Grounding the Virtual Magistrate

Jack Goldsmith and Lawrence Lessig

Associate Professor, University of Virginia School of Law; Professor of Law, University of Chicago Law School.

The virtual magistrate must be grounded to real magistrates; it must be anchored in real world legal systems, both within the United States, and internationally. The question we ask in this paper is how. Our aim is to outline the structures of private international law that would allow the determinations of the Virtual Magistrate (VM) to be enforceable in the widest range of legal contexts. Our conclusion is that systems such as the VM could be made enforceable, with the adaptation of very conventional legal materials.

This distinguishes our argument from the argument of cy- berspaceís most prominent sovereignty spokesmen - John Perry Barlow, 1 and David Johnson, 2 for example. We do not argue that cyberspace in general, or VM in particular, should ground its authority on some special claim to sovereign authority. We avoid such arguments because we donít believe they can work - at least, we donít believe they can work just now. Why they canít work we explain in part I below; what could work we outline in parts II and III.

I. The Limits on Sovereignty

Cyberspace is everywhere, but no where in particular. People enter from everywhere; their transactions pass through everywhere; anywhere can touch a part of these transactions, and hence, under traditional notions of sovereign jurisdiction, just about anywhere can claim the right to regulate aspects of any particular transaction. The result, conceptually, is a confusion. It is as if the world has flattened out, with all its many jurisdictions sitting side by side on one plane, and above it hovers a separate space, singular and coextensive with the world below it; and over time, the many jurisdictions lying below interfere with the singular jurisdiction above, with the result, a mess.

The specter of this confusion has led some to argue that law should recognize in this space a separate jurisdiction, or in some cases, a separate sovereignty, and that respect or comity should be granted the legal determinations of this sovereign. Whether through comity, 3 or through formal recognition, these arguments push for the recognition of a new and unified authority in cyberspace, grounded in the law of no particular sovereign, but extending to all. As perhaps befits something as radically new as cyberspace, these people argue, something radically new in legality is needed as well.

Or perhaps something radically old. For the precedent for such authority (not grounded in any particular sovereign but binding a cross sovereignties) is not hard to find historically. At the founding of the United States, bodies of customary law floated across American jurisdictions without any real connection to the sovereignty of a particular jurisdiction. Customary international law (formerly known as the law of nations), the law merchant, the common law - these were all bodies of law that were in different ways enforceable in both state and federal courts in America. 4 They functioned in just the way these citizens of cyberspace desire a law of cyberspace to function - as a unified and universal set of authority for resolving disputes, without the conflict of real space law.

Swift v. Tyson 5 is the most famous example of this body of customary law, though it is certainly not the only example. In Swift, the Supreme Court, Justice Story writing for the Court, held that the Judiciary Act of 1789, 6 when it spoke of the "laws of the several states" did not mean to refer to the common law of the states. Hence federal courts could ignore determinations by state courts about the meaning of the common law. Rather this law, the common law, was distinct from the statutory law of any particular state, and because not within the federal legislative author- ity, it was distinct from the statutory law of the nation as well. Because distinct, it could be interpreted, and developed, by federal courts. And thus was born the practice of general federal common law in the federal courts.

Swift thus recognized a body of authority that was binding in local courts, but which had its origin in the actions of no particular sovereign - just the kind of authority that cyberspace activists want. 7 What these jurists want, whether modeled on the law merchant, or common law generally, is a structure of authority in- dependent of any particular sovereignty, yet binding within every sovereignty. What they want, that is, is a structure of common authority to simplify the conflicts among diverse real world au- thorities.

But customary law in the sense that Swift understood the term is gone. We cannot recover it in it s original sense, and it is important to understand just why. We canít go back to the authorities of the world of Swift, because the meaning of "authority", or "law" that they had has changed. Swift comes from an era of legal thought that was pre-positivist. 8 Positivism drives the jurist to locate the source of law in a sovereign before that law can be recognized as authoritative. But Swift, and customary law in general, comes from a period before this theory of law had become dominant-when legal theorists didnít much care to trace law to its source. Law could be recognized as normative, and authoritative, without locating the ultimate agent responsible for this normativity. Custom, whatever its source and however it had been transformed, could bind.

We donít see law like this anymore. As Justice Brandeis said in Erie R.R. v. Tompkins, 9 "law in the sense in which courts speak of it today does not exist without some definite authority behind it." 10 We just canít see law, or at least authoritative law, unless we can trace it to some source. And the problem with common law, or the law merchant, or customary international law, or with customary law generally, is that each is a law that has no clear source in the actions of a sovereign.

