1 OIL Casebook Topic I: Power 1 OIL Casebook Topic I: Power

This section introduces internet law. It examines different perspectives about the regulation of internet law, as well as approaches to cross-national and state regulation

1.1 Looking Back at the Law of the Horse: Why Cyberlaw and the Rule of Law are Important - Andrew Murray 1.1 Looking Back at the Law of the Horse: Why Cyberlaw and the Rule of Law are Important - Andrew Murray

© Andrew Murray 2013

Reproduced, Formatted and Excerpted under Creative Commons License: Attribution, Non-Commercial, No Derivative Works 2.5 UK

Original: script-ed.org/?p=1157

1. Opening

This is not the speech I planned to give – please accept my apologies. I had a great and rousing speech planned and I would write it between the 1st of February when I delivered the second edition of Information Technology Law to OUP and today. The problem was that the text delivery was delayed, isn’t it always, and I eventually delivered it on the 3rd of March. This left not enough time to do everything I planned so instead today’s speech is more of an outline or plan of where I see Cyberlaw research going rather than the original concept of delivering tangible results in this direction. For this I apologise and beg your indulgence.

The role of an opening keynote speaker is an important one. They set the tone for the conference by in essence “getting their view in first”. It is an honour therefore to be given this role and it shows trust on the part of the conference organiser. I must therefore begin by thanking the conference organisers for trusting me with this role. In thirty minutes or so Joseph will know whether his decision has been vindicated, but the reward for me is greater as I will have completed my contribution to the conference programme and will be able to enjoy the remainder of the conference without wondering whether my paper will go down well. So if you enjoy the next thirty minutes or so, come up afterwards and let me know. If you didn’t send me a note after the conference so that for two days at least I can imagine everyone liked what they heard.

I am following in a line of illustrious keynote speakers – in 2010 there was Josef Azizi, a Judge of the Court of the European Union, in 2011 Paul Maharg formerly BILETA executive chair and Iain Mitchell QC, UK representative on the IT Committee of the Bars and Law Societies of Europe. In 2012, the amazing double header of Professor Richard Susskind and Professor Chris Reed.

In truth to follow this line up seems impossible and I would have refused Joseph’s invitation were it not for two things: (1) he asked so very nicely and (2) my memory of my very first BILETA conference in York in 1999, when one of the keynote speakers was Geoff Hoon, future Defence Secretary, Transport Secretary, Leader of the House of Commons and Labour Party Chief Whip. At the time, he was Minister of State in the Lord Chancellor’s Office responsible for eCommerce. He gave an opening address which almost sent the audience to sleep at 10am on day one of the conference. So when Joseph asked, I thought even if it is the worst speech BILETA has ever heard there is a future in politics for me.

Anyway returning to my central theme – my own insecurities. I realised that while I may be better than Geoff Hoon, few in the room would remember that fateful day in York in 1999. Many would remember that last year in Newcastle, Chris Reed not only gave a talk he also entertained by playing the Ukulele. How does one follow a keynote speaker who not only plays a musical instrument but also makes them and flies gliders in his spare time?

I realised that today the renaissance law professor must do more than just teach and research. I also realised I had little skill in the usual fields. I have no artistic ability, am unmusical to the point where cats hide from my singing (really they do) and my dancing is so embarrassing that I simply don’t dance ever. Realising I wouldn’t get very far on Britain’s Got X-Talent Factor Voice, I turned to my wife of ten years to ask her what talents I had!.

She thought and said – you can tell stories. My first thought was – after ten years of marriage is that really the best thing she can say about me talent-wise? My second was great if Jackanory Idol ever gets made (I have it all planned out in my head the Judges would be Bernard Cribbins, Prunella Scales and the Ghost of Kenneth Williams) I’ll be in with a shot of winning. Then I slept on it and thought about it some more and I realised she had paid me a compliment. Academia is all about communication and in particular the telling of complex tales though simple tropes. The better the story teller, the more equipped one is for academia. Thus today I’m going to do what I do best, I’m going to tell you a story, a story of how Cyberlaw started off on the wrong foot and has (arguably) been on the wrong foot ever since and how this conference could be a place to start thinking about our future as Cyberlawyers before we go the same way as railroad lawyers.

2. The Chicago conference

The problem with my thesis is that the best storyteller sets the agenda and unfortunately there is some anecdotal evidence of this. The law evolves constantly but usually within defined substantive boundaries: criminal law, public law, torts, property and intellectual property, commercial law etc.

Occasionally an academic discipline will emerge which follows the vocational model, the most obvious being media law, telecommunications law and financial services law, but usually such vocational-driven academic models die out in a similar manner to railroad law, aviation law and space law.

The emergence of information technology and the Internet led to the slow evolution of Cyberlaw. Probably the first book on Computer Law (in the UK at least) was Colin Tapper’s 1978 edition, it paved the way for others to follow like Chris Reed (1990); Ian Lloyd (1993) and David Bainbridge (1990). In each case though the common theme was a book which collected what one may call “computer world problems” which required the application of traditional legal norms – computer contracts, computer evidence, data protection, copyright etc. The argument could easily be made that there was no such thing as computer law (or Cyberlaw as it would become) as it was not a substantive legal subject, nor, unlike media law, telecommunications law or financial services law was it vocational.

It looked (and looks) like those subjects which no longer exist (or exist only in niches) railroad law, aviation law and space law: an area of legal study determined by technology rather than norms or industry.

This point was made forcefully by Professor Frank Easterbrook at the 1996 University of Chicago conference “Law of Cyberspace”. Using powerful rhetoric he established himself clearly as the best storyteller in the room. Telling a vivid tale entitled “Cyberspace and The Law of the Horse” Easterbrook pretty much damned the nascent study of Cyberlaw with colourful flourishes:

When asked to talk about “Property in Cyberspace,” my immediate reaction was, “Isn’t this just the law of the horse?” I don’t know much about cyberspace; what I do know will be outdated in five years (if not five months!); and my predictions about the direction of change are worthless, making any effort to tailor the law to the subject futile. And if I did know something about computer networks, all I could do in discussing “Property in Cyberspace” would be to isolate the subject from the rest of the law of intellectual property, making the assessment weaker.

To make the point clear to anyone who missed the subtlety – Professor Easterbrook set out his argument in full (please excuse the rather long quote):

When he was dean of this law school, Gerhard Casper was proud that the University of Chicago did not offer a course in The Law of the Horse. He did not mean by this that Illinois specializes in grain rather than livestock. His point, rather, was that “Law and . . .courses should be limited to subjects that could illuminate the entire law. Instead of offering courses suited to dilettantes”. The University of Chicago offered courses in Law and Economics, and Law and Literature, taught by people who could be appointed to the world’s top economics and literature departments—even win the Nobel Prize in economics, as Ronald Coase has done. I regret to report that no one at this Symposium is going to win a Nobel Prize any time soon for advances in computer science. We are at risk of multidisciplinary dilettantism, or, as one of my mentors called it, the cross sterilization of ideas. Put together two fields about which you know little and get the worst of both worlds. Well, let me be modest. I am at risk of dilettantism, and I suspect that I am not alone. Beliefs lawyers hold about computers, and predictions they make about new technology, are highly likely to be false. This should make us hesitate to prescribe legal adaptations for cyberspace. The blind are not good trailblazers.

Dean Casper’s remark had a second meaning–that the best way to learn the law applicable to specialised endeavours is to study general rules. Lots of cases deal with sales of horses; others deal with people kicked by horses; still more deal with the licensing and racing of horses, or with the care veterinarians give to horses, or with prizes at horse shows. Any effort to collect these strands into a course on The Law of the Horse is doomed to be shallow and to miss unifying principles. Teaching 100 percent of the cases on people kicked by horses will not convey the law of torts very well. Far better for most students – better, even, for those who plan to go into the horse trade – to take courses in property, torts, commercial transactions, and the like, adding to the diet of horse cases a smattering of transactions in cucumbers, cats, coal, and cribs. Only by putting the law of the horse in the context of broader rules about commercial endeavours could one really understand the law about horses.

We can sit here and say this no longer applies – it is over sixteen years since Professor Easterbrook set out his manifesto. We can say that although we may not yet have a Nobel Prize for advances in computer science, we do have the Queen Elizabeth Prize for Engineering which was recently won by Tim Berners-Lee, Robert Kahn, Vint Cerf, Louis Pouzin and Marc Andreessen. To do so though would miss the point of Easterbrook’s speech. The power was the rhetoric not the detail of the message. It was his storytelling, not his story that was important.

3. The shadow

Professor Easterbrook continues to cast a long shadow over us and our work. We continue to be, consciously or unconsciously affected by what he said that day. It has split us (that is mainstream Cyberlawyers as opposed to ICT legal educationalists) into two genera and within each genera into several species.

The key is in the genera. One genus is the regulatory cyberlawyer (it is to this genus that I myself belong); the other is the (for want of a better word) applied cyberlawyer. There are no doubt many of you in the room. You research in defined areas of intersection between law and cyberspace – cyber-defamation, cyber-privacy, cyber-indecency or perhaps eCommerce. Both genera operate within the long Easterbrookian shadow.

It is perhaps most obvious in the regulatory cyberlaw (or cyber governance) field. The spiritual leader of those of us in this field is of course Professor Lawrence Lessig who answered Professor Easterbrook’s challenge in his “New Chicago School” model, found (among other places) in his Harvard Law Review paper The Law of the Horse: What Cyberlaw Might Teach. Lessig here introduces us to his now famous modalities of regulation thesis which posits that because cyberspace is distinct from realspace in the way it is designed, and in particular the malleability of the “code” – the design of the environment we learn as lawyers by studying cyberspace problems and therefore cyberspace law. Of course this massively simplifies Lessig’s position but strictures of time requires that I do so here. As I have written elsewhere, it is my belief that Lessig failed to rebut the key indictments in Easterbrook’s challenge to the Cyberlaw community, instead he simply pled “special circumstances”. By demonstrating that the Cyber-regulatory community can give something back to the general legal debate Lessig bought some time for Cyberlawyers but I fear that time is almost up.

Others have attempted to rebut Easterbrook. I myself put forward my own answer to him in The Regulation of Cyberspace but I have come to believe that the outcome of the collected labours of Cyber-regulatory or Cybergovernance theorists including myself, Jon Bing, Lee Bygrave, Roger Brownsword, Tim Wu, Jack Goldsmith, Ian Brown, Chris Marsden, Han Somsen, Paul DeHert and too many others to name merely give strength to the Easterbrook argument.

Again and again we return to the same well. We discuss the unique characteristics of digitisation and cyberspace and argue our case for cyber-regulation or governance theory, we employ academic heavyweights – Michael Foucault, Bruno Latour, Niklas Luhmann – and a number of legal academic cruiserweights – Gunther Teubner, Cass Sunstein, Neil MacCormick – to make our point that cyberspace and cyber-regulation is special. The problem is we continue to use the language and rhetoric of social policy, sociology and political philosophy. We refer to the literature of communications theorists like Manuel Castells and Nicholas Negroponte, we apply Latourian, Foucaultian or Frankfurt School language to the study of our little part of the social world. We become social scientists not lawyers.

We make Easterbrook’s argument for him – we are no longer cyberlawyers we are now cyber political scientists (or cyber political theorists). This was brought home to me recently when a colleague said to me quite baldly – what you do isn’t law; you could be in almost any social science department and do the same research. I argued he was wrong, but he wasn’t. The work I do is no different to that done by Robin Mansell (Media and Communications); Ian Brown (Information Systems) or John Naughton (Systems Engineering). The entire movement of Cyber-governance as populated by lawyers like myself, Jonathan Zittrain, Jack Goldsmith or Lee Bygrave demonstrates the need for cyberlawyers to justify themselves with reference to a wider debate; vitally a debate not founded on jurisprudence but upon other (non-legal) philosophical foundations. In so doing we differentiate ourselves from other lawyers.

A property law text is likely to cite Bentham, Birks, Dworkin, Hart, Hegel, Honore, Hohfeld, Hobbs, Kelsen, MacCormick, Posner, Raz and Waldron. A criminal law text is likely to cite Bentham, Blackstone, Denning, Garland, Hart, Honore, Horder, Kant, Locke and Williams. A cyberlaw text is significantly less likely to engage in this traditional jurisprudential debate. In the past we have told ourselves this is good. While our colleagues sit in the dark, narrow confines of legal orthodoxy we are in the light, taking an expansive, some may say cosmopolitan approach, but the truth is while we may say this inwardly, from the outside property lawyers and tax lawyers scratch their heads.

So is the answer to be found in that other genus, the applied cyberlawyer? This is more difficult for me to answer as I’m less familiar with this branch of the family. As someone who has dabbled in this area though I feel the answer is no. Mainstream privacy lawyers do not disrespect cyberprivacy lawyers but at all times remember (not consciously) Easterbrook’s charge (as amended) that “all I could do in discussing ‘Privacy in Cyberspace’ would be to isolate the subject from the rest of the law of privacy, making the assessment weaker”.

The same is unfortunately true of all other applied areas. Mainstream privacy lawyers, property lawyers, criminal lawyers and commercial lawyers see the application of their discipline within the “cyber” realm as a case-study for their subject not a separate or viable freestanding topic of study. The cyberlawyer who studies online defamation is a valuable colleague who takes time to learn the technical parameters of ISPs, SNPs and blog hosts. It saves them the time of learning the peculiarities of this place but they are not seen to illuminate the wider sphere of libel and defamation law. Cases like Tamiz v Google are interesting but do not get to the heart of their subject.

In other words this genus of the Cyberlaw family fulfils Easterbrook’s claim of multidisciplinary dilettantism, the cross sterilisation of ideas where one puts together two fields to get the worst of both worlds.

However one approaches cyberlaw in the traditional sense therefore one finds oneself in Easterbrook’s shadow. You are either a multidisciplinary dilettante or an apologist in the Lessig vein: a political or social scientist masquerading as a lawyer. This is hard to hear but it is how many colleagues outwith the field of cyberlaw or IT Law view our work. Either too narrow to contribute meaningfully to the subject or too wide and therefore not grounded in jurisprudence.

4. The rule of law

I do not believe cyberlaw is a cul-de-sac subject. I don’t believe we are destined to be as short lived a discipline as railroad law. As cyberspace becomes ever more central to our everyday lives the daily issues of indecency, harm, taxation, privacy, and security (both financial and personal) are discussed and legislated for.

In the media in the last week of March we find a discussion of the Spamhaus DDOS attack, the ACLU challenge to the deployment of IMSI capture under the 4th Amendment, copyright and online piracy, the sexualisation of children through online content, BitCoin online currency, the delivery of government services and the digital divide and the privacy challenge of Google glass.

Although cyberlaw issues may not yet displace Kim Kardashian and Justin Bieber in the affections of tabloid newspaper editors, public discourse in the subjects we have been discussing for twenty years is at an all time high.

Regulators are also aware of the issues and have been quietly building a legal framework to manage these issues. To give but one example, let’s take the measures introduced by the UK government to attempt to stem the tide of illegal file sharing and other forms of online copyright infringement (the issue is not whether these provisions will be effective just to examine why they were created).

The key support of the government’s strategy is the Digital Economy Act (at least it was prior to the Newzbin litigation). How did this come about? More pertinently what was the input of the Cyberlawyer – surely the expert in this area? Well as we know the catalyst for the Digital Economy Act was the Digital Britain report. This was produced by a committee chaired by Lord Carter of Barnes, Minister for Communications, Technology and Broadcasting, who although he has an LLB spent his entire professional career in advertising and marketing before becoming Chief Executive of OFCOM. He was assisted by a team of twenty five civil servants and a steering board of industry experts. They were Peter Black (network engineer); Tanya Byron (clinical psychologist); Francesco Caio (banker/telecommunications executive); Andrew Chitty (digital media production); Barry Cox (Journalist/Broadcaster); Matthew d’Ancona (Journalist); Robin Foster (Economist); Andrew Gowers (Journalist/Banker); Ian McCulloch (Broadcaster); Peter Phillips (Ofcom) and Stephen Temple (electrical engineer). Note that while there were three journalists and two bankers on the steering panel there were no cyberlawyers.

Regulators do not see cyberlawyers as central to the debate on the regulation of activity in cyberspace. This can be seen in the activity of the Joint Committee on Privacy and Injunctions where one of the driving forces behind the formation of the committee was the events of spring 2011 – what one may call the Twitter spring – in which privacy injunctions were breached on an almost daily basis by Twitter users (as well as on Facebook, Wikipedia and other social platforms). This committee was assisted by three specialist advisers: Professor Eric Barendt, Emeritus Professor of Media Law at University College London; Sir Charles Gray, former High Court judge who specialised in media work; and Paul Potts CBE, Visiting Professor of Journalism at Sheffield University and former chief executive of the Press Association.

Again mainstream “old” media is represented throughout but cyberlawyers are conspicuously absent. In the evidence-gathering phase twenty seven days were given over to “old media” interests and three days to “new media”. We are in danger of being marginalised by both academic colleagues and regulators unless we make ourselves relevant

How does one reset this? It is time for the rousing music and the beating of battle drums. Many of the markers for the future of Cyberlaw are already in place: we just need the confidence to follow them. The key is to re-engage with traditional jurisprudential models and thus to make ourselves relevant to lawmakers and lawyers in the way media lawyers have done.

Chris Reed began this with the work he did in the lead up to his book Making Laws for Cyberspace and in his keynote address here last year. Reed reintroduces legal positivism and perhaps rather counter intuitively given that, brings Lon Fuller and more intuitively Joseph Raz into the Cyberlaw analysis. I think to be honest in introducing elements of Raz and Fuller in a single text he offered an exciting opportunity but in his final analysis he was unfortunately rather timid.

His Modern Law Review article How to Make Bad Law: Lessons from Cyberspace was in many elements a more successful synthesis of the Fuller application at least. Here Reed attempts to deal with that most difficult question of Cyberlaw, or indeed any law question, what makes a law or lawmaker legitimate? This, as we all know is a particularly pertinent problem in cyberspace (see David Post) but of course is also a perennial jurisprudential question.

Reed focuses on Fuller’s Morality of Law and notes that “Fuller has asserted that a minimum internal morality is a necessary prerequisite for a purported law-system to be effective in ‘the enterprise of subjecting human conduct to the governance of rules.’” He points out that in Cyberlaw much applicable law (or rules) breaches three of Fuller’s eight basic principles of internal morality.

Principle 4: The law’s rules should be understandable by those who have to comply with them

Principle 5: Rules should not be contradictory, and

Principle 7: Rules must not be changed too frequently to permit compliance

From this Reed draws general principles of good lawmaking, applicable in general, from the lessons and failures of Cyberspace concluding that “an alternative approach to lawmaking which addresses human behaviour and beliefs, rather than specifying compliance in precise and detailed terms, can produce a law which is not immediately implausible and which more closely meets the test for quality.”

For me, this conclusion is gratifying as it is not dissimilar to the foundations of my symbiotic regulation model. Reed of course does not fully agree with symbiotic regulation, and neither he should, for it is a regulatory not a legal model. And so in many ways Reed shows us the past, but for me with two vital weaknesses – (1) in both the article and in his book, to me, he underplays Raz and the rule of law analysis and (2) he falls into the Lessig apologist position – the study of Cyberlaw is important because it illuminates the law elsewhere.

5. Cyberlaw and the rule of the law

To justify and the subject of Cyberlaw as a standalone legal subject we must stop apologising and start engaging jurisprudentially. To begin, I think the basic question of legitimacy and the rule of law in cyberspace offers a unique opportunity to kick start this discussion and to prove our worth to regulators. Regulators are regulating on the assumption that they have legitimacy and the right to do so. I saw this first hand at the Joint Parliamentary Committee on Privacy and Injunctions when Ian Brown and myself caused a senior Parliamentarian to turn puce with rage at the suggestion that “Her Majesty’s Government for the United Kingdom of Great Britain and Northern Ireland were not vested with the authority to regulate activities in the wider Internet of UK citizens”. The Rule of Law question – which should be primary and prior is being assumed based on historical sovereignty.

I think an analysis of the actions of regulatory actors online applying both Raz and Fuller is long overdue- this is the project I hoped to report on here but instead deliver an outline in the hope of inspiring others. Raz identified several principles that may be associated with the Rule of Law in some (but not all) societies.

Raz’s principles encompass the requirements of guiding the individual’s behaviour and minimising the danger that results from the exercise of discretionary power in an arbitrary fashion. In this last respect he shares common ground with the constitutional theorists A. V. Dicey, Friedrich Hayek and E. P. Thompson. Raz’s principles are:

  • Laws should be prospective rather than retroactive.
  • Laws should be stable and not changed too frequently, as lack of awareness of the law prevents one from being guided by it.
  • There should be clear rules and procedures for making laws.
  • The independence of the judiciary has to be guaranteed.
  • The principles of natural justice should be observed, particularly those concerning the right to a fair hearing.
  • The courts should have the power of judicial review over the way in which the other principles are implemented.
  • The courts should be accessible; no man may be denied justice.
  • The discretion of law enforcement and crime prevention agencies should not be allowed to pervert the law.
  • According to Raz, the validity of these principles depends upon the particular circumstances of different societies, whereas the rule of law, generally “is not to be confused with democracy, justice, equality (before the law or otherwise), human rights of any kind or respect for persons or for the dignity of man”.

    If the cyberlibertarians are right though there is no “law” for Cyberspace thus no “rule of law”. To now turn Raz on his head we can look again at Lon Fuller and his eight routes of failure for any legal system:

  • The lack of rules or law, which leads to ad-hoc and inconsistent adjudication.
  • Failure to publicise or make known the rules of law.
  • Unclear or obscure legislation that is impossible to understand.
  • Retroactive legislation.
  • Contradictions in the law.
  • Demands that are beyond the power of the subjects and the ruled.
  • Unstable legislation (ex. daily revisions of laws).
  • Divergence between adjudication/administration and legislation.
  • It appears that there may commonly exist failures 1,5,6,7 in Cyberspace and maybe also failure 2 as given multiple jurisdictions one cannot know all relevant rules of law.

    Thus both Raz and Fuller suggest problems with the Rule of Law and principles of natural justice – yet all too often regulators assume historical jurisdiction gives them extensive, expansive and legitimate authority to regulate activity online (without determining the spillover effects of their actions) – as Ian Brown and I saw first hand.

    Questions should be asked afresh of regulators of all types: Interstate actors such as ISOC, ICANN, WSIS and WIPO; Supranational bodies such as the EU, EEA or MERCOSUR, to non-state actors such as Google, Microsoft or Apple and most pertinently and importantly to States Regulators such as the UK Parliament, OFCOM or to the Courts who act from historical authority and legitimacy – are your actions legitimate, proportionate, and in accordance with the Rule of Law? In particular can you be sure your acts only affect individuals within your jurisdictional authority?

    The problem it seems to me is that we have mostly thought of Cyberspace as a single place – a monolithic place which can be defined rather than a world populated by divergent communities and cultures and representative of (mostly) a domestic space. Regulators imagine it as a (more) fragmented environment – if you will a UK Cyberspace, a French Cyberspace, a Greek Cyberspace etc. each within their jurisdictional remit and each within their sovereign right to rule. This is the root of laws such as the Digital Economy Act or the HADOPI Law. It is this which leads to the Defamation Bill and the Data Protection Regulation. The problem with this fragmentation approach is overlapping controls and over regulation arguably in breach of Raz’s and Fuller’s principles.

    Against the fragmented background – it is easy to imagine legal rules (and systems) which comply with both Fuller and Raz but we must not assume all laws passed by traditionally sovereign lawmakers automatically make the grade.

    There can be, and indeed is, a functioning positivistic legal system for Cyberspace. Our role now is to be part of the debate, and not just as activists. By discussing the rule of law and its role online we can reposition ourselves as lawyers in the Cyberlaw environment who have a role in the lawmaking process – that is the domestic lawmaking process and how it affects us online.

    The question of whether cyberspace is regulable is in the past for lawmakers and regulators. The question is how to most efficiently and effectively regulate it by law. Our role must be to ensure the rule of law is preserved. We must move to the fourth wave of Cyberlaw research and development. The first wave was cyberlibertarianism. The second wave was cyberpaternalism. The third wave was decentred regulation and regulatory theory – it is time to move to the fourth wave – Cyberlaw.

    1.2 Law And Borders: The Rise of Law in Cyberspace David R. Johnson and David G. Post, 48 Stanford Law Review 1367 (1996) (excerpted) 1.2 Law And Borders: The Rise of Law in Cyberspace David R. Johnson and David G. Post, 48 Stanford Law Review 1367 (1996) (excerpted)

     

    Law And Borders: The Rise of Law in Cyberspace
    David R. Johnson and David G. Post, 48 Stanford Law Review 1367  (1996)

     

    Introduction

    Global computer-based communications cut across territorial borders, creating a new realm of human activity and undermining the feasibility--and legitimacy--of applying laws based on geographic boundaries. While these electronic communications play havoc with geographic boundaries, a new boundary, made up of the screens and passwords that separate the virtual world from the "real world" of atoms, emerges. This new boundary defines a distinct Cyberspace that needs and can create new law and legal institutions of its own. Territorially-based law-making and law-enforcing authorities find this new environment deeply threatening. But established territorial authorities may yet learn to defer to the self-regulatory efforts of Cyberspace participants who care most deeply about this new digital trade in ideas, information, and services. Separated from doctrine tied to territorial jurisdictions, new rules will emerge, in a variety of online spaces, to govern a wide range of new phenomena that have no clear parallel in the nonvirtual world. These new rules will play the role of law by defining legal personhood and property, resolving disputes, and crystallizing a collective conversation about core values.

    I. Breaking Down Territorial Borders

     ...

    B. The Absence of Territorial Borders in Cyberspace

    Cyberspace radically undermines the relationship between legally significant (online) phenomena and physical location. The rise of the global computer network is destroying the link between geographical location and: (1) the power of local governments to assert control over online behavior; (2) the effects of online behavior on individuals or things; (3) the legitimacy of the efforts of a local sovereign to enforce rules applicable to global phenomena; and (4) the ability of physical location to give notice of which sets of rules apply.

    The Net thus radically subverts a system of rule-making based on borders between physical spaces, at least with respect to the claim that cyberspace should naturally be governed by territorially defined rules.

    Cyberspace has no territorially-based boundaries, because the cost and speed of message transmission on the Net is almost entirely independent of physical location: Messages can be transmitted from any physical location to any other location without degradation, decay, or substantial delay, and without any physical cues or barriers that might otherwise keep certain geographically remote places and people separate from one another. ... The Net enables transactions between people who do not know, and in many cases cannot know, the physical location of the other party. Location remains vitally important, but only location within a virtual space consisting of the "addresses" of the machines between which messages and information are routed.

    The system is indifferent to the physical location of those machines, and there is no necessary connection between an Internet address and a physical jurisdiction.

    Although a domain name, when initially assigned to a given machine, may be associated with a particular Internet Protocol address corresponding to the territory within which the machine is physically located (e.g., a ".uk" domain name extension), the machine may move in physical space without any movement in the logical domain name space of the Net. Or, alternatively, the owner of the domain name might request that the name become associated with an entirely different machine, in a different physical location. ... Thus, a server with a ".uk" domain name may not necessarily be located in the United Kingdom, a server with a ".com" domain name may be anywhere, and users, generally speaking, are not even aware of the location of the server that stores the content that they read. Physical borders no longer can function as signposts informing individuals of the obligations assumed by entering into a new, legally significant, place, because individuals are unaware of the existence of those borders as they move through virtual space.

    The power to control activity in Cyberspace has only the most tenuous connections to physical location. Many governments first respond to electronic communications crossing their territorial borders by trying to stop or regulate that flow of information as it crosses their borders. Rather than deferring to efforts by participants in online transactions to regulate their own affairs, many governments establish trade barriers, seek to tax any border-crossing cargo, and respond especially sympathetically to claims that information coming into the jurisdiction might prove harmful to local residents. Efforts to stem the flow increase as online information becomes more important to local citizens. In particular, resistance to "transborder data flow" (TDF) reflects the concerns of sovereign nations that the development and use of TDF's will undermine their "informational sovereignty," will negatively impact on the privacy of local citizens, and will upset private property interests in information. Even local governments in the United States have expressed concern about their loss of control over information and transactions flowing across their borders.

    But efforts to control the flow of electronic information across physical borders--to map local regulation and physical boundaries onto Cyberspace--are likely to prove futile, at least in countries that hope to participate in global commerce. Individual electrons can easily, and without any realistic prospect of detection, "enter" any sovereign's territory. The volume of electronic communications crossing territorial boundaries is just too great in relation to the resources available to government authorities to permit meaningful control.

    ...

    Faced with their inability to control the flow of electrons across physical borders, some authorities strive to inject their boundaries into the new electronic medium through filtering mechanisms and the establishment of electronic barriers. Others have been quick to assert the right to regulate all online trade insofar as it might adversely impact local citizens. The Attorney General of Minnesota, for example, has asserted the right to regulate gambling that occurs on a foreign web page that was accessed and "brought into" the state by a local resident. The New Jersey securities regulatory agency has similarly asserted the right to shut down any offending Web page accessible from within the state.

    But such protective schemes will likely fail as well.

    First, the determined seeker of prohibited communications can simply reconfigure his connection so as to appear to reside in a different location, outside the particular locality, state, or country. Because the Net is engineered to work on the basis of "logical," not geographical, locations, any attempt to defeat the independence of messages from physical locations would be as futile as an effort to tie an atom and a bit together. And, moreover, assertions of law-making authority over Net activities on the ground that those activities constitute "entry into" the physical jurisdiction can just as easily be made by any territorially-based authority.

    If Minnesota law applies to gambling operations conducted on the World Wide Web because such operations foreseeably affect Minnesota residents, so, too, must the law of any physical jurisdiction from which those operations can be accessed. By asserting a right to regulate whatever its citizens may access on the Net, these local authorities are laying the predicate for an argument that Singapore or Iraq or any other sovereign can regulate the activities of U.S. companies operating in cyberspace from a location physically within the United States.

    All such Web-based activity, in this view, must be subject simultaneously to the laws of all territorial sovereigns.

    Nor are the effects of online activities tied to geographically proximate locations. Information available on the World Wide Web is available simultaneously to anyone with a connection to the global network. The notion that the effects of an activity taking place on that Web site radiate from a physical location over a geographic map in concentric circles of decreasing intensity, however sensible that may be in the nonvirtual world, is incoherent when applied to Cyberspace. A Web site physically located in Brazil, to continue with that example, has no more of an effect on individuals in Brazil than does a Web site physically located in Belgium or Belize that is accessible in Brazil. Usenet discussion groups, to take another example, consist of continuously changing collections of messages that are routed from one network to another, with no centralized location at all; they exist, in effect, everywhere, nowhere in particular, and only on the Net.

    Nor can the legitimacy of any rules governing online activities be naturally traced to a geographically situated polity. There is no geographically localized set of constituents with a stronger claim to regulate it than any other local group; the strongest claim to control comes from the participants themselves, and they could be anywhere.

    The rise of an electronic medium that disregards geographical boundaries also throws the law into disarray by creating entirely new phenomena that need to become the subject of clear legal rules but that cannot be governed, satisfactorily, by any current territorially-based sovereign. For example, electronic communications create vast new quantities of transactional records and pose serious questions regarding the nature and adequacy of privacy protections. Yet the communications that create these records may pass through or even simultaneously exist in many different territorial jurisdictions. What substantive law should we apply to protect this new, vulnerable body of transactional data? May a French policeman lawfully access the records of communications traveling across the Net from the United States to Japan? Similarly, whether it is permissible for a commercial entity to publish a record of all of any given individual's postings to Usenet newsgroups, or whether it is permissible to implement an interactive Web page application that inspects a user's "bookmarks" to determine which other pages that user has visited, are questions not readily addressed by existing legal regimes--both because the phenomena are novel and because any given local territorial sovereign cannot readily control the relevant, globally dispersed, actors and actions.

    Because events on the Net occur everywhere but nowhere in particular, are engaged in by online personae who are both "real" (possessing reputations, able to perform services, and deploy intellectual assets) and "intangible" (not necessarily or traceably tied to any particular person in the physical sense), and concern "things" (messages, databases, standing relationships) that are not necessarily separated from one another by any physical boundaries, no physical jurisdiction has a more compelling claim than any other to subject these events exclusively to its laws.

    ...

    II. A New Boundary for Cyberspace

    ...

    A. Cyberspace as a Place

    Many of the jurisdictional and substantive quandaries raised by border-crossing electronic communications could be resolved by one simple principle: conceiving of Cyberspace as a distinct "place" for purposes of legal analysis by recognizing a legally significant border between Cyberspace and the "real world."

    Using this new approach, we would no longer ask the unanswerable question "where" in the geographical world a Net-based transaction occurred. Instead, the more salient questions become: What rules are best suited to the often unique characteristics of this new place and the expectations of those who are engaged in various activities there? What mechanisms exist or need to be developed to determine the content of those rules and the mechanisms by which they can enforced?

    Answers to these questions will permit the development of rules better suited to the new phenomena in question, more likely to be made by those who understand and participate in those phenomena, and more likely to be enforced by means that the new global communications media make available and effective.

    1. The New Boundary is Real.

    Treating Cyberspace as a separate "space" to which distinct laws apply should come naturally, because entry into this world of stored online communications occurs through a screen and (usually) a "password" boundary. There is a "placeness" to Cyberspace because the messages accessed there are persistent and accessible to many people. You know when you are "there." No one accidentally strays across the border into Cyberspace. To be sure, Cyberspace is not a homogenous place; groups and activities found at various online locations possess their own unique characteristics and distinctions, and each area will likely develop its own set of distinct rules. But the line that separates online transactions from our dealings in the real world is just as distinct as the physical boundaries between our territorial governments--perhaps more so.

    Crossing into Cyberspace is a meaningful act that would make application of a distinct "law of Cyberspace" fair to those who pass over the electronic boundary.

    ...

    IV. Local Authorities, Foreign Rules: Reconciling Conflicts

    What should happen when conflicts arise between the local territorial law (applicable to persons or entities by virtue of their location in a particular area of physical space) and the law applicable to particular activities on the Net? The doctrine of "comity," as well as principles applied when delegating authority to self-regulatory organizations, provide us with guidance for reconciling such disputes.

    The doctrine of comity, in the Supreme Court's classic formulation, is "the recognition which one nation allows within its territory to the legislative, executive, or judicial acts of another nation, having due regard both to international duty and convenience, and to the rights of its own citizens or of other persons who are under the protections of its law."

    It is incorporated into the principles set forth in the Restatement (Third) of Foreign Relations Law of the United States, in particular Section 403, which provides that "a state may not exercise jurisdiction to prescribe law with respect to a person or activity having connections with another state when the exercise of such jurisdiction is unreasonable," and that when a conflict between the laws of two states arises, "each state has an obligation to evaluate its own as well as the other state's interest in exercising jurisdiction [and] should defer to the other state if that state's interest is clearly greater.").

    It arose as an attempt to mitigate some of the harsher features of a world in which lawmaking is an attribute of control over physical space but in which persons, things, and actions may move across physical boundaries, and it functions as a constraint on the strict application of territorial principles that attempts to reconcile "the principle of absolute territorial sovereignty [with] the fact that intercourse between nations often demand[s] the recognition of one sovereign's lawmaking acts in the forum of another." In general, comity reflects the view that those who care more deeply about and better understand the disputed activity should determine the outcome. Accordingly, it may be ideally suited to handle, by extension, the new conflicts between the a-territorial nature of cyberspace activities and the legitimate needs of territorial sovereigns and of those whose interests they protect on the other side of the cyberspace border. This doctrine does not disable territorial sovereigns from protecting the interests of those individuals located within their spheres of control, but it calls upon them to exercise a significant degree of restraint when doing so.

    Local officials handling conflicts can also learn from the many examples of delegating authority to self-regulatory organizations. Churches are allowed to make religious law. Clubs and social organizations can, within broad limits, define rules that govern activities within their spheres of interest. Securities exchanges can establish commercial rules, so long as they protect the vital interests of the surrounding communities.

    In these cases, government has seen the wisdom of allocating rule-making functions to those who best understand a complex phenomenon and who have an interest in assuring the growth and health of their shared enterprise.

    Cyberspace represents a new permutation of the underlying issue: How much should local authorities defer to a new, self-regulating activity arising independently of local control and reaching beyond the limited physical boundaries of the sovereign. This mixing of both tangible and intangible boundaries leads to a convergence of the intellectual categories of comity in international relations and the local delegation by a sovereign to self-regulatory groups. In applying both the doctrine of "comity" and the idea of "delegation" to Cyberspace, a local sovereign is called upon to defer to the self-regulatory judgments of a population partly, but not wholly, composed of its own subjects.

    Despite the seeming contradiction of a sovereign deferring to the authority of those who are not its own subjects, such a policy makes sense, especially in light of the underlying purposes of both doctrines. Comity and delegation represent the wise conservation of governmental resources and allocate decisions to those who most fully understand the special needs and characteristics of a particular "sphere" of being. Although Cyberspace represents a new sphere that cuts across national boundaries, the fundamental principle remains.

    If the sysops and users who collectively inhabit and control a particular area of the Net want to establish special rules to govern conduct there, and if that rule set does not fundamentally impinge upon the vital interests of others who never visit this new space, then the law of sovereigns in the physical world should defer to this new form of self-government.

    ...

    Because controlling the flow of electrons across physical boundaries is so difficult, a local jurisdiction that seeks to prevent its citizens from accessing specific materials must either outlaw all access to the Net--thereby cutting itself off from the new global trade--or seek to impose its will on the Net as a whole. This would be the modern equivalent of a local lord in medieval times either trying to prevent the silk trade from passing through his boundaries (to the dismay of local customers and merchants) or purporting to assert jurisdiction over the known world. It may be most difficult to envision local territorial sovereigns deferring to the law of the Net when the perceived threat to local interests arises from the very free flow of information that is the Net's most fundamental characteristic--when, for example, local sovereigns assert an interest in seeing that their citizens are not adversely affected by information that the local jurisdiction deems harmful but that is freely (and lawfully) available elsewhere.

    Examples include the German government's attempts to prevent its citizens access to prohibited materials, or the prosecution of a California bulletin board operator for making material offensive to local "community standards" available for downloading in Tennessee.

    Local sovereigns may insist that their interest (in protecting their citizens from harm) is paramount, and easily outweighs any purported interest in making this kind of material freely available. But the opposing interest is not simply the interest in seeing that individuals have access to ostensibly obscene material, it is the "meta-interest" of Net citizens in preserving the global free flow of information.

    If there is one central principle on which all local authorities within the Net should agree, it must be that territorially local claims to restrict online transactions (in ways unrelated to vital and localized interests of a territorial government) should be resisted. This is the Net equivalent of the First Amendment, a principle already recognized in the form of the international human rights doctrine protecting the right to communicate.

    Participants in the new online trade must oppose external regulation designed to obstruct this flow. This naturally central principle of online law bears importantly on the "comity" analysis, because it makes clear that the need to preserve a free flow of information across the Net is just as vital to the interests of the Net as the need to protect local citizens against the impacts of unwelcome information may appear from the perspective of a local territorial sovereign.

    For the Net to realize its full promise, online rule-making authorities must not respect the claims of territorial sovereigns to restrict online communications when unrelated to vital and localized governmental interests.

    V. Internal Diversity

    ...

    The ability of inhabitants of Cyberspace to cross borders at will between legally significant territories, many times in a single day, is unsettling. This power seems to undercut the validity of developing distinct laws for online culture and commerce: How can these rules be "law" if participants can literally turn them on and off with a switch? Frequent online travel might subject relatively mobile human beings to a far larger number of rule sets than they would encounter traveling through the physical world over the same period. Established authorities, contemplating the rise of a new law applicable to online activities, might object that we cannot easily live in a world with too many different sources and types of law, particularly those made by private (non-governmental) parties, without breeding confusion and allowing anti-social actors to escape effective regulation.

    But the speed with which we can cross legally meaningful borders or adopt and then shed legally significant roles should not reduce our willingness to recognize multiple rule sets. Rapid travel between spheres of being does not detract from the distinctiveness of the boundaries, as long as participants realize the rules are changing. Nor does it detract from the appropriateness of rules applying within any given place, any more than changing commercial or organizational roles in the physical world detracts from a person's ability to obey and distinguish rules as a member of many different institutional affiliations and to know which rules are appropriate for which roles Nor does it lower the enforceability of any given rule set within its appropriate boundaries, as long as groups can control unauthorized boundary crossing of groups or messages.

    Alternating between different legal identities many times during a day may confuse those for whom cyberspace remains an alien territory, but for those for whom cyberspace is a more natural habitat in which they spend increasing amounts of time it may become second nature. Legal systems must learn to accommodate a more mobile kind of legal person.

    V1. Conclusion

    Global electronic communications have created new spaces in which distinct rule sets will evolve. We can reconcile the new law created in this space with current territorially-based legal systems by treating it as a distinct doctrine, applicable to a clearly demarcated sphere, created primarily by legitimate, self-regulatory processes, and entitled to appropriate deference--but also subject to limitations when it oversteps its appropriate sphere.

    The law of any given place must take into account the special characteristics of the space it regulates and the types of persons, places, and things found there. Just as a country's jurisprudence reflects its unique historical experience and culture, the law of Cyberspace will reflect its special character, which differs markedly from anything found in the physical world. For example, the law of the Net must deal with persons who "exist" in Cyberspace only in the form of an email address and whose purported identity may or may not accurately correspond to physical characteristics in the real world. In fact, an e-mail address might not even belong to a single person. Accordingly, if Cyberspace law is to recognize the nature of its "subjects," it cannot rest on the same doctrines that give geographically based sovereigns jurisdiction over "whole," locatable, physical persons. The law of the Net must be prepared to deal with persons who manifest themselves only by means of a particular ID, user account, or domain name.

    Moreover, if rights and duties attach to an account itself, rather than an underlying real world person, traditional concepts such as "equality," "discrimination," or even "rights and duties" may not work as we normally understand them. New angles on these ideas may develop. For example, when AOL users joined the Net in large numbers, other Cyberspace users often ridiculed them based on the ".aol" tag on their email addresses--a form of "domainism" that might be discouraged by new forms of Netiquette. If a doctrine of Cyberspace law accords rights to users, we will need to decide whether those rights adhere only to particular types of online appearances, as distinct from attaching to particular individuals in the real world.

    Similarly, the types of "properties" that can become the subject of legal discussion in Cyberspace will differ from real world real estate or tangible objects. For example, in the real world the physical covers of a book delineate the boundaries of a "work" for purposes of copyright law; those limits may disappear entirely when the same materials are part of a large electronic database. Thus, we may have to change the "fair use" doctrine in copyright law that previously depended on calculating what portion of the physical work was copied. Similarly, a web page's "location" in Cyberspace may take on a value unrelated to the physical place where the disk holding that Web page resides, and efforts to regulate web pages by attempting to control physical objects may only cause the relevant bits to move from one place to another. On the other hand, the boundaries set by "URLs" (Uniform Resource Locators, the location of a document on the World Wide Web) may need special protection against confiscation or confusingly similar addresses. And, because these online "places" may contain offensive material, we may need rules requiring (or allowing) groups to post certain signs or markings at these places' outer borders.

    The boundaries that separate persons and things behave differently in the virtual world but are nonetheless legally significant. Messages posted under one e-mail name will not affect the reputation of another e-mail address, even if the same physical person authors both messages. Materials separated by a password will be accessible to different sets of users, even if those materials physically exist on the very same hard drive. A user's claim to a right to a particular online identity or to redress when that identity's reputation suffers harm, may be valid even if that identity does not correspond exactly to that of any single person in the real world.

    Clear boundaries make law possible, encouraging rapid differentiation between rule sets and defining the subjects of legal discussion. New abilities to travel or exchange information rapidly across old borders may change the legal frame of reference and require fundamental changes in legal institutions. Fundamental activities of lawmaking--accommodating conflicting claims, defining property rights, establishing rules to guide conduct, enforcing those rules, and resolving disputes--remain very much alive within the newly defined, intangible territory of Cyberspace. At the same time, the newly emerging law challenges the core idea of a current law-making authority--the territorial nation state, with substantial but legally restrained powers.

    If the rules of Cyberspace thus emerge from consensually based rule sets, and the subjects of such laws remain free to move among many differing online spaces, then considering the actions of Cyberspace's system administrators as the exercise of a power akin to "sovereignty" may be inappropriate. Under a legal framework where the top level imposes physical order on those below it and depends for its continued effectiveness on the inability of its citizens to fight back or leave the territory, the legal and political doctrines we have evolved over the centuries are essential to constrain such power. In that situation, where exit is impossible, costly, or painful, then a right to a voice for the people is essential. But when the "persons" in question are not whole people, when their "property" is intangible and portable, and when all concerned may readily escape a jurisdiction they do not find empowering, the relationship between the "citizen" and the "state" changes radically. Law, defined as a thoughtful group conversation about core values, will persist. But it will not, could not, and should not be the same law as that applicable to physical, geographically-defined territories.

     (c)1996 David R. Johnson and David G. Post. Permission granted to redistribute freely, in whole or in part, with this notice attached.

    1.3 Against Cyberanarchy- 65 U. Chi. L. Rev. 1199 (1998) - Jack L. Goldsmith 1.3 Against Cyberanarchy- 65 U. Chi. L. Rev. 1199 (1998) - Jack L. Goldsmith

    This article responds to the Johnson & Post article along with other cyberexceptionalist articles? Who has the better argument? Note that Post wrote a followup called “Against ‘Against Cyberanarchy.'”

    University of Chicago Law Review 
    Fall 1998

    AGAINST CYBERANARCHY

    Jack L. Goldsmith [FNd1]

    Copyright © 1998 University of Chicago; Jack L. Goldsmith

         The Supreme Court's partial invalidation of the Communications Decency Act on First Amendment grounds [FN1] raises the more fundamental question of whether the state can regulate cyberspace at all. [FN2] Several commentators, whom I shall call "regulation skeptics," have argued that it cannot. [FN3] Some courts have also expressed skepticism. [FN4] The popular and technical press are full of similar claims. [FN5]

      *1200 The regulation skeptics make both descriptive and normative claims. On the descriptive side, they claim that the application of geographically based conceptions of legal regulation and choice of law to a-geographical cyberspace activity either makes no sense or leads to hopeless confusion. On the normative side, they argue that because cyberspace transactions occur "simultaneously and equally" in all national jurisdictions, regulation of the flow of this information by any particular national jurisdiction illegitimately produces significant negative spillover effects in other jurisdictions. They also claim that the architecture of cyberspace precludes notice of governing law that is crucial to the law's legitimacy. In contrast, they argue, cyberspace participants are much better positioned than national regulators to design comprehensive legal rules that would both internalize the costs of cyberspace activity and give proper notice to cyberspace participants. The regulation skeptics conclude from these arguments that national regulators should "defer to the self-regulatory efforts of Cyberspace participants." [FN6]

      This Article challenges the skeptics' arguments and their conclusion. The skeptics make three basic errors. First, they overstate the differences between cyberspace transactions and other transnational transactions. Both involve people in real space in one territorial jurisdiction transacting with people in real space in another territorial jurisdiction in a way that sometimes causes real-world harms. In both contexts, the state in which the harms are suffered has a legitimate interest in regulating the activity that produces the harms. Second, the skeptics do not attend to the distinction between default laws and mandatory laws. Their ultimate normative claim that cyberspace should be self- regulated makes sense with respect to default laws that, by definition, private parties can modify to fit their needs. It makes much less sense with respect to mandatory or regulatory laws that, for paternalistic reasons or in order to protect third parties, place limits on private legal ordering. Third, the skeptics underestimate the potential of traditional legal tools and technology to resolve the multijurisdictional regulatory problems implicated *1201 by cyberspace. Cyberspace transactions do not inherently warrant any more deference by national regulators, and are not significantly less resistant to the tools of conflict of laws, than other transnational transactions.

      Some caveats are in order up front. This Article argues only that regulation of cyberspace is feasible and legitimate from the perspective of jurisdiction and choice of law. It does not argue that cyberspace regulation is a good idea, and it does not take a position on the merits of particular regulations beyond their jurisdictional legitimacy. For example, it does not examine whether particular national regulations of the Internet promote democracy, or are efficient, or are good or bad for humanity. Similarly, the Article does not consider substantive limitations on cyberspace regulation such as may be found in the Bill of Rights or international human rights law. Resolution of these substantive regulatory issues turns in part on contested normative judgments and difficult context-specific, cost-benefit analyses that are far beyond this Article's scope. But resolution of these issues also turns on how we understand the jurisdictional confusions that arise when national regulation, which has traditionally been understood primarily in geographical terms, applies to a phenomenon that appears to resist geographical orientation. This jurisdictional puzzle is the focus of this Article.

      In addition, the Article does not deny that the new communication technologies known as cyberspace will lead to changes in governmental regulation. Such changes are to be expected when the speed of communication dramatically increases and the cost of communication dramatically decreases. The invention of the telegraph, the telephone, the radio, the television, and the satellite, among many other communications advances, all possessed these characteristics. And they all gave rise to societal and regulatory changes. [FN7] So too will cyberspace. But the skeptics claim much more than that cyberspace necessitates changes in governmental regulation. They claim that cyberspace is so different from other communication media that it will, or should, resist all governmental regulation. My aim here is to show why this claim is flawed, and to explain in general terms how traditional tools of jurisdiction and choice of law apply to cyberspace transactions.

    ****

    I. The Regulation Skeptics' Claims

      People transacting in cyberspace do things that would be regulated by state, national, or international law if they occurred in person or by telephone or mail. They defame, invade privacy, harass, and commit business torts. [FN8] They make and breach contracts. [FN9] They distribute pornography and swap bombmaking tips. [FN10] They infringe trademarks, violate copyrights, and steal data. [FN11] They issue fraudulent securities and restrict competition. [FN12] And so on.

      Are these and other cyberspace activities governed by the same laws that govern similar transnational activities mediated in person, or by phone, or by mail? If so, which jurisdiction's law governs? If not, what governs instead?

      The regulation skeptics' analysis of these questions makes two sets of assumptions. The first concerns the nature of legal regulation of non- cyberspace events. [FN13] The skeptics tend to conceptualize a nation's legal authority as extending to its territorial borders and not beyond. This conception makes them skeptical about the legitimacy of one nation regulating activities that take place in another. And it leads them to believe that transnational *1203 disputes must be resolved by choice-of-law rules that select a unique governing law on the basis of where an event occurs or where transacting parties are located. On this view, tort liability is governed by the law of the place where the tort occurred and the validity of a contract is governed by the law of the place where the contract was made. Such choice-of- law rules are thought to promote rule-of-law values like uniformity (that is, every forum will apply the same law in a given case), predictability, and certainty. And they are supposed to give the parties to transnational transactions reasonable notice of governing law.

      The skeptics' second set of assumptions concerns the architecture of cyberspace. They view cyberspace as a unique "boundary-destroying" means of communication. Internet protocol addresses do not necessarily correlate with a physical location. As a result, the skeptics assert, persons transacting in cyberspace often do not, and cannot, know each other's physical location. [FN14] In addition, information mediated by certain cyberspace services appears "simultaneously and equally in all jursidictions" around the world. [FN15] A web page in Illinois can be accessed from and thus appear in any geographical jurisdiction that is plugged in to the World Wide Web. When I participate in an online discussion group, my messages can appear simultaneously in every geographical jurisdiction where persons participate in the group. In neither case can I control, or even know about, the geographical flow of the information that I upload or transmit.

      It is against this background that the skeptics make their descriptive and normative claims. Descriptively, they claim that cyberspace is a borderless medium that resists regulation conceived in geographical terms. [FN16] One reason is that information transmitted via cyberspace can easily flow across national borders without detection. [FN17] Another reason is that it is senseless to apply geographically configured choice-of-law rules to a- geographical cyberspace activities. [FN18] A third reason is that regulation of the local effects of cyberspace information flows permits all nations simultaneously to regulate all web-based transactions. [FN19] The result is multiple and inconsistent regulation of the same activity. *1204 A final reason is that the architecture of cyberspace enables its users to route around or otherwise evade territorial regulation. [FN20]

      The skeptics' normative arguments build on these assumptions. Their essential normative claim is that it is illegitimate for any particular nation to regulate the local effects of multijurisdictional cyberspace activity. This is so for three reasons. First, such regulation will often apply to acts abroad, and will thus be impermissibly extraterritorial. [FN21] Second, because cyberspace information flows appear in every jurisdiction simultaneously, unilateral regulation of these flows will illegitimately affect the regulatory efforts of other nations and the cyberspace activities of parties in other jurisdictions. [FN22] Third is the problem of notice. The skeptics argue that because a person transacting in cyberspace does not know when or whether her activity produces effects in a particular jurisdiction, she lacks notice about governing law and therefore cannot conform her behavior to it. [FN23] They claim that under these conditions, it is unfair to apply law to her cyberspace activities. The skeptics believe that all three of these problems can be avoided by cyberspace self-regulation.

      To make these claims more concrete, consider the predicament of one of the scores of companies that offer, sell, and deliver products on the World Wide Web. Assume that the web page of a fictional Seattle-based company, Digitalbook.com, offers digital books for sale and delivery over the Web. One book it offers for sale is Lady Chatterley's Lover. This offer extends to, and can be accepted by, computer users in every country with access to the Web. Assume that in Singapore the sale and distribution of pornography is criminal, and that Singapore deems Lady Chatterley's Lover to be pornographic. Assume further that Digitalbook.com's terms of sale contain a term that violates English consumer protection laws, and that the publication of Digitalbook.com's Lady Chatterley's Lover in England would infringe upon the rights of the novel's English copyright owner. Digitalbook.com sells and sends copies of Lady Chatterley's Lover to two people whose addresses (say, anonymous@aol.com and anonymous@msn.com) do not reveal their physical location but who, *1205 unbeknownst to Digitalbook.com, live and receive the book in Singapore and London, respectively.

      The skeptics claim that it is difficult for courts in Singapore or England to regulate disputes involving these transactions in accordance with geographical choice-of-law rules. In addition, they argue that English and Singaporean regulations will expose Digitalbook.com to potentially inconsistent obligations. Finally, the skeptics claim that Digitalbook.com can easily evade the Singaporean and English regulations by sending unstoppable digital information into these countries from a locale beyond their enforcement jurisdiction.

      On the normative side, the skeptics are concerned that the application of English and Singaporean law to regulate Digitalbook.com's transactions constitutes an impermissible extraterritorial regulation of a U.S. corporation. Because Digitalbook.com might bow to the English and Singaporean regulations, and because the company cannot limit its cyberspace information flows by geography, the English and Singaporean regulations might cause it to withdraw Lady Chatterley's Lover everywhere or to raise its price. The English and Singaporean regulations would thus affect Digitalbook.com's behavior in the United States and adversely affect the purchasing opportunities of parties in other countries. The skeptics believe these negative spillover effects of the national regulations are illegitimate. They also think it is unfair for England and Singapore to apply their laws in this situation because Digitalbook.com had no way of knowing that it sold and delivered a book to consumers in these countries.

    II. "Real-space" Jurisdictional Conflict Management

      The skeptics are in the grip of a nineteenth century territorialist conception of how "real space" is regulated and how "real-space" conflicts of law are resolved. [FN24] This conception was repudiated in the middle of this century. [FN25] The skeptics' first mistake, therefore, is to measure the feasibility and legitimacy of national regulation of cyberspace against a repudiated yardstick. This Section *1206 offers a more accurate picture of real-space jurisdictional conflict management as a prelude to analysis of the skeptics' claims.

      Three factors led to the overthrow of the traditional approach to choice of law. [FN26] The first was significant changes in the world. Changes in transportation, communication, and in the scope of corporate activity led to an unprecedented increase in multijurisdictional activity. These changes put pressure on the rigid territorialist conception, which purported to identify a single legitimate governing law for transborder activity based on discrete territorial contacts. So too did the rise of the regulatory state, which led to more caustic public policy differences among jurisdictions, and which pressured the interested forum to apply local regulations whenever possible. [FN27]

      A second factor, legal realism, contributed to the demise of hermetic territorialism. All conflict-of-laws problems by definition have connections to two or more territorial jurisdictions. The legal realists showed that nothing in the logic of territorialism justified legal regulation by any one of these territories rather than another. [FN28] They also argued that a forum's decision to apply foreign law was always determined by local domestic policies. [FN29] This established the theoretical foundation for the lex fori orientation that has dominated choice of law ever since.

