion whatsoever. Inclusion can be limited by inequalities from outside of discourse - by formal or informal restrictions to access. It can also be limited by inequalities within discourse, where some dominate discourse and others struggle to get their voices heard.

These six requirements provide an analytical template by which to evaluate the claim that the Internet is enhancing and extending the public sphere of rational-critical deliberation. The idea is not to simply tick off whether each of these factors is satisfied by online discourse. Rather, these requirements are idealizations that offer a democratic standard by which we can gauge the democratic quality of discourse. They allow us to identify those factors enabling or inhibiting the advancement of the public sphere via cyberspace, and thus provide the starting point for improving discourse. But how can such an evaluation be undertaken given the vast amount of diverse activity taking place through many thousands of Internet sites? Many studies of online discourse, including cyber-democracy research, limit their focus to very specific aspects or sites of cyber-communications that can be closely scrutinized with particular social science research tools.4 Unfortunately, the findings of narrowly focused cyber-research are also quite limited. Given the many and complex dimensions of online discourse, generalizations about the relationship between the Internet and society at large cannot be made from research into single sites and/or by applying single methods. But such generalizations are necessary for determining how to develop stronger democratic arrangements. Broad understandings of Internet practices also add to the meaning and value of more focused studies by situating them within larger socio-cultural processes, enabling them to contribute further to the overall picture. Moreover, general analyses provide the basis for the more specific research projects, not only by providing the wider context but by highlighting the important questions requiring further investigation.

To undertake a general analysis and characterization of cyberspace in relation to the public sphere we must look at a whole range of online practices. However, given that cyberculture is so varied and my personal resources are so limited, I obviously cannot undertake such a general analysis directly using ‘primary data’. This is where more focused studies are invaluable. An array of Internet research, including content analyses, case study research, participant observations, survey materials, and interviews can be drawn upon and combined to enable the required assessment of those aspects either facilitating or retarding the development of the public sphere through the Internet.5

My task now is to undertake this general analysis by comparing cyber-communications with each of the six criteria of the public sphere conception in turn. The first requirement - autonomy from state and economic power - will establish if indeed we can speak of public online interactive spaces. The second requirement - reciprocal critique of validity claims - will determine whether such interactions are actually deliberative. The final four requirements will establish the quality of these deliberations.

The Six Requirements

Autonomy from state and economic power

State and corporate colonization of cyberspace threatens the autonomy of online public fora by replacing rational communication with instrumental rationality. State censorship of the Internet and online surveillance continues to threaten free speech and public interaction online, whether it be in the form of official blocks to access or hidden monitoring of messages. An even greater threat to public discursive spaces online may be coming from the increasing privatization and commercialization of cyberspace.6 The rapid commercialization of cyberspace and increasing control of Internet infrastructure and content by major corporate players, driven by the integration of media industries as exemplified by the massive Time Warner-America Online and Vodafone-Mannesmann mergers of 2000, is leading towards a consumer-oriented cyberspace that promises to either marginalize online public discourse or incorporate it within privatized and individualized forms of interaction: online commerce, entertainment, and business communication. It is becoming more and more difficult for non-commercial sites to compete for the attention of online participants. Even democratically-oriented Internet sites are increasingly being hosted or directly run by corporate ventures promoting an individualized consumer-oriented politics that allows politicians to sell their messages directly to citizens online without the mediation of public discourse.7

However, a diverse array of spaces of non-commercial, non-state controlled interaction continues to exist, and even multiply, while riding on the back of the privatized infrastructure. Thousands of people participate daily in a plethora of non-commercial online fora (through e-mail lists, Usenet groups, chat lines, and Web publishing) that are unaffiliated to any political party, interest group, or corporate concern. Some of this communication actually facilitates the growth and coordination of a global culture of resistance to the corporate takeover of cyberspace and of public life in general. The success of this culture in pushing back corporate colonization of everyday life can be seen in those online interactions that have been central to the derailment of the MAI and the partial shut-down of recent WTO, IMF, and other meetings of the transnational economic and political elite. Even more than the mass media, the Internet provides spaces that stimulate critical debate. While such spaces cannot be said to be fully autonomous since they are always situated in a complex relationship with corporate and state power, they can offer the basis for an expansion of public deliberation. Of course, we should not underestimate the problems and threats posed for public discourse by state and capitalist incursions into cyberspace. But full autonomy is not required before the online discourse that currently exists in restricted form can begin to contribute to the development of rational-critical discourse and the public sphere at large.

We have yet to establish, however, if in fact the observed online interaction takes the form of communication required in the public sphere. It may be, for example, that online communication lacks rational reflexivity and simply relies upon a realm of pre-discursive values. We need to evaluate the form of interaction taking place within these online spaces through comparison with the other five public sphere requirements. For a start, we must ask if online discourse actually follows the reciprocal structure of critique demanded by the public sphere conception.

Thematization and critique of critizable moral-practical validity claims

The ‘rhythm’ of computer-mediated communication (CMC), Kolb (1996, pp. 15–16) explains, encourages a pattern of discussion that clearly parallels the rational, dialogic form of conversation required within the public sphere. With CMC,

[t]opics get developed in several exchanges of shorter messages rather than in one exchange of long position statements…. Instead of long argumentative lines developed then presented all at once, we find point-for-point statements and rebuttals. The texts feel more like animated conversation …. [T]he rhythm of e-mail and mailing list exchange encourages opposing manifestos and summaries but also quick movement from what you just said to the arguments and presuppositions behind it. Positions get examined from a variety of angles, and there will be demand for backing on specific points. This makes e-mail a good medium for the kind of dialogue that Habermas speaks of, which demands justification for each speech act and inquires into the validity and sincerity of claims.

