The "information highway" makes possible unprecedented forms of mediated communications. Never before have there been means of communication which have provided so many individuals with the ease and ability to engage in instantaneous, interactive communications with a broad and diverse public. The import of these new technologies, however, lies not simply in the novel forms of communication. What is also of profound significance is the ways in which these communicative possibilities lead to new types of identities and social relationships. Previously unimaginable, millions of individuals are increasingly interacting across time and space, and forming mutual bonds with others, most of whom they neither have met nor will ever meet face-to-face.

A striking characteristic of these new interactive encounters is the general potential for anonymity. In fora of face-to-face dialogue, individuals must confront one another and generally reveal their identities, or at least their appearances. On the telephone, conversants can hear one another's voices, a characteristic that often reveals, for example, gender, nationality, [1] or age, and can easily detect emotions, such as fear, rage, or delight. Online, however, participants in discussion have substantially more control over their self-presentations. If they desire, speakers can be neither seen, nor heard, nor touched. [2] They can easily send and receive instantaneous typographical messages that bear minimal imprint of their social identities. While one may argue that the content of the messages reveals much about the nature of the speakers, nonetheless, the messages tend to stand alone, allowing the potential for identity confusion and deception.

The pseudonymity offered by many information service providers, who maintain records of the real identities of their customers, can be enhanced further to create an almost absolute type of anonymity. This possibility is available through the technical capabilities offered by anonymous remailers. For every message that is sent via the Internet, there is usually a header, the equivalent of an envelope, that precedes the text. This header includes the user name, the domain name, and the time, date, and subject of the post. By sending messages through an anonymous remailer, users can send messages stripped of their original headers and assigned instead random replacement headers. Given the ability to send one message through numerous remailers, the originator of a message may become virtually untraceable.

These messages, which have been the subject of widespread media attention, [3] are the subject of this essay. Users and network managers are currently engaged in a heated debate, either exhorting the virtues or deploring the vices of anonymous communications in cyberspace. In this piece, the arguments for and against anonymous messages will be set forth, and then a limited proposal will be offered on how to address some of the perplexing problems arising from anonymous messages while still encouraging their benefits. Although by no means a solution to the myriad serious issues arising from anonymity online, my proposal, that anonymously mailed messages be required to be identified clearly as anonymous on the header, involves a minimal regulation that offers significant benefits.

The Virtues and Vices of Anonymous Messages

The desirability of anonymous communications within the "marketplace of ideas" has been historically a source of significant dispute. In 1960 the United States Supreme Court in Talley v. California [4] invalidated, on first amendment grounds, a Los Angeles ordinance requiring that pamphlets bear the name of their authors. This decision encouraged limited discussion in legal discourse about questions of anonymity and compulsory disclosure. [5] In 1995 the Supreme Court in McIntyre v. Ohio Elections Commissions [6] once again considered the desirability of anonymous communications in public discourse. In striking down a state statute prohibiting the distribution of campaign literature that does not indicate the name and address of those responsible for its issuance, the Court employed broad language touting the role of anonymous communications: anonymity "exemplifies the purpose behind the Bill of Rights, and of the First Amendment in particular: to protect unpopular individuals from retaliation -- and their ideas from suppression -- at the hand of an intolerant society." FN [Id. at 1524]. Yet, at the same time, the Court suggested that its holding was quite narrow, limited to the particular facts of the McIntyre case. After describing the facts of McIntyre -- an individual community member distributing unsigned leaflets that urged defeat of a school levy on car windshields in a school parking lot -- the Court expressly acknowledged that differing circumstances might "justify a more limited identification requirement." FN [Id. at 1522]. The situation in McIntyre summons up the most traditional of images of anonymous communications. It bears little resemblance to the anonymous communications now taking place online. Here, users have virtually limitless ability to communicate anonymously and instantaneously with a broad and diverse public. As the Networld is giving rise to a flourishing of anonymous interactive public discussions, rigorous analysis informed by modern contexts ought to be devoted to understanding its consequences. [7]

The arguments for and against anonymous communications generally fall into three main categories. They focus on the informative, grouppressure, and enforcement effects of anonymity versus disclosure. [8] First, critics of anonymity argue, disclosure serves an important informative effect. Identification of the speaker may help audiences to assess the truthfulness or accuracy of the speech. Listeners may be better able to evaluate the speech, based on the speaker's reputation for truthfulness or the interests that she serves. For example, knowledge of whether a statement describing a drug's causal relationship to birth defects is being made by a disinterested scientist or by a public relations officer of the drug's manufacturer can help consumers to determine the appropriate amount of weight or deference to be given such information. Last year, an anonymous person on one of Prodigy's electronic bulletin boards asserted that a company, which was giving a public offering, was led by a president who had engaged in fraudulent and criminal activity. [9] The allegedly libelous assertions were said to lead to a plummeting of the value of the stock. Had the readers known the identity of the speaker and been able to engage with him, however, perhaps they could have better evaluated the truthfulness of his assertions. Anonymity, critics argue, precludes this type of informed analysis.