Perhaps, however, one could argue that the source of authority did come from the sovereign, acting through the courts. And indeed, simultaneous with the rise of positivism was the rise of a Holmsian realism: no longer could courts pretend that they were merely "finding" the common law; more and more, it became clear that they were "making" it. But for federal courts at least, this new understanding presented a double barreled problem: not only was it increasingly embarrassing for federal courts to be seen to be making law, it was al so constitutionally embarrassing for federal courts to be seen to be making what all conceded was state law. The rise of positivism, and its concomitant focus on the generative authority behind any legal authority, rendered customary law more and more tenuous. 11

For the Supreme Court, what that meant in 1937 was that the common law making that federal courts had been engaging for 96 years we now seen to be unconstitutional, and had to be relegated again to state courts alone. 12 What it meant for theorists of customary law generally-and customary international law in particular-was that authority for customary international law had to be rediscovered through the notion of consent. 13 This was Joseph Bealeís tack with customary international law, and it is the clue to what we argue is the only tack possible with the law of cyberspace. Contract then replaced custom and provided the source of authority in law, as it must again. For again, "law in the way we speak of it today" would not exist without such a source of authority.

We may be on the edge of a post-positivist legal culture. We might, that is, entering again a stage where we can understand a rich customary authority not grounded in the actions of any particular sovereign. Such is the emerging understanding of certain aspects of international human rights law, and such indeed may be the product of the emerging norms of cyberspace itself. Cyberspace, that is, may become a model for understanding again how there could be authority without a law giver. 14 But our argument is that any system that today wishes to establish itself as an enforceable system of authority must acknowledge the positivist premise within this legal culture. And to acknowledge this premise means to look for an authority for a law of cyberspace that is grounded quite directly in the law of real space.

That is our tack here. In the sections that follows, we map out a path with which VM judgments can be considered authoritative within a wide range of legal systems, without requiring these sys- tems to accept some radically different understanding of legal au- thority.

II. The possibilities of the Technology

While law "in the sense in which courts speak of it today" may limit the possibilities of developing a customary authority enforceable within existing sovereignties, technology expands these possibilities. It expands them because the technology of cyberspace is a technology for contracting, and through a system of contracting, law-in the sense of enforceable obligations-can be constructed. This is the key to overcoming the limits for a law of cyberspace imposed by a positivist legal culture. In this section we sketch just how this key fits. 15

By a technology of contracting, we mean just this: Contracts are self-imposed obligations, enforceable through a legal system. 16 Each component in that definition is essential: The authority for an obligation comes from its being self-imposed; its significance comes through its enforceability.

"Enforceability," however, is itself underspecified, for a contract can be enforced either through self-enforcement or state-en- forcement. An efficient regime of enforcement would embrace that type of enforcement that is most effective at the cheapest cost. Often, obviously, self-enforcement (or generally, non-state en- forcement) is cheaper than state enforcement. The technology of cyberspace makes this even more so. For many of the mechanisms of enforcement can be built into the architecture of cyberspace itself, through conditioning (through software) access based upon conforming to the obligations of enforcement.

The point is a commonplace: One enters cyberspace only through software; that software defines the terms on which one enters. These terms are the terms of a contract, and to some extent, these terms can be enforced by the software itself. If the terms of access limit the size of oneís email in box to 500 messages, in a sense that is a mutual obligation between the parties to the on-line service contract; but it is an obligation of a special kind. The limit is self- enforced; the code does the enforcement. No external authority is required to assure that no more than 500 messages are collected in the in box. After 500, messages are bounced. Obligations are "enforced" through the privatized system of enforcement called the computer code.

Of course one might well question whether such "terms" should be enforceable, at least by courts. To the extent that there is no dickering over such terms, some argue that these "contracts" would be contracts of adhesion. But we reject that view. In our view, that so long as there is a range of alternative options for access to cyberspace, the doctrine of adhesion should have no place in the law of cyberspace contracts. The predicate for a sensible application of a doctrine of adhesion is market power. But cyberspace abhors a monopoly, and given the low cost of startup, there is little reason to fear what cyberspace abhors.

If cyberspace facilitates the self-enforcement of contractual obligations, then, it makes it easier to imagine contractual regimes in places where there is no adequate state-enforcement mechanism. If much of the enforcement of obligations entered into can be encoded within the system of exchange itself, then the system need not rely to any significant extent on external systems of enforcement. While of course, at some point, a link to an external legal system is essential, that "some point" need not be very close to the core of the obligation producing exchange. The software of cyberspace makes less significant the softness of international legal space.

