MICHAEL W. CARROLL ^††

While unwanted commercial solicitations have been a fact of life for a long time, never before have they threatened the viability of an entire mode of communication. Because the marginal costs of producing and distributing electronic junk mail are very low, the incentives for advertisers to flood the network with unsolicited commercial solicitations are substantial. Left unchecked, this flood of advertisements could produce a tragedy of the commons; advertisers, acting in their rational self-interest, will distribute as many unsolicited advertisements as they can until most users of the medium find that the effort of sifting through unwanted solicitations has become too great. At a minimum, this result would substantially impede communication through the medium.

This article considers the recognized means to avoid the tragedy of the commons-self-regulation, regulation by market forces, and government regulation-and concludes that some government regulation of unsolicited commercial solicitations in a unified medium is likely to be necessary and will be permissible under the prevailing interpretation of the First Amendment.

Before proceeding further, the terminology used in this article merits a short discussion. The object under study is the practice of sending unsolicited commercial solicitations; these include the credit card applications and catalogues you often receive in the mail, the phone call from the long distance phone company asking you if you want to switch your service, or the Internet email inquiring about your interest in purchasing a cassette tape of erotic readings.² The advertising industry considers these all to be forms of "direct marketing" because they are solicitations sent directly from the advertiser to a known universe of consumers. Direct marketing is distinguished from mass marketing, by which advertisements are transmitted via mass media to an imprecisely known universe of consumers.³ This article will not adopt the industry phrase "direct marketing" because in the eyes of most recipients these solicitations are better described as junk.⁴ A more accurate legal description would be unsolicited commercial solicitations, but repetitious use of this phrase-or worse, its acronym (UCS)-would be cumbersome for reader and writer alike. Therefore, for the sake of variety, this article also uses the terms "junk communications," "unsolicited advertisements" and "spam"⁵ to designate unsolicited commercial solicitations sent directly from advertiser to consumer.

Part II briefly surveys the permissible scope for regulation of unsolicited commercial solicitations in developed media-door-to-door solicitations, junk mail, junk phone and junk fax. The First Amendment,⁶ as currently interpreted, requires courts to apply intermediate scrutiny to content-based regulations of commercial speech. Under that test, complete bans on junk communications through any media have been held to be unconstitutionally overbroad-except in the case of junk fax. But restrictions allowing individuals to stop the flow of junk communications or to limit certain technologies that can be used for unsolicited communications have been upheld.

Part III first reviews current attempts by private parties to regulate electronic junk mail on the Internet. These attempts have been mostly unsuccessful and probably will continue to fail absent some government regulation. Instead of considering how the government might regulate the Internet as it is presently constituted, this part anticipates that a new, unified medium will emerge in which most, if not all, of our communications will consist of paperless, digital bits of information moving along an open network and funneled through a single box in our homes and offices. This part next considers what might happen were there no regulation of unsolicited solicitations in this medium. Concluding that unregulated electronic junk mail could substantially impede the medium's functioning, the part then considers what market responses might develop to prevent such an outcome. Part III concludes with an assertion that market regulation alone will not sufficiently respond to unsolicited commercial solicitations, and that some form of government regulation will be necessary.

Part IV explores whether the principles and doctrine of the law on junk communications set out in part II can readily support government regulation of unsolicited commercial solicitations in a unified medium. This section indicates that some forms of regulation, such as a requirement that unsolicited commercial solicitations bear a label identifying them as such, probably would be supportable under existing First Amendment doctrine. A more difficult issue arises if the government attempts to ban junk communications altogether.

Litigation has arisen with regard to door-to-door sales,⁷ unsolicited mail,⁸ unsolicited telephone calls,⁹ and unsolicited fax transmissions.¹⁰ Interestingly, many of the issues have been developed in the context of unsolicited contacts made by attorneys.¹¹

With regard to unsolicited commercial solicitations made door-to-door, by mail and by phone, the dominant theme of the litigation has been to balance the First Amendment rights of the advertisers to speak and individuals to hear that speech against other individuals' rights to privacy in their homes. With regard to junk fax, legislatures and courts have employed a different analysis because most persons affected by unsolicited fax transmissions have been businesses, whose privacy interests carry less weight. In the junk fax context, courts have allowed government regulation because junk faxes use the recipient's paper and ink (a conversion or unjustified cost-shifting rationale) and the junk fax transmission precludes other, desired, transmissions from getting through (a scarcity rationale).

A problem that arises in many of the challenges to content-based government regulation of unsolicited commercial solicitations is that the regulations are stated overbroadly, such that they reach unsolicited religious or political solicitations, which are protected by a stricter judicial test for legitimacy. While this is likely to remain an issue with respect to junk communications in a unified medium, the focus of this article is on regulating those unsolicited communications that are clearly commercial in nature.

This section briefly discusses the First Amendment test for government regulation of commercial speech and then reviews how that test has thus far been applied to government regulation of unsolicited commercial solicitations. From this review, I draw some general principles about the permissible scope of government regulation of junk communications.

With the development of our information economy, however, the Court has come to read the First Amendment to provide broader protection over the nexus between the marketplace of ideas and the marketplace for goods and services. Thus, in 1976, the Court invalidated a state statute barring pharmacists from advertising prescription drug prices on the ground that the First Amendment protected consumers' interest in a free flow of commercial information to better assist their purchasing decisions.¹³ Some commentators have challenged the distinct treatment of commercial speech under the First Amendment,¹⁴ but the Court appears firmly committed to its course.

