Paul Gowder

"There you go, talking about class again" ¹

On the net, there are very few actual benefits to the vaunted "cheap speech." The term itself is a complete misnomer. "Speech" has always been free. It is entirely costless to yell one’s brilliant ideas into a well, and hope for a reply. The proponents of "cheap speech" are, in reality, referring to "cheap audience" – the idea that it is easier, in terms of time and effort, to reach people and influence them with ideas over the net.

This paper critically examines, and ultimately rejects, this idea. The unavoidable conclusion is that, to the extent audiences are cheaper over the net, they are only cheaper for those people (corporations and the wealthy) who already have the means to reach audiences in physical space. Finally, this paper proposes several solutions for a society as a whole, interested in using the net as a tool to advance weak ² viewpoints, promote minority businesses, and otherwise create a greater pool of social contribution. A few suggestions are also offered for those disempowered people who wish to take advantage of the internet to attract their own audience.

It is at this point useful to mention a couple of guiding principles for this paper. First, and most important, is the concept that I will call "marginality." This is the idea that, as players with functionally unlimited wealth compete for scarce resources, the price of the resource rises to that point at which the buyer receives $1.01 from each additional dollar spent. This is called marginality because economists, when describing behavior like that in various contexts, have only one thing in common, the description of that $1.01 point as the "marginal" point, or rate of return, or cost, or something similar. ³ This is extremely important on the internet because of one special "how the net is different" factor.

This is the second guiding principle of this paper: the competition between commercial and noncommercial speakers. ⁴ In non-internet space, marginality does not pose a problem for small speakers because the quantity of speech is limited – speech is the scarce resource, not listeners. ⁵ In cyberspace, however, the ability for "anyone to speak", and the adaptability of the medium for all forms of communication, from commercial to political, has heavily changed the speaker-listener ratio. As a result, speakers have to compete with each other for audience members. ⁶ The cost of the audience members is expressed many ways (the entirety, in fact, of part one of this paper), but is always a uniform cost for both commercial and noncommercial speakers. Offline, the costs of noncommercial speakers are significantly lower, because they only compete with each other.

For example, the cost of putting a noncommercial statement in the newspaper is limited to the cost of the writing ability and postage necessary to convince the editorial staff to include your opinion in the "letters to the editor" section. A commercial speaker must pay the advertisement rates and occupy that space. It would be as unusual to see political messages in a corporate retreat as it would to see Ford advertising by writing in chalk on the sidewalks. While public communication is still at a disadvantage, there are at least some places that are given solely to the group. On the net, there are no such "zones" dividing the editorial page from the classified ads from front page bottom corner. While this is a fine method for choosing between purely economic actors, it is a poor method for choosing between economic and social actors. This paper proposes creation of such differentiated spaces as one way of minimizing the competition ⁷ between commercial and noncommercial speakers.

The third major principle is what I will call "credibility disintermediation" (or filter dislocation). ⁸ This refers to the fact that, offline, listeners can judge the credibility of speakers (or, expressed in the language of this paper, speakers can earn the currency of credibility necessary to get listener time) through third party intermediaries. Thus, the credibility of a restaurant review is established once it is printed in the New York Times. This gives a certain amount of power to the credibility granters: off the net, the powerful players effectively act as filters for information: in order to get the credibility of the NYT, you have to be hired by the NYT. On the net, viewers are forced to judge this for themselves. While this may seem more democratic: the NYT can no longer control credibility, it is merely an illusion. If viewers judge credibility from (in the language of this paper, sell their attention based on) links from the NYT web site, or the presence of expensive site design tools that can make their site look like the NYT site, the goals of democratization are lost. If it is the case that the internet does allow more people to put communications "out", to potentially reach large numbers of people, there must be some additional filtering mechanism to protect readers from untrustworthy information. It seems that the filtering mechanism has simply shifted from an editor to a mass of individuals. ⁹

It is a relatively simple matter to isolate those needed to communicate a message of any kind to an audience of willing or unwilling listeners. One needs to have a message to communicate, to some general pool of members out of which your audience may be chosen. One needs to have a means of offering (or propelling) that message to the audience. At that point, one needs a group of people out of that audience willing and able to listen to the specific message (as opposed to the entire pool, who may be willing to listen to the offer of the message: the "click on this link to reach my site"). Finally, one needs the credibility to influence audience members to be, at the very least, willing to take the speech seriously, and evaluate it on the merits. Each of these "audience economy" factors has its own cost which must be paid for effective communication. In addition, success has its own ex post price.

Pool

If a tree falls in usenet, does it make a sound?

The first hurdle which any speaker will encounter is simply finding a group of people who are in a situation to be able to reach their message. On the net, this requires three things. First, that the intended audience be on the internet. Second, the intended audience must be able to find the offer of ideas: the communication must be accessible, not only technically, but easily and cheaply for the potential reader. Finally, they must convince the reader that the communication is worth their time.