Yet knowledge of a speaker's identity may also lead listeners to misjudge the veracity or persuasive force of an argument. As a substantial body of social science research indicates, one of the key indicia of an individual's persuasive potential is social status, as expressed through characteristics such as race, gender, education, or occupation. "[I]n the absence of knowing much else about a person's abilities, people assume, for instance, that people who have attributes that are high on one or more status dimensions, are more able and knowledgeable than people who are low on the same status dimensions." [10] A classic study of jury dynamics, for example, found that high-status males were most likely to assume leadership roles during the course of jury deliberations. [11] The particular significance of the findings in the jury context lies in the fact that "the status characteristics predictive of eventual power and prestige [had] no relevance to the task at hand." [12] Under conditions of anonymity, participants in discussion are unaware, to a large extent, of others' status characteristics. [13] As awareness of status diminishes, the concomitant biases and prejudices that tend to advantage dominant groups while disadvantaging marginal others then become mitigated. Anonymity, as opposed to disclosure, may thus help to diminish the oft irrational biases and prejudices that disadvantage "outsider" or low-status groups.

An additional aspect of the informative effect of disclosure is simply to disallow impersonation. That is, if forced to disclose one's true identity, one cannot assume another person's identity and attempt to speak in another's name. The veil of anonymity, however, invites false impersonation. Numerous episodes on the Internet attest to this possibility. For example, an anonymous user presented himself as an Associated Press reporter, and distributed a news release stating that "the Microsoft Corporation [had] agreed to acquire the Roman Catholic Church in exchange for 'an unspecified number of shares of Microsoft common stock.'" [14] Although such outlandish statements may seem unbelievable and thus pose little harm, other acts of impersonation can result in particular damage or harm to personal reputations. [15] So long as people do not assume particular individuals' identities, the free play of identity experimentation may be desirable. Under conditions of anonymity, individuals are free to introduce themselves with whatever attributes they desire. A taxi driver can become an independently wealthy aristocrat. An invalid senior citizen can become a robust teen. An African-American woman can become a white man. Anonymous conditions online have "created an entire social world in which it doesn't matter what you look like. Looks are absolutely irrelevant." [16] The ability to reinvent oneself, to wear different masks, enables people to experiment with new persona and fantasize in potentially affirmative ways. Physically disabled individuals, for example, may enjoy regular participation in ongoing social interaction with others who have no knowledge of their disabilities.

The second area of debate around anonymity revolves around the effect of group pressure. On the one hand, mandatory disclosure tends to inhibit the expression of socially undesirable speech and conduct. Public exposure tends to assure that people who speak believe strongly, or at least are comfortable enough, in their claims so as to stand accountable and responsible, willing to confront their opponents. Ideas that are an embarrassment to the originator will not be injected into public discourse.

Anonymity, by contrast, relieves people of the public consequences of their speech and may encourage speech that the speakers may know to be harmful or offensive. The recent rise of hate messages espousing white supremacy and white separatism on the Internet, for example, has clearly been facilitated by conditions of anonymity; as a former Ku Klux Klan leader has admitted, "The access is anonymous and there is unlimited ability to communicate with others of a like mind." [17] Anonymity facilitates not only the general spread of messages of hatred, but also targeted forms of personal harassment. [18]

On the other hand, one may argue that group pressure inhibits those who are most in need of first amendment protection. Disclosure often undermines the free expression of ideas because those with unpopular views may become subject not only to belittlement or social ostracism, [19] but also to physical and economic reprisals. In contrast, anonymity, by shielding individuals from personal retaliation, can encourage speech which otherwise would remain stifled. This otherwise silenced speech has surfaced on the Internet, for example, from political and religious dissidents, whistle-blowers, human rights advocates, and members of various marginalized social groups. [20] As one Internet user has said with respect to his need for anonymity, "To be blunt, I am a bisexual, a pervert and a witch. I also live in Alabama, where at least two of the three are illegal. In a worst-case scenario, I could lose my job, have my career ruined, face prosecution and possibly even have to deal with violence." [21] That non-anonymous speakers may choose to suppress their speech may have little relationship with the speakers' truthfulness or strength of conviction but instead with defects in the community of listeners. Without the guise of anonymity, influential figures in early American history, for example, including Alexander Hamilton, James Madison, and Chief Justice Marshall, all may have chosen not to express their controversial but worthy ideas. [22]

