Both the government and appellee Network Solutions, Inc. ("NSI") seek to portray appellant Name.Space, Inc., formerly pgMedia, Inc. ("Name.Space"), as improperly pursuing "a unilateral effort to make fundamental changes to the Internet’s infrastructure … that the [government] has said should be made by the Internet community." Indeed, the National Science Foundation ("NSF") states flatly that "[t]his lawsuit was plaintiff’s effort to derail the carefully developed [Domain Name System] privatization process."

These characterizations are both incorrect and legally irrelevant. First, it is simply untrue that the litigation was instituted to end run the government’s policy deliberations. To the contrary, both NSF and NSI concede, as they must, that Name.Space’s request for access to the Internet "root zone," NSI’s subsequent denial of that request, and the March 1997 filing of the complaint in this action preceded by many months the initiation of the government’s DNS "policy" process.

More significantly, neither NSF nor NSI reveals to the Court that the government’s June 1998 "policy statement" on DNS privatization is not a legally binding decision. It is expressly not "a substantive regulatory regime for the domain name system . . . . [It] is not a substantive rule, does not contain mandatory provisions and does not itself have the force and effect of law." 63 Fed. Reg. at 31748 (J.A. A622.) Accordingly, the fact is that the government – which refused to take any legally binding action in the Commerce Department policy debate and which declined NSI’s request to clarify Internet jurisprudence by asserting governmental authority over the DNS – now wants this Court to find the very legal authority it declined to exercise in the "policy statement."

NSF’s defense of Judge Patterson’s summary judgment decision is equally unavailing, because the government cannot legitimately support the district court’s sole First Amendment holding that TLDs and domain names are not constitutionally protected speech. The district court’s conclusion that TLDs are simply Internet "area codes," with no expressive content, was not argued by the government below and directly contradicts the sole factual evidence on this question in the record. It is settled that, on summary judgment, the distrust court cannot rule against a non-moving party unless that party fails to adduce any evidence in support of its claim; the court’s role is to construe the evidence and all inferences in favor of the non-moving party. The district court’s sub silentio

1. NSI’s THEORY OF IMPLIED ANTITRUST IMMUNITY IS CONTRARY TO THIS COURT’S STROBL JURISPRUDENCE, IMPROPERLY EXTENDS THE SO-CALLED "FEDERAL INSTRUMENTALITY CASES, AND WOULD LEAD TO INCONGRUOUS LEGAL RESULTS

NSI argues it has "long been settled" that antitrust law "provides absolute immunity to private parties when they act under contract with a government agency to implement a government policy or program in a manner authorized or approved by that agency." NSI Br. at 36. This contention contradicts this Court’s own settled antitrust jurisprudence, equates a scattered handful of (mainly district court) cases with formal antitrust doctrine, and extends those few cases vastly beyond recognition. Even if it were to be adopted by this Court – a question of first impression in this appeal – the so-called "federal instrumentality doctrine" cannot do what NSI would have it do, not the least because its consequences would undermine the Congressional supremacy in authorizing agency grants of antitrust immunity and would protect private actors in their dealings not with the government, but as in this case with marketplace competitors.

1. To Affirm the District Court’s Judgment, this Court Would Need to Overrule Its Own Precedent And Establish A New Legal Theory of Implied Antitrust Immunity

As discussed in appellant’s opening brief, implied antitrust immunity is "strongly disfavored, and [has] only been found in cases of plain repugnancy between the antitrust and regulatory provisions." United States v. Philadelphia Nat’l Bank, 374 U.S. 321, 350-51 (1963). Under this Court’s precedent, implied antitrust immunity is found only where there is a "pervasive regulatory scheme," where an antitrust exemption is "necessary to make the [statute] work," and "even then only to the minimum extent necessary." Strobl v. New York Mercantile Exchange, 768 F.2d 22, 26 (2d Cir.), cert. denied, 474 U.S. 1006 (1985).

Despite NSI’s efforts to paint the "federal instrumentality" doctrine as an exception to Strobl, this Court has never adopted the view that implied antitrust immunity can arise merely from an agency’s unilateral contracting decisions. Strobl makes clear that the ground for an implied exception to antitrust obligations for private companies is some direct inconsistency between the operation of a congressional statutory scheme and application of the antitrust laws. Yet here, the government has disclaimed any regulatory role over the Internet’s DNS system and NSI does not contends that either NSF or the Commerce Department is authorized to extend antitrust immunity to private parties. Thus, the irreconcilable conflict with antitrust liability that forms the basis for Strobl immunity is plainly not present in this case.