The first factor is especially difficult for minority messages. ¹⁰ While it is extremely difficult to find information on the precise composition of race, age, religion, etc. on the net, it is likely that the internet is predominantly white, middle-class, and American. This can be assumed simply from the amount of money still required to get on the internet for any significant period of time (20 minutes at the local public library, for example, would not be significant.). A modern computer from a major, easy to reach, retailer still costs between one and two thousand dollars. ¹¹ That about sums it up – unless a small speaker is employed at one of the high-paying jobs involving computers with internet access, or at an institute of higher education offering the same, they are fundamentally shut out from internet access. Whole sections of the potential audience are lost to speech. A speaker who wishes to promote unionization among migrant laborers can shout and shout all she wants on the internet, without that communication ever reaching its intended audience.

To reach this pool, one must purchase advertising on the major portal sites. This is the clearest application of the marginality principle. Portal sites having an oligopoly on eyeballs can charge up to the point where the richest speakers are not willing to purchase their entire stock. This creates a deadweight loss for the smaller purchasers, who can not get any eyeballs from the portals. ¹²

Credibility

"I DON’T LIKE SPAM!" ¹³

The most tragic factor for minority speakers with ideas or goods and services to contribute to the group is inability to attain credibility. As a general matter, credibility can be bought. This can be seen by the simple expedient of comparing the perceived honesty of a businessman on Rodeo Drive in Beverly Hills with one on Normandy in South Central LA. Likewise, a speaker giving a harangue in Harvard Square is more credible in an Armani suit than in the rags of a homeless person. On the internet this factor, obvious to anyone, is translated into more subtle appearance and location factors. ¹⁴

One oft ignored problem with the internet as a mode of transmission is the credibility gap between the internet itself, and more traditional modes of communication. Cyberspace has not earned full "rights" as a medium for credible communications. For example, many businesses do not communicate with potential clients, employees, or contractors over the internet, considering it unprofessional or unnecessary. ¹⁵ Even on the internet, there is a heightened awareness of "scams," ¹⁶ false information, dishonesty, and "kooks" ¹⁷ that would not be present were the same messages communicated in non-electronic form.

Even those parts of the medium which have exact analogs in physical space are disadvantaged on the internet. For example, "junk mail," when transmitted through the United States Postal Service third class mail, is considered obnoxious and is often thrown away unread. Junk mail, when send over the internet by people not equipped to do bulk physical mailings, is met with ISP complaints, denial of service attacks, lawsuits ¹⁸ and legislation. ¹⁹

Finally, the usual offline problems with credibility of weaker speakers are repeated online. Since poor people are less likely to be able to afford education, and education increases earning power, it is just as difficult online as offline for socially weaker individuals to communicate their ideas. This is even exacerbated on the internet because the vast majority of communication is done through text, requiring strong written communication skills. ²⁰ In addition to the flaws in education that can be detected and filtered through the bias of a listener to a spoken discussion, the additional formal rules of written English (like spelling and punctuation) can operate on the internet to further reduce the credibility of persons with a lack of formal education. Furthermore, much communication on the internet requires a certain minimum of technical knowledge – a working command of HTML to build a presentable web page, basic competence in the normal software of internet access. These too may be out of the reach of a less empowered internet user, and in an environment such as usenet where users are regularly castigated for setting their carriage return at anything greater than 80 characters, these small problems can have a huge impact on the credibility of a speaker. This impact is intensified when the audience compares it to a more mainstream information source, such as the New York Times, ²¹ which produces easily readable web pages most likely built by professional web designers.

One additional difficulty for the weak speaker is the lack of access to the tools of greater speakers. The cliche of free press for those who own one holds on the internet, but is no longer a means of distinguishing not between those who own presses and are capable of mass producing documents and those who are not. Rather, it distinguishes between those who have the copy of Adobe Photoshop ²² to produce attractive web pages, and those who do not. This is partly a problem with the legal system’s protection of intellectual property, and partly a problem of marketplaces for technical skill. One major limited resource in the process of the internet (for lack of a better term) is the technical skill (and venture capital) necessary to produce top quality products. As such, the real competition in the case of applications such as Photoshop is between buyers. Commercial speakers drive up the price of such applications by sheer ability and willingness to pay, leaving the weak speakers mired in marginality.

Ex-post price

"You have the right to free speech as long as you’re not dumb enough to actually try it." ²³

The law itself does not impose back-end costs (costs after successful communication) on audiences. Rather, the current maladjustment of the legal system allows people opposed to the communication in question to impose additional, often malicious, costs against the speakers. This creates a cruel paradox where the more effective the attempt at getting an audience is, the more painful the eventual backlash from the legal system is.

The first, and most severe, imposition of additional back-end costs to effective communication on the internet is the current confused state of affairs in personal jurisdiction. Many courts have held that publishing a web page subjects a speaker to personal jurisdiction everywhere the page can be read – basically, the entire world, with the possible exception of Antarctica. ²⁴ The standards used to determine jurisdiction on the internet are likely to be highly detrimental to small speakers.