By minimizing group pressure, conditions of anonymity also may encourage expression from those who fear not retaliation but self-revelation. Some people hesitate to speak because they are wary of revealing too much about themselves. Those who are shy or reserved, or those who think they do not speak effectively, often choose to refrain from public conversation. Yet, empowered through anonymity, "[a] person can invent and present a much more confident persona online." [23] People who want, or need, to talk about intensely personal or private matters also may restrain from speaking unless they have access to anonymity. The proliferation of confidential telephone hotlines offers evidence of the success of anonymity in enabling those who are, for example, victims of sexual or substance abuse to communicate their concerns and to seek help. The popularity of various Usenet newsgroups centered around similar topics further confirms the advantages of anonymity. [24]

The third effect of anonymity that is receiving particular attention is that of enforcement. In situations where certain types of speech are illegal, disclosure can facilitate the enforcement of the laws. Public identification of speakers will aid not simply in the pursuit of those guilty of libel, obscenity, false advertising, or copyright infringement, but also in the prevention of such crimes from the start. Individuals who fear arrest will generally choose to avoid speech that is legally prohibited. The potential to hide one's identity, and thus avoid responsibility for illegal speech, has led both to a burgeoning of fraudulent messages online and to a serious hampering of law enforcement efforts. [25]

Because the disclosure requirement is meant to discourage those types of speech which are already the subject of substantive prohibitions, one might suspect that disclosure in this context would not implicate first amendment concerns. Yet censorship questions are raised here because of the potential for chilling speech. FN [See Frederick F. Schauer, Fear, Risk and the First Amendment: Unraveling the Chilling Effect, 58 B.U. L. Rev. 685, 689-94 (1978)] Individuals may refrain from engaging in legally acceptable speech and conduct on the basis of a mistaken fear of their illegality. The so-called enforcement effect thus may actually be coupled with detrimental consequences to full and open democratic discourse.

A Limited Measure to Address Anonymous Messages

In formulating public policy to address the phenomenon of anonymous messages online, policymakers face a number of challenges. First, they need to find a way to combat the negative consequences of anonymous messages while still preserving the positive and affirming aspects. Second, given the general resistance among cyberspace users to any form of legal regulation, policymakers should implement only minimal restrictions. Third, because of the heightened ability to evade regulations in the Networld, policymakers need to adopt regulations that can be enforced.

A simple proposal that meets all three of these requirements would be to require anonymous remailers to identify messages that they resend as anonymous on the header. For example, the domain address could be required to include the term "anon." [26] Currently, because no such regulation exists, anonymously resent messages have the potential for "invisible anonymity." The term "invisible anonymity" is meant to suggest the quality of being anonymous without ever being recognized as being such. Through anonymous remailers, users can send messages that do not appear to be anonymous at all. The messages can have "normal" headers, with a random user name and domain name that give no indication that they have been sent through a remailer. In contrast, this proposal would require "visible anonymity," that anonymous messages be identified as being anonymous.

Although no panacea for the host of problems raised by anonymous messages in the Networld, visible anonymity tends to preserve the messages' benefits while lessening some of their harms. Because it still veils the identity of the speaker, visible anonymity allows speakers to avoid the irrational biases and prejudices that are typically associated with status characteristics such as race, gender, age, or occupation. It also still enables the free play of identity experimentation. Moreover, visible anonymity continues to create low-threat communicative environments in which people are less fearful of personal retaliation or self-revelation. Finally, speakers communicating under conditions of visible anonymity will not become excessively cautious because they can still evade the legal consequences of their speech.

The possibilities of visible anonymity in combating the harms of anonymous communications online are twofold. It can protect users who want to avoid anonymous communications altogether, and it can lessen the harms that result from particular anonymous messages. First, if all messages sent through anonymous remailers were identified as anonymous on their address, recipients of such mail could easily choose to filter them out. Currently people can create "kill files," which automatically delete messages from a user's directory. [27] Users can specify particular subject headings or user addresses, the messages of which will be placed in the kill file and no longer be received by the user. This self-regulatory mechanism, however, is only available when users can specify in advance the terms in the subject heading or user address. [28] Thus, they can only choose to avoid anonymous messages if they are able to specify certain terms in the message heading that indicate their anonymous origins.