Kolb here speaks of e-mail and mailing lists, but the dialogic structure of CMC is similar on newsgroups, chat lines, Web fora, and other Internet media. Many thousands of individuals from around the globe are engaged in this sort of critical dialogue, putting forward and critiquing claims on every conceivable question on a myriad of online groups. This lively exchange of opinions has been documented by a number of cyberculture observers including Katz (1997), Hauben and Hauben (1997), and Rheingold (1993). Moreover, the critical-dialogic nature of much online interaction has been confirmed by more in-depth online research. Hill and Hughes (1998, p. 49, 58, 114) found, from their research of explicitly political Usenet groups and America Online (AOL) chat, that debate–“people with different opinions” clashing in “a battle of ideas”–constituted the majority of the content of these online interactions. The results of research into the actual extent to which dialogue takes place depend, of course, upon definitions and counting methods. Wilhelm's (1999) content analysis of 10 political newsgroups found that only about 20 percent of a random sample of 50 messages were actually directed to a previous message. In contrast, Rafaeli and Sudweeks' (1997) study of 44 randomly sampled bulletin boards found that more than 60 percent of 4322 messages responded to previous postings. Despite such discrepancies, it is clear that the medium facilitates the critical exchange required by the public sphere conception and that this exchange is taking place within many online spaces.

Furthermore, the expectation that exchange of validity claims be accompanied by reasoned justification is frequently fulfilled in online fora. As Katz (1997, pp. 49–50, 190–191) notes, online debate is often “buttressed by information from Web sites, published research, and archived data,” which points the way towards the development of “a more rational, less dogmatic approach to politics.” This observation is confirmed by Hill and Hughes (1998, p. 125) whose research found that ‘sourced’ information was being drawn upon in over 60% of the ‘political’ Usenet groups they studied.8Wilhelm's (1999, p.173) study backs this further, finding that three out of four postings on the Usenet and AOL political fora studied provide reasons justifying their statements.

It is clear then that the exchange of validity claims with reasons is taking place within many Internet fora. This conclusion establishes that a deliberative structure exists online. However, the deliberative quality of online exchanges has yet to be determined. Accordingly, the extent to which online debate approximates the four remaining requirements of the public sphere conception must now be considered.


At the core of the rational-critical discourse that constitutes the public sphere is reflexivity. Reflexivity, as used here, means the process of standing back from, critically reflecting upon, and changing one's position when faced by ‘the better argument’. Such a process is necessary in order to transform privately oriented individuals into publicly-oriented citizens. Unfortunately, reflexivity is difficult to detect given that it is largely an internal process and changes in people's positions take place over long periods of time. Despite such difficulties, we can gain some appreciation of the level of reflexivity by looking at the structure and content of online debate.

For various reasons, and particularly in comparison to face-to-face conversation, the form of Internet exchange may be seen as facilitating reflexivity. For example, the effort it takes to put forward arguments in written form, in comparison with spoken communication, often encourages participants to think more carefully about their positions. This reflection is aided by the record of exchanges often available to participants in online debate which allows careful consideration of the development of ongoing arguments. Furthermore, online interactions are largely asynchronous exchanges and thus provide participants with time for reflection before presenting their own contributions. However, Internet debates also have a number of characteristics that could be seen as retarding the operation of reflexivity, most notably the bite-sized postings often involved, the non-linear structure of conversations, and the rapidity of the exchanges.

The rhythm of CMC, as Kolb (1996) notes, is such that shorter points are exchanged rather than longer position statements. The length of online postings is restricted by what Millard (1997, p. 159) calls ‘chrono-economic stress,’ that is, by “the psycholinguistic effects of an online writer's awareness of the limits to the time, bandwidth, money, attention, and any other resources that he or she can devote to any given piece of discourse.” Bandwidth limitation discourages posting long articles. Cost restricts time spent online. Indeed, time itself may be the most limited resource people have and may deter their participation in online deliberations altogether given that it takes longer to develop an argument through CMC than through face-to-face conversations or via the phone. Closely related to time limitations is attention. Readers will often skip or skim through posts that contain more than a few lines of text. These resource scarcities encourage ‘bite size’ postings. With CMC, Arterton (1987, p. 173) argues, “loss of nuance and detail are traded-off for brevity.” However, short posts do not necessarily translate into a lack of reflexivity, just as the publication of long position statements cannot be equated with its presence. Short posts not only maximize inclusion by giving others more opportunity to ‘speak’ but can also contribute to sustained arguments that accumulate much thought and reflection. Furthermore, the point-for-point engagement of much CMC can encourage thorough reflection by helping interlocutors to look at multiple aspects of their positions.

A second possible inhibition to reflexivity is the non-linear form of much CMC. Discussions can multiply, mutate, diverge, branch out, and overlap, undermining the ability of participants to follow and reflect carefully upon their own positions. Again, however, such claims often rely upon a fallacious assumption, the assumption that reflection requires linearly structured argument. This is not so. Non-linear arguments, common in offline contexts, can provide different angles from which to look at otherwise ‘obvious truths.’ Furthermore, a digital record of exchanges is normally available and this helps participants follow and examine ongoing debates critically in their own time and in their own way.

A possibly greater inhibition to reflexivity arising from the form of CMC is the high speed at which exchanges often take place. Rapid exchange limits the time and space available for deliberating upon claims and critiques. Apart from the costs of being online (time and money), such rapid exchange is encouraged by a sense of urgency that often develops in online interactions due to the pressure to attract and keep the limited attention of other participants. This urgency tends to be greatest within synchronous modes of online communication (chat rooms and role playing domains) where lines of conversation may disappear off the screen within seconds of being posted and many people may be speaking at once. The pressure to converse at speed in these fora is further increased by the fact that reputations can be built simply on the ability to think and write fast. From their study of AOL chat, Hill and Hughes (1998, p. 30) concluded that

[c]hat rooms are a difficult format for thoughtful discussion. The short line space and the fast pace require people to make snap comments, not thoughtful ones. We see this in the low level of information and the small amount of issue discussion. Most chat room conversations appear to focus on the actions of people, not on the government and its role in society.