The significance of this point about contracting for the VM project should be plain. As the VM project is described, it is to re- solve disputes arising out of network interaction, or more generally, cyberspace. The VM project has not set itself up simply as a more efficient dispute resolution board, resolving disputes of whatever nature. Its scope is limited, and this limitation is valuable. For by focusing the magistrates on their comparative advantage, the project both facilitates respect through reputation, and more importantly, jurisdiction through agreement.

The former (respect through reputation) will attract cases after conflicts have occurred-ex post dispute resolutions. This jurisdic- tion will be important, though not as significant as the jurisdiction we describe below. For even though both parties will often want a speedy resolution of the case, often there will be a clear advantages for one party to delay the resolution of a dispute. In those cases, the VM needs a mechanism to push parties into arbitration.

But it is the jurisdiction through agreement that will constitute the most significant part of VMís future, and here the relevance of the cyber-contracts that we described becomes plain: For the aim of the VM project should be not only to provide a forum for ex post dispute resolution; it should also be to work to induce networks to include reference to the VM ex ante in the terms of agreement for joining the network. Access to the network through these service providers would then be conditioned upon agreeing to VM resolution of on-line disputes. 17

This conditioning would function in two ways. First, simply by making VM dispute resolution (VMDR) a part of the contract, the effect would be to channel disputes to the VM. But more im- portantly, and second, because access has been conditioned upon agreement to resolve disputes through the VM, the service provider can compel cooperation once an event has been noticed for VM resolution. This compulsion can be either by limiting access, or more effectively, by charging against a bond that the user may have presented upon gaining access to the system. (And bond, here, need be little more than a credit card.)

In both ways, the system of cyberspace contracting provides a way to induce people to accept, as a condition of access, an arbitra- tion rulings system to resolve disputes raised on-line. This would be a big step towards the establishment of a significant cyberspace jurisdiction over cyberspace disputes. But even if fully successful, it may be that this will not be enough to assure participation by a sufficiently large percentage of on-line users. A second step may then be to add this ex ante conditioning at the network level. At this stage, networks (through the adjustment of network protocol structures) would screen out networks where the participants have not agreed to some form of arbitration as a means to resolve dis- putes.

It is beyond the present paper to sketch how this network discrimination might be implemented. It is enough simply to suggest it as a possibility. Both function as a type of mandatory insurance for the infobahn. In both cases, what the insurance required does is assure a simple way to resolve claims, and to provide some assurance that the claims can be satisfied.

III. Anchoring to Real Space Law

In this final section, then, we outline the legal structures needed to make the proposed VM system work in the real world. The analysis makes three assumptions. The first is that cyberspace cannot simply opt out of real-world legal and dispute resolution regimes. (This is the point about positivism that we discussed in the first part.) The second assumption is that there is a workable mechanism for parties to consent ex ante to both a dispute resolution regime and a governing law regime. (This is the point about conditioning access that we discussed in part two.) The third as- sumption is that the VM system aspires to be a genuine dispute resolution mechanism, as opposed to a form of non-binding dispute resolution (like mediation or negotiation). There is some question about this latter point in the proposal, which states that "Ď[u]se of the Virtual Magistrate for immediate resolution of disputes would not preclude traditional litigation." We believe that the VM project can be something more than mere mediation, and we assume that is its aim in the analysis that follows.

We address the problem of how to make the VM system work in three parts. We will first discuss in very general terms the legal structure necessary to render VM judgments binding and enforce- able. We then demonstrate that the legal regime that currently governs international arbitration provides a well-tested model for making the VM system work, and we propose a system by analogy to this system. Finally, we discuss some residual problems with the international arbitration analogy. These problems can all be grouped under the heading of governing law.

1. How to Make VM Rulings Binding: General Principles

To make VM rulings binding, the parties must, quite simply, be able to agree to make them binding. For this to happen, three conditions must be satisfied.

First, the parties to the dispute must consent (ex ante or, less effectively, ex post) to have their disputes resolved by the VM. We have already discussed how one might achieve an ex ante agreement to VM dispute resolution (VMDR) in cyberspace. An agreement to VMDR would then be a condition of access to the net through a participating access provider. This condition would track the contractual terms in real world agreements, providing, for example, that "all disputes arising out of or related to the contract shall be resolved by arbitration." 18 Usually the place of arbitration (e.g. New York), the arbitral rules (e.g. the ICC Rules) and governing law are also specified. We can imagine at least two different types of consent to VMDR: (a) an open ended jurisdiction (e.g. "all disputes that arise out of or relate to transactions in cyberspace"); or (b) a more specifically-delineated jurisdiction (e.g. "all disputes that arise out of or relate to" a specific list of transactions).