By 1980, the Court had settled on intermediate scrutiny as the appropriate standard of review for restrictions on most commercial speech. Under this test, commercial speech that neither concerns unlawful activity nor is misleading may be regulated if: (1) the government asserts a substantial interest in support of its regulation; (2) the government demonstrates that its restriction on commercial speech directly and materially advances that interest; and (3) the regulation is narrowly drawn.¹⁵

For a state interest in a regulation to be substantial, the government must show that the regulation is directed at a "real" harm rather than a conjectural harm.¹⁶ To show that a regulation directly and materially advances the state interest, the government must demonstrate that the regulation will in fact alleviate the real harm to a material degree.¹⁷ To be narrowly drawn, a government regulation does not have to be the least restrictive means available; rather, there must be a "reasonable fit" between the government interest and the regulation.¹⁸ In determining whether a fit is reasonable, the presence of "numerous and obvious less-burdensome alternatives" to the restriction is relevant.¹⁹

Nevertheless, sufficient case law exists to tease out the general principles on which courts rely to assess government regulation of junk communications. A recurrent issue that will not receive much discussion here is whether content-based regulations are narrow enough to effectively disaggregate and regulate only commercial solicitations but not solicitations by political or religious organizations. The focus of the discussion is on the courts' analyses once they are satisfied that the regulation reaches only unsolicited solicitations that are purely commercial.

A large number of Americans consider junk mail to be annoying or even offensive.³⁵ Many are willing to pay to avoid it.³⁶ Additionally, concern is growing about the harm done to the environment by the resource consumption involved in producing and disposing of junk mail.³⁷ This concern remains somewhat diffuse because the environmental groups most likely to put the concern in focus are themselves junk mailers.³⁸

This discontent has led to a governmental regulatory response. Numerous pieces of legislation have been proposed to regulate junk mail,³⁹ but because the phenomenon is so pervasive and so many interests rely on junk mail for revenue, the proposed legislation very rarely becomes law. Thus, to the extent that junk mail is regulated, it is done so primarily by forces other than the law.

Members of the direct mail industry have responded to public anger with limited self-regulation: the Direct Marketing Association (DMA) is the trade association for the industry, and it assists consumers in removing their names from DMA-member solicitation lists.⁴⁰ As of 1995, 3.5 million consumers were on the list, which was expanding at a rate of 50,000 per month.⁴¹ In addition to self-regulation by the industry, individuals have occasionally found other means to respond to junk mail; for example, one creative person successfully used state contract law to strike back.⁴²

Another emerging regulatory tactic is attempting to restrict marketers from buying and selling names, addresses, and other personal information. In a novel Virginia case, the plaintiff claimed a right under the state's right-of-publicity statute to be asked for consent before a marketer traffics in his personal information, and a right to be compensated if it does.⁴³ In addition, the New Jersey legislature considered a measure that would allow marketers to traffic in the personal information of only those state residents who consented to such use of their personal information.⁴⁴

However, when the government attempts to protect individual privacy by preventing an entire category of unsolicited commercial solicitations from being sent through the mail on the ground that the solicitation is offensive, the government violates the First Amendment.⁴⁹ The First Amendment does not permit the government to prohibit speech as intrusive unless the "captive" audience cannot avoid objectionable speech.⁵⁰ Recipients of junk mail can avoid offensive speech because the "'short, though regular, journey from mail box to trash can . . . is an acceptable burden . . . so far as the Constitution is concerned.'"⁵¹ Moreover, there may be some willing junk mail recipients whose interests in communication also would be frustrated by a government ban on certain classes of junk mail.

The government may prohibit attorneys from sending unsolicited mail for a brief period of time when that mail is likely to be a significant invasion of privacy and when attorneys have available alternative means of advertising their services.⁵³ In addition, because communications from attorneys have the potential to have a coercive effect, the government may require attorneys to label their unsolicited commercial solicitations as an "Advertisement."⁵⁴

The labeling requirement potentially could be applied to all commercial solicitations. In upholding labeling requirements for attorney solicitations, the decisions appear to reflect a broader principle that a government labeling requirement places only a slight burden on the speaker and the consumer has a strong interest in knowing the nature of a communication received. In addition, commercial speech is entitled to intermediate judicial scrutiny only to the extent that the speech is not misleading.⁵⁵ A government labeling requirement may actually provide greater protection for the sender by curing what might otherwise be misleading commercial speech.

Unsolicited telephone solicitations are intrinsically more invasive of privacy than junk mail.⁶¹ A junk call requires an immediate response because-at least for the time being-there is a cultural expectation that a person should answer a telephone call at the time it comes in.⁶² Some also argue that aural communication is inherently more intrusive than visual communication.⁶³ The perceived intrusion of a junk call is worse when the caller is not even a person but a machine.⁶⁴

As with junk mail, telemarketers have engaged in some self-regulation,⁶⁵ but this effort has been largely ineffective at stemming public frustration with junk phone calls.⁶⁶ Perhaps because the intrusiveness of junk calls is treated as a legislative fact,⁶⁷ government has responded by regulating junk phone calls far more extensively than it has regulated junk mail.