The facility to screen out anonymous messages would be beneficial to individual users as well as newsgroups. In situations of anonymous personal harassment, for example, users might avoid ever receiving any anonymous messages. One could argue that when users begin to read anonymous harassing messages, they can just stop reading. Yet, at that point, the harm may already be done. One could also argue that if users want to avoid messages from unknown harassers, they could simply avoid reading all messages that come from unknown user addresses. Yet that would impose too great a restriction upon those who wish to communicate with sources unknown to them. Identification of anonymous sources would permit users to block messages only from those unfamiliar sources who want to remain shielded from responsibility for their postings. [29]

Newsgroups also would benefit from the ability to screen out anonymously generated messages. Self-governing entities, newsgroups users for example, should be able to create the type of environment in which they want to associate. For some groups, such as sexual or substance abuse recovery groups, anonymity may be essential. For other groups, such as money investing groups, anonymous messages may be unwelcome or disruptive. Some newsgroups already maintain a policy of not entertaining anonymous messages. Yet this policy is only realizable when the manager of the newsgroup can identify which messages are anonymous. Currently, because the bulk of anonymously generated messages go through a remailer in Finland, which already identifies its messages as anonymous, [30] managers can monitor and control the flow of most anonymous messages. A requirement that remailers identify their messages as anonymous would facilitate the managers' ability to enforce the policies of self-governing cybercommunities and to protect their autonomy over their information domain.

Beyond enhancing the ability to screen out anonymous messages, visible anonymity also can lessen the harms that result from particular anonymous messages. The fact that a message is anonymous is itself of informational value to recipients. Knowledge that its originator wants to avoid identification and thus accountability may lead readers to be more wary or cautious of accepting its truth value. Users may think twice or seek out additional information before determining the weight to place on assertions. Awareness of the anonymous origins of a message is especially relevant in the case of defamatory or libelous messages, when the truth or falsehood of statements or accusations proves essential.

One might argue that visible anonymity may lead to irrational biases against anonymously generated messages. In conditions of invisible anonymity, where readers do not know whether a message is anonymous, readers of messages will evaluate all messages equally. Yet in situations of visible anonymity, users may be prejudiced unfairly against anonymous messages. While it is true that biases and prejudices may result, it is unclear whether such perceptions would be irrational or unfair. It seems quite rational, indeed desirable, that people may approach anonymous, as opposed to signed, messages with more caution or doubt. If no one wants to assert a message publicly, then greater scrutiny of the message may be justified. Although anonymous communicators may have legitimate interests in avoiding identification with true assertions, their desire that their claims not be subject to exacting scrutiny seems less persuasive.

As mentioned in the beginning of this section, any plausible policy proposals for the Networld must be minimal and enforceable. This proposed regulation is rather minimal because it only requires remailers to attach to the header some indication that the message has been anonymously resent. Such a regulation may not raise too much uproar for two reasons. First, it preserves the benefits normally associated with anonymity. Second, it is already a current practice of some, including the most popular, remailers. The most stable and widely used remailer, operated by Johan Helsingius in Finland, follows this protocol. [31] On each message resent through his server, the user name is assigned a number after the letters "an", and the domain name is "anon.penet.fi". [32] As soon as users see such an address, they know that it has been sent through an anonymous remailer. Because not all remailers attach such a clear marker to their resent messages, [33] however, such a regulation should be implemented. This regulation would be relatively easy to enforce because "cybercops" need only monitor the remailers. [34] The operators of remailers themselves would be held liable if they resent messages that lacked the anonymity indicator. To enforce this regulation, cybercops do not need to be constantly vigilant over the remailers. Instead, they can periodically post messages through the remailers and then determine whether the message at its final address is visibly anonymous. The enforcement of this measure thus entails minimal time and cost.

Conclusion

This essay has attempted to achieve two aims. First, it has attempted to elucidate systematically the advantages and disadvantages of anonymity online and to introduce the concepts of visible and invisible anonymity. Second, it has proposed a simple regulation, one which requires that anonymously remailed messages be visibly anonymous, to lessen some of the harms resulting from anonymous communications online. Clearly, however, even with the implementation of this measure, numerous problems associated with anonymity will remain. Hatred, harassment, violence, impersonation, libel, and copyright infringement are just a few of the serious concerns that must be addressed. More extensive, critical reflection about the emerging social and communicative relations occurring along the information highways is clearly imperative.
 
 

Footnotes

1. When I refer to nationality here, I mean to allude to the possibility of detection of accents.

2. If speakers desire otherwise, however, with the possibility of sending visual images and audio messages, they can choose to reveal more about themselves.

3. See e.g., Daniel Akst, Postcard from Cyberspace: The Cutting Edge, The Helsinki Incident and the Right to Anonymity, L.A. TIMES, Feb. 22, 1995, at D1.