Consent to VMDR is by itself worth very little. There must also be background enforcement mechanisms that make the consent to VMDR valid and enforceable. Some cyberspace sanctions (through bonds or software, as described above) will provide the initial thrust of enforceability. But these sanctions will ultimately work only if the initial VMDR agreement is itself valid in virtue of a governing law and ultimately enforceable in some national court. 19 Otherwise, the party that wants to resist VMDR in spite of the agreement could initiate litigation in a national court and argue that the agreement is unenforceable. 20 Unless the court recognizes the agreement as valid, the agreement is ineffectual. Thus, to ensure enforcement of the agreement, there must be a national court that is willing (a) treat the agreement as valid under the governing law; (b) specifically enforce the agreement; 21 and (c) enjoin litigation in derogation of the agreement. Moreover, it is not enough that a single national court be available to enforce VMDR agreements. There must be coordinated enforcement among other national courts. Otherwise, a recalcitrant party could litigate in derogation of the VMDR agreement in a non- VDMR-enforcing national court and the VMDR agreement will be ineffectual.

Finally, national courts must recognize and enforce the VMDR ruling. Recognition is a shield: it means that the VMDR ruling will operate as a res judicata bar in subsequent national court litigation. Enforcement is a sword: it means that a national court will treat the VMDR ruling as a valid judgment and, for example, execute the judgment against assets of a recalcitrant party.

Before turning to consider whether and, if so, how these three conditions might be satisfied, it is worth noting that this system is not as litigious, and not as intrusive on cyberspace, as it might at first glance appear. It might seem that under this system every cy- berspace dispute will end up in a national court. This is untrue. Once the system is in place and the validity and enforcement rules are settled, the VMís authority will be established and recourse to national courts will be infrequent. The VM system will work, like most contractual relations, without the need for judicial intervention.

The point can be seen by considering the analogy of arbitration. Arbitration is a private dispute resolution mechanism that derives its authority from the consent of the parties. Almost all arbitrations - whether international or domestic, labor or commercial - rely on a three-part legal superstructure like the one outlined above. Although arbitration disputes sometimes end up in court, these court cases constitute only a very small percentage of all arbitrations. Most arbitrations operate as a matter of course within their own institutional framework without the need for judicial review. But this mostly self-contained system of private dispute resolution is made possible only because arbitration agreements are legally valid in virtue of some law and because a judge will, if asked, enforce arbitration agreements and awards.

2. How to Make VM Rulings Binding: The International Arbitration Analogy

Can these three conditions - ex ante consent to VMDR, judicial enforcement of VMDR agreements, and judicial enforcement of VMDR rulings - plausibly be satisfied? The answer is yes, but not without difficulty. The best place to start is with the well-tested and workable regime that governs international arbitration. 22 This analogy is useful because the problems solved by international arbitration are very much like the problems faced in cyberspace. The goal of both systems is the resolution of disputes that arise from transactions with multiple jurisdictional contacts in a manner that is shaped by the parties, that is neutral and efficient, and that involves minimal intervention of national law and national courts.

The legal regime that makes possible international arbitration is complex. But it has been tested for almost 40 years, and it works. It involves the interplay of four layers of legal regulation.

The first layer of legal regulation is the law of the contract as embodied in the arbitration agreement. An arbitration agreement is a provision in a contract in which the parties to the contract agree to resolve by arbitration certain disputes that arise out of or relate to the contract. The arbitration agreement usually specifies the place of arbitration and the procedural rules that govern the arbitration. The parties also usually specify the law that governs the contract.

The second layer of legal regulation is the procedural arbitration rules. These rules are adopted by the consent of the parties, and are usually specified in the arbitration agreement. These rules govern all aspects of arbitral procedure, such as the appointment and challenge of arbitrators, pleadings, discovery, hearings, and the form of the final award. Arbitral rules come in a variety of prepackaged forms. 23

The third layer of legal regulation is the national arbitration law. (In the United States, this law is the Federal Arbitration Act. 24 ) A national arbitration law defines the scope of permissible arbitration within the country, renders arbitration agreements within this scope valid, and provides various forms of judicial assistance for, and judicial review of, arbitration.

