April 22, 2002
The battle over ICANN reform is in reality a clash between two often-unstated and wholly incompatible myths regarding the nature and purpose of the organization. We all live our lives by myths -- so we shouldn't dispense entirely with the notion of an institutional ideal. But we can't subscribe simultaneously to two incompatible stories about what ICANN ought to be. To reform ICANN, we need affirmatively to choose which of these two basic models the ICANN community should strive to uphold.
Like all myths, this one has holes. Why these principles should apply only to generic top level domains, and not to ccTLDs or to second and higher level domains, is never explained. IBM has a contractual "right" to ibm.com, if it pays its bill, and doesnít have to operate that domain in the public interest. And where ICANN got the power to be the sole source of recommendations to the Department of Commerce about which top level domains will be "delegated" is never explained.
Once operation of a TLD registry is defined as a "delegated" privilege that is continually conditioned on ICANN's protection of the "public interest," the entire apparatus of ICANN policy-making can be justified. Everything not permitted is prohibited. There can be no suggestion that registries are free to operate their businesses as they see fit unless some binding rule has been adopted with their consent. To the contrary, ICANN's "delegation" to the registry is seen as granting only limited powers -- so that permission is necessary before any new powers are enjoyed (or any services may be offered) -- and a determination of the public interest is required even before two-party contracts between ICANN and a registry may be amended.
But this story is coherent, because it suggests that any registry or registrar ought to be willing to agree to go along with future rules that have the support of most of those affected by such rules. This fundamental bargain is accompanied by the notion that, subject only to technical limitations, promises about escrow, and some reasonable minimum financial qualifications, registries ought to be allowed to create and run new top level domains so as to increase competition and user choice.
This myth is most often attacked on the ground that it cannot lead to an "effective" ICANN (meaning, it appears, an ICANN that can itself make many decisions rapidly). But this story of ICANN is being misunderstood. A consensus-based ICANN would allow decisions to be made to the maximum extent possible on a decentralized, diverse basis. A consensus-based ICANN would make global, mandatory policies only when there is wide agreement that they are necessary, leaving everything else to much more rapid and responsive decisions by registries and registrars, who must compete for registrant customers in the marketplace. The public interest in finding desirable domain names under desirable terms and conditions will be met, this myth suggests, by allowing competition to proceed. Competitors' attempts to attract business will assure registrants of maximum choice among diverse alternatives. In this version of the story, the registry wins a right to continued operation by pleasing customers. Properly understood, this myth makes clear that the Board's role is to intervene only when it receives a clear documented showing that most affected parties agree that some wrongful action must be curtailed.
Similarly, the "delegation and public interest" myth invokes the need to protect a broader social good, but is espoused most often by those who deny democratic participation in ICANNís processes. And the "consensus and competition" story is championed most often by those who support "at large" participation on ICANN's Board, even though their myth would give that Board very little power to make binding rules.
There are many alternative myths that might be advanced to explain what ICANN ought to be when it grows up. It could be a trade association, an international administrative body, or a purely technical standard setting body without the power to require anyone to adopt the standards it develops. But these variations lack the essential elements that make the two current deeply conflicting competing myths so attractive and so central to the debate. These other variations either (1) lack any mechanism for adopting any binding rules (without which ICANN becomes just another private actor that has to compete for customers, and not the bold experiment in "governance" that participants have been hoping for), or (2) lack any mechanism to assure that the participants that create domain name businesses and domain name spaces (registries, registrars and registrants) will have the predominant influence on the selection of rules that are binding on them -- thereby eliminating any mechanism to guarantee meaningful "consent of the governed."
There is no way to reconcile these two core competing myths, because their stories fundamentally diverge about where legitimate governmental/regulatory power comes from and what the consequences should be of inaction by an institution that purports to exercise such power. Either "everything not permitted is prohibited" or "everything not prohibited is permitted." We are either creating ICANN to (1) carefully choose new TLDs so as to serve some "public interest" or (2) set rules that prevent free competition among diverse sets of rules only when there is some clear threat to technical interoperability or widespread agreement among affected parties that an action imposing unjustified harm on others should be banned. We cannot have it both ways.
Like all myths, both of these stories are untrue to some extent. As for the "delegation" story, it's clear that, as a practical matter, a useful top level domain is created by private investment and marketing -- not by fiat. The only real function of ICANN's "delegation" is to avoid collisions. And ICANN knows well that local consumer protection agencies can protect the public interest better than it can. Extraneous interests, like intellectual property rights, can more reasonably be required to go to court if the issue raised by a dispute over rights to a name is more controversial than simple (hold-up style) cybersquatting. If ICANN aspired to serve as the sole source of the right to create important online identifiers and to protect the public interest in the operation of such systems, it could not possibly achieve that mission.
The "consensus" story is also clearly untrue, as a practical matter, particularly as applied to the way ICANN has functioned up until now. This myth's proponents must admit that, even if the consensus-documentation process worked the way it is supposed to, a consensus-based ICANN would produce very few policies. But, when paired with the "competition" part of the story, this is perhaps a feature, not a bug. A consensus-based ICANN producing low numbers of policies would have legitimacy -- and might even be an institution to which the ccTLDs would like to belong. An ICANN allowing easy access to new competition at the registry level might also be one that many of those who disagree with particular ICANN policies could more readily support.
But we respectfully submit that any "effective" reform effort must rest on an acknowledged, express choice between these two incompatible stories. Many specific details fall out naturally from that choice. Failure to address this conflict will, in contrast, lead to an incoherent jumble of political compromises that can only perpetuate the sense of illegitimacy that has dogged ICANN for so long.
So, before you submit your final reform proposal, ask yourself: are we creating an authority to make rules to protect the public interest with respect to a resource that derives from a central delegator (operation of which should be treated as a privilege, and innovation in which should require permission)? Or are we creating an opportunity for anyone who wants to create a new domain name business to do so, subject only to showing some minimum technical and financial qualifications, making promises about data safety in the event of business failure, and agreeing to follow consensus policies shown to be supported by most of those affected?
You must decide. Once you do, the details will be easier to develop (and, indeed, may not matter quite so much).