      A third factor, legal positivism, exacerbated the problem of finding a unique governing law in transactional cases. Courts avoided many choice-of-law problems in such cases by applying universal customary laws tied to no particular sovereign authority, such as the law merchant, the law maritime, and the law of *1207 nations. [FN30] But positivism's insistence on a sovereign source for every rule of decision undermined judicial reliance on these laws. [FN31] It also contributed to the waning of universal choice-of- law rules that courts applied in circumstances in which transnational customary laws did not govern. In the United States, for example, the general uniformity of choice-of-law approaches that characterized the nineteenth century gave way in the twentieth century to a plethora of choice-of-law regimes. [FN32] As different jurisdictions adopted different choice-of-law regimes, the goal of a single governing law for transjurisdictional transactions was further frustrated. [FN33]

      These factors did not completely undermine traditional views about territorial regulation. But they did lead to an expansion of the permissible bases for territorial jurisdiction. Today, the Constitution permits a state to apply its law if it has a "significant contact or significant aggregation of contacts, creating state interests, such that choice of its law is neither arbitrary nor fundamentally unfair." [FN34] In practice, this standard is notoriously easy to satisfy. [FN35] It prohibits the application of local law only when the forum state has no interest in the case because the substance of the lawsuit has no relationship to the state. Customary international law limits on a nation's regulation of extraterritorial events are less clear because there are few international decisions on point, and because state practice does not reveal a settled *1208 custom. Nonetheless, it seems clear that customary international law, like the United States Constitution, permits a nation to apply its law to extraterritorial behavior with substantial local effects. [FN36] In addition, both the Constitution and international law permit a nation or state to regulate the extraterritorial conduct of a citizen or domiciliary. [FN37] In short, in modern times a transaction can legitimately be regulated by the jurisdiction where the transaction occurs, the jurisdictions where significant effects of the transaction are felt, and the jurisdictions where the parties burdened by the regulation are from.

      This expansion of the permissible bases for the application of local law has revolutionized conflict of laws in the second half of this century. Any number of choice-of-law regimes are now consistent with constitutional and international law. The earlier belief in a unique governing law for all transnational activities has given way to the view that more than one jurisdiction can legitimately apply its law to the sametransnational activity. [FN38] The uniformity promised by the traditional approach has thus been replaced by the reality of overlapping jurisdictional authority. This means that the application of one jurisdiction's law often comes at the expense of the nonapplication of the conflicting laws of other interested jurisdictions. Because choice-of-law rules often differ from jurisdiction to jurisdiction, and because a forum applies its own choice-of-law rules, the choice of forum is now often critical to the selection of governing law. In this milieu, ex ante notice of a specific governing law is no longer a realistic goal in many transnational situations. Not surprisingly, the Constitution and international law impose very weak notice requirements on the application of local law to extraterritorial activity.

      *1209 This modern world of jurisdictional conflict poses obvious difficulties for participants in transnational transactions. To understand these problems and their resolution, it is important to distinguish between default laws and mandatory laws. For present purposes, a default law can be understood as one that presumptively governs a particular relationship or transaction, but that can be modified or circumvented by the parties in the relationship or transaction. The default laws of different countries can create a conflict of laws. For example, the estate of a U.S. national who dies intestate in England, his domicile, could potentially be subject to the succession rules of either country. Similarly, a contract made in one country for delivery of products in another could be subject to the remedies regime of either country.

      Parties in such transnational relationships can alleviate choice-of-law uncertainty with respect to default rules by contracting for specific terms, by selecting a governing law, or both. [FN39] Most contractual choice-of-law clauses govern the contracts within which they are embedded. But the scope of this private legal control is not limited to traditional contractual issues. In many circumstances, parties can agree to a governing law for torts and related actions that arise from their contractual relations. [FN40] They can also specify the governing law for matters ranging from intellectual property to trusts and estates to internal corporate affairs. [FN41]

      The possibilities for private legal ordering are not limitless. Every nation has mandatory laws that govern particular transactions or relationships regardless of the wishes of the parties. The primary justifications for such laws are paternalism and protection of third parties. [FN42] Mandatory laws range from limits on contractual *1210 capacity to criminal law to securities and antitrust law. Like default laws, they differ in content and scope from jurisdiction to jurisdiction. Unlike conflicts of default laws, conflicts of mandatory laws cannot be resolved easily by private contract. [FN43] They can, in theory, be resolved by public contract--international agreements that embrace uniform international rules [FN44] or uniform choice-of-law rules. [FN45] Such solutions are increasingly prominent but still relatively rare. Moreover, these attempts at international uniformity are often limited to default rules, and are littered with mandatory law exceptions. [FN46]

      This discussion shows that conflicts of law can arise when parties to a transnational transaction do not specify the governing default law, or when the transaction implicates a mandatory law that conflicts with the otherwise- applicable law. Absent a governing international law, transnational activity in these contexts will usually be governed by the law of a single jurisdiction. [FN47] And absent international choice-of-law rules, the forum's choice-of- law rules will determine the governing law. In regulatory contexts, the forum will invariably apply local law. [FN48] But regardless of which substantive law the forum applies, the application of that law will frequently create spillover effects on activities in other countries and on the ability of other interested nations to apply their own law. In our increasingly integrated world, these spillover effects are likely to extend to many countries. [FN49]

      *1211 Consider, for example, the Supreme Court's decision in Hartford Fire Insurance Co v California. [FN50] The Court held that the concerted refusal by London reinsurers to sell certain types of reinsurance to insurers in the United States violated the Sherman Act. The reinsurers' acts in England were legal under English law. But the Court determined that the reinsurers were nonetheless subject to U.S. regulation because their actions "produced substantial effect(s)" in the United States. [FN51] U.S. law thus regulated the activities of English companies in England at the expense of the nonapplication of English law. Similarly, had an English court applied English law to adjudge the reinsurers' acts to be legal, it would have produced spillover effects on consumers in the United States, and would have come at the expense of the nonapplication of U.S. law. No matter which law governed the reinsurers' acts, the application of that law would have produced spillover effects on the English reinsurers' activities in other jurisdictions, and on the activities of persons in other jurisdictions adversely affected by the reinsurers' acts.

      A similar phenomenon occurs in many domestic and international conflicts contexts. For example, the European Commission recently imposed strict conditions on a merger (already approved by the Federal Trade Commission) between two American companies with no manufacturing facilities in Europe. [FN52] Minnesota applied its pro-plaintiff stacking rules for automobile insurance coverage to an accident in Wisconsin among Wisconsin residents. [FN53] A United States federal grand jury ordered the local branch of a foreign bank, a nonparty, to disclose bank records in the Bahamas in possible violation of Bahamian law. [FN54] California applied its workmen's compensation law to benefit an employee of a California corporation who suffered a tort while working in Alaska--even though Alaska purported to make its worker's compensation scheme exclusive, and even though the employment contract specified that Alaska law governed. [FN55] New York *1212 applied its tort law to a car accident in Canada. [FN56] California taxed a British corporation based on the California portion of its world profits. [FN57]

      In these situations and countless others, one jurisdiction regulates extraterritorial conduct in a way that invariably affects individual behavior and regulatory efforts in other jurisdictions. These spillover effects constitute the central problem of modern conflict of laws. The problem is pervasive. It is also inevitable, because the price of eliminating these spillovers--abolishing national or subnational lawmaking entities, or eliminating transnational activity--is prohibitively high. Most of the dizzying array of modern choice-of-law methodologies are devoted to minimizing these spillovers while at the same time preserving the sovereign prerogative to regulate effects within national borders. [FN58] International harmonization efforts seek to achieve similar aims, often at the expense of national prerogatives. [FN59]

      There is widespread debate about which approach, or combination of approaches, is preferable. Resolution of this debate is less important for present purposes than two uncontested assumptions that underlie it. The first assumption is that in the absence of consensual international solutions, prevailing concepts of territorial sovereignty permit a nation to regulate the local effects of extraterritorial conduct even if this regulation produces spillover effects in other jurisdictions. The second assumption is that such spillover effects are a commonplace consequence of the unilateral application of any particular law to transnational activity in our increasingly interconnected world. It is against this background that the skeptics' descriptive and normative claims must be assessed.

    III. Is Cyberspace Regulation Feasible?

      This Section argues that the skeptics' claims about the infeasibility of national regulation of cyberspace rest on an underappreciation *1213 of the realities of modern conflict of laws, and of the legal and technological tools available to resolve multijurisdictional cyberspace conflicts. From the perspective of jurisdiction and choice of law, regulation of cyberspace transactions is no less feasible than regulation of other transnational transactions.

    A. Default Laws and Private Ordering in Cyberspace

      Cyberspace transactions that implicate default laws, like other transnational transactions that implicate such laws, are subject to private legal ordering. The architecture of cyberspace facilitates this private ordering and thus enables cyberspace participants to avoid many transnational conflicts of law.

      At the most basic level, private ordering is facilitated by the technical standards that define and limit cyberspace. [FN60] To participate in the Internet function known as the World Wide Web, users must consent to the TCP/IP standards that define the Internet as well as to the HTML standards that more particularly define the Web. Similarly, sending e-mail over the Internet requires the sender to use TCP/IP standards and particular e-mail protocols. One's experience of cyberspace is further defined and limited by the more particular communication standards embedded in software. [FN61] For example, within the range of what TCP/IP and HTML permit, an individual's communication via the World Wide Web will be shaped and limited by (among many other things) her choice of browsers and search engines. These and countless other technical standard choices order behavior in cyberspace. In this sense, access to different cyberspace networks and communities is always conditioned on the accessors' consent to the array of technical standards that define these networks and communities.

      Technical standards cannot comprehensively specify acceptable behavior in cyberspace. Within the range of what these standards permit, information flows might violate network norms or territorial laws. Many network norms are promulgated and enforced *1214 informally. A more formal method to establish private legal orders in cyberspace is to condition access to particular networks on consent to a particular legal regime.

      This regime could take several forms. It could be a local, national, or international law. When you buy a Dell computer through the company's web page from anywhere in the world, you agree that "(a)ny claim relating to, and the use of, this Site and the materials contained herein is governed by the laws of the state of Texas." [FN62] Alternatively, the chosen law could be a free- standing model law attached to no particular sovereign but available to be incorporated by contract. For example, parties to a commercial transaction over the Internet could agree that their transaction is governed by UNIDROIT Principles or the Uniform Customs and Practice for Documentary Credits. [FN63] Or the governing law could be the contractual terms themselves. [FN64] Waivers and exclusions operate as private law in this way. So too do chat rooms, discussion lists, and local area networks that condition participation on the user's consent to community norms specified in a contract.

      Cyberspace architecture can also help to establish other aspects of a private legal order. Through conditioned access, cyberspace users can consent to have subsequent disputes resolved by courts, arbitrators, systems operators, or even "virtual magistrates." [FN65] They can also establish private enforcement regimes. Technical standards operate as an enforcer of sorts by defining and limiting cyberspace activity. For example, software filters can block or condition access to certain information, and various technologies perform compliance monitoring functions. [FN66] In addition, the gatekeeper of each cyberspace community can cut off entry for noncompliance with the community rules, or punish a user for bad acts by drawing on a bond (perhaps simply a credit card) put up as a condition on the user's entry. [FN67]

      *1215 Many have proposed a structure for private legal ordering of cyberspace along the lines just sketched. [FN68] There is nothing remarkable about this structure. It differs little from the legal structure of other private groups, such as churches, merchants, families, clubs, and corporations, which have analogous consent-based governing laws, dispute resolution mechanisms, and private enforcement regimes. [FN69] But just as private ordering is often not a comprehensive solution to the regulation of "real- space" private groups, it will not be a comprehensive solution to the regulation of cyberspace either.

      In part this is because it remains an open question how to generate consent across cyberspace networks. Conditioning access on consent to a governing legal regime is relatively easy at the entry point of a cyberspace network. In theory, it is just as easy to generate such consent at the interface between networks. It is commonplace to click on a hypertext link and be greeted by a message that conditions further access on presentation of an identification code, or credit card number, or personal information such as age and address. A similar demand for consent to a particular legal regime could be added as a condition for access. However, this process might become confusing; the technological and conceptual details of consenting to and coordinating different legal regimes as one works one's way through dozens of cyberspace networks remain to be worked out. [FN70] In addition, the generation of legal consent across networks will impose time and other costs that are anathema to many cyberspace users.

      An important additional difficulty is that many cyberspace activities affect non-cyberspace participants with whom ex ante consent to a private legal regime will not be possible. Cyberspace is not, as the skeptics often assume, a self-enclosed regime. A communication in cyberspace often has consequences for persons outside the computer network in which the communication took place. For example: a book uploaded on the Net can violate an author's copyright; a chat room participant can defame someone outside the chat room; terrorists can promulgate bomb making or kidnapping tips; merchants can conspire to fix prices by e-mail; a corporation can issue a fraudulent security; a pornographer can *1216 sell kiddie porn; Internet gambling can decrease in-state gambling revenues and cause family strife; and so on. In these and many other ways, communications via cyberspace produce harmful, real-world effects on those who have not consented to the private ordering of the cyberspace community.

      Finally, even if the hurdles to consent can be surmounted, consent-based legal orders are limited by a variety of national mandatory law restrictions. [FN71] These mandatory laws define who may consent to these private regimes. For example, they prevent persons of certain ages from entering into certain types of contracts. They also limit the form and scope of such consent. The consideration requirement and limitations on liquidated damages clauses fall into this category, as do requirements that the law chosen by the parties have a reasonable relationship to the subject matter of the contract. Some mandatory laws also limit the internal and external activities of the group's activities. Criminal law, for example, falls in this category.

      Private legal ordering thus has the potential to resolve many, but not all, of the challenges posed by multijurisdictional cyberspace activity. Cyberspace activities for which ex ante consent to a governing legal regime is either infeasible or unenforceable are not amenable to private ordering. Such activities remain subject to the skeptics' concerns about multiple or extraterritorial national regulation. [FN72]

    B. The Limits of Enforcement Jurisdiction

      The skeptics' concerns are further attenuated, however, by limitations on every nation's ability to enforce its laws. A nation can purport to regulate activity that takes place anywhere. The Island of Tobago can enact a law that purports to bind the rights of the whole world. [FN73] But the effective scope of this law depends on Tobago's ability to enforce it. And in general a nation can only enforce its laws against: (i) persons with a presence or assets in the nation's territory; (ii) persons over whom the nation can obtain personal jurisdiction and enforce a default judgment against abroad; or (iii) persons whom the nation can successfully extradite. [FN74]

      *1217 A defendant's physical presence or assets within the territory remains the primary basis for a nation or state to enforce its laws. The large majority of persons who transact in cyberspace have no presence or assets in the jurisdictions that wish to regulate their information flows in cyberspace. Such regulations are thus likely to apply primarily to Internet service providers and Internet users with a physical presence in the regulating jurisdiction. Cyberspace users in other territorial jurisdictions will indirectly feel the effect of the regulations to the extent that they are dependent on service or content providers with a presence in the regulating jurisdiction. [FN75] But for almost all users, there will be no threat of extraterritorial legal liability because of a lack of presence in the regulating jurisdictions.

      A nation or state can also enforce its laws over an entity with no local presence or assets if it can obtain personal jurisdiction over the entity and enforce a local default judgment against that entity abroad. The domestic interstate context presents a much greater threat in this regard than does the international context. This is because the Full Faith and Credit Clause requires a state to enforce the default judgment of a sister state that had personal jurisdiction over the defendant. [FN76] This threat is attenuated, however, by constitutional limits on a state's assertion of personal jurisdiction. The Due Process Clauses prohibit a state from asserting personal jurisdiction over an entity with no local presence unless the entity has purposefully directed its activities to the forum state and the assertion of jurisdiction is reasonable. [FN77]

      Application of this standard to cyberspace activities presents special difficulties. Under standard assumptions about cyberspace *1218 architecture, persons can upload or transmit information knowing that it could reach any and all jurisdictions, but not knowing which particular jurisdiction it might reach. Can every state where these transmissions appear assert specific personal jurisdiction over the agent of the information under the purposeful availment and reasonableness tests?

      Full consideration of this issue is far beyond this Article's scope.  [FN78] I simply wish to point out why there is relatively little reason at present, and even less reason in the near future, to believe that the mere introduction of information into cyberspace will by itself suffice for personal jurisdiction over the agent of the transmission in every state where the information appears. Most courts have required something more than mere placement of information on a web page in one state as a basis for personal jurisdiction in another state where the web page is accessed. [FN79] For a variety of reasons, these decisions have limited specific personal jurisdiction to cases in which there are independent indicia that the out-of-state defendant knowingly and purposefully directed the effects of out-of-state conduct to a particular state where the acts were deemed illegal.

      Given the skeptics' assumptions about cyberspace architecture, this conclusion appears appropriate. It seems unfair to expose a content provider to personal jurisdiction in all fifty states for the mere act of uploading information on a computer if she cannot take affordable precautions to avoid simultaneous multi-jurisdictional effects. But we shall see below that the skeptics' architectual assumptions are inaccurate. It is already possible for content providers to take measures to achieve significant control over information flows. And filtering and identification technology promise greater control at less cost. [FN80] In cyberspace as in real space, the ultimate meaning of "purposeful availment" and "reasonableness" will depend on the cost and feasibility of information flow control. [FN81] As such control becomes more feasible and less costly, personal jurisdiction over cyberspace activities will become *1219 functionally identical to personal jurisdiction over real- space activities.

      This detour into the technicalities of personal jurisdiction was necessitated by a worry about the extraterritorial enforcement of local default judgments against nonlocal cyberspace users within the American federal system. Such concerns are less pronounced in the international context. In contrast to the domestic interstate context, customary international law imposes few enforceable controls on a country's assertion of personal jurisdiction, and there are few treaties on the subject. [FN82] However, also in contrast to domestic law, there is no full faith and credit obligation to enforce foreign judgments in the international sphere. [FN83] If one country exercises personal jurisdiction on an exorbitant basis, the resulting judgment is unlikely to be enforced in another country. [FN84] In addition, local public policy exceptions to the enforcement of foreign judgments are relatively commonplace in the international sphere, especially when the foreign judgment flies in the face of the enforcing state's regulatory regime. [FN85] For these reasons, there is little concern that a foreign default judgment *1220 will be enforceable against cyberspace users who live outside the regulating jurisdiction.

      The final way that a nation can enforce its regulations against persons outside its jurisdiction is by seeking extradition. In the United States, extradition among the several states is regulated by Article IV of the Constitution and the federal extradition law. [FN86] As a general matter, State A must accede to the proper demand of State B for the surrender of a fugitive who committed an act in State B that State B considers a crime. Nonetheless, a person who in State A transmits information flows that appear in and constitute a crime in State B will not likely be subject to extradition to State B under these provisions. This is because the extradition obligation only extends to fugitives who have fled State B, and these terms have long been limited to persons who were physically present in the demanding state at the time of the crime's commission. [FN87] A different, but equally forceful, limitation applies to international extradition. International extradition is governed largely by treaty. [FN88] A pervasive feature of modern extradition treaties is the principle of double criminality. This principle requires that the charged offense be criminal in both the requesting and the requested jurisdictions. [FN89] This principle, and its animating rationale, make it unlikely that there will be international cooperation in the enforcement of exorbitant unilateral criminal regulations of cyberspace events.

      This review of transnational enforcement jurisdiction makes clear that the skeptics exaggerate the threat of multiple regulation of cyberspace information flows. This threat must be measured by a regulation's enforceable scope, not by its putative scope. And the enforceable scope is relatively narrow. It extends only to individual users or system operators with presence or assets in the enforcement jurisdiction, or (in the U.S.) to entities that take extra steps to target cyberspace information flows to states where such information flows are illegal. Such regulatory exposure is a significant concern for cyberspace participants. But it is precisely how regulatory exposure operates in "real space." And it is far *1221 less significant than the skeptics' hyperbolic claim that all users of the Web will be simultaneously subject to all national regulations. [FN90]

      Even with these limitations, the skeptics worry that an individual cyberspace content provider in one jurisdiction faces potential liability in another jurisdiction when she places information on the Internet. This potential liability can become an unforeseen reality when the provider travels to the regulating jurisdiction, or moves assets there. Such potential liability in turn affects the providers' activities at home and thus can be viewed as a weak form of extraterritorial regulation. This form of regulation is a theoretical possibility, but it should not be exaggerated. No nation has as yet imposed liability on a content provider for unforeseen effects in an unknown jurisdiction. The threat of such liability will lessen as content providers continue to gain means to control information flows. [FN91] It is also conceivable that weak normative limitations might exist or develop to prevent a jurisdiction from regulating local effects that were truly unforeseeable or uncontrollable. [FN92] The point for now is that even in the absence of such limits, this potential threat of liability is relatively insignificant and does not come close to the skeptics' broad descriptive claims about massive multiple regulation of individual users.

    C. Indirect Regulation of Extraterritorial Activity

      Indeed, if the limits on enforcement jurisdiction support any of the skeptics' descriptive claims, it is their somewhat different claim that because of the potential for regulation evasion, cyberspace transactions are beyond the regulatory powers of territorial governments. [FN93] Cyberspace content providers can, at some cost, shift the source of their information flows to jurisdictions beyond the enforceable scope of national regulation and thus continue information transmissions into the regulating jurisdiction. [FN94] For example, they can relocate in geographical space, or employ telnet or anonymous remailers to make the geographical source of their *1222 content difficult to discern. [FN95] These and related regulatory evasion techniques can make it difficult for a nation to regulate the extraterritorial supply side of harmful cyberspace activity.

      Regulation evasion of this sort is not limited to cyberspace. For example, corporations reincorporate to avoid mandatory laws and criminals launder money offshore. Closer to point, offshore regulation evasion has been a prominent characteristic of other communication media. For example, Radio-Free Europe broadcast from western Europe into the former Soviet Union but lacked a regulatable presence there. [FN96] Similarly, television signals are sometimes broadcast from abroad by an entity with no local presence. The extraterritorial source of these and many other non-cyberspace activities is beyond the enforceable scope of local regulation. But this does not mean that local regulation is inefficacious. In cyberspace as in real space, offshore regulation evasion does not prevent a nation from regulating the extraterritorial activity.

      This is so because a nation can regulate people and equipment in its territory to control the local effects of the extraterritorial activity. Such indirect regulation is how nations have, with varying degrees of success, regulated local harms caused by other communications media with offshore sources and no local presence. [FN97] And it is how nations have begun to regulate local harms caused by offshore Internet content providers. For example, nations penalize in-state end users who obtain and use illegal content or who otherwise participate in an illegal cyberspace transaction. [FN98] They also regulate the local means through which foreign *1223 content is transmitted. For example, they impose screening obligations on in-state Internet service providers and other entities that supply or transmit information. [FN99] Or they regulate in-state hardware and software through which such transmissions are received. [FN100] Or they regulate the local financial intermediaries that make commercial transactions on the Internet possible. [FN101]

      These and related regulations of domestic persons and property make it more costly, and thus more difficult, for in-state users to obtain content from, or transact with, regulation evaders abroad. In this fashion a nation can indirectly regulate the extraterritorial supply of prohibited content even though the source of the content is beyond its enforcement jurisdiction and even though it cannot easily stop transmission at the border. These various forms of indirect regulation will not be perfect in the sense of eliminating regulation evasion. But few regulations are perfect in this sense, and regulation need not be perfect in this sense to be effective. [FN102] The question is always whether the regulation will heighten the costs of the activity sufficiently to achieve *1224 its acceptable control from whatever normative perspective is appropriate.

      In the cyberspace regulation context, the answer to this question depends on empirical and technological issues that are unresolved and that will vary from context to context. The prodigious criticism of and lobbying efforts against proposed regulation of (among other things) digital goods, Internet gambling, and encryption technology suggest that governments can raise the costs of many cyberspace transactions to a significant degree. And of course unilateral national regulation is one of many regulation strategies at a nation's disposal. [FN103] The point for now is simply that offshore regulation evasion does not, as the skeptics think, undermine a nation's ability to regulate cyberspace transactions. Although a nation will sometimes have difficulty in imposing liability on extraterritorial content providers, it can still significantly regulate the local effects of these providers' activities through laws aimed at local persons and entities.

    D. Filtering

      We have seen that the skeptics' worries about multiple or extraterritorial regulation of cyberspace activity do not extend to matters for which it is feasible and legal for cyberspace communities to establish private legal regimes, or to matters beyond a nation's enforcement jurisdiction.

      But the possibility of extraterritorial and multiple regulations remains. Consider the Bavarian Justice Ministry's threat in December of 1995 to prosecute CompuServe for carrying online discussion groups containing material that violated German anti-pornography laws. [FN104] CompuServe responded by blocking access to these discussion groups in Germany. Because of the state of then-available technology, this action had the effect of blocking access to these discussion groups for all CompuServe users worldwide. [FN105] This is precisely what the skeptics fear from unilateral regulation of cyberspace. Germany enforced a mandatory law against an international access provider with a presence (office, staff, servers, etc.) in Germany. Faced with multiple regulatory regimes in the many places where it did business, CompuServe bowed to the most restrictive. The consequence was massive extraterritorial regulation, for the German regulation interrupted *1225 the flow and availability of the discussion groups for CompuServe clients everywhere in the world.

      The skeptics frequently recount this story to show how unilateral national regulation of cyberspace can have multijurisdictional consequences. [FN106] But the rest of the story suggests a somewhat different lesson. After closing down transmission of the offending discussions, CompuServe offered its German users software that enabled them to block access to the offending discussion groups. [FN107] The company then began to search for a more centralized way to filter the illegal newsgroups in Germany alone. German prosecutors subsequently indicted a CompuServe executive, alleging that the company failed to implement such national-level filtering technology to prevent dissemination of other illegal information in Germany. [FN108] At about the same time, the German parliament enacted a law clarifying that cyberspace access providers are liable "if they are aware of the content" and fail to use "technically possible and reasonable" means to block it. [FN109]

      The subsequent events of the CompuServe controversy, like the response to the Supreme Court's invalidation of the Communications Decency Act in Reno, [FN110] make clear the growing importance of information discrimination technology to the cyberspace regulation debate. Many jurisdictional challenges presented by cyberspace result from the purported inability of content providers to prevent information flows from appearing simultaneously in every jurisdiction. Thus far I have assumed, with the skeptics, that this is a necessary (and accurate) feature of cyberspace architecture. But it is not. [FN111] Cyberspace information can only appear *1226 in a geographical jurisdiction by virtue of hardware and software physically present in the jurisdiction. Available technology already permits governments and private entities to regulate the design and function of hardware and software to facilitate discrimination of cyberspace information flows along a variety of dimensions, including geography, network, and content. [FN112] This technology is relatively new and still relatively crude, but it is growing very quickly in both sophistication and effectiveness. This technology facilitates discrimination and control of information flows at any of several junctures along the cyberspace information stream.

      At the most basic level, the content provider can take steps to control the flow of the information. This happens, for example, whenever a web page operator conditions access to the page on the users' presentation of information. Consider the many precautions taken by adult web pages. Some pages simply warn minors or persons from certain geographical locations not to view or enter, and disclaim legal liability if they do. [FN113] Others condition access on proof of age or on membership in one of dozens of private *1227 age-verification services. [FN114] Others require potential end- users to send by fax or telephone information specifying age and geographical location. [FN115] Still others label or rate their pages in order to accommodate end-use filtering software, as described below. Finally, digital identification technology developed for Internet commerce provides a way to authenticate the identity of a party in a cyberspace transaction. [FN116] Although digital identification is usually used to verify who someone is, it can also be used to verify other facts about cyberspace users, such as their nationality, domicile, or permanent address.

      At the other end of the distribution chain, end-users can employ software filters to block out or discriminate among information flows. [FN117] Parental control software is the most prominent example of an end-user filter, but many businesses and other local area networks also employ these technologies. Content filters also can be imposed at junctures along the cyberspace information stream between content providers and end-users. They can be imposed, for example, at the network level or at the level of the Internet service provider. They can also assist governments in filtering information at the national level. [FN118] A government can choose to have no Internet links whatsoever and to regulate telephone *1228 and other communication lines to access providers in other countries. [FN119] China, Singapore, and the United Arab Emirates have taken the somewhat less severe steps of (i) regulating access to the Net through centralized filtered servers, and (ii) requiring filters for in-state Internet service providers and end-users. [FN120] We have seen that Germany has chosen to hold liable Internet access providers who have knowledge of illegal content and fail to use "technically possible and reasonable" means to filter it. [FN121] The Federal Communications Commission recently required V-chip blocking technology to be placed in computers capable of receiving video broadcasting, [FN122] and pending anti-spam legislation would impose identification requirements on commercial e-mail senders and filtering requirements on Internet service providers. [FN123] There are numerous other possibilities. [FN124]

      Although technological predictions are precarious, it seems likely that the techniques and technologies for controlling cyberspace information flows will continue to develop in scope and sophistication, and will play an important role in resolving the jurisdictional quandaries presented by the "borderless" medium. Information is not particularly useful unless people can organize, select, and block it. [FN125] This is one reason why information filtering is an essential component of all communications media. [FN126] Filtering is especially important for cyberspace, where the costs of information production and dissemination are extremely low, and thus information overload is a serious concern. Indeed, the explosive growth of the World Wide Web is directly attributable to the *1229 invention of identification and filtering technologies that made it possible to organize and select from the morass of available information. [FN127]

      An additional reason that techniques for controlling cyberspace information flows are likely to be at least moderately successful is that so many participants in the cyberspace regulation debate--parents, businesses, content suppliers, service providers, governments, and even some anticensorship civil libertarians [FN128]--desire such control. As Resnick has pointed out, "meta-data systems . . . are going to be an important part of the Web, because they enable more sophisticated commerce . . ., communication, indexing, and searching services." [FN129] Many jurisdictions have already mandated the use of filtering and identification mechanisms. [FN130] Even in the absence of government mandates, content filtering and digital identification technologies have flourished for commerical reasons and in response to the threat of regulation, and have become de facto standards in many cyberspace contexts.

      Many commentators are skeptical about these filtering and identification technologies. [FN131] They argue that content filters invariably both over- and under-filter; that identification technologies sometimes misidentify; and that some hackers will access prohibited information. These worries are to some degree well-founded. What is not well-founded, however, is the belief that imperfect regulation means ineffective regulation. [FN132] Real space is filled with similarly imperfect filtering and identification techniques: criminals crack safes and escape from jail, fifteen year olds visit bars with fake IDs, secret information is leaked to the press, and so on. In cyberspace as in real space, imperfections in filtering and identification regimes do not render the regimes ineffective. [FN133] Although the ultimate accuracy of cyberspace filtering *1230 and identification technologies remains an open question, there is little doubt that such technologies will contribute significantly to cyberspace regulation by enabling governments, content providers, end-users, and service providers to raise significantly the cost of accessing certain information. Indeed, this has already happened throughout cyberspace, where content filtering, conditioned access, and identification codes are pervasive.

      The ability to control information flows alleviates the many cyberspace regulation problems that are premised on the assumption that information in cyberspace appears simultaneously in every jurisdiction. To see why, consider one set of differences between a newspaper publisher and a cyberspace content provider. It is relatively uncontroversial that a newspaper publisher is liable for harms caused wherever the newspaper is published or distributed. This seems appropriate because, among other reasons, we think the publisher can control the geographical locus of publication and distribution. Requiring such control imposes modest costs on the publisher; she must, for example, keep abreast of regulatory developments in different jurisdictions and take steps to exclude publication and distribution in places where she wants to avoid liability.

      Now consider the cyberspace content provider. Many have an intuition that such content providers should not be liable for harms caused wherever the content appears. [FN134] The primary basis for this intuition is that the content provider cannot control the geographical and network distribution of his information flows. But this latter point is groundless. Content providers already have several means to control information flows. [FN135] As the cost of such control continues to drop, and the accuracy and ease of this control increases, cyberspace content providers will come to occupy the same position as the newspaper publisher. It will thus be appropriate in cyberspace, as in real space, for the law to impose small costs on both types of publisher to ensure that content does not appear in jurisdictions and networks where it is illegal.

    E. International Harmonization

      Private legal ordering, the limitations on enforcement jurisdiction, indirect regulation, and effective information flow control, taken together, go a long way toward redressing the skeptics' descriptive claims about the infeasibility of cyberspace regulation. These techniques will not resolve all conflict of laws in cyberspace *1231 any more than they do in real space. Nor will they definitively resolve the problem of the relative ease by which information suppliers can "relocate" into a safe haven outside of the regulating jurisdiction, a problem that also has many real-space analogies. [FN136] When similar spillover and evasion problems have occurred with respect to non- cyberspace transactions, nations have responded with a variety of international harmonization strategies.

      The same harmonization strategies are being used today to address the challenges presented by cyberspace transactions. A few examples will suffice. Several recent treaties and related multinational edicts have strengthened digital content owners' right to control the distribution and presentation of their property online. [FN137] These harmonization efforts grow out of an international copyright regime that is over one hundred years old. [FN138] The G8 economic powers have recently begun to coordinate regulatory efforts concerning cyberspace-related crimes in five areas: pedophilia and sexual exploitation; drug-trafficking; money-laundering; electronic fraud; and industrial and state espionage. [FN139] These initiatives mirror similar efforts to redress similar regulatory leakage problems in real-space contexts such as environmental policy, banking and insurance supervision, and antitrust regulation. [FN140] Several international organizations have drafted model laws and guidelines to facilitate Internet commerce and related digital certification issues. [FN141] There are scores of other international efforts in a variety of cyberspace-related contexts.

      *1232 International harmonization is not always (or even usually) the best response to the spillovers and evasions that result from unilateral regulation. [FN142] And harmonization is often not easy to achieve. However, the proliferation of international organizations, in combination with modern means of communication and transportation, has helped to facilitate international harmonization. Harmonization is especially likely in those contexts--like many aspects of criminal law enforcement--where nations' interests converge and the gains from cooperation are high. But nations sometimes lack the incentive to participate in international regimes, and there are often international and domestic political economy obstacles to harmonization. [FN143] It is too early to tell how successful international efforts will be in addressing the challenges of cyberspace. It is clear, however, that international harmonization will play an important role in nations' overall cyberspace-regulation strategy.

    F. Residual Choice-of-Law Tools

      The skeptics' implicit goal of eliminating all conflicts of laws that arise from cyberspace transactions is unrealistic. Private legal ordering, the limits of enforcement jurisdiction, indirect regulation of extraterritorial activity, filtering and identification technology, and international cooperation facilitate and rationalize legal regulation of cyberspace. These tools, however, will not eliminate all conflicts of laws in cyberspace any more than they do in real space. Transnational activity is too complex. As mentioned above, the elimination of conflict of laws would require the elimination of decentralized lawmaking or of transnational activity. [FN144] In this light, the enormous increases in the pervasiveness and complexity of conflict of laws in this century can be viewed as *1233 an acceptable cost to a world that wishes to expand transnational activity while retaining decentralized lawmaking. As persistent conflicts become prohibitively costly to private parties and regulating nations, public or private international coordination or technological innovation becomes more attractive and thus more likely.

      Short of these developments, transnational transactions in cyberspace, like transnational transactions mediated by telephone and mail, will continue to give rise to disputes that present challenging choice-of-law issues. For example: "Whose substantive legal rules apply to a defamatory message that is written by someone in Mexico, read by someone in Israel by means of an Internet server located in the United States, injuring the reputation of a Norwegian?" [FN145] Similarly,

      (w)hich of the many plausibly applicable bodies of copyright law do we consult to determine whether a hyperlink on a World Wide Web page located on a server in France and constructed by a Filipino citizen, which points to a server in Brazil that contains materials protected by German and French (but not Brazilian) copyright law, which is downloaded to a server in the United States and reposted to a Usenet newsgroup, constitutes a remediable infringement of copyright? [FN146]

      It would be silly to try to formulate a general theory of how such issues should be resolved. One lesson of this century's many failures in top-down choice-of-law theorizing is that choice-of-law rules are most effective when they are grounded in and sensitive to the concrete details of particular legal contexts. This does not mean that standards are better than rules in this context. It simply means that in designing choice-of-law rules or standards, it is better to begin at the micro rather than macro level, and to examine recurrent fact patterns and implicated interests in discrete legal contexts rather than devise a general context-transcendent theory of conflicts. [FN147]

      With these caveats in mind, I want to explain in very general terms why the residual choice-of-law problems implicated by cyberspace *1234 are not significantly different from those that are non-cyberspace conflicts. Cyberspace presents two related choice-of-law problems. The first is the problem of complexity. This is the problem of how to choose a single governing law for cyberspace activity that has multijurisdictional contacts. The second problem concerns situs. This is the problem of how to choose a governing law when the locus of activity cannot easily be pinpointed in geographical space. Both problems raise similar concerns. The choice of any dispositive geographical contact or any particular law in these cases will often seem arbitrary because several jurisdictions have a legitimate claim to apply their law. Whatever law is chosen, seemingly genuine regulatory interests of the nations whose laws are not applied may be impaired.

      The problems of complexity and situs are genuine. They are not, however, unique to cyberspace. Identical problems arise all the time in real space. In fact, they inhere in every true conflict of laws. Consider the problem of complexity. The hypotheticals concerning copyright infringements and multistate libels in cyberspace are no more complex than the same issues in real space. [FN148] They also are no more complex or challenging than similar issues presented by increasingly prevalent real-space events such as airplane crashes, mass torts, multistate insurance coverage, or multinational commercial transactions, all of which form the bread and butter of modern conflict of laws. [FN149] Indeed, they are no more complex than a simple products liability suit arising from a two-car accident among residents of the same state, which can implicate the laws of several states, including the place of the accident, the states where the car and tire manufacturers are headquartered, the states where the car and tires were manufactured, and the state where the car was purchased. [FN150]

      Resolution of choice-of-law problems in these contexts is challenging. But the skeptics overstate the challenge. Not every geographical contact is of equal significance. For example, in the copyright hypothetical above, the laws of the source country and the end-use countries have a much greater claim to governing the copyright action than the laws of the country of the person who *1235 built the server and the country of the server whose hyperlink pointed to the server that contained the infringing material. [FN151] The limits on enforcement jurisdiction may further minimize the scope of the conflict. [FN152] In addition, even in extraordinarily complex cases where numerous laws potentially apply, these laws will often involve similar legal standards, thus limiting the actual choice of law to two or perhaps three options. [FN153] Finally, these complex transactions need not be governed by a single law. Applying different laws to different aspects of a complex transaction is a perfectly legitimate choice-of-law technique. [FN154]

      The application of a single law to complex multijurisdictional conflicts will sometimes seem arbitrary and will invariably produce spillover effects. But as explained above, the arbitrariness of the chosen law, and the spillovers produced by application of this law, inhere in all conflict situations in which two or more nations, on the basis of territorial or domiciliary contacts, have a legitimate claim to apply their law. When in particular contexts the arbitrariness and spillovers become too severe, a uniform international solution remains possible. Short of such harmonization, the choice-of-law issues implicated by cyberspace transactions are no more complex than the issues raised by functionally identical multijurisdictional transactions that occur in real space all the time.

      Like the problem of complexity, the situs problem is a pervasive and familiar feature of real-space jurisdictional conflicts. A classic difficulty is the situs of intangibles like a debt or a bank deposit. [FN155] More generally, the situs problem arises whenever legally significant activity touches on two or more states. For example, when adultery committed in one state alienates the affections of a spouse in another, the situs of the tort is not self-evident. It depends on what contact the forum's choice-of-law rule deems dispositive. Similar locus difficulties arise when the tort takes place over many states, such as when poison is administered *1236 in one state, takes effect in another, and kills in a third. The situs problem even arises when a bodily injury occurs in one state based on negligence committed in another, for there is no logical reason why the place of injury should be viewed as the place of the tort any more than should the place of negligence. [FN156] In all of these situations, the importance of any particular geographical contact is never self-evident; it is a legal rather than a factual consideration that is built into the forum's choice-of-law rules. As the geographical contacts of a transaction proliferate, the choice of any one contact as dispositive runs the risk of appearing arbitrary. But again, this problem pervades real-space conflicts of law and is not unique to cyberspace conflicts.

      So the complexity and situs problems inhere to some degree in all transnational conflicts, and are exacerbated in real space and cyberspace alike as jurisdictional contacts proliferate. No choice-of-law rule will prove wholly satisfactory in these situations. However, several factors diminish the skeptics' concerns about the infeasibility of applying traditional choice-of- law tools to cyberspace. For example, the skeptics are wrong to the extent that they believe that cyberspace transactions must be resolved on the basis of geographical choice-of-law criteria that are sometimes difficult to apply to cyberspace, such as where events occur or where people are located at the time of the transaction. But these are not the only choice-of-law criteria, and certainly not the best in contexts where the geographical locus of events is so unclear. Domicile (and its cognates, such as citizenship, principal place of business, habitual residence, and so on) are also valid choice-of-law criteria that have particular relevance to problems, like those in cyberspace, that involve the regulation of intangibles or of multinational transactions.

      The skeptics are further mistaken to the extent that their arguments assume that all choice-of-law problems must be resolved by multilateral choice-of-law methodologies. A multilateral methodology asks which of several possible laws governs a transaction, and selects one of these laws on the basis of specified criteria. Multilateral methods accentuate the situs and complexity problems. But the regulatory issues that are most relevant to the cyberspace governance debate almost always involve unilateral *1237 choice-of-law methods that alleviate these problems. [FN157] A unilateral method considers only whether the dispute at issue has close enough connections to the forum to justify the application of local law. [FN158] If so, local law applies; if not, the case is dismissed and the potential applicability of foreign law is not considered. For example, a jurisdiction typically does not apply foreign criminal law. If a Tennessee court has personal jurisdiction over someone from across the Virginia border who shot and killed an in-stater, the court does not consider whether Tennessee or Virginia law applies. It considers only whether Tennessee law applies. If so, the case proceeds; if not, it is dismissed. [FN159]

      Unilateral choice-of-law methods make the complexity and situs problems less significant. They do not require a determination of which of a number of possible laws apply. Nor do they require a court to identify where certain events occurred. What matters is simply whether the activity has local effects that are significant enough to implicate local law. By failing to recognize that courts can and will use unilateral rather than multilateral choice-of-law methods to resolve cyberspace conflicts, the skeptics again exaggerate the challenge of cyberspace regulation.

    G. Number and Velocity of Transactions

      The skeptics' final descriptive claim is that even if cyberspace transactions appear like real-space transnational transactions in other respects, they differ significantly with respect to the velocity and number of transactions. [FN160] Cyberspace dramatically lowers the costs of multinational communication. With only a computer and Internet access, anyone in the world can communicate with anyone, and potentially everyone, in the world. The skeptics believe communications via cyberspace will be so prevalent that governments will not find it cost-effective to regulate them. [FN161]

      A dramatic increase in the number and speed of transactions might well multiply the aggregate harms from such transactions. But this increases rather than decreases a nation's incentives to regulate. Consider Internet gambling. In pre-Internet days, individuals *1238 in the United States could gamble from home or work via telephone with domestic and offshore bookies. Although this form of gambling was regulated by a variety of state and federal statutes, the statutes were filled with loopholes and rarely enforced because transactions were relatively infrequent. [FN162] Internet gambling makes it significantly easier to gamble from home or work. This has led to a dramatic increase in gambling and a related rise in the costs of gambling that governments worry about: fraud, diminution in local gambling and other entertainment expenditures, loss of tax revenues, decreased productivity, gambling by children, and so on. Not surprisingly, federal and state governments are beginning to regulate gambling much more extensively, and seriously, than ever. [FN163]

      Even with governments' heightened incentives to regulate Internet transactions, some believe that the sheer number of transactions will overwhelm governments' ability to regulate. A related argument is that because individuals can so easily engage in transnational communications via the Internet, governmental regulation will be less effective; for individuals operating on the Internet are hard to identify, isolate, and thus sanction. Once again, the conclusion that regulation is infeasible simply does not follow from these premises. The mistake here is the belief that governments regulate only through direct sanctioning of individuals. But of course this is not the only way, or even the usual way, that regulation works. Governments regulate an activity by raising the activity's costs in a manner that achieves desired ends. This can be accomplished through several means other than individual sanctions. Governments can, for example, try to alter the social meaning of the activity, regulate the hardware and software through which the activity takes place, make individual penalties severe and notorious, or impose liability on intermediaries like Internet service providers or credit card companies.

      In short, a dramatic increase in the number and velocity of transactions by itself says very little about the feasibility of governmental regulation. Numerous communication advances, beginning with the telegraph, dramatically increased the velocity and number of communications, and lowered their costs. The *1239 skeptics have provided no reason to think that the differences between cyberspace and prior communication technology are so much greater than the differences between pre-and post-telegraph technology (which reduced communication time from weeks and months to hours and minutes), or between pre- and post-telephone technology (which also dramatically reduced the cost and enhanced the frequency and privacy of transjurisdictional communication) to justify the conclusion that governmental regulation will be nonefficacious.

    IV. Is Cyberspace Regulation Legitimate?

      Section III explored some of the many ways that nations might regulate cyberspace transactions. This Section considers the skeptics' normative claim that such regulation is illegitimate. This claim is directed primarily to the application of mandatory laws. The skeptics argue that cyberspace should be self-regulated, and that national mandatory laws should not limit these private legal orders. This argument subsumes three closely related claims: (i) unilateral regulation of cyberspace is extraterritorial; (ii) unilateral regulation of cyberspace produces significant spillover effects; and (iii) the structure of cyberspace makes effective notice of territorial regulation impossible. I address each claim in turn.

    A. Extraterritoriality

      In the Digitalbook.com example above, Singapore and England regulated the local effects of Digitalbook.com's activities in the United States. [FN164] In the CompuServe example, Germany regulated transmission flows from other countries. [FN165] These are the types of extraterritorial regulation that worry the skeptics. But such extraterritorial regulation is commonplace in the modern world. As we saw above, it is settled with respect to real-space activity that a nation's right to control events within its territory and to protect its citizens permits it to regulate the local effects of extraterritorial acts. [FN166]

      The same rationale applies to cyberspace because cyberspace is for these purposes no different than real space. Transactions in cyberspace involve real people in one territorial jurisdiction either (i) transacting with real people in other territorial jurisdictions or (ii) engaging in activity in one jurisdiction that causes *1240 real-world effects in another territorial jurisdiction. To this extent, activity in cyberspace is functionally identical to transnational activity mediated by other means, such as mail or telephone or smoke signal. The new medium of communication is richer, more complex, and much more efficient. But in terms of real-space acts in one jurisdiction that produce real-space effects in another, it is no different from other forms of transnational transaction and communication. And the justification for and legitimacy of regulating local effects is no different. Under current conceptions of territorial sovereignty, a jurisdiction is allowed to regulate extraterritorial acts that cause harmful local effects unless and until it has consented to a higher law (for example, international law or constitutional law) that specifies otherwise.

    B. Spillover Effects

      The skeptics argue that unilateral extraterritorial regulation of cyberspace differs from similar regulation of real-space activities because of the regulation's spillover effects in other jurisdictions. These effects are inevitable, they think, because information flows in cyberspace appear simultaneously in all territorial jurisdictions. As a result, unilateral territorial regulation of the local effects of cyberspace transmission flows will sometimes affect the flow and regulation of web information in other countries. This is especially true when the regulation is directed at a multijurisdictional access provider, as was the case with Germany's regulation of CompuServe.

      Section III described how technology and international cooperation can diminish these spillover effects. But even without these mitigating factors, there is nothing extraordinary or illegitimate about unilateral regulation of transnational activity that affects activity and regulation in other countries. Germany's regulation of CompuServe is no less legitimate than the United States' regulation of the competitiveness of the English reinsurance market, which has worldwide effects on the availability and price of reinsurance. [FN167] Nor is it any different in this regard from national regulation of transborder pollution, or from national consumer protection regulation of transnational contracts, or from national criminal prohibitions on transnational drug activities, all of which produce spillovers. In many contexts, there are powerful reasons for nations to surrender their regulatory prerogatives in order to reduce spillover and other costs. But at least *1241 under our current conceptions of territorial sovereignty, such reforms must proceed by national consent. The need for such consent begins from the premise that in its absence, national regulation of local effects is a legitimate incident of sovereignty, even if such regulation produces spillover effects.

      Germany's regulation of CompuServe is not just a legitimate incident of territorial sovereignty. It is also fair to CompuServe under a straightforward reciprocal benefits rationale. CompuServe reaps financial and other benefits from its presence in Germany. [FN168] Without this presence, German enforcement threats would be largely empty. CompuServe need not remain in Germany; it could close its shop there. Its decision to stay in Germany and comply with German regulations might increase the price of its services in Germany and elsewhere. For CompuServe this is a cost of doing business via a new communication medium. The desire to reduce this and related costs is driving the development of technology that permits geographical and other forms of discrimination on the Internet. [FN169] But even in the absence of such technologies, Germany's local regulation of CompuServe remains within traditional reciprocity-based justifications for regulating local effects.

      What about CompuServe users in other countries who are affected by the German regulation? It is hard to see how the German regulation unfairly burdens them. They remain free to choose among dozens of Internet access services that are not affected by the German regulation. Consider further the German perspective. Germany bans certain forms of pornography within its borders. If the medium of this pornography were paper, there would be no fairness-based jurisdictional objection to a German prohibition on the pornography's entry at the border or to German punishment of those who are later discovered to have smuggled it in. [FN170] From Germany's perspective, it makes no difference whether the pornography enters the nation via cyberspace or the postal service. The rationale for the regulation is the same in both cases: something is happening within Germany that implicates the government's paternalistic concerns or that harms third parties within its borders. The fact that the local regulation might affect the cost or availability of pornography in other countries *1242 is, from this perspective, irrelevant. Fairness does not require Germany to yield local control over its territory in order to accommodate the users of anew communication technology in other countries. Nor does it require Germany to absorb the local costs of foreign activity because of the costs that the German regulation might impose on such activity.

      This latter point sheds light on one of the major fallacies of the skeptics' normative project. The skeptics argue that the spillover effects caused by territorial regulation of cyberspace justify cyberspace self- regulation. Spillover-minimization is not the criterion of legitimacy for national regulation of harmful local effects. [FN171] But even if it were, the skeptics' conclusions would not follow. For the skeptics completely ignore the spillover effects of cyberspace activity itself. They do not consider these effects because they take it as an article of faith that cyberspace participants form a self-contained group that can internalize the costs of its activity. [FN172] But this assumption is false. Cyberspace participants are no more self-contained than telephone users, members of the Catholic Church, corporations, and other private groups with activities that transcend jurisdictional borders. They are real people in real space transacting in a fashion that produces real-world effects on cyberspace participants and nonparticipants alike. Cyberspace users solicit and deliver kiddie porn, launder money, sexually harass, defraud, and so on. It is these and many other real-space costs--costs that cyberspace communities cannot effectively internalize--that national regulatory regimes worry about and aim to regulate.

      So the spillover argument runs in both directions. Cyberspace activity outside of Germany produces spillovers in Germany, and German regulation produces spillovers on cyberspace activity beyond its borders. The legitimacy and fairness of Germany's territorial regulation does not depend on minimization of these costs. But even if it did, the skeptics' desired normative conclusion that cyberspace should be self-regulated would only follow if the costs of cyberspace self-regulation were less significant than the costs of territorial regulation. The skeptics have not begun to try to demonstrate that this is true. And any such attempt is very unlikely to succeed at the level of generality at which their arguments are invariably pitched.*1243

    C. Notice

      The skeptics' final normative argument against mandatory law regulation of cyberspace concerns notice. In real space, parties can direct the flow of their transnational transactions and can in most cases avoid jurisdictions that prohibit the transactions. The skeptics claim that this cannot be done in cyberspace. They worry that cyberspace participants therefore lack notice about governing mandatory law and hence cannot conform their behavior to it. The skeptics claim this lack of notice violates basic norms of fairness.

      This argument rests on a number of empirical assumptions that have been questioned in Section III. The assumption that cyberspace involves uncontrollable universal information flows is inaccurate today and will become even less accurate with time. Information flows can be directed and controlled in a variety of ways, with varying costs that will almost certainly decrease in the future. [FN173] Concerns about notice are further attenuated by the many limitations on enforcement jurisdiction that effectively limit the application of mandatory laws to entities with a local presence. [FN174] In none of the many cases in which regulations have been enforced against cyberspace transactions has an out-of-state defendant had a basis to claim unfair surprise.

      It is nonetheless worth considering how the notice issue will play out in cyberspace. The Constitution and international law impose weak notice requirements on the application of local law to extraterritorial conduct. The Constitution permits a state with significant contacts to the case to apply its law if the defendant could have reasonably foreseen its application. [FN175] International law might impose a similar restraint on legislative jurisdiction.

      This requirement of reasonable foreseeability does not mean that harmful local effects of extraterritorial activity are automatically immune from local regulation just because they were accidental, or because the agent of the activity did not know the precise locus of the effects. "Reasonable foreseeability" is a dynamic *1244 concept. A manufacturer that pollutes in one state is not immune from the antipollution laws of other states where the pollution causes harm just because it cannot predict which way the wind blows. Similarly, a cyberspace content provider cannot necessarily claim ignorance about the geographical flow of information as a defense to the application of the law of the place where the information appears. At first glance it appears unfair to expose Digitalbook.com to the antipornography laws of Singapore. But it would not seem unfair if Digitalbook.com could at a small cost prevent its information from entering Singapore. Nor would it seem unfair to expose Digitalbook.com to liability for the damage caused in Singapore by a virus that it released into cyberspace that destroyed every Apple computer hard drive connected to the Internet.

      These intuitions show that, like the related personal jurisdiction question,  [FN176] the standard of foreseeability depends on a complex mixture of what the content provider knows or reasonably should have known about the geographical consequences of its acts, the significance of the extrajursidictional harms caused by the acts, and the costs of precautions. [FN177] Content providers can already achieve pretty reliable information flow control by conditioning access to content on telephone or facsimile proof of geographical location. To many this is an unacceptable burden on Internet communication. But there is nothing sacrosanct about Internet speed and ease, and dimunitions in speed and ease might be warranted by the social costs imposed by uncontrolled information flows. And in any event, as filtering and identification technologies continue to raise the feasibility and lower the costs of information flow control, the problem of notice in cyberspace will look much like the problem of notice in real space.

    V. Grounding Cyberspace in Real-Space Law

      I have argued that national and international regulations of cyberspace transactions are legitimate and feasible. I have not argued for any particular regulation, or that such regulation should be pervasive. I have tried to show only that the skeptics' global arguments against national and international regulation of cyberspace are unfounded. Cyberspace self-regulation will often *1245 be difficult to achieve. And like non-cyberspace transactions, cyberspace transactions will in any event be limited by national mandatory rules.

      The challenging issue from a jurisdictional perspective is to develop a legal structure that both facilitates private legal ordering of cyberspace transactions and accommodates national mandatory law limitations. Consider the predicament of a person in England who wants to buy a security from a web page on a server in Japan. The parties want the sale to circumvent U.S. securities regulations, and, more broadly, the interference of national courts. The parties thus agree that the sale will be governed by Japanese law and that any disputes will be resolved by private arbitration in Japan. Will this contract be enforceable? This example involves a commercial transaction. But the problem is generic, for, as we saw above, parties can by contract create governing legal structures for a variety of non-commercial activities. Thus, for example, the same basic problem arises when chat room participants from different countries agree that the tort law principles of the state of Illinois govern chat room activities, and that all disputes will be resolved privately. Will the parties' ex ante consent be respected when a chat room participant from France claims in his national court that he suffered a tort in the chat room in violation of French law?

      To avoid national court litigation and minimize national regulation, the parties to these transactions need to satisfy the following conditions. They must consent ex ante to a governing law, a private method of dispute resolution, and a private enforcement regime. The consent must be consistent with the mandatory law applied by any national court where a defector from the contract might seek to have any part of the contract declared invalid. [FN178] To ensure the sanctity of the private order and to discourage *1246 such defection, the national court must be willing to (i) treat the consent to the private order as valid, (ii) enjoin litigation in derogation of the contract, and, sometimes, (iii) specifically enforce the defector's agreement to abide by the private order. Moreover, it is not enough that the courts of a single country will enforce the contract. There must be coordinated enforcement among national courts in every country in which the recalcitrant party might go to seek to avoid the obligation. Finally, national courts must subsequently recognize the validity of the private dispute resolution process. They must enjoin subsequent litigation in derogation of the results of the private dispute resolution, and enforce any judgments that cannot be done so privately.