The fourth layer of legal regulation is the international en- forcement treaty. By far the most important such treaty is the New York Convention on the Recognition and Enforcement of Arbitral Awards. 25 The Convention obligates the national courts of signatory states to recognize and enforce arbitration agreements and awards, subject to limited exceptions. 26

This four-layer legal regime could be modified to apply to cy- berspace. The first two layers - consent to VMDR and to particular procedural rules - could be accomplished, as explained above, at the point of entry. National arbitration laws could be modified to include dispute resolution in cyberspace. 27 For example, the FAA would require modification in only two important respects. 28 First, the FAAís requirement that the arbitration agreement be made in writing would need to be modified to accommodate cyberspace realities. Second, some of the FAAís rules turn on the place of the arbitration. 29 These rules would need to be modified to account for cyberspaceís non- geographical locus. 30 Similarly, the New York Convention could be modified to serve VMDR. It would require modification of the writing requirement and certain geographical requirements in a manner similar to national laws.

So the good news is that international arbitration presents an analogous and potentially workable model for VMDR. Unfortunately, there are serious disanalogies. The cyberspace problem is more complex than the international arbitration problem. Below we touch on some of these problems. They can all be grouped under the heading of governing law.

3. Governing Law Problems

We have already addressed one governing law problem: The law that governs, and validates, the agreement to VMDR. As we have explained, the validity of the VMDR agreement will depend on the interaction between national arbitration laws and an international enforcement treaty.

An even more challenging problem is presented by the law that the VM will apply in resolving the merits of cyberspace disputes. This is a complex problem. But it must be faced if the VM system is to work. For the moment we will assume that the only legal problems that will arise in cyberspace involve the making and enforcement of contracts. Even on this assumption, the task of designing a governing law regime is hard. And of course this assumption is false. We will remove the assumption at the end to discuss different problems.

There are several possibilities for a governing contract law regime. One possibility is to let the VM apply extant national choice of law rules to determine the law that governs the merits of a dispute. This is a bad idea, because it is unclear which countryís choice of law rules would govern any particular dispute, and because modern choice of law regimes in any event are not sufficiently determinant even in the real world, much less as applied in cyberspace. A different possible solution is a uniform choice of law treaty regime. 31 This isnít a very good idea either, because agreement will be difficult to obtain and because a choice of law process in the cyberspace context will invariably be difficult, even if the rules were relatively determinate (which they rarely are in modern times) and uniformly applied.

Rather than establish a choice of law regime, the better idea is to establish a uniform law so that all disputes in cyberspace are governed by the same law . This approach makes (most of) the choice of law problem go away. Uniform law has proved to be a useful antidote to choice of law problems in similar transnational contexts, ranging from the ancient commercial communities that developed the law merchant and the law maritime, 32 to the UCC, to the Convention on the International Sale of Goods (CISG). 33

The decision to develop a uniform law still leaves open many options. First, nations of the world might enter into a treaty, like CISG, in which they agree that certain sales disputes in cyberspace shall be governed by the uniform international law embodied in the treaty. This option is probably not a great idea. It might work for sales disputes; CISG, like the New York Convention, can perhaps be modified to suit cyberspace. But CISG addresses only a subset of all contract law; for the plethora of disputes that arise in cyberspace, this option is no less daunting than the creation of a world government. And in any event, it is unlikely that governments will establish law that cyberspace users find conducive.

A better idea is to develop a private legal regime. This can be done in one of two ways. First, the cyberspace sovereign could de- velop a uniform law that would govern all contractual disputes in cyberspace. 34 The parties would consent to this law just as they consent to VMDR. This uniform cyberspace law could be mandatory law or default law. 35 Second, if the uniform cyberspace law is merely default law, the parties might instead choose to design their own legal regime. They could do so by either by opting out of certain default cyberspace rules, or by choosing a particular national law to govern their contract (also with the possibility of selective opt-outs). Each of these private legal options has strengths and weaknesses.

Here is where the real problems begin. Countries do not as a general matter permit parties to have plenary control over their contractual relations. Instead, certain mandatory "public policy" laws apply regardless of party intent or consent. For example, the United States will not permit parties within its jurisdiction to enter into contracts in violation of the Sherman Act. 36 Such contracts are simply not enforceable. What is a VM to do when uniform cy- berspace law, or the law otherwise chosen by the parties, does not include this mandatory law? 37 Can she ignore the mandatory law? Will there be adverse consequences for subsequent enforcement of her ruling if she does? Does she even have the authority to apply the mandatory law? If she does apply it, hasnít she violated the rule of party consent? Will this cause enforcement problems?