There is also pressure to reply quickly within asynchronous modes of online participation (Usenet groups, Web fora, and e-mail lists), especially given that discussions tend to be dynamic, beginning and ending rapidly and moving off in new directions without much warning. Moreover, participants often feel compelled to respond to messages at a rate that reflects the speed of the network. Knowing that the medium cannot generally be held responsible for the delay, people at times become annoyed with those who do not respond within a day or two of postings. Furthermore, Usenet messages must be read within a few days because they are only held on servers for a short time, and e-mail list messages need to be ‘dealt with’ or they result in unmanageably congested e-mail ‘in-boxes’.

This rapid turn-around rate may be seen as limiting the likelihood of participants' carefully considering and re-developing their positions. However, a quick response is not absolutely required, particularly within asynchronous modes of CMC. While the technology enables nearly instant communication and online discussions often seem to proceed very quickly, participants often take care and time over postings, learning from their own and other's embarrassing online experiences such as sending posts to the wrong person, sending grammatically and logically incoherent posts, and being pulled up for badly misrepresenting other participants' positions. Participants may also learn that greater attention and respect can be gained through the quality rather than the quantity of posts. Moreover, a quick response cannot be simply deemed unreflexive just as a slow response cannot be assumed to be reflexive. To get a better idea of the extent to which reflexivity is taking place within online deliberations we must turn to actual deliberations.

Reflexivity can be detected taking place within some online debates. Observers can witness times and places when “positions soften and change” (Katz, 1997, pp. 49–50). Many participants within online fora are conscious of the self-transformative and democratic possibilities of cyber-interaction. Some explicitly see themselves as having been transformed through such interaction (Rheingold, 1993; Turkle, 1995). These individuals and groups demonstrate that reflexive deliberation is not only possible but is taking place (to some extent) online. However, the level of reflexivity in most online fora falls well short of what is demanded by the notion of the public sphere. While observations confirm that many participants put forward and staunchly defend their own claims and vigorously critique other positions, few participants acknowledge the strength of criticisms directed towards them and even fewer seem moved to change or compromise their positions in the course of argumentation. Debates often turn into repetitive exchanges between dogmatic interlocutors unprepared to reconsider their positions. Usenet discussions, Streck (1998, p. 44) observes, “are woefully circular; discussion reduces to the same people saying the same things in the same ways.” Reflexivity needs to be increased in online discourse if the public sphere conception is to be more fully approximated. An important task in improving reflexivity is getting participants to listen respectfully to, and thus come to an understandings of, their interlocutor's position. What is required is ideal role-taking, which is the next public sphere condition that needs to be evaluated.

Ideal role-taking

Publicly-oriented discourse involves ideal role-taking, that is, participants attempting to put themselves in the position of the other so as to come to an understanding of the other's perspective. Such a process requires a commitment to an ongoing dialogue with difference in which interlocutors respectively listen to one another in spite of, and because of, their social and cultural differences. Cyberspace is a place where difference is not hard to find. The Internet as a whole is extremely plural. One can move between newsgroups, discussion lists, and chat lines, finding groups of every conceivable description. There is the opportunity online to encounter a much greater social diversity than is possible in most offline situations. Given this potential, what can be said about the level of respectful listening in online deliberative situations where difference is encountered? Moreover, what can be said about the extent to which individuals actually commit themselves to such situations and undertake the ongoing ideal role-taking required to secure understanding?

As already established, cyberspace is filled with vibrant exchanges of opinion. However, in general there tends to be far too much talk and not enough respectful listening. As Streck (1998) reports, a sense of democratic responsibility or duty to listen to others is rare within present online discourse. Cyberspace “on a day to day basis is about as interactive as a shouting match” (p. 45). At worst sites can be dogged by flaming, which is antithetical to respectful listening. Flaming has been attributed by some observers to the ‘disinhibiting effects’ of CMC - the bounds of constraint on expression being weakened in cyberspace with the elimination of many social context cues (Dutton, 1996). With the possibility of rapid response, single flames can quickly escalate into flame wars (Dery, 1994; Sproull & Faraj, 1997, p. 40). As Millard (1997, p. 148) notes,

online writing's unique balance between bodiless abstraction and dynamic public drama creates a rhythm of attacks and responses that is all but guaranteed to amplify disagreement into dissonance, disturbance, and finally … full-fledged dissing.

Hill and Hughes (1998, p. 59) measured the extent to which the explicitly political discussion groups that they studied deteriorated beyond a minimum level of acceptable democratic interaction by counting what they called ‘flame fests:’“to qualify as a flame-fest, the majority of messages in the thread must be flames.” Individual flame messages that do not manage to drag the whole thread down with them were not counted. Hill and Hughes (pp. 57–62, 115) found that 39% of Usenet debates and 32% of AOL chat could be defined as a ‘flame fest.’ In contrast, however, Rafaeli and Sudweeks (1997) concluded from their research on newsgroups (including Usenet) that “content on the net is less confrontational than is popularly believed.” Data reported in Dahlberg (2000a) suggests a level of flaming somewhat less than that found in Hill and Hughes9

The different levels of flaming found by researchers can be accounted for by different operationalizations of the phenomena and the great variation of discourse found across cyber-fora. However, even strongly regulated and democratically-oriented online deliberations can be susceptible to disruption by abusive postings. Santa Monica's PEN, often seen as a model of cyber-democracy, was affected by flaming to the extent that use of the system's conferencing facilities was abandoned (see Dutton, 1996, p. 275–278). Fortunately, ways of effectively dealing with flaming, particularly the elaboration of rules of discourse, which will be discussed further below, are increasingly being formulated and applied successfully.

