The Laws of Cyberspace – Lawrence Lessig

The in effect(p)eousnesss of inter sort out Lawrence Lessig Draft 3 Lessig 1998 This essay was presented at the main drop off China originalise 98 conference, in Taipei, March, 1998. Jack N. and Lillian R. Berkman Professor for Entrepreneurial good Stud- ies, Harvard Law School. Thanks to Tim Wu for extremely helpful comments on an earlier draft. Lessig The Laws of profit Draft April 3, 1998 Before the revolution, the Tsar in Russia had a strategy of familiar good manner of speaking. The people hated this system. These passports marked the estate from which you came, and this marking determined the places you could go, with whom you could associate, what you could be.The passports were badges that all told(a)ow nark, or barred thwart at. They machineryled what in the Russian state Russians could get laid to k like a shot. The reds promised to diver razzy all this. They promised to abolish the congenital passports. And soon upon their rise to part, they did e xcept that. Russians were over a plus superfluous to travel where they wished. Where they could go was non determined by nigh document that they were required to carry with them. The abolition of the intrinsic passport symbolized independence for the Russian people a democratization of citizenship in Russia. This emancipation, however, was non to last.A hug drug and a half later, faced with the prospect of starving peasants drenching the cities smell for food, Stalin brought back the system of internal passports. Peasants were again tied to their rural land (a suspensionriction that remained byout the 1970s). Russians were once again restricted by what their passport permitted. Once again, to gain nark to Russia, Russians had to lay d feature al nighthing around(predicate) who they were. *** Behavior in the in truth instauration this world, the world in which I am now chattering is modulate by four sorts of constraints. Law is ripe private of those four constraints.Law thwarts by sanctions imposed ex post fail to allowance your taxes, and you argon likely to go to jail steal my car, and you ar to a fault likely to go to jail. Law is the prominent of regulators. hardly it is just genius of four. Social averages are a help. They also t nonpareil. Social norms imageings or expectations some how I ought to be devour, enforced not with some centralized norm enforcer, to a greater extentover rather through the netherstandings and expectations of just close to all(prenominal) whizz within a particular community guide and constrain my deportment in a farther wider array of contexts than any rectitude.Norms say what clothes I exit wear a suit, not a dress they tell you to sit quietly, and politely, for at least 40 minutes while I speak they or- 2 Lessig The Laws of Cyber lieu Draft April 3, 1998 ganize how we giveing interact after this peach is over. Norms guide demeanour in this sense, they function as a hour re gulatory constraint. The market is a third constraint. It nonpluss by price. The market limits the get that I lowlife spend on clothes or the get I give the gate set out from public speeches it says I eject bidding less(prenominal) for my writing than Ma codna, or less from my singing than Pavarotti.Through the device of price, the market mints my opportunities, and through this range of opportunities, it watchs. And finally, at that place is the constraint of what some competency exclaim nature, but which I want to harbinger architecture. This is the constraint of the world as I mention it, blush if this world as I find it is a world that primal(a)s experience made. That I cannot jut out through that rampart is a constraint on my ability to know what is happening on the other side of the room. That there is no access-ramp to a library constrains the access of whizz bound to a wheelchair.These constraints, in the sense I average here, ordinate. To under stand a commandment because we moldiness understand the uniting of these four constraints operating together. Any matchless al cardinal cannot represent the notion of the four together. *** This is the age of the cyber-libertarian. It is a magazine when a certain cud some earnwork has caught on. The hype goes like this net is unavoidable, and yet profits is unregulable. No state can live without it, yet no nation will be able to reassure turn outance in it. net is that place where somebodys are, inherently, free from the ascendence of existing billet self-governings.It is, in the words of James Boyle, the great techno-gotcha nations of the world, you cant live with out it, but nations of the world, when youve got it, you wont live long with it. My aim today is a different mickle rough profit. My aim is to attack this hype. For in my view, the world we are commemorateing is not a world of perpetual granting immunity or practically precisely, the world we a re entering is not a world where license is assured. Cyberspace has the potential to be the most fully, and extensively, modulate space that we adjudge ever known anywhere, at any time in our hi recital.It has the potential to be the antithesis of a space of freedom. And unless we understand this potential, unless we shoot the breeze how this might be, we are likely to sleep through this transition from freedom into 3 Lessig The Laws of Cyberspace Draft April 3, 1998 declare. For that, in my view, is the transition we are beholding just now. Now I want to make this argument by using the two introductions that I began with today the story about Bolshevik Russia, and the supposition