.DOC File
Observations on Recommendation 6: Morality and Public Order
Cheryl B. Preston
ICANN has requested public input on a draft of the Request for Proposals (RFP), or Applicant Guidebook, regarding new generic top-level domains (gTLDs). One policy recommendation in the RFP is that the new TLD strings, or names, “not be contrary to legally accepted norms related to morality and public order that are recognized under international principles of law.” (GNSO Recommendation No. 6).
The RFP proposes the following three as categories fitting the language of Recommendation No. 6.
1. Incitement to violent lawless action;
2. Incitement to or promotion of discrimination based on race, color, gender, ethnicity, religion, or national origin;
3. Incitement to or promotion of child pornography or other sexual abuse of children.
The following five comments address this issue:
1. It is time to move on. Those seeking a new gTLD will, at best, have waited two years for ICANN to hash out the requirements for applications. The ICANN staff has done extensive research and consulting. Those who doubt the truth of this claim should ask for supporting documentation or names, rather than dismiss the staff’s conclusions without an equal good faith showing of competing research and expertise. The limitations as stated are extraordinarily minimal. Treating them as a serious violation of free expression principles warranting this on-going attention saps resources and support from real speech issues such as the repression of political speech in China.
2. The limitations protect against “mission creep” of the worst kind. The value of giving a for-profit company a shocking commercial icon does not outweigh the risks of making ICANN and the Internet the obvious target of the culture wars. As illustrated by the milder example of the proposal for a .XXX domain, forays into contested, high-profile symbolic battles diverts ICANN into the kind of social and cultural sensitivities it is ill equipped to handle. The furor of the victims of Nazi Germany need not be unleashed at ICANN. By drawing minimal limitations, ICANN can dodge the worst of these battles. Moreover, ICANN can use the resources, volunteer time, and good will to forward more significant objectives.
There is an interesting tension between the insignificant speech value of string names and their potential symbolic enormity in public discourse. Names of gTLDs are few and obvious. Some become part of the popular vernacular. They may convey little information but provide, in the minds of some, significant social symbolic meaning. What important human freedoms and rights are sacrificed if one of a hundred gTLDS cannot be .spic (or another ethnic slur)? The Internet, and other forums, offer myriad alternatives for freely discussing one’s racial views. Yet, by elevating .spic to gTLD status, the implied approval of racial hatred is unnecessarily amplified.
Some argue that a short string of letters could never be sufficiently offensive to warrant restriction. But some fairly short letter strings could raise scorching public and governmental backlash. Is it necessary to embroil ICANN in the controversy that would accompany, for example, .nazi, .jihad, .KKK, .nigger, .Alah, .Jesus, .kidrape, .fckboys? Is there enough speech value to warrant the costs of having to publicly debate such words?
3. U.S. law is not the test. Opponents to Recommendation 6 frequently assert that it does not reflect U.S. law. The world need not be subject to U.S. law when the deeply held traditions of most other countries place other values on equal or higher footing. The RFP states that the three public order guidelines are based on standards in “most, if not all, countries.” It need not be unanimous. No logic compels the particularities of U.S. free expression law as the conclusive test. A global standard typically requires some chipping at the extreme edges of competing systems. We would not accept the speech restrictions of the ten (or fifty) most restrictive countries; we should not insist on the speech rights of the ten most liberal, especially when the speech value of a TLD string is so minimal. As glorious as free speech is to an American, we must respect that others may equally value stopping race discrimination, protecting children from sexual exploitation, and respecting the lives of Nazi victims.
4. Even if U.S. law were the test, these three guidelines are not unconstitutional.
A. The proper analogy in U.S. law is to broadcast law anyway. If indeed U.S. law is the appropriate standard for ICANN, then the law to apply is the broadcast indecency rule upheld by the Supreme Court in United States v. Pacifica, 438 U.S. 726 (1978). The statute permits the Federal Communications Commission (FCC) to punish broadcast media that allow “indecent” content, a far broader standard than those in Recommendation 6.
It is true that the Supreme Court in Reno v. ACLU, 521 U.S. 844, 845-46 (1997), refused to apply the broadcast indecency rule to ordinary publications made on the Internet. But the Court’s carefully explained reasons in Reno for not applying Pacfica are entirely absent in the case of a TLD. The application of Pacific turns on three things: pervasiveness, scarcity, and history.
Consider the phrase “.com.” It appears on every kind of public signage, radio and television advertising, emails, and talk shows. Its “pervasiveness” and accessibility to children far exceeds the particular content of a single online post that must be sought out and digitally requested, even if one accepts the Reno Court’s determination that the Internet, as a general matter, is not as pervasive as radio and television.
More importantly, the Reno Court relied on the fact that television and radio station licenses are scarce and subject to significant barriers to entry. In Red Lion Broadcasting Co. v. F.C.C., 395 U.S. 367 (1969), the Court explained:
Where there are substantially more individuals who want to broadcast than there are frequencies to allocate, it is idle to posit an unabridgeable First Amendment right to broadcast comparable to the right of every individual to speak, write, or publish. If 100 persons want broadcast licenses but there are only 10 frequencies to allocate, all of them may have the same ‘right’ to a license; but if there is to be any effective communication by radio, only a few can be licensed and the rest must be barred from the airwaves. Id. at 388-89.