1. "Commercial" speech. The courts generally seem to look dimly on commercial activities over the internet, holding that attempting to do business on the net subjects a user to personal jurisdiction in any states where he may have done business. This does not simply extend to the location of customers, but includes the location of any ISP that is used to reach the customers. ²⁵ This is problematic to small speakers because, without trying to use the internet for revenue, many will not have the resources to get their message to audiences. Large speakers are, obviously, more likely to have the resources outside of the internet to generate enough funds to promote their message and limit their market transactions for jurisdictional purposes. They do not need to accept the jurisdictional consequences sell advertising or products over the entire net, and can limit their activity to reduce their costs, while still being able to afford to gather an audience. Thus, this artifact of the legal system privileges more powerful speakers on the internet.
2. Interactivity. In non-commercial cases, the courts have seen fit to establish jurisdiction by testing the "interactivity" of a site. ²⁶ Beyond the artificiality of designating web sites as more or less interactive, it is likely that this will also prejudice small users. Assume a web site that collects mailing addresses from users, then uses the US mail to send them more political or commercial information. This would allow the web site owner to distinguish between states he wanted to communicate with for jurisdictional purposes and states to which he did not want to risk subjecting himself. This comes with the increased cost of using the mail. If the web site owner needs to save these costs by responding with e-mail, she does not have geographical tags. This small user is now subject to much higher potential costs to be inflicted by any offended party in any state. This is clearly absurd.

Wide jurisdiction is one of the major causes of the paradox mentioned earlier, that a speaker on the net is punished more as the reach and persuasion of her speech increases. If the internet is to fulfill its potential as the ground for "cheap speech", plaintiffs should not be allowed to retroactively increase the costs of an audience by increasing the potential for punishment as the chances of successful communication increase. The jurisdictional muddle on the internet should be solved by federal legislation, outlined below, or by judicial action.

More "substantive" law also causes a severe increase in back-end costs. The two major substantive areas of difficulty are libel and intellectual property law. This is a short discussion, because these problems have been explored many times before.

Libel law is the most obvious manifestation of the paradox of increasing success equaling increasing costs. The more audience members who can or do receive and are convinced by the offending communication, the more "harm" is allegedly done by the defendant, and the more liability attaches to the conduct. However, this is not as important as the increase in the plausibility of a suit. Assuming a libel suit is likely to lose, with the increasing exposure of the information bringing about the suit, the credibility of the plaintiff’s case can rise, and with this, the costs to the defendant. Cases that could be dealt with in summary judgment may have to go through expensive discovery first. And so on through the system, until the defendant is forced to take down the offending communication or reach bankruptcy. ²⁷

This is perfectly reasonable when reach strictly correlates with wealth. When the New York times is accused of libel, it can afford to defend itself against the ensuing lawsuit. However, a single person with a political manifesto on the internet cannot. Worse still, any speaker who is not aware of these increased costs may not be prepared to pay them, and can then be severely punished through the private law system. A speaker who is aware of these possibilities, of course, will be chilled in the traditional sense.

The same chill can occur in intellectual property law. In this case, however, it may come from two ends, with two different purposes. Trademark law could have the effect of chilling the most effective ways to reach audiences. Copyright law could have the effect of chilling the most effective ways to convince them. This is primarily important in criticism and satire. Again, because this ground has been walked over many times, I will only provide a short summary of the basic arguments.

Trademark law can be used to lock up the most effective ways to reach an audience, in any number of ways. A person wishing to criticize McDonalds, say, for environmentally unfriendly practices, would be most effective registering the domain McDonald’s.com. However, the moment they do this, McDonald’s lawyers will undoubtedly be on the courthouse steps with a trademark infringement suit. ²⁸ Likewise, the person wishing to criticize McDonalds may realize the domain name is already taken. The next step may be the registration of their criticism site in the major search engines under "McDonalds." Come the resulting trademark infringement suit. Then they may advertise their site using the McDonalds trademark. Meeting again the requisite lawyers. The suits, again, do not need to be meritous to silence dissent, just plausible. Trademark is also subject to the paradox illustrated in libel law: the more consumers reach the criticism site instead of McDonald’s, the more "consumer confusion" there is, increasing the plausibility of any suit filed.

Copyright works the same effect on the content of the messages themselves. ²⁹ The organization criticizing McDonalds may decide to copy policy statements that passed through McDonalds headquarters, to the effect of "lets use styrofoam, it’s cheaper and hey, we’re all old, and we don’t have kids, so forget the environment." Doing so immediately subjects them to an expensive copyright suit as the price of convincing their audiences. ³⁰

Another method of increasing back-end costs is through the device of contract. There are many new forms of adhesive contract on the net, with an increasing disassociation from the actual ability of users to consent. On the internet, contract sneaks in many forms – from the "shrinkwrap" used to sell any new software product, through "clickwrap" agreements where a user is told that "clicking ok" binds them to several different "terms of service" and "user agreements" on different sites, to agreements imposed in almost comically unilateral ways. ³¹ The problem is in the very ubiquity of these contracts on the internet. In a real world commercial transaction, it is clear what circumstances invoke contract, and what do not. It is also generally clear what terms are present in a contract. Walking into a store and purchasing a commercial good is a familiar experience to users, who understand that they are agreeing to pay a fixed price for a known service. On the internet, where there is very little "trade usage" or other common practice support, it is considerably more difficulty for a "little guy" to know what he is agreeing to. Contracts are imposed on the internet for many things that are not normally within the experience of the user – eg. "by viewing the next page you agree to" style clickwrap – and exacerbate the standard problems with adhesive contracts. This becomes important in the realm of speech when ISP’s and quasi ISP’s (such as free e-mail providers) step into the picture. For example, America Online still refuses to publish its list of "dirty words" of which a user consents to the silencing. Likewise, ISP’s as a matter of course bar spam – even though "spam" is very loosely defined. (This would be akin to UPS banning bulk commercial mail.) This increases the cost of speech ex ante by requiring that the user who inadvertently runs afoul of these restrictions go to court or find an alternate way of speaking. ³²