4. Peter H. Lewis, Computer Jokes and Threats Ignite Debate on Anonymity, N.Y. TIMES, Dec. 31, 1994, at 1.4362 U.S. 60 (1960).

5. See e.g., Comment, The Constitutional Right to Anonymity: Free Speech, Disclosure and the Devil, 70 YALE L.J. 1084 (1961); Note, Disclosure as a Legislative Device, 76 HARV. L. REV. 1273 (1963).

 6. 1995 WL 227810 (U.S. 1995) (holding unconstitutional a state law that made it a crime to distribute anonymously any "political communication").

7. Two recent works that do give some attention to this topic include Anne Branscomb, Anonymity, Autonomy and Accountability: Challenges to the First Amendment in Cyberspace, 104 YALE L.J. 1639 (1995), and George P. Long, III, Note, Who Are You? Identity and Anonymity in Cyberspace, 55 U. PITT. L. REV. 1177 (1994). Both works focus primarily on how anonymity ought to be preserved online, without satisfactorily examining the justifications for its preservation.

8. See Note, Disclosure as a Legislative Device, 76 HARV. L. REV. 1273 (1963) (classifying the effects of disclosure into these three categories).

9. See Robert B. Charles, Outside Counsel: Computer Libel Questions in Stratton v.Prodigy, N.Y. L. J., Dec 13, 1994, at 1. In this case, the readers did not know that the person was actually anonymous; he had used someone else's user name.

10. John H. Ellard & Douglas D. Bates, Evidence for the Role of the Justice Motive in Status Generalization Processes, SOC. JUST. RES., June 1990, at 115, 117.

11. See Fred L. Strodtbeck, Rita M. James & Charles Hawkins, Social Status in Jury Deliberations, 22 AM. SOC. REV. 713 (1957).

12. See Ellard & Bates, supra, note 10, at 116.

13. Of course, the message itself, for example, its command of the English language, may reveal some status characteristics.

14. See Lewis, supra note 3, at 1.

15. See Stacey Anderson, Cyber-Security Bugs UW, CAPITAL TIMES (Madison, Wisconsin)(describing an incident where someone impersonated a professor and sent out a racist message under his name to 20,000 users nationwide).

16. All Things Considered: Lawless Frontiers of Cyberspace Invite Trouble, (National Public Radio broadcast, July 12, 1994).

17. Keith Schneider, Hate Groups Use Tools of the Electronic Trade, N.Y. TIMES, Mar. 13, 1995, at A12.

18. Cf. Megan Garvey, Crossing the Line on the Info Highway: He Put His Ugly Fantasy on the Internet, WASH. POST, Mar. 11, 1995 (describing the story of a college student who described his fantasy of raping a fellow dorm mate on the Internet).

19. Consider, for example, the current trend of "political correctness." Many university students on American campuses today are unwilling to express their ideas for fear that they will be ridiculed or labeled as "politically incorrect."

20. See Lewis, supra note 3, at 1 (discussing, for example, self-identified former members of the Church of Scientology who criticize and post official church documents anonymously because they fear retribution from the Church).

21. Akst, supra note 3, at D1.

22. See Comment, supra note 5, at 1085 (recounting that Hamilton and Madison wrote the Federalist papers; Madison, the Letters of Pacifus defending Washington's proclamation of neutrality; and Chief Justice Marshall, as a "friend of the Republic," a defense of certain Supreme Court decisions).

23. All Things Considered, supra note 16.

24. See Lewis, supra note 3, at 1 ("Hundreds of people each day, for instance, use the cloak of electronic anonymity to share their deepest secrets about drug addition, childhood sexual abuse or other sensitive topics with sympathetic strangers on electronic bulletin boards or computer-network chat rooms.")

25. e.g., Charles, supra note 9, at 1 (alleged libel).

26. The most popular anonymous remailer already does clearly label all of its resent messages as anonymous. See infra notes 31-33 and accompanying text.

27. See Long, supra note 7, at 1202.

28. In his note on anonymity in cyberspace, George Long incorrectly asserts that "One could 'kill'all messages that are anonymously originated." Id. In fact, one can only kill anonymously sent messages if one knows elements of the anonymous headers in advance.

 29. I do not mean to suggest here that being "shielded from responsibility" is necessarily a negative quality. As discussed above, this shield or protection serves important purposes.

30. See Akst, supra note 3, at D1.

31. See id.

32. For example, an address could be "an115@anon.penet.fi".

 33. Other anonymous sites include, for example, "alumni.caltech.edu", "soda.berkeley.edu", or "mead.u.washington.edu." See Long, supra note 7, at 1213.

34. It is possible, however, that a hacker would change every header that indicated anonymity to a random header. Cybercops would then have to search for the hacker.