A possibly more serious problem for productive deliberations is the ease at which participants can simply opt out of ideal role-taking and so avoid commitment to the ongoing and rather demanding process of attempting to come to an understanding of different points of view. Some commentators argue that meaningful commitment to online groups is very difficult to develop and sustain due to the lack of bodily markers in cyberspace. The loss of grounded identity perceived within cyber-interactions, Robins warns (1996, p.14), undermines real-world intersubjectivity: “(I)t is the continuity of grounded identity that underpins and underwrites moral obligation and commitment.”Sclove (1995, p. 79) likewise fears the abstraction of interaction from ‘Real Life’: “electronically mediated communication filters out and alters nuance, warmth, contextuality, and so on that seem important to fully human, morally engaged interaction.”Nguyen and Alexander (1996, p. 116) believe that cyber-interactions “lack the responsibility of an actual bodily commitment” and thus cyberspace “diminishes the range and quality of human encounters” needed for meaningful association. Lajorie (1996, p. 168) argues, in sum, that political collectivities as we presently know them are undermined by “the erosion of the local, the effacement of bodies, and the pure extension of interface.”

In contrast to these warnings, research about the Internet has generally shown that strong commitment to online groups is still possible, and can be observed, despite the hiddenness of bodies and extension of interaction over time and place.10 In fact, the lack of bodily markers can encourage people to become closer and lead to intimate and long-lasting relationships: “For many Net communicants it is the absence of a visual dimension that affords depth and an integrity to computer-mediated conversation” (Clark, 1997, p. 171). Participants may become committed to others in online groups to the extent that they support each other in birth, sickness, and death. Research has shown that such commitment takes place across the various modes of CMC, whether in e-mail lists, role playing domains (MUDs), newsgroups, or commercial conferencing systems like the WELL.11 One would expect the least attachment to occur in chat groups given their greater ephemerality.12 However, even with online chat participants tend to return to the same ‘rooms’ to look for the (virtual) persons with whom they have developed (virtual) bonds.

Clearly, virtual communities of mutual support develop online. There are now literally tens of thousands of virtual communities in cyberspace, flourishing through e-mail lists, electronic bulletin boards, online chat groups, and MUDs. However, the interaction taking place through these communities does not generally take the form of rational-critical debate required to extend the public sphere. Participants in these groups are not necessarily moved to offer criticizable validity claims or forced to critique their strongly held assumptions. There is little demand on participants to engage in ideal role-taking as these groups are relatively homogenous. Although they feature disagreements, virtual communities are often based upon people getting together with similar values, interests, and concerns in order to provide emotional support, companionship, and advice (Wellman & Gulia, 1999). The aim is to reinforce other group members, rather than evaluate their claims critically. Studies of online bulletin boards by Hill and Hughes (1998) and Wilhelm (1999) show that even groups focused upon (political) issues expected to involve diverse opinions often simply develop into ideologically homogeneous ‘communities of interest.’ Furthermore, those who browse the Web for information, who may be expected to encounter a myriad of challenging ideas on their journeys through cyberspace, tend to seek out information (often aided by information-gathering software) that simply supports their personal interests and values.13

Of course, as already noted, there are online spaces where people do encounter and take part in discourse that challenges them to rethink their positions. But typically there is little ideal role-taking involved in the arguments within these groups. In addition, when ideal role-taking does take place, it is often short-lived. In comparison with offline groups, participants can leave a discourse with relative ease if disagreements become too challenging to their values or demanding of their attention. On many occasions, critical dialogues do not get much further than a few exchanges before drying up. This observation is confirmed by Wilhelm's (1999, p. 174) study of AOL and Usenet groups oriented towards explicitly political issues, “which found that sustained dialogue among all participants on a single topic or line of inquiry is uncommon.”

In general, it can be concluded that online interaction is largely fragmented into groups of similar values and interests and, moreover, that in online debate which does involve difference, respectful listening is rarely maintained to the extent necessary for coming to understanding. Of course, it is unreasonable to expect that the major proportion of cyber-interaction will involve the rational-critical form of discourse of the public sphere, given the demands that such discourse makes on participants. However, to exploit online interaction more fully for the enhancement of the public sphere, there must be spaces where a broad range of problems arising from living with difference are confronted by interlocutors committed to understanding the position of the other. Few Internet fora presently involve any great proportion of such exchanges. More deliberative spaces are needed where participation involves ongoing, respectful listening if the public sphere is to be enhanced through the Internet to any significant extent. Furthermore, to be meaningful, such ideal role-taking must be accompanied by an openness or sincerity on the part of those putting forward claims and reasons. The extent to which this latter requirement is fulfilled in online discourse will be examined below.


For understanding and rational assessment of positions to be possible, discursive participants must make a sincere effort to make known all relevant information and their true intentions, interests, needs, and desires. Intentionally misleading others about one's claims, including relevant information about one's identity, undermines the whole deliberative process. To determine the quality of deliberation in online fora it is important to examine carefully the extent to which falsification of claims takes place, particularly given the degree of control of the presentation of self and information made possible by CMC.

Identity remains salient within cyberspace despite the hiddenness of bodily markers. Identity online develops via multiple signifiers. One's e-mail address signifies gender, ethnicity, institutional affiliation, nationality, location, and so on. Language and writing style may indicate class, culture, and gender.14 Nicknames are often gendered. The content of posts are riddled with identity cues including the posters' interests, positions on various issues, lifestyles, and offline relationships. The signature that can be attached to the end of posts operates as one of “the online world's most deliberate identity signals,” often used to “anchor the virtual persona to the real-world person” (Donath, 1999, p. 40). All these signifiers build up identity online.

These signifying mechanisms indicate that offline identity seeps into cyberspace, whether participants like it or not. But these signifiers also provide participants in cyberspace with quite a degree of control over self-presentation. This can encourage identity experimentation and the expression of parts of the self repressed in offline interactions (Danet, 1998, p. 131). Yet few online participants actually take up the opportunity to depart significantly in their online presentations from the faces they present offline (Kitchen, 1998, p. 83). Although performance of identity is an inescapable part of any interaction, explicit identity play is mostly confined to the peripheral zones of cyberspace, particularly MUDs and chat groups.15 Research by Rafaeli and Sudweeks (1997) confirms a high level of voluntary self-disclosure in online interactions. This explicit disclosure, together with the unconscious signifying of identity, indicates that the hiddenness of bodily identity markers in CMC does not necessarily undermine the public sphere requirement for sincerity.