Pursuant to the TCPA, the Federal Communications Commission (FCC) engaged in rulemaking to implement the Act.⁷⁰ Some of the significant regulations promulgated pursuant to the TCPA include a "labeling" requirement by which a telemarketer must identify himself,⁷¹ time restrictions limiting calls to between 8 a.m. and 9 p.m.,⁷² and a provision requiring telemarketers to record and respect do-not-call requests.⁷³ On reconsideration, the FCC has maintained all of these restrictions.⁷⁴

The TCPA also directed the FCC to explore the possibility of compiling a national database of those who do not want to receive unsolicited telephone solicitations.⁷⁵ The FCC considered the option but decided that the costs of compiling and maintaining such a database would exceed the benefits to the public.⁷⁶

The most far-reaching regulation of junk phone calls in the TCPA is a prohibition on artificial or prerecorded voice message calls made to residences without prior express consent.⁷⁷ Congress provided a private right of action to assist in the enforcement of this provision.⁷⁸ The breadth of the prohibition on automated calls prompted a swift court challenge, resulting in the decision in Moser v. Federal Communications Comm'n [Moser I].⁷⁹ The Moser I court enjoined the FCC from enforcing the prohibition on prerecorded message calls to residences on the ground that the prohibition was a content-based regulation that did not reflect a reasonable fit between the government's substantial interest in protecting individual privacy and the means chosen, which still left open the possibility of auto-dialed message from noncommercial speakers.⁸⁰

The FCC appealed the decision and won.⁸¹ In Moser II, the Ninth Circuit reversed the district court, holding that the ban on auto-dialing machines was a content-neutral time, place, and manner restriction that was narrowly tailored and which left abundant alternatives open to advertisers.⁸² Judge Fletcher, author of the Moser II opinion, read the statute to be content-neutral because it bans all automated calls, regardless of content.⁸³ While the statute uses permissive language that allows the FCC to exempt calls not made for a commercial purpose,⁸⁴ this permissive language alone does not convert the all-encompassing ban into a content-based prohibition.⁸⁵

The Moser II court recognized that choosing to view the statute as a content-neutral regulation rather than as a content-based regulation of commercial speech was a distinction without a difference in First Amendment terms,⁸⁶ but the Moser II court appears to have made the distinction to imply that Congress could have banned telemarketing calls altogether. For authority to uphold the selective ban on automated telephone calls, the court relied on a series of Supreme Court decisions holding that underinclusive government regulations on speech are permissible so long as the government's distinction does not amount to viewpoint discrimination.⁸⁷ The court found the statute to be reasonable because "Congress may reduce the volume of intrusive telemarketing calls without completely eliminating the calls."⁸⁸ Having prefaced this comment with a review of permissible underinclusive regulations, the court implied that Congress acted reasonably in banning automated calls because it could have banned all telemarketing calls as unwarranted invasions of residential privacy.

In Destination Ventures, the Ninth Circuit held that the government has a substantial interest in preventing the shift of advertising costs from sender to recipient-here the relevant costs are paper and toner-and in preventing unsolicited advertisements from precluding receipt of desired communications. The court concluded that the interest in preventing cost-shifting is directly advanced by a total ban on fax advertisements and the ban is a reasonable fit with the interest; the government is not required to adopt less restrictive regulations such as time or page restrictions on unsolicited fax advertisements.⁹⁶ By using an underinclusion rationale similar to that in her Moser II opinion, Judge Fletcher found that Congress could reasonably choose to ban unsolicited fax advertising without addressing other unsolicited fax transmissions such as those with political messages since junk faxes impose the bulk of the cost-shifting problem.⁹⁷ Interestingly, the court was not swayed by the plaintiff's evidence that there soon may be technological solutions to avoid the "real harm" of cost-shifting.⁹⁸

Regulations that enable individuals to choose not to receive unsolicited advertisements are reasonable,¹⁰⁴ but the government may not supplant that private choice with a decision that an entire mode of communication should be foreclosed¹⁰⁵ or engage in content discrimination by banning unsolicited advertisements that the government deems to be offensive.¹⁰⁶ However, when Congress finds that use of a technology for advertising is particularly intrusive on residential privacy, and when recipients have no effective means of opting out of receipt of such intrusive communications, Congress may supplant private choice and ban advertising by those means so long as advertisers have alternative avenues open to them.¹⁰⁷ Congress also may supplant private choice and ban altogether a mode of advertising when that advertising necessarily shifts advertising costs from sender to recipient.¹⁰⁸

This subsection briefly describes current attempts to privately regulate spam on the Internet. Concluding that these efforts alone cannot stem the rising tide of unwanted commercial solicitations made by electronic mail, this subsection then describes the constitutional issues that might arise should the government attempt to regulate electronic junk mail. Because these issues are likely to evolve as the medium develops, as discussed in part II.B., a full analysis of them is deferred to part III.