To accept NSI’s position and adopt the federal instrumentality doctrine, this Court would need to overrule Strobl and fashion an ad hoc, policy-driven theory of antitrust immunity. NSI argues, without citation, that Strobl is "an entirely different kind of antitrust immunity," NSI Br. at 52, but fails to recognize that under Strobl, it is Congress, not the courts, that is in charge of immunity issues. When Congress legislates in a way that makes application of the antitrust laws impossible or irrational, the courts assume that the legislature impliedly exempted the statutory scheme from the irreconcilable dictates of antitrust. In this odd way, NSI is right; Strobl is indeed "entirely different" from federal instrumentality

immunity because the former is driven by classic judicial concerns and a reflection of the courts’ limited role in interpreting law, while the latter is based far more on perceived notions of what makes "good sense" from an antitrust policy perspective.

There is nothing at all inconsistent with an agency entering into a contract with a private vendor and the imposition of antitrust obligations on that government contractor. General Motors is not immune from antitrust liability to its competitors merely because it sells tanks to the Department of Defense. There are numerous examples of congressionally enacted grants of authority to administrative agencies to immunize private conduct from antitrust scrutiny; had Congress wanted NSF (or the Commerce Department) to exercise such a role, it could easily have added such a provision to the agencies’ governing statutes. In the absence of any such authority, however, Strobl teaches that implied agency immunization should be found only in the narrowest of cases where Congress plainly foresaw a different mechanism than antitrust for regulation of the marketplace. That predicate is completely absent in this case.

Even under NSI’s unprincipled view of antitrust immunity, however, there is no basis to distinguish the role of NSF (and nor Commerce) in DNS from that of any agency which lacks regulatory power over a market or statutory authority for granting antitrust immunity. NSI asserts, again without citation, that the federal instrumentality doctrine is sensible because the private parties "are acting for the government, not because they are caught in a conflict between a regulatory scheme and the antitrust laws." NSI Br. at 54 (emphasis in original). That is wrong for two simple reasons. First, as the D.C. Circuit explained in refusing to accept the doctrine as a ground for affirming an antitrust judgment for NSI, TLDs and domain name registration are not a "governmental function" because, while the National Science Foundation Act "may, or may not, permit NSF to register and renew domain names, . . .we are certain that it does not require NSF to do so." Thomas, 176 F.3d at 511. Second, despite its focus on the "Cooperative Agreement" with the government, NSI fails to tell the Court that although the Federal Grant and Cooperative Agreement Act permits such agreements for a "public purpose," 31 U.S.C. § 6306, neither that Act nor any other statute suggests that cooperative agreements can immunize private recipients. Appellant’s Br. at 21.

Even if NSI were correct that its conduct was "part of a government policy or program," NSI Br. at 41, it therefore does not necessarily follow that NSI should be absolutely immune from all antitrust liability. Such a holding would conflict with this Court’s settled antitrust jurisprudence by creating implied antitrust immunity arising from agency actions for which Congress has not either legislated inconsistent regulation or authorized agency grants of immunity. Particularly disingenuous is NSI’s claim that implied immunity should be found merely because it operated as "part of" the NSF program, when the Cooperative Agreement itself was executed pursuant to a congressional statute that does not permit, and has never been construed to authorize, a contracting agency to confer antitrust immunity. As the Seventh Circuit has explained (in a case cited by Name.Space but not discussed by NSI), antitrust defendants "cannot avoid liability for their actions because requested by some public official to do so. . . . Only Congress can lift the restrictions which find expression in the Sherman Act." Eugene Dietzgen Corp. v. FTC, 142 F.2d 321, 329 (7^th Cir.), cert. denied, 323 U.S. 730 (1944); see Consumers Union v. Rogers, 352 F. Supp. 1319, 1323 (D.D.C. 1973)(President cannot grant antitrust immunity without statutory authority); Appellant’s Br. at 23.

2. The Analytical Basis for NSI’s Proposed "Federal Instrumentality Doctrine" Conflicts With Settled Antitrust Law and Cannot Immunize a Private Party’s Actions Directed Against Third-Party Marketplace Competitors

NSI seeks to elevate its federal instrumentality theory to the status of settled antitrust doctrine by analogizing to Noerr-Pennington immunity and Parker v. Brown "state action" immunity. NSI Br. at 40, 44, 46, 49 & n.31. Of course, while these recognized bases of antitrust immunity are so well-established that they have been affirmed by the Supreme Court, analyzed in law review articles and summarized in leading antitrust treatises, there are no citations to any similar authorities in NSI’s brief. That is because the so-called "doctrine" NSI asserts is merely a smattering of cases, only a few of which have even reached the appellate level, that cannot be reconciled with basic antitrust tenets.