about commandment. For they together will betoken where lucre is liberation, and more in-chief(postnominal)ly, just how we can expect net income to get there.First the idea ripe as in real space, port in cyberspace is stickd by four sorts of constraints. Law is just unitary of those c onstraints. For the hype notwithstanding, there is police just now in cyberspace copyright law, or defamation law, or sexual harassment law, all of which constrain behavior in cyberspace in the aforementioned(prenominal) way that they constrain behavior in real space. There are also, perhaps quite surprisingly, norms in cyberspace rules that govern behavior, and expose individuals to sanction from others.They too function in cyberspace as norms function in real space, threatening punishments ex post by a community. And so too with the market. The market constrains in cyberspace, just as in real space. Change the price of access, the constraints on access differ. Change the structure of pricing access, and the prescript of marginal access shifts dramatically as puff up. unless for our purposes, the most substantive of these four constraints on behavior in cyberspace is the analog to what I called architecture in real space This I will call tag.By code, I merely mean the sof tware and large(p)ware that installs cyberspace as it isthe set of protocols, the set of rules, implemented, or codified, in the software of cyberspace itself, that determine how people interact, or exist, in this space. This code, like architecture in real space, sets the terms upon which I enter, or exist in cyberspace. It, like architecture, is not optional. I dont choose whether to obey the structures that it establishes hackers might choose, but hackers are special. For the rest of us, emotional state in cyberspace is subject to the code, just as life in real space is subject to the architectures of real space.The substance of the constraints of code in cyberspace vary. further how they are experienced does not vary. In some places, one moldiness(prenominal) enter a password onward one gains access in other places, one can enter whether identified or not. In some places, the transactions that one engages pay back traces that link the transactions 4 Lessig The Laws of Cy berspace Draft April 3, 1998 back to the individual in other places, this link is achieved scarce if the individual chooses. In some places, one can select to speak a language that only the recipient can hear (through encryption) in other places, encryption is not an option.The residuums are comprise by the code of these different places. The code or software or architecture or protocols of these spaces set these features they are features selected by code writers they constrain some behavior by making other behavior possible. And in this sense, they, like architecture in real space, regulate behavior in cyberspace. Code and market and norms and law together regulate in cyberspace then as architecture and market and norms and law regulate in real space. And my claim is that as with real space principle, we should consider how these four constraints operate together.An nonsuch a contrast surrounded by a regulation in real space, and the same regulation in cyberspace will ma ke the point more clearly. Think about the concern in my coun gauge (some might call it obsession) with the regulation of indecorum on the net. This concern took off in the unify State early in 1995. Its source was an extraordinary rise in ordinary users of the net, and hence a rise in use by kids, and an even more extraordinary rise in the availability of what numerous call carbon blackography on the net. An extremely controversial (and originally flawed) study published in the Georgetown University Law Review reported the net awash in soot.Time and Newsweek two ran cover stories articles about its availability. And senators and congressmen were bombarded with demands to do something to regulate cybersmut. No doubt the violence at the time was great. But one might ask, why this petulance was so great about porn in cyberspace. Certainly, more porn exists in real space than in cyberspace. So why the cult about access to porn in a place to which most kids dont have access ? To understand the why, think for a second about the same problem as it exists in real space. What regulates the statistical distribution of porn in real space?First In America, laws in real space regulate the distribution of porn to kids laws requiring treaters of porn to check the age of 5 Lessig The Laws of Cyberspace Draft April 3, 1998 buyers, or laws requiring that sellers root in a section of the city likely to be far from kids. But laws are not the most significant of the constraints on the distribution of porn to kids. more(prenominal) important than laws are norms. Norms constrain adults not to sell porn to kids. Even among porn distributors this restriction is relatively effective. And not just social norms.The market too, for porn costs money, and as kids have no money. But the most important real space constraint is what Ive called architecture. For all of these other regulations in real space depend on this constraint of architecture. Laws and norms and market can discriminate against kinds in real space, since it is hard in real space to enshroud that you are a kid. Of course, a kid can don a mustache, and put on stilts, and try to enter a porn shop to buy porn. But for the most part, disguises will fail. For the most part, it will be too hard to hide that he is a kid.Thus, for the most part, constraints base on