As the number of Internet users has mushroomed, online culture and the customary law on the Internet are in a period of rapid, contentious development.¹¹⁷ Battles over electronic junk mail capture aspects of this culture clash.¹¹⁸ A general netiquette principle is that one's communications should not waste other people's time.¹¹⁹ A second principle has been that the Internet is a non-commercial forum.¹²⁰ This latter principle has waned with the rise of commercial online service providers, but remnants remain. These two principles are brought together in the netiquette rule against unsolicited advertisements sent by electronic mail.¹²¹

While breaches of the netiquette prohibition on spamming continue to increase, the prohibition still acts as a constraint on some marketers.¹²² For now, the netiquette prohibition increases the cost of spamming by increasing the risk that a spammer will receive a negative return on his investment. Even if one percent of the spam recipients purchase the spammer's goods or services, there is a real chance that the spam will drive away otherwise potential customers among the remaining ninety-nine percent of recipients.¹²³

This constraint, however, is limited. The online culture is in the midst of rapid transformation. While some newcomers are likely to adopt current netiquette norms, many others are likely to bring the norms that govern our physical spaces into cyberspace. One of these norms may be an acceptance of junk mail as an inevitable nuisance. If all marketers engage in spamming, the risk of a negative return will decrease because consumers will not blame Sears for spamming them when they also receive spam from J.C. Penney, Macy's and all the other department stores.

Generally, the techniques vigilantes use to enforce the norm are: attempting to deny a spammer continued access to the Internet; putting the spammer on a blacklist; "flaming" the spammer (i.e., inundating his mailbox with hostile mail);¹²⁵ and using utility programs known as Cancelbots that automatically delete multiple postings to online discussion groups.¹²⁶ If the vigilantes cannot identify the spammer, they may seek to exact vengeance on the service provider that allowed the spam to get through.¹²⁷

Vigilante justice generally is a troubling form of regulation because results are unpredictable and often appear more arbitrary than the results under a more formal system. For example, one marketing company claims that it was wrongly blacklisted, and those who have posted the blacklist have not provided a contact number or a process for getting off the list.¹²⁸ Vigilante regulation, then, is not likely to provide a systemic solution to the problem of unsolicited electronic junk mail, and there may be a backlash against the vigilantes if their actions produce arbitrary enforcement.

Because both senders and recipients of electronic junk mail gain access to the Internet through ISPs, ISPs can regulate spam at both its origin and destination. To regulate spam at its origin, numerous ISPs, including those who operate proprietary networks that are also linked to the Internet, include terms that make it a violation of the service agreements to send electronic junk mail.¹³⁰ To regulate spam at its destination, ISPs can attempt to filter out the junk in their subscribers' incoming electronic mail. America Online, the largest commercial online service, recently began to filter subscribers' incoming mail by bouncing certain junk mail back to the sender, and, at the time of this writing, litigation is ongoing with respect to that policy.¹³¹

One commentator advances the view that America Online should be allowed to take this action and that ISPs should be allowed to use contract as a kind of marketplace for the law of the Internet.¹³² However, this form of structural regulation of spam probably is of limited utility for practical, market and legal reasons. The practical constraints are that identifying the source of electronic junk mail is often difficult,¹³³ and junk mailers can themselves become ISPs. Spammers have incentives to disguise the origin of their mass mailings-to avoid being flamed among other reasons. As of this writing, litigation is in process over a leading spammer's practice of designating a different reply address than the originating address to avoid flames and the return of undeliverable messages.¹³⁴

The market reason is that some of the larger service providers take an equivocal position toward spam. On the one hand, the service agreements of most major providers make it a breach of the agreement to spam other members.¹³⁵ On the other hand, these same providers can and do profit from selling subscriber profiles to spammers¹³⁶ or from spamming their own subscribers.¹³⁷

The legal reason is that, even if the providers decided to crack down on spammers, to do so effectively they would have to act in concert,¹³⁸ and such action could raise antitrust problems. Regulation of electronic junk mail by online service providers is therefore not likely to stem the rising tide of spam.

Each of the four aspects of private regulation on the Internet deter some spamming, but the costs imposed on spammers are fairly low in comparison to the potential benefits from spamming. Assuming rational action, we should expect spamming to become an increasingly common practice. We might reasonably expect to see some government regulation in response. As a historical matter, government at either the local, state and/or federal level has acted to regulate every other form of commercial solicitation.¹⁴⁰ As a practical matter, spamming potentially could become so pervasive that, left unregulated by government, it could drive people away from the network altogether.

The TCPA makes it unlawful for any person within the United States "to use any telephone facsimile machine, computer or other device to send an unsolicited advertisement to a telephone facsimile machine."¹⁴² Thus, the focus is on the receiving device; to violate the TCPA, a spammer using a computer must simply send an "unsolicited advertisement" to a "telephone facsimile machine." An "unsolicited advertisement" is broadly defined as "any material advertising the commercial availability or quality of any property, goods, or services which is transmitted to any person without that person's prior express invitation or permission.".¹⁴³ The TCPA defines a "telephone facsimile machine" as:

As appealing as this argument may be to those who wish to make spamming illegal, it presents a problem. If the government were to rely on this textual argument to apply the TCPA to spamming on the Internet, a court likely would find that application to violate the First Amendment. Most of the elements that persuaded the Ninth Circuit to uphold the ban as applied to conventional fax machines are not present in the context of junk mail on the Internet.¹⁴⁶

In the fax context, the government persuaded the court that the ban was reasonably fitted to the government's interest in preventing unfair cost-shifting and preclusion of desired communications by unsolicited advertisements.¹⁴⁷ Neither of these harms are equally presented by unsolicited electronic mail.