There are so many different verbal formulations of this so-called "doctrine" in NSI’s briefs and the cases it cites that specific rebuttal of each, which Name.Space discussed in detail in its opening brief, would be a waste of the Court’s limited time. It suffices to say that NSI uses a simple syllogism: because the federal government is not a "person" for antitrust purposes, and because a contract cannot be lawful for one party but not another, when the government contracts with a private party that contract cannot be challenged under the antitrust laws. NSI Br. at 42-43. If that were all NSI argued, Name.Space would have no disagreement. As we conceded in our opening brief, that is a "quite unremarkable proposition." Appellant’s Br. at 27-28.

Stripped of the dicta paraded by NSI, the "doctrine" holds only that if a federal agency or federally-created entity (e.g., a federal territory or public corporation) contracts with a third-party private entity to exclusively perform a service (such as public telephone service), a resulting antitrust challenge to the exclusivity of the contract as an unlawful agreement or conspiracy to monopolize must fail. Appellant’s Br. at 27. To suggest, however, that those contractors should be absolutely immune for all activities is a gross misreading of these cases and a dramatic departure from the roots of the doctrine NSI proposes. Certainly, a government corporation or agency is just as absolutely immune from antitrust liability as the government itself. E.g., Sea-Land Service, Inc. v. Alaska Railroad, 659 F.2d 243, 244 (D.C. Cir. 1981)(Alaska Railroad was "an entity wholly owned and operated by the United States"); IT&E Overseas, Inc. v. RCA Global Communications, 747 F. Supp. 6, 7 (D.D.C. 1990) (Guam Telephone Authority was "created a not-for-profit public corporation"). That does not mean, however, that to the extent a "federal instrumentality" is involved, it can provide blanket antitrust immunity to any party with which it enters into a government contract. As the D.C. Circuit itself – which NSI admits decided the "leading case recognizing the ‘federal instrumentality’ doctrine" – has recently indicated, "[w]hether there is, or should be, any such ‘federal instrumentality doctrine’ in this context is not clearly settled." Thomas, 176 F.3d at 508.

It is somewhat surprising that NSI almost completely disregards the Thomas decision, arguing in a single paragraph near the end of its brief that Thomas only relates to whether there should be immunity where the contractor has discretion to perform the contract in a number of ways. NSI Br. at 58. Contrary to NSI’s implication, the D.C. Circuit not only questioned the basic applicability of the doctrine to private parties, but expressly referenced this litigation in stating that "[i]t is not obvious to us . . . that a private party automatically shares the federal agency’s immunity simply because the contractor’s allegedly anticompetitive conduct occurred – as Network Solutions puts it and some courts suggest – ‘pursuant’ to a government contract." 176 F.3d at 509 & n.13 (citing district court decision in this case).

That is precisely the point, which NSI cavalierly overlooks. If an agency cannot by statute immunize its contractor, on what basis should courts imply immunity to conduct taken by the contractor vis-à-vis third-party marketplace competitors? Here is where NSI’s syllogism breaks down. Even if the contractor is equally immune for the contract itself – if the grant of an exclusive monopoly

cannot be challenged – that does not extend to the contractor’s use of the monopoly to harm rivals. NSI’s explanation of the cases is simply false. NSI Br. as 55-57. Champaign-Urbana News Agency, Inc. v. J.L. Cummins News Co., 632 F.2d 680, 681 (7^th Cir. 1980), the plaintiff challenged defendant’s exclusive contract to provide books and magazines to Army and Air Force Exchange Services, but did not challenge use of that monopoly power against rivals. Likewise, in Sakamoto v. Duty Free Shoppers, Ltd., 764 F.2d 1285, 1286 (9^th Cir 1985), the plaintiffs challenged the legality of exclusive concession agreement between defendant and Guam Airport Terminal, but did not claim that the duty-free concessionaire had denied it access to an essential facility for competition. And even if NSI were correct that some cases have allowed the conduct of a government contractor to escape antitrust scrutiny, such a conclusion makes no sense. As discussed in the next section, the district court’s holding below – even when constrained by NSI’s limiting conditions, added only after its Thomas set-back – has no limiting principle and would create serious consequences for administration of the antitrust laws.

