by Ellen Spertus 
May 5, 1996
Copyright (c) 1996 by Ellen Spertus
NOTICE: THIS WORK MAY BE PROTECTED BY COPYRIGHT
YOU ARE REQUIRED TO READ THE COPYRIGHT NOTICE AT THIS LINK BEFORE YOU READ THE FOLLOWING WORK, THAT IS AVAILABLE SOLELY FOR PRIVATE STUDY, SCHOLARSHIP OR RESEARCH PURSUANT TO 17 U.S.C. SECTION 107 AND 108. IN THE EVENT THAT THE LIBRARY DETERMINES THAT UNLAWFUL COPYING OF THIS WORK HAS OCCURRED, THE LIBRARY HAS THE RIGHT TO BLOCK THE I.P. ADDRESS AT WHICH THE UNLAWFUL COPYING APPEARED TO HAVE OCCURRED. THANK YOU FOR RESPECTING THE RIGHTS OF COPYRIGHT OWNERS.
Reports of on-line harassment of women have caused concern to many and have led some to advocate government control over the Internet. This paper describes social and technical - rather than legal - defenses already used by women on-line or that will soon be available. These include blacklists, explicit reputations, secure authentication, private or moderated mailing lists, programs for filtering messages based on their contents or sender, and public replies to harassers. Freedom of speech is shown to be a valuable tool against harassment, both in allowing individuals to not read unwelcome material and in allowing intended victims to publicly embarrass would-be harassers. The ethics of showing offensive on-line material to the author's off-line associates is also discussed.
Because sexism and harassment are problems in off-line society, it should come as no surprise that they are on-line [Val 96], although the forms vary. In this paper, I discuss the experiences of women on-line and the social and technical tools in use or on the horizon for fighting harassment. While it is of course not always the case that women are the targets of harassment and men the aggressors, I focus here on this much-publicized stereotypical pattern.
Online harassment can be divided into two categories: material received by a woman and material posted about a woman (or about women in general). The first category is analogous to when a woman off-line receives repeated unwanted telephone calls or mailings or when she unwittingly encounters material she considers offensive, such as on billboards or in conversation. This category will be discussed immediately, and the second in the latter half of this paper.
Material Received by an Individual
Answering machines and caller identification are two technologies that help to protect women against telephone harassment, although these are of limited effectiveness. In contrast, the potential exists on-line to completely block contact from unwanted individuals with tools for different on-line media:
• Programs to read Usenet news support kill files, used to automatically bypass messages listed as being from a certain individual or meeting other criteria specified by the user. This allows an individual to choose not to see further messages in a given discussion "thread" or posted from a specified user account or machine. People can choose to share their kill files with others in order to warn them about offensive individuals.
• Real-time discussion forums, such as MUDs and Internet Relay Chat (IRC), allow a user to block receiving messages from a specified user. Similar technology could be used to allow blocking messages containing words that the user considers unwelcome. Individuals can also be banned from forums at the operators' discretion.
• Programs have existed for years to automatically discard (or file or forward) e-mail based on its contents or sender and are now coming into widespread use. The second generation of filtering tools is being developed. The LISTSERV list maintenance software [Lsoft 96] contains heuristics to detect impermissible advertisements, and an experimental system, Smokey, recognizes "flames" (insulting email) [Spertus 96].
• Numerous tools exist to selectively prevent access to World-Wide Web sites. While the simplest ones, such as SurfWatch, maintain a central database of pages that they deem unsuitable for children, others are more sophisticated. SafeSurf rates pages on several different criteria, allowing a user to specify that sites containing nudity are acceptable if artistic or educational but that no sites depicting violence are acceptable, for example. Net Nanny provides a starting dictionary of offensive sites, which the user can edit. The user can also specify that pages containing certain words or phrases should not be downloaded.
What is elegant about all these tools is that they allow the user to decide what material is acceptable to her, without censoring anyone else. Freedom of speech is not violated: Anyone can write whatever he likes; whether others choose to read it is their choice, part of their freedom of speech. Just as freedom of religion includes freedom from religion, freedom of speech can now be implemented as freedom from material one chooses not to see.