      Such a structure might appear hopelessly complicated and thus fanciful. But this appearance is deceiving. The essentials for such a regime already exist in the system that governs international commercial arbitration. [FN179] This system works through the interplay of three layers of law. The first layer is the private law of the parties' contract. In the contract, the parties specify the law governing the transaction (in the examples above, the laws of Japan and Illinois), agree to use private arbitration to resolve certain disputes that arise out of or relate to the transaction, and choose the place for the arbitration and the procedures that govern it. [FN180] The second layer is the national arbitration law. [FN181] 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. Most nations have generally similar national arbitration laws that ensure harmonization of enforcement across jurisdictions. This harmonization is substantially bolstered by the third layer of legal regulation: the international *1247 enforcement treaty. By far the most important such treaty is the New York Convention on the Recognition and Enforcement of Arbitral Awards, which almost every nation has signed. [FN182] The Convention obligates the national courts of signatory states to recognize and enforce arbitration agreements and awards, subject to limited exceptions. [FN183]

      The basic structure of international commercial arbitration could easily be modified to cyberspace. As explained above, the law of the contract-- boththe substantive law and the dispute resolution mechanism--could be agreed to as an incident of the securities transaction or as a condition of access to the chat room. [FN184] National arbitration laws could be modified to include dispute resolution in cyberspace. For example, the Federal Arbitration Act ("FAA") would require modification in only two important respects. First, the FAA's requirement that the arbitration agreement be made in writing might need to be amended to accommodate cyberspace realities. [FN185] Second, FAA rules that turn on the place of the arbitration [FN186] require modification for virtual arbitrations that lack a geographical locus. [FN187] The New York Convention would likely require similar amendment.

      *1248 The accommodation of mandatory laws presents special challenges. In the securities example, assume that the English purchaser is unhappy with the security, and defects from the contractual agreement to arbitrate by bringing a private securities action in a U.S. court that alleges that the sale was fraudulent and in violation of U.S. securities law. This raises two basic mandatory law issues. The first is whether the U.S. court will enforce the agreement to arbitrate, or will instead adjudicate the mandatory law (and perhaps other) claims. Assuming the court enforces the arbitration agreement, the second question is whether the arbitrator can apply the U.S. mandatory law consistent with the jurisdictional limits imposed by the parties' contractual choice of Japanese law.

      Both difficulties frequently arise with respect to non-cyberspace transnational transactions, and can be addressed within the framework of international commercial arbitration. As for the first problem, national courts increasingly permit private arbitrators to resolve claims involving economic regulation and quasi-criminal laws subject to subsequent, deferential judicial review. [FN188] The deferential nature of such review, combined with the costs of seeking it, mean that private arbitrators will often have the final say. As for the second problem, arbitrators have established a number of devices grounded in (often fictional) party consent *1249 that permit them to apply a mandatory law of a country other than the one specifically chosen by the parties' contract. [FN189]

      To many it will seem ironic and damning that my description of a legal regime that supposedly promotes private ordering focuses so much on the role of national courts and national laws. This focus is misleading. Much of the regulation of these private matters is and will continue to be governed by a variety of privately enforceable rules, norms, and enforcement mechanisms. Yet the overarching national and international legal regimes remain necessary for two reasons. First, they provide a ready-made coordination and enforcement regime that transnational parties can invoke in the many situations in which information-gathering and related costs of purely private enforcement are prohibitively high. Second, they give private parties enormous flexibility in creating a private regime in a fashion that can accommodate and minimize the intrusion of oft-conflicting mandatory laws. In this connection, it should be emphasized that the international commercial arbitration model is not as litigious, and would not be as intrusive on private cyberspace orders, as it might at first glance appear. If real-space commercial arbitrations are any guide, recourse to national courts will be relatively infrequent as the background public enforcement patterns become relatively clear.

      I do not mean to suggest that international commercial arbitration is a comprehensive panacea for the jurisdictional challenges of cyberspace. It is not. Many, probably most, cyberspace transactions will have such a low value that affected parties will not bother to enter into contractual relations, much less contract for governing law and private enforcement. In addition, cyberspace transactions that adversely affect third parties are beyond the ken of international commercial arbitration, which depends upon ex ante consent for its effectiveness and legitimacy. Related-ly, although the international arbitration regime has taken steps to privatize the enforcement of mandatory laws, many mandatory laws--most prominently traditional criminal laws and certain limits on contractual capacity--are not subject to enforceable international arbitration. Indeed, some might object that cyberspace-related choice-of-law and private arbitration agreements that are not dickered should be viewed as unenforceable contracts of adhesion. [FN190]

      *1250 These limitations on the international commercial law regime are not, of course, unique to cyberspace transactions. These same limitations characterize real-space transnational transactions. Such limitations are inevitable when it is difficult for parties to transnational transactions to craft private legal regimes ex ante, or when these transactions harm third parties or implicate the paternalistic interests of affected nations. The important point is that these limitations are difficult to overcome in "real space" and cyberspace alike. My modest aim has been to show that the governing law challenges presented by cyberspace are not significantly different from the ones presented by other transnational transactions.

    Conclusion

      Cyberspace transactions are no different from "real-space" transnational transactions. They involve people in real space in one jurisdiction communicating with people in real space in other jurisdictions in a way that often does good but sometimes causes harm. There is no general normative argument that supports the immunization of cyberspace activities from territorial regulation. And there is every reason to believe that nations can exercise territorial authority to achieve significant regulatory control over cyberspace transactions. Resolution of the choice-of-law problems presented by cyberspace transactions will be challenging, but no more challenging than similar problems raised in other transnational contexts.

    FOOTNOTES

    [FNd1]. Associate Professor of Law, The University of Chicago. For their comments and discussion, I thank Bill Arms, Caroline Arms, Curtis Bradley, Stephen Choi, Richard Craswell, David Currie, Larry Downes, Richard Epstein, Michael Froomkin, Elizabeth Garrett, Andrew Guzman, Larry Kramer, Larry Lessig, Doug Lichtman, Richard Posner, David Post, Cass Sunstein, Tim Wu, and participants at workshops at the University of Chicago and the University of California (Boalt Hall). I also thank Kyle Gehrmann and Greg Jacob for excellent research, and the Arnold and Frieda Shure Research Fund for support.

    [FN1]. See Communications Decency Act of 1996 ("CDA"), Pub L No 104-104, 110 Stat 133, codified at 47 USCA §§ 223, 230, 303, 560-61, 609 (1991 & Supp 1998); Reno v ACLU, 117 § Ct 2329, 2346 (1997) (holding that CDA's prohibition on Internet transmission of indecent or offensive messages to minors violates the First Amendment).

    [FN2]. I shall use the terms "state," "nation," and "jurisdiction" interchangeably to refer to national, as opposed to subnational, legal authority. I shall indicate when the analysis differs for subnational units. Although the term "cyberspace" has a broader meaning, I shall use it here loosely as a synonym for the Internet--the transnational network of computer networks.

    [FN3]. See James Boyle, Foucault in Cyberspace: Surveillance, Sovereignty, and Hardwired Censors, 66 U Cin L Rev 177, 178 (1997) ("For a long time, the Internet's enthusiasts have believed that it would be largely immune from state regulation."). The leading regulation skeptics, and this Article's primary targets, are David Post and David Johnson. See David R. Johnson and David Post, Law And Borders--The Rise of Law in Cyberspace, 48 Stan L Rev 1367, 1367 (1996). See also David Post and David R. Johnson, Borders, Spillovers, and Complexity: Rule-making Processes in Cyberspace (and Elsewhere), draft presented at the Olin Law & Economics Symposium on "International Economic Regulation" at Georgetown University Law Center (Apr 5, 1997) (copy on file with U Chi L Rev); David Post and David R. Johnson, The New 'Civic Virtue' of the Internet (also published in The Emerging Internet, Feb 1998, the Annual Review of the Institute for Information Studies), available online at < www.cli.org.paper4.htm> (visited Sept 28, 1998); David G. Post, Governing Cyberspace, 43 Wayne L Rev 155 (1996); David G. Post, Anarchy, State, and the Internet: An Essay on Law-Making in Cyberspace, 1995 J Online L, Article 3, available online at <www.wm.edu/law/publications/jol/post.html> (visited Sept 10, 1998). Commentators who have made similar arguments include John T. Delacourt, The International Impact of Internet Regulation, 38 Harv Intl L J 207 (1997); John Parry Barlow, A Cyberspace Independence Declaration, available online at <www.eff.org/barlow> (visited Sept 10, 1998); Dan L. Burk, Federalism in Cyberspace, 28 Conn L Rev 1095 (1996); Joel R. Reidenberg, Governing Networks and Rule-making in Cyberspace, 45 Emory L J 911 (1996).

    [FN4]. See, for example, ACLU v Reno, 929 F Supp 824, 832 (E D Pa 1995), affd, 117 § Ct 2329, 2348 (1997); Digital Equipment Corp v Altavista Technology, Inc, 960 F Supp 456, 462 (D Mass 1997); American Libraries Associations v Pataki, 969 F Supp 160, 170 (S D NY 1997).

    [FN5]. See, for example, Thomas E. Weber, The Internet (A Special Report): Debate: Does Anything Go? Limiting free speech on the Net, Wall Street J (Dec 8, 1997); Vinton G. Cerf, Building an Internet Free of Barriers, NY Times § 3 p 12 (July 27, 1997); George Black, Call for Controls: The Internet Must Regulate Itself, Fin Times part 4 p 12 (Apr 1, 1998).

    [FN6]. Johnson and Post, 48 Stan L Rev at 1367 (cited in note 3).

    [FN7]. See generally Irwin Lebow, Information Highways and Byways: From the Telegraph to the 21st Century (IEEE 1995); Dan Lacy, From Grunts to Gigabytes: Communications and Society (Illinois 1996).

    [FN8]. See, for example, Naxos Resources (U.S.A.) Ltd v Southam, Inc, 1996 US Dist LEXIS 21757, *13-15 (C D Cal) (defamation); Panavision International, LP v Toeppen, 938 F Supp 616, 619 (C D Cal 1996) (interference with economic advantage); Sally Greenberg, Threats, Harassment, and Hate Online: Recent Developments, 6 BU Pub Intl L J 673, 673-75, 680-84 (1997) (harassment and threats).

    [FN9]. See, for example, Thompson v Handa-Lopez, Inc, 998 F Supp 738 (W D Tex 1998) (contract made online).

    [FN10]. See, for example, United States v Thomas, 74 F3d 701, 704-05 (6th Cir), cert denied, 117 § Ct 74 (1996) (pornography); < www.personal.psu.edu/users/j/m/jm

    [FN11]. See, for example, Zippo Manufacturing Co v Zippo Dot Com, Inc, 952 F Supp 1119, 1121 (W D Pa 1997) (trademark infringement); Religious Technology Center v F.A.C.T.-NET, Inc, 901 F Supp 1519, 1521-22 (D Colo 1995) (copyright infringement); Anatomy of a Cyber Break-in, Newsweek 63 (Feb 27, 1995) (data theft).

    [FN12]. See, for example, Maritz, Inc v CyberGold, Inc, 947 F Supp 1328, 1329 (E D Mo 1996) (unfair competition); Robert A. Robertson, Personal Investing in Cyberspace and the Federal Securities Laws, 23 Sec Reg L J 347, 397-405 (1996) (fraudulent securities).

    [FN13]. The arguments from this paragraph are drawn from Johnson and Post,  48 Stan L Rev at 1368-70 (cited in note 3); Post and Johnson, The New 'Civic Virtue' at 5-6 (cited in note 3); Reidenberg, 45 Emory L J at 912-16 (cited in note 3).

    [FN14]. See Johnson and Post, 48 Stan L Rev at 1374-75 (cited in note 3); Burk, 28 Conn L Rev at 1098, 1110-12 (cited in note 3).

    [FN15]. Post and Johnson, Borders, Spillovers, and Complexity at 5 (cited in note 3).

    [FN16]. See Johnson and Post, 48 Stan L Rev at 1370-72 (cited in note 3); Post and Johnson, Borders, Spillovers, and Complexity at 6 (cited in note 3).

    [FN17]. See Johnson and Post, 48 Stan L Rev at 1372-73 (cited in note 3).

    [FN18]. See id at 1374-76.

    [FN19]. See id at 1374.

    [FN20]. See Post, 1995 J Online L, Article 3, para 39-40 (cited in note 3).

    [FN21]. See Johnson and Post, 48 Stan L Rev at 1376 (cited in note 3); Burk, 28 Conn L Rev at 1123-34 (cited in note 3).

    [FN22]. See Post and Johnson, Borders, Spillovers, and Complexity at 38  (cited in note 3); Post and Johnson, The New 'Civic Virtue' at 5-6 (cited in note 3); Burk, 28 Conn L Rev at 1123-34 (cited in note 3).

    [FN23]. See Johnson and Post, 48 Stan L Rev at 1370, 1379 & n 33 (cited in note 3).

    [FN24]. The skeptics' views about territorialism and choice of law are remarkably similar to Story's and Beale's. See, for example, Joseph Story, Commentaries on the Conflict of Laws 7 (Little, Brown 2d ed 1841); Joseph Henry Beale, A Treatise on the Conflict of Laws or Private International Law 118 (Harvard 1916).

    [FN25]. The claim that the territorialist premises of the traditional approach to choice of law were flawed does not necessarily mean that the traditional choice-of-law rules that were based on these premises cannot in some circumstances be justified on independent grounds. See Alfred Hill, The Judicial Function in Choice of Law, 85 Colum L Rev 1585, 1619-36 (1985).

    [FN26]. The classic criticisms of the traditional view are Brainerd Currie, Selected Essays on the Conflict of Laws (Duke 1963), and Walter Wheeler Cook, The Logical and Legal Bases of the Conflict of Laws (Harvard 1949).

    [FN27]. This is one reason why so many of the transformative midcentury constitutional choice-of-law decisions involved public regulations rather than private law. See, for example, Clay v Sun Insurance Office, Ltd, 377 US 179, 182-83 (1964) (insurance); Watson v Employers Liability Assurance Corp, 348 US 66, 72-73 (1954) (insurance); United States v Aluminum Co of America, 148 F2d 416, 444 (2d Cir 1945) (antitrust); Pacific Employers Insurance Co v Industrial Accident Commission of California, 306 US 493, 497 (1939) (workmen's compensation); Alaska Packers Association v Industrial Accident Commission of California, 294 US 532, 538 (1935) (workmen's compensation); Bradford Electric Light Co v Clapper, 286 US 145, 150-51 (1932) (workmen's compensation); Home Insurance Co v Dick, 281 US 397, 405-08 (1930) (insurance).

    [FN28]. See, for example, Cook, The Logical and Legal Bases at 311-22, 354- 70, 433-37 (cited in note 26); Walter Wheeler Cook, The Jurisdiction of Sovereign States and the Conflict of Laws, 31 Colum L Rev 368, 372-80 (1931); Ernest G. Lorenzen, Selected Articles on the Conflict of Laws 305-21 (Yale 1947).

    [FN29]. See Cook, The Logical and Legal Bases at 35-36 (cited in note 26); Hessel E. Yntema, The Hornbook Method and the Conflict of Laws, 37 Yale L J 468, 478 (1928).

    [FN30]. See Leon E. Trakman, The Law Merchant: The Evolution of Commercial Law 39-44 (Fred B. Rothman 1983); Bradford R. Clark, Federal Common Law: A Structural Reinterpretation, 144 U Pa L Rev 1245, 1280-81 (1996).

    [FN31]. See Friedrich K. Juenger, American Conflicts Scholarship and the New Law Merchant, 28 Vand J Transnatl L 487, 491 (1995).

    [FN32]. The main approaches used by the several states today are the traditional vested rights approach, interest analysis, the Second Restatement, comparative impairment, and the better law approach. See Lea Brilmayer, Conflict of Laws: Cases and Materials 203-314 (Little, Brown 4th ed 1995). Even states that purport to use the same methodology--for example, interest analysis or the Second Restatement--often do so in name only, with important differences in practice.

    [FN33]. In the United States, the horizontal nonuniformity fostered by different choice-of-law regimes in different states was exacerbated by the rule that federal courts sitting in diversity apply state choice-of-law rules. See Klaxon Co v Stentor Electric Manufacturing Co, Inc, 313 US 487, 496 (1941).

    [FN34]. Phillips Petroleum Co v Shutts, 472 US 797, 818 (1985), quoting  Allstate Insurance Co v Hague, 449 US 302, 312-13 (1981).

    [FN35]. For example, in the case in which this modern standard was formulated, the Supreme Court held that Minnesota could apply its plaintiff- favoring insurance law to an accident in Wisconsin among Wisconsin residents based on the fact that the decedent worked in Minnesota, the insurance company did business there, and the beneficiary moved there from Wisconsin after the accident. See Hague, 449 US at 315-20. On the weaknesses and uncertainties of the Hague test, see Brilmayer, Conflict of Laws at 140-43 (cited in note 32).

    [FN36]. The Permanent Court of International Justice famously established a very weak effects test for extraterritorial jurisdiction and suggested a default rule that favored extraterritorial jurisdiction. See The case of the S.S. "Lotus", 1927 P C I J (ser A) No 10 at 18-25. Section 403 of the Restatement (Third) of the Foreign Relations Law (ALI 1987), recognized the effects test as a basis for extraterritorial jurisdiction, but added the caveat that a state may not exercise such jurisdiction when it would be "unreasonable" to do so. This reasonableness requirement has little basis in state practice and does not reflect customary international law. See William S. Dodge, Extraterritoriality and Conflict-of-Laws Theory: An Argument for Judicial Unilateralism, 39 Harv Intl L J 101, 139-40 & nn 241-42 (1998).

    [FN37]. See, for example, Blackmer v United States, 284 US 421, 436  (1932); United States v Reeh, 780 F2d 1541, 1543 n 2 (11th Cir 1986); Restatement (Third) of the Foreign Relations Law § 402(2). International law also permits a nation to regulate extraterritorial conduct that threatens local security, Restatement (Third) of the Foreign Relations Law § 402(3), and might permit a nation to regulate certain extraterritorial acts against its citizens, id at comment g.

    [FN38]. See Shutts, 472 US at 823; Hague, 449 US at 307; Restatement  (Third) of the Foreign Relations Law § 403(3).

    [FN39]. Some courts will not enforce choice-of-law agreements in which the  "chosen state has no substantial relationship to the parties or the transaction and there is no other reasonable basis for the parties' choice." Restatement (Second) of Conflict of Laws § 187(2)(a) (1971). This restriction has less force in transnational contexts in which there are often good reasons for parties to choose a neutral law unrelated to the parties. See id § 187 comment f.

    [FN40]. See, for example, Moses v Business Card Express, Inc, 929 F2d 1131, 1138 (6th Cir 1991).

    [FN41]. See, for example, Hague Convention on the Law Applicable to Succession to the Estates of Deceased Persons, Art 5, 28 ILM 146, 150 (1989) (providing that an individual may designate either the law of habitual residence or the law of nationality to govern succession); Hague Convention on the Law Applicable to Trusts and on Their Recognition, Art 6, 23 ILM 1389 (1984) ("A trust shall be governed by the law chosen by the settlor."); McDermott Inc v Lewis, 531 A2d 206, 215 (Del 1987) (holding that law of place of incorporation governs internal corporate affairs); William Grantham, Comment, The Arbitrability of International Intellectual Property Disputes, 14 Berkeley J Intl L 173, 190-95 (1996) (describing how parties' choice of law governs intellectual property disputes).

    [FN42]. See Michael J. Trebilcock, The Limits of Freedom of Contract 58-77, 145-63 (Harvard 1993).

    [FN43]. See, for example, Vimar Seguros y Reaseguros, SA v M/V Sky Reefer, 515 US 528, 540-41 (1995) (noting that transnational parties cannot reduce their liability under the Carriage of Goods at Sea Act by contracting around its provisions); Mitsubishi Motors Corp v Soler Chrysler-Plymouth, Inc, 473 US 614, 637 n 19 (1985) (noting that transnational parties cannot contract around the Sherman Act). Private parties can, of course, circumvent mandatory laws to the extent that they can shift the location or effects of their activities beyond the mandatory law's enforceable scope. For further discussion, see note 178.

    [FN44]. See, for example, United Nations Convention on Contracts for the International Sale of Goods ("CISG"), UN Doc A/CONF.97/18, reprinted at 19 ILM 671 (1980). A related solution is to develop uniform laws like the Uniform Commercial Code, which minimize choice-of-law difficulties by ensuring that every jurisdiction's local law is (in theory) the same.

    [FN45]. See, for example, Convention on the Law Applicable to Contractual Obligations ("Rome Convention"), June 19, 1980 (80/934/EEC) 1980 OJ (L266/1), p 1.

    [FN46]. See, for example, id Arts 3(3), 5(2), 6(1), and 7(1)- (2) (acknowledging various mandatory law restrictions on choice of law governing contracts).

    [FN47]. I say "usually" because sometimes there will be parallel litigation of the same matter in two nations, each of which attempts to apply its own law. See, for example, Laker Airways Ltd v Sabena, 731 F2d 909, 917-20 (DC Cir 1984).

    [FN48]. See Andreas F. Lowenfeld, International Litigation and the Quest for Reasonableness: Essays in Private International Law 5 (Clarendon 1996).

    [FN49]. In my discussion here and throughout the Article, I shall follow the skeptics in assuming that the spillovers produced by unilateral regulation of transnational activity are negative spillovers. This will not always be true, but it will usually be true in situations in which one state regulates extraterritorial conduct that the territorial government would regulate differently.

    [FN50]. 509 US 764 (1993).

    [FN51]. Id at 796.

    [FN52]. See McDonnell Douglas-Boeing Link Gets Europe Approval, NY Times D4 (July 31, 1997).

    [FN53]. See Hague, 449 US at 306, 319-20.

    [FN54]. See In re Grand Jury Proceedings United States v Bank of Nova Scotia, 691 F2d 1384, 1391 (11th Cir 1982).

    [FN55]. See Alaska Packers, 294 US at 539-44.

    [FN56]. See Babcock v Jackson, 12 NY2d 473, 240 NYS2d 743, 191 NE2d 279, 284-85 (1963).

    [FN57]. See Barclays Bank PLC v Franchise Tax Board, 512 US 298, 310-15  (1994).

    [FN58]. This is the goal, for example, of such different approaches as the Restatement of the Foreign Relations Law's interest-balancing approach, see Restatement (Third) of the Foreign Relations Law § 403; William Baxter's comparative impairment approach, see William F. Baxter, Choice of Law and the Federal System, 16 Stan L Rev 1, 4-20 (1963); Larry Kramer's multistate canons of construction, see Larry Kramer, Rethinking Choice of Law, 90 Colum L Rev 277, 319-38 (1990); and Lea Brilmayer's strategy to maximize state policy objectives, see Brilmayer, Conflict of Laws at 169-218 (cited in note 32).

    [FN59]. See David W. Leebron, Lying Down with Procrustes: An Analysis of Harmonization Claims, in Jagdish N. Bhagwati and Robert E. Hudec, eds, 1 Fair Trade and Harmonization 41, 43-50 (MIT 1996).

    [FN60]. For more general discussions of this point, see Joel R. Reidenberg, Lex Informatica: The Formulation of Information Policy Rules Through Technology, 76 Tex L Rev 553 (1998); Lawrence Lessig, Reading the Constitution in Cyberspace, 45 Emory L J 869, 895-99 (1996); M. Ethan Katsh, Software Worlds and the First Amendment: Virtual Doorkeepers in Cyberspace, 1996 U Chi Legal F 335, 339-47; Post, 1995 J Online L, Article 3, paras 20- 21 (cited in note 3).

    [FN61]. Of course, computer hardware--keyboards, monitors, modems, disk drives, processors, and the like--also affects how individuals experience cyberspace. Many software instructions that are interpreted by a computer could be instantiated in hardware rather than software. For the most part, however, hardware is less significant than software in creating and shaping one's experience of cyberspace. See Katsh, 1996 U Chi Legal F at 339-43 (cited in note 60).

    [FN62]. <www.dell.com/dell/legal/disclwww.htm> (visited Apr 1, 1998).

    [FN63]. International Institute for the Unification of Private Law  ("UNIDROIT"), Principles of International Commercial Contracts (UNIDROIT 1994); ICC, Uniform customs and practice for documentary credits, ICC Pub No 500 (1993).

    [FN64]. Such a regime will invariably be underspecified and will require supplementation by some default law regime.

    [FN65]. "Virtual Magistrate" is the name of the decisionmaker in a relatively new online project "for resolving disputes that arise on worldwide computer networks about online messages, postings, and files . . . ." The Virtual Magistrate Project Concept Paper, available online at < vmag.vcilp.org/docs/vmpaper.html> (visited Apr 1, 1998).

    [FN66]. See Reidenberg, 76 Tex L Rev at 558-68 (cited in note 60).

    [FN67]. See Jack Goldsmith and Lawrence Lessig, Grounding the Virtual Magistrate 4, available online at <www.law.vill.edu/ncair/disres/groundvm.htm> (visited Apr 1, 1998). For further development of these points, see Section V.

    [FN68]. See, for example, Johnson and Post, 48 Stan L Rev at 1387-91  (cited in note 3); I. Trotter Hardy, The Proper Legal Regime for "Cyberspace", 55 U Pitt L Rev 993, 1028-33 (1994).

    [FN69]. See generally Eric A. Posner, The Regulation of Groups: The Influence of Legal and Nonlegal Sanctions on Collective Action, 63 U Chi L Rev 133, 165-97 (1996).

    [FN70]. See Goldsmith and Lessig, Grounding the Virtual Magistrate at 3-4  (cited in note 67); Johnson and Post, 48 Stan L Rev at 1395-1400 & nn 102- 03 (cited in note 3).

    [FN71]. The skeptics challenge the normative basis for nations to apply mandatory laws to regulate the private legal regimes of cyberspace. I consider these arguments in Section IV.

    [FN72]. I discuss private legal ordering in cyberspace in greater detail in Section V.

    [FN73]. See Buchanan v Rucker, 9 East 192, 103 Eng Rep 546, 547 (KB 1808) ("Can the Island of Tobago pass a law to bind the rights of the whole world?").

    [FN74]. I set aside for present purposes two other relatively rare methods of extraterritorial enforcement: military invasion, see, for example, United States v Noriega, 746 F Supp 1506 (S D Fla 1990), and secondary boycotts, see, for example, Cuban Liberty and Democratic Solidarity (Libertad) Act of 1996 ("Helms-Burton Act"), Pub L No 104-114, 110 Stat 785, codified at 22 USCA §§ 6021-91 (1994 & Supp 1998) (sanctioning nations that engage in certain transactions with Cuba).

    [FN75]. I explain below why local regulation of service or content providers that produces multijurisdictional spillover effects is legitimate and fair, see Section IV.B; my goal for now is to show that the scope of national regulation of cyberspace is much narrower than the skeptics claim.

    [FN76]. See US Const, Art IV, § 1; Roger C. Crampton, et al, Conflict of Laws: Cases-Comments-Questions 735-37 (West 5th ed 1993).

    [FN77]. See Asahi Metal Industry Co v Superior Court, 480 US 102, 108-09  (1987). This is the test for specific jurisdiction; such jurisdiction is limited to cases in which the cause of action arises out of or relates to the defendant's contacts with the forum. A court may also assert general personal jurisdiction over a defendant for a cause of action that accrued anywhere. General jurisdiction is normally limited to the defendant's domicile and anywhere else where it may have "continuous and systematic . . . contacts." Helicopteros Nacionales de Columbia v Hall, 466 US 408, 414-16 (1984). Courts are unanimous that a web page accessible in a jurisdiction does not by itself establish general jurisdiction there. See, for example, Weber v Jolly Hotels, 977 F Supp 327, 333-34 (D NJ 1997).

    [FN78]. For broader treatments, see Henry H. Perritt, Jr., Jurisdiction in Cyberspace, 41 Vill L Rev 1, 13-25 (1996); Burk, 28 Conn L Rev at 1107- 23 (cited in note 3).

    [FN79]. See, for example, Cybersell, Inc v Cybersell, Inc, 130 F3d 414, 419-20 (9th Cir 1997); Weber, 977 F Supp at 334; Panavision International, LP v Toeppen, 938 F Supp 616, 622 (C D Cal 1996). For more comprehensive analyses of the many Internet personal jurisdiction cases, see Howard B. Stravitz, Personal Jurisdiction in Cyberspace: Something More is Required on the Electronic Stream of Commerce, 49 SC L Rev 925 (1998); Christopher W. Meyer, Note, World Wide Web Advertising: Personal Jurisdiction Around the Whole Wide World?, 54 Wash & Lee L Rev 1269 (1997).

    [FN80]. See Section III.D.

    [FN81]. See Burk, 28 Conn L Rev at 1117-20 (cited in note 3).

    [FN82]. I should emphasize the term "enforceable" here, because many commentators talk as if there are (or should be) customary international law limits on exorbitant assertions of personal jurisdiction. See, for example, Restatement (Third) of the Foreign Relations Law § 421. This talk does not appear to be supported by state practice followed from a sense of legal obligation, the usual requirements for a rule of customary international law. The Brussels and Lugano Conventions are treaties that specify the legal bases for personal jurisdiction among members of the European Union. See Convention on Jurisdiction and the Enforcement of Judgments in Civil and Commerical Matters, Sept 27, 1968, 1990 OJ (C 189) 2 (consolidated); Convention on Jurisdiction and the Enforcement of Judgments in Civil and Commercial Matters, Sept 16, 1988, 1988 OJ (L 319) 9. These rare treaties on the subject prove the point that there is no effective or established customary international law that regulates personal jurisdiction, for the Brussels-Lugano regime permits exorbitant assertions of personal jurisdiction against defendants from non- European countries. See Friedrich Juenger, Judicial Jurisdiction in the United States and in the European Communities: A Comparison, 82 Mich L Rev 1195, 1211 (1984).

    [FN83]. The U.S. Constitution's Full Faith and Credit obligation does not extend to judgments of foreign nations. See US Const, Art IV, § 1 (requiring states to give full faith and credit to acts, records, and proceedings "of every other State") (emphasis added). The enforceability of these judgments is generally regulated by state law and is weaker than the obligation imposed by the Full Faith and Credit Clause. See Gary B. Born, International Civil Litigation in United States Courts: Commentary & Materials 938-62 (Kluwer 3d ed 1996). In Europe, the Brussels and Lugano Conventions, which regulate the enforcement of foreign judgments among European Union members, are again an exception to the general rule.

    [FN84]. See Born, International Civil Litigation at 942-43 (cited in note 83) (discussing examples from Japan, Germany, and England).

    [FN85]. See, for example, Bachchan v India Abroad Publications Inc, 154 Misc 2d 228, 585 NYS2d 661, 664-65 (NY Sup Ct 1992) (declining to enforce English money judgment for libel against a newspaper whose activities would have been protected by the First Amendment in the United States). See generally Born, International Civil Litigation at 942-43 (cited in note 83).

    [FN86]. See US Const, Art IV, § 2, cl 2; 18 USC § 3182 (1994).

    [FN87]. See Innes v Tobin, 240 US 127, 131 (1916); Hyatt v People, 188 US 691, 711-12 (1903); Gee v Kansas, 912 F2d 414, 418 (10th Cir 1990). This jurisdictional limitation does not apply, of course, when a person in one state commits a federal crime in another. See United States v Thomas, 74 F3d 701, 709-10 (6th Cir 1996).

    [FN88]. For an overview, see I.A. Shearer, Extradition in International Law (Manchester 1971).

    [FN89]. See John T. Soma, Thomas F. Muther, Jr., and Heidi M.L. Brissette, Transnational Extradition for Computer Crimes: Are New Treaties and Laws Needed?, 34 Harv J Leg 317, 323-26 (1997).

    [FN90]. See, for example, Johnson and Post, 48 Stan L Rev at 1374 (cited in note 3).

    [FN91]. This is the topic of the next Section.

    [FN92]. I examine this normative question in Section IV.C.

    [FN93]. This component of the skeptics' argument is in tension with their concerns about the threat of multiple regulation. If, as they claim, cyberspace users can easily relocate the source of their transmissions to evade legal enforcement, and if, as they further claim, the physical location of parties transacting in cyberspace is "indeterminate both ex ante and ex post," see Post and Johnson, Borders, Spillovers, and Complexity at 3 (cited in note 3), then cyberspace users have little to fear from multiple national regulation.

    [FN94]. See Post, 1995 J Online L, Article 3 at para 40 (cited in note 3).

    [FN95]. Telnet allows a computer user to log into a remote computer over the Internet. Once connected to the foreign computer, the user can perform any Internet function, such as sending e-mail or "telnetting" to yet other servers, as though she were logged on to a terminal at the foreign computer's location. An anonymous remailer is a service that allows the sender of an e-mail to remain anonymous by sending the message through an intermediary that strips the message of the identifying characteristics of the original sender. The receiver of the e-mail can respond to the e-mail by sending an e-mail to the intermediary, which then forwards it to the sender.

    [FN96]. See generally Stephen D. Krasner, Global Communications and National Power: Life on the Pareto Frontier, 43 World Pol 336, 343-46 (1991).

    [FN97]. Id.

    [FN98]. Consider two of many examples. Pending legislation in the United States Congress would impose criminal penalties on persons in the United States who gamble on the Internet. See Internet Gambling Prohibition Act of 1998, § Amend 3266 to § 2260, 105th Cong, 2d Sess (July 22, 1998); Internet Gambling Prohibition Act of 1998, HR 4427, 105th Cong, 2d Sess (Aug 6, 1998). Chinese law punishes in-state Internet users who access or transmit a broader array of prohibited information. See Computer Information Network and Internet Security, Protection and Management Regulations (approved by the State Council on December 11, 1997 and promulgated by the Ministry of Public Security on December 30, 1997), available online at <www.gilc.org/speech/china/net-regs- 1297.html> (visited Sept 11, 1998).

    [FN99]. For example, a new German law imposes liability on Internet service providers if they knowingly offer a venue for content illegal in Germany and fail to use technically possible and reasonable means to block it. See Germany to Enforce Child-Friendly Internet, Chi Trib 4 (July 5, 1997). Australia is about to implement a similar law. See Electronic Frontiers Australia, Internet Regulation in Australia, available online at < www.efa.org.au/Issues/Censor/cens1.html> (visited Sept 10, 1998). In the United States, pending federal Internet gambling legislation would authorize the federal government to order service providers, at risk of penalty, to discontinue the availability of illegal gambling sites. See Internet Gambling Prohibition Act of 1998, § Amend 3266 to § 2260 (cited in note 98). Legislation is also pending that would require senders of unsolicited commercial e-mail to identify themselves in a way that would enable Internet service providers to filter such messages. See Jeri Clausing, Compressed Data; House E-Mail Effort Raises Censorship Issues, NY Times D3 (Aug 10, 1998). And some states have held Internet service providers liable for facilitating the transmission of illegal extraterritorial content into the regulating jurisdiction. See, for example, Stratton Oakmont, Inc v Prodigy Services Co, 1995 WL 323710, *4-5 (NY Sup Ct).

    [FN100]. This form of regulation is explored in detail in Section III.D.

    [FN101]. Compare Matt Beer, The wagers of the Web; Lawsuit could unravel on- line gaming industry, San Fran Examiner B1 (Aug 17, 1998) (describing lawsuit by Internet bettor against credit card companies that financed online gambling).

    [FN102]. As Lessig notes: 
    A regulation need not be absolutely effective to be sufficiently effective. It need not raise the cost of the prohibited activity to infinity in order to reduce the level of that activity quite substantially. If regulation increases the cost of access to this kind of information, it will reduce access to this information, even if it doesn't reduce it to zero. . . . If government regulation had to show that it was perfect before it was justified, then indeed there would be little regulation of cyberspace, or of real space either. But regulation, whether for the good or the bad, has a lower burden to meet. 
    Lawrence Lessig, The Zones of Cyberspace, 48 Stan L Rev 1403, 1405 (1996).

    [FN103]. As we shall see in Section III.E, nations can further regulate extraterritorial supply through international harmonization.

    [FN104]. See Nathaniel Nash, Holding Compuserve Responsible, NY Times D4  (Jan 15, 1996).

    [FN105]. See id.

    [FN106]. See, for example, Johnson and Post, 48 Stan L Rev at 1373 & n 20 (cited in note 3).

    [FN107]. See Edmund L. Andrews, Germany's Effortsto Police Web Are Upsetting Business, NY Times A1 (June 6, 1997).

    [FN108]. See id; Germany Brings Criminal Charges Against CompuServe Manager, Eurowatch (May 2, 1997). The CompuServe executive was later convicted, even though at trial's end the prosecution sought acquittal because it agreed with the defense that, at the time of the indictment, CompuServe lacked the technological means to block the illegal material. Battle of the Somm (May 29, 1998), available online at <www.wired.com/news/ news/politics/story/12607.html> (visited Sept 10, 1998).

    [FN109]. Jordan Bonfante, The Internet Trials: Germany Makes an Early Attempt at Taming the Wide, Wild Web. But Many are Crying Foul--or Folly, Time 30 (Intl ed July 14, 1997) (emphasis added).

    [FN110]. See ACLU White Paper, Fahrenheit 451.2: Is Cyberspace Burning? How Rating and Blocking Proposals May Torch Free Speech on the Internet (1997), available online at <www.aclu.org/issues/cyber/burning.html> (visited Apr 1, 1998) (warning of censorship threats posed by rating and filtering proposals that flourished in wake of Reno); Lawrence Lessig, Tyranny in the Infrastructure, Wired (July 1997), available online at < www.wired.com/wired/5.07/cyber_rights.html> (visited Apr 1, 1998) (same).

    [FN111]. See Lessig, Tyranny in the Infrastructure, Wired (cited in note 110); Boyle, 66 U Cin L Rev at 191-96 (cited in note 3).

    [FN112]. In addition to the discussion below, see Jonathan Weinberg, Rating the Net, 19 Hastings Comm/Enter L J 453 (1997); Paul Resnick, Filtering Information on the Internet, Scientific Am 62-64 (Mar 1997); Reidenberg, 76 Tex L Rev at 556-68 (cited in note 60).

    [FN113]. For example, Sexroulette.com includes the following conditions upon entry to its pages: 
    WARNING: You are about to enter an ADULT ONLY area. You must agree to the following terms before proceeding: . . . . If you are under the age of eighteen years . . . you are not authorized to download any materials from XPICS and any and all such downloading shall constitute intentional infringement of XPICS's rights in such materials. 
    All materials, messages, and other communications contained at XPICS are intended for distribution exclusively to consenting adults in locations where such materials, messages and other communications do not violate any community standards or any federal, state or local law or regulation of the United States or any other country. No materials from any parts of XPICS designated as "XXX" are authorized to or otherwise may be downloaded to persons located in the following areas: Alabama, Florida, except Ft. Lauderdale, Miami, and St. Petersburg, Georgia, except Atlanta, Kansas, except Kansas City, Kentucky, Minnesota, Missouri, Mississippi, North Carolina, Ohio, except Cleveland and Cincinnati, Pennsylvania, except Philadelphia and Pittsburgh, South Carolina, Tennessee, except for Nashville, Utah, Afghanistan, Kuwait, Iran, Iraq, Japan, Jordan, Libya, Pakistan, The Republic of China, Singapore, Saudi Arabia, Syria, The United Arab Emirates, or any other place in which to do so would constitute a violation of any law, regulation, rule or custom. Any and all unauthorized downloading of materials from XPICS shall constitute intentional infringement of XPICS's rights in such materials. 
    . . . 
    If you agree with the above, you may ENTER. If you don't agree, you must EXIT. 
    <members.sexroulette.com> (visited Apr 1, 1998).

    [FN114]. See, for example, The Adult Check System, <www.adultcheck.com>  (visited Sept 10, 1998).

    [FN115]. In many of the well-known cyberspace regulation cases, the defendants knew that they were sending content into the regulating jurisdiction because they had conditioned the users' request for the content on the presentation of information (including geographical identification) by fax or mail. See, for example, Thomas, 74 F3d at 705; Playboy Enterprises, Inc v Chuckleberry Publishing, Inc, 939 F Supp 1032, 1035 (S D NY 1996); State v Granite Gate Resorts, Inc, 1996 WL 767431, *4 (Minn Dist Ct), affd, 568 NW2d 715 (Minn App 1997).

    [FN116]. See generally Introduction to Client Digital ID's SM, available online at <www.verisign.com/repository/brwidint.htm> (visited Apr 1, 1998); A. Michael Froomkin, The Essential Role of Trusted Third Parties In Electronic Commerce, 75 Or L Rev 49, 55-60 (1996).

    [FN117]. First generation software filters blocked access to individually compiled lists of prohibited Internet addresses. See Weinberg, 19 Hastings Comm/Enter L J at 457 (cited in note 112). More recently, a much more sophisticated industry-wide standard for labeling, rating, and filtering Internet information has emerged. This standard, known as PICS, establishes content-neutral labeling formats and distribution methods. The PICS format does not specify a labeling vocabulary or what should be done with the labels. Instead, the PICS format allows both content providers and independent entities to label content along several dimensions. Selection software then decides what to do with these labels--whether to block them, restrict access, highlight them, organize them in certain ways, or whatever else the software is designed for.

    [FN118]. See Timothy S. Wu, Note, Cyberspace Sovereignty?--The Internet and the International System, 10 Harv J L & Tech 647, 649-56 (1997); Delacourt, 38 Harv J Intl L at 208-19 (cited in note 3); Paul Resnick, PICS, Censorship, & Intellectual Freedom FAQ, version 1.14, available online at < www.si.umich.edu/ <<degrees>> presnick/pics/intfree/faq.htm> (visited Apr 1, 1998).

    [FN119]. See Wu, Note, 10 Harv J L & Tech at 651 (cited in note 118) ("As of July 1996, at least thirty-three states were completely unconnected.").

    [FN120]. See id at 652-54 (China, Singapore); Madanmohan Rao, Persian Gulf Net Censorship: Governments Force Server Blockades (Oct 3, 1997), available online at <media info.elpress.com/ephome/news/newshtm/webnews/glob1003.htm> (visited Apr 1, 1998) (United Arab Emirates).

    [FN121]. See text accompanying notes 104-09.

    [FN122]. See Christopher Stern, V-chip on fast track as FCC ok's tech spex, Variety 27 (Mar 16-22, 1998); Brooks Boliek, Television sharpens bite: V-chip wins FCC approval, Hollywood Reporter 15 (Mar 13, 1998).

    [FN123]. See Clausing, Compressed Data, NY Times D3 (cited in note 99).

    [FN124]. Many predict that Congress will more broadly require filtering or digital identification technology to be built into the architecture of cyberspace. See Boyle, 66 U Cin L Rev at 193, 202-04 (cited in note 3); Lawrence Lessig, What Things Regulate Speech: CDA 2.0 vs. Filtering, 38 Jurimetrics J 629 (forthcoming 1998). Even in the absence of direct governmental mandates, the threat of such regulation has already spurred the development and adoption of an array of private de facto Internet discrimination standards that facilitate extensive private regulation of the Net.

    [FN125]. See J.M. Balkin, Media Filters, the V-Chip, and the Foundations of Broadcast Regulation, 45 Duke L J 1131, 1141-44 (1996) (describing the filtering of information in print and broadcast media).

    [FN126]. Id at 1143.

    [FN127]. See Robert H. Reid, Architects of the Web: 1000 Days that Built the Future of Business xxiii-xxiv (Wiley 1997).

    [FN128]. Some civil libertarians favor information filtering technologies because they allow individuals--rather than the government--to decide what information is appropriate for their own (or their children's) consumption. See, for example, Brief of Feminists for Free Expression as Amicus Curiae in support of Appellees, Reno v ACLU, 117 § Ct 2329 (1997), available at 1997 WL 74382, *15-16.

    [FN129]. Resnick, PICS, Censorship & Intellectual Freedom at 3 (cited in note 118).

    [FN130]. See Johnson and Post, 48 Stan L Rev at 1373-74 & n 20 (cited in note 3); notes 122-24 and accompanying text.

    [FN131]. See, for example, Weinberg, 19 Hastings Comm/Enter L J at 459-70  (cited in note 112); Johnson and Post, 48 Stan L Rev at 1373-74 (cited in note 3); Electronic Privacy Information Center, Faulty Filters: How Content Filters Block Access to Kid-Friendly Information on the Internet, available online at < www2.epic.org/reports/filter_report.htm> (visited Apr 1, 1998).

    [FN132]. See, for example, Johnson and Post, 48 Stan L Rev at 1372-74  (cited in note 3).

    [FN133]. See note 102 and accompanying text.

    [FN134]. See, for example, Johnson and Post, 48 Stan L Rev at 1375-76  (cited in note 3).

    [FN135]. See notes 113-16 and accompanying text.

    [FN136]. See discussion in Section II.

    [FN137]. In December 1996, the World Intellectual Property Organization  ("WIPO") reached agreement on a treaty that significantly extended international copyright protection for digital property. See WIPO Copyright Treaty, adopted Dec 20, 1996, WIPO Pub No 226(E) (WIPO 1997); Seth Schiesel, Global Agreement Reached to Widen Law on Copyright, NY Times 1 (Dec 21, 1996). Within a year, the European Commission issued a draft directive to bring European law into line with these international obligations. See Draft EC Directive Provides Strong Online Copyright Protection, Outlaws Devices Facilitating Infringement, BNA Electronic Commerce and L Rep (Mar 13, 1998), available online at < www.bna.com/e-law/main.htm> (visited Apr 1, 1998). The United States is in the process of enacting similar legislation. See WIPO Copyright Treaties Implementation Act, HR 2281, 105th Cong, 1st Sess (July 29, 1997); Digital Millenium Copyright Act of 1998, 2037, 105th Cong, 2d Sess (May 6, 1998).

    [FN138]. The digital protection treaty signed in Geneva operates as a protocol to the Berne Convention for the Protection of Literary and Artistic Works, a treaty regime that began in 1886. See Berne Convention for the Protection of Literary and Artistic Works (WIPO 1970).

    [FN139]. See Clifford Krauss, 8 Countries Join in an Effort To Catch Computer Criminals, NY Times A12 (Dec 11, 1997).

    [FN140]. See Anne-Marie Slaughter, The Real New World Order, Foreign Affairs 183, 189-92 (Sep/Oct 1997).

    [FN141]. For example, in February 1997, the United Nations Commission on International Trade Law ("UNCITRAL") began to draft model international digital signature legislation. See Report of the Working Group on Electronic Commerce, Thirty-First Session (New York, Feb 12-28, 1997). See also UNCITRAL Working Group on Electronic Commerce, Planning Of Future Work on Electronic Commerce: Digital Signatures, Certification Authorities and Related Legal Issues A/CN.9/WG.IV/WP.71 (Dec 31, 1996). Similarly, in November 1997, the International Chamber of Commerce issued the General Usage for International Digitally Ensured Commerce ("GUIDEC"), a set of guidelines for ensuring trustworthy digital transactions over the Internet, available online at < www.iccwbo.org/guidec2/htm> (visited Apr 5, 1998). And the Organisation for Economic Co-operation and Development ("OECD") recently adopted principles to guide countries in formulating their own policies and legislation relating to the use of cryptography. See OECD Cryptography Policy: The Guidelines and the Issues, Unclassified OCDE/GD(97)204 (1997), available online at <www.oecd.org/ dsti/sti/it/secur/prod/GD97-204.htm> (visited Apr 2, 1998).

    [FN142]. See Leebron, Lying Down with Procrustes (cited in note 59).

    [FN143]. See, for example, Krasner, 43 World Pol at 337-60 (cited in note 96); Andrew T. Guzman, Is International Antitrust Possible?, 73 NYU L Rev (forthcoming 1998).

    [FN144]. See text accompanying notes 58-59.

    [FN145]. Perritt, 41 Vill L Rev at 3 (cited in note 78).

    [FN146]. Post and Johnson, Borders, Spillovers, and Complexity at 2-3 (cited in note 3).

    [FN147]. Many European conflict systems demonstrate that it is both possible and useful to design choice-of-law rules that are context-sensitive and not beholden to any grand choice-of-law theory. See, for example, Swiss Private International Law Statute of December 18, 1987, translated in Andreas Bucher and Pierre-Yves Tschanz, International Arbitration in Switzerland 225 (Helbing & Lichtenhahn 1988).

    [FN148]. See London Film Productions v Intercontinental Communications, Inc, 580 F Supp 47, 48-49 (S D NY 1984) (involving a British corporation suing an American corporation for copyright infringement in Chile, Venezuela, Peru, Equador, Costa Rica, and Panama); Eugene F. Scoles and Peter Hay, Conflict of Laws 631 (West 2d ed 1992) (describing choice-of-law rules for multistate libel).

    [FN149]. See generally Larry Kramer, Choice of Law in Complex Litigation, 71 NYU L Rev 547, 551-65 (1996).

    [FN150]. See, for example, Rutherford v Goodyear Tire and Rubber Co, 943 F Supp 789, 790-91 (W D Ky 1996), affd, 142 F3d 436 (6th Cir 1998).

    [FN151]. For an excellent analysis of how traditional choice-of-law rules might apply to copyright violations in cyberspace, see Jane C. Ginsburg, Copyright Without Borders?: Choice of Forum and Choice of Law for Copyright Infringement in Cyberspace, 15 Cardozo Arts & Enter L J 153, 168-74 (1997).

    [FN152]. See Section III.B.

    [FN153]. See Kramer, 71 NYU L Rev at 583 (cited in note 149).

    [FN154]. This is known as "depecage." See Brilmayer, Conflict of Laws at 366 (cited in note 32).

    [FN155]. See Peter S. Smedresman and Andreas F. Lowenfeld, Eurodollars, Multinational Banks, and National Laws, 64 NYU L Rev 733, 734-37 (1989) (discussing the bank deposit problem); Andreas F. Lowenfeld, In Search of the Intangible: A Comment on Shaffer v. Heitner, 53 NYU L Rev 102, 115-17, 122-24 (1978) (exploring situs of debt problem largely in context of personal jurisdiction).

    [FN156]. See Larry Kramer, Vestiges of Beale: Extraterritorial Application of American Law, 1991 § Ct Rev 179, 190 n 36. A similar problem is presented by multistate contracts. When contractual negotiations and signings take place in two states, the place of the contract might be either state, depending on what contact the forum's choice-of-law rule deems dispositive for this purpose.

    [FN157]. See Lowenfeld, International Litigation and the Quest for Reasonableness at 5 (cited in note 48).

    [FN158]. See Dodge, 39 Harv Intl L J at 108-10 (cited in note 36).

    [FN159]. The same analysis applies in the international context for the extraterritorial application of other regulatory laws like RICO and the antitrust and securities laws. The question in these cases is whether Congress intended for federal law to apply to conduct abroad. If so, these laws apply. If not, the court dismisses the case without considering the application of foreign law.

    [FN160]. See Johnson and Post, 48 Stan L Rev at 1372-73 (cited in note 3).

    [FN161]. See id.

    [FN162]. See Britta Gordon, Gaming on the Internet: The Odds are on the House, But How Long Will it Last? 5, available online at < www.cyberlaw.law.ttu.edu/cyberspc/jour9.htm> (visited Apr 1, 1998).

    [FN163]. There is currently legislation pending in Congress that would extensively regulate Internet gambling by, among other things, penalizing online bettors and authorizing governmental officials to order Internet service providers to shut down offending online cites. See note 98.

    [FN164]. See text following note 23.

    [FN165]. See text accompanying notes 104-09.

    [FN166]. See notes 34-37 and accompanying text.

    [FN167]. See Hartford Fire Insurance Co, 509 US at 795-99.

    [FN168]. In late 1996, CompuServe had 335,000 German subscribers and employed over 250 workers there. Compuserve May Curb German Operations, NY Times D6 (Nov 19, 1996).

    [FN169]. See Section III.D.

    [FN170]. There might of course be substantive objections akin to the First Amendment found either in German law or in international human rights law. As I mentioned at the outset, such substantive limitations on cyberspace regulation are not my concern here.

    [FN171]. See discussion in Section II.

    [FN172]. See Johnson and Post, 48 Stan L Rev at 1378-91 (cited in note 3).

    [FN173]. See Section III.D.

    [FN174]. See Section III.B.

    [FN175]. I glean this formulation from Phillips Petroleum Co v Shutts, 472 US 797, 807 (1985); Allstate Insurance Co v Hague, 449 US 302, 312-13 (1981) (plurality opinion); Clay v Sun Insurance Office, Ltd, 377 US 179, 182 (1964); Watson v Employers Liability Assurance Corp, 348 US 66, 72-73 (1954); and Home Insurance Co v Dick, 281 US 397, 410 (1930). Like all formulations of constitutional limitations on choice of law, this one is open to debate because the Court's analysis in these decisions is maddeningly vague, and because the Court has mixed due process and full faith and credit concerns. See note 34 and accompanying text. But at the very least, the formulation in the text is close enough to the constitutional requirement of notice to consider the application of this test to cyberspace.

    [FN176]. See text accompanying notes 78-85.

    [FN177]. What is foreseeable will also be informed, in a circular fashion, by what the law requires. If the law permits one jurisdiction to hold a content provider in another jurisdiction strictly liable for the mere act of placing information on a web page, then this result is foreseeable. The pertinent question is what level of foreseeability the law should require, and this analysis is informed by the factors listed in the text.

    [FN178]. There are of course some private legal orders that are relatively immune from this requirement, either because the costs of defecting from the private order are greater than the costs of continued participation, see Lisa Bernstein, Opting Out of the Legal System: Extralegal Contractual Relations in the Diamond Industry, 21 J Legal Stud 115 (1992); Posner, 63 U Chi L Rev at 165-97 (cited in note 69), or, relatedly, because the members of the order structure their affairs to circumvent mandatory laws. See Frank H. Easterbrook and Daniel R. Fischel, The Economic Structure of Corporate Law 3 (Harvard 1991) (corporate context). Even these prominent examples of mandatory law circumvention, however, are not clear-cut. For example, Bernstein's paradigmatic examples of private legal orders are sometimes subject to mandatory law interventions. See Bernstein, 21 J Legal Stud at 125 & n 24, 129 & n 35 (noting that arbitration awards can be vacated for procedural irregularity and that the group's actions must conform with antitrust regulations). Similarly, a corporation's ability to circumvent mandatory law restrictions is dependent in large part on the settled but by no means inevitable choice-of-law rule that the law of the place of incorporation governs internal corporate affairs. Mandatory corporate laws would be harder to avoid if nations and states instead applied local corporate law on the basis of the local effects of corporate transactions. See, for example, Western Airlines, Inc v Sobieski, 191 Cal App 2d 399, 12 Cal Rptr 719, 727-29 (1961) (applying California cumulative voting rule to Delaware corporation doing significant business in California). Because for a variety of reasons it frequently will be difficult for cyberspace communities to completely circumvent mandatory law restrictions, I will set aside this possibility in the analysis.

    [FN179]. Two excellent introductions to international commercial arbitration are Gary B. Born, International Commercial Arbitration in the United States: Commentary and Materials (Kluwer 1994), and Alan Redfern and Martin Hunter, Law and Practice of International Commercial Arbitration (Sweet & Maxwell 2d ed 1991).

    [FN180]. 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 UNCITRAL. These rules are generally similar but contain important differences. See Born, International Commercial Arbitration at 10-16, 50-96 (cited in note 179).

    [FN181]. In the United States, this law is the Federal Arbitration Act, Pub L No 282, 61 Stat 669 (1947), codified at 9 USC §§ 1 et seq (1994).

    [FN182]. The Convention is reproduced at 9 USC § 201 (1994). For general commentary, see Born, International Commercial Arbitration at 18-20 (cited in note 179); A. Jan van den Berg, The New York Arbitration Convention of 1958: Towards a Uniform Judicial Interpretation (Kluwer 1981).