3. As Adopted by the District Court, the Federal Instrumentality Doctrine Would Have No Limiting Principle and Conflicts With the Primacy of Congress Over Federal Antitrust Immunity

The district court held that that "private entities under contract with federal instrumentalities are immune from Sherman Act liability for their actions taken pursuant to that contract." (J.A. A853.) That is precisely the formulation advanced by NSI below and in the Thomas litigation. Indeed, under the district court’s rule, federal agencies could immunize conduct without any predicate requirements whatsoever except the existence of contracting authority. (J.A. A856.) ("NSF had the authority to enter into the Cooperative Agreement with NSI and to assign the registration tasks to NSI which could otherwise have been done by NSF itself."). Now, however, NSI would like to amend the district court’s holding to limit it to parties acting "in compliance with the specific direction" of a government agency. NSI Br. at 1, 41, 61. That is perhaps understandable, but nonetheless improper.

The "specific direction" of a government agency is an insufficient basis on which to rest the untoward consequences of adopting a federal instrumentality doctrine as broad as that proposed by NSI. First, as the Supreme Court has made clear, "government contracting officers do not have the power to grant [antitrust] immunity." Otter Tail, 410 U.S. at 378-79. NSI’s argument that in Otter Tail the private actor sought "to advance its own private interest" is hardly distinguishable, as the same is manifestly true here. NSI Br. at 50. Moreover, in Otter Tail an electric power company with monopoly control over the facilities necessary to "wheel" electricity was required under the antitrust laws to provide wheeling access to potential competitors, 410 U.S. at 368, notwithstanding the fact that its contract with a federal agency included restrictions against the practice. Id. at 374-75. Thus, even express contractual directives for a monopolist not to deal with competitors – precisely what Amendment No. 11 finally established in this case in late 1998 – are no defense to the same type of "essential facility" Section 2 Sherman Act claim Name.Space raises in this litigation against NSI.

Second, basic antitrust principles dictate that implied immunity is to be narrowly construed and rarely granted. In the context of "compliance with" federal contracts and the "specific direction" of government agencies, NSI’s approach would subvert Congressional supremacy over the scope of agency authority. Even where, as here, Congress had chosen not to empower agencies to grant antitrust immunity, they would still be permitted to do so merely by giving "specific direction" to a contractor. Thus, using the previous example, the Defense Department could immunize General Motors from price fixing charges by specifically directing it to exchange bidding information with other contractors. This is an absurd amount of power – essentially unreviewable by the courts – to bestow upon contracting agencies. At the very least, such a radical change in the scope of delegated immunization power should, if at all, come from Congress, not the courts. Simply put, NSI is asking this Court to act in a legislative, not judicial role by inventing a new form of agency power where none is needed; Congress can easily do so itself it so desires.

4. Even if NSI Were Correct, the District Court Erred in Applying the Federal Instrumentality Doctrine to Pre-"Amendment No. 11" Liability and To Issues Other Than Preliminary Injunctive Relief

The district court found that until Amendment No. 11 in 1998, NSI had discretion to add new TLDs to the root server. (J.A. 859-60.) That finding is completely in accord with the uncontradicted evidence below, which the district court was required to credit on summary judgment. Contrary to NSI’s ad hominem attack, counsel for Name.Space has not "play[ed] fast and loose with the facts." NSI Br. at 61. To the contrary, appellant made this factual argument below, pointed out that the district court was required to accept it for purposes of NSI’s summary judgment motion, and forthrightly explained to this Court the interplay between the parties’ stipulation and the district court’s holding below. Appellant’s Br. at 32-33 & n.23 (indicating that under stipulation, from the effective date of Amendment No. 11 onward, appellant can only prevail on its request for injunctive relief if the NSF/NSI contract provides no implied antitrust immunity).

1. The District Court Violated Rule 56 in Rejecting the Uncontested Record Evidence Demonstrating that TLDs and Internet Domain Names Have Expressive Purposes

The only evidence in the record on the expressive purposes and functions of TLDs and Internet domain names was the declaration of Professor Milton Mueller, one of the country’s foremost experts on Internet domain names, submitted by appellant Name.Space. Reviewing the usage and treatment of domain names, Prof. Mueller opined that domain names constitute expressive speech as does any other form of Internet communication. E.g., Mueller Decl. ¶ 3. Under Rule 56, the district court is required to construe all evidence in the light most favorable to the non-moving party. Anderson, 477 U.S. at 248; Joyce, 171 F.3d at 133. Since the government failed to introduce any evidence showing that domain names are not expressive or political speech, under Fed. R. Civ. P. 56 these factual issues should have been resolved, as a matter of law, in favor of plaintiff.

The district court not only did not do so, but instead made a series of findings analogizing TLDs to area codes that not supported by any evidence in the record. This error requires reversal. NSF suggests in a footnote that the district court was "acting well within its discretion in rejecting Professor Mueller’s conclusions applying First Amendment doctrine." NSF Br. at 26 n.*. Yet Prof. Mueller’s declaration contained ample factual evidence in addition to his conclusions on applicability of the First Amendment. Even if, as NSF argues, the legal opinion evidence was inadmissible, the portions of the declaration that discussed TLD usage as expression, rather than "appl[ication] of First Amendment doctrine, were plainly admissible and relevant.