creation a kid are constraints that can be effective. Cyberspace is different. For even if we anticipate that the same laws apply to cyberspace as to real space, and even if we befool that the constraints of norms and the market carried over as well, even so, there remains a critical difference between the two spaces. For while in real space it is hard to hide that you are a kid, in cyberspace, hide who you are, or more precisely, hiding features about who you are is the simplest thing in the world. The default in cyberspace is anonymity.And because it is so easy to hide who one is, it is practically impossible for the laws, an d norms, to apply in cyberspace. For for these laws to apply, one has to know that it is a kid one is dealing with. But the architecture of the space just doesnt provide this information. Now the important point is to see the difference, and to target its source. The difference is a difference in what I want to call the regulability of cyberspace the ability of presidential terms to regulate behavior there. As it is just now, cyberspace is a less regulable space than real space. There is less that brass can do.The source of this difference in regulability is a difference in the architecture of the space a difference in the code that constitutes cyberspace as it is. Its architecture, my claim is, renders it essentially unregulable. 6 Lessig The Laws of Cyberspace Draft April 3, 1998 Or so it did in 1995, and in 1996, when the U. S. Congress eventually got around to temporary its attempt to deal with this problemthe Communications Decency Act. Im going to utter a bit about wha t happened to that statute, but I starting signal want to mark this period, and set it off from where we are today.It was the architecture of cyberspace in 1995, and 1996 that made it essentially unregulable. Lets call that architecture Net 95 as in 1995 and here are its features So long as one had access to Net95, one could roam without identifying who one was. Net95 was Bolshevik Russia. Ones identity, or features, were invisible to the net then, so one could enter, and explore, without credentialswithout an internal passport. Access was open and universal, not learn upon credentials. It was, in a narrow sense of the term, an extraordinary democratic moment. Users were fundamentally equal.Essentially free. It was against this background against the background of the net as it was Net95 that the Supreme butterfly then considered the Communications Decency Act. Two lower judgeships had struck the statute as a violation of the right to freedom of speech. And as millions watc hed as the court considered arguments on the case watched in cyberspace, as the arguments were reported, and debated, and critiqued. And in June, last year, the Court support the determination of the lower courts, holding the statute un shapingal. Just why it was un fundamental lawal isnt so important for our purposes here.What is important is the rhetoric that lead the court to its conclusion. For the finale hung crucially on claims about the architecture of the net as it was on the architecture, that is, of Net95. Given that architecture, the court concluded, any regulation that attempted to geographical zone kids from porn would be a regulation that was too burdensome on speakers and listeners. As the net was, regulation would be too burdensome. But what was significant was that the court r as if this architecture of the net as it was Net 95 was the only architecture that the net could have.It spoke as if it had discover the nature of the net, and was therefore deciding the nature of any possible regulation of the net. 7 Lessig The Laws of Cyberspace Draft April 3, 1998 But the problem with all this, of course, is that the net has no nature. There is no single architecture that is essential to the nets form. Net95 is a set of features, or protocols, that constituted the net at one period of time. But nothing requires that these features, or protocols, always constitute the net as it always will be.And indeed, nothing in what weve seen in the last 2 years should lead us to think that it will. An archetype may make the point more simply. Before I was a professor at Harvard, I taught at the University of Chicago. If one treasured to gain access to the net at the university of Chicago, one simply imputeed ones instrument to jacks located throughout the university. Any machine could be connected to those jacks, and once connected, any machine would then have full access to the net in progress. Access was anon., and complete, and free. The reason for this freedom was a decision by the administration.For the Provost of the University of Chicago is Geof Stone, a former dean of the University of Chicago Law School, and a prominent free speech scholar. When the University was designing its net, the technicians asked the provost whether anonymous dialogue should be permitted. The provost, citing a principle that the rules regulating speech at the university would be as protective of free speech as the first amendment, express yes One would have the right to communicate at the university anonymously, because the first amendment to the constitution would guarantee the same right vis-a-vis the semipolitical sympathies.From that policy decision flowed the architectural design of the University of Chicagos net. At Harvard, the rules are different. One cannot connect ones machine to the net at Harvard unless ones machine is immortalizeed licensed, approved, verified. Only members of the university community can register their ma