    [FN183]. 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 noncommercial arbitrations. New York Convention on the Recognition and Enforcement of Foreign Arbitral Awards, Arts I, II, June 10, 1958, 330 UNTS 38. Second, the obligation to enforce arbitration agreements and awards does not extend to matters that, under national law, are nonarbitrable or violate a strong public policy. Id at Arts II(1), V(2). I discuss this exception further below. See note 188. Third, the duty to enforce arbitral awards does not extend to awards rendered without minimal due process protections (such as notice). Id at Art V(1)(b). And fourth, the duty to enforce arbitral awards does not extend to ultra vires awards. Id at Art V(1)(c).

    [FN184]. See Section III.A.

    [FN185]. It is possible, however, that an agreement in cyberspace constitutes an agreement in writing. For an overview of various responses to this problem, see Michael E. Schneider and Christopher Kuner, Dispute Resolution in International Electronic Commerce, 14 J Intl Arb 5, 13-15 (1997); Jasna Arsic, International Commercial Arbitration on the Internet: Has the Future Come Too Early?, 14 J Intl Arb 209, 215-17 (1997).

    [FN186]. See, for example, 9 USC § 4 (1994) (authorizing federal court to order arbitration "within the district in which the petition for an order directing such arbitration is filed"); 9 USC § 10(b) (1994) (providing for limited judicial review by "(t)he United States district court for the district wherein an award was made").

    [FN187]. 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 place within the jurisdiction. Because the locus of an arbitration in cyberspace is difficult to identify, many jurisdictions might assert the power of judicial review and assistance. One answer to this problem is coordination of the judicial assistance function. Such enforceable coordination could be accomplished most effectively by international treaty. Indeed, the New York Convention's judicial enforcement provisions could be modified to cover judicial assistance. Another (less effective) possibility is to make national arbitration laws uniform, so that it doesn't matter which court provides judicial assistance. This is the basic strategy of the UNCITRAL model arbitration law. See generally Born, International Commercial Arbitration at 37-38 (cited in note 179). For an overview of other solutions to these difficulties, see Arsic, 14 J Intl Arb at 217-20 (cited in note 185).

    [FN188]. The Supreme Court, for example, has ruled that antitrust, RICO, and securities claims are arbitrable. Shearson/American Express, Inc v McMahon, 482 US 220, 242 (1987) (holding RICO claims to be arbitrable); Mitsubishi Motors Corp v Soler Chrysler-Plymouth, Inc, 473 US 614, 632-40 (1985) (holding Sherman Act claims to be arbitrable); Scherk v Alberto-Culver Co, 417 US 506, 518-20 (1974) (holding claims under the Securities and Exchange Act of 1934 to be arbitrable). See generally Born, International Commercial Arbitration at 322-66 (cited in note 179). These decisions are not required by the New York Convention. To the contrary, The New York Convention and national arbitration laws permit an exception to national courts' obligation to enforce arbitration agreements and awards when the arbitration involves a nonarbitrable subject. See New York Convention, Arts II(1), V(2) (cited in note 183). But this exception has been construed in an increasingly narrow fashion, as nations increasingly delegate the task of enforcing mandatory laws to private arbitrators.

    [FN189]. See Born, International Commercial Arbitration at 147-52 (cited in note 179).

    [FN190]. I have tried to avoid analysis of the merits of particular regulatory regimes in this article, but it is perhaps worth noting that this adhesion contracts concern has relatively little force in the cyberspace context, where users have an array of options for the large majority of functions and services.

    END OF DOCUMENT 
     

    1.4 Yahoo! Inc. v. La Ligue Contre Le Racisme et L'antisemitisme 1.4 Yahoo! Inc. v. La Ligue Contre Le Racisme et L'antisemitisme

    This short case note illustrates some of the issues associated with cross-border disputes

    433 F.3d 1199 (2006)

    YAHOO! INC., a Delaware corporation, Plaintiff-Appellee,
    v.
    LA LIGUE CONTRE LE RACISME ET L'ANTISEMITISME, a French association; L'Union Des Etudiants Juifs De France, a French association, Defendants-Appellants.

    No. 01-17424.

    United States Court of Appeals, Ninth Circuit.

    Argued and Submitted March 24, 2005.
    Filed January 12, 2006.

    [1200] Randol Schoenberg, Burris & Schoenberg, Los Angeles, CA; Robert A. Christopher, Coudert Brothers, Palo Alto, CA; Mark D. Lebow, Sokolow Carreras, New York, NY, for the defendants-appellants.

    Michael Traynor, Cooley, Godward, Castro, Huddelson & Tatum, San Francisco, CA; Robert C. Vanderet, O'Melveney & Myers, Los Angeles, CA, for the plaintiff-appellee.

    Ann Brick, ACLU, San Francisco, California; John B. Morris, Jr., Alan B. Davidson, Center for Democracy & Technology, Washington, DC, for amici American Booksellers Foundation for Free Expression, et al.

    Jodie L. Kelley, Brian Hauck, Jenner & Block, Washington, DC; Stephen A. Bokat, Robin S. Conrad, Joshua A. Ulman, National Chamber Litigation Center, Washington, DC, Robert Corn-Revere, Davis Wright Tremaine LLP, Washington, [1201] DC, for amici Chamber of Commerce of the United States, et al., and for amicus Center for Democracy.

    Before SCHROEDER, Chief Judge, and FERGUSON, O'SCANNLAIN, HAWKINS, TASHIMA, W. FLETCHER, FISHER, GOULD, PAEZ, CLIFTON, and BEA, Circuit Judges.

    PER CURIAM.

    A majority of the en banc court (Judge W.A. Fletcher, joined by Chief Judge Schroeder and Judges Hawkins, Fisher, Gould, Paez, Clifton, and Bea) concludes that the district court had personal jurisdiction over the defendants. Of that majority, three judges (Chief Judge Schroeder, and Judges W.A. Fletcher and Gould) conclude that the action should be dismissed for lack of ripeness. Five judges (Judge Fisher, joined by Judges Hawkins, Paez, Clifton, and Bea) conclude that the case is ripe for adjudication. The three remaining judges (Judges Ferguson, O'Scannlain, and Tashima) conclude that the action should be dismissed because the district court lacked personal jurisdiction over the defendants.

    A majority of the en banc court having voted therefor, the judgment of the district court is REVERSED and the case REMANDED with directions to dismiss the action without prejudice.

    W. FLETCHER, Circuit Judge, with whom SCHROEDER, Chief Circuit Judge, and GOULD, Circuit Judge, join as to the entire opinion, and with whom HAWKINS, FISHER, PAEZ, CLIFTON and BEA, Circuit Judges, join as to Parts I and II:

    Yahoo!, an American Internet service provider, brought suit in federal district court in diversity against La Ligue Contre Le Racisme et L'Antisemitisme ("LICRA") and L'Union des Etudiants Juifs de France ("UEJF") seeking a declaratory judgment that two interim orders by a French court are unrecognizable and unenforceable. The district court held that the exercise of personal jurisdiction over LICRA and UEJF was proper, that the dispute was ripe, that abstention was unnecessary, and that the French orders are not enforceable in the United States because such enforcement would violate the First Amendment. The district court did not reach the question whether the orders are recognizable. LICRA and UEJF appeal only the personal jurisdiction, ripeness, and abstention holdings. A majority of the en banc panel holds, as explained in Part II of this opinion, that the district court properly exercised personal jurisdiction over LICRA and UEJF. A plurality of the panel concludes, as explained in Part III of this opinion, that the case is not ripe under the criteria of Abbott Laboratories v. Gardner, 387 U.S. 136, 149, 87 S.Ct. 1507, 18 L.Ed.2d 681 (1967). We do not reach the abstention question.

    I. Background

    Yahoo! is a Delaware corporation with its principal place of business in California. Through its United States-based website yahoo.com, Yahoo! makes available a variety of Internet services, including a search engine, e-mail, web page hosting, instant messaging, auctions, and chat rooms. While some of these services rely on content created by Yahoo!, others are forums and platforms for user-generated content.

    Yahoo! users can, for example, design their own web pages, share opinions on social and political message boards, play fantasy baseball games, and post items to be auctioned for sale. Yahoo! does not monitor such user-created content before [1202] it is posted on the web through Yahoo! sites.

    Yahoo!'s United States website is written in English. It targets users in the United States and relies on servers located in California. Yahoo!'s foreign subsidiaries, such as Yahoo! France, Yahoo! U.K., and Yahoo! India, have comparable websites for their respective countries. The Internet addresses of these foreign-based websites contain their two-letter country designations, such as fr.yahoo.com, uk.yahoo.com, and in.yahoo.com. Yahoo!'s foreign subsidiaries' sites provide content in the local language, target local citizens, and adopt policies that comply with local law and customs. In actual practice, however, national boundaries are highly permeable. For example, any user in the United States can type www.fr.yahoo.com into his or her web browser and thereby reach Yahoo! France's website. Conversely, any user in France can type www.yahoo.com into his or her browser, or click the link to Yahoo.com on the Yahoo! France home page, and thereby reach yahoo.com.

    Sometime in early April 2000, LICRA's chairman sent by mail and fax a cease and desist letter, dated April 5, 2000, to Yahoo!'s headquarters in Santa Clara, California. The letter, written in English, stated in part:

    [W]e are particularly choked [sic] to see that your Company keeps on presenting every day hundreds of nazi symbols or objects for sale on the Web.
    This practice is illegal according to French legislation and it is incumbent upon you to stop it, at least on the French Territory.
    Unless you cease presenting nazi objects for sale within 8 days, we shall size [sic] the competent jurisdiction to force your company to abide by the law.

    On April 10, five (rather than eight) days after the date on the letter, LICRA filed suit against Yahoo! and Yahoo! France in the Tribunal de Grande Instance de Paris. On April 20, UEJF joined LICRA's suit in the French court. LICRA and UEJF used United States Marshals to serve process on Yahoo! in California.

    After a hearing on May 15, 2000, the French court issued an "interim" order on May 22 requiring Yahoo! to "take all necessary measures to dissuade and render impossible any access [from French territory] via Yahoo.com to the Nazi artifact auction service and to any other site or service that may be construed as constituting an apology for Nazism or a contesting of Nazi crimes" (emphasis added).[1] Among other things, the French court required Yahoo! to take particular specified actions "[b]y way of interim precautionary measures." Yahoo! was required "to cease all hosting and availability in the territory of [France] from the `Yahoo.com' site . . . of messages, images and text relating to Nazi objects, relics, insignia, emblems and flags, or which evoke Nazism," and of "Web pages displaying text, extracts, or quotes from `Mein Kampf' and the `[Protocols of the Elders of Zion]'" at two specified Internet addresses. Yahoo! was further required to remove from "all browser directories accessible in the territory of the French Republic" the "index heading [1203] entitled `negationists'" and any link "bringing together, equating, or presenting directly or indirectly as equivalent" sites about the Holocaust and sites by Holocaust deniers.

    The May 22 interim order required Yahoo! France (as distinct from Yahoo!) to remove the "negationists" index heading and the link to negationist sites, described above, from fr.yahoo.com. The order further required Yahoo! France to post a warning on fr.yahoo.com stating to any user of that website that, in the event the user accessed prohibited material through a search on Yahoo.com, he or she must "desist from viewing the site concerned[,] subject to imposition of the penalties provided in French legislation or the bringing of legal action against him."

    The order stated that both Yahoo! and Yahoo! France were subject to a penalty of 100,000 Euros per day of delay or per confirmed violation, and stated that the "possibility of liquidation of the penalties thus pronounced" was "reserve[d]." The order also awarded 1 Franc in "provisional damages," payable by Yahoo! and Yahoo! France to UEJF, and awarded an additional 1 Franc against Yahoo! and Yahoo! France for expenses under Article 700 of the New Code of Civil Procedure. The French court also awarded 10,000 Francs against Yahoo! for expenses under Article 700, payable to LICRA, and 10,000 Francs each against Yahoo! and Yahoo! France under Article 700 (a total of 20,000 Francs), payable to UEJF.

    Yahoo! objected to the May 22 order. It contended, among other things, that "there was no technical solution which would enable it to comply fully with the terms of the court order." (Emphasis added.) In response, the French court obtained a written report from three experts. The report concluded that under current conditions approximately 70% of Yahoo! users operating from computer sites in France could be identified. The report specifically noted that Yahoo! already used such identification of French users to display advertising banners in French. The 70% number applied irrespective of whether a Yahoo! user sought access to an auction site, or to a site denying the existence of the Holocaust or constituting an apology for Nazism.

    With respect to auction sites, the report concluded that it would be possible to identify additional users. Two out of the three experts concluded that approximately an additional 20% of users seeking access to auction sites offering Nazi-related items for sale could be identified through an honor system in which the user would be asked to state his or her nationality. In all, the two experts estimated that almost 90% of such auction site users in France could be identified: "The combination of the two procedures, namely geographical identification of the IP address and declaration of nationality, would be likely to achieve a filtering success rate approaching 90%." The third expert expressed doubts about the number of additional users of the auction site who would respond truthfully under the honor system. He did not, however, specify an alternative number of users — say, 15% or 10% — who would respond truthfully.

    With respect to sites denying the existence of the Holocaust or constituting an apology for Nazism, the report was not able to "propose suitable and effective technical solutions" because no "grievance" against those sites had been made with "sufficient precision." In consequence, as to these non-auction sites, the report did not estimate how many Yahoo! users above the base 70% number could be identified by an honor system.

    In a second interim order, issued on November 20, 2000, the French court reaffirmed its May 22 order and directed Yahoo! [1204] to comply within three months, "subject to a penalty of 100,000 Francs per day of delay effective from the first day following expiry of the 3 month period." (The May 22 order had specified a penalty of 100,000 Euros rather than 100,000 Francs.) The court "reserve[d] the possible liquidation of the penalty" against Yahoo!. The French court's November 20 order required Yahoo! France (as distinct from Yahoo!) to display "a warning to surfers even before they have made use of the link to Yahoo.com, to be brought into effect within 2 months following notification of the present order." However, the French court found "that YAHOO FRANCE has complied in large measure with the spirit and letter of the order of 22nd May 2000[.]" (Emphasis added.)

    The November 20 order required Yahoo! to pay 10,000 Francs for a report, to be prepared in the future by one of the experts previously appointed by the court, to determine whether Yahoo! was in compliance with the court's orders. It also awarded a total of 20,000 Francs against Yahoo! for expenses under Article 700, payable to LICRA and UEJF, and an unspecified amount of costs against Yahoo!, payable to LICRA and UEJF. The court specifically stated that it was not awarding any expenses or costs against Yahoo! France (which it had found to have complied "in large measure" with its order). LICRA and UEJF used United States Marshals to serve both orders on Yahoo! in Santa Clara, California.

    Yahoo! did not pursue appeals of either interim order.

    The French court has not imposed any penalty on Yahoo! for violations of the May 22 or November 20 orders. Nor has either LICRA or UEJF returned to the French court to seek the imposition of a penalty. Both organizations affirmatively represent to us that they have no intention of doing so if Yahoo! maintains its current level of compliance. Yet neither organization is willing to ask the French court to vacate its orders. As LICRA and UEJF's counsel made clear at oral argument, "My clients will not give up the right to go to France and enforce the French judgment against Yahoo! in France if they revert to their old ways and violate French law."

    The record reveals that the French "public prosecutor" participated in the proceedings against Yahoo! and Yahoo! France in the French court, but it does not reveal whether he has the authority to seek a penalty against Yahoo! under the interim orders, either on his own or pursuant to a request by LICRA and/or UEJF. The public prosecutor was not made a party to the suit in the district court, and has made no appearance in the district court or on appeal to this court. If LICRA, UEJF, or the public prosecutor were to seek the imposition of a penalty by the French court pursuant to the interim orders, that court would have to determine the extent of Yahoo!'s violation, if any, of the orders, as well as the amount of any penalty, before an award of a penalty could be entered.

    On December 21, 2000, Yahoo! filed suit against LICRA and UEJF in federal district court, seeking a declaratory judgment that the interim orders of the French court are not recognizable or enforceable in the United States. Subject matter jurisdiction is based solely on diversity of citizenship. 28 U.S.C. § 1332(a)(2). In a thoughtful opinion, the district court concluded that it had personal jurisdiction over LICRA and UEJF. Yahoo! Inc. v. La Ligue Contre Le Racisme Et L'Antisemitisme, 145 F.Supp.2d 1168, 1180 (N.D.Cal.2001). Several months later, in another thoughtful opinion, the district court concluded that the suit was ripe, that abstention was not warranted, and that "the First Amendment precludes enforcement within the [1205] United States." Yahoo!, Inc. v. La Ligue Contre Le Racisme et L'Antisemitisme, 169 F.Supp.2d 1181, 1194 (N.D.Cal.2001).

    In early 2001, after both interim orders had been entered by the French court, and after Yahoo! had filed suit in federal district court, Yahoo! adopted a new policy prohibiting use of auctions or classified advertisements on Yahoo.com "to offer or trade in items that are associated with or could be used to promote or glorify groups that are known principally for hateful and violent positions directed at others based on race or similar factors." Yahoo! has represented, in this court and elsewhere, that its new policy has not been adopted in response to the French court's orders, but rather for independent reasons. Yahoo's new policy eliminates much of the conduct prohibited by the French orders. However, after conducting its own Internet research on yahoo.com, the district court found that even after this policy change, Yahoo! "appear[s]" not to have fully complied with the orders with respect to its auction site. 169 F.Supp.2d at 1185. For example, the district court found that Yahoo! continued to allow the sale of items such as a copy of Mein Kampf and stamps and coins from the Nazi period on which the swastika is depicted. Id. The district court also found that access was available through yahoo.com to various sites in response to searches such as "Holocaust/5 did not happen." Id.

    LICRA and UEJF timely appealed the district court's rulings on personal jurisdiction, ripeness, and abstention.

    II. Personal Jurisdiction

    The only bases for personal jurisdiction over LICRA and UEJF in the district court are the actions they have taken in connection with their French suit against Yahoo!. Those actions are sending a cease and desist letter to Yahoo! at its headquarters in Santa Clara, California; serving process on Yahoo! in Santa Clara to commence the French suit; obtaining two interim orders from the French court; and serving the two orders on Yahoo! in Santa Clara.

    Where, as here, there is no applicable federal statute governing personal jurisdiction, the district court applies the law of the state in which the district court sits. See Fed.R.Civ.P. 4(k)(1)(A); Panavision Int'l, L.P. v. Toeppen, 141 F.3d 1316, 1320 (9th Cir.1998). Because California's long-arm jurisdictional statute is coextensive with federal due process requirements, the jurisdictional analyses under state law and federal due process are the same. Id. at 1320 (citing Cal.Civ.Proc.Code § 410.10).

    In International Shoe Co. v. Washington, 326 U.S. 310, 66 S.Ct. 154, 90 L.Ed. 95 (1945), the Supreme Court held that a court may exercise personal jurisdiction over a defendant consistent with due process only if he or she has "certain minimum contacts" with the relevant forum "such that the maintenance of the suit does not offend `traditional notions of fair play and substantial justice.'" Id. at 316, 66 S.Ct. 154 (quoting Milliken v. Meyer, 311 U.S. 457, 463, 61 S.Ct. 339, 85 L.Ed. 278 (1940)). Unless a defendant's contacts with a forum are so substantial, continuous, and systematic that the defendant can be deemed to be "present" in that forum for all purposes, a forum may exercise only "specific" jurisdiction — that is, jurisdiction based on the relationship between the defendant's forum contacts and the plaintiff's claim. The parties agree that only specific jurisdiction is at issue in this case.

    In this circuit, we analyze specific jurisdiction according to a three-prong test:

    (1) The non-resident defendant must purposefully direct his activities or consummate [1206] some transaction with the forum or resident thereof; or perform some act by which he purposefully avails himself of the privilege of conducting activities in the forum, thereby invoking the benefits and protections of its laws;
    (2) the claim must be one which arises out of or relates to the defendant's forum-related activities; and
    (3) the exercise of jurisdiction must comport with fair play and substantial justice, i.e. it must be reasonable.

    Schwarzenegger v. Fred Martin Motor Co., 374 F.3d 797, 802 (9th Cir.2004) (quoting Lake v. Lake, 817 F.2d 1416, 1421 (9th Cir.1987)). The first prong is determinative in this case. We have sometimes referred to it, in shorthand fashion, as the "purposeful availment" prong. Schwarzenegger, 374 F.3d at 802. Despite its label, this prong includes both purposeful availment and purposeful direction. It may be satisfied by purposeful availment of the privilege of doing business in the forum; by purposeful direction of activities at the forum; or by some combination thereof.

    We have typically treated "purposeful availment" somewhat differently in tort and contract cases. In tort cases, we typically inquire whether a defendant "purposefully direct[s] his activities" at the forum state, applying an "effects" test that focuses on the forum in which the defendant's actions were felt, whether or not the actions themselves occurred within the forum. See Schwarzenegger, 374 F.3d at 803 (citing Calder v. Jones, 465 U.S. 783, 789-90, 104 S.Ct. 1482, 79 L.Ed.2d 804 (1984)). By contrast, in contract cases, we typically inquire whether a defendant "purposefully avails itself of the privilege of conducting activities" or "consummate[s][a] transaction" in the forum, focusing on activities such as delivering goods or executing a contract. See Schwarzenegger, 374 F.3d at 802. However, this case is neither a tort nor a contract case. Rather, it is a case in which Yahoo! argues, based on the First Amendment, that the French court's interim orders are unenforceable by an American court.

    LICRA and UEJF contend that we must base our analysis on the so-called "effects" test of Calder v. Jones, 465 U.S. 783, 104 S.Ct. 1482, 79 L.Ed.2d 804 (1984), which is normally employed in purposeful direction cases. See, e.g., CE Distrib., LLC v. New Sensor Corp., 380 F.3d 1107, 1111 (9th Cir.2004); Schwarzenegger, 374 F.3d at 803; Dole Food Co. v. Watts, 303 F.3d 1104, 1111 (9th Cir.2002). In Calder, a California-based entertainer sued the National Enquirer and various individual defendants for an allegedly defamatory article published in the Enquirer. The article had been written and edited in Florida, and the defendants had few contacts with California. The Court nonetheless upheld the exercise of personal jurisdiction in California because the defendants knew that the article would have an effect in that state. In the words of the Court, the defendants had not engaged in "mere untargeted negligence"; rather, their "intentional, and allegedly tortious, actions were expressly aimed at California." 465 U.S. at 789, 104 S.Ct. 1482.

    In this circuit, we construe Calder to impose three requirements: "the defendant allegedly [must] have (1) committed an intentional act, (2) expressly aimed at the forum state, (3) causing harm that the defendant knows is likely to be suffered in the forum state." Schwarzenegger, 374 F.3d at 803 (quoting Dole Food, 303 F.3d at 1111). In some of our cases, we have employed a slightly different formulation of the third requirement, specifying that the act must have "caused harm, the brunt of which is suffered and which the defendant knows is likely to be suffered in the forum state." Bancroft & Masters, Inc. v. [1207] Augusta Nat'l Inc., 223 F.3d 1082, 1087 (9th Cir.2000) (emphasis added). The "brunt" of the harm formulation originated in the principal opinion in Core-Vent Corp. v. Nobel Indus. AB, 11 F.3d 1482 (9th Cir.1993). That opinion required that the "brunt" of the harm be suffered in the forum state; based on that requirement, it concluded that there was no purposeful availment by the defendant. Id. at 1486. A dissenting judge would have found purposeful availment. Relying on the Supreme Court's opinion in Keeton v. Hustler Magazine, 465 U.S. 770, 104 S.Ct. 1473, 79 L.Ed.2d 790 (1984), he specifically disavowed the "brunt" of the harm formulation. Core-Vent, 11 F.3d at 1492 (Wallace, C.J., dissenting) ("[T]he Supreme Court has already rejected the proposition that the brunt of the harm must be suffered in the forum."). Without discussing the disputed "brunt" of the harm formulation, a concurring judge agreed with the dissenter that purposeful availment could be found. Id. at 1491 (Fernandez, J., concurring) ("I agree with Chief Judge Wallace that purposeful availment can be found in this case."). Later opinions picked up the "brunt" of the harm formulation of the principal opinion in Core-Vent without noting that at least one, and possibly two, of the judges on the panel disagreed with it. See, e.g., Bancroft & Masters, 223 F.3d at 1087; Panavision, 141 F.3d at 1321; Caruth v. Int'l Psychoanalytical Ass'n, 59 F.3d 126, 128 (9th Cir.1995).

    We take this opportunity to clarify our law and to state that the "brunt" of the harm need not be suffered in the forum state. If a jurisdictionally sufficient amount of harm is suffered in the forum state, it does not matter that even more harm might have been suffered in another state. In so stating we are following Keeton, decided the same day as Calder, in which the Court sustained the exercise of personal jurisdiction in New Hampshire even though "[i]t is undoubtedly true that the bulk of the harm done to petitioner occurred outside New Hampshire." 465 U.S. at 780, 104 S.Ct. 1473.

    LICRA and UEJF contend that the Calder effects test is not satisfied because, in their view, Calder requires that the actions expressly aimed at and causing harm in California be tortious or otherwise wrongful. LICRA and UEJF contend that they have done no more than vindicate their rights under French law, and that their behavior has therefore not been wrongful. They conclude that their behavior therefore does not confer personal jurisdiction in California. We agree with LICRA and UEJF that the Calder effects test is appropriately applied to the interim orders of the French court. But we disagree with them about the meaning and application of Calder.

    In any personal jurisdiction case we must evaluate all of a defendant's contacts with the forum state, whether or not those contacts involve wrongful activity by the defendant. See, e.g., Quill Corp. v. North Dakota, 504 U.S. 298, 308, 112 S.Ct. 1904, 119 L.Ed.2d 91 (1992) (upholding jurisdiction to enforce state tax on out-of-state corporation that sent catalogs and goods to forum); Burger King Corp. v. Rudzewicz, 471 U.S. 462, 479, 105 S.Ct. 2174, 85 L.Ed.2d 528 (1985) (upholding personal jurisdiction based on a course of dealing related to a franchise agreement). Many cases in which the Calder effects test is used will indeed involve wrongful conduct by the defendant. See, e.g., Calder, 465 U.S. at 790, 104 S.Ct. 1482, (allegedly defamatory publication purposefully directed at California); Bancroft & Masters, 223 F.3d at 1088 (wrongful interference with California corporation's use of domain name); Sinatra v. Nat'l Enquirer, Inc., 854 F.2d 1191, 1192 (9th Cir.1988) (unauthorized use of celebrity's name and likeness to promote Swiss clinic); Lake, [1208] 817 F.2d at 1422-23 (provision of legal services to secure allegedly improper custody order). But we do not read Calder necessarily to require in purposeful direction cases that all (or even any) jurisdictionally relevant effects have been caused by wrongful acts. We do not see how we could do so, for if an allegedly wrongful act were the basis for jurisdiction, a holding on the merits that the act was not wrongful would deprive the court of jurisdiction.

    We therefore analyze all of LICRA and UEJF's contacts with California relating to its dispute with Yahoo!, irrespective of whether they involve wrongful actions by LICRA and UEJF. There are three such contacts. The first two contacts, taken by themselves, do not provide a sufficient basis for jurisdiction. However, the third contact, considered in conjunction with the first two, does provide such a basis.

    The first contact is the cease and desist letter that LICRA sent to Yahoo!, demanding that Yahoo! alter its behavior in California to conform to what LICRA contended were the commands of French law. A cease and desist letter is not in and of itself sufficient to establish personal jurisdiction over the sender of the letter. Red Wing Shoe Co. v. Hockerson-Halberstadt, Inc., 148 F.3d 1355, 1361 (Fed.Cir.1998) ("A patentee should not subject itself to personal jurisdiction in a forum solely by informing a party who happens to be located there of suspected infringement."). There are strong policy reasons to encourage cease and desist letters. They are normally used to warn an alleged rights infringer that its conduct, if continued, will be challenged in a legal proceeding, and to facilitate resolution of a dispute without resort to litigation. If the price of sending a cease and desist letter is that the sender thereby subjects itself to jurisdiction in the forum of the alleged rights infringer, the rights holder will be strongly encouraged to file suit in its home forum without attempting first to resolve the dispute informally by means of a letter. See Red Wing Shoe, 148 F.3d at 1360-1361; Cascade Corp. v. Hiab-Foco AB, 619 F.2d 36, 38 (9th Cir.1980); Douglas Furniture Co. of Cal., Inc. v. Wood Dimensions, Inc., 963 F.Supp. 899, 903 (C.D.Cal.1997) ("If any attempt by an intellectual property holder to put an alleged wrongdoer on notice forced the property holder to submit to the jurisdiction of the alleged wrongdoer's forum, an intellectual property owner would be forced to file an action in his own jurisdiction in order to avoid the threat of being haled before a court in another, possibly distant state.").

    This is not to say that a cease and desist letter can never be the basis for personal jurisdiction. For example, in Bancroft & Masters, we upheld jurisdiction based on two letters sent by Augusta National Inc. ("ANI"), based in Georgia, contending that Bancroft & Masters, Inc. ("B & M") was improperly using its domain name. One letter was sent to Network Solutions, Inc. ("NSI") in Virginia. NSI was then the sole registrar of domain names. The other, a cease and desist letter, was sent to B & M at its corporate offices in California. B & M sued ANI in federal district court in California seeking a declaratory judgment that it had the right to the disputed domain name. On the assumption that B & M's factual allegation was true, we held that the letters were intended to trigger NSI's dispute resolution procedures, to interfere wrongfully with B & M's use of its domain name, and to misappropriate that name for ANI's own use. 223 F.3d at 1087. We therefore upheld jurisdiction under Calder based on the letters.

    LICRA's letter was not used to facilitate settlement. Although it stated that LICRA would file suit in eight days if Yahoo! had not complied with LICRA's demands, [1209] LICRA filed suit five days after the date of the letter. Nonetheless, LICRA's letter to Yahoo! was more like a normal cease and desist letter than the letters at issue in Bancroft & Masters, for it was not abusive, tortious or otherwise wrongful. Rather, it simply alerted Yahoo! to its view of French law and stated its intent to file suit in France to enforce that law against Yahoo!.

    Under these circumstances, we do not believe that LICRA's letter is a contact that would, if considered alone, justify the exercise of personal jurisdiction.

    LICRA and UEJF's second contact (or, more precisely, set of contacts) with California was service of process on Yahoo! in California. LICRA first effected service of process to commence the French suit. LICRA and UEJF later effected service of the French court's two interim orders. We do not regard the service of documents in connection with a suit brought in a foreign court as contacts that by themselves justify the exercise of personal jurisdiction over a foreign litigant in a United States court. If we were to hold that such service were a sufficient basis for jurisdiction, we would be providing a forum-choice tool by which any United States resident sued in a foreign country and served in the United States could bring suit in the United States, regardless of any other basis for jurisdiction. We are unaware of any case so holding, and Yahoo! has cited none.

    Third, and most important, LICRA and UEJF have obtained two interim orders from the French court directing Yahoo! to take actions in California, on threat of a substantial penalty. We agree with LICRA and UEJF that the French court's orders are appropriately analyzed under the Calder effects test.

    The first two requirements are that LICRA and UEJF "have `(1) committed an intentional act, [which was] (2) expressly aimed at the forum state[.]'" Schwarzenegger, 374 F.3d at 805 (quoting Dole Food, 303 F.3d at 1111). It is obvious that both requirements are satisfied. LICRA intentionally filed suit in the French court. Indeed, it had previously signaled its intent to file suit in its April 5 letter to Yahoo!. UEJF intentionally joined LICRA's suit ten days later. Further, LICRA and UEJF's suit was expressly aimed at California. The suit sought, and the French court granted, orders directing Yahoo! to perform significant acts in California. It is of course true that the effect desired by the French court would be felt in France, but that does not change the fact that significant acts were to be performed in California. The servers that support yahoo.com are located in California, and compliance with the French court's orders necessarily would require Yahoo! to make some changes to those servers. Further, to the extent that any financial penalty might be imposed pursuant to the French court's orders, the impact of that penalty would be felt by Yahoo! at its corporate headquarters in California. See Dole Food, 303 F.3d at 1113-14.

    The third requirement is that LICRA and UEJF's acts "`caus[e] harm that the defendant knows is likely to be suffered in the forum state.'" Id. This requirement is somewhat problematic, for Yahoo! has not shown or even alleged any specific way in which it has altered its behavior in response to the French court's interim orders. Yahoo! changed its policy with respect to Yahoo.com after the French court's orders were entered, but Yahoo! has consistently maintained that the change was unrelated to the orders. Therefore, even if we were persuaded that Yahoo!'s change of policy harmed it in some way, Yahoo! itself has represented that such harm was not caused by any action of LICRA or UEJF. Nor is it clear that, absent the interim orders, Yahoo! [1210] would change its policy in the future. Indeed, Yahoo! represented to us during oral argument that there is nothing that it would like to do, but is now refraining from doing, because of the interim orders.

    Yahoo!, however, points to the possibility that a substantial penalty will be assessed under the French court's November 20 interim order. It points in particular to the provision in that order specifying that the potential amount of the penalty increases by 100,000 Francs for every day that Yahoo! is in violation of the court's orders. Yahoo! represents to us that even now, after its change of policy, it is acting in plain violation of the orders. It contends that a declaratory judgment determining the enforceability by an American court of the French court's orders will allow it to determine an appropriate course of conduct with respect to the activities in which it continues to engage. The district court found that, notwithstanding its new policy,

    the Yahoo.com auction site still offers certain items for sale (such as stamps, coins, and a copy of Mein Kampf) which appear to violate the French Order. While Yahoo! has removed the Protocol of the Elders of Zion from its auction site, it has not prevented access to numerous other sites which reasonably "may be construed as constituting an apology for Nazism or a contesting of Nazi crimes."

    169 F.Supp.2d at 1185 (emphasis added).

    In both this court and the district court, LICRA and UEJF have represented that, in their view, Yahoo! is in what they call "substantial compliance" with the French court's orders. They have further represented that they will not seek enforcement of the penalty provision if Yahoo! continues its present level of compliance with the orders. However, LICRA and UEJF have stopped short of making a binding contractual commitment that they will not enforce the orders, and they have taken no action to have the orders withdrawn. As their counsel made clear at oral argument, LICRA and UEJF want to be able to return to the French court for enforcement if Yahoo! returns to its "old ways." For its part, while Yahoo! does not independently wish to take steps to comply more fully with the French court's orders, it states that it fears that it may be subject to a substantial (and increasing) fine if it does not. Yahoo! maintains that in these circumstances it has a legally cognizable interest in knowing whether the French court's orders are enforceable in this country.

    In a specific jurisdiction inquiry, we consider the extent of the defendant's contacts with the forum and the degree to which the plaintiff's suit is related to those contacts. A strong showing on one axis will permit a lesser showing on the other.

    A single forum state contact can support jurisdiction if "the cause of action . . . arise[s] out of that particular purposeful contact of the defendant with the forum state." See Lake, 817 F.2d at 1421. The case before us is the classic polar case for specific jurisdiction described in International Shoe, in which there are very few contacts but in which those few contacts are directly related to the suit. See 326 U.S. at 318, 66 S.Ct. 154 ("[S]ome single or occasional acts of the corporate agent in a state . . . because of their nature and quality and the circumstances of their commission, may be deemed sufficient to render the corporation liable to suit."). All of the contacts with the forum state in this case are either the interim orders themselves or contacts directly related to those orders.

    LICRA and UEJF have not sought enforcement of the French court's orders in this country, and they have stated that [1211] they will not seek enforcement or penalties so long as Yahoo! continues its current course of conduct. However, LICRA and UEJF have not sought to vacate the French court's orders, and it is at least possible that they might later seek enforcement based on a continuation of Yahoo!'s current conduct. Or more likely, they might seek enforcement if Yahoo! changes it conduct in the future. But even if LICRA and UEJF seek enforcement at some time in the future, and even if the French court finds a violation that warrants the imposition of a penalty, enforcement of that penalty is extremely unlikely in the United States. Enforcement is unlikely not because of the First Amendment, but rather because of the general principle of comity under which American courts do not enforce monetary fines or penalties awarded by foreign courts.

    Finally, Yahoo! contends that it has a legally protected interest, based on the First Amendment, in continuing its current policy with respect to Nazi memorabilia and Holocaust-related anti-semitic materials. Until that contention is endorsed by the judgment of an American court, it is only a contention. But even if the French court's orders are not enforced against Yahoo!, the very existence of those orders may be thought to cast a shadow on the legality of Yahoo!'s current policy.

    It is a close question whether LICRA and UEJF are subject to personal jurisdiction in California in this suit. But considering the direct relationship between LICRA and UEJF's contacts with the forum and the substance of the suit brought by Yahoo!, as well as the impact and potential impact of the French court's orders on Yahoo!, we hold that there is personal jurisdiction.

    III. Ripeness

    Because we conclude that the exercise of personal jurisdiction over LICRA and UEJF is proper, we turn to the question of ripeness. Ripeness doctrine is "`drawn both from Article III limitations on judicial power and from prudential reasons for refusing to exercise jurisdiction.'" Nat'l Park Hospitality Ass'n v. DOI, 538 U.S. 803, 808, 123 S.Ct. 2026, 155 L.Ed.2d 1017 (2003) (quoting Reno v. Catholic Social Servs., Inc., 509 U.S. 43, 57 n. 18, 113 S.Ct. 2485, 125 L.Ed.2d 38 (1993)). Even where jurisdiction is present in the Article III sense, courts are obliged to dismiss a case when considerations of prudential ripeness are not satisfied. Socialist Labor Party v. Gilligan, 406 U.S. 583, 588, 92 S.Ct. 1716, 32 L.Ed.2d 317 (1972) ("Problems of prematurity and abstractness may well present `insuperable obstacles' to the exercise of the Court's jurisdiction, even though that jurisdiction is technically present.") (citing Rescue Army v. Municipal Court, 331 U.S. 549, 574, 67 S.Ct. 1409, 91 L.Ed. 1666 (1947)).

    The existence of Article III subject matter jurisdiction is, like personal jurisdiction, a close question, but we agree with the district court that the effect of the French court's orders on Yahoo! is sufficient to create a case or controversy within the meaning of Article III. See 169 F.Supp.2d at 1187-91. However, we disagree with the district court's conclusion that there is prudential ripeness. In its current form, this case presents the sort of "[p]roblems of prematurity and abstractness" that counsel against reaching the First Amendment question that Yahoo! insists is presented by this case. See Socialist Labor Party, 406 U.S. at 588, 92 S.Ct. 1716.

    In determining whether a case satisfies prudential requirements for ripeness, we consider two factors: "the fitness of the issues for judicial decision," and "the hardship to the parties of withholding court consideration." Abbott Labs. v. Gardner, [1212] 387 U.S. 136, 149, 87 S.Ct. 1507, 18 L.Ed.2d 681 (1967); Pac. Gas & Elec. Co. v. State Energy Res. Conservation & Dev. Comm'n, 461 U.S. 190, 201, 103 S.Ct. 1713, 75 L.Ed.2d 752 (1983) (quoting Abbott Labs.). We address these two factors in turn.

    A. Fitness of the Issue for Judicial Decision

    1. The Substantive Legal Question at Issue

    Whether a dispute is sufficiently ripe to be fit for judicial decision depends not only on the state of the factual record. It depends also on the substantive legal question to be decided. If the legal question is straightforward, relatively little factual development may be necessary. As we wrote in San Diego County Gun Rights Comm. v. Reno, 98 F.3d 1121, 1132 (9th Cir.1996), "[P]ure legal questions that require little factual development are more likely to be ripe." By contrast, if the legal question depends on numerous factors for its resolution, extensive factual development may be necessary.

    A noted example is Adler v. Bd. of Educ., 342 U.S. 485, 72 S.Ct. 380, 96 L.Ed. 517 (1952), in which Justice Frankfurter disagreed with the other justices about the precise legal question presented, and, as a consequence, disagreed about ripeness. Because the legal question, as Justice Frankfurter understood it, required fine-grained and subtle judgments based on extensive factual development, he concluded that the suit was not ripe. Id. at 506-07, 72 S.Ct. 380 (Frankfurter, J., dissenting). In the view of the other justices, however, the legal question was different. In their view, this different legal question was relatively simple, requiring little factual development. For them (and for this different legal question), the suit was ripe. Id. at 492-93, 72 S.Ct. 380 (maj.op.); 342 U.S. at 508-09, 72 S.Ct. 380 (Douglas, J., dissenting). See Fritz W. Scharpf, Judicial Review and the Political Question: A Functional Analysis, 75 Yale L.J. 517, 532 (1966). See also United Public Workers v. Mitchell, 330 U.S. 75, 90-91, 67 S.Ct. 556, 91 L.Ed. 754 (1947) (dismissing suit as unripe); id. at 109, 67 S.Ct. 556 (Black, J., dissenting); id. at 116-17, 67 S.Ct. 556 (Douglas, J., dissenting).

    It is thus important to a ripeness analysis that we specify the precise legal question to be answered. Depending on the legal question, the case may be ripe or unripe. If we ask the wrong legal question, we risk getting the wrong answer to the ripeness question. The legal question presented by this case is whether the two interim orders of the French court are enforceable in this country. These orders, by their explicit terms, require only that Yahoo! restrict access by Internet users located in France. The orders say nothing whatsoever about restricting access by Internet users in the United States. We are asked to decide whether enforcement of these interim orders would be "repugnant" to California public policy.

    There is currently no federal statute governing recognition of foreign judgments in the federal courts. See American Law Institute, Recognition and Enforcement of Foreign Judgments: Analysis and Proposed Federal Statute (April 11, 2005) (proposed final draft). The federal full faith and credit statute, 28 U.S.C. § 1738, governs only judgments rendered by courts of states within the United States. In diversity cases, enforceability of judgments of courts of other countries is generally governed by the law of the state in which enforcement is sought. Bank of Montreal v. Kough, 612 F.2d 467, 469-70 (9th Cir.1980); see also Southwest Livestock & Trucking Co. v. Ramon, 169 F.3d 317, 320 (5th Cir.1999); Choi v. Kim, 50 F.3d 244, 248 (3d Cir.1995); S.A. Andes v. [1213] Versant Corp., 878 F.2d 147, 150 (4th Cir.1989); Ingersoll Milling Mach. Co. v. Granger, 833 F.2d 680, 686 (7th Cir.1987); Branca v. Security Benefit Life Ins. Co., 773 F.2d 1158, 1161 (11th Cir.1985). This is a diversity suit, brought by Yahoo! in federal district court in California.

    In a typical enforcement case, the party in whose favor the foreign judgment was granted comes to an American court affirmatively seeking enforcement. The standard rule in such a case is that the federal court sitting in diversity applies the law of the state in which it sits. However, this is not the typical case, for the successful plaintiffs in the French court do not seek enforcement. Rather, Yahoo!, the unsuccessful defendant in France, seeks a declaratory judgment that the French court's interim orders are unenforceable anywhere in this country.

    Insofar as the issue is whether the French court's orders are enforceable in California, it is clear that California law governs. However, it is less clear whose law governs when enforceability in other states is at issue. This is a potentially difficult choice-of-law question, but we do not need to answer it in order to decide ripeness. First, the central issue is enforceability in California. Therefore, if the suit is unripe under California law, we should not decide the case, irrespective of whether it might be ripe under the law of some other state. To do otherwise would be to allow the tail to wag the dog. Second, in any event, the law of virtually all other states appears to be similar, or even identical, to California law. We may thus safely proceed with our ripeness analysis based on the California law of enforceability.

    California, along with many other states, has adopted the Uniform Foreign Money-Judgments Recognition Act ("Uniform Act" or "Act"). Cal.Civ.Proc.Code §§ 1713-1713.8. The relevant standard for enforceability under the Act is whether "the cause of action or defense on which the judgment is based is repugnant to the public policy of this state." Id. § 1713.4(b)(3) (emphasis added). However, the Act is not directly applicable to this case, for it does not authorize enforcement of injunctions. See id. § 1713.1(2) ("`Foreign judgment' means any judgment of a foreign state granting or denying recovery of a sum of money, other than . . . a fine or other penalty[.]") But neither does the Uniform Act prevent enforcement of injunctions, for its savings clause specifies that the Act does not foreclose enforcement of foreign judgments "in situations not covered by [the Act]." Id. § 1713.7.

    Because the Uniform Act does not cover injunctions, we look to general principles of comity followed by the California courts. We may appropriately consult the Restatement (Third) of the Foreign Relations Law of the United States ("Third Restatement" or "Restatement"), given that California courts frequently cite the Restatement, as well as earlier Restatements, as sources of law. See, e.g., Renoir v. Redstar Corp., 123 Cal.App.4th 1145, 1150, 20 Cal.Rptr.3d 603 (2004) (Third Restatement); American Home Assurance Co. v. Sociét é Commerciale Toutélectric, 104 Cal.App.4th 406, 424, 128 Cal.Rptr.2d 430 (2003) (same); Smith v. Hopland Band of Pomo Indians, 95 Cal.App.4th 1, 10, 115 Cal.Rptr.2d 455 (2002) (same); Pecaflor Construction, Inc. v. Landes, 198 Cal.App.3d 342, 349, 243 Cal.Rptr. 605 (1988) (Second Restatement). The general principle of enforceability under the Third Restatement is the same as under California's Uniform Act. That is, an American court will not enforce a judgment if "the cause of action on which the judgment was based, or the judgment itself, is repugnant to the public policy of the United States or of the State where recognition is sought[.]" Restatement § 482(2)(d) (emphasis added); see also Restatement [1214] (Second) of the Conflict of Laws § 117 cmt. c (1971) ("[E]nforcement will usually be accorded [a] judgment [of a foreign court] except in situations where the original claim is repugnant to fundamental notions of what is decent and just in the State where enforcement is sought.") (emphasis added).

    There is very little case law in California dealing with enforceability of foreign country injunctions under general principles of comity, but that law is consistent with the repugnancy standard of the Restatement. We have found only one case in which a California court has ruled on the enforceability of an injunction granted in another country. In In re Stephanie M., 7 Cal.4th 295, 27 Cal.Rptr.2d 595, 867 P.2d 706 (1994), a Mexican court had entered a guardianship decree purporting to authorize the named guardian to take immediate custody of a child and to return her to Mexico. The California Supreme Court recognized that an injunction could be enforced by the California courts as a matter of comity, but it declined to order enforcement in this particular case because the Mexican decree conflicted with California public policy. Id. at 314, 27 Cal.Rptr.2d 595, 867 P.2d 706.

    California courts have also relied on public policy in the analogous context of injunctions entered by other American courts. In Smith v. Superior Court, 41 Cal.App.4th 1014, 49 Cal.Rptr.2d 20 (1996), plaintiff Smith had been badly injured, and her husband and two children killed, when their General Motors ("GM") vehicle burst into flames after a collision. Smith brought a product liability suit in California against GM. Elwell had been an engineer for GM for many years and had extensive knowledge about the design of GM vehicles. An earlier wrongful termination suit between Elwell and GM in Michigan had been dismissed after the parties stipulated to a permanent injunction forbidding Elwell from testifying in any suit about GM vehicles. Smith sought to call Elwell as an expert witness in her California suit. The California Court of Appeal declined to enforce the Michigan injunction on the ground that it "blatantly and irreconcilably conflicts with our fundamental public policy against the suppression of evidence." Id. at 1025, 49 Cal.Rptr.2d 20; see also Baker v. General Motors Corp., 522 U.S. 222, 118 S.Ct. 657, 139 L.Ed.2d 580 (1998) (Missouri state court not required by 28 U.S.C. § 1738 to enforce the same Michigan injunction against Elwell when such enforcement would violate Missouri public policy).

    The repugnancy standard is also generally followed in states other than California. See, e.g., Hilkmann v. Hilkmann, 579 Pa. 563, 575, 858 A.2d 58 (2004) (observing that the Restatement's repugnancy standard has been incorporated into Pennsylvania common law); Alberta Sec. Comm'n v. Ryckman, 200 Ariz. 540, 549, 30 P.3d 121 (2001) (stating that foreign judgments are not enforceable under Arizona common law if they are repugnant to public policy); Panama Processes, S.A. v. Cities Serv. Co., 796 P.2d 276, 283 (Okla.1990) (declaring that a judgment must not be enforced if repugnant to public policy); Greschler v. Greschler, 51 N.Y.2d 368, 377, 434 N.Y.S.2d 194, 414 N.E.2d 694 (1980) ("[T]he public policy exception to the doctrine of comity is usually invoked . . . when the original claim is repugnant to fundamental notions of what is decent and just in the State where enforcement is sought.") (internal quotation omitted). Further, federal courts sometimes cite general principles of comity without reference to particular state laws. See, e.g., Jaffe v. Accredited Sur. & Cas. Co., 294 F.3d 584, 593 (4th Cir.2002) (declaring that a judgment will not be enforced if repugnant to public policy); In re Schimmelpenninck, 183 F.3d 347, 365 (5th Cir.1999) (to [1215] be enforceable, "foreign laws need not be identical to . . . the laws of the United States; they merely must not be repugnant to our laws and policies"); Turner Entertainment Co. v. Degeto Film GmbH, 25 F.3d 1512, 1519 (11th Cir.1994) ("General comity concerns include . . . whether the foreign judgment is prejudicial, in the sense of violating American public policy because it is repugnant to fundamental principles of what is decent and just."); see also Hilton v. Guyot, 159 U.S. 113, 16 S.Ct. 139, 40 L.Ed. 95 (1895) (discussing principles of comity governing enforcement of foreign judgments).

    Under the repugnancy standard, American courts sometimes enforce judgments that conflict with American public policy or are based on foreign law that differs substantially from American state or federal law. See, e.g., In re Hashim, 213 F.3d 1169, 1172 (9th Cir.2000) (reversing bankruptcy court's refusal to enforce English court's award of $10 million in costs against debtors whose assets had been frozen by Saddam Hussein); Milhoux v. Linder, 902 P.2d 856, 861-62 (Colo.Ct.App.1995) (affirming recognition of Belgian judgment as a matter of comity, even though it was based on a 30-year Belgian statute of limitations). Inconsistency with American law is not necessarily enough to prevent recognition and enforcement of a foreign judgment in the United States. The foreign judgment must be, in addition, repugnant to public policy.

    2. Fitness of the Question for Judicial Decision

    With the suit in its current state, it is difficult to know whether enforcement of the French court's interim orders would be repugnant to California public policy. The first difficulty is evident. As indicated by the label "interim," the French court contemplated that it might enter later orders. We cannot know whether it might modify these "interim" orders before any attempt is made to enforce them in the United States.

    A second, more important, difficulty is that we do not know whether the French court would hold that Yahoo! is now violating its two interim orders. After the French court entered the orders, Yahoo! voluntarily changed its policy to comply with them, at least to some extent. There is some reason to believe that the French court will not insist on full and literal compliance with its interim orders, and that Yahoo!'s changed policy may amount to sufficient compliance.

    In its interim second order, entered on November 20, the French court found that Yahoo! France had "complied in large measure with the spirit and letter" of its May 22 order. (Emphasis added.) Based on that level of compliance, the French court was satisfied. It declined to enter any further orders against Yahoo! France. It also declined to award any expenses or costs against Yahoo! France, even though in that same order it awarded expenses and costs against Yahoo!. We thus know from this second order that compliance "in large measure" by Yahoo! is very likely to be satisfactory to the French court, just as compliance "in large measure" by Yahoo! France was satisfactory.

    LICRA and UEJF insist that Yahoo! has now, in their words, "substantially complied" with the French court's orders. We take this to be a statement that, in their view, Yahoo! has complied "in large measure" with the orders. For its part, however, Yahoo! insists that it continues to be in serious violation of the orders. The district court did not hold that Yahoo! is in violation, substantial or otherwise, of the French court's orders. It wrote only that Yahoo! does not "appear" to be in full compliance with the French court's order with respect to its auction site, and that [1216] various anti-semitic sites continue to be accessible through yahoo.com. 169 F.Supp.2d at 1185. There is only one court that can authoritatively tell us whether Yahoo! has now complied "in large measure" with the French court's interim orders. That is, of course, the French court.

    To the extent that we are uncertain about whether Yahoo! has complied "in large measure" with the French court's orders, the responsibility for that uncertainty can be laid at Yahoo!'s door. In its November 20 interim order, the French court ordered the appointment of one of the experts who had previously reported on the technical feasibility of restricting access by French users to Yahoo.com. Under the November 20 order, Yahoo! was required to pay the expert, who would be charged "to undertake an assignment to prepare a consultancy report on the conditions of fulfillment of the terms of the aforementioned order." Yahoo! has placed nothing in the record to tell us whether Yahoo! has paid the expert; whether the expert has prepared a report for the French court; and, if a report has been prepared, what it says. There is also nothing in the record to indicate what other steps, if any, Yahoo! has taken to obtain an indication from the French court whether it believes that Yahoo! is in compliance, "in large measure" or otherwise, with the terms of its interim orders. All we know for certain is that Yahoo! abandoned its appeal of the May 22 interim order and declined to appeal the November 20 interim order, and that on December 21, a month and a day after entry of the second interim order, it came home to file suit in the Northern District of California.

    A third difficulty is related to the second. Because we do not know whether Yahoo! has complied "in large measure" with the French court's orders, we cannot know what effect, if any, compliance with the French court's orders would have on Yahoo!'s protected speech-related activities. We emphasize that the French court's orders require, by their terms, only a limitation on access to anti-semitic materials by users located in France. The orders do not by their terms limit access by users outside France in any way. Yahoo! contended in the French court that it was technically too difficult to distinguish between users inside and outside France. As described above, the French court commissioned a report by three experts to determine if Yahoo!'s contention were true. The experts disagreed with Yahoo!, concluding that Yahoo! is readily able to distinguish between most users inside and outside France.

    With respect to users seeking access to forbidden auction sites, two out of the three experts concluded that Yahoo! could identify almost 90% of its users located in France. The third expert did not dispute that 70% of such auction site users could be identified, but expressed doubt about how many additional such users could be identified. With respect to users seeking access to sites of Holocaust deniers and Nazi apologists, the experts declined to propose any solution by which a greater number than 70% of users located in France could be identified.

    In its briefing to this court, Yahoo! contends that restricting access by French Internet users in a manner sufficient to satisfy the French court would in some unspecified fashion require Yahoo! simultaneously to restrict access by Internet users in the United States. This may or may not be true. It is almost certainly not true if Yahoo! is now complying "in large measure" with the French court's orders, for in that event the French court will almost certainly hold that no further compliance is necessary. Even if the measures [1217] Yahoo! has already taken restrict access by American Internet users to antisemitic materials, this has no bearing on Yahoo!'s First Amendment argument. By its own admission, Yahoo! has taken these measures entirely of its own volition, for reasons entirely independent of the French court's orders.

    However, it is possible, as Yahoo! contends, that it has not complied "in large measure" with the French court orders, and that the French court would require further compliance. It is also possible, as Yahoo! contends, that further compliance might have the necessary consequence of requiring Yahoo! to restrict access by American Internet users. But Yahoo! has been vague in telling us in what ways, and for what reasons, it believes further compliance might have that consequence. One possible reason for Yahoo!'s vagueness might be that its contention is ill-founded, and that a detailed explanation would reveal that fact. We are not now in a position to judge this. Another, more important, reason — not merely a possible reason — for its vagueness is that Yahoo! has no way of knowing what further compliance might be required by the French court. Until it knows what further compliance (if any) the French court will require, Yahoo! simply cannot know what effect (if any) further compliance might have on access by American users.

    The possible — but at this point highly speculative — impact of further compliance with the French court's orders on access by American users would be highly relevant to the question whether enforcement of the orders would be repugnant to California public policy. But we cannot get to that question without knowing whether the French court would find that Yahoo! has already complied "in large measure," for only on a finding of current noncompliance would the issue of further compliance, and possible impact on American users, arise.