One of the biggest limitations to the above techniques is the computer's difficulty in determining whether a message is offensive. Many of the above tools use string matching and will not recognize a phrase as offensive if it is misspelled or restated in other words. Few systems use more sophisticated techniques. Smokey recognizes that "you" followed by a noun phrase is usually insulting, but such heuristics have limited accuracy, especially if they are publicly known. For this reason, the LISTSERV rules for recognizing advertisements are kept secret.
The other major difficulty is determining the sender of a message. It is easy to send anonymous messages or to forge a return address. In many discussion forums, a user who gets blocked can get through with a different user name (although this would disrupt any other conversations he is in). The technology exists for users to place unforgeable signatures on their work, such as PGP, which is available for free in the United States (but cannot legally be exported and is outlawed in some countries) [Licquia 95]. If such systems were widely used, perfect filtering by sender could be done. As with caller identification, a user could choose to block messages with the signatures of certain users or all messages without signatures. This technology prevents users from needing "unlisted" e-mail addresses. A person could have one publicly-available e-mail address and alert her friends to include their signatures, allowing her to accept those messages but to reject messages from uninvited individuals (or send them to a program or human assistant for further analysis).
Reputations and social pressure play a part in maintaining civility off-line. So too can they be used on-line, especially in combination with the technology described above. Sharing a kill file or blacklist is a simple example. If I get sent a message that I consider offensive, I could add the user to a blacklist shared with individuals with similar sensibilities. I could set my e-mail program to ignore messages from everyone placed on the list, or I could choose to only filter someone listed by multiple individuals or by an individual whose judgment I particularly trusted.
Alternately, instead of a blacklist, groups could build lists of individuals trusted by at least one member of the group. When considering interacting with a stranger, one could check this database.[May] This is analogous to people's tendency in the off-line world to feel a level of safety with friends of their friends, whom they might meet through explicit introductions or while socializing with the mutual friend. People also sometimes choose movies or films recommended by a friend or critic they trust. Already on-line are collections of reviews of movies, restaurants, and auto-repair shops, to allow people to benefit from others' experiences. Systems such as Movie Critic (http://www.moviecritic.com) give personalized results by determining which users have tastes most similar to yours on items you've both rated and giving you their recommendations on items you have not yet judged. Similar techniques could be used for rating people.
As Amy Bruckman has written [Bruckman 96], whether a woman is uncomfortable on-line depends on which woman and where on-line. Just as people socialize in different places off-line, so do they on-line. Private MUDs and mailing lists allow a group of people to talk among themselves using rules they establish. There are also moderated lists, where a designated leader or team evaluates each message submitted to a list, only forwarding it if it meets the group's standards. Numerous women-only groups already exist.
Most public discussions on-line (as opposed to in the popular press) of "net abuse" are not about sexual harassment but about unsolicited advertisements, which are known as "spam." Spam is easy to produce because it is trivial to send a message to thousands or even hundreds of thousands users. Accordingly, defensive techniques have been developed. If the LISTSERV software decides that a message sent to a mailing list is spam, it intercepts it and prevents the sender from posting anything directly to a LISTSERV discussion list for the next forty-eight hours, unless approved by a human moderator [Lsoft 96]. Similar technology could allow women who are sent an obscene message to broadcast a warning that would advise other users' mail programs to warn them if the same piece of mail or one by the same user arrives in their mailbox.
While many Internet service providers will remove the accounts of users caught spamming, others have no such policy. Letters of complaint and boycotts of these providers has been advocated and, in some cases, appear to have been effective [Ragsdale]. In addition to complaining, one could set one's system to reject all messages sent from an offending system with a form reply explaining the reason why. This would cause users to pressure their provider to adopt anti-spam policies. This same approach could work for harassment. If psuvax.edu is known as being the source of harassing messages, women could choose to automatically reject messages from that machine. Internet Relay Chat already supports the banning of not just individuals but of machines and sites [Rose]. Users who received such a rejection might exert social pressure on their peers to behave responsibly or on their system administrators to deal with complaints in a more effective manner. Because the ability to communicate with women is so highly valued by most of the men on-line, "girl"cott techniques should be effective.