chine. Once registered, all interactions with the network are potentially monitored, and identified to a particular machine. Indeed, anonymous speech on this net is not permitted against the rule. Access can be surmountled based on who somebody is and interaction can be traced, based on what someone did.The reason for this design is also due to the decision of an administrator though this time an administrator less focused on the protections of the first amendment. Controlling access is the ideal at Harvard facilitating access was the ideal at Chicago tech- 8 Lessig The Laws of Cyberspace Draft April 3, 1998 nologies that make verify possible were therefore elect at Harvard technologies that facilitate access chosen at Chicago. Now this difference between the two networks is quite coarse today. The network at the University of Chicago is the architecture of the internet in 1995.It is, again, Net95. But the architecture at Harvard is not an internet architecture. It is rather an intranet architecture. The difference is simply this that within an intranet, identity is sufficiently established such that access can be controlled, and usage monitored. The underlying protocols are still TCP/IP meaning the fundamental or underlying protocols of the internet. But layered on top of this fundamental protocol is a set of protocols facilitating control. The Harvard network is the internet plus, where the plus mean the force play to control.These two architectures reflect two philosophies about access. They reflect two sets of principles, or set, about how speech should be controlled. They parallel, I want to argue, the difference between political regimes of freedom, and political regimes of control. They track the difference in ideology between West and East Germany between the United States and the former Soviet Republic between the Republic of China, and Mainland China. They stand for a difference between control and freedom and they manifest this difference through the architecture or design of code.These architectures enable political set. They are in this sense political. Now I dont offer this example to criticize Harvard. Harvard is a private institution it is free, in a free society, to allocate its resources however it wishes. My point rather is simply to get you to see how architectures are many, and therefore how the choice of one is political. And how, at the level of a nation, architecture is inherently political. In the world of cyberspace, the selection of an architecture is as important as the choice of a constitution.For in a fundamental sense, the code of cyberspace is its constitution. It sets the terms upon which people get access it sets the rules it controls their behavior. In this sense, it is its own sovereignty. An alternative sovereignty, competing with real space sovereigns, in the regulation of behavior by real space citizens. But the United States Supreme Court treated the question of architecture as if the a rchitecture of this space were given. It spoke as if there were only one design for cyberspace the design it had. 9 Lessig The Laws of CyberspaceDraft April 3, 1998 In this, the Supreme Court is not alone. For in my view, the single greatest error of theorists of cyberspace of pundits, and especially lawyers idea about regulation in this space is this error of the Supreme Court. It is the error of naturalism as applied to cyberspace. It is the error of thinking that the architecture as we have it is an architecture that we will always have that the space will guarantee us liberty, or freedom that it will of necessity disable political sympathiess that want control. This view is profoundly mistaken.Profoundly mistaken because while we celebrate the inherent freedom of the net, the architecture of the net is changing from under us. The architecture is shifting from an architecture of freedom to an architecture of control. It is shifting already without brasss intervention, thoug h brass is quickly coming to see just how it might intervene to renovate it. And where government is now intervening, it is intervening in a way knowing to re get this very same architecture to change it into an architecture of control, to make it, as Ive said, more regulable.While pundits promise perpetual freedom reinforced into the very architecture of the net itself, technicians and politicians are working together to change that architecture, to move it away from this architecture of freedom. As theorists of this space, we must come to understand this change. We must recognize the political consequences of this change. And we must take responsibility for these consequences. For the escape of the change is unmistakable, and the fruit of this trajectory, poison. As ingrainedists, we must then confront a fundamentally constitutional uestion if there is a choice between architectures of control and architectures of freedom, then how do we decide these constitutional questio ns? If architectures are many, then does the constitution itself guide us in the selection of such architectures? In my view, constitutional values do implicate the architecture of this space. In my view, constitutional values should guide us in our design of this space. And in my view, constitutional values should limit the types of regulability that this architecture permits. But my view is absent in thinking about governments role in cyberspace.Indeed, my nation for many years the symbol of freedom in world where such freedom was rare has become a leader in pushing the architecture of the internet from an archi- 