Control over self-presentation also offers the opportunity for conscious deception of identity. Deception here refers to neither the routine (and inescapable) performance of identity online nor those cyber-sites where playing with identity, including taking on multiple personae, is an acceptable (and often central) part of interaction. Deception in deliberation occurs in situations in which a participant intentionally misleads others into believing that intentions, needs, desires, and interests have been honestly presented. This is quite a widespread problem in online discussion groups. Sometimes called trolling, identity deception in cyberspace aims to embarrass, anger, and disrupt. It is often undertaken merely for amusement, but is sometimes driven by more ‘serious’ motives including political goals.16 Trolls infiltrate a particular group by assuming the identity of a person who would pass as a “legitimate participant, sharing the group's common interests and concerns” (Donath, 1999, p. 45). A troll may even attempt to impersonate an existing participant within the group concerned. After developing their false identity and becoming accepted within a group, the troll sets about disrupting proceedings while trying to maintain his or her cover.

Such identity deception undermines the trust within online groups. Suspicion of a troll can lead to participants' re-evaluating how they post and how seriously they take other posters. Participants may become cautious about self-revelations and about believing the revelations of others. Participants may even decide to keep silent, or to withdraw altogether from the group concerned. In this way, trolls can cause a lot of damage to online deliberations. To safe-guard deliberations trolls must be exposed and ejected from groups as soon as possible. Of course, the risk of detection keeps ‘the game’ exciting for the troll. Furthermore, it can be difficult to detect well-disguised deception. However, the costs involved in trolling tend to deter many would-be trolls. Trolling requires a considerable expenditure of time and energy in order to control consistently for all the different signifiers of identity that make up a non-contradictory character. Moreover, the detection of deception can lead to a variety of punishments, from humiliation of perpetrators by means of online publicity to suspension from the group and even the conferencing system concerned.17 These impediments to trolling reduce identity deception. Unfortunately, any amount of identity deception aimed at misleading discursive participants can be a major disruption to deliberative proceedings and so needs to be curbed as far as possible.

Other aspects of information deception may be even more pervasive online than the intentional misrepresentation of identity. Many discussion groups, including those dedicated to ‘serious’ political issues, face the problem of postings aimed to misinform, embarrass, self-promote, provoke, gossip, trivialize, and so on. Discussions of the arrest of Malaysia's deputy PM Anwar Ibrahim in 1998 were dominated by sodomy jokes, racist slanging, and political personality bashing (Erickson, 1998, p. 44). Usenet discussions are particularly full of conspiracies, propaganda, rumor, and scandal. They can be worse than television talk shows because the posters do not have to reveal themselves and risk public humiliation for scandal-mongering. On the Web, documents, photographic evidence, and whole organizations can be readily fabricated. Warnick's (1998, p. 321) study of political parody Web sites utilising “uncorroborated sources and digitized pastiche to make their point” shows that “the Web environment lends itself particularly well to plagiarism, anonymous information, distortion, and digitized [distorted] images.” In 1996, conspiracy Web sites claiming the TWA flight 800 was destroyed by U.S. Navy missiles were so convincing that even the veteran newsman Pierre Salinger was duped (Selnow, 1998, p. 173).

Online information may be falsified simply as a ‘good prank’. At other times, some strong political motivation may lie behind misinformation online. The atrocities committed against Chinese-Indonesians during the Indonesian riots of 1998 gained increased attention and condemnation through Internet news reports, including postings showing pictures of rape victims and publishing victim's emotional testimonies. Unfortunately, unverifiable materials were posted. It seems that at least some of the material was intentionally falsified, “particularly photos that seemed to have been either downloaded from porn sites or were unrelated to the Jakarta incidents” (Erickson, 1998, p.46). It is hard to know if these reports were posted to stir up popular sentiment against the human rights abuses that had taken place or to undermine the credibility of the claims of abuse. Verifiable online evidence is often hardest to come by in cases where support for claims is most crucial. Web sites enabling the dissemination of Kosovo-Serbian news during the media blackout in Yugoslavia in March 1999 generally could not verify the sources of reports received, which to some extent undermined the power of the stories posted.

These verification problems can inhibit online interactions from realizing the deliberative conception where only ‘the force of better argument’ decides outcomes. The problem does not have to be debilitating. What we are faced with here is a variation of the problem of authentication of claims and supporting information in any deliberative situation. Just as with offline discussions, and as demanded by the deliberative conception, participants in online fora need to remain sceptical of unverifiable claims and information. This scepticism is evident in many online fora. Aware of the possibilities of fraud, participants often challenge any claims and supporting information that are not convincingly substantiated. Although it is sometimes a difficult task, claimants are expected to provide convincing support (from either offline or online sources) for their assertions before their positions become accepted by other participants. Survey results from The Pew Research Center (1999) show that online users, just as in offline situations, are quite confident in discriminating between accurate and inaccurate information, good and bad arguments. However, it cannot be guaranteed that participants within online deliberations will detect deception aimed at misleading them. The detection of fraudulent postings is hampered, among other things, by inequalities and exclusions that limit some individuals and groups participating in the critical evaluation of the claims under consideration. Such inequalities will now be considered.

Discursive Equality and Inclusion

Inclusion in online discourses is inhibited by social inequalities and cultural differences outside cyberspace. Despite formal accessibility and the rapid expansion of the Internet, much of the world's adult population still does not have any real chance of Internet access due to poverty, poor telecommunications infrastructures, and state censorship. When access is available, many people simply do not have the time, cultural capital, or community support to engage in online political deliberation.18 Given such exclusions, the Internet can at best support an elite public sphere. But what happens once people do manage to become involved in online discourse? Is inclusion and equality experienced within cyber-fora?