The code also assists in the imposition of back-end prices. First, to the extent that smaller speakers can use the internet to reach a wide range of audiences, larger speakers can use the same methods to reach even more people and attack the message. Second, the powerful interests can use any success in reaching people to easily discover the offending speech and take legal or extra-legal action.

This second problem is because most accepted communication on the internet is untargeted. With the exception of extremely expensive targeted search engine ads, and sanctionable bulk mailing, ³³ all mass-communication on the internet can not be targeted. While the cost to the speaker is debatable, the cost to an audience member is near zero, except in time. The powerful interest wishing to find hostile information no longer needs to subscribe to the right magazines, but only needs to use 1 search engine to search the entire web, perhaps another to search usenet, and can discover any communication that, in a pre internet world, would perhaps have been limited to a niche audience. This is a variety of the "perfect enforcement" problem in public law. In private law, we have a "perfect information" problem, because powerful interests can now attack all information they find hostile. A disaffected employee, pre-net, could distribute negative, yet true, information about her corporate employer through word-of-mouth, anonymous written communication, and other difficult and less effective techniques. On the internet, however, the employee is tempted by the apparent ease of diffusion to many, and the anonymity she believes she has, to post those negative messages on, for example, a stock valuation website, where the company, concerned about its image, can easily search the web for such postings, and, through any number of means (such as legal action, financial action, or simple intimidation or cooperation), get the true identity of the disaffected employee, for easy termination.

Small speakers can take certain "self-help" steps to increase their own exposure. Organization to share vital resources is the most important, and the only one addressed here. The role of the government is somewhat more extensive. The government can directly address those laws which unduly burden small speakers, but can also subsidize low-income speakers and their audiences.

Organization of small speakers

"If love is just a big charge account, why it has to be mine?" ³⁴

The most effective method of self help for small speakers is too become large speakers. Perhaps the best comparison is to the labor unions. An individual group of workers does not have the resources to survive in the event of a strike, nor to compete with "scabs" and control the employment market. When labor unions organize into large groups, some striking at various times (eg. the AFL-CIO), the funds of the members not striking can be used to support the striking members. The unions can also gain more attention to the workplace problems their action is concerned with by virtue of sheer numbers. Similarly, if environmental activists on the internet wish to support each other in an attack on a polluting corporation, their best move would be to contract to support each other in the event one of them gets sued by a powerful opponent. The environmental groups on the internet can create linking programs and search engines to lower the cost for users to discover their speech and increase their joint readership through network effects.

In the case of ex post legal sanctions, the money to defend themselves and survive should be an adequate deterrent for many suits. Especially when the opponent is a corporation or other profit-maximizing entity (such as the church of scientology), the inability to deter critics with ultimately unsuccessful suits will cause the suits to stop. A simple game theoretic model will perhaps demonstrate this best.

Assume a profit making organization, C, wants to deter an attack by B, an individual with a bone to pick, and A, an organization with pooled resources to defend against suit. The cost of an effective attack is 1 million dollars in stock drops, lost business, etc. The cost of filing a lawsuit to each side is 500,000 dollars. B has unallocated defense resources of 50,000 dollars. A has unallocated defense resources of 2 million dollars. B values the ability to cost a million dollars in lost profits at $50,000, the amount of money he is willing and able to spend on legal defense. A, because of the diminishing marginal utility of wealth, is willing to spend more to attack C, and it is worth $700,000 dollars to A to cause a million dollars of loss to C.

Extensive form game of this scenario:

B is unwilling to pay the –500k costs that a lawsuit from C will impose. As a result, if C accepts a certain strategy of always filing suit, B will not attack C. If C builds a reputation for always filing suit, this can actually save C money in the long run. The cost of filing suit twice is recovered in one attack deterred. A, however, because it has more resources, is willing and able to handle the lawsuit costs. As a result, no matter how litiginous C becomes, A will not be deterred as long as the harm (1.5m) creates enough utility (1,050,000 in this case, at a rate of 700k per million) to compensate for the cost of a lawsuit. Because C is a profit-maximizing actor, there is no reason for C to spend an additional 500k when it will deter nobody. The suits will not occur.

This reasoning suggests that minority interests on the net should pool their resources in the event of conflict. ³⁵ However, they can also pool their resources to encourage the effective speech which got them into trouble in the first place.

