Anonymity, Structuration, and the “Wild West”
Food for Thought
Use of “wild west” metaphors was fairly common early on the the development of the Internet and its component networks. There was, this metaphor asserted, no law on the Internet and research suggested that users of interactive computer media tended to abandon such social niceties as politeness, consideration of others, and other well established norms of behavior (Keisler's “flaming”).
Of course where some saw "lawlessness", others saw opportunities in, among other things, forms of anonymous interaction that are all but impossible in face to face interaction. An Internet user can pretend to be almost anyone, allowing experimentation with self-presentation in a relatively low risk environment. If a guy wants to experience something what it is like to act and be treated like a gal (or vice versa), they can simply adopt a female (or male) persona and interact as one in chat rooms, forums, moos, interactive games, or other social environments on the net. While the interaction may not be entirely realistic, it provides at least a taste of what it means to walk in another's shoes. Advocates of the enabling whistle blowing and a more widespread freedom of speech and interaction (especially in countries like China that restrict the free exchange of information) see other additional value in systems that enable anonymity. Indeed, some see anonymity as an ultimate expression of privacy online.
This treasure of this anonymity comes with its share of trash, as can be seen in the many quasi-anonymous e-mail spam programs. A properly designed e-mail broadcaster can assume any identity, which is one of the reasons why I receive so much spam that is sent by (or so the header says) me. When anybody can say anything without consequence, there are always at least some number of people who are willing to say almost anything. Slander, always difficult to prove and often damaging to the target of the slander, becomes almost impossible to track down to a source, never mind prove. False news stories become easy to plant and difficult, and sometimes all but impossible, to track down and counter.The kinds of antisocial behavior associated with flaming become, in effect, consequence free.
The “wild west” metaphor isn't often applied to the Internet anymore. Both users and serious scholars of Internet media seem to be increasingly unaware of even of the idea of “flaming”. The most important of several reasons for this is that that as “wild wests” are settled, the rule of law inevitably follows. This was well understood at the 1992 National Science Foundation panel and 1993 symposium on “The Rights and Responsibilities of Participants in Networked communities. I argued, based on a long term ethnographic observation of the evolution of computer conferencing in IBM, that participants in Internet media would negotiate and enforce rules based on their own self-interest. Richard Godwin, then working for the Electronic Frontiers Foundation (EFF) argued emphatically that, from the perspective of law, nothing new was needed. His point was straightforward: if you call someone a pedophile on the Internet they can still sue you and win. No new law, Godwin insisted, was needed.
Before going further, it should be noted that these are two very different answers to the general question. The first says that regardless of law, people structurate the social environments that they interact in. They create rules to govern behavior and establish social mechanisms for enforcing those rules. The mechanisms may be as simple as pointing out the socially unacceptable behavior (this often works) and can escalate through group interventions to making someone feel unwelcome to generally ignoring the individual to blocking that individual. Not every computer-mediated communication system supports this last of these options, but the rest are not only under the control of participants, but reflect similar strategies of socialization and enforcement that can be observed in non-Internet media, including face to face communication. Law (legislation and the administrative rules that are created based on that legislation) is generally unnecessary at this level. It is only when problems occur that require more extreme measures or encompass concerns beyond those of an immediate community of users that legislative bodies need become involved.
Godwin's concerns were at this “higher” or more encompassing level, and it is hard to disagree with his fundamental argument. The internet media of 1992 weren't breaking new ground relative to the wide array of broader concerns that were already the subject of legislation, administrative rules, and judicial tort, at least in the United States, which at that time still claimed far more Internet users than the rest of the world combined. It was easy to see limitations in Godwin's assertion even then, but those limitations didn't relate to the need for the U.S. Government to enact legislation so much as to the increasingly global character of the Internet, which make it difficult for any government to restrict behavior in any serious way. Inevitably, such legislation has been passed anyway, and much of it, including the Digital Millennium Copyright Act, has focused more on the narrow interests of a small group rather than any real problem associated with the Internet.
Anonymity on the Internet increasingly challenges both of these assertions. Godwin's premise is challenged by the problem of enforcing existing law when you can't identify those who have violated it. To be clear, there are many Internet contexts where this is not a problem, Even where they offer the appearance of anonymity with anonymous IDs and profiles, most Internet contexts have collected and verified at least some information that can be tied back to, if not an individual, certainly one or more locations on the internet and the machines they have used. Often, the profile of the individual is much more detailed, including mailing address, credit card, and other unique information. Maintaining the reasonable appearance of maintaining anonymity is important to many sites, however, and getting that information will often require a court order. While Yahoo, for instance, advises posters to their forums that users should not assume that their posts are truly anonymous, they also vigorously and successfully resist such court orders. If, following Godwin's example, a party who feels they have been slandered cannot learn the identity of their accuser, there is no real possibility of defending ones reputation under existing law.