NSF refuses to acknowledge, as did the district court, the clear precedent in this Circuit that states that domain names can have expressive function. Planned Parenthood Fed. of Am. v. Bucci, No. 97 Civ. 0629, 1997 LEXIS 3338, *34 (S.D.N.Y. Mar. 19, 1997), aff’d, 1998 WL 336163 (2d Cir. Feb. 9, 1998). Name.Space’s domain names, including "for.mayor" and "microsoft.free.zone," undoubtedly evince the "expressive purpose" that the Second Circuit has required of a communicative message. United We Stand America v. United We Stand, America New York, 128 F.3d 86, 92 (2d Cir. 1997). Thus, appellant has satisfied its burden of showing that its speech warrants First Amendment protection. See Clark v. Community for Creative Non-Violence, 468 U.S. 288, 293 n.5 (1984).

2. Amendment No. 11 is a Clear and Unjustified Prior Restraint on Constitutionally Protected Internet Expression

NSF incorrectly argues that "[p]laintiff has failed to demonstrate any infringement on the ability of plaintiff (or anyone else) to ‘speak’ through use of domain names." NSF Br. at 24. The law is to the contrary. Once Name.Space showed, as it did, that TLDs have a First Amendment function, the burden shifted to the government to justify the prior restraint imposed by Amendment No.11 – under which NSF/Commerce reserve the right to approve or disapprove all new TLDs, and as yet have not allowed any new TLDs to be used on the Internet.

This is a classic prior restraint case. The purported fear that the addition of new TLDs to the root will endanger the "stable transition of the DNS," NSF Br. at 35, cannot compare to the rare situations in which the Supreme Court has upheld prior restraints of speech. Schenck v. United States, 249 U.S. 47, 52 (1919)(preventing publication of troop movements in time of war); Near v. Minnesota, 372 U.S. 58, 70 (1963)("We have tolerated such a system only where it operated under judicial superintendence and assured an almost immediate judicial determination of the validity of the restraint."). Because NSF lacks any equally weighty basis upon which to justify its suppression of pgMedia’s speech, it has committed (and is still committing) an unconstitutional prior restraint of speech on the Internet

3. This Court Cannot Affirm the District Court’s Judgment on Any of the Alternative Theories Advanced by the Government

The Supreme Court defined the inquiry establishing standing in Warth v. Seldin as "whether the plaintiff has ‘alleged such a personal stake in the outcome of the controversy’ as to warrant his invocation of federal-court jurisdiction and to justify exercise of the court’s remedial powers on his behalf. 422 U.S. 490, 498-99 (1975). The injury suffered by Name.Space certainly passes this test.

NSF answers Name.Space’s claim of unlawfully compelled speech by arguing that the First Amendment prohibits only the compulsion of political or ideological speech. This argument is, at the very least, a novel interpretation of Supreme Court precedent.

In the landmark case of Miami Herald Publishing Company v. Tornillo, 418 U.S. 241 (1974), the Supreme Court struck down the Florida right-of-reply statute foursquare on grounds that "the Florida statute fails to clear barriers of the First Amendment because of its intrusion into the function of editors." 418 U.S. at 258. The Court went on to state that "[t]he choice of material to go into a newspaper, and the decisions made as to limitations on the size and content of the paper, and treatment of public issues and public officials — whether fair or unfair — constitute the exercise of editorial control and judgment." Id. Indeed, NSI exercises this control and judgment every day in its self-proclaimed editorial role in domain name registration.

Nor do the additional cases cited by NSF hold that compelled speech is violative of the First Amendment only when the speech is political in nature. For example, the Supreme Court stated in Wooley v. Maynard (cited at NSF Br. at 29) simply that "[w]e begin with the proposition that the right of freedom of thought protected by the First Amendment against state action includes both the right to speak freely and the right to refrain from speaking at all." 430 U.S. 705, 714 (1977). In another case relied upon by NSF, the Supreme Court held that the First Amendment protects individuals from "requir[ing] [them] to repeat an objectionable message out of their own mouths." Glickman v. Wileman Bros. & Elliott, Inc., 521 U.S. 457, 466 (1997). Therefore, the district court erred in holding that forcing Name.Space to register its domain names according to NSI’s own prescriptions is "[f]ar from being compelled speech, the TLD is simply a routing instruction that helps computers find each other." (J.A. A865.)

For all these reasons, the judgment of the district court should be reversed.