There may be some cost-shifting for those who subscribe to commercial online services and who pay by the minute-these people pay for the time it takes to delete and/or download the junk mail. But many others have Internet access for which they pay a flat rate, which means that they incur no marginal cost in dollars and cents for the time it takes to throw out the spam.

Arguing that spam precludes other electronic mail messages is even more difficult. At present, the scale of the problem has not reached the level at which the network's electronic memory has been filled to capacity with electronic junk mail. Generally, electronic mail messages do not go undelivered because the addressee's box is overflowing with spam. The best argument for why message-preclusion is as big a problem with electronic junk mail as it is with fax solicitations is that in both cases what is precluded is timely receipt of a desired communication.

In the fax context, if a business does not receive a desired communication because junk faxes are tying up the line, that communication may eventually get through, but the business will have been precluded from receiving the message in a timely fashion. Similarly, with electronic junk mail, a business has not effectively "received" a desired communication until it is identified. If someone in the business has to scroll through and delete large amounts of electronic junk mail on a daily basis, that junk mail is precluding the business from effectively receiving its desired communications in a timely fashion.

This argument is problematic because it proposes a more subjective standard for preclusion. In the fax context, an unsolicited advertisement precludes a desired communication from reaching the recipient's "premises" at the time the fax is sent because the two machines cannot connect. In the electronic mail context, a desired communication reaches the recipient's "premises" to the same extent that all other electronic mail messages do because the two machines, via the network, connect.

To withstand intermediate scrutiny, the government must assert a substantial interest in imposing the regulation, and the regulation must alleviate a real harm.¹⁴⁸ The government interest in avoiding message preclusion is probably in preventing desired communications from being precluded from reaching the recipient's premises. This would be the case because the government risks its claim to a "substantial" interest in preventing message-preclusion if that interest is defined in the vague terms of preventing only the "untimely receipt" of desired communications. If preventing preclusion from the premises is the government interest, then that interest is present in the ban on unsolicited faxes but not present if the ban is applied to electronic junk mail. Thus, governments probably cannot enforce the TCPA as a ban on electronic junk mail.¹⁴⁹ The government may be able to ban electronic junk mail, but specific findings and potentially a different rationale may be necessary for such a ban to be constitutionally permissible.

This analysis proceeds on the basis of two assumptions: that most communications will be digitized and sent through a unified medium, and that the structure of that medium will be a network with an open architecture.¹⁵⁰ "Open architecture" refers to a decentralized network structure, analogous to the telephone system, within which communications can take a number of paths to reach their destination. In this structure there is no single network operator and no single concentrated distribution point-as there is in a typical cable television system-over which a network operator can exercise private control to restrict communications. An additional assumption is that abundant memory capacity for electronic communications and sufficiently abundant channels of communication will be present so that there is no reasonable argument that the medium suffers from a "scarcity" problem that would provide the basis for content-based government regulation analogous to that in the broadcasting area.¹⁵¹

While this exercise requires a degree of speculation, some recent developments provide a basis for imagining how communication may take place in a unified medium. Evidence suggests that the Internet, particularly the World Wide Web, is beginning to take on attributes that may bring it closer to television as a medium of communication.¹⁵² For example, some content providers have begun to assign "channels" for their information and to package it in the form of "shows."¹⁵³

But because information on the Internet currently travels point-to-point rather than over the airwaves or via satellite signal, some content providers are exploring ways to exploit that attribute of the medium to send customized information to users in what are essentially personalized broadcasts.¹⁵⁴ At the same time, other companies have developed a means of using radio frequencies designated by the FCC for personal communication services (PCS) to send customized information from designated Web sites to computers equipped with an appropriate receiver.¹⁵⁵ While these developments demonstrate how consumers can increasingly use a computer as a kind of television, other developments allow the use of a television as a computer connected to the Internet.¹⁵⁶

In light of these developments, we can fairly assume that instead of our recognized categories of communication-such as mass media, community newsletters, private mail-we will have a continuum of communications on the network, with familiar aspects at both ends of the spectrum. In terms of private communications, we will probably have one-on-one audio-visual conversations, protected by the same degree of privacy enjoyed by those who currently still use wireline telephones. At the mass media end, live video coverage of major events will continue to exist, which millions of viewers may simultaneously access.¹⁵⁷ But in between, one can imagine the presence of a great deal more variety. For example, the category of an "unsolicited" commercial solicitation may become contentious as people can increasingly customize the information they receive via the medium.¹⁵⁸

Small electronic discussion groups among friends may come to include strangers. In these groups, members may put on "performances,"¹⁵⁹ and they may begin to charge for access to these performances. Alternatively, members may include product endorsements or other advertising in their performances in exchange for money, which may be necessary for the production of the performances.¹⁶⁰ Mass media presentations also may be less uniform. The current model of particular shows or newspapers being sent to our homes at particular times could be inverted. Rather than being sent into our homes by producers, information might be made available for us to retrieve at our leisure.¹⁶¹ Assume that standardized, "mass-produced"¹⁶² entertainment and non-entertainment "programs"¹⁶³ will continue to be available. Because we are the ones accessing the program, it need not be the same program for each of us.