A common theme within Internet literature is that social hierarchies and power relations are levelled out by the ‘blindness’ of cyberspace to bodily identity, thus allowing people to interact as if they were equals. Arguments are said to be assessed by the value of the claims themselves and not the social position of the poster. Yet, as seen above, identity becomes just as salient online as offline. And the development of identity differences leads to the reassertion of authority and subsequently power differentials online.19 This re-constitution of authoritative power online needs to be explored in order to determine if and how it may limit realization of the requirement of discursive equality and inclusion.

Status develops online with every post, even when offline identities are not revealed. Factors contributing to the development of the reputations of participants in online fora include time spent online, displays of technical expertise, frequency of postings, consistency and reliability in reciprocating other posters' messages, being a group moderator, offering and effectively helping new users, and posting messages of the style and content admired by the group concerned (Donath, 1999, p. 31). Developing a positive reputation also builds authority, which in turn empowers certain individuals to do things online. The varying authoritative power that develops between individuals contributes towards a stratification of online populations along the lines of ‘cyberpower’ (Jordan, 1999; Streck, 1998). These differentials of power are linked to and extended by offline social hierarchies and identities. The ability of an individual to assert authority online is inextricably linked to the extent to which he or she possesses the resources (time, money, and skills) necessary for success in cyberculture. Moreover, online status is often directly reinforced by the revelation of offline identities that are, as seen above, readily brought into cyberspace.

This stratification of authoritative power means that some participants are able to make their voices heard more than others, which leads to the domination of discourse by certain individuals or groups and thus to discursive inequalities. Domination of discourse takes place in three ways: abusive postings, monopolization of attention, and the control of the agenda and style of discourse. The most rudimentary way in which some individuals or groups come to silence others is through abusive postings, postings that may not only involve the biting and sarcastic content that the term flaming encompasses but that may aim to belittle and humiliate others. Such abuse is often targeted at those with less power in cyberspace and predictably, given the ease with which offline identity and power are brought online, such targeting overlaps with those marginalized in offline discourse: new users, women, and non-white ethnic groups (Cherny & Weise, 1996; Tadmore-Shimony, 1995). Even when individual identities are unknown, racist, sexist, and other forms of abuse can be broadcast to thousands of people through e-mail lists, Usenet groups, and chat rooms. Such abuse can be extremely upsetting for participants of online proceedings and can lead to their silencing and even withdrawal from cyber-interactions.

Possibly more pervasive and damaging to egalitarian discourse is the monopolization of attention within online discourse by particular individuals and groups. In many groups a small number of participants are responsible for most posts while the majority of subscribers post infrequently or simply ‘lurk,’ reading messages without posting.20 However, an uneven distribution of the number and length of postings across a group's population does not necessarily indicate discursive inequalities and exclusions. There are many reasons for lurking or posting sporadically other than the fear of being put down or not being listened to.21 Furthermore, infrequent posters and lurkers are not necessarily unengaged, passive, or non-participating. They may be as committed, attentively listening, and reflexive as are frequent posters. Moreover, regular posters cannot necessarily be said to dominate discourse or to exclude other participants from posting. Theoretically, all subscribers to a group have equal opportunity to post. In addition, given the lack of bodily signifiers, an online discussion's cohesion and very existence needs regular posters who can always be relied upon to have an opinion and so keep conversation going. However, sometimes inconsiderate and noisy individuals do monopolize the very limited attention available in online discussions. As Watson (1997) notes in his study of a Usenet group, online interactions may be dominated by a few excessive posters who post for the sake of it, without having much to say. These posters tend to have the same identities (educated, white, English-speaking, and male) as those who frequently monopolize attention in offline conversation.

There are a number of ways to minimize abusive postings and the monopolization of attention. These will be considered below. More difficult to deal with, and thus more problematic for achieving egalitarian and inclusive discourse, is the qualitative domination of conversations by particular individuals and groups. Here, dominant participants may not necessarily be directly abusive or say more; rather, they assert their influence and sideline other participant's views by dictating the agenda and style of dialogue. The dominant voices are those who have developed online authority, and these again tend to be educated, white, English-speaking, and male.

The development of discursive inequality online can be illustrated by looking at the inequalities that result from the reproduction of gender online. Some feminists would argue that the Internet offers greater expression for women (Light, 1995). In particular, cyberfeminist polemics speak of online liberation through the transcending of real-world binary oppositions and essentialist gendering.22 Yet gender is the most readily reproduced difference in cyberspace. ‘RUMorF?’ and ‘stats?’ (the statistics that describe a person's offline physical appearance) are two of the most popular questions within chat groups. Gendered stereotypes tend to be reinforced even where participants have the opportunity to experiment with identity. Online identity play is notorious for reproducing offline identity, with participants regularly cross-gendering in terms of Ken and Barbi models (O'Brien, 1999). Gender is not merely reproduced but ‘intensified discursively’ resulting in ‘hyper-gendering’ (Hall, 1996; O'Brien, 1999).

Numerous inequalities within online discourse follow from this gendering of participants. Observations of online interactions suggest that although women in cyberspace (compared with offline) often find it easier to have their voices heard and that men may be more receptive, women also “still attract the unwanted attention of men, are still sexually harassed and receive abusive messages and are still expected to adopt the same gender roles as in real space” (Kitchen, 1998, pp. 68–69).23 Although there are many spaces and dialogues which do not involve such direct abuse, men still dominate many online interactions in terms of the quantity of postings and their influence over what is said and how it is said. In particular, as Susan Herring's (1993, 1996, 1999) extensive research on gender and CMC shows, gendered inequality in online discourse arises from the dominance of a ‘male style’ of interaction. The male style is characterized by messages that are longer and more frequent, issue-oriented, assertive, authoritative, adversarial, sarcastic, and self-promoting. The female-gendered style, on the other hand, tends to be shorter, personally-oriented, questioning, tentative, apologetic, and supportive. Research by Baym (1996), Savicki, et al. (1996a), Soukup (1999), and Sutton (1994) supports these findings. The agonistic male style tends to be accepted and even encouraged in many online groups. The dominance of this style tends to impede women's participation more than men's, given that women are on the whole less accustomed to and willing to engage in such forms of interaction. Tired of being pushed aside and intimidated within online discourse, many women leave, become passive observers, post self-censored messages for fear of reprisal, or start women-only groups (Brail, 1996; Sutton, 1996).