    Without a finding that further compliance with the French court's orders would necessarily result in restrictions on access by users in the United States, the only question in this case is whether California public policy and the First Amendment require unrestricted access by Internet users in France. In other words, the only question would involve a determination whether the First Amendment has extraterritorial application. The extent of First Amendment protection of speech accessible solely by those outside the United States is a difficult and, to some degree, unresolved issue. Compare, e.g., Desai v. Hersh, 719 F.Supp. 670, 676 (N.D.Ill.1989) ("[F]or purposes of suits brought in the United States courts, first amendment protections do not apply to all extraterritorial publications by persons under the protections of the Constitution."), and Laker Airways Ltd. v. Pan American Airways, Inc., 604 F.Supp. 280, 287 (D.D.C.1984) ("It is less clear, however, whether even American citizens are protected specifically by the First Amendment with respect to their activities abroad[.]"), with Bullfrog Films, Inc. v. Wick, 646 F.Supp. 492, 502 (C.D.Cal.1986) ("[T]here can be no question that, in the absence of some overriding governmental interest such as national security, the First Amendment protects communications with foreign audiences to the same extent as communications within our borders."), aff'd, 847 F.2d 502 (9th Cir.1988).

    We are thus uncertain about whether, or in what form, a First Amendment question might be presented to us. If the French court were to hold that Yahoo!'s voluntary change of policy has already brought it into compliance with its interim orders "in large measure," no First Amendment question would be presented at all. Further, if the French court were to require additional compliance with respect to users in France, but that additional compliance [1218] would not require any restriction on access by users in the United States, Yahoo! would only be asserting a right to extraterritorial application of the First Amendment. Finally, if the French court were to require additional compliance with respect to users in France, and that additional compliance would have the necessary consequence of restricting access by users in the United States, Yahoo! would have both a domestic and an extraterritorial First Amendment argument. The legal analysis of these different questions is different, and the answers are likely to be different as well.

    B. Hardship to the Parties

    We next consider "the hardship to the parties of withholding court consideration." Abbott Labs., 387 U.S. at 149, 87 S.Ct. 1507. As discussed above, we believe that Yahoo! has suffered sufficient harm to justify (though not by a wide margin) the exercise of personal jurisdiction over LICRA and UEJF. The threshold requirement for hardship for purposes of personal jurisdiction, however, is not necessarily the same as the threshold for purposes of prudential ripeness. Particularly where, as here, there are substantial uncertainties bearing on the legal analysis to be performed, there is a high threshold requirement for hardship.

    Yahoo! contends that it will suffer real hardship if we do not decide its suit at this time. Yahoo! makes essentially two arguments. First, it argues that the potential monetary penalty under the French court's orders is mounting every day, and that the enforcement of a penalty against it here could be extremely onerous. Second, it argues that the French court's orders substantially limit speech that is protected by the First Amendment. We take these arguments in turn.

    1. Enforceability of the Monetary Penalty

    Yahoo! contends that the threat of a monetary penalty hangs like the sword of Damocles. However, it is exceedingly unlikely that the sword will ever fall. We may say with some confidence that, for reasons entirely independent of the First Amendment, the French court's orders are not likely to result in the enforcement of a monetary penalty in the United States. The French court's orders threaten monetary sanctions against Yahoo!, which that court explicitly labels "penalties." In order to obtain an award of a penalty from the French court, LICRA and UEJF would have to return to the French court, to explain to the French court why they believe Yahoo! has violated its interim orders, and to persuade the French court that Yahoo!'s violation merits the imposition of a penalty. In the nearly five years since the entry of the French court's second interim order and Yahoo!'s change of policy, LICRA and UEJF have taken none of these steps. Further, LICRA and UEJF have represented that they have no intention of seeking a monetary penalty by the French court so long as Yahoo! does not revert to its "old ways."

    More important, even if the French court were to impose a monetary penalty against Yahoo!, it is exceedingly unlikely that any court in California — or indeed elsewhere in the United States — would enforce it. California's Uniform Act does not authorize enforcement of "fines or other penalties." Cal.Civ.Proc.Code § 1713.1(2). The Act includes a savings clause, see Cal.Civ.Proc.Code § 1713.7, but the fine is equally unenforceable under California common law doctrine.

    California courts follow the generally-observed rule that, "`[u]nless required to do so by treaty, no state [i.e., country] enforces the penal judgments of other states [i.e., countries].'" In re Manuel P., [1219] 215 Cal.App.3d 48, 81, 263 Cal.Rptr. 447 (1989) (Wiener, J., dissenting) (quoting Restatement § 483 cmt. 3); see also In re Marriage of Gray, 204 Cal.App.3d 1239, 1253, 251 Cal.Rptr. 846 (1988). This is consistent with the Restatement's declaration that "[c]ourts in the United States are not required . . . to enforce judgments [from foreign countries] for the collection of . . . fines[ ] or other penalties." Restatement § 483; see also 30 Am.Jur.2d Execution and Enforcement of Judgments § 846 (2004) ("Courts in the United States will not recognize or enforce a penal judgment rendered in another nation."). A number of states have adopted an identical version of California's Uniform Act, see Enforcing Foreign Judgments in the United States and United States Judgments Abroad 28-32 (Ronald A. Brand ed., 1992), and the common law rule against the enforcement of penal judgments is venerable and widely-recognized. See Huntington v. Attrill, 146 U.S. 657, 673-74, 13 S.Ct. 224, 36 L.Ed. 1123 (1892); see also 18 James Wm. Moore et al., Moore's Federal Practice § 130.05 (2002).

    Penal judgments are those intended "`to punish an offense against the public justice of the [foreign] state[.]'" Chavarria v. Superior Court, 40 Cal.App.3d 1073, 1077, 115 Cal.Rptr. 549 (1974) (quoting Huntington, 146 U.S. at 673-74, 13 S.Ct. 224). The test to determine a judgment's nature

    is not by what name the statute [on which the judgment is based] is called by the legislature or the courts of the State in which it was passed, but whether it appears to the tribunal which is called upon to enforce it to be, in its essential character and effect, a punishment of an offense against the public, or a grant of a civil right to a private person.

    Huntington, 146 U.S. at 682, 13 S.Ct. 224.

    There are a number of indications that the French judgments are penal in nature. First, the word used by the French court ("astreinte") is consistently translated as "penalty" in the record in this case. For example, the May 22 order provides that Yahoo! and Yahoo! France are "subject to a penalty of 100,000 Euros per day of delay and per confirmed violation[.]" The November 20 order provides that Yahoo! is "subject to a penalty of 100,000 Francs per day of delay[.]"

    Second, the French court held that Yahoo! was violating Section R645-1 of the French Penal Code, which declares it a "crime" to exhibit or display Nazi emblems, and which prescribes a set of "criminal penalties," including fines. Fr. C. Pén. § R645-1, translation available at http://www.lex2k.org/ yahoo/art645.pdf. The monetary penalties against Yahoo! do not lose their character as "penalties" simply because they were obtained in a civil action. See Wisconsin v. Pelican Ins. Co., 127 U.S. 265, 299, 8 S.Ct. 1370, 32 L.Ed. 239 (1888). Nor do they lose their character because private litigants initiated the action. A civil remedy is penal, as the term is understood in private international law, if it awards a penalty "to a member of the public, suing in the interest of the whole community to redress a public wrong." Weiss v. Glemp, 792 F.Supp. 215, 227 (S.D.N.Y.1992); see also Loucks v. Standard Oil Co., 224 N.Y. 99, 101, 120 N.E. 198 (1918) (Cardozo, J.). In short, the label "civil" does not strip a remedy of its penal nature. Thus, for example, an American court is not required to enforce an order of contempt or an award of punitive damages in a civil action. Cf. Frank v. Reese, 594 S.W.2d 119, 121 (Tex.Civ.App.1979) ("Other jurisdictions are reluctant to give full faith and credit to an order for contempt due to its punitive nature [.]"); Republic of Philippines v. Westinghouse Elec. Corp., 821 F.Supp. 292, 295 (D.N.J.1993) (refusing to enforce Philippine law providing for punitive damages); see also [1220] Third Restatement § 483 cmt. b ("Some states consider judgments penal for purposes of non-recognition if multiple, punitive, or exemplary damages are awarded, even when no governmental agency is a party.").

    Third, the penalties the French court imposed on Yahoo! are primarily designed to deter Yahoo! from creating, in the words of the November 20 order, "a threat to internal public order." The penalties are payable to the government and not designed to compensate the French student groups for losses suffered. See Farmers & Merchants Trust Co. v. Madeira, 261 Cal.App.2d 503, 510, 68 Cal.Rptr. 184 (1968) (suggesting that a judgment is penal if it is designed to punish a defendant "for an offense committed against the public justice" of the jurisdiction). Judgments designed to deter conduct that constitutes a threat to the public order are typically penal in nature. Cf. Kennedy v. Mendoza-Martinez, 372 U.S. 144, 168, 83 S.Ct. 554, 9 L.Ed.2d 644 (1963).

    The French court awarded nominal damages of one Franc to LICRA and UEJF in its first (but not its second) order. Balanced against the far more substantial penalties payable to the government (up to 100,000 Francs per day under the second order), this award of one Franc cannot render the orders primarily remedial rather than punitive in nature. See Ducharme v. Hunnewell, 411 Mass. 711, 714, 585 N.E.2d 321 (1992) (determining that whether a judgment requires enforcement "depends on whether its purpose is remedial in nature, affording a private remedy to an injured person, or penal in nature, punishing an offense against the public justice"). Even the "restitution" the court ordered — the printing of its judgment in publications of UEJF's and LICRA's choosing — benefits the general public and does not specifically compensate the two student groups for a particular injury.

    2. First Amendment

    Yahoo! argues that any restriction on speech and speech-related activities resulting from the French court's orders is a substantial harm under the First Amendment. We are acutely aware that this case implicates the First Amendment, and we are particularly sensitive to the harm that may result from chilling effects on protected speech or expressive conduct. In this case, however, the harm to First Amendment interests — if such harm exists at all — may be nowhere near as great as Yahoo! would have us believe. Yahoo! has taken pains to tell us that its adoption of a new hate speech policy after the entry of the French court's interim orders was motivated by considerations independent of those orders. Further, Yahoo! refuses to point to anything that it is now not doing but would do if permitted by the orders. In other words, Yahoo! itself has told us that there is no First Amendment violation with respect either to its previous (but now abandoned) speech-related activities, or to its future (but not currently engaged in) speech-related activities. Any restraint on such activities is entirely voluntary and self-imposed.

    The only potential First Amendment violation comes from the restriction imposed by the interim orders — if indeed they impose any restrictions — on the speech-related activities in which Yahoo! is now engaged, and which might be restricted if further compliance with the French court's orders is required. For example, Yahoo! continues to allow auctions of copies of Mein Kampf, and it maintains that the French court's orders prohibit it from doing so. The French court might find that Yahoo! has not yet complied "in large measure" with its orders, and that Yahoo! is [1221] prohibited by its orders from allowing auctions of copies of Mein Kampf.

    Even if the French court took this step, Yahoo!'s claim to First Amendment protection would be limited. We emphasize that the French court's interim orders do not by their terms require Yahoo! to restrict access by Internet users in the United States. They only require it to restrict access by users located in France. That is, with respect to the Mein Kampf example, the French court's orders — even if further compliance is required — would by their terms only prohibit Yahoo! from allowing auctions of copies of Mein Kampf to users in France.

    The core of Yahoo!'s hardship argument may thus be that it has a First Amendment interest in allowing access by users in France. Yet under French criminal law, Internet service providers are forbidden to permit French users to have access to the materials specified in the French court's orders. French users, for their part, are criminally forbidden to obtain such access. In other words, as to the French users, Yahoo! is necessarily arguing that it has a First Amendment right to violate French criminal law and to facilitate the violation of French criminal law by others. As we indicated above, the extent — indeed the very existence — of such an extraterritorial right under the First Amendment is uncertain.

    3. Summary

    In sum, it is extremely unlikely that any penalty, if assessed, could ever be enforced against Yahoo! in the United States. Further, First Amendment harm may not exist at all, given the possibility that Yahoo! has now "in large measure" complied with the French court's orders through its voluntary actions, unrelated to the orders. Alternatively, if Yahoo! has not "in large measure" complied with the orders, its violation lies in the fact that it has insufficiently restricted access to anti-semitic materials by Internet users located in France. There is some possibility that in further restricting access to these French users, Yahoo! might have to restrict access by American users. But this possibility is, at this point, highly speculative. This level of harm is not sufficient to overcome the factual uncertainty bearing on the legal question presented and thereby to render this suit ripe.

    C. The Dissent Addressed to Ripeness

    The dissent addressed to the question of ripeness makes two principal contentions. First, it contends that the French court's interim orders are unconstitutional on their face, and that further factual development is therefore not needed. Second, it contends that if any further factual development is necessary, we should remand to the district court for that purpose. We take these contentions in turn.

    1. Unconstitutionality of the French Court's Orders

    The dissent repeatedly states that the French court's interim orders are facially unconstitutional. It writes, "The French orders on their face . . . violate the First Amendment and are plainly contrary to one of America's, and by extension California's, most cherished public policies." (Dissent at 1239.) It later refers to the French court's orders as "foreign court orders that so obviously violate the First Amendment." (Id. at 1239-40.) It writes further, "[T]he absence of a discernible line between the permitted and the unpermitted . . . makes the orders facially unconstitutional." (Id. at 1244)

    The dissent is able to conclude that the French court's interim orders are facially unconstitutional only by ignoring what they say. The dissent appears to assume that the orders, on their face, require Yahoo! [1222] to block access by United States users. It writes, "[T]he question we face in this federal lawsuit is whether our own country's fundamental constitutional guarantee of freedom of speech protects Yahoo! (and, derivatively, at least its users in the United States) against some or all of the restraints the French defendants have deliberately imposed upon it within the United States." (Id. at 1234-1235) (emphasis in original). Further, "Yahoo! confront[s] the dilemma of whether or not to stand by its United States constitutional rights or constrain its speech and that of its user[.]" (Id. at 1238.) "Legions of cases permit First Amendment challenges to governmental actions or decrees that on their face are vague, overbroad and threaten to chill protected speech. Indeed, the sweeping injunction here presents just such a paradigmatic case." (Id. at 1238.) Still further, "Under the principles articulated today, a foreign party can use a foreign court decree to censor free speech here in the United States[.]" (Id. at 1240.)

    If it were true that the French court's orders by their terms require Yahoo! to block access by users in the United States, this would be a different and much easier case. In that event, we would be inclined to agree with the dissent. See, e.g., Sarl Louis Feraud Int'l v. Viewfinder Inc., No. 04 Civ. 9760, 2005 WL 2420525, 2005 U.S. Dist. LEXIS 22242 (S.D.N.Y. Sept. 29, 2005) (holding unenforceable as contrary to the First Amendment a French damage judgment based on photographs posted on the Internet freely accessible to American viewers). But this is not the case. The French court's orders, by their terms, require only that Yahoo! restrict access by users in France. The boundary line between what is permitted and not permitted is somewhat uncertain for users in France. But there is no uncertainty about whether the orders apply to access by users in the United States. They do not. They say nothing whatsoever about restricting access by users in the United States.

    The dissent's conclusion that the French court's orders are unconstitutional may be based in part on an assumption that a necessary consequence of compliance with the French court's orders will be restricted access by users in the United States. But if this is the basis for the dissent's conclusion, it could hardly say that the orders are unconstitutional "on their face." Whether restricted access by users in the United States is a necessary consequence of the French court's orders is a factual question that we cannot answer on the current record.

    If the only consequence of compliance with the French court's orders is to restrict access by Internet users in France, Yahoo!'s only argument is that the First Amendment has extraterritorial effect. The dissent fails to acknowledge that this is inescapably a central part of Yahoo!'s argument, let alone acknowledge that it may be Yahoo!'s only argument.

    2. Remand to the District Court

    As a fallback position, the dissent contends that we should remand to the district court for a determination whether a necessary consequence of compliance with the French court's orders would be restriction on access by users in the United States. This fallback contention is, of course, in tension with the dissent's conclusion that the French court's orders are unconstitutional on their face.

    If a necessary consequence of compliance with the French court's orders were a restriction on access by American users, this would be a different and much easier case. The dissent argues that we should remand to the district court to determine whether this is a necessary consequence. But we cannot obtain this determination merely by remanding to the district court. [1223] Before the district court can engage in useful factfinding, it must know whether (or to what extent) Yahoo! has already sufficiently complied with the French court's interim orders. There are two alternative scenarios.

    First, if the French court were to conclude, as LICRA and UEJF contend, that Yahoo! has already complied "in large measure" with the French court's orders, Yahoo! simply has no First Amendment argument. Yahoo! has explicitly stated that its change of policy after the entry of the second interim order was undertaken for reasons entirely independent of the French court's orders. Under this scenario, the question of compliance would disappear, and the district court would have no factfinding role.

    Second, if the French court were to determine, contrary to LICRA and UEJF's contention, that Yahoo! has not complied "in large measure," the question of the necessary consequences for American users would then arise. If and when the French court determines what further compliance is necessary, there might be some appropriate factfinding role for the district court on that question. But even under this scenario, we first need to get a determination from the French court as to what further compliance is necessary, for the district court's factfinding role is dependent on there having been such a prior determination by the French court.

    Under either scenario, the essential initial step is to find out from the French court whether Yahoo! has complied "in large measure" with its orders, and, if not, what further compliance is required. Until we know that, the district court cannot perform any useful factfinding on the question of whether a necessary consequence of compliance with the French court's orders will be to restrict access by Internet users in the United States.

    Conclusion

    First Amendment issues arising out of international Internet use are new, important and difficult. We should not rush to decide such issues based on an inadequate, incomplete or unclear record. We should proceed carefully, with awareness of the limitations of our judicial competence, in this undeveloped area of the law. Precisely because of the novelty, importance and difficulty of the First Amendment issues Yahoo! seeks to litigate, we should scrupulously observe the prudential limitations on the exercise of our power.

    Yahoo! wants a decision providing broad First Amendment protection for speech and speech-related activities on the Internet that might violate the laws or offend the sensibilities of other countries. As currently framed, however, Yahoo!'s suit comes perilously close to a request for a forbidden advisory opinion. There was a live dispute when Yahoo! first filed suit in federal district court, but Yahoo! soon thereafter voluntarily changed its policy to comply, at least in part, with the commands of the French court's interim orders. This change in policy may or may not have mooted Yahoo!'s federal suit, but it has at least come close. Unless and until Yahoo! changes its policy again, and thereby more clearly violates the French court's orders, it is unclear how much is now actually in dispute.

    It is possible that because of Yahoo!'s voluntary change of policy it has now complied "in large measure" with the French court's orders. It is also possible that Yahoo! has not yet complied "in large measure." If further compliance is required, Yahoo! will have to impose further restrictions on access by French users. The necessary consequence of such further restrictions on French users may or may not be that Yahoo! will have to impose restrictions on access by American users. Until we know whether further restrictions on [1224] access by French, and possibly American, users are required, we cannot decide whether or to what degree the First Amendment might be violated by enforcement of the French court's orders, and whether such enforcement would be repugnant to California public policy. We do not know whether further restrictions are required, and what they might be, because Yahoo! has chosen not to ask the French court. Instead, it has chosen to come home to ask for a declaratory judgment that the French court's orders — whatever they may or may not require, and whatever First Amendment questions they may or may not present — are unenforceable in the United States.

    An eight-judge majority of the en banc panel holds, as explained in Part II of this opinion, that the district court properly exercised specific personal jurisdiction over defendants LICRA and UEJF under the criteria of Calder. A three-judge plurality of the panel concludes, as explained in Part III of this opinion, that the suit is unripe for decision under the criteria of Abbott Laboratories. When the votes of the three judges who conclude that the suit is unripe are combined with the votes of the three dissenting judges who conclude that there is no personal jurisdiction over LICRA and UEJF, there are six votes to dismiss Yahoo!'s suit.

    We therefore REVERSE and REMAND to the district court with instructions to dismiss without prejudice.

    FERGUSON, Circuit Judge, with whom O'SCANNLAIN and TASHIMA, Circuit Judges, join with respect to Part I, concurring in the judgment:

    I concur that the District Court judgment in favor of Yahoo! should be reversed and the case dismissed, but I do so based on reasons other than those set forth by the majority. I do not believe that lack of ripeness is the proper ground to dismiss Yahoo!'s suit. Instead, I believe that the District Court did not properly exercise personal jurisdiction over the defendants and also should have abstained from deciding Yahoo!'s claims. Yahoo!'s suit should be dismissed, therefore, either under Rule 12(b)(2) or Rule 12(b)(6) of the Federal Rules of Civil Procedure.

    I.

    The District Court did not properly exercise personal jurisdiction over La Ligue Contre Le Racisme et L'Antisemitisme ("LICRA") and L'Union des Etudiants Juifs de France ("UEJF"). LICRA and UEJF's suit was not "expressly aimed" at California under the "effects" test of Calder v. Jones, 465 U.S. 783, 789-90, 104 S.Ct. 1482, 79 L.Ed.2d 804 (1984), which, I agree with Judge Fletcher, governs this case and may be appropriately applied to the French court orders.

    An intentional act aimed exclusively at a location other than the forum state, which results in harm to a plaintiff in the forum state, does not satisfy the "express aiming" requirement under Calder. In Schwarzenegger v. Fred Martin Motor Co., 374 F.3d 797, 799 (9th Cir.2004), an Ohio car dealer ran an advertisement in the Akron Beacon Journal that featured Arnold Schwarzenegger as "the terminator" without first seeking Schwarzenegger's permission. We held that the advertisement, though it wrongfully depicted Schwarzenegger, a California resident, "was expressly aimed at Ohio rather than California." Id. at 807. Because the dealer's "express aim was local," the district court lacked jurisdiction to hear Schwarzenegger's complaint. Id. Cf. Dole Food Co., Inc. v. Watts, 303 F.3d 1104, 1112 (9th Cir.2002) (finding that European defendants "expressly aimed" at California, the forum state, since they "communicated directly with Dole's California managers to [fraudulently] induce them . . . to enter into significant and detrimental contractual arrangements"); [1225] Bancroft & Masters, Inc. v. Augusta Nat'l, Inc., 223 F.3d 1082, 1088 (9th Cir.2000) (deciding that defendant's "letter was expressly aimed at California[,]" the forum state, "because it individually targeted [Bancroft & Masters], a California corporation doing business almost exclusively in California").

    The majority provides a one-sentence explanation for why LICRA and UEJF's suit was expressly aimed at California: "The suit sought, and the French court granted, orders directing Yahoo! to perform significant acts in California." Maj. op. at 1209.

    That is not true. LICRA and UEJF's suit sought French court orders directing Yahoo! to perform significant acts locally in France, not in California. The May 22, 2000 interim order declares: "[B]y permitting[anti-Semitic] objects to be viewed in France and allowing surfers located in France to participate in such a display of items for sale, the Company Yahoo! Inc. is therefore committing a wrong in the territory of France, a wrong whose unintentional character is averred but which has caused damage to be suffered by LICRA and UEJF, both of whom are dedicated to combating all forms of promotion of Nazism in France." (emphases added).

    To comply with French law, Yahoo! would need "to prevent surfers calling from France from viewing these [anti-Semitic] services on their computer screen"; "to identify the geographical origin of a visiting site from the caller's IP address, which should enable it to prevent surfers calling from France . . . from accessing services and sites which[,] when displayed on a screen installed in France [,] . . . is liable to be deemed an offence in France and/or to constitute a manifestly unlawful trouble [under French law]"; and "to take all measures to dissuade and make impossible any access by a surfer calling from France to disputed sites and services of which the title and/or content constitutes a threat to internal public order." (emphases added).

    There is no evidence whatsoever that LICRA and UEJF had any intention to expressly aim their suit at California. The majority believes that because the effect of the French court orders was for Yahoo! to perform significant acts in California, express aiming on the part of LICRA and UEJF was "obvious." Maj. op. at 1209. But the majority fails to recognize what Schwarzenegger makes clear: express aiming requires intentional conduct by a party directed at the forum state. LICRA and UEJF are two anti-racist French civil liberties organizations. Yahoo! is a global Internet service. At the time LICRA and UEJF brought their suit, they could not precisely have known of Yahoo!'s server locations, security capabilities, or technical procedures or, more important, how they relate to Yahoo!'s California-based operations. LICRA and UEJF had one aim and one aim only: to prevent French citizens from using "Yahoo.fr" and "Yahoo.com" to access illegal anti-Semitic hate merchandise in France. They were plainly concerned with Yahoo!'s actions within France, regardless of where those actions emanated from.

    "It may be true that [LICRA and UEJF]'s intentional [suit] eventually caused harm to [Yahoo!] in California, and [LICRA and UEJF] may have known that [Yahoo!] [was based] in California. But this does not confer jurisdiction, for [LICRA and UEJF]'s express aim was local." Schwarzenegger, 374 F.3d at 807.

    II.

    The District Court should have also abstained from deciding Yahoo!'s claims.

    The common law act of state doctrine specifies:

    Every foreign state is bound to respect the independence of every other sovereign [1226] state, and the court of one country will not sit in judgment on the acts of government of another, done within its own territory. Redress of grievances by reason of such acts must be obtained through the means open to be availed of by sovereign powers as between themselves.

    Underhill v. Hernandez, 168 U.S. 250, 252, 18 S.Ct. 83, 42 L.Ed. 456 (1897). "Judicial . . . engagement in the task of passing on the validity of foreign acts of state may hinder the conduct of foreign affairs." Banco Nacional de Cuba v. Sabbatino, 376 U.S. 398, 423, 84 S.Ct. 923, 11 L.Ed.2d 804 (1964). The act of state doctrine therefore "mandates [judicial] abstention." Liu v. Republic of China, 892 F.2d 1419, 1432 (9th Cir.1989); see also West v. Multibanco Comermex, S.A., 807 F.2d 820, 827 (9th Cir.1987) ("The act of state doctrine is a combination justiciability and abstention rule . . . ").

    While a foreign court judgment arising out of private litigation is generally not an act of state, it can be when it gives effect to the public interest of the foreign government. See Philippine Nat'l Bank v. U.S. Dist. Ct. of Hawaii, 397 F.3d 768, 773 (9th Cir.2005); Liu, 892 F.2d at 1433-34 & n. 2 (citing Restatement (Second) of Foreign Relations of the United States § 41 cmt. d (1965) ("A judgement of a court may be an act of state")).

    In Philippine Nat'l Bank, a dispute arose between a class of plaintiffs and the Republic of Philippines over the right to the assets of Philippine President Ferdinand Marcos's estate. Id. at 770. The class obtained a large judgment in a federal district court in Hawaii against the Marcos estate for human rights violations by the Marcos regime. At the same time, the Republic of Philippines brought suit in the Philippines seeking forfeiture of the Marcos estate's assets on the ground that they were stolen by Marcos from the Philippine government and its people. Id. at 771. The Philippine Supreme Court agreed with the Republic of Philippines and ordered the assets to be forfeited to the Philippine Government. Id. A federal district court in Hawaii, however, ruled that the Philippine Supreme Court judgment violated the due process rights of the class of plaintiffs and was entitled to no judicial deference. Id. at 772.

    We disagreed and held that the Philippine Supreme Court judgment was an act of state because it effectuated the "statutory mandate [of the Philippine government] to recover property allegedly stolen from the treasury." Id. at 773 (quoting In re Estate of Ferdinand Marcos Human Rights Litig., 94 F.3d 539, 546 (9th Cir.1996)). Significantly, we held that the "collection efforts of the Republic [of Philippines]," even though they extended beyond Philippine's borders into Singapore, were "governmental," and the Philippine Supreme Court decision upholding those efforts was therefore an act of state. Philippine Nat'l Bank, 397 F.3d at 773 ("[T]he Republic's `interest in the enforcement of its law does not end at its borders' . . .") (quoting Callejo v. Bancomer, S.A., 764 F.2d 1101, 1121-25 (5th Cir.1985)).

    Like the Philippine forfeiture judgment, both French court orders at issue in this case constitute acts of state. Three factors lead to this conclusion. First, while LICRA and UEJF were private French litigants, they were acting as non-governmental, anti-racist associations and institutional partners with the French government in fighting anti-Semitism.[2] Their injunctive actions against Yahoo! clearly [1227] followed the French government's mandate to enforce Le Nouveau Code Penal Art. R. 645-2 ("Nazi Symbols Act"), a criminal provision. The record makes clear, for example, that LICRA and UEJF litigated with the assistance of Mr. Pierre Dillange, First Deputy Prosecutor representing the office of the Public Prosecutor to the County Court of Paris. Dillange, in fact, "demand[ed]" to the French court "that the reality of the damages suffered by [LICRA and UEJF] be recognised." Prior to the issuance of the French court orders, Dillange publicly condemned the sale of Nazi memorabilia on Yahoo.fr and Yahoo.com calling for "constraints and an injunction" against Yahoo!.[3] LICRA and UEJF litigated their claims in accordance with the demands of the French public prosecutor.

    Second, French justice Jean-Jacques Gomez expressly recognized in his court orders the compelling interest of France to rid its country of anti-Semitic merchandise and speech within its borders. In his May 22, 2000 interim order, for example, he called Yahoo.com "the largest vehicle in existence for the promotion [of] Nazism" and described the commercial sale of Nazi objects as "an affront to the collective memory of a country profoundly traumatized by the atrocities committed by and in the name of the criminal Nazi regime against its citizens." Access to Nazi memorabilia on Yahoo!'s auction sites "constitute[d] a threat to internal public order" and a "wrong in the territory of France." Like the Philippine Supreme Court, the French court here gave clear effect to the collective efforts of French civil liberties organizations, the French government, and French law enforcement to enforce French criminal provisions against anti-Semitism. Justice Gomez's opinion sets forth the moral judgment of France itself.

    Third, the French court orders reflected judicial enforcement of a robust French state policy against racism, xenophobia, and anti-Semitism. France has acceded to the International Convention on the Elimination of all Forms of Racial Discrimination (ICEFRD) (1965) and the International Covenant on Civil and Political Rights (ICCPR) (1966), both of which include provisions against racist speech. See ICCPR, Art. 20-2; ICEFRD, Art. 4(a). Since World War II, France has introduced sweeping legislation to combat anti-Semitism. In July 1972 it passed "Loi Pléven," which criminalized a range of racist behavior from racial defamation and provocation to racial hatred and violence, and in July 1990 it passed "Loi Fabius-Gayssot," which criminalized speech that denied the existence of the Holocaust or that celebrated Nazism. The Nazi Symbols Act, which Yahoo! was found guilty of violating, encompassed France's earlier dramatic efforts to criminalize racist speech within its borders.

    It is apparent then that the French court orders were not merely private judgments but, in fact, reflected the sentiments of two French civil liberties organizations, the French public prosecutor, and, indeed, France itself. They were acts of state.[4]

    [1228] The District Judge sitting in San Jose, California did not have the authority to second guess these orders and should have abstained from invalidating them. He should have deferred to the Executive and Congress to assess the foreign consequences of France's broad policy against anti-Semitic hate speech. See Siderman de Blake v. Republic of Argentina, 965 F.2d 699, 707 (9th Cir.1992) ("The [act of state] doctrine reflects the prudential concern that the courts, if they question the validity of foreign acts taken by sovereign states, may be interfering with the conduct of American foreign policy by the Executive and Congress.") (footnote and citations omitted). Our current government, in fact, is already "fully committed to monitoring and combating anti-Semitism throughout the world."[5]

    The criminal statutes of most nations do not comport with the U.S. Constitution. That does not give judges in this country the unfettered authority to pass critical judgment on their validity, especially where, as here, the criminal statute embodies the determined will of a foreign sovereign to protect its borders from what it deems as morally reprehensible speech of the worst order.

    O'SCANNLAIN, Circuit Judge, with whom FERGUSON and TASHIMA, Circuit Judges, join, concurring only in the judgment:

    Our requirement that a defendant have "purposefully availed" himself of the protections and benefits of the forum state, or have "purposefully directed" his activities into the forum state, must be read in light of the Supreme Court's admonition in Milliken v. Meyer, 311 U.S. 457, 61 S.Ct. 339, 85 L.Ed. 278 (1940), that the exercise of personal jurisdiction must comport with "traditional notions of fair play and substantial justice." Id. at 463, 61 S.Ct. 339. Because I cannot agree that California's exercise of personal jurisdiction over La Ligue Contre Le Racisme et L'Antisemitisme ("LICRA") and L'Union des Etudiants Juifs de France ("UEJF") comports with those basic principles, I respectfully dissent from the majority's opinion while concurring in its conclusion that Yahoo!'s suit must be dismissed. For similar reasons, I concur in Judge Tashima's concurrence and in Part I of Judge Ferguson's concurrence.

    I

    A State's jurisdiction is defined not by force or influence but by physical territory and its judicial power traditionally extended over only those persons and property within its borders. See Pennoyer v. Neff, 95 U.S. 714, 720-22, 24 L.Ed. 565 (1878). The idea of "minimum contacts" developed as a surrogate for actual presence in a State but did not alter the essentially territorial nature of jurisdiction. The question in every personal jurisdiction case, then, is whether an individual's contacts with the forum State are so substantial that they render the extension of sovereign power [1229] just, notwithstanding his lack of physical presence there.

    A

    The personal jurisdiction requirement is not merely a rule of civil procedure; it is a constitutional constraint on the powers of a State, as exercised by its courts, in favor of the due process rights of the individual. See Omni Capital Int'l v. Rudolf Wolff & Co., 484 U.S. 97, 104, 108 S.Ct. 404, 98 L.Ed.2d 415 (1987) ("The requirement that a court have personal jurisdiction flows not from [Article] III, but from the Due Process Clause. It represents a restriction on judicial power not as a matter of sovereignty, but as a matter of individual liberty."). Grounded in the Fourteenth Amendment's protection of the processes necessary to ensure basic fairness in the application of the law, the requirement that an individual have "certain minimum contacts" with the relevant forum "such that the maintenance of the suit does not offend `traditional notions of fair play and substantial justice,'" International Shoe v. Washington, 326 U.S. 310, 316, 66 S.Ct. 154, 90 L.Ed. 95 (1945) (quoting Milliken, 311 U.S. at 463, 61 S.Ct. 339), protects him from the unpredictable and burdensome exercise of authority by foreign courts. It follows from this that the rights and interests of Yahoo! and the interests of the State of California, if not irrelevant to the inquiry, are clearly subordinate to the rights of LICRA and UEJF, the parties against whom jurisdiction is asserted and whose rights are protected by the Due Process Clause.

    The Supreme Court has advised that

    the constitutional touchstone remains whether the defendant purposefully established "minimum contacts" in the forum State. Although it has been argued that foreseeability of causing injury in another State should be sufficient to establish such contacts there when policy considerations so require, the Court has consistently held that this kind of foreseeability is not a "sufficient benchmark" for exercising personal jurisdiction. Instead, the foreseeability that is critical to due process analysis is that the defendant's conduct and connection with the forum State are such that he should reasonably anticipate being haled into court there.

    Burger King v. Rudzewicz, 471 U.S. 462, 474, 105 S.Ct. 2174, 85 L.Ed.2d 528 (1985) (emphases added). By requiring that individuals have "fair warning that a particular activity may subject [them] to the jurisdiction of a foreign sovereign," Shaffer v. Heitner, 433 U.S. 186, 218, 97 S.Ct. 2569, 53 L.Ed.2d 683 (1977) (STEVENS, J., concurring in judgment), the Due Process Clause "gives a degree of predictability to the legal system that allows potential defendants to structure their primary conduct with some minimum assurance as to where that conduct will and will not render them liable to suit." World-Wide Volkswagen Corp. v. Woodson, 444 U.S. 286, 297, 100 S.Ct. 559, 62 L.Ed.2d 490 (1980).

    B

    The Supreme Court has never approved such a radical extension of personal jurisdiction as would sanction the majority's holding that, by litigating a bona fide claim in a foreign court and receiving a favorable judgment, a foreign party automatically assents to being haled into court in the other litigant's home forum. Such a result cannot be reconciled with the "constitutional touchstone" of foreseeability: that the defendant "should reasonably anticipate being haled into court" in the forum. Burger King, 471 U.S. at 474, 105 S.Ct. 2174, 85 L.Ed.2d 528.

    In Calder v. Jones, 465 U.S. 783, 104 S.Ct. 1482, 79 L.Ed.2d 804 (1984), the defendants should reasonably have expected that, by circulating a libelous story in California [1230] about a California celebrity, they would be haled into court in California to answer for their tortious behavior. And in Burger King, because the defendants' business ties with the State of Florida were "shielded by the `benefits and protections'" of Florida's laws, it was "presumptively not unreasonable to require [them] to submit to the burdens of litigation [there] as well." 471 U.S. at 543, 105 S.Ct. 2218. These cases stake out the limits of personal jurisdiction as approved by the Supreme Court.

    LICRA's and UEJF's actions lie beyond that limit. Neither party has ever carried on business or any other activity through which they have availed themselves of the benefits and protections of California's laws,[6] nor should either party have reasonably anticipated that it would be haled into court in California to answer for the legitimate exercise of its rights in France.

    II

    This case was reheard en banc primarily for the purpose of answering the question of whether the underlying action in a non-contract case must be tortious or otherwise wrongful to justify the exercise of personal jurisdiction, or whether the "express aiming" of any action, regardless of culpability, will suffice.[7] Although the resolution of that question does not affect my conclusion that California cannot exercise personal jurisdiction over LICRA or UEJF, I respectfully disagree with the majority's interpretation of Calder on this point.

    A

    Under the majority's reading of Calder, acts giving rise to personal jurisdiction in a non-contract case need not be wrongful. Maj. op. at 1208 ("[W]e do not read Calder necessarily to require in purposeful direction cases that all (or even any) jurisdictionally relevant effects have been caused by wrongful acts."). That conclusion is undermined by the language of Calder itself and requires the majority to divorce that case's holding from its fact — always a dubious exercise. In Calder, the Supreme Court affirmed a decision that had "concluded that a valid basis for jurisdiction existed on the theory that petitioners intended to, and did, cause tortious injury to respondent in California." Calder, 104 S.Ct. at 1485 (emphasis added). The Court itself held that "[i]n this case, petitioners are primary participants in an alleged wrongdoing intentionally directed at a California resident, and jurisdiction is proper on that basis." Id. at 1487 (emphasis added). The wrongfulness of the defendants' acts was, therefore, a key element in the jurisdictional calculus, possibly because a person who has committed a wrongful act should expect to be haled into court by his victim in the victim's home State. Although the Court might have reached the same result if the act in question had not been wrongful — as the majority [1231] apparently presumes it would — it is reckless of us to proceed on the basis of such speculation beyond what is currently the farthest reach of personal jurisdiction approved by the Court.

    B

    The majority's jurisdictional legerdemain is nimble but, like any trick, does not stand up to close scrutiny. It begins innocuously enough by noting that the traditional analysis of minimum contacts depends on whether the disputed act sounds in tort or in contract. In tort cases, "we typically inquire whether a defendant `purposefully direct[s] his activities' at the forum state," maj. op. at 1206. And in commercial and contract cases, "we typically inquire whether a defendant `purposefully avails itself [sic] of the privilege of conducting activities' or `consummate[s][a] transaction' in the forum." Id. and do not require that the defendants actions be wrongful. However, that traditional distinction is abruptly jettisoned when the majority next asserts that "in any personal jurisdiction case we must evaluate all of a defendant's contacts with the forum state, whether or not those contacts involve wrongful activity by the defendant." Id. at 1207 (emphases added).

    The majority's statement is, quite literally, unprecedented. With a stroke of its pen, the majority extends the analysis previously applied only to commercial and contract cases to all assertions of personal jurisdiction. Tellingly, the only cases that the majority musters in support of its novel assertion are commercial or contract-related "purposeful availment" cases. In Quill Corp. v. North Dakota, 504 U.S. 298, 112 S.Ct. 1904, 119 L.Ed.2d 91 (1992), the Supreme Court held that when an out-of-state mail order company "purposefully avails itself of the benefits of an economic market in the forum State, it may subject itself to the State's in personam jurisdiction even if it has no physical presence in the State." 504 U.S. at 302, 112 S.Ct. 1904. And, in Burger King, the Court held that jurisdiction was proper on the grounds that defendants' business ties with the State of Florida were "shielded by the `benefits and protections'" of Florida's laws. 471 U.S. at 543, 105 S.Ct. 2218. In sharp contrast, every "purposeful direction" case that the majority cites in its opinion involved tortious or otherwise wrongful acts by the defendants.

    Given our long line of precedent applying the "purposeful availment" test only in contract and commercial cases, and the majority's concession that this case should be analyzed under Calder's "purposeful direction" test, see maj. op. at 1208, the majority's conflation of the elements of these two tests is an unseemly act of judicial slight of hand. LICRA and UEJF are, indisputably, non-commercial actors who have never purposefully availed themselves of the benefits or protections of California's laws. Therefore, neither Calder nor any other Supreme Court precedent justifies California's assertion of personal jurisdiction over them.

    III

    LICRA's and UEJF's actions and contacts with the State of California were, at most, incidental to the legitimate exercise of their rights under French law. They should not have reasonably anticipated being haled into court in California to answer for their prosecution of a lawsuit in France. Because California's exercise of personal jurisdiction over them on that basis would violate traditional notions of fair play and substantial justice and, therefore, the procedural guarantees of the Due Process Clause, I would remand the case with instructions to dismiss for want of personal jurisdiction and not reach the issue of ripeness.

    [1232] Thus, while I must dissent from its rationale, I concur in the majority's conclusion that the district court's opinion must be reversed.

    TASHIMA, Circuit Judge, with whom FERGUSON and O'SCANNLAIN, Circuit Judges, join, concurring in the judgment:

    I concur in the judgment reversing and remanding with instructions to dismiss this action, but I dissent from the majority's conclusion that personal jurisdiction exists over La Ligue Contre Le Racisme et L'Antisemitisme ("LICRA") and L'Union des Etudiants Juifs de France ("UEJF"). I therefore concur in Part I of Judge Ferguson's concurring opinion — that a district court located in California cannot exercise personal jurisdiction over LICRA and UEJF.

    Because I believe that the district court lacked in personam jurisdiction, I would not reach the issues discussed in Part III of the majority opinion[8] — ripeness — and Part II of Judge Ferguson's concurring opinion — whether, even if it had jurisdiction over the defendants, the district court should have abstained from deciding this case. I do believe, however, that Judge Ferguson's eloquent discussion in Part II of the reasons why he would hold that abstention is proper further supports why personal jurisdiction is lacking in this case.

    LICRA and UEJF ("defendants") had only three contacts with California. These contacts were a cease and desist letter, the service of process to commence the French action, and the subsequent service of two interim orders on Yahoo!. Service was made in accordance with the requirements of the Hague Convention on the service abroad of judicial documents. As the majority rightly acknowledges, these contacts are an insufficient basis for the exercise of personal jurisdiction over defendants. Maj. op. at 1207-1209.

    The majority goes on, however, to find a sufficient basis for the exercise of personal jurisdiction over defendants in two interim orders issued by the French court because those orders "direct[ed] Yahoo! to take actions in California, on threat of a substantial penalty." Id. at 1209. The majority's conclusion is not based on any contact with California, but on acts which it contends were "expressly aimed at the forum state." Id. (quoting Schwarzenegger v. Fred Martin Motor Co., 374 F.3d 797, 805 (9th Cir.2004)). But neither Schwarzenegger nor any other case relied on by the majority based a finding of specific jurisdiction on conduct expressly aimed at the forum state which conduct was not also a contact with the forum state. Here, for the first time, the majority completely divorces the expressly-aimed conduct from the requirement that that conduct also be a contact with the forum state. Thus, I submit that the finding of personal jurisdiction on the basis of Calder's[9] "effects" test in the circumstances of this case is a radical extension of that doctrine.

    It is self-evident that the orders are the orders of the French court, not acts of defendants. Thus, more precisely, the majority's finding of personal jurisdiction is, in fact, based on LICRA and UEJF petitioning the French court for relief under French law. But should the petitioning by a citizen of the courts of his or her own country to uphold the laws of that country form the sole basis of personal jurisdiction over that citizen by the courts of a foreign [1233] country? The majority's answer is yes. That answer, seems to me, to be perverse. First, the bringing and prosecuting of an action in a French court are all acts done wholly in France. None of these acts constitutes a "contact" with California. Second, no citizen of any country can safely sue a foreign defendant under the majority's theory of specific jurisdiction because the sought judgment, including an ordinary money judgment for injury or damages, will have an adverse "effect" on the defendant's purse or treasury in that defendant's home country. In this sense, every lawsuit naming a foreign defendant can be said to be expressly aimed at that defendant's home state (or nation). Thus, unless it is anchored to a contact with the forum, express aiming becomes a meaningless test in terms of due process.

    Moreover, courts, even when acting at the behest of a private petitioner, have an independent interest and obligation to uphold their nations' domestic laws, particularly when, as here, those laws are designed to carry out an important and strongly-held national policy. Thus, as Judge Ferguson reminds us, it is the manner in which the French courts have determined to vindicate French national policy — that "state action" — that has the adverse "effect" in California that Yahoo! is complaining about, not the acts of defendants in petitioning for French anti-Semitism laws to be upheld. It was not defendants who determined the terms and scope of injunctive relief, nor was it defendants who determined that continuing non-compliance should be "subject to a penalty," or the amount of such a penalty. Needless to say, defendants will not be the ones who decide whether such penalties ultimately will have to be paid or waived.[10]

    Whatever other conduct Calder's "effects" test was intended to encompass, it surely was not intended to include attribution of the effects of an intervening court's order when a citizen does no more that petition a court in his own country for relief under domestic law, particularly in a case, such as this, in which defendants have had no contact that would "provide a sufficient basis for jurisdiction."[11] Maj. op. at 1208. For these additional reasons, I concur in Part I of Judge Ferguson's concurring opinion.

    FISHER, Circuit Judge, with whom HAWKINS, PAEZ, CLIFTON and BEA, Circuit Judges, join, concurring in part and dissenting in part:[12]

    I. Overview

    Stated simply, the issue before us is whether a United States Internet service [1234] provider, whose published content has been restricted by a foreign court injunction, may look to the United States federal courts to determine the enforceability of those restrictions under the United States Constitution's First Amendment. The French injunctive orders — backed by substantial, retroactive monetary penalties for noncompliance — require Yahoo! to block access from French territory to Nazi-related material on its website.[13] Some prohibited content is readily identifiable, such as Nazi artifacts or copies of Mein Kampf. Much, however, is not. The orders impose the following sweeping mandate:

    We order the Company YAHOO! Inc. to take all necessary measures to dissuade and render impossible any access via Yahoo.com to the Nazi artifact auction service and to any other site or service that may be construed as constituting an apology for Nazism or a contesting of Nazi crimes.

    (Emphasis added.) In traditional First Amendment terms, this injunctive mandate is a prior restraint on what Yahoo! may post (or control access to) on its U.S.-located server — imposed under principles of French law and in such facially vague and overbroad terms that even the majority does not know "whether further restrictions on access by French, and possibly American, users are required" to comply with the French orders. (Op. at 1223.) Yahoo! can either hope to comply with what the French court (and the defendants here) deems to be inappropriate content by attempting to block access to material Yahoo! thinks the orders cover or by simply removing any questionable content altogether. Or Yahoo! can ignore the French court's mandate in whole or in part and accept the risk of substantial accruing fines. The majority, however, is unmoved. For it, Yahoo!'s proper recourse is to take its case back to France. We cannot agree.

    As the district court readily concluded in its thoughtful opinion, "[a] United States court constitutionally could not make such an order." Yahoo!, Inc. v. La Ligue Contre Le Racisme et L'Antisemitisme, 169 F.Supp.2d 1181, 1189 (N.D.Cal.2001) (hereinafter "Yahoo II"). It specifically found that the orders are "far too general and imprecise to survive the strict scrutiny required by the First Amendment," and that "[p]hrases such as `all necessary measures' and `render impossible' instruct Yahoo! to undertake efforts that will impermissibly chill and perhaps even censor protected speech." Yahoo II, 169 F.Supp.2d at 1189-90 (citing Bd. of Airport Comm'rs v. Jews for Jesus, 482 U.S. 569, 107 S.Ct. 2568, 96 L.Ed.2d 500 (1987); and Gooding v. Wilson, 405 U.S. 518, 92 S.Ct. 1103, 31 L.Ed.2d 408 (1972)). The district court emphasized that "`[t]he loss of First Amendment freedoms, for even minimal periods of time, unquestionably constitutes irreparable injury.'" Id. at 1190 (quoting Elrod v. Burns, 427 U.S. 347, 373, 96 S.Ct. 2673, 49 L.Ed.2d 547 (1976) (citing New York Times Co. v. United States, 403 U.S. 713, 91 S.Ct. 2140, 29 L.Ed.2d 822 (1971))).

    The issue is not whether the French defendants who obtained the injunctive orders, or the French court that issued them, are justified in trying to suppress hateful speech. We of course recognize the horrors of the Holocaust and the scourge of anti-Semitism, and France's understandable interest in protecting its citizens from those who would defend or glorify either. Nor is the issue one of extra-territorial application of the First Amendment; if anything, it is the extra-territorial application of French law to the United [1235] States. We do not question the validity of the French orders on French soil, and Yahoo! has complied with the orders as they relate to its website. Rather the question we face in this federal lawsuit is whether our own country's fundamental constitutional guarantee of freedom of speech protects Yahoo! (and, derivatively, at least its users in the United States) against some or all of the restraints the French defendants have deliberately imposed upon it within the United States. "`[P]rior restraints on speech and publication are the most serious and the least tolerable infringement on First Amendment rights.'" Tory v. Cochran, ___ U.S. ___, ___, 125 S.Ct. 2108, 2111, 161 L.Ed.2d 1042 (2005) (quoting Neb. Press Ass'n v. Stuart, 427 U.S. 539, 559, 96 S.Ct. 2791, 49 L.Ed.2d 683 (1976)).

    The majority, after properly opening the door to the federal courthouse by upholding personal jurisdiction, nonetheless turns a blind eye to the constitutional free speech interests of Yahoo!, throwing it out of court because those interests are not "ripe" for adjudication. The majority's thesis rests on the contention that the French "orders do not by their terms limit access by users outside France in any way." (Op. at 1216.) But as the majority recognizes elsewhere in its opinion (Op. at 1216-1218), the crux of this case is not in the words of the order alone, but in their application. And to assess the effects of the orders, one cannot simply disregard the "what" of the orders and focus only on their "who."

    As we shall explain later, we disagree with the majority's conclusion that uncertainties about whether Yahoo! can technologically isolate the effects of the orders only to France-based users compel us to withdraw the case from the district court. Even assuming such uncertainties exist and are material, the district court is fully capable of exercising its factfinding role to resolve them. But there is no uncertainty that the mandate imposed on Yahoo! is also content based, and the orders identify that content in terms that on their face are overbroad and vague. They require Yahoo! to guess what has to be censored on its Internet services here in the United States, under threat of monetary sanction if it guesses wrong. In that respect, the orders are facially unconstitutional.

    By their terms, the orders reach "any other site or service [in addition to the auction service] that may be construed as constituting an apology for Nazism or a contesting of Nazi crimes." (Emphasis added.) As the district court rightly understood, this is the crux of Yahoo!'s facial overbreadth and vagueness concern:

    Yahoo! seeks protection for its actions in the United States, specifically the ways in which it configures and operates its auction and Yahoo.com sites. Moreover, the French order requires Yahoo! not only to render it impossible for French citizens to access the proscribed content but also to interpret an impermissibly overbroad and vague definition of the content that is proscribed. . . . In light of the Court's conclusion that enforcement of the French order by a United States court would be inconsistent with the First Amendment, the factual question of whether Yahoo! possesses the technology to comply with the order is immaterial. Even assuming for purposes of the present motion that Yahoo! does possess such technology, compliance still would involve an impermissible restriction on speech. . . .

    Yahoo II, 169 F.Supp.2d at 1193-94 (emphasis added).

    Surely the majority is not suggesting that Yahoo! has no First Amendment protection from being sanctioned when it could not guess or it guessed wrong as to what it was supposed to censor on its [1236] domestic servers — even if limited to France-based users. (And if not so limited, so much the worse.) Yet the majority faults Yahoo! because — like Yahoo! itself — we do not know whether its current activities are permitted by the orders. (Op. at 1216.) This is to apply First Amendment precedents exactly backwards. As the majority admits, "[t]he boundary line between what is permitted and not permitted is somewhat uncertain for users in France." (Op. at 1222.) Under such circumstances, we blame the law, not the speaker.

    Instead, the majority effectively imposes an exhaustion requirement on Yahoo! to litigate this issue in France, confirm that it is still is not in compliance with the orders (just as it was not on May 22 and November 20, 2000) and obtain a "final" adverse judgment before the majority will consider this case ripe. In doing so, the majority imposes a heightened standard on a U.S. plaintiff seeking to vindicate its First Amendment rights when that plaintiff is challenging a foreign prior restraint. Principles of ripeness (or comity) do not require this result. The extraordinary hurdles the majority creates are inconsistent with our established jurisprudence protecting this country's tradition of free expression. See, e.g., City of Lakewood v. Plain Dealer Publ'g Co., 486 U.S. 750, 755-56, 108 S.Ct. 2138, 100 L.Ed.2d 771 (1988) (holding that a plaintiff need not have applied and been denied a newspaper rack license before challenging a city ordinance as an unconstitutional prior restraint on speech). To say so is not to deny France's interests in protecting its own citizens from harmful speech, but only to recognize that federal courts have the duty to adjudicate and uphold the legitimate constitutional rights of litigants who have properly invoked our federal jurisdiction.

    In correctly sustaining personal jurisdiction over the defendants and in finding an Article III case or controversy, the majority concedes the central dilemma Yahoo! faces as a result of the French injunction. "[W]hile Yahoo! does not independently wish to take steps to comply more fully with the French court's orders, it states that it fears that it may be subject to a substantial (and increasing) fine if it does not." (Op. at 1210.) Acknowledging the obvious chilling effect of the injunction, the majority recognizes that "[e]ven if the French court's orders are not enforced against Yahoo!, the very existence of those orders may be thought to cast a shadow on the legality of Yahoo!'s current policy." (Op. at 1210-1211.)

    But unfortunately the majority then stops short, concluding that the "level of harm [suffered by Yahoo!] is not sufficient to overcome the factual uncertainty bearing on the legal question presented and thereby to render this suit ripe." (Op. at 1221.) With respect, the majority creates its own factual dilemma — and bad First Amendment precedent — in its attempt to find daylight between its holdings on personal jurisdiction and ripeness. We agree that the Calder "effects" test, see Schwarzenegger v. Fred Martin Motor Co., 374 F.3d 797, 803 (9th Cir.2004) (citing Calder v. Jones, 465 U.S. 783, 104 S.Ct. 1482, 79 L.Ed.2d 804 (1984)), need not be satisfied by the same degree of harm as ripeness (Op. at 1218), but the majority's rationale for finding the harm sufficient in one instance and deficient in the other is seriously flawed.

    By peremptorily terminating Yahoo!'s access to federal court, the majority establishes a new and burdensome standard for vindicating First Amendment rights in the Internet context, threatening the Internet's vitality as a medium for robust, open debate. It also bypasses the factfinding role of the district court — failing to credit much of what the district court found on [1237] the record as litigated below, and removing the district court from the process of resolving the factual issues the majority now finds so vital to Yahoo!'s First Amendment claims. Accordingly, although we concur in that part of the majority's opinion upholding personal and Article III jurisdiction, we respectfully dissent from its ultimate holding that this case is not ripe for adjudication.

    II. Prudential Ripeness

    The majority invokes prudential ripeness because it finds Yahoo!'s circumstances suffer from "prematurity and abstractness" that preclude our reaching Yahoo!'s claim that the French injunction on its face violates the First Amendment. (Op. at 1211-1212.) As did the district court, we conclude otherwise.

    A. Fitness of the issues for judicial resolution

    1. A "purely legal" question

    The majority holds this case unfit for judicial resolution by suggesting that it does not involve a "purely legal" question, Abbott Laboratories v. Gardner, 387 U.S. 136, 149, 87 S.Ct. 1507, 18 L.Ed.2d 681 (1967), but instead requires us to sort through factual uncertainties, which ultimately make adjudication inappropriate. (Op. at 1212.) Yet even if the majority were correct that Yahoo!'s case suffers from a lack of factual development, it does not follow that the suit is therefore rendered unripe. When a dispositive fact is missing from the district court record, we usually remand for further factfinding. We do not peremptorily throw litigants out of court and expect them to petition a foreign court for relief.