These and many other related problems arise in international arbitration. There are as yet no perfect solutions, but there are pretty good ones. As a general matter, the New York Convention permits an exception to the obligation to enforce arbitration agreements and awards when the arbitration involves a nonarbitrable subject or when enforcement of the award would violate a strong public policy. 38 These exceptions - especially the one concerning nonarbitrability - have in general been narrowly construed, and increasingly so. 39 Moreover, arbitrators have been very clever at making arguments, grounded in party consent, that permit them to apply mandatory law not expressly chosen by the parties. 40

Many of these solutions might apply in the VM context. But there are two additional problems in the VM context not present in the arbitration context. The first has to do with the fact that cyberspace implicates many more legal problems than contractual relations. International arbitration is primarily limited to the resolution of contractual disputes. 41 Can we design a dispute resolution regime to deal with legal problems more comprehensively? The governing law problem is immense. Will cyberspace sovereigns offer such comprehensive law? Will nations permit the development of this law?

Assuming that these problems can be resolved, the problem of mandatory national laws remains. And the problem is exacerbated because the typical cyberspace transaction potentially implicates many more national jurisdictions than the typical real-world commercial contract, which usually implicates two national juris- dictions. This means that many more nations will arguably have pertinent mandatory law to be taken into account. Solutions here will be harder. Uniformity of mandatory laws is desirable but not likely (although this is what is slowly being achieved, albeit indirectly, by the New York Convention enforcement regime). And the choice of applicable mandatory law is a conceptual nightmare when multiple jurisdictions are involved.

But here again the possibilities of contract - at both the private and public levels - present themselves. For to the extent that there can be some self-enforcement of these agreements using the network itself, conflicts with national law can be made less signifi- cant. And there might be ways to coordinate, or make more uniform, certain mandatory laws at the treaty level. No doubt there is no way in principle to know whether debilitating conflicts will nonetheless remain. But the strategy for reducing them should be clear, and it is that strategy that has been our focus.

4. Comparative Effectiveness

We have thus far focused on the resolution of contractual dis- putes. The international arbitration model is most helpful in this context. But as we have already suggested, the model also applies to non-contractual issues. Indeed, arbitrators frequently apply law common law other than contract law (especially tort law), as well as a variety of statutory laws. 42 They do so by broadly interpreting the parties agreement to arbitrate "all disputes arising out of or related to the contract to include these noncontractual issues. And national laws are increasingly construed to permit the arbitrability of these non- concontractual issues.

What this means is that in theory, an entire legal regime gov- erning most legal issues that might arise in cyberspace could be grounded in party consent. But the expansion of the international arbitration model outside of the contract law context presents two special problems for cyberspace that should by now be familiar. First, there are some mandatory law issues - criminal law is an obvious example - that nations will simply not permit the VM to apply. And second, even if nations are generous in permitting the VM to apply mandatory law, the question of which of many mandatory laws remains to be solved. As the areas of legal regulation in cyberspace expand, the coordination and harmonization of the choice of law problem will prove very difficult.

Conclusion

Revolutions promise drama, but change comes in the detail. We have argued that the revolutions of cyberspace may promise drama, but that changes in the source of legal authority governing life in cyberspace will come at first through detail. Our aim has been to sketch a path that will make it possible for the VM project to gain ground in the effort to establish an efficient and effective system of dispute resolution in cyberspace. And our strategy has been to map a path that begins with the ordinary legal material that already exists.

It should be plain, however, that the success of this new regime of authority depends upon the success of project to more effectively deliver justice. One suspects that the relative success of the law merchant, or the early common law, to command obedience, the lack of a sovereign notwithstanding, came in large part from the sensibility of the authority this custom developed. The VM project has limited itself appropriately, so it too could develop a set of authority viewed as similarly sensible. Our expectation is that this success will breed its own success.

Footnotes

1 See John Perry Barlowís "Cyberspace Independence Declaration", www.eff.org/~barlow.

2 See David Johnson and David Post, Fencing the New Frontier: "The Rise of Law in Cyberspace," STAN. L. REV. (forthcoming 1996). Johnson and Post are not arguing for a new sovereignty, though they do advance an argument for a separate law of cyberspace.