Such investigations of gender-based inequalities and exclusions demonstrate how cyber-discourse often fails to approximate the requirements of equality and inclusion. However, some hope for deliberative equality can be found within online discourse. For a start, the extent of inequalities within online interaction varies across cyberspace. There are numerous online deliberative fora in which there is no noticeable abuse or domination.24 Discursive inequalities may also diminish with the influx of women and other marginalised groups into cyberspace.

Furthermore, the increasing formalisation of netiquette (online rules of etiquette) may curb the worst excesses of online coercion.25 However, attempts to maintain a minimal level of reasonable interaction through either the enforcement of netiquette via participant self-regulation or official group moderators is often undermined by a combination of cleverly forged posts, users operating from rogue sites that do not maintain acceptable use policies, and anonymous postings (Pfaffenberger, 1996). Participants ejected from groups because of continual violations of discourse rules can return under a pseudonym and if necessary with a different Internet address. As a result, moderation may only have limited success. For instance, Hill and Hughes (1998) found that (weak) moderation of politically-oriented chats on AOL had little impact on flaming. Furthermore, a libertarian ‘free speech’ ethos permeates cyberculture to such an extent that the acceptable behavior norms of online fora often permit dominant posters, moderate hostility, and even harassment.26 Even when rules of discourse are such as to eliminate flaming, they tend to say or do little about the risk that the agenda and discursive style may become dominated by certain individuals and groups.

Discursive inequalities and exclusions continue to occur in online interaction, despite the so-called bracketing of identity, the development of netiquette, and moderation. Discursive inequalities and exclusions result from the uneven distribution of power in the wider society, a distribution that is reproduced in online relations. These inequalities may dissipate with time as more women and minorities come online and as the rules of discourse further develop. However, at present online discourse, including the very rules of discourse themselves, tend to be biased in favor of those individuals and groups that dominate offline discourse.

Online Discourse and the Public Sphere: Where to from here?

This paper has examined the claim that the Internet's interactive spaces are enhancing and extending the public sphere of rational-critical discourse as conceived by advocates of deliberative democracy. This examination has been carried out by comparing Internet practices, particularly focusing upon publicly-oriented online deliberative fora, with a model of the public sphere developed from Habermas' theory of democratic communication. This comparison has been undertaken at a general level so as to cover extensive areas of cyberculture, a broad-ranging analysis achieved by drawing upon a diversity of Internet research.

The evaluation shows that exchange and critique of political claims can be found to be taking place everyday on thousands of Usenet groups, e-mail lists, Web fora, chat groups, and through Web publishing. These communicative acts confirm that the Internet is facilitating discourse that replicates the basic structure of rational-critical debate and that in various ways approximates the requirements of the public sphere. As such, we can say that the Internet facilitates an expansion of the public sphere that is constituted whenever people enter into deliberation on political questions. However, as can be expected given the critical nature of the analysis, observations of cyber-discourse also show that the quality of such discourse falls short of the requirements of public sphere model. First, the increasing commodification of cyberspace threatens the autonomy of public interaction online. Second, reflexivity is often a very minimal part of cyber-deliberations. Third, many online fora experience a lack of respectful listening to others and minimal commitment to working with difference. Fourth, there is difficulty verifying identity claims and information put forward. Fifth, extensive exclusions from online fora result from social inequalities. Finally, discourse tends to be quantitatively and qualitatively dominated by certain individuals and groups.

The three most significant mechanisms for moving online discourse towards the desired form arise from the technology employed, the ‘rules of discourse’ instituted, and the type of forum management undertaken. In terms of the technology employed, the actual technological artifact (the hardware and software of the Internet) produces a bias towards some uses over others: it encourages certain forms of online interaction rather than others. For instance, synchronous chat-group software promotes more rapid exchange than asynchronous computer conferencing systems. It is thus important to choose, adapt, and develop technology that maximizes rational-critical discourse. Rules of discourse can also be used to structure online debate. For the purposes of enhancing the public sphere, these rules can be developed with reference to netiquette which, like the requirements of rational-critical discourse, has evolved from communicative practices. However, at its present stage of development netiquette does not directly map onto the requirements of rational-critical discourse. Netiquette contains a bias towards particular, agonistic forms of discourse. As such, netiquette needs to be modified so as to draw out rational-critical discourse from online interaction. The type of forum management employed can also have significant impact. In particular, forum management can shape discourse by the way it chooses to police the rules of communicative exchange and by taking a lead role in setting the tone of the discourse.

These solutions, along with others, need to be further explored and developed if the Internet's deliberative promise is to be more fully utilized. To do this it would be useful to look at online cyber-democracy initiatives that actively seek to structure online deliberation in the direction of the public sphere model. Such studies would not only help further specify limitations to rational-critical discourse faced by cyber-publics but would also help identify ways of overcoming the problems identified here.


  • 1

    See, for example, Aikens (1997), Barber (1998a, 1998b), Fernback (1997), Hauben and Hauben (1997), Kellner (1999), Moore (1999), Noveck (1999), Rheingold (1993), and Wilhelm (2000).

  • 2

    For Habermas, moral-practical discourse is just one form of problem-solving communication. He also refers to theoretical discourse that seeks truth in scientific investigation, aesthetic criticism that seeks beauty in artistic endeavors, and therapeutic critique that seeks understanding in self-reflection.