Currently, there are only 21 TLDs in the world. Although ICANN has proposed opening more, the number is still extremely limited compared to the number of speakers who would want to register a TLD. The Internet root system cannot accommodate all comers. To be one of the lucky few, an applicant must pay and estimated $185,000 at the outset and $75,000 a year or 5% of revenue, Cost Considerations of the New gTLD Program, Oct. 24, 2008, http://www.icann.org/en/topics/new-gtlds/cost-considerations-23oct08-en.pdf. This is in addition to establishing evidence of significant economic resources, support systems, employees, and expertise – not to mention the competition with others wanting a similar string at an auction and the difficulty of understanding even the application process. When there are scarcity and barriers, the governing authority has a legitimate interest in assuring that the few available outlets are used for higher and more public-serving purposes.
Finally, the Reno opinion stressed the significance of the unregulated prior history of the Internet and the kind of speech already available on existing webpages. This comparison also fails with respect to gTLDs. None of the 21 gTLDs in the history of the Internet use strings that would be circumscribed by the guidelines ICANN proposes for the new gTLDs. Permitting these minimal restrictions would not be an effort to put the cat back in the bag or otherwise undo expectations and use patterns.
B. The Brandenburg Rule does not defeat these guidelines. One issue is how a TLD string could actually “incite” anything (other than public protest and outrage). This, of course, turns on the definition of “incite.” Some argue that the U.S. Supreme Court case of Brandenburg v. Ohio, 395 U.S. 444 (1969), requires “immediate” violence or harm in response to speech to warrant restrictions on speech. This argument is flawed for a number of reasons.
First, the rest of the world need not be subject to a pre-technology statement of a Court that inconsistently applies that “rule” itself, especially in cases dealing with child pornography. This summer, the majority of the U.S. Supreme Court upheld in Williams v. United States, 128 S. Ct. 1830 (2008), a statute prohibiting offering to sell a computer-generated or non-existent sexually explicit image of a minor. Because no actual child was hurt in the image production, the statute must rest on other harms. The dissent was vehement that the statute fails under the Brandenburg rule and that the social harms of child pornography in general are too remote. The majority easily distinguished speech related to pandering, or commercial sales transactions. Similarly, in New York v. Ferber, 458 U.S. 747 (1982), decided more than a decade after Brandenburg, the Court upheld a speech restriction based on long-term inducements, rather than immediate incitement. The Court found the distribution of images depicting sexual activity by children “is intrinsically related to the sexual abuse of children in that it contributes to the distribution network, which must be closed if the sexual exploitation of children is to end.” Id. at 759. The harms of a “distribution network” are not immediate.
Second, other courts and scholars in the United States have made powerful arguments that Brandenburg is inapplicable to Internet communications, a question the Supreme Court has not addressed. For instance, John Cronan argues that the Brandenburg Court in 1969 could not have foreseen the methods of communication that the Internet would produce. See John P. Cronan, The Next Challenge for the First Amendment: The Framework For an Internet Incitement Standard, 51 CATH. U. L. REV. 425 (2002). Internet communications are generally delayed, and the speaker and hearer are almost always in different locations as well as confronting the speech at different times. Cronan predicts the courts will modify the Brandenburg imminence requirement for the Internet, focusing on the listener rather than the speaker, the intent of speaker to incite, and other elements of the standard, rather than immediacy. See also Steven Penaro, Reconciling Morse with Brandenburg, 77 FORDHAM L. REV. 251, 270 (2008).
I do not raise these cases or articles with the intent of suggesting they are the only “True” statement of U.S. law on incitement. (I do not believe U.S. law is the required standard in any event.) I raise them to show that the Brandenburg case does not resolve the question of U.S. law in this area, even if U.S. law were the standard. The nuances in the application of Brandenburg, particularly to the Internet, are such that no definitive answer exists that would warrant ICANN to make policy based on that case.
C. At best, a gTLD string is commercial speech, a pointer, an address, or a license subject to a lesser standard of scrutiny. A gTLD string is not the kind of “communication” that would be subject to strict scrutiny under the First Amendment in the United States. It is a basis of distinguishing the paths of digital packets through electronic routers. The purpose of that speech is as a gatekeeper to control access to a public place used by citizens of all ages and cultures. A city council can refuse to name a street something offensive to many of its citizens, even if it could not prevent a citizen from standing on that street and saying those words.
Outside of the FCC and broadcasting, the most appropriate analogy to granting a TLD is to licensing. Many state statutes explicitly cover public decency in the granting of entitlements. In Texas, an alcohol sales permit is subject to preventing conduct that is “lewd, immoral, or offensive to public decency.” Texas Alcoholic Beverage Code §104.01. The California Department of Motor Vehicles “may refuse to issue any combination of letters or numbers on personalized license plates … that may carry connotations offensive to good taste and decency.” California Vehicle Code §5105. Both of these statutes have been challenged on First Amendment grounds and upheld as constitutional. Although some restrictions that were much broader or subject to unguided agency discretion have been found unconstitutional, ICANN’s gTLD limitations in Recommendation 6 are sufficiently specific, minimal, and clear.
Other state and federal laws also govern public morality in a commercial context. For example, federal debt collectors may not use “obscene or profane language,” when collecting a debt. (15 U.S.C. § 1692d). California statute prohibits any statements or words that are obscene, indecent, or of immoral character on any advertising structure or sign. (California Business & Professional Code § 5402). In Ridley v. Massachusetts Bay Transportation Authority, 390 F.3d 65 (2004), the First Circuit Court of Appeals held that quasi-government companies can prohibit advertisements containing profanity on public buses and subway trains.
In conclusion, arguments asserting that Recommendation 6 violates United States law are based on superficial assumptions about the First Amendment rather than specific legal analysis of the complex case law. This should not be a basis for continuing to delay the new gTLD process.