    To begin with, this case fundamentally involves a straight-forward legal question: whether the French injunction as ordered against Yahoo! runs afoul of the First Amendment. The answer calls for a legal application of free speech doctrine to final orders that on their face are vague and overbroad. True, the defendants must take steps in the French court to initiate actual enforcement, but Yahoo! is subject to the orders and to a retrospective financial penalty for noncompliance. The majority's argument that we should give weight to the label "interim" because it indicates that "the French court contemplated that it might enter later orders" is a make-weight. (Op. at 1215.) A court may contemplate issuing subsequent orders whether or not a prior order on the subject is called "interim" or "final." We need not be distracted by the label "interim," because, as the district court found, "there is no dispute that the French order is valid under French law and that the French Court may fix a penalty retroactive to the date of the order." Yahoo II, 169 F.Supp.2d at 1190.

    Cases involving far less definitive or targeted mandates — not yet enforced against the complaining party — have been treated as final actions ripe for adjudication. In Abbott Laboratories, one of the majority's lynchpin cases, drug manufacturers challenged the Food and Drug Commissioner's regulation requiring that their products' labels show both a drug's generic and its brand name. The Supreme Court, addressing the "purely legal" issue presented, held that the regulation was a final agency action, even though it was a statement of general applicability and violations of the new rule could be enforced only by the Attorney General authorizing criminal and seizure actions. 387 U.S. at 151-52, 87 S.Ct. 1507. The Court held the case ripe for pre-enforcement review because the Commissioner's labeling order placed "petitioners in a dilemma that it was the very purpose of the Declaratory Judgment Act to ameliorate." Id. at 152, 87 S.Ct. 1507. The Court cited the district court's finding that petitioners either "`must comply [1238] [with the label changeovers] . . . or they must follow their present course and risk prosecution'" and concluded that the latter "course would risk serious criminal and civil penalties for the unlawful distribution of `misbranded' drugs." Id. at 152-53, 87 S.Ct. 1507. See also Frozen Food Express v. United States, 351 U.S. 40, 43-44, 76 S.Ct. 569, 100 L.Ed. 910 (1956) (holding justiciable a challenge to an Interstate Commerce Commission rule because violations could be punished by criminal sanctions and the rule itself would cause companies to conform their behavior to the regulation); cf. United States v. Storer Broadcasting Co., 351 U.S. 192, 198, 76 S.Ct. 763, 100 L.Ed. 1081 (1956) (finding standing to challenge a Federal Communications Commission rule limiting radio licenses even though the broadcaster had not yet received an unfavorable decision).[14]

    The final, targeted injunction before us presents the same kind of purely legal issue — with Yahoo! confronting the dilemma of whether or not to stand by its United States constitutional rights or constrain its speech and that of its users to avoid a French-imposed penalty. Legions of cases permit First Amendment challenges to governmental actions or decrees that on their face are vague, overbroad and threaten to chill protected speech. Indeed, the sweeping injunction here presents just such a paradigmatic case. See, e.g., Freedom to Travel Campaign v. Newcomb, 82 F.3d 1431, 1434-35 (9th Cir.1996) (rejecting a ripeness defense to a facial attack on blanket travel restrictions to Cuba under the First and Fifth Amendments, even though the plaintiff group had never applied for a license, because the case presented purely legal questions); see also Forsyth County, Ga. v. Nationalist Movement, 505 U.S. 123, 129-30, 112 S.Ct. 2395, 120 L.Ed.2d 101 (1992) (addressing a facial First Amendment challenge to a licensing scheme even though the plaintiff had never applied for a permit, citing numerous First Amendment cases involving facial claims); Steffel v. Thompson, 415 U.S. 452, 459, 94 S.Ct. 1209, 39 L.Ed.2d 505 (1974) (concluding that petitioner had established an actual controversy based on his threatened criminal trespass arrest by state police for distributing political handbills and holding that he did not need to "first expose himself to actual arrest or prosecution to be entitled to challenge a statute that he claims deters the exercise of his constitutional rights"). Yahoo! seeks nothing more than for a United States court to resolve its legal claim that the French court injunction by its very nature — in whole or in part — threatens Yahoo!'s [1239] protected speech. See NAACP, W. Region v. City of Richmond, 743 F.2d 1346, 1352, 1358 (9th Cir.1984) (upholding standing to bring facial challenge to "substantially overbroad" city parade ordinance).

    2. Comity and the repugnance of unconstitutional injunctions

    We do not agree with the majority's professed uncertainties as to whether a California court, under principles of comity, would be inclined to enforce a foreign court order that infringes upon a U.S. corporation's First Amendment rights. The "repugnancy" standard the majority invokes is easily satisfied here. California's case law and its federal underpinnings tell us to honor foreign court judgments unless they "prejudice the rights of United States citizens or violate domestic public policy." In re Stephanie M., 7 Cal.4th 295, 27 Cal.Rptr.2d 595, 867 P.2d 706, 716 (1994) (citing Hilton v. Guyot, 159 U.S. 113, 202-03, 16 S.Ct. 139, 40 L.Ed. 95 (1895); and Victrix S.S. Co. v. Salen Dry Cargo A.B., 825 F.2d 709, 713 (2d Cir.1987)). The French orders on their face — and by putting Yahoo! at risk of substantial penalties — violate the First Amendment and are plainly contrary to one of America's, and by extension California's, most cherished public policies.[15] In short, they constitute a foreign judgment that is "repugnant to public policy." (Op. at 1215.)

    The district court considered the role of comity but ultimately found that it was outweighed by U.S. constitutional freedoms. "Although France has the sovereign right to regulate what speech is permissible in France, this Court may not enforce a foreign order that violates the protections of the United States Constitution by chilling protected speech that occurs simultaneously within our borders." Yahoo II, 169 F.Supp.2d at 1192. This finding does not mean that every foreign court judgment implicating speech in the United States would be deemed repugnant to American public policy and therefore unenforceable, but this particular judgment is so vague and overbroad that it fails the repugnancy analysis. Significantly, the defendants do not argue to us that the French injunction comports with the First Amendment. Indeed, they did not even appeal the district court's ultimate finding that the orders are unconstitutional.

    The majority goes to great lengths to avoid labeling a prior restraint on speech — overbroad and vague by its terms — as "repugnant to public policy" and is content to leave in place foreign [1240] court orders that so obviously violate the First Amendment. (Op. at 1215.) In reaching this result, the majority has succumbed to an error of logic. It has conflated foreign orders that are somewhat inconsistent with U.S. law with those that violate U.S. law. It is one thing for U.S. courts to pass on foreign attorney's fees larger than what domestic laws would award, see In re Hashim, 213 F.3d 1169, 1172 (9th Cir.2000), or to recognize a judgment pursuant to a foreign statute of limitations longer than that of its domestic analogue, see Milhoux v. Linder, 902 P.2d 856, 861-62 (Colo.Ct.App.1995). It is quite another to imply, as the majority does, that a violation of the U.S. Constitution is no different from any other "[i]nconsistency with American law," which the majority claims "is not necessarily enough to prevent recognition and enforcement of a foreign judgment in the United States." (Op. at 1215.)

    Neither In re Hashim nor Milhoux implicated federal or state constitutional rights. Indeed, both cases held that the foreign judgments being challenged were not repugnant to the public policy of either Arizona or Colorado, respectively. Where a foreign judgment was held to be repugnant to California's public policy, the repugnancy was based on the violation of California's Uniform Child Custody Jurisdiction Act that would have resulted had the foreign order been enforced. See In re Stephanie M., 27 Cal.Rptr.2d 595, 867 P.2d at 716. The majority provides no explanation why the California courts would refuse to enforce a foreign judgment that violated a state statute, yet be willing to enforce a foreign judgment that violates the federal (and perhaps the state) Constitution.

    The majority's dictum implying that foreign judgments that would be unconstitutional if entered by a U.S. court may nonetheless be enforceable is troubling. Under the principles articulated today, a foreign party can use a foreign court decree to censor free speech here in the United States on any range of subjects it finds objectionable — religion, democracy, gender equality — in the name of enforcing its own country's laws. The good intentions of even sympathetic foreign parties such as LICRA and UEJF in this case are not the standard. How could a California court honor the French defendants' good intentions in proscribing pro-Nazi speech when the City of St. Paul's good intentions did not cure its anti-hate speech code of viewpoint discrimination and constitutional infirmity even when directed at cross-burnings? See R.A.V. v. City of St. Paul, 505 U.S. 377, 392, 112 S.Ct. 2538, 120 L.Ed.2d 305 (1992) ("St. Paul has no such authority to license one side of a debate to fight freestyle, while requiring the other to follow Marquis of Queensberry rules."); see also Collin v. Smith, 578 F.2d 1197, 1201 (7th Cir.), cert. denied, 439 U.S. 916, 99 S.Ct. 291, 58 L.Ed.2d 264 (1978) (striking down on First Amendment grounds several Skokie, Illinois ordinances prohibiting the National Socialist Party of America from marching through the town: "First Amendment rights are truly precious and fundamental to our national life. . . . It is, after all, in part the fact that our constitutional system protects minorities unpopular at a particular time or place from governmental harassment and intimidation, that distinguishes life in this country from life under the Third Reich.")

    People in the United States and France should abhor anti-Semitism and the horrors perpetrated by the Nazi Party. Nonetheless, our constitutional law differs from French jurisprudence in our approach to hate speech. Our law reflects deeply held political beliefs about freedom of expression in this country. Borrowing Justice Brandeis's formulation, "the remedy to be applied [to expose falsehood and fallacies] is more speech, not enforced silence." [1241] Whitney v. California, 274 U.S. 357, 377, 47 S.Ct. 641, 71 L.Ed. 1095 (1927) (Brandeis, J., concurring).

    3. The alleged lack of factual development

    Even accepting the majority's assumption that this case does not turn on purely legal issues, the concerns the majority invokes as reasons to withhold judicial resolution are either unconvincing or at most reasons for remand. For instance, the majority seems to call into question whether the French court's injunction is sufficiently final because the orders are labeled "interim," notwithstanding their unconditional and mandatory language. (Op. at 1215.) In considering whether the injunction survives U.S. laws, we must take the orders issued by the French court as final actions, reflecting that court's view of Yahoo!'s conduct and current obligations under French law. There is no reason for us to assume that the French court intends something different from the words of its own mandatory orders — just as we would not assume that a U.S. federal or state court would not stand by an injunctive order it has issued.

    Moreover, by insisting on withholding judicial resolution, the majority disregards the district court's factual determinations, and its role in resolving factual disputes. First, with respect to the content at issue, the majority minimizes the district court findings that Yahoo! hosts content violating the specific terms of the orders. As the district court found, Yahoo! "continues to offer at least some Third Reich memorabilia as well as Mein Kampf on its auction site and permits access to numerous web pages with Nazi-related and anti-Semitic content." Yahoo II, 169 F.Supp.2d at 1189. The district court took judicial notice from its own search of the site (in October 2001) that using the keyword "Nazi" called up 69 Nazi-related items posted for sale, such as stamps, coins and a copy of Mein Kampf. Id. at 1185 n. 3. The district court also conducted keyword searches on Yahoo!'s general website, finding thousands of sites referring to "Jewish conspiracy," promoting modern-day Nazism or suggesting the Holocaust did not happen. Id. at n. 4.[16]

    Clouding the majority's view of the facts are the defendants' assertions before us and in the district court that they "have no present intention of taking legal action against Yahoo! in the United States" because they consider Yahoo! to be in "substantial compliance with the French order." Yahoo II, 169 F.Supp.2d at 1188. But the French court has never made such a determination of Yahoo!'s alleged compliance. Instead, the majority speculates that because Yahoo! France has "complied [in France] in large measure with the spirit and letter" of the May 22 French order, "compliance `in large measure' by Yahoo! is very likely to be satisfactory to the French court." (Op. at 1215-1216.) But Yahoo! is not Yahoo! France, and the French court did not explain the factual basis for its finding of compliance.

    Nor have the defendants ever taken any steps to stipulate in a legal forum that Yahoo! is in compliance with the injunction. Thus the district court properly gave no weight to the defendants' professions of Yahoo!'s substantial compliance. The court pointedly observed that the defendants "have not taken steps available to [1242] them under French law to seek withdrawal of the orders or to petition the French court to absolve Yahoo! from any penalty," Yahoo II, 169 F.Supp.2d at 1188, and they gave no indication they would pursue such measures when pressed on the subject. Id. at 1189 n. 7.

    During oral argument before us, defense counsel conceded that the defendants did not want to foreclose their options by agreeing to such a stipulation. As the majority recognizes (Op. at 1204), should Yahoo! alter its content in a way that the defendants disapprove of, they want the judicial authority to seek relief and mandate Yahoo!'s compliance. (Oral Arg. 1:02.) The majority in large part hinges its analysis on the defendants' litigation position of saying that they have no problem now with Yahoo!'s conduct but declining to take any steps to eliminate the speech injunction or accruing financial penalties. See Abbott Laboratories, 387 U.S. at 154, 87 S.Ct. 1507 (concluding that the "subsequent representation of the Department of Justice" that it was likely to impose only civil sanctions for violations, thus mitigating the harm to the plaintiff, "should not suffice to defeat" the claim); see also Culinary Workers Union, Local 226 v. Del Papa, 200 F.3d 614, 617-18 (9th Cir.1999) (disregarding attorney general's claim that she lacked authority to carry out specific threat of prosecution in holding that a real controversy existed for purposes of Article III).

    The majority claims that "we do not know whether the French court would hold that Yahoo! is now violating its two interim orders." (Op at 1215.) Ironically, the majority thereby highlights the very threat Yahoo! faces. Uncertainty about whether the sword of Damocles might fall is precisely the reason Yahoo! seeks a determination of its First Amendment rights in federal court. See Metro. Wash. Airports Auth. v. Citizens for Abatement of Aircraft Noise, Inc., 501 U.S. 252, 265 n. 13, 111 S.Ct. 2298, 115 L.Ed.2d 236 (1991); Chang v. United States, 327 F.3d 911, 921 (9th Cir.2003) (recognizing that this court does not require "Damocles's sword to fall" before it will adjudicate a case).

    In sum, the uncertainties Yahoo! faces are not reasons to delay adjudication. Rather, they provide a compelling basis for a federal court to hear Yahoo!'s First Amendment challenge at this time, as the district court did.

    The fact that Yahoo! does not know whether its efforts to date have met the French Court's mandate is the precise harm against which the Declaratory Judgment Act is designed to protect. The Declaratory Judgment Act was designed to relieve potential defendants from the Damoclean threat of impending litigation which a harassing adversary might brandish, while initiating suit at his leisure or never.

    Yahoo II, 169 F.Supp.2d at 1189 (emphasis added).[17] Instead, the majority turns Yahoo!'s [1243] uncertainties against it — relegating it to the French courts for clarification and absolution.

    B. Substantial hardship of withholding judicial consideration

    Even more perplexing is the majority's conclusion that Yahoo! does not face "substantial hardship" because of our unwillingness to adjudicate its First Amendment claim. The majority attempts to avoid the obvious chilling effect of an overbroad and vague injunction in two creative and troubling ways. First, the majority opines "with some confidence" that Yahoo! need not fear the enforcement of a fine because "it is exceedingly unlikely that the sword [of Damocles] will ever fall" (Op. at 1218) — another speculative assessment, we submit. It also faults Yahoo! for failing to proffer examples of "anything that it is now not doing but would do if permitted by the orders" (Op. at 1220) and thereby imposes a new, higher burden on a First Amendment plaintiff to establish a chilling effect.

    1. The French orders chill speech

    First, the majority overlooks Yahoo!'s claim that it faces actual abridgment of its current speech — not just a chilling effect on its ever-changing Web content. As the majority does acknowledge, Yahoo! hosts content on its auction site, including the sale of Mein Kampf, that is specifically prohibited by the terms of the injunction. The district court's findings of impermissible material still present on the auction site demonstrate that Yahoo! is currently engaged in speech that the French orders — by their terms — compel it to foreclose to some users or forgo entirely. Yahoo! opts not to accede to the injunction, thereby incurring daily accumulating fines should its current or future behavior displease LICRA or UEJF. Certainly Yahoo! should not have to abstain from conduct it believes is constitutionally protected solely for us to find its claim ripe. Cf. City of Auburn, 260 F.3d at 1173 (9th Cir.2001) (noting that finding case unripe would require party to comply with "costly and cumbersome" franchise requirements, only for the party to then raise "exactly the same argument that it makes here").

    More importantly, the majority largely ignores the broad and diffuse scope of the French injunction — which extends well beyond Yahoo!'s auction site and clearly raises the question whether it is substantively possible for Yahoo! to comply. Apart from entirely obvious cases, how can one determine with any certainty whether something "may be construed as constituting an apology for Nazism or a contesting of Nazi crimes"? The majority makes the rather startling assertion that "[b]efore the district court can engage in useful factfinding, it must know whether (or to what extent) Yahoo! has already sufficiently complied with the French court's interim orders." (Op. at 1222.) Of course, this is precisely the crux of Yahoo!'s predicament — and [1244] highlights the vagueness and overbreadth of the orders. We know the actions Yahoo! has taken and not taken with respect to Nazi paraphernalia appearing on its site. The only reason we cannot determine "whether (or to what extent) Yahoo! has already sufficiently complied" with the French orders is because we cannot assess the scope of the orders themselves.[18] It is this very kind of uncertainty that epitomizes a purely legal question of facial infringement of First Amendment rights and the harms routinely associated with such an infringement.

    In plain terms, if no one but the French court can decipher the meaning of its injunction aimed at Yahoo!'s speech, how can Yahoo! comply? Yahoo! has to know what content it has to screen from France-based users. The French orders contain no meaningful instructions for Yahoo! to winnow permitted speech from unpermitted speech. It is the absence of a discernible line between the permitted and the unpermitted that makes the orders facially unconstitutional. As the district court concluded, and as discussed previously, "compliance would still involve an impermissible restriction on speech" because it would require Yahoo! to interpret the vague and overbroad injunction as to what content is prohibited and which users should be denied access, on pain of substantial penalty should it guess wrong. See Yahoo II, 169 F.Supp.2d at 1193-94.

    Ultimately, the majority's parsimonious treatment of the free speech issues here culminates with its reducing Yahoo!'s argument to an interest in merely "allowing access by users in France" to Nazi materials. (Op. at 1221.) Yahoo! is allegedly seeking "a First Amendment right to violate French criminal law and to facilitate the violation of French criminal law by others."[19] (Op. at 1221.) Notably, even the defendants have not construed Yahoo!'s First Amendment argument in such crabbed terms.

    But suppose Yahoo! really were concerned only with not having to act in the United States as an enforcer of France's restrictions on Internet access by France-based users. That would not make the constitutional implications of the effects on Yahoo!'s United States operations go away. Yahoo! cannot merely act in France to restrict access by users located in France; the French orders require Yahoo! to make changes to its servers and protocols in the United States. That Yahoo! seeks First Amendment protection from having to compromise its domestic operations to comply with a foreign injunction [1245] does not translate into its seeking the right simply to violate French law. This case is not about the extra-territorial application of the First Amendment; it is about the extra-territorial application of France's anti-Holocaust denial speech codes and the extent to which compliance may infringe Yahoo!'s rights of free speech here in the United States.

    The majority, however, views the French orders as concerning "speech accessible solely by those outside the United States." (Op. at 1217.) Additionally, it accepts that Yahoo! can screen out access to any prohibited materials by "most" — estimated to be 70-90% — of France-based users. (Op. at 1216.) This reasoning is flawed in several respects.

    First, Yahoo! does not target specific users by initiating content directed solely at them. Rather, anyone who logs on to, including users in France, gains access to material on Yahoo!'s message boards, search engines, auction sites and other services. It is the accessing of vaguely and overbroadly described content — by anyone in French territory — that the orders prohibit and hold Yahoo! responsible for preventing. Thus, even if one could readily and reliably limit the universe of Internet users whose access must be censored — an assumption the record before us does not justify — Yahoo! would still be at a loss to define the universe of content it must censor.

    Second, the factual question of whether it is technologically feasible for Yahoo! to monitor the postings and filter the millions of users accessing the website — assuming such technology actually bears on Yahoo!'s First Amendment claims — is an unresolved issue that should be returned to the district court. The parties have not addressed the specifics of technical feasibility issue on this appeal, nor the validity of the experts' report. Thus the 70% and 90% figures the majority adopts from that report depend solely on the majority's reading of a translated technical and ambiguous document, the scientific merits of which have not been addressed even in the district court. LICRA and UEJF did raise the issue of feasibility below, but the district court denied them discovery regarding technological feasibility of screening France-based users because it deemed the issue immaterial to the court's First Amendment ruling. See Yahoo II, 169 F.Supp.2d at 1194. The defendants have not appealed either the district court's First Amendment decision or its discovery ruling. To the extent that the technological feasibility issue has been argued at all on appeal, Yahoo! has said that it "could not monitor the content of these millions of postings and listings to its U.S.-based Internet services" and that it essentially faces a binary choice between self-censorship and paying the French fines.

    On the record before us — lacking expert testimony and cross-examination, much less district court findings of fact — we do not believe we as appellate judges can or should accept as a given that Yahoo! can readily and reliably identify 70% of the users it must censor, "irrespective of whether a Yahoo! user sought access to an auction site, or to a site denying the existence of the Holocaust or constituting an apology for Nazism." (Op. at 1203-1204.)

    This is particularly true given that the experts' report is replete with hearsay, technological assumptions and disclaimers. Most importantly, the experts explicitly limited their analysis to how an Internet "surfer" in France could be prevented from accessing prohibited content only on Yahoo!'s auction site, not all such content that might find its way onto generally. As the experts emphasized — echoing Yahoo!'s own concern about the imprecision of the orders:

    The decisions of the [French] court and the demands made are precisely directed [1246] against the auctions site. No grievance against any other Yahoo! sites or services is formulated with sufficient precision to enable the consultants to propose suitable and effective technical solutions. In these circumstances, the consultants will therefore confine their answers to the matter of the auctions site. . . . [20]

    (Emphasis added.) The experts also emphasized, "[t]he measures to be taken depend upon the particular case in point. They cannot be generalised to all sites and services on the Internet. In this case, the site in question is pages.auctions.yahoo.com." (Emphasis added.)

    Of course, the French orders do not solely prohibit content on Yahoo!'s auction site but, by their terms, encompass content on all of Yahoo!'s services. Yahoo!'s services extend far beyond its auction site and include its search engine, e-mail, classified listings, personal Web pages, shopping, message boards, chat rooms and news stories.

    The majority — like the French court itself — seems to credit two of the three experts who estimated as many as 90% of France-based users of Yahoo's auction site could be identified and screened. The methodology underlying this estimate, however, further illustrates the uncertainty of predicting Internet identification and screening, compounded by the vague and overbroad mandate of the court orders. Assuming that "70% of the IP addresses assigned to French surfers can be matched with certainty to a service provider located in France, and can be filtered," all three experts agreed that "no filtering method is capable of identifying all French surfers or surfers connecting from French territory."[21] To reach 90%, two experts relied on a voluntary "sworn declaration of nationality" by a French surfer that "could be made when a first connection is made to a disputed site, in this case the Yahoo auctions site. . . ." (Emphasis added.)[22] They suggested asking for the declaration of nationality at "the home page of the auctions site" or "in the context of a search for Nazi objects if the word `Nazi' is included in the user's request. . . ." In short, the experts' 90% figure depends on the ability to link users to a specific Yahoo! site and to specific content on that site.

    The third expert, Vinton Cerf, a 1997 recipient of the United States National Medal of Technology for co-designing the architecture of the Internet,[23] disavowed relying on users' self-identification at all, concluding that "it does not appear to be [1247] very feasible to rely on discovering the geographic locations of users for purposes of imposing filtering of the kind described in the [French] Court Order."

    Given the orders' broad language, none of the experts could devise a system for screening out France-based users that went beyond the auction site. Therefore, even if were true that Yahoo! can identify up to 70% of all of its France-based users, irrespective of the site or service they are accessing, the evidence is clear that geographical identification alone would not enable Yahoo! to prohibit such users from accessing 100% of the content proscribed by the French orders — indeed, Yahoo! could not even come close on that side of the compliance equation.

    There are other serious questions about the experts' report that should be part of an evidentiary hearing in the district court. For example, the 70% IP-address screening figure was derived in part from information provided by a French Internet association regarding how many of its access providers can identify whether their users are located in France. Such anecdotal data do not demonstrate conclusively that Yahoo! itself has the capability to identify the location of its users.

    Indeed, the method the experts proposed for Yahoo! to identify users is imprecise. The experts noted that for a number of reasons the "real world" location of a user may not be readily identifiable. For instance, a French citizen who uses AOL for Internet service may be shown as having an IP address from Virginia, where AOL's network is located. In other instances, users may choose to mask the geographical origin of their Internet address.

    Thus we cannot assume, as does the majority, that this case is about Yahoo! restricting access only by French users, 70-90% of whom are readily identifiable regardless of what content they may seek out on. The validity of these percentage assumptions not only drives the majority's definition of whose access is restricted, but also its apparent willingness to assume that even if Yahoo! can identify only 70% of the prohibited universe of users, that would be good enough. If technical feasibility is to be the lynchpin on which Yahoo!'s day in federal court depends, then let the parties return to the district court for proper factfinding. Instead, the majority preempts the district court's factfinding function, interpreting the French experts' report as conclusive evidence in order to deny Yahoo! access to the court altogether.

    Lastly, there is the issue of cost of compliance. There can be no dispute that the very nature of the French orders puts Yahoo! to the choice of incurring the costs to develop and implement mechanisms to filter out individual users based on location or removing content from its service altogether. This type of immediate financial burden clearly suffices to make a case ripe for adjudication, even if we accept the majority's proposition that the threat of enforcement is remote. See Pac. Gas & Elec. Co. v. State Energy Res. Conservation & Dev. Comm'n, 461 U.S. 190, 197-98, 103 S.Ct. 1713, 75 L.Ed.2d 752 (1983) (holding ripe for review a preemption challenge to a regulation imposing a moratorium on new nuclear plants because petitioners would face substantial financial hardship if they built plants while hoping the law would be struck down); City of Auburn v. Qwest Corp., 260 F.3d 1160, 1173 (9th Cir.2001) (noting that finding case unripe would require party to comply with "costly and cumbersome" franchise requirements).[24]

    [1248] 2. The enforceability of foreign penal judgments

    Recognizing that the risk of a large monetary penalty must inevitably weigh heavily in Yahoo!'s assessment of its options, the majority tries to neutralize the risk — creating a protective shield by invoking the doctrine that United States courts will not enforce the penal judgments of other countries. It thus assures Yahoo! that "even if the French court were to impose a monetary penalty against Yahoo!, it is exceedingly unlikely that any court in California — or indeed elsewhere in the United States — would enforce it" because it is a penal judgment. (Op. at 1218.)

    It is true as Justice Marshall observed that "[t]he courts of no country execute the penal laws of another," The Antelope, 23 U.S. (10 Wheat.) 66, 123, 6 L.Ed. 268 (1825). But that begs the question whether the French injunction itself or the accruing fines are truly penal. Although we respect the majority's scholarship, this issue has not been the focus of the parties' briefs or arguments, and thus we cannot share the majority's level of confidence that its dictum is sufficiently accurate — or binding — that we should remove the risk of a substantial, retroactive monetary penalty from the First Amendment or ripeness analysis. As with the French defendants' assurances that they consider Yahoo! currently in substantial compliance, absent a binding court order actually freeing Yahoo! from the enforcement of the French orders, Yahoo! remains at serious risk if it fails to conform its web content to the dictates of those orders.

    "The test whether a law is penal, in the strict and primary sense, is whether the wrong sought to be redressed is a wrong to the public, or a wrong to the individual. . . ." Huntington v. Attrill, 146 U.S. 657, 668, 13 S.Ct. 224, 36 L.Ed. 1123 (1892). The Court warned against the "danger of being misled by the different shades of meaning allowed to the word `penal' in our language." Id. at 666, 13 S.Ct. 224.[25] Determining whether a sanction is penal or civil in nature is not always a simple task. Cf. F.J. Hanshaw Enters., Inc. v. Emerald River Dev., Inc., 244 F.3d 1128, 1137-38 (9th Cir.2001) (establishing procedural protections due a party based on whether sanctions were criminal or civil in nature).

    Although LICRA and UEJF's substantive claims against Yahoo! in French court depended in part upon Yahoo!'s violations of French criminal law,[26] the record suggests [1249] that the French lawsuits were civil rather than criminal and, more importantly, that the French orders primarily sought to redress a wrong to LICRA and UEJF rather than a wrong to the French public. Of course, we agree with the majority that "the label `civil' does not strip a remedy of its penal nature." (Op. at 1219.) However, that still begs the question whether or not the French accruing fines were penal. On this point, the majority asserts that there is some language in the November 20 order that supports the characterization of the fines as penal and that in any event the fines are potentially much larger than the nominal damages awarded to UEJF and therefore the "award of one Franc [to UEJF] cannot render the orders primarily remedial rather than punitive in nature." (Op. at 1220.) The majority cites no authority for the novel arithmetic balancing test it proposes to distinguish penal from non-penal orders, and although we admit there is some language in the orders that supports holding the French orders punitive, there is also significant language that supports the conclusion that the orders sought to redress a wrong done to LICRA and UEJF. The proper test for determining whether the French orders are penal is a purposive one, see Huntington, 146 U.S. at 668, 13 S.Ct. 224, and based on the record before us, we do not share the majority's certainty that the orders are undoubtedly penal in nature.

    French law gives standing to public interest, non-governmental organizations dedicated to defending the interests of members of certain victimized groups, including victims of the Holocaust (déportés), to initiate enumerated types of civil actions (but not criminal prosecutions) on behalf of such victims. See, e.g., C. Pr. Pén. arts. 2-4 & 2-5; Law of July 29, 1881 (Law on Freedom of the Press) (2004), art. 48-2. Yahoo!'s challenge to UEJF's standing under Article 48-2 of the French Law on Freedom of the Press and the French court's subsequent finding that LICRA and UEJF "are dedicated to combating all forms of promotion of Nazism in France" suggest that the French trial was a civil proceeding under one of the specialized French standing statutes. This conclusion is further supported by the French court's reliance on Article 809 of the New Code of Civil Procedure for its authority to issue orders.

    Furthermore, the award of damages to UEJF and other relief "by way of restitution" strongly suggests that the French court orders were predominantly civil and remedial rather than penal.[27] The court based its award of damages and other restitution in its May 22 decision on a finding that the exhibition for sale of Nazi objects "has caused damage to be suffered by LICRA and UEJF." The French court reiterated this finding of direct harm in its November 20 decision: "this display [of Nazi objects] clearly causes damage in France to the plaintiff associations who are justified in demanding the cessation and reparation thereof." In this context, the additional relief afforded to the French plaintiffs — an injunction ordering Yahoo! to cease its harmful activity in France — appears to be merely an additional remedy in a civil suit.

    [1250] As with the French injunction, the accruing fines are similarly more likely civil than penal in nature. The most natural reading of the French court's rationale for imposing the accruing fines is that such fines were meant to coerce Yahoo! into compliance with the substance of the French injunction. Rather than assessing the fines retroactively as a court would do when redressing the public wrong Yahoo! had allegedly already committed, the French court made the fines entirely conditional on Yahoo!'s future behavior beginning three months after the date of the second French order.

    The U.S. analogue for such a regime of per diem fines is civil contempt. See Sarl Louis Feraud Int'l v. Viewfinder Inc., 2005 WL 2420525, at *1, *3, 2005 U.S. Dist. LEXIS 22242, at *2, *7 (S.D.N.Y.2005) (characterizing a French court's judgments and a "fine (`astreinte') of 50,000 francs per day for each day that View-finder failed to comply with each judgment" as "an injunction backed by coercive penalties analogous to a civil contempt fine under American law"). "In contrast [to criminal contempt], civil contempt sanctions, or those penalties designed to compel future compliance with a court order, are considered to be coercive and avoidable through obedience." Int'l Union, United Mine Workers v. Bagwell, 512 U.S. 821, 827, 114 S.Ct. 2552, 129 L.Ed.2d 642 (1994). See also 17 C.J.S. Contempt § 64 (2005) ("[C]ontempt proceedings brought to vindicate the dignity and authority of the court may be characterized as criminal in nature, whereas those brought to preserve and enforce the rights of private parties are remedial and civil in character."). Courts have the power to order either imprisonment or the payment of fines when holding a party in civil contempt: "A close analogy to coercive imprisonment is a per diem fine imposed for each day a contemnor fails to comply with an affirmative court order. Like civil imprisonment, such fines exert a constant coercive pressure." Bagwell, 512 U.S. at 829, 114 S.Ct. 2552. See also People v. Gonzalez, 12 Cal.4th 804, 50 Cal.Rptr.2d 74, 910 P.2d 1366, 1373 (1996).[28]

    Yahoo! was afforded a three-month safe harbor to allow it to implement the French court's orders, and only then would any fines be assessed. As with a U.S. civil contempt order, the fines were entirely "avoidable through obedience." Because the French coercive fines' aim is enforcement of an underlying injunction that is civil (preventing the continuation of harm the French court found LICRA and UEJF had already suffered) rather than penal (benefiting French public justice or vindicating [1251] the French court's dignity and authority), the California rule of comity announced in In re Stephanie M. might well apply, were it not for the orders' substantive unconstitutionality.[29]See 27 Cal.Rptr.2d 595, 867 P.2d at 716.

    For these reasons, unlike the majority we cannot take the monetary penalty out of the ripeness analysis and assume that Yahoo! is not harmed by the very threat of the French orders' possible enforcement. Once again, at the least this is another issue that could and should be remanded to the district court for appropriate briefing and factfinding.

    3. A new, higher burden for proving chilling effect

    Finally, the majority dismisses the chilling effect of the orders by placing the burden on Yahoo! to identify other speech it wants to engage in but which is foreclosed by the French orders. What more should Yahoo! have to specify about the exact manner in which the objectionable content would appear on its site? Millions of postings and other material flow through Yahoo!'s networks each day.[30] Yahoo! cannot possibly predict when and how specific content prohibited by the French orders will make its way onto its service. For example, a user could decide at any time to post a message or a link to a website containing impermissible content. Because it acts as a platform for other speakers, Yahoo! cannot, as the majority demands, identify the specific speech it wishes to engage in that is prohibited by the injunction.

    Nor should it have to. To place such a requirement on an Internet provider — essentially forcing it to speculate as to the particular speech activity its millions of users "might" engage in as senders or recipients — is to afford it no First Amendment protection at all. As the Supreme Court has recognized, "`[t]he Internet . . . offer[s] a forum for a true diversity of political discourse, unique opportunities for cultural development, and myriad avenues for intellectual activity.'" Ashcroft v. American Civil Liberties Union, 535 U.S. 564, 566, 122 S.Ct. 1700, 152 L.Ed.2d 771 (2002) (quoting 47 U.S.C. § 230(a)(3) (1994 ed., Supp. V)); see Batzel v. Smith, 333 F.3d 1018, 1027 (9th Cir.2003) (emphasizing that Congress, in insulating Internet service providers from liability for certain content published on their sites, recognized the importance of protecting the "unfettered and unregulated development of free speech on the Internet").[31]

    [1252] The majority would impose on Yahoo! far greater burdens and litigation risks than those alleging First Amendment violations by domestic parties would have to bear. Yahoo! is expected to try to persuade the French court to narrow or eliminate the very injunction Yahoo! has unsuccessfully fought against in France from the beginning. Unconstrained by our First Amendment, the French court might well take the opportunity to sanction Yahoo! for noncompliance — and do nothing to alleviate the sweeping restraint on the content of the website. If the defendants want to narrow the injunction such that it might warrant comity, that burden should fall on them, not Yahoo!.

    But even if Yahoo! went to the French court and obtained a ruling that its current auction site policy and Internet services content comply with the orders, that would not resolve Yahoo!'s First Amendment problem unless the sweeping injunction itself were permanently withdrawn or narrowed. All Yahoo! would obtain would be clearance for its current operations; it would remain exposed to the risk of violating the orders and incurring penalties should it deviate from those current practices or should the defendants decide that Yahoo!'s content has become objectionable. The very nature of Yahoo!'s business is inherently mutable — that is the essence of the Internet, because of the sheer number and constantly changing identity of its users and of the content those users may seek or themselves post on. Only a United States court can provide Yahoo! with a legal resolution of its claim that the injunctive order, as written, cannot be enforced in the United States without infringing the company's First Amendment rights, thereby relieving it of the coercive threat hanging over its website and the operation of its business. By denying adjudication, the majority abdicates our proper role in protecting Yahoo!'s constitutional rights.

    In so doing, it leaves in place a foreign country's vague and overbroad judgment mandating a U.S. company to bar access to prohibited content by Internet users from that country. This astonishing result is itself the strongest argument for finding Yahoo!'s claims ripe for adjudication. Are we to assume that U.S.-based Internet service providers are now the policing agencies for whatever content another country wants to keep from those within its territorial borders — such as, for example, controversial views on democracy, religion or the status of women? If the majority's application of the First Amendment in the global Internet context in this case is to become the standard — whether as a matter of constitutional law or comity — then it should be adopted (or not) after full consideration of the constitutional merits, not as a justification for avoiding the issue altogether as not ripe for adjudication.

    III. Conclusion

    Without doubt, the hateful speech the defendants in this case seek to suppress is to be condemned. But censoring speech we find repugnant does not comport with our cherished First Amendment. It is well-settled that a hate speech code which "prohibits otherwise permitted speech solely on the basis of the subjects the speech addresses" is "facially unconstitutional." R.A.V., 505 U.S. at 381, 112 S.Ct. 2538. Under the majority's reasoning, a party targeted for enforcement of a foreign judgment restricting its speech in the United States will have no recourse but to [1253] appeal to the foreign court, which does not recognize the First Amendment, to try to escape the strictures of the decree — or to demonstrate compliance, either through voluntary action or by submitting to its terms. Only after enduring the decree's chilling effects while this process plays out, and then faced with whatever sanction the foreign court may impose for noncompliance, may the doors of the United States District Court be opened.

    We should not allow a foreign court order to be used as leverage to quash constitutionally protected speech by denying the United States-based target an adjudication of its constitutional rights in federal court. By invoking the doctrine of prudential ripeness — notwithstanding having found both personal jurisdiction over the two foreign defendants and a constitutional case or controversy — the majority does just that, denying Yahoo! the only forum in which it can free itself of a facially unconstitutional injunction. Moreover, in doing so the majority creates a new and troubling precedent for U.S.-based Internet service providers who may be confronted with foreign court orders that require them to police the content accessible to Internet users from another country. We therefore respectfully dissent from the majority's ripeness decision.

    [1] The French court's orders are written in French. We quote from the English translation provided in the record. Counsel for LICRA and UEJF contended at oral argument that the words "all necessary measures" (underlined and italicized above) are a mistranslation of the French text. The original French for the entire phrase (italicized above) is "prendre toutes les mesures de nature à dissuader et à rendre impossible." Counsel contended that the words "toutes les mesures de nature à" are more accurately translated as "all reasonable (or available) measures."

    [2] The French anti-racism Pleven law ("Loi Pléven"), passed in July 1972, expressly permits French anti-racist associations to file legal actions to combat racism. The law confers upon French anti-racist associations official "civil party" status in such matters. The French text of the law is referenced at: http://www.culture.gouv.fr/culture/infos-pratiques/droitculture/cinema/pdf/l-290781.pdf; see also Eric Bleich, RACE POLITICS IN AND FRANCE: IDEAS AND POLICYMAKING SINCE THE 1960S 135-39 (2003).

    [3] Reuters, "Paris Prosecutor Condemns Nazi Auctions on Yahoo," May 15, 2000, available at http://www.icare.to/archivemay2000.html.

    [4] It is also worth noting that the French court orders were final criminal judgments that Yahoo! elected not to appeal through the French court system. Instead, Yahoo! brought the present declaratory relief action for a U.S. district court to invalidate the French court orders based on a violation of Yahoo!'s First Amendment right. In so doing, Yahoo! here is essentially no different than a party losing in state court who seeks to vindicate his or her federal rights by challenging the adverse state court judgment in federal district court. The Supreme Court has barred such opportunistic attempts at relitigation under the Rooker-Feldman doctrine. See Exxon Mobil Corp. v. Saudi Basic Indus. Corp., 544 U.S. 280, 125 S.Ct. 1517, 1521-22, 161 L.Ed.2d 454 (2005).

    [5] Bureau of Democracy, Human Rights, and Labor, U.S. Dept. of State, REPORT ON GLOBAL ANTI-SEMITISM, 5-6, 13-15 (January 2005) (discussing France's efforts to combat anti-Semitism). On October 16, 2004, President George W. Bush signed into law the Global Anti-Semitism Review Act, Pub.L. No. 108-332, which authorized the 2005 report, the first of its kind.

    [6] I agree with the majority that the mailing in good faith of cease and desist letters and the use of the United States Marshal's Office to effect service of process of documents related to the French legal proceedings are not sufficient bases for jurisdiction. Maj. op. at 1208-1209.

    [7] Although the fact is ignored by the majority, this question was settled law in our circuit prior this appeal being reheard en banc. In Bancroft & Masters, Inc. v. Augusta Nat'l Inc., 223 F.3d 1082, 1086 (9th Cir.2000), the panel made it clear that its decision relied on the assumption that the defendant had engaged in tortious conduct. Judge Sneed, writing for a majority of the panel, further held that "[j]urisdiction in California would be ripe for challenge if following the development of trial it should appear that ANI acted reasonably and in good faith to protect its trademark against an infringer." Id. at 1089 (Sneed, J., concurring).

    [8] I refer to the opinion authored by Judge W.A. Fletcher as the "majority opinion," because it commands a majority of the en banc court on the issue of personal jurisdiction, although that is not the majority that controls the disposition of the case.

    [9] Calder v. Jones, 465 U.S. 783, 789-90, 104 S.Ct. 1482, 79 L.Ed.2d 804 (1984).

    [10] Indeed, if any penalties are ever paid, they will not redound to the benefit of defendants, but "are payable to the government." Maj. op at 1220.

    [11] What the majority opinion calls a "third contact," maj. op. at 1208 ("However, the third contact, considered in conjunction with the first two, does provide such a [sufficient] basis [for personal jurisdiction]."), is not a "contact" with California at all. The majority classifies as the "[t]hird, and most important [contact], LICRA and UEJF have obtained two interim orders from the French court directing Yahoo! to take actions in California, on threat of a substantial penalty." Id. at 1209. It cites no authority for the proposition that conduct by LICRA and UEJF which takes place entirely in France can be classified as a "contact" with California.

    [12] Like Judge Tashima, we refer to Judge Fletcher's opinion as the "majority" or the "majority opinion" because an eight-judge majority of the en banc court joins Part II of the opinion on the issue of personal jurisdiction. As the per curiam and Judge Fletcher's opinions explain, however, Judge Fletcher's articulated rationale on ripeness in Part III of his opinion represents a three-judge plurality and does not command a majority of the en banc court. Nevertheless, we refer to Judge Fletcher's opinion as the "majority" throughout our dissent for ease of reference.

    [13] As the majority recognizes, any Internet user in France or a French territory — whether or not a French citizen or resident — can gain access to Yahoo!'s U.S.-based server by typing into her browser or linking through . (Op. at 1202.)

    [14] The majority's citation of Adler v. Bd. of Educ., 342 U.S. 485, 72 S.Ct. 380, 96 L.Ed. 517 (1952), as a "noted example" of a debate over ripeness in the context of speech is inapposite. (Op. at 1212.) Adler, a case affirming limits on the speech of teachers in New York public schools during the post-World War II "Red Scare," not only concerns the constitutionally distinct situation of a state government regulating the speech of its employees (as opposed to a court being asked to enforce a speech-restrictive injunction against a corporation), but also predates important modern free speech precedents establishing the doctrine of facial invalidation. See, e.g., Lakewood, 486 U.S. at 755-56, 108 S.Ct. 2138. Furthermore, Justice Frankfurter was the sole dissenter (and the sole Justice) to question the suit's ripeness; Justices Black and Douglas were convinced the suit was ripe and that New York's laws infringed upon public school teachers' First Amendment rights. See Adler, 342 U.S. at 496-511, 72 S.Ct. 380. Because Adler itself would certainly be reasoned, and likely decided, differently today given cases such as Connick v. Myers, 461 U.S. 138, 103 S.Ct. 1684, 75 L.Ed.2d 708 (1983), and Pickering v. Bd. of Educ., 391 U.S. 563, 88 S.Ct. 1731, 20 L.Ed.2d 811 (1968), which recognized a government employee's interest in commenting on matters of public concern, Adler's approach to ripeness is hardly illuminating even within the narrow confines of government employee speech.

    [15] As the majority correctly notes, it is California's public policy (rather than U.S. public policy) that is relevant to a comity analysis in a federal diversity case. (Op. at 1212.) However, although Yahoo! focused its energies on alleging violations of the federal First Amendment rather than violations of the analogous provision of the California Constitution, see art. I, § 2(a), it is certainly not California's public policy to countenance violations of the United States Constitution. Indeed, the California Supreme Court has held California's free speech clause to be more expansive than the First Amendment. See Golden Gateway Ctr. v. Golden Gateway Tenants Ass'n, 26 Cal.4th 1013, 111 Cal.Rptr.2d 336, 29 P.3d 797, 801 (2001) ("Unlike the United States Constitution, which couches the right to free speech as a limit on congressional power, the California Constitution gives `[e]very person' an affirmative right to free speech. Accordingly, we have held that our free speech clause is `more definitive and inclusive than the First Amendment.'") (internal citations omitted). See also Sarl Louis Feraud Int'l v. Viewfinder Inc., 2005 WL 2420525, *3, 2005 U.S. Dist. LEXIS 22242, at *19 (S.D.N.Y.2005) ("American courts have recognized that foreign judgments that run afoul of First Amendment values are inconsistent with our notions of what is fair and just, and conflict with the strong public policy of our State [New York].") (emphasis in original).

    [16] The French defendants initially objected to a far broader array of content than the limited category of items Yahoo! now excludes under its revised auction site policy. UEJF's plea for relief asked the French court to mandate that Yahoo! remove from all browser directories the index heading entitled "negationists" and any link "bringing together, equating or presenting directly or indirectly as equivalent sites categorized under the heading `Holocaust' and those indexed as negationist."

    [17] Yahoo!'s circumstances are readily distinguishable from those found not ripe in Socialist Labor Party v. Gilligan, 406 U.S. 583, 92 S.Ct. 1716, 32 L.Ed.2d 317 (1972). There, the principal First Amendment claims the party leveled against Ohio's election code were mooted by legislative amendments, leaving only a subsidiary challenge to a loyalty oath. The Court found this claim "singularly sparse in its factual allegations," with no suggestion that the Party had ever refused or would refuse in the future to sign the oath, or that it had suffered or would suffer any injury from the existence of the oath requirement. Id. at 586, 92 S.Ct. 1716.

    Similarly, in American-Arab Anti-Discrimination Committee v. Thornburgh, 970 F.2d 501 (9th Cir.1992), members of the Popular Front for the Liberation of Palestine (PFLP) contended that anti-Communist provisions of the McCarran-Walter Act unconstitutionally put them at risk of deportation for engaging in protected First Amendment activities without the opportunity for a fair and impartial hearing before the INS. We held that the plaintiffs were sufficiently at risk of government prosecution to give them standing; but we found their claims not ripe because there was "a sketchy record . . . with many unknown facts," such as whether the plaintiffs were actually members of the PFLP or what acts the government alleged they had committed, and we emphasized that the INS had not yet interpreted or applied the challenged provisions. Id. at 510-11.

    In marked contrast to these cases, here the French injunction remains extant and as broadly worded as ever; the defendants have refused to stipulate to Yahoo!'s compliance; and the district court has found actual noncompliance with specific terms as well as an overall risk of noncompliance with fatally undefined terms — thereby subjecting Yahoo! to the risk of substantial monetary fines and the chilling effect of the vague and overbroad injunction. Additionally, there is no court or agency — other than this federal court — that can address Yahoo!'s United States constitutional claim.

    [18] It is telling that even the Internet experts relied upon by the French court were unable to recommend a "suitable and effective technical solution" for Yahoo! to screen out France-based users from any of its sites or services, other than the auction site, that may be construed as constituting an apology for Nazism or a contesting of Nazi crimes because "[n]o grievance against any . . . Yahoo! sites or services[other than the auction site] is formulated with sufficient precision."

    [19] According to the majority, "the French court's interim orders do not by their terms require Yahoo! to restrict access by Internet users in the United States." (Op. at 1221.) This is not Yahoo!'s position. The company has asserted that complying with the French orders would compel it to remove prohibited material from its United States-based Internet services and reengineer its servers, also located in the United States, to identify both France-based users and prohibited material that may be posted in the future; therefore, it may not be possible to comply with the French orders without rendering certain content inaccessible to all users, including those in the United States and not just those in France. Nor does Yahoo! appear to be interested in asserting its constitutional rights solely for the sake of violating French law. To comply with the orders as they affect the company's French services, Yahoo! now removes any posted material it becomes aware of on its site that would violate French law.

    [20] Even as to screening content on the auction site, the experts acknowledged that it was not possible for Yahoo! to "exclude a priori items which have not been described by their owner as being of Nazi origin or belonging to the Nazi era." How then would Yahoo! keep the prohibited material from being accessed? The report suggested that a more "radical solution" might be warranted, essentially prohibiting any search containing the word "Nazi" by an identified French user. How such Nazi paraphernalia which has not been described by its owners with the label "Nazi" could be screened remains a mystery.

    [21] Significantly, the experts were at pains to caution that even the 70% figure based on IP addresses has a short shelf life: "[t]he consultants stress that there is no evidence to suggest that the same will apply in the future. Encapsulation is becoming more widespread, service and access providers are becoming more international, and surfers are increasingly intent on protecting their rights to privacy."

    [22] Notably, the French orders compel Yahoo! to prohibit access by any users in French territory, not just French citizens. Thus a declaration of "nationality" does not seem adequate in any event.

    [23] See Technology Administration, Department of Commerce, The National Medal of Technology Recipients, at http://www.technology.gov/Medal/Recipients.htm.

    [24] The mere possibility of future fines can have very real financial consequences for a publicly held corporation like Yahoo!. To the extent it is material to a corporation's financial condition, such companies are required to disclose contingent liabilities in Form 10-Q and 10-K statements filed with the Securities and Exchange Commission. See Securities Exchange Act of 1934, §§ 10(b), 15(d), 15 U.S.C. §§ 78j(b), 78o(d); 17 C.F.R. §§ 240.10b-5, 240.12b-20; see also Financial Accounting Standards Board Statement of Financial Standards No. 5, available at http:// www.fasb.org/pdf/fas5.pdf. Such filings may adversely affect the credit ratings and hence the valuation of shares of such companies. In another context, we have held that financial impacts on a business resulting from legal uncertainty support a finding that a case is ripe. See Chang v. United States, 327 F.3d 911, 922 (9th Cir.2003).

    [25] The Supreme Court's warning in Huntington has even greater salience when we are attempting to determine "the different shades of meaning allowed to the word `penal'" in a language other than our own. (Op. at 1219.)

    [26] LICRA and UEJF's claims are based in part on a French law that criminalizes the public wearing or display of the uniforms, insignias and emblems of any organization declared criminal by the post-World War II International (Nuremberg) Military Tribunal (e.g., the Nazi Party). See C. Pén. R645-1. One of the most serious penalties for violation of this provision of the penal code is a fine. See id. Their claims also appear to rely on the French Law of July 29, 1881 (Law on Freedom of the Press) (2004), which, among other things, criminalizes Holocaust denial, see art. 24 bis, and the incitement of discrimination, hatred or violence on the basis of belonging to a particular ethnic, national, racial group, see art. 24, ¶ 6. Both crimes carry a penalty of one year imprisonment or a fine of 45,000 Euros or both. See id.

    [27] The French court ordered payment by Yahoo! (jointly and severally with Yahoo! France) of provisional damages of 1 Franc to UEJF. As a means of effecting restitution for the harm suffered, the French court also ordered Yahoo! to pay for the publication of one of the French decisions in "five daily or weekly publications at the choice of [UEJF]."

    [28] Admittedly, the characterization of the fines as coercive yet non-penal may depend on whether Yahoo! "is afforded an opportunity to purge" its liability to pay the fines. Bagwell, 512 U.S. at 829, 114 S.Ct. 2552. Alternatively, the fines may be compensatory and therefore non-penal if they are payable to LICRA and UEJF "for losses sustained" rather than to the French government. See id. The majority claims that the "penalties are payable to the government and not designed to compensate the French student groups for losses suffered." (Op. at 1220.) However, nothing in the record indicates to whom the fines are payable. Furthermore, the majority fails to acknowledge the possibility, indeed the probability, that the fines were not designed to punish Yahoo! for its past behavior, but rather to prevent future harm to LICRA and UEJF. Coercive per diem fines need not be "designed to compensate [plaintiffs] for losses suffered" (Op. at 1220), in order to be non-penal, so long as their purpose is to "preserve and enforce the rights of private parties," 17 C.J.S. Contempt § 64 (2005). In any event, such uncertainty concerning the nature of the fines merely reinforces the conclusion that further factfinding by the district court is necessary before we can jump to the conclusion that the French fines are penal and unenforceable.

    [29] Even if the exact accruing fines as calculated by the French orders were not directly enforceable under California law, Yahoo! could face the possibility that a California court would enforce a foreign injunction with its own state contempt proceedings under comity doctrine (again, assuming no substantive constitutional defect). Cf. Biewend v. Biewend, 17 Cal.2d 108, 109 P.2d 701, 704 (1941) ("Upon the basis of comity, however, as distinguished from the requirements of full faith and credit, the California courts have in numerous cases ordered that a foreign decree for future payments of alimony be established as the decree of the California court with the same force and effect as if it had been entered in this state, including punishment for contempt if the defendant fails to comply."), overruled on other grounds by Worthley v. Worthley, 44 Cal.2d 465, 283 P.2d 19, 22-23 (1955).

    [30] The record indicates that as of July 2000, Yahoo! and its subsidiaries had 146 million users worldwide. Each month Yahoo! users added or edited more than 15 million Geocities web pages and posted more than 6 million classified advertisements. There were more than 2.5 million active auction items viewable on Yahoo! each day and 200,000 Yahoo! clubs were accessed each day by members who posted messages, uploaded photos or added Internet links.

    [31] Batzel analyzed the rationale for the provisions protecting Internet providers under 47 U.S.C. § 230, which Yahoo! invoked before the district court as a statutory basis for preventing enforcement of the French court orders here.

    1.5 American Libraries Ass'n v. Pataki 1.5 American Libraries Ass'n v. Pataki

    Pataki is a useful introductory case, not only because of the regulatory issues, but also because it provides a basic description of how the internet is used. While the internet is used much more broadly than at the time of Pataki, the basic description here provides a simple technical foundation for internet law.

    969 F.Supp. 160 (1997)

    AMERICAN LIBRARIES ASSOCIATION; et al., Plaintiffs,
    v.
    George PATAKI, et al., Defendants.

    No. 97 Civ. 0222 (LAP).

    United States District Court, S.D. New York.

    June 20, 1997.

    Latham & Watkins, New York City by Michael K. Hertz, Anat Hakim; American [161] Civil Liberties Union New York City by Christopher A. Hansen, Ann Beeson; New York Civil Liberties Union, New York City by Arthur N. Eisenberg; Sonnenschein Nath & Rosenthal, New York City by Michael A. Bamberger, for Plaintiffs.

    Dennis C. Vacco, Attorney General of the State of New York, New York City by Jeanne Lahiff, James M. Hershler, for defendants.