On the web, there is a site called "Disinformation, the subculture search engine." ³⁶ Disinformation was begun by a group of liberal-minded individuals with technical knowledge and writing skills who wanted to provide links to pages that agree with them, and criticism of those that disagree. To fund this project, they are members of the "at hand network" which places advertising banners on websites and pays per click. This is the ideal way to solve the problem of difficult communication on the web. With adequate publicity, ³⁷ this site could become the portal, or the first place checked, for anyone dissatisfied with the current crop of information. In addition, they should try to gain some commercial leverage, sufficient funds to make a significant advertising dent in major competitors, or some other way of drawing mainstream viewers. The power of such a group effort could influence enough mainstream views to create, through the free market and without government intervention, a competitive "sidewalk" for those dissatisfied with the mainstream "sidewalks." Disinformation does not go nearly far enough. The ideal for a privately run "sidewalk" would be more of a network between fringe sites. This should be open to any site with fewer "hits" per month than some arbitrary number, to track those sites who actually need more effective speech. Resource sharing should also be encouraged – perhaps each member would be asked to contribute something, be it graphics creation ability, writing talent, server space, or some other needed resource, to the good of the group. In addition, the network should engage in some form of revenue raising or sharing activity to enable members to meet legal challenges and meet the costs of site promotion. If resources permit, the fringe network should even attempt to capture more mainstream viewers by setting up a portal site with special preference for those groups who are members of their network.

A full outline of the proposed steps small speakers could take to increase their own exposure is beyond the scope of this paper. However, it should be noted that no such site exists. While many sites exist purporting to create "networks" of other internet speakers, ³⁸ those small speakers are uniformly large – the ACLU, for example, seems to be a member of many of these coalitions. The ACLU does not need the help. Such coalitions should focus on sharing resources between small speakers, not providing yet another list of links for large ones, no matter how socially important those large speakers are.

Government subsidy for low-income computing

2 gigs in every hard drive...

One characteristic of minority viewpoints is that they are not backed up by a lot of money. That is why they are not majority viewpoints. For a poor person in the inner city wanting to speak out against the police, even if they can find a computer, in the library or something, it probably has censorware installed. Even if they can find a computer without censorware, it is probably impossible for them to find webspace. ³⁹ Even if they can find webspace, it is probably impossible for them to find the computer time, in the scarce resources of public libraries, YMCA’s, and similar organizations, to actually publish their views. Even if they manage all that, they probably do not have the knowledge necessary to create web pages, promote them, or successfully post anonymously to usenet. To solve these problems:

1. The government should provide free, no-questions-asked (and content-neutral) web space for socially important issues. (any non-commercial, not purely personal, web pages.) This should probably be restricted to individuals, as organizations can often afford their own. However, if the removal of persons from the potential customers of ISP’s raises the organizational rates very high, perhaps tax-exempt organizations should be given free webspace as well.
2. There should be a separate, government run, free TLD. Perhaps a .soc domain or a .pol domain (or both) would be appropriate. These should not be subject to trademark law, since there would be no risk of consumer confusion in an entirely noncommercial domain. ⁴⁰ Indeed, trademark owners should not be allowed to own domains corresponding to their trademarks, since the purpose of that space on the government TLDs would be free criticism of those products. ⁴¹
3. The government should provide public computers in public places, such as libraries. Federal funds should be allocated to the poorer communities to achieve this.
4. Computer education (as opposed to computer-assisted education, which is an entirely different paper) ⁴² should be provided to all high-school students. This education should focus on computer-assisted communication and information gathering (eg. html, effective information searching) to close the technical ability gap between the elite and the ordinary person. In addition, these classes should be offered in prisons (and computer access should be provided in all prisons) and in any public assistance sort of "job training" programs.
5. A federal sales tax should be placed on internet businesses with revenues over [some amount large enough to not catch any "little guys"] to pay for all of these improvements.

Government paring down of internet-unfriendly laws

‘But congressman... how will this affect the web?"

The final government ⁴³ step that can be taken to insure the availability of the internet for dissent is to eliminate those laws which unfairly burden small speakers. This should be done as follows:

1. Since the courts have been so nice as to hold that the entire internet is subject to the commerce clause, ⁴⁴ the federal government should immediately enact internet-friendly legislation preempting state unfriendly legislation. ⁴⁵ This could swiftly deal with libel and state unfair competition/trademark. The standard used to determine which laws to preempt should be: "does this law increase the costs of communication without regarding the resources of the communicator?"
2. Congress must pass a law limiting the personal jurisdiction of the federal courts, at least, to those circumstances where the contacts with the forum state are more than merely internet contacts. ⁴⁶ Some other contacts, be they telephone, mail, or personal, must be required to establish personal jurisdiction. For state court jurisdiction, there is little that can be done on a national level unless the Supreme Court speaks on "fair play and substantial justice" at some point. Otherwise, the best that can be hoped for is a modification of state long-arm statutes.
3. Intellectual property rights should be severely curtailed on the internet. ⁴⁷ An expedited hearing process in the federal courts should be instituted for all IP claims involving internet communications. The goal of this process should be, by limiting discovery, etc., to create an early stage where IP based claimed on the internet can be dismissed at minimal cost to either party. (The same could be done for defamation law, were it in federal court.) In addition, much stronger "fair use" rights should be recognized in both trademark and copyright contexts, to allow for all social and political comment on the protected works. ⁴⁸
4. This almost goes without saying, but internet censorship should be stopped at all costs. Censorship never affects the mainstream. Is "indecency" ⁴⁹ the price of a politically free society? If so, it is one that everyone should be willing to pay. ⁵⁰ If "Lady Chatterly’s Lover" can grace the shelves of the public libraries, free for perusal by children everywhere, there is no reason why it can not similarly be available for children on the internet. Many political, cultural, and otherwise valuable, works are considered "obscene" or "dangerous" by the majority. To violate the strict principles traditionally held in the area of government-sponsored censorship for the sake of "decency" would be a grave mistake. ⁵¹
5. Private censorship should also be stopped. While there is some advantage to allowing, for example, parents to control their children’s use of the internet, ⁵² it is important that the people controlling are really the parents. If the actual listener has control over the decision (and the education to make that decision) to accept content, perhaps the promise of credibility disintermediation can be fulfilled. "Proxies" for the judgements of parents, whether providers of "dangerous sites" lists, filtering software providers, "safe" search engines and ISP’s, should be held strictly liable for both pecuniary and non-pecuniary damages caused to parties desiring to be read by customers of a proxy. It should be an affirmative defense to liability under this statute, that a reasonable average customer of their software would have an expectation that sites such at the plaintiffs would be censored. ⁵³ In addition, proxies should be required to provide the reason a site is filtered out when refusing access, and should most certainly not be allowed to "invisibly" block sites. ⁵⁴ While this will not solve all of the severe problems with private censorship (for example, there will be no incentives for people who would choose not to visit the sites filtered for "extremism"), it is a start.
6. Privacy should be protected as an important tool to allow unfettered speech.

One strong way the government can contribute to the diversity of voices on the internet is by completely protecting anonymity. In the scenario of a corporate employee who criticizes his company, and is fired after a routine web or usenet search, the government could save the time and expense of a wrongful termination lawsuit by simply making anonymity the default rule. ⁵⁵ There may not be much the legal system can do to lower the technical demands of anonymity. It would be extremely difficult and expensive to require corporations to not keep backup copies of e-mail traffic, for example. Fortunately, the government can take some steps to protect anonymity.

1. Recognize a privacy right in communications stored on the property of others. The government can create, or the courts, can acknowledge, security against unwarranted intrusion into the private e-mail files of persons on a corporate network, by the corporation, a private ISP by their employees, and other ostensibly privately owned forums for intended anonymous communication. To do this, the government would be required to abandon the idea of privacy rights partly based in property rights, as well as the circular "expectation of privacy." Instead, privacy should be based in an idea of the intent of the user – if the user (or a reasonable user in the circumstances) intended the communication, or the identity attached to it, to only reach certain people, then any unauthorized invasion should be considered an invasion of privacy, actionable to the extent of the harm done. ⁵⁶ (Legal costs, termination of employment, etc.)
2. Not actually pass laws forbidding, or deterring, anonymous communication or ISP facilitation of anonymous communication. If ISPs are allowed to choose or not choose to allow anonymity in their systems, if anonymity is desirable to the speaker, the market should quickly make itself heard. ⁵⁷ This will only be the case, however, if the market is open enough to allow small ISP’s to offer anonymity and compete with the larger ISP’s without risk of liability (or the increased cost of suits mentioned in "back-end costs" above) for the actions of their users.

The government has an important interest in democratization of viewpoints. Under traditional "marketplace of ideas" theory, the more different ideas that are allowed to enter the market and receive full consideration, the more effective social policy will be. In addition, effective speech allows persons to actualize themselves – they can express their unique personalities and thoughts for the world at large to see.

Many scholars claim that such a structure exists on the internet right now. However, those scholars uniformly ignore the important cost factors outlined in this paper. Until the cost of effective speech is lowered on the internet, small speakers will have free speech without free communication. Without an audience, the promise of "cheap speech" is nothing but an illusion.

The internet provides a great opportunity to allow small speakers to influence society and express their personalities. The internet also poses a great challenge: society, to take advantage of this opportunity, must break the control over the means of, not production, but communication. The internet breaks control over one of the means of speech, but not the means of audience gathering. If society wishes to take full advantage of the potential of the internet as a democratizing medium, the full "ownership of the means of production" of ideas must be transferred to the people as a whole.

FOOTNOTES

This change in perspective follows naturally from the change in power, and has some interesting dynamics. "Consumers" control the failure or success of net speakers more than they do offline, because the increased supply of substitutable goods creates a very high elasticity of demand in any single good. As such, the cost of satisfying a listener increases, to the point where the speaker is buying a listener, not the other way around. This is further complicated by the growth of such programs as @guard <http://www.atguard.com> (visited 12-22-98), a filter for listeners to remove advertisements. If such a program becomes popular, it will be a rejection by the listeners of the offer of useful content in exchange for eyeballs on other content. Presumably, advertiser supported content will decrease -- the power relationship from such a product will actually shift in favor of the content provider/eyeball consumer, as supply goes down.