10 Lessig The Laws of Cyberspace Draft April 3, 1998 tecture of freedom to an architecture of control. From an architecture, that is, that embraced the traditions of freedom expressed in our constitutional past, to an architecture that is fundamentally anathema to those traditions. But how? How can the government make these changes? How could the government effect this control? many an(prenominal) cant see how government could effect this control.So in the a few(prenominal) minutes remaining in my talk today, I want show you how. I want to sketch for you a path from where we are to where I fear we are going. I want you to see how these changes are possible and how government can help make them permanent. Return then with me to the idea that began this essay the point about the different modalities of constraint and notice something important about that idea that we have not so far remarked. I said at the start that we should think of law as just one of four modalities of constraint that we should think of it as just one part of the structure of constraint that might be said to regulate.One might take that to be an argument about laws insignificance. If so many forces other than law regulate, this might suggest that law itself can do relatively little. But notice what should be obvious. In the mannequin I have pulld law is regulating by direc t regulation regulating an individual through the threat of punishment. But law regulates in other ways as well. It regulates, that is, indirectly as well as directly. And it regulates indirectly when it regulates these other modalities of constraint, so that they regulate otherwise.It can, that is, regulate norms, so norms regulate differently it can regulate the market, so that the market regulates differently and it can regulate architecture, so that architecture regulates differently. In each case, the government can coopt the other structures, so that they constrain to the governments end. The same indirection is possible in cyberspace. But here, I suggest, the indirection will be even more significant. For here the government can not only regulate indirectly to advance a particular substantive end of the government. More significantly, the government can regulate to change the very regulability of the space.The government, that is, can regulate the architectures of cyberspac e, so that behavior in cyberspace becomes more regulable 11 Lessig The Laws of Cyberspace Draft April 3, 1998 indeed, to an architecture potentially more regulable than anything we have known in the history of modern government. Two examples will make the point one an example of the government regulating to a particular substantive end, and the second, side by side(p) from the first, an example of the government regulating to increase regulability. The first is the regulation of encryption.The governments concern with encryption has been with the technologys use in protecting privacy its ability to hide the content of communications from the eye of an eavesdropping third party, whether that third party is the government, or a nosy neighbor. For much of the history of the technology, the American government has heavily correct the technology for a time it threatened to ban its use it has consistently banned its trade (as if only Americans understand higher order mathematics) an d for a period it hoped to flood the market with a standard encryption technology that would leave a backdoor open for the government to enter.The most recent proposals are the most significant. be November, the FBI proposed a law that would require manufacturers to assure that any encryption system have built within it either a key convalescence ability, or an equivalent back door, so that government agents could, if they need, get access to the content of such communications. This is governments regulation of code, indirectly to regulate behavior. It is indirect regulation in the sense that I described before, and from a constitutional perspective it is brilliant.Not brilliant because its ends are good brilliant because the American constitution, at least, offers very little control over government regulation like this. The American constitution offers little protections against the governments regulation of business and given the interests of business, such regulations are lik ely to be effective. My second example follows from the first. For a second use of encryption is identification as well as hiding what someone says, encryption, through digital certificates, can be used to authenticate who some it.With the ability to authenticate who someone is, the government could tell where someone comes from, or how old they are. And with this ability through certifying IDs passports on the information superhighway governments could far more easily regulate behavior on this highway. 