  • 3

    I developed this public sphere model in my doctoral thesis (Dahlberg, 2000a). I drew heavily from Habermas' theories of communicative action, discourse ethics, and deliberative democracy. See, in particular, Habermas (1984, pp. 1–26; 1989; 1990, pp. 43–115; 1996, pp. 267–387). Cooke's (1994) illuminating work on Habermas' idealizations of communicative rationality and Chamber's (1996) insights into the conditions of discourse have also been very useful. This is not the place to explore in depth Habermas' theory of public discourse or to provide a general assessment of the requirements of argumentation in democratic society. The focus here is specifically upon evaluating the possibility of cyber-interactions facilitating the Habermasian model of the public sphere. The formulation of a normative model of rational-critical discourse and the public sphere has been extensively examined elsewhere, particularly within the broader debate on deliberative democracy. See, for instance, contributions to the volumes Benhabib (1996a), Bohman and Rehg (1997), and Elster (1998).

  • 4

    See, for instance, the studies of Schneider (1997) and Wilhelm (1999).

  • 5

    A number of broad evaluations of the relationship between cyber-culture and society drawing upon an array of more focused Internet research have already been undertaken. See, for instance, Jordan (1999) and Kitchen (1998).

  • 6

    I have previously undertaken an assessment of Internet and the public sphere in the context of the commodification of cyberspace; See Dahlberg (1998). My analysis in particular draws upon a critical political economy approach. Recent useful work in this tradition and focusing upon the Internet includes McChesney (1999), McChesney, et al. (1998), and Schiller (1999).

  • 7

    For a brief survey on the expansion of this individualized and commercialized politics online see Dahlberg (2000b).

  • 8

    While the concern here is with developing a general picture of online discourse, it is important to specify whenever possible the particular discursive format (chat, Usenet, e-mail list, etc) under consideration because variations across online fora, and particularly between different discursive formats, can be quite marked. For example, Hill and Hughes (1998, p. 125) found that there was more than five times as much sourced information drawn upon in the Usenet groups than in the AOL chat rooms that they observed.

  • 9

    The observations I refer to here were undertaken for a doctoral thesis (Dahlberg, 2000a). Sites observed included the Usenet groups nz.politics, uk.politics.philosophy, uk.politics.electoral, mn.politics, us.politics, and over the period August 1997 to July 1998.

  • 10

    Studies and reports referred to here include Argyle (1996), Baym (1998), Correll (1995), Hafner (1997), Turkle (1995), Watson (1997), and Wellman and Gulia (1999).

  • 11

    See research referred to in footnote nine above.

  • 12

    Rheingold (1993, p. 178) describes this ephemerality well. ‘Chat systems lack the community memory of a BBS or conferencing system or MUD, where there is some record of what was said or done in your absence. Although words are written and broadcast (and thus can be electronically captured, duplicated, and redistributed by others), they aren't formally stored by the chat system. The discourse is ephemeral.’

  • 13

    The Pew Research Center (1999) found that “[n]early one-fifth of Internet users get customized news reports and an equal number receive e-mailed news.”

  • 14

    On the gendering of online communicative style see Herring (1993, 1996). On racial identity as a feature of many online interactions see Burkhalter (1999).

  • 15

    O'Brien (1999) usefully differentiates sites of ‘authentic fantasy’ from sites of ‘real authenticity’: sites where the intention is to ‘perform’ from sites where the intention is to ‘be’. The former are more often true of MUDs and IRCs, while the latter are much more the case with e-mail, e-mail-lists, Usenet, and on Web sites.

  • 16

    Donath's (1999) compelling study of identity deception on Usenet offers interesting examples of trolling and the damage it can inflict on cyber-interactions.

  • 17

    See Donath (1999) for a more extensive description of what is required to operate successfully as a troll.

  • 18

    Wilhelm (2000) provides a sophisticated exploration of the way in which social and cultural inequalities restrict participation in online deliberation and hence limit the possibility of the Internet extending the public sphere. Useful discussion and evaluation of the ‘information divide’ is also offered by various contributions in Loader (1998).

  • 19

    Power here is taken to be the capacity to transform a given set of circumstances (Giddens, 1979, pp. 88–94, 100–101). Authority, drawing upon Giddens (pp. 100–101), is one resource through which power is mediated, a resource that generates command over persons.

  • 20

    See, for instance, research reported by Kitchen (1998).

  • 21

    Reasons given for non-posting, from research reported by Kitchen (1998), include: reluctance to speak to unknown people, resistance to participate in a group formed prior to one's arrival, being unsure about disclosure of self, fear of participating wrongly or poorly expressing oneself, and fear of being evaluated and receiving negative criticism.

  • 22

    Sadie Plant (1997) exemplifies the cyberfeminist position.

  • 23

    For examples of such observations, see Adams (1996), Brail (1996), Hall (1996), Herring (1999), and Sutton (1996).

  • 24

    In contrast to their previous research, Savicki, et al. (1996b) found no significant correlation between the number of males in a cyber-group and the type of language used. Crowston and Kammerer (1998), from their research into gender style in CMC, report no difference in the way men and women respond to different styles online: men are equally affected by the ‘male style’. Witmer and Katzman's (1997) study, based on the content analysis of 3000 messages from 30 randomly chosen newsgroups and e-mail lists, found that although men tended to use more challenging language women were somewhat more likely to flame. These discrepancies provide a useful reminder of the caution needed in generalizing too far from studies of limited aspects of cyberspace.

  • 25

    This is despite the initial belief of many cyber-enthusiasts that cyber-interactions would develop towards ideal forms of discourse without the need for such rules.

  • 26

    Some self-regulated groups are torn between advocates of libertarian free-speech and those who want greater regulation and sanctioning. See, for instance, Pfaffenberger (1996).