Women's scarcity already gets them superior treatment in some forums. As Phillip Robinson and Nancy Tamosaitis write in The Joy of Cybersex [Robinson and Tamosaitis 93]:
Women in the straight or bisexual adult bulletin board world wield an immensely high level of power. According to Boardwatch Magazine, only 10 percent of bulletin board callers are female. The other 90 percent who are males are eager, often desperate, to talk with female callers. Female callers have their pick in choosing the digital cream of the crop. I was impressed by the male gallantry displayed on the boards (p. 86).
The situation can be similar in forums not devoted to sex, according to Stacy Horn, president of ECHO (East Coast Hangout), which is 37% female:
Men want to be able to say whatever they please, however they please, whenever they please, AND they want women to stick around and listen to it without complaint and like it!
Well, you can't have it all. If you've got a bunch of guys acting like a bunch of jerks, the women are going to go where they can have a decent conversation, where they can make connections, where they can find a sense of community. (p. 104)
Making Use of Freedom of Speech
The most basic social technique - often overlooked but highly effective - is bringing an offensive message to the attention of people who know the sender on- or off-line, causing him to suffer the social or professional consequences of his behavior. For example, I received the following piece of e-mail in response to my web page on women and computer science:
From: "Alexander O. Yuriev" <firstname.lastname@example.org>
Subject: Before I read some of the stuff about Women and Computer I though better about women
Basically, everything is in the subject. Do you honestly belive in what you write and what you have links to? Do you honestly belive that women are capable do as much as men? (In that particular field?) Really?....
Sorry lady, the only place of such studies is a trash can.....
Call me a sexist pig, but when I need to figure out why the whole regional network went down I don't want my partner to be a woman....
The note didn't disturb me. (Not only was an undergraduate's sending such a note to a published female MIT CS doctoral candidate logically absurd, but it was pragmatically stupid. My friends and I will be professors when/if he applies to graduate school, and I will remember his name.) I almost did nothing more with the letter but then saw the following form disclaimer [] at the bottom of the note:
For those who *DO NOT* understand that CIS Laboratories is *JUST* a place I work:
I have no ideas if my boss, CIS Laboratories, Temple University or Chairman of the CIS Lab Committe agrees or not with what I said above. If you need to know that, CHECK WITH THEM!
Obligingly, I forwarded the message, with the following note, to about a dozen people at Temple University, including his boss. the chairs of the department and lab, and the female support staff:
Subject: You might want to clarify whether this is your opinion
I received the below note from Alexander Yuriev. At the bottom, he says he doesn't know whether you agree with the contents of the note. Could you let me know whether you do?
I immediately received several apologies and disavowals of the student's behavior, as well as being told by the department chair that the student was on his way to his office for a talk. Despite the intent of the original message, at no point did I feel victimized: I ended up feeling great, while the student ended up with his reputation lowered in the eyes of people important to him professionally.
Michael Redrain uses web pages to publicly ridicule whose online behavior has offended him. Specifically, he maintains a public list of "lamers" who misuse public bulletin boards in their search for pornography or sex [Redrain]. For example, his "horny AOLers of the week" page includes the following note and his commentary (in italics):
From: email@example.com (JimHotM)
Subject: Re: High school girl wants dirty talk.
Where do u live? Let me know when we can have a dirty talk on phone.
Ohhhh...it's too bad he didn't give out his phone number. That would've made for some interesting times! Why not e-mail him and tell him when you can have a dirty talk on phone. But only if you're in high school! Can we say "Jail bait"? *Clap, clap* I knew we could!
While Redrain did not create the original message from "Kristi," he has created other messages ("trolls") to catch miscreants. Anyone online is free to set such traps and publicize others' behavior.
Women are not the only targets of harassing messages. Webmasters of controversial pages also are frequently flamed. [] Some have taken control of the situation by publicly posting flames they receive complete with sender's name. Anecdotal evidence suggests that announcing that flames will be publicly posted causes fewer flames to be sent. Furthermore, sharing a received flame with the public or with friends can be highly satisfying.