Producers of standardized entertainment and non-entertainment programs might engage in "modular" programming. Under this form of production, a program would be composed of standardized modules that could be assembled differently according to consumer preferences. For example, when choosing to view a movie, viewers may be able to select the rated version they want, with additional scenes of sex and violence added in progressing from the G-rated version to the X-rated version. We may even get a choice of plot structure.¹⁶⁴ Assuming that some degree of modular programming comes into being,¹⁶⁵ advertisers may be better able to target their audiences.¹⁶⁶

Under this scenario, the advertisements we receive will depend on the modules we have selected; the distinction between mass media advertising and junk mail will break down. Rather than selling advertising space directly to specific advertisers, programmers may simply sell blocks of space to advertising syndicates. Because programming will not be entirely standard and because the time constraints of fitting the program into a particular "time slot" will not exist, the limits on how much advertising can be attached to programming will be determined by the mix of government and market regulation that emerges in the unified medium. That is, will consumers be deemed to have consented to receive any and all advertisements attached to programming, such that they have no recourse but to not buy programs that include advertisements for, say, guns or sexual aids? Or might the government have a role in regulating the types and amount of advertising that can be attached to programs?

Given the incentives for advertisers, we should expect that the network would be flooded with unsolicited commercial solicitations.¹⁶⁸ If this comes to pass, the worst-case scenario would be a tragedy of the commons in which the entire network would be rendered useless.¹⁶⁹ Everyone's receiver would be so full of unwanted solicitations that the effort needed to sift the useful from the useless would be far too great to make the effort worthwhile. To avoid the tragedy of the commons, some form of regulation is needed. This could be self-regulation by advertisers, market regulation by other participants in the market, and/or government regulation. Under our constitutional scheme, which places limits on the scope of government regulation of commercial speech but which places no limits on private regulation, our cultural preference appears to be in favor of private regulation.

Although there is some precedent for self-regulation by advertisers, the motivation for that regulation appears to be to preempt government regulation rather than to preserve the commons.¹⁷⁰ If the threat of government regulation were neutralized, we probably would not see effective self-regulation by advertisers.

Information filters would thus reduce the threat of junk mail mucking up the system so badly that it grinds to a halt. Users of opt-out filters presumably could set them either to exclude junk mail altogether or to shunt the junk mail to a holding place where it may or may not be read. Users of opt-in filters would get no unsolicited commercial solicitations unless they either gave their addresses to certain mailers-in which case the mail is not "unsolicited"-or if the primary recipients of their addresses intentionally or unintentionally passed them on to advertisers.

For either type of filter to work effectively, communications need to be accurately labeled. To enable someone to opt-out of junk mail without opting out of desired communications, the filter must reliably be able to distinguish the two categories. This is less an issue with restrictive opt-in filters, but if one uses a broad opt-in filter to receive certain classes of communications, accurate labeling is equally important.

If accurate labeling is not possible, only restrictive opt-in filters may make the medium work. But if this limited opt-in filter is the norm, a number of the advantages of the medium are foregone. Not only will each user have to designate affirmatively all the entities from whom she wishes to receive communications, but she also will not receive unsolicited but desired communications. Moreover, were use of this type of filter the norm, people would need to remain cautious about revealing their addresses. This caution might curtail participation in online discussion groups or any other public communication that would result in disclosure of one's electronic address to strangers. Since this result is undesirable and the only alternative is effective opt-out filtering, accurate labeling of junk mail will be important for the medium to fulfill its potential.

Moreover, for opt-out filters and broadly calibrated opt-in filters to work, accurate labeling is required. Why should an advertiser label his communication as an unsolicited commercial solicitation when the likely result is that it will get filtered out by numerous recipients? Why not circumvent the filter by mislabeling the advertisement as a desired or important communication?¹⁷⁴ Assuming that this is the likely response of most advertisers, how can the market alone regulate mislabeled advertisements? It simply cannot.

Thus, some form of government regulation in addition to market regulation through filters will be necessary to stop junk mailers from constipating the network. The three most likely forms of regulation the government might consider would be a labeling requirement, restrictions on commerce in personal information, and a complete ban on unsolicited electronic advertising. If the government is able to force advertisers to label their solicitations accurately, filters should work to avoid the tragedy of the commons. Further, if the government can restrict the ability of businesses to sell customer profiles to advertisers, junk mailers will have less information with which to identify their targets.¹⁷⁵ A total ban on unsolicited electronic solicitations is obviously the most forceful means to regulate electronic junk mail. If these forms of government regulation become desirable or necessary, the question arises: will they be constitutionally permissible?

Before taking this step, we must ask whether the principles from part I are likely to remain good law in the future. Will our understanding of the First Amendment be changed by our communicative efforts in a unified medium? Influential constitutional scholars who have thought about the issue conclude that First Amendment principles should not be viewed as technology-specific, and therefore the challenge will be applying the tried-and-true principles in new contexts.¹⁷⁶ A number of First Amendment scholars have begun to think about how these principles might be applied to speech communicated through the emerging media.¹⁷⁷ Most commentators assert that, so long as the new medium has an "open" architecture, the dominant feature of the new medium will be greater individual control over sending and receiving speech.¹⁷⁸ There is some suggestion that greater individual control may impact the rules for regulating "mass media," but there is little suggestion that that control will necessitate alteration of the principles for regulating unsolicited commercial solicitations. Therefore, this analysis will proceed on the assumption that the part I principles remain the relevant analytic framework.