It is also important to consider the impact of this on the intellectual property “industry.” If, as I have just claimed, the speaker is “buying” the attention of the audience, the value of her intellectual property comes from what that attention can purchase: as many people have argued, the reason the speaker “buys” attention is not so they can “sell” words, or code, or whatever, but because they then get power in other markets. For example, they “buy” credibility by “buying” attention, and this credibility can then be used to sell books. To put it a different way, one first buys credibility by placing oneself on public access television. If one is successful, one gets picked up by a local television, and is now paid, rather than paying, for the opportunity to speak. Public access television has many fine analogies to small speakers on the internet, one of which is that nobody ever watches it. On the issue of the changing nature of IP on the net, see generally Eric S. Raymond, The Cathedral and the Bazaar, <http://www.tuxedo.org/~esr/writings/cathedral-bazaar/cathedral-bazaar.html> (last updated 11-22-98), John Perry Barlow, IP on The Global Net <http://www.eff.org/~barlow/EconomyOfIdeas.html>. (visited 12-21-98) This will become important when I later argue that one remedy to the audience cost problem is a relaxation of certain IP laws.

Of course, another way to view this is as a disintermediation problem. Where before the consumer’s market for restaurant reviews in New York was mediated by the New York Times hiring process, now it is direct. Two markets existed in the previous example: the “restaurant reviewer” market and the “newspaper” market. Now that these two markets are one, we must use different tools to analyze this new market.

9. Internet users have devised a number of ways to judge the credibility of communication on the internet. One person interviewed for a “scam protection” site recommends judging the quality of a page based on the credibility of the pages linking to them. This, in addition to being somewhat useless, because the credibility and judgment of the linking page must then be judged, is also a means of perpetrating the “lock-in” effects of ideas. It can be compared to the network effects idea in economics: once a group of credible sites link to each other, they bolster each other’s perceived credibility relative to unlinked sites. At the same time, the situation of unlinked sites falls into a viscous circle: the unlinked sites become more difficult to find as the number of linked sites increase to occupy the attention of a reader. As a result, it is less likely that this potentially good information may be discovered and linked. See Scambusters on Y2K <http://www.scambusters.com/Scambusters26.html> (visited 12-22-98)

While the “linking” situation is usually considered a problem of getting eyeballs rather than gaining credibility, for any solution to work, both eyeballs and credibility must be provided. This is why public “sidewalks in cyberspace” are doomed to fail. The government, by creating, for example, “social search engines” to reach dissident information would either have to indiscriminately catalog every site on the net, in which case it would be nothing more than a publicly owned altavista, or it would have to prefer certain sites over others, providing this credibility by government endorsement. Obviously, however, government endorsed dissident groups quickly cease to be dissident groups. This is the core reason why the government’s role should be limited to removing harmful laws and giving poor people access and training. Anything else introduces distortion into the “marketplace” for audiences, and leaves an opening for questionable public motives.

10. Strong businesses and other monied interests, on the other hand, have sophisticated tools to identify their potential audiences. One example is SRI consulting <http://www.sriconsulting.com/> (visited 12-21-98), a business consulting firm who’s parent company claims to have been one of the first two nodes of ARPANET, after UCLA <http://www.sriconsulting.com/sri/sri1C.html> (visited 12-21-98). SRI offers a package of internet services, including “consumer acceptance of technology” studies <http://future.sri.com/vals/cat.method.html> (visited 12-21-98) to help businesses determine their precise markets. For anyone interested in this issue, their “vals” service should be worth exploring. To avoid spending the hours clicking that I did, here is a good starting point: <http://future.sri.com:80/vals/ivals.segs.html> (visited 12-21-98). This describes the “types” on the internet. Note also that you can take a survey to determine your “type”, but this will in all probability include your information in their database. See also Firefly <http://www.ffly.com> (visited 12-21-98).

11. See eg. <http://www.compusa.com/CompUSA/CompUSAPC/pc_default.asp> (visited 11-98) Note also that a move to “network computers” will, while lowering the cost of access to the internet, will probably be more harmful to minority viewpoints than beneficial. Networks supporting the network computers have to be owned by someone - most likely corporations with their own interests in regulating the content of information presented to consumers. So, for example, if it became profitable to market a network with “child protection” censorware products built in at the network level (with considerably less control to the user), it would quickly defeat many of the diversity advantages from making computers more widely available.

14. For example, a column on one of the sites within the “Pathfinder Network” <http://www.pathfinder.com> (visited 12-21-98), because it implies some sort of reintermediation of credibility, has more inherent credibility than a web page on Geocities <http://www.geocities.com> (visited 12-21-98), a free home page provider that has no “standards.” This is just one of many examples of the parallelism between the false promise of economic “disintermediation” and the false promise of noneconomic filter removal.

15. See eg. “FAQ for beginning writers” <http://www.sfwa.org/writing/faqs.htm> (visited 12-21-98) suggesting that publishers do not accept online submissions because “the consensus among editors has been that the process of manuscript submission works fine for them, and they're not interested in changing it... the publishing industry was designed to turn out profitable and high-quality books, not to be convenient for writers.”