The threat anonymity to an online community's ability to structurate is less obvious, but perhaps more problematic. Communities are built on a complex of social intangibles, including the willingness of individuals to participate in, take responsibility for, and develop some level of confidence in, respect for others within that community. Communities not only create (socially construct) rules, but actively enforce them, sometimes using extreme measures like social embarrassment, imprisonment, exclusion, and even killing. Imprisonment and killing are (most would say fortunately) unavailable in distributed online communities, but social embarrassment and exclusion can be, at least in an environment in which identity is reasonably unambiguous. Anonymity, however, allows individuals to bypass such enforcement, and in so doing undermines the ability of a community to structurate around community developed norms.
It is relatively easy to find instances on Internet discussion boards where a “troll” (the choice of term and its application is an evaluative bid for enforcement) ignores attempts by a community to bring the troll's behavior under control. Its not that boards lack enforcement tools. Yahoo groups (which I will use as a primary example through the rest of this essay), for instance, provides a mechanism that allows users to evaluate posts on a one to five star system (one is star is bad and often invisible to most users; five is great and suggests that something is worth reading) and a complaint mechanism that, if successful (it often isn't) can lead to a user ID being blocked for hours, days, or forever. Beyond that, users have the ability to ignore disruptive posters (“Don't feed the trolls” is a popular encapsulation of this idea).
The reality of anonymity, however, is that all of these actions are relatively toothless. Individuals that communities have agreed are “trolls” have continued to post on some groups for over a year, despite being substantially ignored (creating many new topics that draw no responses), almost all of which have drawn one star ratings from multiple posters. If this seems benign when considered in the abstract, consider that these individuals sometimes create dozens of new topics in a day, with the effect of pushing other discussion off the front page, and that many of their posts actively denigrate the community. The effect is not unlike heckling. It disrupts other conversation and encourages hostility. A heckler at a town or school board meeting might initially be tolerated and encouraged to express themselves productively, but will eventually be forceably removed from the room if they continue. On Internet boards there is frequently no way to stop the heckling of a determined individual.
Even if there was, however, anonymity makes it relatively easy for an ostracized individual (or any other individual) to assume another identity. It isn't necessarily easy to identify an individual who is posting under multiple anonymous user Ids. Even when they can be tied together and posted from interchangeably from the same logon (something Yahoo groups allows), there are very few obvious clues to a dual identity other than posting errors that tie two Ids together, recurring “Platonic” dialogs in which the Ids have conversations with one another and, if used, common profile elements (a shared avatar, for instance). While a dialog between two or more Ids controlled by the same person can be very entertaining, especially if the individual has a good sense of characterization and narrative (as Plato did), they also potentially bias the dialog and undermine attempts by an online community to self-structure.
The idea of using multiple ID's this way is hardly new. In 1983, several years before the Internet is created, Orson Scott Card creates a prescient vision of online interaction that involves the use of multiple identities controlled by the same individual in the classic Science Fiction novel “Ender's Game”. In one of the major subplots, the central character's brother exploits multiple identities over a global Internet-like network. These dueling online identities are used to explore issues of the day, with one ID often used to create compelling straw man arguments that the other ID proceeded to exploit. The character exploits these ID's to develop a reputation, create a global political base, and ultimately to achieve great power.
This sort of second ID is sometimes described today as a “sock puppet”. Unlike the novel, in which the second ID is formally owned by a compliant sister who acts as a willing co-conspirator, it is relatively easy for an individual to own almost any number of such Ids. Yahoo, for instance, formally allows the creation of multiple “profiles”, each associated with a distinct name, but all operated from the same login ID. When posting on Yahoo bulletin boards, you simply select the profile that you want to post from. It is fairly trivial, in this system, to create a conversation with yourself that appears to involve several different people. If you want to shed the minimal telltale signs that ID's are common, one need only create a second logon ID. The overhead of logging out and logging in can be minimized by using multiple tabs in a web browser, multiple web browers (e.g. Firefox, Opera, IE), or multiple computers.