The three types of regulation that I consider most likely are: (1) a labeling requirement for unsolicited advertisements so that individuals can use their information filters effectively; (2) restrictions on commercial use of personal information; and (3) a complete prohibition on unsolicited electronic advertisements.

The First Amendment places limitations on the government's ability to require that labels be put on unsolicited commercial solicitations. The requirement could be viewed as either a content-based requirement applied only to commercial solicitations or as a content-neutral time, place, and manner restriction on all unsolicited solicitations; in either case the First Amendment analysis is virtually identical.¹⁷⁹ The government will have to demonstrate a substantial interest that is directly advanced by a labeling requirement, and the fit between the interest and the scope of the regulation must be reasonable.¹⁸⁰

A labeling requirement for unsolicited commercial advertisements would probably be upheld as a constitutional regulation. To ensure such a result, the government should assert three independent substantial interests in regulating-protecting individual privacy, preventing unfair cost-shifting, and preserving the viability of the medium. The reason that the government should assert all three interests is that none of these interests alone is neatly transferred from other contexts to the unified medium. The government should then demonstrate how a labeling requirement reasonably advances these interests.

In seeking to impose a labeling requirement for unsolicited solicitations in the unified medium, the government could argue that such a regulation is a reasonable fit with privacy protection because the locus of decisionmaking remains with the individual who chooses whether or not to filter out the labeled communications. The government could also point to current state law, which requires attorneys to label their solicitations as an "Advertisement" to ameliorate the potentially coercive effects of attorney communications and to protect privacy in the home.¹⁸⁴

But would a labeling requirement for unsolicited advertisements in the unified medium really be protecting privacy in the home? Unlike a knock at the door or a ring on the phone, nothing is similarly disruptive and intrusive about receipt of an electronic message. Whether a government-enforced labeling requirement, which enables effective filtering, is sufficiently analogous to the statute upheld in Rowan, which enabled individuals to filter out offensive junk mail, remains to be seen. Such an analogy may be problematic depending on how we conceive of the "place" to which electronic messages are sent.

If the electronic mailbox is merely a part of our home, which is entitled to the fullest protection of privacy, then the analogy may be a snug fit. But if our electronic mailbox is a more public space than the physical mail slot entering our home, the government's interest in protecting privacy in that more public space will be diminished. As one writer has described it:

If the government sought to use this argument in support of a labeling requirement, it should do so cautiously because the reasoning for why prevention of cost-shifting is such a strong state interest or why a complete ban is reasonably fitted to that interest remains somewhat elusive. Neither Congress nor the courts have thus far considered it necessary to articulate precisely where the substance in the interest lies. The First Amendment merely requires that the state demonstrate that the harms it recites are real and that its restriction will in fact alleviate those harms to a material degree.¹⁹⁰ Thus far, the courts have accepted as evidence of real harm unchallenged testimony before Congress that junk fax forces unwilling recipients to pay for paper and ink and to forgo receipt of other transmissions while the unsolicited transmission is received.¹⁹¹

"Cost-shifting" cannot mean general cost imposition because every form of junk communication imposes some costs on recipients,¹⁹² but the government cannot reasonably ban door-to-door sales or other junk communications in order to avoid those costs. On the one hand, a ban based on a cost-shifting rationale may mean that when junk communications impose objectively quantifiable costs on recipients, the state has a stronger interest in protecting recipients from unwillingly spending money on behalf of advertisers rather than from spending time for the benefit of advertisers.

On the other hand, the focus may be on situations in which advertisers shift costs from themselves to recipients. The time cost imposed on recipients of other forms of junk communication is not a cost the advertiser would otherwise bear. But presumably, the cost of paper and ink are costs an advertiser would bear but for the existence of fax technology that allows the advertiser to shift those costs to recipients. If it is this shifting that is the key aspect of the cost-shifting rationale, the derived principle may be that when an advertiser can use a technology to shift costs that the advertiser otherwise would bear to the recipient, the government may reasonably correct that market failure by prohibiting the advertiser from using the technology altogether to make unsolicited solicitations.

Taken as a whole, the interests in protecting privacy in the home, preventing unfair cost-shifting, and preserving the viability of the medium appear to be fairly substantial even though none are neatly transferred from other contexts to the unified medium.

A labeling requirement would place a modest burden on commercial speakers. If the medium has become so pervasive that other alternatives for unlabelled solicitations such as paper junk mail are no longer viable, then the requirement effectively requires all in-bound communications to be labeled. The material costs for placing a label on the communication would be de minimus, and the label helps consumers better obtain the commercial information they want, which is the First Amendment value that justified heightened protection for commercial speech in the first instance.

In addition, advertisers can also communicate with consumers by maintaining "places" in the unified medium where consumers come to them. As with a home page on the World Wide Web, advertisers will be able to post information about their goods and services to a site on the medium.¹⁹⁸ When consumers seek information by entering a search command into their computer, the advertiser's information-if it fits within the consumer's search parameters-may then reach the consumer. In such a situation, the information travels into a consumer's home not as an unsolicited advertisement but as a solicited communication.