16. See eg. Yahoo Directory, “internet frauds” <http://dir.yahoo.com/Computers_and_Internet/Internet/Business_and_Economics/Fraud/> (visited 12-21-98), Scambusters “Scam Check” <http://www.scambusters.com/scamcheck.html> (visited 12-21-98) (This is the same scambusters that accepts advertising from groups that would be recognized as shifty even off the internet: “7 internet success strategies” <http://www.successplace.com/iba/index.html> (visited 12-21-98))

17. See eg. Usenet group <alt.usenet.kooks>, “Howling at Net.kooks” <'http://www.netizen.org/Howling/Kooks/> (visited 12-21-98). The very titles of these forums ignore the reality that people who are “kooks” on the internet are also “kooks” off the internet. This erroneously implies that “kookery” is somehow more encouraged by the internet, and contributes to a general distrust of the medium. (Sanity is not medium-dependent.) The ignorance of this seemingly obvious point is a result of credibility disintermediation. Were there no need to independently judge speaker credibility, there would be no special “kook” warning. However it comes about, it effectively reduces the credibility of the entire net by casting it as the sole place where one may find “kooks.”

18. See eg. “AOL’s Campaign Against Junk E-mail” <http://legal.web.aol.com/email/jeaolcam/index.html> (visited 12-21-98), Bigfoot <http://www.bigfoot.com> (visited 12-21-98)

19. See eg. summary of proposed bill H.R. 1748 “Netizens Protection act of 1997” <http://thomas.loc.gov/cgi-bin/bdquery/z?d105:HR01748:@@@L|/bss/d105query.html>> See also <http://thomas.loc.gov/cgi-bin/query/D?r105:1:./temp/~r105XN6iBV:e371:> (visited 12-21-98) (there does not seem to be any better way to cite this - however, they may all be reached from AOL, id)

See eg. Real Player <http://www.real.com> (visited 12-21-98) (once merely an internet audio player, now offering streaming audio and video, specially designed “channels”, “bookmarks” and other “large” features favoring stronger players who can afford to cut deals with Real to feature their communication), Pointcast <http://www.pointcast.com> (visited 12-21-98) (more high bandwidth, high commercialization technology), and, astonishingly enough, Neoplanet <http://www.neoplanet.com/> (visited 12-21-98), a browser apparently building a commercially sponsored portal right into the software. This last comes perilously close to a complete private ownership of the internet.

21. <http://www.nytimes.com> (visited 12-21-98)

22. Costing our poor speaker an additional $995. See <http://www.adobe.com/prodindex/photoshop/prodinfo.html> (visited 12-21-98)

31. One of the more amusing examples of the latter variety comes from the Software Publisher’s association. They provide a “button” to use on websites to show support for their antipiracy campaign <http://www.spa.org/piracy/button.htm> (visited 12-15-98). Along with the button is a contract regulating the terms of use of that button, basically taking away fair use rights, but reciting assent as “downloading, copying or otherwise reproducing the SPA BUTTON." Since the SPA must take the position that loading into RAM is a copy (otherwise they could not prevent software being used on a network under a single user license), they are placed in thw position of claiming that a mere view of the button - even on someone else’s web page, and inadvertent - is sufficient to assent to a contract taking away fair use rights.

36. <http://www.disinfo.com>

38. see eg. Global Internet Liberty Coalition <http://www.gilc.org>, Netaction <http://www.netaction.org>

40. Consumers would be on notice that trademark did not apply there, and that brand names in that space would bear no relation beyond the choice of the speaker to the goods or services branded. In effect, this would be an example of government (rather than dangerous private) “zoning”, placing players on notice that different rule sets apply in different locations. When the theoretical justification for these rule sets (for trademark, consumer confusion) is based on notice, this allows the rules to be abandoned. See generally David G. Post, Anarchy, State, and the Internet: An Essay on Law-Making in Cyberspace, 1995 J. Online L. 3, <http://www.law.cornell.edu/jol/post.html> (arguing the inevitability of a “free market” in rule sets based on private zoning)

41. This would also avoid the problems discussed in Dan L. Burk, Trademarks Along the Infobahn: A First Look at the Emerging Law of Cybermarks, 1 RICH. J.L. & TECH. 1 (Apr. 10, 1995) <http://www.urich.edu/~jolt/v1i1/burk.html>' with online misrepresentation.

54. For example, if a site censored http://www.mightyblack.net (to my knowledge, no such site exists), when a user attempts to visit the site (or executes a search that would normally retrieve it), the refusal should read “This site blocked by OppressBot (tm) for: extremism” rather than “This site blocked by OppressBot(tm)”, “document contains no data” or, worst of all, a simple absence from the results of a search engine. This would allow a parent to actually choose to view the “extremism”, being made aware that the site espouses “black power” political views, rather than African-American pornography. This would also be to the advantage of the proxy manufacturers, as they would not be so vulnerable to defamation lawsuits from groups claiming that, because most people only expect filtering software to block pornography, they were associated with unsavory content.