Social and technical means exist or are feasible to reduce the likelihood of an individual's receiving or viewing unwelcome material. In many ways, women are better protected from unwanted speech on-line than off-line. For example, there is no way in the off-line world for a woman to ensure that she not hear certain insulting terms, which could be used by her co-workers or yelled at her on the street. On-line, tools such as Net Nanny ensure that we do not see unwanted words. Developing technologies will allow the on-line implementation of standard social mechanisms such as reputations (good and bad), introductions, and social pressure to behave civilly (however that is defined). More basically, the same freedom of speech that allows someone to send us an offensive message allows us to call it to the attention of others, however this might embarrass the sender.
Material about Women
The above discussion was primarily about materials received by a woman. The other class of potentially harassing behavior is material circulated to others or publicly displayed. If the material about me contains untrue claims, standard defamation or slander law applies. [Loundy 95] If material is falsely attributed to me, secure authentication can vindicate me. Additionally, in forums where forgery is easy, people know to ignore uncharacteristic messages. The difficult areas are when someone posts information about me that is not a lie (which includes not just the truth but also opinion and satire) or when someone posts material that I consider degrading to women.
Babes on the Web
Robert Toups's "Babes on the Web" (BOTW) page [Toups] has offended women both by exploiting their name and picture and for its treatment of women in general. BOTW is a set of pages that display the names and photographs of women with web pages, links to the pages, and Toups's rating of the woman based on her picture. Women were added to the list without permission or notice. At first, Toups refused requests from women to remove information about them. Toups clearly meant to be inflammatory, as evidenced by this excerpt from his original BOTW page:
Along with being a Capitalist Pig, I am a proud Male Chauvinist Pig. As such, I have gathered the World Wide Web sites of Women I could find. Instead of rating them on quality of design, I am grading them on a four Toupsie scale according to their personal pictures. My rating system is totally subjective to my personal tastes and whims. If this page is offensive to you, then go to The National Organization for Women (NOW) Home Page and cry to them. Maybe they will organize a cyber-protest against my page or maybe you will find something else to bitch about. Either way, I won't care. If you have a BABE's page that I don't have on this list please E-Mail it to me using the form at the end of the list. If you found your site on this page and didn't like the rating, put up a better picture, your rating can change in the future.
I reserve the right to post all FLAMES that this page receives.
To avoid violating copyright law, Toups did not copy the women's photographs but included an inlined link to the pictures in his pages, causing them to be loaded from the woman's machine whenever someone visited BOTW. A few women removed their pictures from their own home page in order to prevent such abuses. (Some women had already decided against displaying photographs or even were using their first initials instead of their name to prevent people from knowing they were female.6) Some clever women changed their picture after Toups linked to it, causing male beefcake pictures to appear on his page instead. New "bozo filter" software makes it simple to dishonor unapproved links [Fraser] but was not available then. Toups and his page received a great deal of attention, which he seemed to thrive on.
Brock Meeks wrote a CyberWire Dispatch column on BOTW [Meeks 95], which appears to be entirely based on Toups's account of the controversy and contains gross misrepresentations of women's reactions [Rollins 95]. When a woman pointed out to Meeks the inaccuracy of the information in his column about her (whom he fortunately left unnamed), Meeks neither apologized nor issued a correction but instead further insulted the woman.
The publicity that Toups seemed most proud of, however, was a picture of him (partly undressed) in Wired. My suspicion that Wired would not give positive publicity to a male-bashing woman was later supported by their treatment of Kashka (Shimrit Elisar), author of the "All Men Must Die" page [Elisar]. They printed a horrifically bad (doctored?) picture of her. [Williams 95]
The fundamental problem is not with technology, however, but with power imbalances. That some men post pornographic pictures in workplaces, for examples, shows they either don't realize that many women find them offensive or they don't care. If half of the most powerful people in the workplace were female, they would quickly learn and care. The same phenomenon is at work at Wired, which is aimed at men and written by men. [Borsook 96]
A number of pages were put up protesting or ridiculing Toups and BOTW. The most elaborate, by the gallant Kilroy (Dav Amann) [Amann 95], ridicules Toups and provides the opportunity to rate Toups on a score of one to four hemorrhoids. Nevertheless, Kilroy acknowledges that he is stooping to Toups's level, quoting: "Never wrestle a pig. You both get dirty and the pig likes it." In contrast, the essay I wrote [Spertus 95] includes Eleanor Roosevelt's saying: "No one can make you feel inferior without your consent." In it, I argue that while Toups's behavior is obnoxious, it does not diminish women. Rather, he is publicly making a fool of himself and behaving in a way that might harm his career:
I think Mr. Toups is being stupid to post something offensive. I'm sure his list will be discussed on Systers, which has over 1500 female computer professionals, many in high places. Also, many men disapprove of behavior such as his and would consider him a risk for their company because of his behavior. Using the network to make enemies instead of friends is not rocket science. I have the feeling that in the short term I am more likely to be a sought-after employee than Mr. Toups and that in the long term I'm more likely to become a tenured professor, respected researcher, or successful executive (to say nothing of greater success in personal relationships).