Furthermore, a labeling requirement appears to be one of the least restrictive alternatives available to the government. It would simply require that an advertiser identify in a header or in some other standardized format which could easily be read by the receiving device that the enclosed communication is an unsolicited commercial solicitation. Presumably, the advertiser would have the option to add additional descriptions of the contents-for example, "catalogue enclosed"-so long as those descriptions are not misleading. A labeling requirement is merely an enabling provision that allows private individuals to filter their communications. The requirement by itself does nothing to impede the flow of unsolicited advertisements. Therefore, a labeling requirement would be reasonably fitted to the government's interests in protecting individual privacy, preventing unfair cost-shifting and preserving the viability of the medium.

Quite possibly, a provision restricting the sale of personal information such as a person's address, phone number or email address has no First Amendment implications. What is being regulated is the sale, which under most analyses would be conduct rather than speech. A number of privacy advocates support such a restriction, and at least one state legislature recently considered such a provision.²⁰¹

The government could also directly regulate the sale of personal information. A First Amendment issue might arise if such a regulation places a heavy burden on or effectively precludes communication by commercial speakers. But advertisers are unlikely to have the basis for such a claim. Assuming that some form of telephone directory will be available in the unified medium,²⁰² advertisers would find it difficult to argue that privacy restrictions precluded their ability to speak when such a directory would provide the means to address a large audience. This section simply points out that the government has another indirect means of regulating unsolicited advertisements by increasing the costs of information gathering for marketers through prevention of the sale of marketing lists.

Under current principles and precedents, the government may apparently ban this form of communication only if necessary to avoid unfair cost-shifting or as a reasonable time, place and manner restriction. This subsection considers whether either of these arguments would be likely to succeed in justifying such a ban on the unified medium.

Whether this rationale would support a complete ban on unsolicited advertisements in a unified medium depends on how participation in the medium is financed. If most users incur time-based charges for online usage, then a fairly direct analogy to the fax context can be drawn: just as unsolicited faxes impose the unavoidable and objectively measurable costs of paper and ink on the recipient, so too do unsolicited electronic advertisements impose the unavoidable and objectively measurable costs of the online time required to download and delete the solicitations. If participation is free or financed through fixed monthly charges, then the analogy begins to evaporate. Under such a scheme the costs imposed by unsolicited advertisements would be the inconvenience of sorting through the junk. This inconvenience is analogous to the "short, though regular, journey from mail box to trash can" that the Court considered a reasonable cost for consumers to bear.²⁰⁷

In the fax context, however, a complete ban is also justified by the interest in avoiding the added harm of message preclusion. This interest would not be present in the unified medium.²⁰⁸ Whether this added element is necessary to justify a complete ban probably depends on the alternatives available to both parties. On the one hand the government could make the costs of electronic advertisements avoidable by enacting a labeling requirement, which would allow for effective filtering. On the other hand, advertisers would still have an electronic means of communicating with potential customers through their Web page, and advertisers would have the alternative channels of regular mail (assuming it still exists) and the voice component of the medium (the telephone equivalent), as well as door-to-door sales.

Given the wealth of alternatives for advertisers and the fact that the government does not have to choose the least restrictive means available, a complete ban on electronic advertisements would likely withstand intermediate scrutiny if electronic advertisements imposed unavoidable and objectively measurable costs on recipients, even though the message-preclusion problem is absent. If the medium were financed other than by charges for online time, the government interest in preventing the inconvenience of sorting through the junk would probably be insufficient to justify a complete ban. However, the government may be able to use an alternative rationale to reach the same result.

As discussed above, the government could constitutionally ban all unsolicited electronic advertisements if such a ban prevented the imposition of unavoidable and objectively measurable costs on recipients. Because the costs would be unavoidable, a government ban in effect would not be supplanting any private choice. Although a ban would burden commercial speakers and those willing recipients of unsolicited electronic advertisements, the burden would be permissible because it would avoid unfair cost-shifting while leaving open ample alternative channels to advertisers.

If electronic advertisements do not impose unavoidable costs on recipients, a closer question arises if the government attempts to ban all such advertisements. The government may attempt to justify such a ban as a reasonable time, place, and manner restriction. The precedents support both sides. Those challenging such a ban could point to attempts at similar bans on door-to-door solicitations, which have been invalidated as overinclusive.²¹¹ The argument would be that just as a ban on door-to-door sales is overinclusive because it entirely forecloses a valuable avenue of communication for commercial speakers, so too would a ban on unsolicited electronic advertising be overinclusive because it shuts off the primary means of communication with potential customers. The government could respond that a ban on electronic advertising is no different than a ban on automated telephone calls. Just as the ban on auto-dialers reasonably protected privacy while leaving open the avenue of "live" telephone solicitation, so too would a ban on unsolicited electronic advertising protect privacy while leaving open the avenue of electronic advertising through the Web, telephone or door-to-door solicitation, or mass-media advertising. The issue is too close to predict an outcome without further experience.

All of the benefits and the marvels offered by the emerging media, however, may be unobtainable if we allow ourselves to be buried in a blizzard of electronic clutter. The incentives for advertisers to communicate early and often are great, and relying on self-restraint by advertisers to avoid a tragedy of the commons in the unified medium would not be reasonable. Already the phenomenon of electronic junk mail-spam-is growing fast on the Internet. As more potential customers get wired, online advertisers will likely step up their efforts.

The market is responding with sophisticated information filters, but the government will have to provide an enabling regulatory structure for the filters to be effective. Should the government decide as a matter of sound public policy to ban all unsolicited commercial solicitations in the unified medium, it may be possible to do so under certain circumstances.