The use of such ID's can be fairly benign. Exploring identity by assuming different ones is entirely valid, but other, less benign uses, can also be observed. A second ID can be used, for instance, to frustrate community sentiment towards a “troll”. The second ID, in this scenario, can seed a posting from the first ID with a high rating, thus preventing the community of users from undermining the post with a relatively small number of one star votes. An individual with four ID's can seed that post with a visible five star rating. Second ID's can be used to simply post a straw man question which the first ID then answers, but it can also be used to create replica and parody ID's. Replica ID's, or ghosts, closely duplicate another ID though either character substitution (substituting a “1” for an “l”, an “O” for a “0”, or a “vv” for a “w”, for instance) or a normally truncated suffix (using “myuniqueidisnot” as a duplicate for “myuniqueid”, for instance). Parody ID's attempt to make fun of some alleged aspect of another user that is under attack. The probably work best when a parody posting style has already been established by a replica ID.
If, as asserted above, communities are built on a complex of social intangibles, these uses of multiple anonymous user ID's which appear to be different people, but are actually the same person, have the potential to undermine those intangibles. Replica and parody ID's can be used to mislead people, thus undermining confidence and respect for others. It can be difficult to convince others that you didn't say something when it has been read by others as your comment. It can be difficult to reach the kinds of near social consensus (not votes taken, but a general decision still made) when a single individual can disagree from several different ID's using multiple arguments. Willingness to participate is undermined, for at least some people, when they can be attacked by replica ID's, the ability to take responsibility is actively undermined when individuals use multiple ID's to game the rating system.
“Flaming” has hardly disappeared from online interaction (or from marital or legislative interaction), but students and users of online systems are increasingly unaware of the term. Foul language, name calling, ribald jokes, and indications of relative intolerance of those with divergent views are easy to find on many online discussion systems. But it is also easy to find active discussion groups and systems on which such behavior is all but unheard of. The apparent demographics of these groups suggest that discussions that cater to women, mature adults, and serious subject matters tend to see less flaming. Discussions that are more attractive to males, teenagers and young adults, and contentious subject matters appear to see less. Structuration appears to be a reality in online discussion groups, and different groups structurate differently.
It is unclear what effect, if any, anonymity has on “flaming” behavior, but it seems clear that current deprecation of flaming as a meaningful term for online users reflects the success of online communities in the social construction of social norms and expectancies. Flaming behavior is clearly more common among those that communities identify as trolls; less common among identities that are treated as useful ongoing sources of information; more common in replica and parody Ids and those that make posts that attract one star ratings and few, if any, responses. The ongoings patterns of abuse tat these behaviors represent suggest that however successful communities may be in structurating expectations of behavior, anonymity appears to limit the ability of online communities to enforce those expectations.
In the end, these observations are not suggestive rather than conclusive. Anonymity is not, of itself, an inherently bad thing. There are times when it useful, as it is when individuals explore identity, and even critically important, as it is in whistle blowing and the collection of consequence free feedback. Structuration is not necessarily an inherently a good thing. There are always costs to the individual, and all to often the community, when social norms are defined so narrowly that dissent and free expression are discouraged . It remains that structuration of social environments is as inevitable as the interaction that leads to it, and perhaps as inevitable as the disruptive behavior that social norms attempt to control. There is a necessary balance between individual expression and community expectations.
There is nothing new about this. Creators of content must balance creativity against social expectations if they want to find an audience. Overbalance towards generic content and consumers will find the content predictable and uninteresting. Overbalance towards creative content and consumers will find the content confusing. There is a fine line between creative content and generic content. Most successful content contains a sizable measure of each. Producers of interactive online environments must find a similar balance when shaping and maintaining the “spirit” or “message” of their medium. A successful interactive medium depends on the formation of a core set of users who come back, at least in part, because they know what to expect and gratified when those expectations are satisfied. Users of any medium can be expected to construct social norms and expectations as they attempt to maximize such gratifications. They also also come back, at least in part, because they find the space useful. Usefulness is often found in the unpredictable creative content.
Hence there is a delicate balance that needs to be maintained in the design, implementation, and administration of any interactive communication system. Creativity needs to be encouraged, but structuration needs to supported. A “wild west” is, in the end, simply an inadequate structurated social space. Anonymity can thwart such structuration, but it can also enable creativity. It can enhance experimentation, but can also be used to irresponsibly attack others. It can allow whistle blowing, but creates problems in vetting the information that is provided. It is, in short a language that can not only be used to express truth without fear of retribution, but to lie with impunity. Anonymity operates in the gray area between the right to express oneself and the responsibility, to ones ideas as much as to others, to be heard.
Food for thought.
Davis Foulger
Do you have a response to Davis’s pundits? Have something you would like to ponder in this column? Please submit your responses or Food for thought toto John Howard at howardjo@ecu.edu