The most controversial part of my essay was:
I also looked up the address of [Toups's] employer, Next for Mac. While I would not suggest (or want) them to discipline him for his outside activity, I would consider warning them to be sure he is not using their equipment to do this, or else they could be held responsible. If they subsequently question his ability to work on a team or be a manager, that's their business. (Like Mr. Toups, all I would be doing is pointing people to publicly-available information.)
I still have mixed feelings about this argument. On the one hand, complainers would only be making use of their freedom of speech. On the other hand, one recalls the voice of blacklisted actor: "I am free to say what I think, and they...are free to punish me for what I say. Let them argue with me...But to take my livelihood away! You can freeze to death from such freedom." In Free Speech for Me - But Not for Thee [Hentoff 92], Nat Hentoff discusses the 1970s boycott of Florida orange juice for having Anita Bryant as their spokesperson. As Ira Glasser (then executive director of the New York Civil Liberties Union and now of the ACLU) said: "Anita Bryant has taken certain public positions, and certain people who disagree with those positions are trying to punish her economically. This is exactly what happened during the McCarthy years." Hentoff argues that a boycott for an employee's off-hours speech is fundamentally different from boycotting a business for the employer's actions, as in the grape and lettuce boycott led by Cesar Chavez. Informing co-workers or employers of an employee's behavior is clearly different from mounting a boycott or firing someone, but I still have qualms about it. I don't know if it's merely my internalization of Internet mores or if such behavior really is wrong.
The feeling on the web is that it is acceptable to contact the employer of a spammer (someone who inappropriately sends advertisements). Søren Ragsdale maintains a page about "Spam King" Jeff Slaton [Ragsdale] that includes Slaton's supervisor's name, address, and telephone number, as well as Slaton's own phone number and address and a map to his house. An inspection of Usenet archives and personal correspondence with Ragsdale shows that response to his page was overwhelmingly favorable, with the exception of Wired. It is not clear what the reaction would be to similar treatment of someone writing racist or sexist material.*
Whether or not complaining to an employer is acceptable (which may depend on whether the message was posted from work), publicly posting commentary and criticism certainly is. I have the satisfaction of knowing that my essay is now part of the public record and that anyone who searches on "Rob Toups" will find it. While Toups has toned down the rhetoric on his page, his original inflammatory introduction is available on my page unless I choose to remove it.
The Jake Baker Case
University of Michigan undergraduate Jake Baker e-mailed and posted violent sexual fantasies to the Usenet group alt.sex.stories; in one of them, the name of the victim was a female classmate [Swanson 95]. The University promptly suspended Baker, and he was charged with interstate transmission of a threat, a federal crime. Baker's defenders included the Electronic Frontier Foundation (EFF).[EFF] Charges were dismissed by U.S. District Court Judge Avern Cohn, who said that the writings were "only a rather savage and tasteless piece of fiction." [Cohn 95] In contrast to Cohn and the EFF, Catherine MacKinnon considered the writings a threat, quoting from Baker's e-mail: "It's not enough anymore to think about it, I have to do it." [MTTLRA 95] The government has appealed the dismissal.
While Baker suffered numerous effects off-line and generated much discussion on-line, few responses are known to have been made on-line. One woman wrote and posted a violent story in which Baker was the victim, but I think that response was analogous to countering pornography aimed at men with pornography aimed at women in that the mode of discourse does not emotionally affect both sexes in the same way. A more successful technique, that has been used off-line to counter pornographic pictures of women, was posting homosexual male pornography, which makes many heterosexual men uncomfortable plus fearful that people will think the pictures are theirs. Similar techniques could be used on-line. I suspect that men such as Jake Baker would be more disturbed by the public display of stories depicting them engaged in and enjoying homosexual sex than as victims of heterosexual women. This raises the question of whether to exploit homophobia and other prejudices. This question arose with Rob Toups and lookism. There was public criticism of his body, and, when I appeared in Wired (for unrelated reasons), I was tempted to publicly remark that Wired hadn't asked me to take my shirt off (as Toups did) even though my breasts are even larger than his.