    OPINION

    PRESKA, District Judge:

    The Internet may well be the premier technological innovation of the present age. Judges and legislators faced with adapting existing legal standards to the novel environment of cyberspace struggle with terms and concepts that the average American five-year-old tosses about with breezy familiarity.[1] Not surprisingly, much of the legal analysis of Internet-related issues has focused on seeking a familiar analogy for the unfamiliar. Commentators reporting on the recent oral argument before the Supreme Court of the United States, which is considering a First Amendment challenge to the Communications Decency Act, noted that the Justices seemed bent on finding the appropriate analogy which would tie the Internet to some existing line of First Amendment jurisprudence: is the Internet more like a television? a radio? a newspaper? a 900-line? a village green? See, e.g., Linda Greenhouse, What Level of Protection for Internet Speech? High Court Weighs Decency-Act Case, N.Y. Times, March 24, 1997, at C5; see also Denver Area Educ. Telecommunications Consortium v. Federal Communics. Comm'n, ___ U.S. ___, ___ - ___, 116 S.Ct. 2374, 2419-21, 135 L.Ed.2d 888 (1996) (Thomas, J., concurring in the judgment and dissenting in part) (criticizing the majority for declining to determine whether cable television is more closely analogous, for purposes of First Amendment analysis, to a print medium or a broadcast medium). This case, too, depends on the appropriate analogy. I find, as described more fully below, that the Internet is analogous to a highway or railroad. This determination means that the phrase "information superhighway" is more than a mere buzzword; it has legal significance, because the similarity between the Internet and more traditional instruments of interstate commerce leads to analysis under the Commerce Clause.

    BACKGROUND

    The plaintiffs in the present case filed this action challenging New York Penal Law § 235.21(3) (the "Act" or the "New York Act"), seeking declaratory and injunctive relief. Plaintiffs contend that the Act is unconstitutional both because it unduly burdens free speech in violation of the First Amendment and because it unduly burdens interstate commerce in violation of the Commerce Clause. Plaintiffs moved for a preliminary injunction enjoining enforcement of the Act; defendants opposed the motion. A factual hearing was held from April 3 to April 7, 1997 and oral argument conducted on April 22, 1997. For the reasons that follow, the motion for a preliminary injunction is granted.

    I. Parties to the Action

    Plaintiffs in the present action represent a spectrum of individuals and organizations who use the Internet to communicate, disseminate, display, and access a broad range of communications. All of the plaintiffs communicate online both within and outside the State of New York, and each plaintiff's communications are accessible from within and outside New York. Plaintiffs include:

    American Library Association, Freedom to Read Foundation, Inc., New York Library Association, and Westchester Library System are organizations representing the interests [162] of libraries. Libraries serve as both access and content providers on the Internet, providing their patrons with facilities to access the Internet. Libraries also post their card catalogues, information about upcoming events and online versions of text or art from their collections, as well as sponsoring chat rooms.

    American Booksellers Foundation For Free Expression ("ABFFE") is a national association of general interest and specialized bookstores formed to protect free expression rights. ABFFE has many members who use the Internet and electronic communications to obtain from publishers information and excerpts, some of which may contain sexually explicit passages.

    Association of American Publishers ("AAP") is a national association of publishers of general books, textbooks, and educational materials. AAP has many members who actively use and provide content on the Internet, both creating and posting electronic products and using the Internet as a communication and promotional tool for their print publishing activities.

    BiblioBytes is a private, profit-seeking enterprise that uses the World Wide Web (the "Web") to provide information about and to sell electronic books. BiblioBytes offers titles in a variety of genres, including romance, erotica, classics, adventure, and horror.

    Magazine Publishers of America ("MPA") is a national association of publishers of consumer magazines. MPA's members publish magazines in print form, but are also beginning to offer publications in electronic formats available to the public on the Internet or through online service providers.

    Interactive Digital Software Association ("IDSA") is a non-profit trade association of United States publishers of entertainment software. IDSA has many members who both sell their software in retail outlets and make their entertainment software available to the public on the Internet for demonstration, purchase, and play.

    Public Access Networks Corporation ("Panix") is an Internet service provider serving subscribers located in the New York area. Panix also hosts various organizational Web pages, assists its subscribers in creating individual Web pages, and hosts online discussion groups and chat rooms.

    ECHO is a for-profit Internet service provider that offers a "virtual salon" to Internet users. ECHO and its subscribers provide content on the Internet through the posting of Web sites, including personal home pages, and through over 50 discussion groups oriented to subscribers' interests.

    New York City Net ("NYC Net") is a for-profit Internet service provider catering primarily to lesbians and gay men in the New York area. NYC Net provides access services and content specifically oriented to gay and lesbian interests, including a large number of online discussion groups and chat rooms.

    Art on the Net is a non-profit organization with an international artist site ("art. net") on the Web. Art on the Net assists over 110 artists from all over the world in maintaining online studios.

    Peacefire is an organization whose membership consists primarily of minors. It was formed to protect the rights of citizens under the age of 18 to use the Internet. Peacefire's members use the Internet to communicate and access a wide variety of information. Peacefire's founder points out in his Declaration that Internet access is particularly important to those members who are too young to drive and might otherwise be unable to view materials from museums, libraries, and other institutions to which their families are unwilling to transport them. (See Declaration of Bennett Haselton, sworn to on March 12, 1997, at p. 4).

    American Civil Liberties Union ("ACLU") is a national civil rights organization. The ACLU maintains a Web site on which it posts civil liberties information and resources, including material about arts censorship, obscenity laws, discrimination against lesbians and gays, and reproductive choice. In addition, the ACLU hosts unmoderated online discussion groups that allow citizens to discuss and debate a variety of civil liberties issues.

    [163] Defendants in this case are the Governor and the Attorney General of New York. Defendants have raised the question of whether an injunction against those parties would also bind the sixty-two District Attorneys in New York who would actually be mounting prosecutions against alleged violators of the Act. Fed.R.Civ.P. 65(d) provides:

    Every order granting an injunction ... is binding only upon the parties to the action, their officers, agents, servants, employees, and attorneys, and upon those persons in active concert or participation with them who receive actual notice of the order by personal service or otherwise.

    Thus, parties such as the local District Attorneys who "participate" in the enjoined activities with defendants and who have actual notice of the injunction would be bound. See American Booksellers v. Webb, 590 F.Supp. 677, 693-94 (N.D.Ga.1984) (holding that an injunction against the Attorney General also binds state law enforcement officials who might seek to enforce the challenged Act); see also United Transportation Union v. Long Island RR Co., 634 F.2d 19, 21 (2d Cir.1980) (binding non-party Attorney General to the terms of an injunction against the defendants because Attorney General "undoubtedly had knowledge of the instant action and could have participated therein had he chosen to do so"), rev'd on other grounds, 455 U.S. 678, 102 S.Ct. 1349, 71 L.Ed.2d 547 (1982). Thus, a preliminary injunction would effectively bar enforcement of the Act whether the prosecution happened to be brought directly by the Attorney General's office or by one of the individual District Attorneys.

    II. The Challenged Statute

    The Act in question amended N.Y. Penal Law § 235.21 by adding a new subdivision. The amendment makes it a crime for an individual:

    Knowing the character and content of the communication which, in whole or in part, depicts actual or simulated nudity, sexual conduct or sado-masochistic abuse, and which is harmful to minors, [to] intentionally use[] any computer communication system allowing the input, output, examination or transfer, of computer data or computer programs from one computer to another, to initiate or engage in such communication with a person who is a minor.

    Violation of the Act is a Class E felony, punishable by one to four years of incarceration. The Act applies to both commercial and non-commercial disseminations of material.

    Section 235.20(6) defines "harmful to minors" as:

    that quality of any description or representation, in whatever form, of nudity, sexual conduct, sexual excitement, or sado-masochistic abuse, when it:

    (a) Considered as a whole, appeals to the prurient interest in sex of minors; and

    (b) Is patently offensive to prevailing standards in the adult community as a whole with respect to what is suitable material for minors; and

    (c) Considered as a whole, lacks serious literary, artistic, political and scientific value for minors.

    N.Y. Penal Law § 235.20(6).

    The statute provides six defenses to liability. First, Section 235.15(1) provides the following affirmative defense to prosecution under § 235.21(3):

    In any prosecution for obscenity, or disseminating indecent material to minors in the second degree in violation of subdivision three of section 235.21 of this article, it is an affirmative defense that the persons to whom the allegedly obscene or indecent material was disseminated, or the audience to an allegedly obscene performance, consisted of persons or institutions having scientific, educational, governmental or other similar justification for possessing, disseminating or viewing the same.

    The statute further provides four regular defenses to prosecution:

    (a) The defendant made a reasonable effort to ascertain the true age of the minor and was unable to do so as a result of the actions taken by the minor; or

    (b) The defendant has taken, in good faith, reasonable, effective and appropriate actions under the circumstances to restrict or prevent access by minors to materials [164] specified in such subdivision, which may involve any appropriate measures to restrict minors from access to such communications, including any method which is feasible under available technology; or

    (c) The defendant has restricted access to such materials by requiring use of a verified credit card, debit account, adult access code or adult personal identification number; or

    (d) The defendant has in good faith established a mechanism such that the labelling, segregation or other mechanism enables such material to be automatically blocked or screened by software or other capabilities reasonably available to responsible adults wishing to effect such blocking or screening and the defendant has not otherwise solicited minors not subject to such screening or blocking capabilities to access that material or circumvent any such screening or blocking.

    N.Y. Penal Law § 235.23(3). And, finally, Section 235.24 provides that no individual shall be held liable:

    [S]olely for providing access or connection to or from a facility, system, or network not under that person's control, including transmission, downloading, intermediate storage, access software, or other related capabilities that are incidental to providing such access or connection that do not include the creation of the content of the communication.

    N.Y. Penal Law § 235.24. Exceptions to this defense for conspirators or co-owners and an additional employer liability defense are set forth in Section 235.24(1)(a)-(b) and (2).

    III. The Internet[2]

    The Internet is a decentralized, global communications medium linking people, institutions, corporations, and governments all across the world. ACLU v. Reno, 929 F.Supp. 824 (E.D.Pa.), prob. juris. noted, ___ U.S. ___, 117 S.Ct. 554, 136 L.Ed.2d 436 (1996), argued, March 19, 1997; Shea v. Reno, 930 F.Supp. 916 (S.D.N.Y.1996), argued, March 19, 1997. The nature of the Internet makes it very difficult, if not impossible, to determine its size at any given moment. Undoubtedly, however, the Internet has experienced extraordinary growth in recent years. In 1981, fewer than 300 computers were linked to the Internet; in 1989, the number stood at fewer than 90,000 computers. By 1993, over 1,000,000 computers were linked. Today, over 9,400,000 host computers worldwide, 60% of them located in the United States, are linked to the Internet. This count does not include users who access the Internet via modem link-up from their personal computers. As many as 40 million people worldwide currently enjoy access to the Internet's rich variety of resources, and that number is expected to grow to 200 million by the year 1999.

    The Internet is a network of networks — a decentralized, self-maintaining series of redundant links among computers and computer networks, capable of rapidly transmitting communications without direct human involvement or control. No organization or entity controls the Internet; in fact, the chaotic, random structure of the Internet precludes any exercise of such control.

    The information available on the Internet is "as diverse as human thought," ACLU, 929 F.Supp. at 842. Every facet of art, literature, music, news, and debate is represented. There can be no question that the overwhelming variety of available information includes some sexually explicit materials. Sexually-oriented content is, however, not "the primary type of content on this new medium." Id.

    Individuals obtain access to the Internet via a number of avenues. Students and faculty often obtain access via their educational institutions; similarly, some corporations provide their employees with direct or modem access to the Internet. Individuals in some communities can access the Internet via a community network or a local library that provides direct or modem access to library patrons. Storefront "computer coffee shops" offer another option, serving up access to cyberspace accompanied by coffee [165] and snacks for a small hourly fee. "Internet service providers" typically offer modem telephone access to a computer or computer network linked to the Internet. Many such providers — including plaintiffs Panix, Echo, and NYC NET — are commercial entities offering Internet access for a monthly or hourly fee. Another common way for individuals to access the Internet is through one of the major national commercial "online services" such as America Online, Compuserve, the Microsoft Network, or Prodigy, which collectively service almost twelve million individual subscribers across the United States. These online services offer nationwide computer networks (allowing subscribers to dial in via a local telephone number) and provide both proprietary content and links to the even more extensive resources of the Internet for a monthly or hourly fee. Finally, local dial-in computer services, called "bulletin board systems" or "BBSs" provide Internet access via direct or indirect links.

    The Internet permits a user to communicate pictures and text in several ways including:

    (1) one-to-one messaging (such as "e-mail");

    (2) one-to-many messaging (such as "listserv" or "mail exploder");

    (3) distributed message databases (such as "USENET newsgroups");

    (4) real time remote computer utilization (such as "Internet Relay Chat");

    (5) real time remote computer utilization (such as "telnet"); and

    (6) remote information retrieval (such as "ftp," "gopher," and the Web).

    In addition to transmitting pictures and text, many of these communication methods can be used to transmit data, computer programs, sound, and moving video images.

    Most users of the Internet are provided with a username, password and e-mail address that allow them to sign on to the Internet and communicate with other users. Many usernames are pseudonyms, known as "handles," which provide users with a distinct online identity and preserve anonymity. For example, Ms. Kovacs testified that she uses the handle "Harriet Vane" when communicating with fellow mystery aficionados in the "Dorothy L" listserv and the nom de cyber "Mrs. Archangel" when she's just "goofing off" on the Internet. (4/4/97 Tr., at 58). The username and e-mail address are the only indicators of a user's identity; generally speaking, neither datum discloses a party's age or geographic location.

    E-mail is the simplest method of Internet communication. E-mail allows an online user to address and transmit an electronic message to one or more people. The ACLU court noted that e-mail is "comparable in principle to sending a first class letter." ACLU, 929 F.Supp. at 834. The analogy is not a perfect one, however, for two reasons. First, the sender directs his message to a logical rather than geographic address, and therefore need not know the location of his correspondent in real space. Second, most programs provide for a "reply" option which enables the recipient to respond to the sender's message simply by clicking on a button; the recipient will therefore not even need to type in the sender's e-mail address. A further distinction concerns the level of security that protects a communication. While firstclass letters are sealed, e-mail communications are more easily intercepted. Concerns about the relatively easy accessibility of e-mail communications have led bar associations in some states to require that lawyers encrypt sensitive e-mail messages in order to protect client confidentiality. See Carey Ramos & Curtis Carmack, Beware of Cyberspace Marauders: Internet Security Addressed, N.Y.L.J., February 24, 1997, at S1.

    The Internet also includes a wide variety of online discussion fora that allow groups of users to discuss and debate subjects of interest. The three most common means by which such discussion groups come together are through mail exploders, USENET newsgroups, and chat rooms.

    Mail exploders, also known as "listservs," allow online users to subscribe to automated mailing lists that disseminate information on particular subjects. Subscribers send an e-mail message to the "list," and the mail exploder automatically and simultaneously sends the message to all of the other subscribers on the list. Users of mailing lists [166] can add or delete their names from the list automatically, without any direct human involvement. Id. at 834; Shea, 930 F.Supp. at 927.

    USENET newsgroups are a very popular set of discussion groups arranged according to subject matter and automatically disseminated "using ad hoc peer to peer connections between approximately 200,000 computers ... around the world." ACLU, 929 F.Supp. at 834-35. Users may read or send messages to newsgroups without any prior subscription, and there is no way for a speaker who posts an article to a newsgroup to know who is reading the message. Id.; Shea, 930 F.Supp. at 927-28. Currently, more than 15,000 different subjects are represented in USENET newsgroups, and over 100,000 new messages are posted to these groups every day. ACLU, 929 F.Supp. at 835.

    Chat rooms allow online discussion in real time. Users are able to engage in simultaneous conversations with one or many "occupants" by typing in messages and reading the messages typed by others participating in the chat; the ACLU court analogized this Internet application to a telephone party line. ACLU, 929 F.Supp. at 835; Shea, 930 F.Supp. at 928. There are thousands of different chat rooms available "in which collectively tens of thousands of users are engaging in conversations on a huge range of subjects." ACLU, 929 F.Supp. at 835.

    Finally, perhaps the most well-known method of communicating information online is the Web; many laypeople erroneously believe that the Internet is co-extensive with the Web. The Web is really a publishing forum; it is comprised of millions of separate "Web sites" that display content provided by particular persons or organizations. Any Internet user anywhere in the world with the proper software can create a Web page, view Web pages posted by others, and then read text, look at images and video, and listen to sounds posted at these sites. Many large corporations, banks, brokerage houses, newspapers and magazines provide online editions of their reports and publications or operate independent Web sites. Government agencies and even courts use the Web to disseminate information to the public. At the same time, many individual users and small community organizations have established individual "home pages" on the Web that provide information to any interested person who "surfs by."

    Although information on the Web is contained on innumerable Web sites located on individual computers around the world, each of these Web sites and computers is connected to the Internet by means of protocols that permit the information to become part of a single body of knowledge accessible by all Web visitors. ACLU, 929 F.Supp. at 836, 837. To gain access to the resources of the Web, an individual employs a "browser." A browser is software, such as Netscape Navigator, Mosaic, or Internet Explorer, that allows the user to display, print, and download documents that are formatted in the standard Web formatting language. Shea, 930 F.Supp. at 929.

    There are a number of different ways that Internet users can browse or search for content on the Web. First, every document on the Web has an address that allows users to find and retrieve it, and a user can simply type in the address and go directly to that site. Again, however, the address is a logical rather than geographic concept, and the user will not necessarily know where the site is located in real space. Additionally, a user who wants to conduct a generalized search or wants to reach a particular site but does not know the address, can use a "search engine," which is available free of charge to help users navigate the Web. ACLU, 929 F.Supp. at 837. The user simply types a word or string of words as a search request, and the search engine provides a list of sites that match the search string. Id.

    Finally, online users may "surf" the Web by "linking" from one Web page to another. Almost all Web documents contain "links," segments of text or images that refer to another Web document. Id. at 836. When the user clicks on the link, the linked document is automatically displayed, wherever in the world it is stored. Id. For example, the American Library Association ("ALA") home page contains several links. Some of these links are to other Web pages or documents within the ALA site, including documents [167] entitled "Libraries Online," "Library Promotional Events," and the "ALA Bookstore." Other links from the ALA home page connect the user to sites maintained by other organizations or individuals and stored on other computers around the world. The ALA Web site, for example, provides links to the American Association of Law Libraries, the Art Libraries Society of North America, and the Medical Library Association. "These links from one computer to another, from one document to another across the Internet, are what unify the Web into a single body of knowledge, and what makes the Web unique." Id. at 836-37.

    Regardless of the aspect of the Internet they are using, Internet users have no way to determine the characteristics of their audience that are salient under the New York Act — age and geographic location. In fact, in online communications through newsgroups, mailing lists, chat rooms, and the Web, the user has no way to determine with certainty that any particular person has accessed the user's speech. "Once a provider posts content on the Internet, it is available to all other Internet users worldwide." Id. at 844. A speaker thus has no way of knowing the location of the recipient of his or her communication. As the poet said, "I shot an arrow into the air; it fell to the earth I know not where."

    This highly simplified description of the Internet is not intended to minimize its marvels. While no one should lose sight of the inventiveness that has made this complex of resources available to just about anyone, the innovativeness of the technology does not preclude the application of traditional legal principles — provided that those principles are adaptable to cyberspace. In the present case, as discussed more fully below, the Internet fits easily within the parameters of interests traditionally protected by the Commerce Clause. The New York Act represents an unconstitutional intrusion into interstate commerce; plaintiffs are therefore entitled to the preliminary injunction that they seek.

    DISCUSSION

    I. Standard Applicable to a Preliminary Injunction

    To demonstrate their entitlement to a preliminary injunction, plaintiffs must show (a) that they will suffer irreparable harm and (b) either (i) a likelihood of success on the merits or (ii) sufficiently serious questions going to the merits to make them a fair ground for litigation[3] and a balance of hardships tipping decidedly in the plaintiffs' favor. Paulsen v. County of Nassau, 925 F.2d 65, 68 (2d Cir.1991); Streetwatch v. National R.R. Passenger Corp., 875 F.Supp. 1055, 1058 (S.D.N.Y.1995). In the present case, as discussed more fully below, plaintiffs have amply demonstrated the likelihood of their successful prosecution of their claim that the Act violates the Commerce Clause because it seeks to regulate communications occurring wholly outside New York, imposes a burden on interstate commerce that is disproportionate to the local benefits it is likely to engender, and subjects plaintiffs, as well as other Internet users, to inconsistent state obligations. See Healy v. Beer Institute, 491 U.S. 324, 332, 109 S.Ct. 2491, 2497, 105 L.Ed.2d 275 (1989); Pike v. Bruce Church, Inc., 397 U.S. 137, 142, 90 S.Ct. 844, 847, 25 L.Ed.2d 174 (1970); Southern Pac. Co. v. Arizona ex rel. Sullivan, 325 U.S. 761, 767, 65 S.Ct. 1515, 1519, 89 L.Ed. 1915 (1945).

    Plaintiffs have also shown that they face irreparable injury in the absence of an injunction. Irreparable injury means "the kind of injury for which money cannot compensate," Sperry Int'l Trade, Inc. v. Government of Israel, 670 F.2d 8, 12 (2d Cir.1982), [168] and which is "neither remote nor speculative, but actual and imminent." Tucker Anthony Realty Corp. v. Schlesinger, 888 F.2d 969, 975 (2d Cir.1989). Deprivation of the rights guaranteed under the Commerce Clause constitutes irreparable injury. C & A Carbone, Inc. v. Town of Clarkstown, 770 F.Supp. 848, 854 (S.D.N.Y.1991) (holding that a local waste disposal law caused irreparable injury to the plaintiffs' rights under the Commerce Clause). Thus, by demonstrating that the Act threatens their rights under the Commerce Clause, as will be discussed more fully below, the plaintiffs have shown both irreparable injury and a likelihood of success on the merits.

    II. Federalism and the Internet: The Commerce Clause

    The borderless world of the Internet raises profound questions concerning the relationship among the several states and the relationship of the federal government to each state, questions that go to the heart of "our federalism." See Younger v. Harris, 401 U.S. 37, 44, 91 S.Ct. 746, 750-51, 27 L.Ed.2d 669 (1971) ("[O]ne familiar with the profound debates that ushered our Federal Constitution into existence is bound to respect those who remain loyal to the ideals and dreams of `Our Federalism.' The concept does not mean blind deference to `States' Rights' any more than it means centralization of control over every important issue in our National Government and its courts. The Framers rejected both these courses.") The Act at issue in the present case is only one of many efforts by state legislators to control the chaotic environment of the Internet. For example, the Georgia legislature has enacted a recent law prohibiting Internet users from "falsely identifying" themselves online. Ga. Stat. 16-9-9.1. Similar legislation is pending in California. California Senate Bill SB-1533 (1996); see also Ilana DeBare, State Trademark Bill Ignites Net Turmoil, The Sacramento Bee, March 2, 1991, at F1. Texas and Florida have concluded that law firm web pages (apparently including those of out of state firms) are subject to the rules of professional conduct applicable to attorney advertising. See Texas Bar Advertising Comm., Interpretive Comment on Attorney Internet Advertising (1996); see also Texans Against Censorship v. State Bar of Texas, 888 F.Supp. 1328, 1369-70 (E.D.Tex.1995) (discussing applicability of Texas lawyers advertising regulation to the Internet), aff'd, 100 F.3d 953 (5th Cir.1996); Ethics Update, Fla. Bar News, January 1, 1996. Further, states have adopted widely varying approaches in the application of general laws to communications taking place over the Internet. Minnesota has aggressively pursued out-of-state advertisers and service providers who reach Minnesotans via the Internet; Illinois has also been assertive in using existing laws to reach out-of-state actors whose connection to Illinois occurs only by virtue of an Internet communication. See Mark Eckenwiler, States Get Entangled in the Web, Legal Times, Jan. 22, 1996, at § 35, § 37. Florida has taken the opposite route, declining to venture into online law enforcement until various legal issues (including, perhaps, the one discussed in the present opinion) have been determined. Id. at § 37.[4]

    The unique nature of the Internet highlights the likelihood that a single actor might be subject to haphazard, uncoordinated, and even outright inconsistent regulation by states that the actor never intended to reach and possibly was unaware were being accessed. [169] Typically, states' jurisdictional limits are related to geography; geography, however, is a virtually meaningless construct on the Internet. The menace of inconsistent state regulation invites analysis under the Commerce Clause of the Constitution, because that clause represented the framers' reaction to overreaching by the individual states that might jeopardize the growth of the nation — and in particular, the national infrastructure of communications and trade — as a whole. See Quill Corp. v. North Dakota, 504 U.S. 298, 312, 112 S.Ct. 1904, 1913, 119 L.Ed.2d 91 (1992) ("Under the Articles of Confederation, state taxes and duties hindered and suppressed interstate commerce; the Framers intended the Commerce Clause as a cure for these structural ills."); see also The Federalist Nos. 7, 11 (A.Hamilton).

    The Commerce Clause is more than an affirmative grant of power to Congress. As long ago as 1824, Justice Johnson in his concurring opinion in Gibbons v. Ogden, 9 Wheat. 1, 231-32, 239, 6 L.Ed. 23 (1824), recognized that the Commerce Clause has a negative sweep as well. In what commentators have come to term its negative or "dormant" aspect, the Commerce Clause restricts the individual states' interference with the flow of interstate commerce in two ways. The Clause prohibits discrimination aimed directly at interstate commerce, see, e.g., Philadelphia v. New Jersey, 437 U.S. 617, 98 S.Ct. 2531, 57 L.Ed.2d 475 (1978), and bars state regulations that, although facially non-discriminatory, unduly burden interstate commerce, see, e.g., Kassel v. Consolidated Freightways Corp. of Del., 450 U.S. 662, 101 S.Ct. 1309, 67 L.Ed.2d 580 (1981). Moreover, courts have long held that state regulation of those aspects of commerce that by their unique nature demand cohesive national treatment is offensive to the Commerce Clause. See, e.g., Wabash, St. L. & P. Ry. Co. v. Illinois, 118 U.S. 557, 7 S.Ct. 4, 30 L.Ed. 244 (1886) (holding railroad rates exempt from state regulation).

    Thus, as will be discussed in more detail below, the New York Act is concerned with interstate commerce and contravenes the Commerce Clause for three reasons. First, the Act represents an unconstitutional projection of New York law into conduct that occurs wholly outside New York. Second, the Act is invalid because although protecting children from indecent material is a legitimate and indisputably worthy subject of state legislation, the burdens on interstate commerce resulting from the Act clearly exceed any local benefit derived from it. Finally, the Internet is one of those areas of commerce that must be marked off as a national preserve to protect users from inconsistent legislation that, taken to its most extreme, could paralyze development of the Internet altogether. Thus, the Commerce Clause ordains that only Congress can legislate in this area, subject, of course, to whatever limitations other provisions of the Constitution (such as the First Amendment) may require.

    A. The Act Concerns Interstate Commerce

    At oral argument, the defendants advanced the theory that the Act is aimed solely at intrastate conduct. This argument is unsupportable in light of the text of the statute itself, its legislative history, and the reality of Internet communications. The section in question contains no such limitation; it reads:

    A person is guilty of disseminating indecent material to minors in the second degree when: ...

    (3) Knowing the character and content of the communication which, in whole or in part, depicts actual or simulated nudity, sexual conduct or sado-masochistic abuse, and which is harmful to minors, he intentionally uses any computer communication system allowing the input, output, examination or transfer, of computer data or computer programs from one computer to another, to initiate or engage in such communication with a person who is a minor.

    N.Y. Penal Law § 235.21(3) (McKinney's 1997). Section 235.20, which contains the definitions applicable to the challenged portion of the Act, does not import any restriction that the criminal communication must take place entirely within the State of New York. By its terms, the Act applies to any communication, intrastate or interstate, that [170] fits within the prohibition and over which New York has the capacity to exercise criminal jurisdiction. See Boyd v. Meachum, 77 F.3d 60, 65 (2d Cir.1996) (holding that a criminal court "has personal jurisdiction over any party who appears before it, regardless of how his appearance was obtained"), cert. denied, ___ U.S. ___, 117 S.Ct. 114, 136 L.Ed.2d 66 (1996); see also United States v. Lussier, 929 F.2d 25, 27 (1st Cir.1991); United States v. Stuart, 689 F.2d 759, 762 (8th Cir.1982), cert. denied, 460 U.S. 1037, 103 S.Ct. 1427, 75 L.Ed.2d 788 (1983).

    Further, the legislative history of the Act clearly evidences the legislators' understanding and intent that the Act would apply to communications between New Yorkers and parties outside the State, despite occasional glib references to the Act's "intrastate" applicability. The New York State Senate Introducer's Memorandum in Support of the Act contains a paragraph under the subtitle, "Justification," which states:

    Law enforcement agencies around the nation are becoming increasingly alarmed at the growing use of computer networks and other communications by pedophiles. As one observer noted, "perverts are moving from the playground to the internet." Several cases have come to light wherein a pedophile has traveled clear across the country to have sexual relations with a minor initially contacted and engaged through various computer networks.

    (Affidavit of James Hershler, Exh. D) (emphasis added). A letter from the Bill's sponsor to Governor Pataki characterized sexually-infused Internet communications between adults and minors as "long-distance, hightech sexual abuse." (See Letter dated July 11, 1996 from William Sears to Governor Pataki, designated page 3 in the Bill Jacket, Hershler Aff., Exh. A). Jeanine Pirro, the Westchester County District Attorney, wrote a letter to Governor Pataki dated February 13, 1996 that similarly reflects the expectations of the Act's proponents that it would apply to interstate communications. Ms. Pirro's letter states:

    This bill was proposed partly in response to a Westchester County case wherein an adult male resident of Seattle, Washington, [one Alan Paul Barlow,] communicated about sexually explicit matters by computer with a thirteen year old girl over several months.

    (Hershler Aff., Exh. F); see also John Heileman, The Crusader, The New Yorker, February 24 and March 3, 1997 (detailing Ms. Pirro's "crusade" to achieve the passage of the Act in the aftermath of the Barlow incident).[5] Ms. Pirro's references to this incident, known as the Barlow case, are echoed throughout defendants' memorandum of law. (See Defendants' Memorandum of Law in Opposition to Preliminary Injunction, pp. 15, 16, 17-18). Obviously, however, the Act would be completely ineffective in forestalling a pedophile like Barlow if it applied only to purely intrastate communications.

    The conclusion that the Act must apply to interstate as well as intrastate communications receives perhaps its strongest support from the nature of the Internet itself. The Internet is wholly insensitive to geographic distinctions. In almost every case, users of the Internet neither know nor care about the physical location of the Internet resources they access. Internet protocols were designed to ignore rather than document geographic location; while computers on the network do have "addresses," they are logical addresses on the network rather than geographic addresses in real space. The majority of Internet addresses contain no geographic clues and, even where an Internet address provides such a clue, it may be misleading. [171] For example, in his article, Federalism in Cyberspace, 28 Conn. L.Rev. 1095, 1112 (1996), Professor Dan Burk described how he uses Seton Hall University's computer system to access the Internet, providing anyone who communicates with him (and is aware of Seton Hall's locale) a hint that he is in New Jersey. However, Professor Burk also has a guest account at a university in California which he continues to use even when he is in New Jersey; any clue derived from the California university's name within the Internet address would therefore be deceptive. In a similar vein, Ms. Kovacs testified that as she was using her computer to give an in-court demonstration of various Internet applications, she received an e-mail from a colleague who believed she was sending the message to Cincinnati, Ohio (where Ms. Kovacs is normally located); in fact, Ms. Kovacs was in New York and received the message here. (4/4/97 Tr., p. 61).

    Moreover, no aspect of the Internet can feasibly be closed off to users from another state. An internet user who posts a Web page cannot prevent New Yorkers or Oklahomans or Iowans from accessing that page and will not even know from what state visitors to that site hail. Nor can a participant in a chat room prevent other participants from a particular state from joining the conversation. Someone who uses a mail exploder is similarly unaware of the precise contours of the mailing list that will ultimately determine the recipients of his or her message, because users can add or remove their names from a mailing list automatically. Thus, a person could choose a list believed not to include any New Yorkers, but an after-added New Yorker would still receive the message.[6]

    E-mail, because it is a one-to-one messaging system, stands on a slightly different footing than the other aspects of the Internet. Even in the context of e-mail, however, a message from one New Yorker to another New Yorker may well pass through a number of states en route. The Internet is, as described above, a redundant series of linked computers. Thus, a message from an Internet user sitting at a computer in New York may travel via one or more other states before reaching a recipient who is also sitting at a terminal in New York.

    The system is further complicated by two Internet practices: packet switching and caching. "Packet switching" protocols subdivide individual messages into smaller packets that are then sent independently to the destination, where they are automatically reassembled by the receiving computer. If computers along the route become overloaded, packets may be rerouted to computers with greater capacity. A single message may — but does not always — travel several different pathways before reaching the receiving computer. "Caching" is the Internet practice of storing partial or complete duplicates of materials from frequently accessed sites to avoid repeatedly requesting copies from the original server. The recipient has no means of distinguishing between the cached materials and the original. Thus, the user may be accessing materials at the original site, or he may be accessing copies of those materials cached on a different machine located anywhere in the world.

    The New York Act, therefore, cannot effectively be limited to purely intrastate communications over the Internet because no such communications exist. No user could reliably restrict her communications only to New York recipients. Moreover, no user could avoid liability under the New York Act simply by directing his or her communications elsewhere, given that there is no feasible way to preclude New Yorkers from accessing a Web site, receiving a mail exploder message or a newsgroup posting, or participating in a chat room. Similarly, a user has no way to ensure that an e-mail does not pass through New York even if the ultimate recipient is not located there, or that a message never leaves New York even if both sender and recipient are located there.

    This conclusion receives further support from the unchallenged testimony that plaintiff's introduced in the form of declarations. [172] For example, Stacy Horn, the president of ECHO, an electronic cultural salon, testified that "[c]onference participants do not know, and have no way to determine, the ... geographic location of other participants." (Decl. of Stacy Horn, sworn to on March 12, 1997, at p. 6). Oren Teicher, the President of the American Booksellers Foundation for Free Expression, indicated that:

    Much of the Internet use by booksellers is interstate in nature. For example, any bookseller's Web page can be accessed by Internet users not only throughout the United States, but throughout the world. Similarly, ABFFE members from across the country communicate with one another as well as Internet users across the country via e-mail. Moreover, ABFFE users cannot effectively prevent their Web sites or discussion groups from being accessed by New York users.

    (Decl. of Oren Teicher, sworn to on March 26, 1997, at p. 4). Lawrence J. Kaufman, the Vice President of the Magazine Publishers of America, Inc., a trade association for the consumer magazine industry, noted that "Online users anywhere in the world can access the content provided by MPA members on the Web and via e-mail. These members cannot effectively prevent their Web sites from being accessed by New York users." (Decl. of Lawrence J. Kaufman, sworn to on March 26, 1997, at p. 2).

    The Act is therefore necessarily concerned with interstate communications. See Virginia v. American Booksellers Ass'n, Inc., 484 U.S. 383, 397, 108 S.Ct. 636, 645, 98 L.Ed.2d 782 (1988) (holding that only if a statute is "readily susceptible" to a narrowing construction will the court apply such a construction to save an otherwise unconstitutional law). The next question that requires an answer as a threshold matter is whether the types of communication involved constitute "commerce" within the meaning of the Clause.

    The definition of commerce in the Supreme Court's decisions has been notably broad. Most recently, in Camps Newfound Owatonna, Inc. v. Town of Harrison, Maine, ___ U.S. ___, 117 S.Ct. 1590, 137 L.Ed.2d 852 (1997), the Court rejected defendant's arguments that the Commerce Clause was inapplicable to a discriminatory real estate tax deduction, either because "campers are not `articles of commerce'" or because the plaintiff camp's "product is delivered and `consumed' entirely within Maine." Id. at ___, 117 S.Ct. at 1596. In the past, the Court has held that interstate commerce is affected by private race discrimination that limited access to a hotel and thereby impeded interstate commerce in the form of travel. Heart of Atlanta Motel, Inc. v. United States, 379 U.S. 241, 244, 258, 85 S.Ct. 348, 351, 358, 13 L.Ed.2d 258 (1964).

    In the present case, the parties have stipulated that:

    The Internet is not exclusively, or even primarily, a means of commercial communication. Many commercial entities maintain Web sites to inform potential consumers about their goods and services, or to solicit purchases, but many other Web sites exist solely for the dissemination of non-commercial information. The other forms of Internet communication — e-mail, bulletin boards, newsgroups, and chat rooms — frequently have non-commercial goals. For the economic and technical reasons set forth in the following paragraphs, the Internet is an especially attractive means for not-for-profit entities or public interest groups to reach their desired audiences. There are examples in the plaintiffs' affidavits of some of the noncommercial uses that the Internet serves. Plaintiff Peacefire offers information on its Internet site regarding the rights of minors on the Internet. Plaintiff Art on the Net allows artists to post their works on the World Wide Web. Plaintiff American Civil Liberties Union offers information on civil liberties issues.

    (Joint Stipulation of Facts, ¶ 79). This stipulation, however inartfully worded, cannot insulate the statute at issue from Commerce Clause scrutiny. The non-profit nature of certain entities that use the Internet or of certain transactions that take place over the Internet does not take the Internet outside the Commerce Clause. See Camps Newfound, at ___, 117 S.Ct. at 1597; Hughes v. Oklahoma, 441 U.S. 322, 326 n. 2, 99 S.Ct. [173] 1727, 1731 n. 2, 60 L.Ed.2d 250 (1979); Philadelphia v. New Jersey, 437 U.S. 617, 621-23, 98 S.Ct. 2531, 2534-35, 57 L.Ed.2d 475 (1978).

    The Supreme Court has expressly held that the dormant commerce clause is applicable to activities undertaken without a profit motive. In Edwards v. California, 314 U.S. 160, 62 S.Ct. 164, 86 L.Ed. 119 (1941), the Court examined the constitutionality of a California statute prohibiting the transport of indigent people into the state. The Court struck the statute as violative of the dormant Commerce Clause, reasoning that "the transportation of persons is `commerce,'" and that the California law at issue raised an "unconstitutional barrier to that commerce." Id. at 172-73, 62 S.Ct. at 166. In making its threshold determination, the Court emphasized that "[i]t is immaterial whether or not the transportation is commercial in character." Id. at 172, n. 1, 62 S.Ct. at 166, n. 1; see also Caminetti v. United States, 242 U.S. 470, 491, 37 S.Ct. 192, 196-97, 61 L.Ed. 442 (1917); Hoke v. United States, 227 U.S. 308, 320, 33 S.Ct. 281, 283, 57 L.Ed. 523 (1913).

    Commercial use of the Internet, moreover, is a growing phenomenon. See, e.g., Don Clark, Disney Launching Children's Web Site Only on Microsoft's On-Line Service, Wall St. Journal, March 31, 1997 (describing Disney's efforts to create and market a fee-based Web service); see also Andrew Bowser, Advertising on the Net, New Orleans Citybusiness, March 6, 1995; John Casey, Growing Potential of World Wide Web, Business & Finance, The Irish Times, June 3, 1996. In addition, many of those users who are communicating for private, noncommercial purposes are nonetheless participants in interstate commerce by virtue of their Internet consumption. Many users obtain access to the Internet by means of an on-line service provider, such as America Online, which charges a fee for its services. "Internet service providers," including plaintiffs Panix, Echo, and NYC NET, also offer Internet access for a monthly or hourly fee. Patrons of storefront "computer coffee shops," such as New York's own CyberCafe, similarly pay for their access to the Internet, in addition to partaking of food and beverages sold by the cafe. Dial-in bulletin board systems often charge a fee for access. See Katzenbach v. McClung, 379 U.S. 294, 300-01, 85 S.Ct. 377, 381-82, 13 L.Ed.2d 290 (1964) (holding that an entity that purchases goods used in the provision of its services from interstate sources is an actor in interstate commerce even in connection with the provision of services within a single state).

    The courts have long recognized that railroads, trucks, and highways are themselves "instruments of commerce," because they serve as conduits for the transport of products and services. See Kassel v. Consolidated Freightways Corp., 450 U.S. 662, 101 S.Ct. 1309, 67 L.Ed.2d 580 (1981); Southern Pacific Co. v. Arizona, 325 U.S. 761, 780, 65 S.Ct. 1515, 1525-26, 89 L.Ed. 1915 (1945). The Internet is more than a means of communication; it also serves as a conduit for transporting digitized goods, including software, data, music, graphics, and videos which can be downloaded from the provider's site to the Internet user's computer. For example, plaintiff BiblioBytes and members of plaintiff IDSA both sell and deliver their products over the Internet.

    The inescapable conclusion is that the Internet represents an instrument of interstate commerce, albeit an innovative one; the novelty of the technology should not obscure the fact that regulation of the Internet impels traditional Commerce Clause considerations. The New York Act is therefore closely concerned with interstate commerce, and scrutiny of the Act under the Commerce Clause is entirely appropriate. As discussed in the following sections, the Act cannot survive such scrutiny, because it places an undue burden on interstate traffic, whether that traffic be in goods, services, or ideas.

    B. New York Has Overreached by Enacting a Law That Seeks To Regulate Conduct Occurring Outside its Borders

    The interdiction against direct interference with interstate commerce by state legislative overreaching is apparent in a number of the Supreme Court's decisions. In Baldwin v. G.A.F. Seelig, Inc., 294 U.S. 511, 521, 55 S.Ct. 497, 499-500, 79 L.Ed. 1032 (1935), for example, Justice Cardozo authored an opinion [174] enjoining enforcement of a law that prohibited a dealer from selling within New York milk purchased from the producer in Vermont at less than the minimum price fixed for milk produced in New York. Justice Cardozo sternly admonished, "New York has no power to project its legislation into Vermont by regulating the price to be paid in that state for milk," finding that "[s]uch a power, if exerted, [would] set a barrier to traffic between one state and another as effective as if customs duties, equal to the price differential, had been laid upon the thing transported." Id.

    The Court has more recently confirmed that the Commerce Clause precludes a state from enacting legislation that has the practical effect of exporting that state's domestic policies. In Edgar v. MITE, 457 U.S. 624, 102 S.Ct. 2629, 73 L.Ed.2d 269 (1982), the Court examined the constitutionality of an Illinois anti-takeover statute that required a tender offeror to notify the Secretary of State and the target company of its intent to make a tender offer and the terms of the offer 20 days before the offer became effective. During the twenty-day period, the offeror was barred from communicating its offer to the shareholders, but the target company was free to disseminate information to its shareholders concerning the impending offer. Id. at 633, 102 S.Ct. at 2636. The statute defined "target company" as a corporation of which Illinois shareholders own 10% of the class of securities subject to the takeover offer, or for which any two of the following conditions are met: the corporation has its principal office in Illinois, is organized under Illinois law, or has at least 10% of its stated capital and paid-in surplus within Illinois. Id. at 625, 102 S.Ct. at 2632. The Court acknowledged that states traditionally retained the power to regulate intrastate securities transactions by enacting "blue-sky laws." Id. at 641, 102 S.Ct. at 2640. Nonetheless, the Court asserted that "[t]he Illinois Act differs substantially from state blue-sky laws in that it directly regulates transactions which take place across state lines, even if wholly outside the State of Illinois." Id. In striking the law as violative of the Commerce Clause, the Court found particularly egregious the fact that the Illinois law on its face would apply to a transaction that would not affect a single Illinois shareholder if a corporation fit within the definition of a "target company." Id. at 642, 102 S.Ct. at 2640-41. The Court concluded "the Illinois statute is a direct restraint on interstate commerce and has a sweeping extraterritorial effect," because the statute would prevent a tender offeror from communicating its offer to shareholders both within and outside Illinois. Acceptance of the offer by any of the shareholders would result in interstate transactions; the Illinois statute effectively stifled such transactions during the waiting period and thereby disrupted prospective interstate commerce. Under the Commerce Clause, the projection of these extraterritorial "practical effect[s]," regardless of the legislators' intentions, "`exceeded the inherent limits of the State's power.'" Id. at 642-43, 102 S.Ct. at 2641 (quoting Shaffer v. Heitner, 433 U.S. 186, 197, 97 S.Ct. 2569, 2576, 53 L.Ed.2d 683 (1977)).

    In the present case, a number of witnesses testified to the chill that they felt as a result of the enactment of the New York statute; these witnesses refrained from engaging in particular types of interstate commerce. In particular, I note the testimony of Rudolf Kinsky, an artist with a virtual studio on Art on the Net's Website. Mr. Kinsky testified that he removed several images from his virtual studio because he feared prosecution under the New York Act. (4/7/97 Tr., at 231-35). As described above, no Web siteholder is able to close his site to New Yorkers. Thus, even if Mr. Kinsky were located in California and wanted to display his work to a prospective purchaser in Oregon, he could not employ his virtual studio to do so without risking prosecution under the New York law.

    Oren Teicher, the President of the American Booksellers Foundation for Free Expression, similarly testified to the stifling effects that the Act will have on prospective interstate commerce in books, stating that:

    The Internet is an important source of interstate business for ABFFE members ... [B]ooksellers conduct business over the Internet in a variety of ways. If the Act is not enjoined and ABFFE members are forced to self-censor rather than be [175] subject to criminal liability, they will suffer immeasurable injury because they will lose significant sales and goodwill generated by their use of the Internet with respect to both censored and noncensored materials and resources. If a bookstore must self-censor certain books, it loses the profits from the sale of those particular books generated from the books' listing on the booksellers' Web sites. In addition, the bookstore will lose even more business because it will appear that the bookstore has an incomplete or inadequate listing of books in its inventory and Internet users will choose to buy their books elsewhere.

    (Teicher Decl., pp. 4-5). Lawrence Kaufman, the Vice President of the Magazine Publishers of America, also testified to the interstate nature of the business conducted by MPA over the Internet and to the loss of sales and goodwill that MPA members will suffer if forced to self-censor in order to avoid criminal liability under the Act. In particular, Mr. Kaufman noted that Playboy magazine, an MPA member, occasionally posts electronic versions or excerpts from its magazines that might fall within the Act's prohibition, presumably in an effort to attract new readership and subscribers. (Kaufman Decl. pp. 2-3). Edgar teaches that for New York to attempt to strangle prospective interstate transactions between parties from states other than New York by this means offends the Commerce Clause.

    The "extraterritoriality" analysis of the Edgar opinion commanded only a plurality of the Court. Later majority holdings, however, expressly adopted the underlying principles on which Justice White relied in Edgar. See Healy v. The Beer Institute, 491 U.S. 324, 109 S.Ct. 2491, 105 L.Ed.2d 275 (1989); Brown-Forman Distillers Corp. v. New York State Liquor Authority, 476 U.S. 573, 106 S.Ct. 2080, 90 L.Ed.2d 552 (1986). In Healy, the Court assessed the constitutionality of a Connecticut statute that required that out-of-state beer shippers affirm that their prices were no higher than the prices being charged in the bordering states at the time of the affirmation. The Court derived three guiding principles from its prior cases. First, the Court emphasized that the "Commerce Clause ... precludes the application of a state statute to commerce that takes place wholly outside the State's borders, whether or not the commerce has effects within the state."' Healy, 491 U.S. at 336, 109 S.Ct. at 2499 (citing Edgar, 457 U.S. at 642-43, 102 S.Ct. at 2640-41; Brown-Forman, 476 U.S. at 581-83, 106 S.Ct. at 2085-87). Second, the Court instructed that "a statute that directly controls commerce occurring wholly outside the boundaries of a State exceeds the inherent limits of the enacting State's authority and is invalid regardless of whether the statute's extraterritorial reach was intended by the legislature. The critical inquiry is whether the practical effect of the regulation is to control conduct beyond the boundaries of the State." Id. (citing Brown-Forman, 476 U.S. at 579, 106 S.Ct. at 2084). Finally, "the practical effect of the statute must be evaluated not only by considering the consequences of the statute itself, but also by considering how the challenged statute may interact with the legitimate regulatory regimes of other States and what effect would arise if not one, but many or every, State adopted similar legislation. Generally speaking, the Commerce Clause protects against inconsistent legislation arising from the projection of one state regulatory regime into the jurisdiction of another State." Id.; cf. CTS Corp. v. Dynamics Corp. of America, 481 U.S. 69, 88-89, 107 S.Ct. 1637, 1649-50, 95 L.Ed.2d 67 (1987).

    Applying these principles to the Connecticut price-affirmation statute, the Court held that the statute had the undeniable and impermissible effect of controlling commercial activity occurring wholly outside Connecticut. In particular, the Court examined the practical impact of the statute, in light of the regulations prevailing in the neighboring states of Massachusetts and New York and determined that the affirmation law, when taken in conjunction with the laws that had been or might be enacted in neighboring states, created "just the kind of competing and interlocking local economic regulation that the Commerce Clause was meant to preclude." Healy, 491 U.S. at 337, 109 S.Ct. at 2500.

    The Edgar/Healy extraterritoriality analysis rests on the premise that the Commerce [176] Clause has two aspects: it subordinates each state's authority over interstate commerce to the federal power of regulation (a vertical limitation), and it embodies a principle of comity that mandates that one state not expand its regulatory powers in a manner that encroaches upon the sovereignty of its fellow states (a horizontal limitation). The Court most recently recognized this duality in BMW of North America, Inc. v. Gore, ___ U.S. ___, 116 S.Ct. 1589, 134 L.Ed.2d 809 (1996). In a seminal case concerning an American's most precious possession (if not his most precious rights), a BMW purchaser in Alabama sued after discovering that his new BMW had been repainted prior to sale, alleging that the failure to disclose the repainting constituted fraud under Alabama law. Although the difference caused by the repainting was apparently imperceptible to the layperson, when the purchaser brought his car to "Slick Finish," an independent detailer, to make it look "snazzier than it normally would appear," BMW of North America Inc. v. Gore, 646 So.2d 619, 621 (Ala.1994), Mr. Slick, the aptly yclept proprietor, detected evidence that the car had been repainted. The plaintiff alleged that he had suffered $4,000 in actual damages, relying on the testimony of a former BMW dealer who estimated that the value of a repainted BMW was approximately 10% less than one that was "showroom new." Plaintiff further argued that a punitive damage award of $4 million was an appropriate penalty in light of evidence he introduced that BMW had sold 983 refinished cars as new, including 14 in Alabama.

    At trial, BMW acknowledged that it had adopted a nationwide policy of disclosing predelivery repairs only when the cost of the repairs exceeded 3% of the car's suggested retail price. The jury returned a verdict finding BMW liable for compensatory damages of $4,000 and punitive damages of $4 million, apparently calculated by multiplying the number of sales in all states of refinished cars by $4,000. BMW filed a post-trial motion to set aside the punitive damages award, contending that its nondisclosure policy was consistent with the laws of 25 states defining the disclosure obligations of automobile manufacturers; BMW asserted that the punitive damages were excessive because they were computed on the basis of sales that took place in jurisdictions where its conduct was perfectly legal.

    The Supreme Court agreed. The Court indicated that while Congress could enact a law requiring full disclosure of every presale repair to an automobile, no single state could impose such a policy nationwide by imposing economic sanctions aimed at changing the conduct of a tortfeasor in other states. Id. at ___, 116 S.Ct. at 1596. Speaking emphatically of the need to confine state legislation to its proper constitutional sphere, the Court stated:

    [O]ne State's power to impose burdens on the interstate market for automobiles is not only subordinate to the federal power over interstate commerce, Gibbons v. Ogden, 9 Wheat. 1, 194-96, 6 L.Ed. 23 (1824), but is also constrained by the need to respect the interests of other States, see, e.g., Healy v. Beer Institute, 491 U.S. 324, 335-36, 109 S.Ct. 2491, 2498-99, 105 L.Ed.2d 275 (1989) (the Constitution has a "special concern both with the maintenance of a national economic union unfettered by state-imposed limitations on interstate commerce and with the autonomy of the individual States within their respective spheres" (footnote omitted)); Edgar v. MITE Corp., 457 U.S. 624, 643, 102 S.Ct. 2629, 2641, 73 L.Ed.2d 269 (1982).

    Id. The need to contain individual state over-reaching thus arises not from any disrespect for the plenary authority of each state over its own internal affairs but out of a recognition that true protection of each state's respective authority is only possible when such limits are observed by all states.[7]

    [177] The nature of the Internet makes it impossible to restrict the effects of the New York Act to conduct occurring within New York. An Internet user may not intend that a message be accessible to New Yorkers, but lacks the ability to prevent New Yorkers from visiting a particular Website or viewing a particular newsgroup posting or receiving a particular mail exploder. Thus, conduct that may be legal in the state in which the user acts can subject the user to prosecution in New York and thus subordinate the user's home state's policy — perhaps favoring freedom of expression over a more protective stance — to New York's local concerns. See Bigelow v. Virginia, 421 U.S. 809, 824, 95 S.Ct. 2222, 2234, 44 L.Ed.2d 600 (1975) ("A State does not acquire power or supervision over the internal affairs of another State merely because the welfare and health of its own citizens may be affected when they travel to that State."). New York has deliberately imposed its legislation on the Internet and, by doing so, projected its law into other states whose citizens use the Net. See Southern Pacific Co. v. Arizona ex rel. Sullivan, 325 U.S. 761, 774, 65 S.Ct. 1515, 1522-23, 89 L.Ed. 1915 (1945) ("If one state may regulate train lengths, so may all others, and they need not prescribe the same maximum limitation. The practical effect of [a law limiting train lengths] is to control train operations beyond the boundaries of the state exacting it because of the necessity of breaking up and reassembling long trains at the nearest terminal points, before entering and after leaving the regulating state."). This encroachment upon the authority which the Constitution specifically confers upon the federal government and upon the sovereignty of New York's sister states is per se violative of the Commerce Clause.

    C. The Burdens the Act Imposes on Interstate Commerce Exceed Any Local Benefit

    Even if the Act were not a per se violation of the Commerce Clause by virtue of its extraterritorial effects, the Act would nonetheless be an invalid indirect regulation of interstate commerce, because the burdens it imposes on interstate commerce are excessive in relation to the local benefits it confers. The Supreme Court set forth the balancing test applicable to indirect regulations of interstate commerce in Pike v. Bruce Church, 397 U.S. 137, 142, 90 S.Ct. 844, 847, 25 L.Ed.2d 174 (1970).[8] Pike requires a twofold inquiry. The first level of examination is directed at the legitimacy of the state's interest. The next, and more difficult, determination weighs the burden on interstate commerce in light of the local benefit derived from the statute.

    In the present case, I accept that the protection of children against pedophilia is a quintessentially legitimate state objective — a proposition with which I believe even the plaintiffs have expressed no quarrel. See New York v. Ferber, 458 U.S. 747, 756-57, 102 S.Ct. 3348, 3354, 73 L.Ed.2d 1113 (1982) ("It is evident beyond the need for elaboration that a State's interest in `safeguarding the physical and psychological well-being of a minor' is `compelling.'") (quoting Globe Newspaper Co. v. Superior Court, 457 U.S. 596, 607, 102 S.Ct. 2613, 2620, 73 L.Ed.2d 248 (1982)); see also Sable v. Federal Communications Commission, 492 U.S. 115, 126, 109 S.Ct. 2829, 2836-37, 106 L.Ed.2d 93 (1989) ("[T]here is a compelling interest in protecting the physical and psychological well-being of minors. This interest extends to shielding minors from the influence of literature that is not obscene by adult standards."). The defendants spent considerable time in their [178] Memorandum and at argument asserting the legitimacy of the state's interest. Even with the fullest recognition that the protection of children from sexual exploitation is an indisputably valid state goal, however, the present statute cannot survive even the lesser scrutiny to which indirect regulations of interstate commerce are subject under the Constitution. The State cannot avoid the second stage of the inquiry simply by invoking the legitimate state interest underlying the Act. See Hunt v. Washington State Apple Advertising Comm'n, 432 U.S. 333, 350, 97 S.Ct. 2434, 2445, 53 L.Ed.2d 383 (1977) ("[A] finding that state legislation furthers matters of legitimate local concern, even in the health and consumer protection areas, does not end the inquiry."); Bibb v. Navajo Freight Lines, 359 U.S. 520, 528, 79 S.Ct. 962, 967, 3 L.Ed.2d 1003 (1959) (holding that "local safety measures that are nondiscriminatory [can] place an unconstitutional burden on interstate commerce"); see also Dean Milk Co. v. Madison, 340 U.S. 349, 354, 71 S.Ct. 295, 297-98, 95 L.Ed. 329 (1951) (holding that permitting a state to discriminate against interstate commerce to promote the health and safety of its citizens "would mean that the Commerce Clause of itself imposes no limitations on state action ... save for the rare instances where a state artlessly discloses an avowed purpose to discriminate against interstate goods."); Southern Pac. Co. v. Arizona, ex rel. Sullivan, 325 U.S. 761, 779, 65 S.Ct. 1515, 1525, 89 L.Ed. 1915 (1945) ("The principle that, without controlling Congressional action, a state may not regulate interstate commerce so as substantially to affect its flow or deprive it of needed uniformity in its regulation is not to be avoided by `simply invoking the convenient apologetics of the police power.'") (quoting Kansas City Southern Ry. v. Kaw Valley Drainage Dist., 233 U.S. 75, 79, 34 S.Ct. 564, 565, 58 L.Ed. 857 (1914)).