3 This again is Johnson and Postís argument. See supra.

4 See Curtis Bradley and Jack Goldsmith , Customery International Law as Federal Common Law: A Critique of the Modern Position, HARV. L. REV. (forthcoming, 1997). On the law merchant, see Friedrich K. Juenger, American Conflicts Scholarship and the New Law Merchant, 28 VAND. J. TRANSATíL L. 487 (1995); Chris Williams, The Search for Bases of Decision in Commercial Law, 97 Harv. L. Rev. 1495 (1984) (Book review). See also Matther R. Burnstein, Conflicts on the Net: Choice of Law in Transnational Cyberspace, 29 VAND. J. TRANSNATíL L. 75, 108-10 (1996); I. Trotter Hardy, The Proper Legal Regime for "Cyberspace", 55 U. PITT. L. REV. 993, 1019-20, 1022-25 (1994).

5 41 U.S. (16 Pet.) 1 (1842).

6 Ch. 20, ß34, 1 Stat. 73, 92 (1789).

7 The law merchant and law of nations were not immutable transnational law. That is, sovereigns could, and did, change the domestic force of these laws through positive law. Thus, even in its heyday, the law of nations was, in most contexts anyway, alterable by the sovereign. For an explanation of this point, see The Antelope, 10 Wheat. 66 (1825), in which Marshall reasoned that although the law of nations permitted the slave trade, many sovereigns had, through positive statutes, abrogated the application of that rule.

8 For a more general discussion of this change, see Lawrence Lessig, Understanding Changed Readings: Fidelity and Theory, 47 STAN. L. REV. 395, 426-33 (1995).

9 304 U.S. 64 (1938).

10 Id., at 79.

11 See Lessig, supra, at 428.

12 Again, this is what Erie held.

13 A good explanation of this change (which you should look at in any event for your Erie-effect stuff) is Edwin Dickinson, Changing Concepts and the Doctrine of Incorporation, 26 Am. J. Intíl L. 239 (1932).

14 A post- positivist law would rely upon the understandings of normative authority drawn from analysis such as ROBERT ELLICKSON, ORDER WITHOUT LAW: HOW NEIGHBORS SETTLE DISPUTES (1991). Compare Henry H. Perritt, Dispute Resolution in Electronic Network Communities, 38 VILL. L. REV. 349 359-60 (1993).

15 For a different view, see Henry H. Perritt, Dispute Resolution in Electronic Network Communities, 38 VILL. L. REV. 349, 354 (1993) (arguing "technology makes enforcement more difficult at the same time it increases pressures for overarching rules.").

16 Compare David R. Johnson & Kevin Marks, Mapping Electronic Data Communications Onto Existing Legal Metaphors: Should We Let Our Conscience (and Our Contracts) Be Our Guide?, 38 VILL. L. REV. 487 (1993); Perritt, supra, at 356, 363.

17 On the use of agreements such as this, see Matther R. Burnstein, Conflicts on the Net: Choice of Law in Transnational Cyberspace, 29 VAND. J. TRANSNATíL L. 75, 100-02 (1996); I. Trotter Hardy, The Proper Legal Regime for "Cyberspace", 55 U. PITT. L. REV. 993, 1028-32 (1994).

18 Phrases like "All disputes arising out of or related to the contract" have been interpreted to include not just contract claims, but also common law tort claims and a variety of statutory claims like antitrust, securities, and RICO. See, e.g., Mitsubishi Motors Corp. v. Soler Chrysler- Plymouth, 473 U.S. 614 (1985) (antitrust); Roby v. Lloyds Corporation, 996 F.2d 1353 (2d Cir. 1993) (securities and RICO); Fleck v. E.F. Hutton Group, Inc., 891 F.2d 1047 (2d Cir. 1989) (common law tort).

19 The cyberspace sovereignís power to enforce sanctions (by bonds or exclusion) is only valid if the private parties have the power to confer such power on the cyberspace sovereign. And whether private parties have such power depends on the background national laws governing contracts and the enforcement of contracts.

20 The recalcitrant party might also challenge any subsequent VMDR ruling in court. The defense of the cyberspace plaintiff (and the VM) is the recalcitrant partyís consent to the VMDR, a defense that is useless if initial VMDR agreement is not valid and enforceable.

21 That is, by ordering the recalcitrant party - the party that wants to litigate in a national court rather than, as agreed, in cyberspace - to participate in the VMDR.

22 For excellent introductions to international arbitration, see GARY BORN, INTERNATIONAL COMMERCIAL ARBITRATION IN THE UNITED STATES (1994); ALAN REDFERN AND MARTIN HUNTER, INTERNATIONAL COMMERCIAL ARBITRATION (2d ed. 1991).