I believe that social means could have been used to get Baker to change or remove the offending story. The first step could have been for someone to tell Baker that (some) women find it threatening and disturbing for their name to be used in such a manner and offering to help him think of a new one. Even if Baker did not cooperate, such a move would be helpful in evaluating whether Baker was threatening or harassing this specific woman, which would make a criminal or civil case more feasible. A next social step might be to publicly call attention to Baker and his story, such as by printing a letter or advertisement in the student newspaper, exposing Baker to the censure of the community. Certainly if Baker can use a woman's name in fiction without her permission, his name can be used in non-fiction. More militant activity would be posting signs in Baker's dorm or in women's bathrooms announcing that Jake Baker has published fantasies about raping and killing women and has used the name of a female student acquaintance. Since the statement is true (assuming it has been established that Baker's posting had not been forged), this action would be legal, although people might disagree about whether it would be ethical, since it could be interpreted as an invitation to harass him. That leads to the question of whether it is appropriate to respond to harassment with harassment, as well as what one's definition of harassment is. Undoubtedly, different people will come to different conclusions, with some people "fighting dirty" and others taking the moral high ground. Already, a wide variety of women's and feminist groups have formed on-line, from the respectable Systers mailing list to the outrageous Heartless Bitches International [HBI 96]. Women have written documents on on-line safety, such as Lucretia's "Take Back the Net" [Voisin 95], and Internet scholar Phil Agre has written about abusive games people play on-line and how to deal with them [Agre 96].
While women cannot be absolutely protected from offensive material on-line, there are numerous technical tools that prevent them from receiving unwanted material. These tools can also be used socially to implement on-line reputations which individuals will be concerned with preserving. Because men want to interact socially with women on-line, it is unlikely that most men would say things offensive to many women in their on-line community, since this could get them blacklisted. Nobody is forbidden to write what he chooses; others are empowered to read what they choose.
A concern of mine about filtering systems is that they will further fragment society by allowing people to only read material that supports their prejudices, just as people who read differently-slanted magazines grow even further apart. Despite my concern, I have to acknowledge not only that the technology is inevitable but that people's freedom to choose what they read should be just as respected as their freedom to write. Just as feminists need to treat with respect all women's tastes, we need to honor everyone's tastes, including those of men or of anti-feminists. (When I told a friend that it was offensive for a woman to receive a sexually-explicit message in response to a call for friendship, he asked me if it would be offensive for a man looking for sex to be sent messages offering a relationship.) I do not doubt that I would be blacklisted in certain communities, and I accept that. When we consider whether to try to publicly shame an author of material we find offensive, we should remember that this technique will also be used against us. While free-speech absolutists might argue that social and economic pressure should not be brought on people because of their speech, social consequences (reputations) are important off-line and should properly play a role on-line.
The collection of on-line communities will continue to grow. Some will be feminist-friendly; others will not. How women are spoken of in the most public of forums will depend on how they are regarded in society at large. If men are overwhelmingly powerful in the media, they will continue to be the arbiters of what's cool, although individual women will vote with their feet (fingertips). Not only do women have popularity, some have authority and all have intelligence. Sociolinguistic Susan Herring and her colleagues have studied the language of men and women on-line, documenting and quantifying what so many women already felt, such as that when women try to participate at the same rate as men that men complain that women are dominating the conversation and employ a variety of silencing strategies. In a sublime example of using the master's tools on the master's house, they documented such behavior's taking place on an e-mail list for professional linguists Herring et al. 95. Herring's other papers explore the differences between men's and women's on-line styles and examine whose style of appropriate conduct is enforced [Herring 95a, Herring 95b] Social scientists should recognize the Internet as a fantastic source of data that can be used to evaluate their theories and as a new medium that will transform the fields of applied gender studies.