    The local benefits likely to result from the New York Act are not overwhelming. The Act can have no effect on communications originating outside the United States. As the three-judge panel that struck the federal analog of the New York Act, the Communications Decency Act, on First Amendment grounds concluded:

    [The Act] will almost certainly fail to accomplish the Government's interest in shielding children from pornography on the Internet. Nearly half of Internet communications originate outside the United States, and some percentage of that figure represents pornography. Pornography from, say, Amsterdam, will be no less appealing to a child on the Internet than pornography from New York City, and residents of Amsterdam have little incentive to comply with the [Act].

    American Civil Liberties Union v. Reno, 929 F.Supp. 824, 882 (E.D.Pa.1996). Further, in the present case, New York's prosecution of parties from out of state who have allegedly violated the Act, but whose only contact with New York occurs via the Internet, is beset with practical difficulties, even if New York is able to exercise criminal jurisdiction over such parties. The prospect of New York bounty hunters dragging pedophiles from the other 49 states into New York is not consistent with traditional concepts of comity.

    Moreover, the State has espoused an interpretation of the Act that, if accepted,[9] would [179] further undermine its effectiveness. According to defendant, the Act reaches only pictorial messages that are harmful to minors and has no impact on purely textual communications. Were this interpretation adopted, Mr. Barlow, whose conduct supposedly motivated the supporters of the Act, would escape prosecution because his messages were verbal. See The Crusader, supra, at 122 (reporting Barlow's message to New York girl as "I'm feeling really horny — I think Oscar is making a `statement.' We both want you very much. I'm thinking about you, & he's thinking about Love Bunny & tingling like mad.")

    The Act is, of course, not the only law in New York's statute books designed to protect children against sexual exploitation. The State is able to protect children through vigorous enforcement of the existing laws criminalizing obscenity and child pornography. See United States v. Thomas, 74 F.3d 701, 704-05 (6th Cir.1996), cert. denied, ___ U.S. ___, 117 S.Ct. 74, 136 L.Ed.2d 33 (1996). Moreover, plaintiffs do not challenge the sections of the statute that criminalize the sale of obscene materials to children, over the Internet or otherwise, and prohibit adults from luring children into sexual contact by communicating with them via the Internet. See N.Y. Penal Law § 235.21(1); N.Y. Penal Law § 235.22(2). The local benefit to be derived from the challenged section of the statute is therefore confined to that narrow class of cases that does not fit within the parameters of any other law. The efficacy of the statute is further limited, as discussed above, to those cases which New York is realistically able to prosecute.

    The conclusion that the New York Act has a very limited effect was bolstered by the testimony of Michael McCartney, an investigator with the New York State Attorney General's office. Mr. McCartney testified that he personally had logged over 600 hours investigating on-line criminal activity. (4/3/97 Tr., p. 12). Despite this extensive investment of time, Mr. McCartney admitted that he had investigated only two cases involving the dissemination of indecent materials to minors over the Internet that did not fall into the category of child pornography (which is, of course, subject to prosecution under other laws). (Id., p. 36). In one case, further investigation disclosed that the e-mail conversation actually took place between two adults and thus was outside the terms of the Act. (Id.). In the second case, Mr. McCartney was never able to determine which of the people in the household that held the Internet access account was responsible for sending the messages and pictures in question; he therefore never determined whether the sender was an adult. (Id., p. 37). In neither case did the Attorney General's office institute a prosecution. In fact, the Attorney General to date has not brought any prosecutions under the Act at all. (Id., p. 15). By contrast, Mr. McCartney described with justifiable pride his participation in the sting operation that resulted in the arrest of a student at SUNY who was using the Internet to contact a child; the defendant in that case, however, was charged under N.Y. Penal Law § 263, which prohibits an adult from promoting the sexual performance of a child. (Id., p. 12).

    Balanced against the limited local benefits resulting from the Act is an extreme burden on interstate commerce. The New York Act casts its net worldwide; moreover, the chilling effect that it produces is bound to exceed the actual cases that are likely to be prosecuted, as Internet users will steer clear of the Act by significant margin. See ACLU, 929 F.Supp. at 863 (holding that individuals, uncertain of the reach of the CDA, will undoubtedly "`steer far wider of the unlawful zone'") (citing Baggett v. Bullitt, 377 U.S. 360, 372, 84 S.Ct. 1316, 1322-23, 12 L.Ed.2d 377 (1964)); see also testimony of Maurice J. Freedman, Director of Westchester Library System, 4/7/97 Tr., at p. 209 ("My concern about prosecution in the context of this court proceeding is in relation to this Act. When I became aware of this Act and its implications for public libraries, as I perceived those implications, I at that point became quite concerned [180] — and scared might be another word — for being arrested or being in violation."). At oral argument, the State asserted that only a small percentage of Internet communications are "harmful to minors" and would fall within the proscriptions of the statute; therefore, the State argued, the burden on interstate commerce is small. On the record before me, I conclude that the range of Internet communications potentially affected by the Act is far broader than the State suggests. I note that in the past, various communities within the United States have found works including I Know Why the Caged Bird Sings by Maya Angelou, Funhouse by Dean Koontz, The Adventures of Huckleberry Finn by Mark Twain, and The Color Purple by Alice Walker to be indecent. (Teicher Decl., p. 3). Even assuming, arguendo, that the Act applies only to pictures, a number of Internet users take advantage of the medium's capabilities to communicate images to one another and, again, I find that the range of images that might subject the communicator to prosecution (or reasonably cause a communicator to fear prosecution) is far broader than defendants assert. For example, many libraries, museums and academic institutions post art on the Internet that some might conclude was "harmful to minors." Famous nude works by Botticelli, Manet, Matisse, Cezanne and others can be found on the Internet. In this regard, I point out that a famous painting by Manet which shows a nude woman having lunch with two fully clothed men was the subject of considerable protest when it first was unveiled in Paris, as many observers believed that it was "scandalous." (Declaration of Judith F. Krug, sworn to in March, 1997, at p. 5). Lesser known artists who post work over the Internet may face an even greater risk of prosecution, because the mantle of respectability that has descended on Manet is not associated with their as yet obscure names. Lile Elam, the founder of Art on the Net, submitted a Declaration that included samples of the types of work found on Art on the Net's site; certain of the images might be considered harmful to minors in some communities, including several nudes and a very dark, disturbing short story entitled "Two Running Rails of Mercury," accompanied by a picture of a woman's nude body dissolving into railroad tracks. (Declaration of Lile Elam, sworn to on March 13, 1997, Exh. 6). Rudolf Kinsky testified to his perception of the greater risk run by an unrenowned artist who posts controversial images on the Internet; when he was asked by defendants if a work by Corbet could subject the artist to prosecution, he answered, "His works are established; they are known. This is a different situation. Could be or could not, but my situation, when I am at the beginning of my career, and someone can, because I am not known, I have no established name and everything, I can still be prosecuted." (4/7/97 Tr., at 250). Individuals who wish to communicate images that might fall within the Act's proscriptions must thus self-censor or risk prosecution, a Hobson's choice that imposes an unreasonable restriction on interstate commerce. See Allen B. Dumont Labs., Inc. v. Carroll, 86 F.Supp. 813, 816 (1949) (holding that Pennsylvania state law requiring that motion pictures be submitted for review by a censorship board prior to being exhibited in the state imposed an undue and unreasonable burden on interstate commerce), aff'd, 184 F.2d 153 (3d Cir.1950), cert. denied, 340 U.S. 929, 71 S.Ct. 490, 95 L.Ed. 670 (1951).

    Moreover, as both three-judge panels that struck the federal statute have found, the costs associated with Internet users' attempts to comply with the terms of the defenses that the Act provides are excessive. Both courts that addressed the Communications Decency Act found that these costs of compliance, coupled with the threat of serious criminal sanctions for failure to comply, could drive some Internet users off the Internet altogether. See ACLU, 929 F.Supp. at 855-56 ("Many speakers who display arguably indecent content on the Internet must choose between silence and the risk of prosecution ... [the] defenses are not technologically or economically feasible for most providers"); Shea, 930 F.Supp. at 942-48 (finding that the defenses provided by the CDA do not offer a safe harbor to Internet users, who are then faced with the choice between complying, despite economic and technological barriers, or refraining from the Internet posting that potentially subjects [181] them to prosecution). While the defenses in the Act are not identical to those present in the CDA, the cost analysis undertaken by the ACLU and Shea courts is equally applicable to both statutes.

    The severe burden on interstate commerce resulting from the New York statute is not justifiable in light of the attenuated local benefits arising from it. The alternative analysis of the Act as an indirect regulation on interstate commerce therefore also mandates the issuance of the preliminary injunction sought by plaintiffs.

    D. The Act Unconstitutionally Subjects Interstate Use of the Internet to Inconsistent Regulations

    Finally, a third mode of Commerce Clause analysis further confirms that the plaintiffs are likely to succeed on the merits of their claim that the New York Act is unconstitutional. The courts have long recognized that certain types of commerce demand consistent treatment and are therefore susceptible to regulation only on a national level. The Internet represents one of those areas; effective regulation will require national, and more likely global, cooperation. Regulation by any single state can only result in chaos, because at least some states will likely enact laws subjecting Internet users to conflicting obligations. Without the limitation's imposed by the Commerce Clause, these inconsistent regulatory schemes could paralyze the development of the Internet altogether.

    In numerous cases, the Supreme Court has acknowledged the need for coordination in the regulation of certain areas of commerce. As long ago as 1886, the Supreme Court stated:

    Commerce with foreign countries and among the states, strictly considered, consists in intercourse and traffic, including in these terms navigation, and the transportation and transit of persons and property, as well as the purchase, sale, and exchange of commodities. For the regulation of commerce, as thus defined, there can be only one system of rules, applicable alike to the whole country; and the authority which can act for the whole country can alone adopt such a system. Action upon it by separate states is not, therefore, permissible.

    Wabash, St. L. & P. Ry. Co. v. Illinois, 118 U.S. 557, 574-75, 7 S.Ct. 4, 12, 30 L.Ed. 244 (1886). The Court in Wabash struck the Illinois statute at issue, which purported to establish interstate railway rates, stating "[t]hat this species of regulation is one which must be, if established at all, of a general and national character, and cannot be safely and wisely remitted to local rules and regulations, we think is clear from what has already been said." Id. at 577, 7 S.Ct. at 13.

    Similarly, in Southern Pac. Co. v. Arizona ex rel. Sullivan, 325 U.S. 761, 65 S.Ct. 1515, 89 L.Ed. 1915 (1945), the Court addressed the constitutionality of an Arizona statute that limited the length of trains within the state to fourteen passenger and seventy freight cars. The lower court's findings demonstrated that 93% of the freight traffic and 95% of the passenger traffic in Arizona was interstate; moreover, the Court endorsed the findings that travel by trains of more than fourteen passenger cars and more than seventy freight cars over the main lines of the United States was standard practice, and that the Arizona law had the effect of forcing railroads to decouple their trains in Texas or New Mexico and reform the train at full length in California. Id. at 774, 65 S.Ct. at 1522-23. Thus, the practical impact of the Arizona law was to control the length of trains, as the Court put it, "all the way from Los Angeles to El Paso." Id. The Court concluded that the Arizona train limit law imposed a serious burden on interstate commerce, noting that various states had imposed varying limits. The Court stated:

    With such laws in force in states which are interspersed with those having no limit on train lengths, the confusion and difficulty with which interstate operations would be burdened under the varied system of state regulation and the unsatisfied need for uniformity in such regulation, if any, are evident.

    Id. at 773-74, 65 S.Ct. at 1522. In striking the Arizona law as an unconstitutional intrusion on interstate commerce, the Court relied on a long-established rule barring the states from regulating "those phases of the national [182] commerce which, because of the need of national uniformity, demand that their regulation, if any, be prescribed by a single authority." Id. at 767, 65 S.Ct. at 1519 (citing Gibbons v. Ogden, 9 Wheat. 1, 6 L.Ed. 23 (1824); Cooley v. Board of Wardens, 12 How. 299, 319, 13 L.Ed. 996 (1851); Leisy v. Hardin, 135 U.S. 100, 108-09, 10 S.Ct. 681, 683-84, 34 L.Ed. 128 (1890); Minnesota Rate Cases, 230 U.S. 352, 400, 33 S.Ct. 729, 740, 57 L.Ed. 1511 (1913); Edwards v. People of State of California, 314 U.S. 160, 176, 62 S.Ct. 164, 168, 86 L.Ed. 119 (1941)).

    In Bibb v. Navajo Freight Lines, Inc., 359 U.S. 520, 79 S.Ct. 962, 3 L.Ed.2d 1003 (1959), the Court examined an Illinois statute that required the use of contour mudguards on trucks in Illinois. The Court took note of the fact that straight or conventional mudguards were permissible in most other states and actually required in Arkansas. Id. at 526, 79 S.Ct. at 966. Recognizing the need for coordinated legislation, the Court stated that "[t]he conflict between the Arkansas regulation and the Illinois regulation ... suggests that this regulation of mudguards is not one of those matters `admitting of diversity of treatment, according to the special requirements of local conditions.'" Id. at 529, 79 S.Ct. at 968 (quoting Sproles v. Binford, 286 U.S. 374, 390, 52 S.Ct. 581, 585-86, 76 L.Ed. 1167 (1932)). The Court struck the Illinois law as imposing an undue burden on interstate commerce, in part because Illinois was insisting upon "a design out of line with the requirements of almost all the other states." Id.

    The Internet, like the rail and highway traffic at issue in the cited cases, requires a cohesive national scheme of regulation so that users are reasonably able to determine their obligations. Regulation on a local Level, by contrast, will leave users lost in a welter of inconsistent laws, imposed by different states with different priorities. New York is not the only state to enact a law purporting to regulate the content of communications on the Internet. Already Oklahoma and Georgia have enacted laws designed to protect minors from indecent communications over the Internet; as might be expected, the states have selected different methods to accomplish their aims. Georgia has made it a crime to communicate anonymously over the Internet, while Oklahoma, like New York, has prohibited the online transmission of material deemed harmful to minors. See Ga.Code Ann. § 16-19-93.1 (1996); Okla. Stat. tit. 21, § 1040.76 (1996).

    Moreover, the regulation of communications that may be "harmful to minors" taking place over the Internet poses particular difficulties. New York has defined "harmful to minors" as including:

    that quality of any description or representation, in whatever form, of nudity, sexual conduct, sexual excitement, or sado-masochistic abuse, when it:

    (a) Considered as a whole, appeals to the prurient interest in sex of minors; and

    (b) Is patently offensive to prevailing standards in the adult community as a whole with respect to what is suitable material for minors; and

    (c) Considered as a whole, lacks serious literary, artistic, political and scientific value for minors.

    N.Y. Penal Law § 235.20(6). Courts have long recognized, however, that there is no single "prevailing community standard" in the United States. Thus, even were all 50 states to enact laws that were verbatim copies of the New York Act, Internet users would still be subject to discordant responsibilities. To use an example cited by the court in ACLU v. Reno, the Broadway play Angels in America, which concerns homosexuality and AIDS and features graphic language, was immensely popular in New York and in fact earned two Tony awards and a Pulitzer prize. ACLU, 929 F.Supp. at 852-53. In Charlotte, North Carolina, however, a production of the drama caused such a public outcry that the Mecklenberg County Commission voted to withhold all public funding from arts organizations whose works "expose the public to perverted forms of sexuality." Eric Harrison, Charlotte Ban on Funding Questions Community Culture Commission — Boycotts "Perverted Sexuality", Milwaukee J. & Sentinel, April 21, 1997, at 3. The Supreme Court has always recognized that "our nation is simply too big and too [183] diverse for this Court to reasonably expect that such standards [of what is patently offensive] could be articulated for all 50 states in a single formulation." Miller, 413 U.S. at 30, 93 S.Ct. at 2618.

    As discussed at length above, an Internet user cannot foreclose access to her work from certain states or send differing versions of her communication to different jurisdictions. In this sense, the Internet user is in a worse position than the truck driver or train engineer who can steer around Illinois or Arizona, or change the mudguard or train configuration at the state line; the Internet user has no ability to bypass any particular state. The user must thus comply with the regulation imposed by the state with the most stringent standard or forego Internet communication of the message that might or might not subject her to prosecution. For example, a teacher might invite discussion of Angels In America from a Usenet newsgroup dedicated to the literary interests of high school students. Quotations from the play might not subject her to prosecution in New York[10] — but could qualify as "harmful to minors" according to the community standards prevailing in Oklahoma. The teacher cannot tailor her message on a community-specific basis and thus must take her chances or avoid the discussion altogether.

    Further development of the Internet requires that users be able to predict the results of their Internet use with some degree of assurance. Haphazard and uncoordinated state regulation can only frustrate the growth of cyberspace. The need for uniformity in this unique sphere of commerce requires that New York's law be stricken as a violation of the Commerce Clause.

    III. The First Amendment and the Internet

    Plaintiffs have also asserted their entitlement to a preliminary injunction on the grounds that the Act unconstitutionally burdens free speech. Plaintiffs' ready ability to demonstrate the Act's unconstitutionality under the Commerce Clause, however, provides fully adequate support for the issuance of a preliminary injunction at this time. Moreover, the Supreme Court heard argument on a First Amendment challenge to the federal statute, the CDA, on March 19, 1997. The State vigorously argues that its law was designed to avoid the constitutional pitfalls presented by the CDA; however, the New York Act was clearly modelled on the CDA, and numerous provisions of the New York Act mirror their federal counterparts. See New York State Executive Charter Memorandum, annexed as Exhibit A to Declaration of Anat Hakim, sworn to on March 21, 1997 ("This bill ... is consistent with the federal statute"); Letter from William Sears to Governor Pataki, dated July 11, 1996, annexed as Exhibit A to Hershler Aff't ("This bill is consistent with the Federal Communications Decency Act"); Introducer's Memorandum in Support of Amended Senate Bill S. 210-E and Assembly Bill A. 3967-C, annexed as Exhibit G to Hershler Aff't ("Amendments were necessary for the bill to be consistent with the recently passed Federal Communications Decency Act.... Furthermore, it should be noted that the `harmful to minors' standard contained in the charging language of the offense is consistent with the Federal law...."). I believe any determination of plaintiffs' First Amendment challenge should therefore await the guidance to be provided by the Supreme Court's forthcoming opinion.

    CONCLUSION

    The protection of children from pedophilia is an entirely valid and laudable goal of State legislation. The New York Act's attempts to effectuate that goal, however, fall afoul of the Commerce Clause for three reasons. First, the practical impact of the New York Act results in the extraterritorial application of New York law to transactions involving citizens of other states and is therefore per se [184] violative of the Commerce Clause. Second, the benefits derived from the Act are inconsequential in relation to the severe burdens it imposes on interstate commerce. Finally, the unique nature of cyberspace necessitates uniform national treatment and bars the states from enacting inconsistent regulatory schemes. Because plaintiffs have demonstrated that they are likely to succeed on the merits of their claim under the Commerce Clause and that they face irreparable injury in the absence of an injunction, the motion for a preliminary injunction is granted.

    Defendants are enjoined from institutinq any prosecutions under the Act, until further Order of this Court. Plaintiffs shall submit a proposed form of injunction on two days' notice.

    SO ORDERED:

    [1] I recall in this respect a particularly confusing item of testimony elicited at the evidentiary hearing. Ms. Kovacs, plaintiffs' expert witness with respect to the Internet, testified that on one occasion while she was in a MUD (a Multi-User Dungeon), a malefactor sicced his "virtual dog" on her because she had trespassed on his domain. Fortunately, the other inhabitants of the MUD came to her rescue, vehemently protesting the unfriendliness of the virtual canine attack. Relieved as I was that the story had a happy ending, I must admit that it afforded me a window into an entirely unknown world. (4/4/97 Tr., p. 95).

    [2] Where information in this subsection is not cited to ACLU or Shea, it was derived from the parties' Joint Stipulation of Facts.

    [3] Because I find, as discussed below, that plaintiffs have demonstrated a likelihood of success on the merits of their claim that the New York Act violates the Commerce Clause, I do not rely on the "fair ground for litigation" standard. I note, however, that the standard would be applicable to this case because: (1) the action alleges constitutional violations, Almonte v. Pierce, 666 F.Supp. 517, 526 (S.D.N.Y.1987); (2) the public interest in a free flow of interstate commerce served by an injunction against enforcement of the Act counterbalances the public interest in protecting children served by the Act, see Carey v. Klutznick, 637 F.2d 834, 839 (2d Cir.1980); and (3) the New York Legislature did not engage in any fact-finding regarding the public interest served by the Act before promulgating it. Able v. United States, 44 F.3d 128, 131 (2d Cir.1995).

    [4] Other jurisdictions internationally have also gotten into the act. In January, 1997, two associations dedicated to the preservation of France's linguistic purity filed suit against two private corporations and Georgia Tech Lorraine, a French university affiliated with the Georgia Institute of Technology, claiming that the defendants violated a French law that prohibits advertising in any language other than French by operating English-language sites on the World Wide Web. See Complaint filed by L'Association "Avenir de la Lengue Francaise" and L'Association "Defense de La Lengue Francaise," Jan. 6, 1996; see also E. Schneiderman & R. Kornreich, Personal Jurisdiction and Internet Commerce, N.Y.L.J., June 4, 1997, at 1. The French court dismissed the action as to Georgia Tech, but other efforts by foreign jurisdictions to regulate the Internet are likely to follow. In addition, Germany made headlines recently when its anti-pornography laws forced Compuserve to close access to over 200 Internet sites from anywhere in the world. See John Markoff, Compuserve Bars Access to Internet Sex: German Laws Prompt the Provider to Block Pictures and Chat Groups, Orange County Register, December 29, 1995.

    [5] The defendants proposed Ms. Pirro as a witness for the evidentiary hearing, but then withdrew the proposal. Ms. Pirro's letter, which preceded the bill's signature into law by the Governor, is properly considered as part of the legislative history. See Civil Service Employees Association, Inc. v. Oneida, 78 A.D.2d 1004, 1005, 433 N.Y.S.2d 907 (4th Dep't 1980), appeal denied, 53 N.Y.2d 603, 439 N.Y.S.2d 1027, 421 N.E.2d 854 (1981). Ms. Pirro's testimony, on the other hand, would be a hindsight, post-enactment review of legislative intent by a non-legislator and would carry no probative weight. See Bread Political Action Committee v. Federal Election Committee, 455 U.S. 577, 582 n. 3, 102 S.Ct. 1235, 1238 n. 3, 71 L.Ed.2d 432 (1982); Frontier Ins. Co. v. New York, 160 Misc.2d 437, 609 N.Y.S.2d 748, 752 (Ct.Cl.1993), aff'd, 197 A.D.2d 177, 610 N.Y.S.2d 647 (3d Dep't 1994).

    [6] Judge Stein recently concluded that these realities meant that one whose only contact with the forum occurs via the Internet is not susceptible to suit there. Bensusan Restaurant Corp. v. King, 937 F.Supp. 295 (S.D.N.Y.1996).

    [7] The Court's injunction against extraterritorial regulation is long-established. See Huntington v. Attrill, 146 U.S. 657, 669, 13 S.Ct. 224, 228, 36 L.Ed. 1123 (1892) ("Laws have no force of themselves beyond the jurisdiction of the State which enacts them, and can have extraterritorial effect only by the comity of other States"); New York Life Ins. Co. v. Head, 234 U.S. 149, 161, 34 S.Ct. 879, 881-82, 58 L.Ed. 1259 (1914) ("[I]t would be impossible to permit the statutes of Missouri to operate beyond the jurisdiction of that State ... without throwing down the constitutional barriers by which all the States are restricted within the orbits of their lawful authority and upon the preservation of which the Government under the Constitution depends. This is so obviously the necessary result of the Constitution that it has rarely been called in question and hence authorities dealing directly with it do not abound.").

    [8] The distinction between direct regulations of interstate commerce, which are subject to a per se rule of invalidation, and indirect regulations subject to the less stringent balancing test has never been sharply defined. In either situation, however, the "critical consideration is the overall effect of the statute on both local and interstate activity." See Brown-Forman, 476 U.S. at 579, 106 S.Ct. at 2084; Raymond Motor Transportation, Inc. v. Rice, 434 U.S. 429, 440-41, 98 S.Ct. 787, 793-94, 54 L.Ed.2d 664 (1978).

    [9] The state's construction of the Act is unsupportable in light of the plain language of the statute and the interpretation that has been applied to closely related statutes. The Act applies to "communication[s] which, in whole or in part, depict[] actual or simulated nudity, sexual conduct or sado-masochistic abuse, and which [are] harmful to minors." The defendants contend that "depict" embraces only pictorial images. The dictionary definition of "depict," however, includes both visual representations and "description." Webster's Third New International Dictionary 605 (1981). The Act itself defines material that is harmful to minors as including any "description or representation," supporting an interpretation of the word "depict" that includes both text and pictures. Further, the Act is intended to extend liability under the statute as it existed prior to amendment. Cases brought under the prior law confirmed its applicability to sexually frank text, as well as pictures. See People v. Lida, 42 Misc.2d 56, 247 N.Y.S.2d 421 (N.Y.City Crim.Ct.) (finding that magazine containing short stories dealing with sex as well as photographs showing nude and partially nude women fell within the prohibition of Penal Law 1909 § 484-h [now Penal Law § 235.21], proscribing sale of magazines to minors), aff'd, 43 Misc.2d 692, 252 N.Y.S.2d 142 (1964); see also People v. Ginsberg, 56 Misc.2d 882, 290 N.Y.S.2d 239 (Dist.1966) (holding that evidence demonstrating that defendant, knowing buyer to be under 17, sold material containing pictures and photographs depicting female nudity and verbal descriptions and narrative accounts of sexual conduct and excitement was sufficient to sustain conviction for selling material harmful to minors), aff'd, 390 U.S. 629, 88 S.Ct. 1274, 20 L.Ed.2d 195 (1968).

    [10] Further distinctions may exist within the state of New York. The community standards prevailing in New York City may well be different than the community standards prevailing in, for example, Rensselaer County. See, e.g., United States v. Various Articles of Obscene Merchandise, Schedule No. 2102, 709 F.2d 132, 134, 137 (2d Cir. 1983) (upholding the district court's conclusion that "detailed portrayals of genitalia, sexual intercourse, fellatio, and masturbation" including the film "Deep Throat" and other pornographic films and magazines, are not obscene, "in light of the community standards prevailing in New York City.")

    1.6 State v. Heckel 1.6 State v. Heckel

    Heckel rules on dormant commerce clause issues as well, but comes out differently than Pataki. Why?

    24 P.3d 404 (2001)
    143 Wash.2d 824

    STATE of Washington, Appellant,
    v.
    Jason HECKEL, doing business as Natural Instincts, Respondent.

    No. 69416-8.

    Supreme Court of Washington, En Banc.

    Argued March 20, 2001.
    Decided June 7, 2001.

    [405] Honorable Christine Gregoire, Attorney General, Paula Lillian Selis, Helen Regina Cullen, W. Stuart Hirshfeld, Assts., Seattle, Jay Douglas Geck, Asst., Olympia, for Appellant.

    Van Siclen & Stocks, Robert Craig Van Siclen, Auburn, Dale L. Crandall, Charese Rhony, Salem, for Respondent.

    Miller, Nash, Brian William Esler, Richard J. Busch, Seattle, Amicus Curiae on Behalf of Washington Association of Internet Service Providers.

    OWENS, J.

    The State of Washington filed suit against Oregon resident Jason Heckel, alleging that his transmissions of electronic mail (e-mail) to Washington residents violated Washington's commercial electronic mail act, chapter 19.190 RCW (the Act). On cross-motions for summary judgment, the trial court dismissed the State's suit against Heckel, concluding that the Act violated the dormant Commerce Clause of the United States Constitution. [406] This court granted the State's request for direct review. We hold that the Act does not unduly burden interstate commerce. We reverse the trial court's dismissal of the State's suit, vacate the order on attorney fees, and remand this matter for trial.

    FACTS

    As early as February 1996, defendant Jason Heckel, an Oregon resident doing business as Natural Instincts, began sending unsolicited commercial e-mail (UCE), or "spam," over the Internet.[1] In 1997, Heckel developed a 46 page on-line booklet entitled "How to Profit from the Internet." The booklet described how to set up an on-line promotional business, acquire free e-mail accounts, and obtain software for sending bulk e-mail. From June 1998, Heckel marketed the booklet by sending between 100,000 and 1,000,000 UCE messages per week. To acquire the large volume of e-mail addresses,[2] Heckel used the Extractor Pro software program, which harvests e-mail addresses from various on-line sources and enables a spammer to direct a bulk-mail message to those addresses by entering a simple command. The Extractor Pro program requires the spammer to enter a return e-mail address, a subject line,[3] and the text of the message to be sent. The text of Heckel's UCE was a lengthy sales pitch that included testimonials from satisfied purchasers and culminated in an order form that the recipient could download and print. The order form included the Salem, Oregon, mailing address for Natural Instincts. Charging $39.95 for the booklet, Heckel made 30 to 50 sales per month.

    In June 1998, the Consumer Protection Division of the Washington State Attorney General's Office received complaints from Washington recipients of Heckel's UCE messages. The complaints alleged that Heckel's messages contained misleading subject lines and false transmission paths.[4] Responding [407] to the June complaints, David Hill, an inspector from the Consumer Protection Division, sent Heckel a letter advising him of the existence of the Act. The Act provides that anyone sending a commercial e-mail message from a computer located in Washington or to an e-mail address held by a Washington resident may not use a third-party's domain name without permission,[5] misrepresent or disguise in any other way the message's point of origin or transmission path, or use a misleading subject line.[6] RCW 19.190.030 makes a violation of the Act a per se violation of the Consumer Protection Act, chapter 19.86 RCW (CPA).

    Responding to Hill's letter, Heckel telephoned Hill on or around June 25, 1998. According to Hill, he discussed with Heckel the provisions of the Act and the procedures bulk e-mailers can follow to identify e-mail addressees who are Washington residents. Nevertheless, the Attorney General's Office continued to receive consumer complaints alleging that Heckel's bulk e-mailings from Natural Instincts appeared to contain misleading subject lines, false or unusable return e-mail addresses, and false or misleading transmission paths. Between June and September 1998, the Consumer Protection Division of the Attorney General's Office documented 20 complaints from 17 recipients of Heckel's UCE messages.

    On October 22, 1998, the State filed suit against Heckel, stating three causes of action. First, the State alleged that Heckel had violated RCW 19.190.020(1)(b) and, in turn, the CPA, by using false or misleading information in the subject line of his UCE messages. Heckel used one of two subject lines to introduce his solicitations: "Did I get the right e-mail address?" and "For your review—HANDS OFF!" Clerk's Papers (CP) at 6, 92, 113. In the State's view, the first subject line falsely suggested that an acquaintance of the recipient was trying to make contact, while the second subject line invited the misperception that the message contained classified information for the particular recipient's review.

    As its second cause of action, the State alleged that Heckel had violated RCW 19.190.020(1)(a), and thus the CPA, by misrepresenting information defining the transmission paths of his UCE messages. Heckel routed his spam through at least a dozen different domain names without receiving permission to do so from the registered owners of those names. For example, of the 20 complaints the Attorney General's Office received concerning Heckel's spam, 9 of the messages showed "13.com" as the initial ISP to transmit his spam. CP at 44, 113. The 13.com domain name, however, was registered as early as November 1995 to another individual, from whom Heckel had not sought or received permission to use the registered name. In fact, because the owner of 13.com had not yet even activated that domain name, no messages could have been sent or received through 13.com.

    Additionally, the State alleged that Heckel had violated the CPA by failing to provide a valid return e-mail address to which bulk-mail recipients could respond. When Heckel created his spam with the Extractor Pro software, he used at least a dozen different return e-mail addresses with the domain name "juno.com" (Heckel used the Juno accounts in part because they were free). CP at 88-89. None of the Juno e-mail accounts was readily identifiable as belonging to Heckel; the user names that he registered generally [408] consisted of a name or a name plus a number (e.g., "marlin1374," "cindyt5667," "howardwesley13," "johnjacobson1374," and "sjtowns"). CP at 88-89. During August and September 1998, Heckel's Juno addresses were canceled within two days of his sending out a bulk e-mail message on the account. According to Heckel, when Juno canceled one e-mail account, he would simply open a new one and send out another bulk mailing. Because Heckel's accounts were canceled so rapidly, recipients who attempted to reply were unsuccessful. The State thus contended that Heckel's practice of cycling through e-mail addresses ensured that those addresses were useless to the recipients of his UCE messages.[7] During the months that Heckel was sending out bulk e-mail solicitations on the Juno accounts, he maintained a personal e-mail account from which he sent no spam, but that e-mail address was not included in any of his spam messages. The State asserted that Heckel's use of such ephemeral e-mail addresses in his UCE amounted to a deceptive practice in violation of RCW 19.86.020.

    The State sought a permanent injunction and, pursuant to RCW 19.86.140 and .080 of the CPA, requested civil penalties, as well as costs and a reasonable attorney fee. In early 2000, the parties cross-moved for summary judgment. On March 10, 2000, the trial court entered an order granting Heckel's motion and denying the State's cross motion. The court found that the Act violated the Commerce Clause (U.S. CONST. art. I, § 8, cl. 3) and was "unduly restrictive and burdensome." CP at 175. The order permitted Heckel to "present a cost bill for recovery of his costs and statutory attorneys fees." CP at 175. Heckel then moved the court for a fee award of $49,897.50. Denying Heckel's request for fees under RCW 19.86.080 of the CPA, the court limited Heckel's award to statutory costs under RCW 4.84.030.

    Challenging the trial court's finding that the Act violated the Commerce Clause, the State sought this court's direct review. Heckel cross-appealed, seeking reversal of the trial court's denial of his attorney fee request under the CPA. We granted direct review.

    ISSUE

    Does the Act, which prohibits misrepresentation in the subject line or transmission path of any commercial e-mail message sent to Washington residents or from a Washington computer, unconstitutionally burden interstate commerce?

    ANALYSIS

    Standard of Review. The State seeks review of the trial court's decision on summary judgment that the Act violated the dormant Commerce Clause. This court reviews de novo a trial court's grant of summary judgment and views all facts in the light most favorable to the party challenging the summary dismissal. Lybbert v. Grant County, 141 Wash.2d 29, 34, 1 P.3d 1124 (2000). A legislative act is presumptively constitutional, "and the party challenging it bears the burden of proving it unconstitutional beyond a reasonable doubt." State v. Brayman, 110 Wash.2d 183, 193, 751 P.2d 294 (1988); see also Frach v. Schoettler, 46 Wash.2d 281, 280 P.2d 1038, cert. denied, 350 U.S. 838, 76 S.Ct. 75, 100 L.Ed. 747 (1955). A party meets the standard "if argument and research show that there is no reasonable doubt that the statute violates the constitution." Amalgamated Transit Union Local 587 v. State, 142 Wash.2d 183, 205, 11 P.3d [409] 762 (2000) (citing Belas v. Kiga, 135 Wash.2d 913, 920, 959 P.2d 1037 (1998)).

    Heckel's Challenge under the Commerce Clause. The Commerce Clause grants Congress the "power ... [t]o regulate commerce with foreign nations, and among the several states." U.S. CONST. art. I, § 8, cl. 3. Implicit in this affirmative grant is the negative or "dormant" Commerce Clause-the principle that the states impermissibly intrude on this federal power when they enact laws that unduly burden interstate commerce. See Franks & Son, Inc. v. State, 136 Wash.2d 737, 747, 966 P.2d 1232 (1998). Analysis of a state law under the dormant Commerce Clause generally follows a two-step process. We first determine whether the state law openly discriminates against interstate commerce in favor of intrastate economic interests. If the law is facially neutral, applying impartially to in-state and out-of-state businesses, the analysis moves to the second step, a balancing of the local benefits against the interstate burdens:

    Where the statute regulates evenhandedly to effectuate a legitimate local public interest, and its effects on interstate commerce are only incidental, it will be upheld unless the burden imposed on such commerce is clearly excessive in relation to the putative local benefits. If a legitimate local purpose is found, then the question becomes one of degree. And the extent of the burden that will be tolerated will of course depend on the nature of the local interest involved, and on whether it could be promoted as well with a lesser impact on interstate activities....

    Id. at 754, 966 P.2d 1232 (quoting Pike v. Bruce Church, Inc., 397 U.S. 137, 142, 90 S.Ct. 844, 25 L.Ed.2d 174 (1970)).

    The Act is not facially discriminatory. The Act applies evenhandedly to in-state and out-of-state spammers: "No person" may transmit the proscribed commercial e-mail messages "from a computer located in Washington or to an electronic mail address that the sender knows, or has reason to know, is held by a Washington resident." RCW 19.190.020(1) (emphasis added). Thus, just as the statute applied to Heckel, an Oregon resident, it is enforceable against a Washington business engaging in the same practices.

    Because we conclude that the Act's local benefits surpass any alleged burden on interstate commerce, the statute likewise survives the Pike balancing test. The Act protects the interests of three groups—ISPs, actual owners of forged domain names, and e-mail users. The problems that spam causes have been discussed in prior cases and legislative hearings. A federal district court described the harms a mass e-mailer caused ISP CompuServe:

    In the present case, any value CompuServe realizes from its computer equipment is wholly derived from the extent to which that equipment can serve its subscriber base.... [H]andling the enormous volume of mass mailings that CompuServe receives places a tremendous burden on its equipment. Defendants' more recent practice of evading CompuServe's filters by disguising the origin of their messages commandeers even more computer resources because CompuServe's computers are forced to store undeliverable e-mail messages and labor in vain to return the messages to an address that does not exist. To the extent that defendants' multitudinous electronic mailings demand the disk space and drain the processing power of plaintiff's computer equipment, those resources are not available to serve CompuServe subscribers. Therefore, the value of that equipment to CompuServe is diminished even though it is not physically damaged by defendants' conduct.

    CompuServe Inc. v. Cyber Promotions, Inc., 962 F.Supp. 1015, 1022 (S.D.Ohio 1997) (citations omitted) (granting preliminary injunction against bulk e-mailer on theory of trespass to chattels); see also Am. Online, Inc. v. IMS, 24 F.Supp.2d 548, 550 (E.D.Va.1998) ("rely[ing] on the reasoning of CompuServe" and finding that bulk e-mailer "injured AOL's business goodwill and diminished the value of its possessory interest in its computer network"). To handle the increased e-mail traffic attributable to deceptive spam, ISPs must invest in more computer equipment.[8] [410] Operational costs likewise increase as ISPs hire more customer service representatives to field spam complaints and more system administrators to detect accounts being used to send spam.[9]

    Along with ISPs, the owners of impermissibly used domain names and e-mail addresses suffer economic harm. For example, the registered owner of "localhost.com" alleged that his computer system was shut down for three days by 7,000 responses to a bulk-mail message in which the spammer had forged the e-mail address "nobody@localhost.com" into his spam's header. Seidl v. Greentree Mortgage Co., 30 F.Supp.2d 1292, 1297-98 (D.Colo.1998); see also Spamming: The E-Mail You Want to Can: Hearing Before the Subcomm. on Telecommunications, Trade, and Consumer Protection of the Comm. on Commerce, 106th Cong. 9 (1999) (statement of Rep. Gary G. Miller) (attached as App. 4, Br. of Amicus WAISP); 146 CONG. REC. H6373 (daily ed. July 18, 2000) (statement of Rep. Miller), available at http://thomas.loc.gov/home/ 106query.html (recounting similar experience of California constituent).

    Deceptive spam harms individual Internet users as well. When a spammer distorts the point of origin or transmission path of the message, e-mail recipients cannot promptly and effectively respond to the message (and thereby opt out of future mailings); their efforts to respond take time, cause frustration, and compound the problems that ISPs face in delivering and storing the bulk messages. And the use of false or misleading subject lines further hampers an individual's ability to use computer time most efficiently. When spammers use subject lines "such as `Hi There!,' `Information Request,' and `Your Business Records,'" it becomes "virtually impossible" to distinguish spam from legitimate personal or business messages.[10] Individuals who do not have flat-rate plans for Internet access but pay instead by the minute or hour are harmed more directly, but all Internet users (along with their ISPs) bear the cost of deceptive spam.

    This cost-shifting—from deceptive spammers to businesses and e-mail users—has been likened to sending junk mail with postage due or making telemarketing calls to someone's pay-per-minute cellular phone.[11] In a case involving the analogous practice of junk faxing (sending unsolicited faxes that contain advertisements), the Ninth Circuit acknowledged "the government's substantial interest in preventing the shifting of advertising costs to consumers." Destination Ventures, Ltd. v. F.C.C., 46 F.3d 54, 56 (9th Cir.1995) (holding that the Telephone Consumer Protection Act's (47 U.S.C. § 227) limitations on commercial speech did not violate the First Amendment). We thus recognize that the Act serves the "legitimate local purpose" of banning the cost-shifting inherent in the sending of deceptive spam.

    Under the Pike balancing test, "[i]f a legitimate local purpose is found, then the question [411] becomes one of degree." 397 U.S. at 142, 90 S.Ct. 844. In the present case, the trial court questioned whether the Act's requirement of truthfulness (in the subject lines and header information) would redress the costs associated with bulk e-mailings. As legal commentators have observed, however, "the truthfulness requirements (such as the requirement not to misrepresent the message's Internet origin) make spamming unattractive to the many fraudulent spammers, thereby reducing the volume of spam." Jack L. Goldsmith & Alan O. Sykes, The Internet and the Dormant Commerce Clause, 110 Yale L.J. 785, 819 (2001). Calling "simply wrong" the trial court's view "that truthful identification in the subject header would do little to relieve the annoyance of spam," the commentators assert that "[t]his identification alone would allow many people to delete the message without opening it (which takes time) and perhaps being offended by the content." Id. The Act's truthfulness requirements thus appear to advance the Act's aim of protecting ISPs and consumers from the problems associated with commercial bulk e-mail.

    To be weighed against the Act's local benefits, the only burden the Act places on spammers is the requirement of truthfulness, a requirement that does not burden commerce at all but actually "facilitates it by eliminating fraud and deception." Id. Spammers must use an accurate, nonmisleading subject line, and they must not manipulate the transmission path to disguise the origin of their commercial messages. While spammers incur no costs in complying with the Act, they do incur costs for noncompliance, because they must take steps to introduce forged information into the header of their message.[12] In finding the Act "unduly burdensome," CP at 175, the trial court apparently focused not on what spammers must do to comply with the Act but on what they must do if they choose to use deceptive subject lines or to falsify elements in the transmission path. To initiate deceptive spam without violating the Act, a spammer must weed out Washington residents by contacting the registrant of the domain name contained in the recipient's e-mail address.[13] This focus on the burden of noncompliance is contrary to the approach in the Pike balancing test, where the United States Supreme Court assessed the cost of compliance with a challenged statute. Pike, 397 U.S. at 143, 90 S.Ct. 844. Indeed, the trial court could have appropriately considered the filtering requirement a burden only if Washington's statute had banned outright the sending of UCE messages to Washington residents. We therefore conclude that Heckel has failed to prove that "the burden imposed on ... commerce [by the Act] is clearly excessive in relation to the putative local benefits." Id. at 142, 90 S.Ct. 844 (emphasis added).

    Drawing on two "unsettled and poorly understood" aspects of the dormant Commerce Clause analysis, Heckel contended that the Act (1) created inconsistency among the states and (2) regulated conduct occurring wholly outside of Washington.[14] The inconsistent-regulations test and the extraterritoriality analysis are appropriately regarded as facets of the Pike balancing test.[15] The Act survives both inquiries. At present, 17 other states have passed legislation regulating [412] electronic solicitations.[16] The truthfulness requirements of the Act do not conflict with any of the requirements in the other states' statutes, and it is inconceivable that any state would ever pass a law requiring spammers to use misleading subject lines or transmission paths. Some states' statutes do include additional requirements; for example, some statutes require spammers to provide contact information (for opt-out purposes) or to introduce subject lines with such labels as "ADV" or "ADV-ADLT." But because such statutes "merely create additional, but not irreconcilable, obligations," they "are not considered to be `inconsistent'" for purposes of the dormant Commerce Clause analysis. Instructional Sys., Inc. v. Computer Curriculum Corp., 35 F.3d 813, 826 (3d Cir.1994). The inquiry under the dormant Commerce Clause is not whether the states have enacted different anti-spam statutes but whether those differences create compliance costs that are "clearly excessive in relation to the putative local benefits." Pike, 397 U.S. at 142, 90 S.Ct. 844. We do not believe that the differences between the Act and the anti-spam laws of other states impose extraordinary costs on businesses deploying spam.[17]

    Nor does the Act violate the extraterritoriality principle in the dormant Commerce Clause analysis. Here, there is no "sweeping extraterritorial effect" that would outweigh the local benefits of the Act. Edgar v. MITE Corp., 457 U.S. 624, 642, 102 S.Ct. 2629, 73 L.Ed.2d 269 (1982). Heckel offers the hypothetical of a Washington resident who downloads and reads the deceptive spam while in Portland or Denver. He contends that the dormant Commerce Clause is offended because the Act would regulate the recipient's conduct while out of state. However, the Act does not burden interstate commerce by regulating when or where recipients may open the proscribed UCE messages. Rather, the Act addresses the conduct of spammers in targeting Washington consumers. Moreover, the hypothetical mistakenly presumes that the Act must be construed to apply to Washington residents when they are out of state, a construction that creates a jurisdictional question not at issue in this case.

    In sum, we reject the trial court's conclusion that the Act violates the dormant Commerce Clause. Although the trial court found particularly persuasive American Libraries Association v. Pataki, 969 F.Supp. 160 (S.D.N.Y.1997), that decision—the first to apply the dormant Commerce Clause to a state law on Internet use—is distinguishable in a key respect.[18] At issue in American Libraries was a New York statute that made it a crime to use a computer to distribute harmful, sexually explicit content to minors. The statute applied not just to initiation of e-mail messages but to all Internet activity, including the creation of websites. Thus, under the New York statute, a website creator in California could inadvertently violate the law simply because the site could be viewed in New York. Concerned with the statute's "chilling effect," id. at 179, the court observed that, if an artist "were located in California and wanted to display his work to a prospective purchaser in Oregon, he could not employ his virtual [Internet] studio to do so without risking prosecution under the New York law." Id. at 174. In contrast to the New York statute, which could reach all content posted on the Internet and therefore [413] subject individuals to liability based on unintended access, the Act reaches only those deceptive UCE messages directed to a Washington resident or initiated from a computer located in Washington; in other words, the Act does not impose liability for messages that are merely routed through Washington or that are read by a Washington resident who was not the actual addressee.

    CONCLUSION

    The Act limits the harm that deceptive commercial e-mail causes Washington businesses and citizens. The Act prohibits e-mail solicitors from using misleading information in the subject line or transmission path of any commercial e-mail message sent to Washington residents or from a computer located in Washington. We find that the local benefits of the Act outweigh any conceivable burdens the Act places on those sending commercial e-mail messages. Consequently, we hold that the Act does not violate the dormant Commerce Clause of the United States Constitution. We reverse the trial court and remand the matter for trial. The trial court's order on attorney fees is vacated.

    ALEXANDER, C.J., SMITH, JOHNSON, MADSEN, SANDERS, IRELAND, CHAMBERS and BRIDGE, JJ., concur.

    [1] "`Commercial electronic mail message' means an electronic mail message sent for the purpose of promoting real property, goods, or services for sale or lease." RCW 19.190.010(2). The term "spam" refers broadly to unsolicited bulk e-mail (or "`junk' e-mail"), which "can be either commercial (such as an advertisement) or noncommercial (such as a joke or chain letter)." Sabra Anne Kelin, State Regulation of Unsolicited Commercial E-Mail, 16 Berkeley Tech. L.J. 435, 436 & n. 10 (2001). Use of the term "spam" as Internet jargon for this seemingly ubiquitous junk e-mail arose out of a skit by the British comedy troupe Monty Python, in which a waitress can offer a patron no single menu item that does not include spam: "Well, there's spam, egg, sausage and spam. That's not got much spam in it." 2 Graham Chapman et al., The Complete Monty Python's Flying Circus: All the Words 27 (Pantheon Books 1989); see also Kadow's Internet Dictionary, at http://www.msg.net/kadow/answers/s.html (last visited May 7, 2001). Hormel Foods Corporation, which debuted its SPAM® luncheon meat in 1937, has dropped any defensiveness about this use of the term and now celebrates its product with a website (www. spam.com). See Hormel Objects to Cyber Promotions' Use of "SPAM" Mark, 4 No. 1 Andrews Intell. Prop. Litig. Rep. 19 (1997); Laurie J. Flynn, Gracious Concession on Internet "Spam," N.Y. Times, Aug. 17, 1998, at D3. Because the term has been widely adopted by Internet users, legislators, and legal commentators, we use the term herein, along with its useful derivatives "spammer" and "spamming."

    [2] "`Electronic mail address' means a destination, commonly expressed as a string of characters, to which electronic mail may be sent or delivered." RCW 19.190.010(3).

    [3] The subject line, similar to the "RE" line of a letter or memorandum, is generally displayed (at least in part) alongside the sender's name in the recipient's e-mail inbox.

    [4] Each e-mail message, which is simply a computer data file, contains so-called "header" information in the "To," "From," and "Received" fields. When an e-mail message is transmitted from one e-mail address to another, the message generally passes through at least four computers: from the sender's computer, the message travels to the mail server computer of the sender's Internet Service Provider (ISP); that computer delivers the message to the mail server computer of the recipient's ISP, where it remains until the recipient retrieves it onto his or her own computer. Every computer on the Internet has a unique numerical address (an Internet Protocol or IP address), which is associated with a more readily recognizable domain name (such as "mysite.com"). As the e-mail message travels from sender to recipient, each computer transmitting the message attaches identifying data to the "Received" field in the header. The information serves as a kind of electronic postmark for the handling of the message. See Clerk's Papers (CP) at 130-34. It is possible for a sender to alter (or "spoof") the header information by misidentifying either the computer from which the message originated or other computers along the transmission path. See Kelin, supra note 1, at 445.

    [5] See RCW 19.190.010(6) (defining "Internet domain name").

    [6] "(1) No person may initiate the transmission, conspire with another to initiate the transmission, or assist the transmission, of a commercial electronic mail message from a computer located in Washington or to an electronic mail address that the sender knows, or has reason to know, is held by a Washington resident that:

    "(a) Uses a third party's internet domain name without permission of the third party, or otherwise misrepresents or obscures any information in identifying the point of origin or the transmission path of a commercial electronic mail message; or

    "(b) Contains false or misleading information in the subject line.

    "(2) For purposes of this section, a person knows that the intended recipient of a commercial electronic mail message is a Washington resident if that information is available, upon request, from the registrant of the Internet domain name contained in the recipient's electronic mail address." RCW 19.190.020.

    [7] The experience of 1 of the 17 complainants to the Attorney General's Office is illustrative. Nancy Smith received Heckel's spam on September 1, 1998; the message was sent from a Juno account with the user name "apollo1113," and the subject line read "For your review—HANDS OFF." CP at 140. On or about September 1, 1998, Smith sent a copy of the Natural Instincts order form with a check for $39.95 by U.S. Mail to the Salem, Oregon, address provided on the order form. Hearing nothing for some weeks, Smith sent a message by return e-mail on September 30, 1998, but within a minute she received a return e-mail from Juno stating that the attempt had failed due to termination of the account. Unable to find any information about Natural Instincts on the Internet, Smith contacted her bank and learned that the check had cleared two weeks earlier. Smith then contacted the Attorney General's Office. CP at 140-41, 149-50.

    [8] "[W]hen Internet users attempt to reply to deceptive spam that has a fraudulent return address or domain name, one e-mail message (and the ISP[`s] related computer log entry) instantly becomes three separate e-mail messages (and additional computer log entries) because: (1) the ISP server that is the victim of the fraudulent return address or domain name sends an error message back to the Internet user and their ISP announcing that the return path was invalid, (2) a message is sent to the server administrator requesting an investigation of the return address for potential problems, and (3) a message is sent to the server log in case the ISP wishes to track down the problem later. With bulk spam, these messages snowball to clog ISP resources, and ISPs have little choice but to purchase additional equipment at a significant cost." Br. of Amicus Washington Association of Internet Service Providers (WAISP) at 11-12.

    [9] See Br. of Amicus WAISP at 12-13; see also Spamming: The E-Mail You Want to Can: Hearing Before the Subcomm. on Telecommunications, Trade, and Consumer Protection of the Comm. on Commerce, 106th Cong. 41-42 (1999) (statement of Michael Russina, Director of Systems Operations, SBC Internet Services) (attached as App. 4, Br. of Amicus WAISP).

    [10] Testimony of Ed McNichol at Hearing on H.B. 2752 Before the Washington House Comm. on Energy and Utilities (Jan. 28, 1998) (partial transcript attached as App. 2, Br. of Amicus WAISP; audio also available at http://1 98.239.32.162/ramgen/199801/XXXXXXXXXX.ra).

    [11] See Spamming: The E-Mail You Want to Can, supra note 9, at 1 (statement of Rep. W.J. Tauzin, Chairman, Subcomm. on Telecommunications, Trade, and Consumer Protection) (attached as App. 4, Br. of Amicus WAISP).

    [12] "This generally involves paying a bulk re-mailing service to forge e-mail headers and send out the spammer's message, or at least running additional software programs to alter the e-mail messages' address and domain name information." Br. of Amicus WAISP at 8.

    [13] See RCW 19.190.020(2). The Washington Association of Internet Service Providers (WAISP) and the Washington Attorney General co-sponsor a registry of Washington residents who do not want to receive spam. See WAISP Registry Page, at http://registry.waisp.org (last visited May 7, 2001).

    [14] Jack L. Goldsmith & Alan O. Sykes, The Internet and the Dormant Commerce Clause, 110 Yale L.J. 785, 789 (2001).

    [15] See Goldsmith & Sykes, supra note 14, at 808 (concluding that "inconsistent-regulations cases, like extraterritoriality cases, should be viewed as just another variant of balancing analysis"); see also William Lee Biddle, State Regulation of the Internet: Where Does the Balance of Federalist Power Lie? 37 Cal. W.L.Rev. 161, 167 (2000) (suggesting that "[t]he burden placed on interstate commerce through inconsistent local regulation is more appropriately placed as part of the Pike balancing test, rather than its own, separate line of inquiry").

    [16] See David E. Sorkin, Spam Laws, at http://www.spamlaws.com/state/index.html; see also Max P. Ochoa, Legislative Note: Recent State Laws Regulating Unsolicited Electronic Mail, 16 Santa Clara Computer & High Tech. L.J. 459 (2000); Br. of Appellant at 23 and App. A, B. Proposed federal legislation, the Unsolicited Commercial Electronic Mail Act of 2000, H.R. 3113, 106th Cong. (2000), was passed by the House on July 18, 2000, and has been referred to the Senate Committee on Commerce, Science, and Transportation. The text of the bill may be accessed through http://thomas.loc.gov/home/c106query.html.

    [17] As the State notes, "[p]resently, mail and phone solicitors are expected to abide by different states' telemarketing laws and other consumer protection laws. E-mail solicitors should not be excused from the burden of complying with a state's law simply because of the case of sending bulk e-mail solicitations in relation to other forms of commercial solicitation." CP at 53.

    [18] See CP at 216. At least 10 other cases have distinguished American Libraries. See, e.g., Hatch v. Super. Ct., 80 Cal.App.4th 170, 94 Cal. Rptr.2d 453 (2000); People v. Hsu, 82 Cal. App.4th 976, 99 Cal.Rptr.2d 184 (2000); Ford Motor Co. v. Tex. Dep't of Transp., 106 F.Supp.2d 905, 909 (W.D.Tex.2000).