23 The most prominent international arbitration rules are those promulgated by the International Chamber of Commerce, the American Arbitration Association, the London Court of International Arbitration, and the United Nations Commission on International Trade Law ("UNCITRAL"). These rules are generally similar but contain important differences. For an overview of these similarities and differences, see Gary Born, supra note 22, at 10-16, 50-96.

24 9 U.S.C. ß 1 et seq.

25 The Convention is reproduced at 9 U.S.C. ß 201. For general commentary, see Born, supra note 22; A. VAN DEN BERG, THE NEW YORK CONVENTION OF 1958 (1981).

26 The exceptions, which are narrowly construed, can be grouped in four categories. First, the Convention has certain jurisdiction prerequisites. For example, it does not apply to oral arbitration agreements, to domestic arbitrations, or to non-commercial arbitrations. New York Convention, arts. I, II. Second, the obligation to enforce arbitration agreements and awards does not extend to matters that, under national law, is nonarbitrable or violates a strong public policy. Id. at art. II(1); V(2). (This exception is discussed further below.) Third, the duty to enforce arbitral awards does extend to awards rendered without minimal due process protections (such as notice). Id. at art. V(2)(b). And fourth, the duty to enforce arbitral awards does not extend to ultra vires awards. Id. at V(2)(c).

27 For reasons that it would take too long here to explain, there would probably have to be some coordination of changes among many countries for national laws to effectively deal with the cyberspace challenge.

28 The FAA would have to be modified in several other minor respects not addressed here.

29 See, e.g., 9 U.S.C. ß 4 (authorizing federal court to order arbitration "within the district within the petition for an order directing such arbitration is filed); 9 U.S.C. 10 (providing for limited judicial review by a court "in and for the district wherein the award was made").

30 This task might prove to be difficult. Under the current international arbitration legal regime, national arbitration laws govern many issues of judicial assistance other than enforcement of the agreement and award. (These issues include, for example, certain aspects of discovery, the selection of arbitrators when the parties have not done so, and provisional relief.) National court jurisdiction over these issues is almost always determined by the fact that the arbitration takes places within the jurisdiction. Because the locus of VMDR is difficult, if not conceptually impossible, to identify, many jurisdictions might assert the power of judicial review and assistance. The answer to this problem is consensus coordination of the judicial assistance function. Such enforceable coordination could most effectively be accomplished by international treaty. Indeed, the New York Conventionís judicial enforcement provisions could be modified to cover judicial assistance. Another (less effective) possibility is to make national arbitration laws uniform, so that it doesnít matter which court provides judicial assistance. This is the basic strategy of the UNCITRAL model arbitration law. See generally Born, supra note 22, at 37-38.

31 A model for this approach is the European Convention on the Law Applicable to Contractual Obligations (the Rome Convention), 1980 Official Journal of the European Communities No. L266/1.

32 See generally Edwin Dickinson, THE LAW OF NATIONS AS PART OF THE NATIONAL LAW OF THE UNITED STATES, 101 PENN. L. REV. 26, 26-34 (1952)

33 CISG, United Nations Conventionon Contracts for the International Sale of Goods, Apr. 11, 1980, U.N. Doc. A/Conf. 97/18 (1980), reprinted in 19 I.L.M. 668 (1980).

34 This law might be statutory or common law. A model that is available in the contract context is the UNIDROIT Principles of International Commercial Contracts, see International Institution for the Unification of Private Law (UNIDROIT) (1994).

35 There are arguments both ways here.

36 Other examples include securities laws, criminal laws, laws against slavery contracts, etc.

37 One solution, of course, is to make it a condition of entry into cyberspace that such mandatory laws will apply. The problem is: Which mandatory laws? This problem is addressed at the end of the paper.

38 See New York Convention art. II(1), V(2).

39 The Supreme Court, for example, has ruled that antitrust, RICO, and securities claims are arbitrable. See Shearson/American Express, Inc. v. McMahon, 482 U.S. 220 (1987) (RICO claims arbitrable); Mitsubisho Motors Corp. v. Soler Chrysler-Plymouth, Inc., 473 U.S. 614 (1985) (Sherman Act claims arbitrable); Scherk v. Alberto-Culver Co., 417 U.S. 506 (1974) (1934 Securities Exchange Act claims arbitrable).

40 See generally Born, supra note 22, at 147-152.

41 Arbitrators increasingly apply non-contract law, but these legal issues all arise from contractual relations.

42 See generally Born, supra note 22, at 405-08.