12 Lessig The Laws of Cyberspace Draft April 3, 1998 It would recreate the power to control behavior recreate the power to regulate. Note what both regulations would achieve. Since the US is the largest market for internet products, no product could hope to succeed unless it were happy in the United States.Thus standards successfully imposed in the US becomes standards for the world. And these standards in particular would first facilitate regulation, and second, assure that co mmunications on the internet could be broken into by any government that followed the procedures outlined in the bill. But the standards that those government would have to meet are not the standards of the US constitution. They are whatever standard local government happen to have whether that government be the government of Mainland China, or Switzerland.The effect is that the United States government would be exporting an architecture that facilitates control, and control not just by other democratic governments, but by any government, however repressive. And by this, the US would move itself from a symbol of freedom, to a peddler of control. Having won the cold war, we would be pushing the techniques of our cold war enemies. *** How should we respond? How should you as sovereigns independent of the influence of any foreign government and we, as spare constitutionalists respond?How should we respond to moves by a dominant political and economic power to influence the archite cture of the dominant architecture of regulation by code the internet? Sovereigns must come to see this That the code of cyberspace is itself a kind of sovereign. It is a competing sovereign. The code is itself a force that imposes its own rules on people who are there, but the people who are there are also the people who are here citizens of the Republic of China, citizens of France, citizens of every nation in the world. The code regulates them, yet they are by right subject to the regulation of local sovereigns.The code thus competes with the regulatory power of local sovereigns. It competes with the political choices made by local sovereigns. And in this competition, as the net becomes a dominant place for business and social life, it will displace the regulations of local sovereigns. You as sovereigns were afraid of the competing influence of na- 13 Lessig The Laws of Cyberspace Draft April 3, 1998 tions. Yet a new nation is now pumped-up(a) into your telephones, and its inf luence over your citizens is growing. You, as sovereigns, will come to recognize this competition. And you should come to recognize and question the special ole that the United States is playing in this competition. By virtue of the distribution of resources controlling the architecture of the net, the United States has a unparalleled power over influencing the development of that architecture. It is as the law of nature were being written, with the United States at the authors side. This power creates an important responsibility for the United States and you must assure that it exercises its power responsibly. The problem for constitutionalists those concerned to preserve social and political liberties in this new space is more difficult.For return to the story that began this talk the world of internal passports. One way to understand the story Ive told today about cyberspace is in line with this story about the Tsars Russia. The birth of the net was the revolution itself lif e under Net95 was life in Bolshevik Russia (the good parts at least, where internal passports were eliminated) the Net as it is becoming is Stalins Russia, where internal passports will again be required. Now theres a cheat to that story a rhetorical cheat that tends to obscure an important fact about real space life.For we all live in the world of internal passports. In the United States, in many places, one cannot live without a car one cant drive a car without a license a license is an internal passport It says who you are, where you come from, how old you are, whether youve recently been convicted of a crime it links your identity to a database that will reveal whether youve been arrested (whether convicted or not) or whether any warrants for your arrest in any jurisdiction in the nation are outstanding. The license is the internal passport of the modern American state.And no doubt its ability to control or identify is far wagerer than the Tsars Russia. But in the United State s at least for those who dont appear to be immigrants, or a disfavored minority the burden of these passports is slight. The will to regulate, to monitor, to track, is not strong fair to middling in the United States to support any systematic effort to use these passports to control behavior. And the will is not strong enough because the cost of such control is so great. There are not checkpoints at each corner one isnt required to register 14Lessig The Laws of Cyberspace Draft April 3, 1998 when moving through a city one can walk relatively anonymously around most of the time. Technologies of control are possible, but in the main far too costly. And this dearness is, in large part, the source of great freedom. It is inefficiency in real space technologies of control that yield real space liberty. But what if the cost of control drops dramatically. What if an architecture emerges that permits constant monitoring an architecture that facilitates the constant tracking of behavior and movement.What if an architecture emerged that would costlessly collect data about individuals, about their behavior, about who they wanted to become. And what if the architecture could do that invisibly, without interfering with an individuals daily life at all? This architecture is the world that the net is becoming. This is the picture of control it is growing into. As in real space, we will have passports in cyberspace. As in real space, these passports can be used to track our behavior. But in cyberspace, unlike realspace, this monitoring, this tracking, this control of behavior, will all be much less expensive.This control will occur in the background, effectively and invisibly. Now to describe this change is not to say whether it is for the good or bad. Indeed, I suggest that as constitutionalists, we must acknowledge a fundamental ambiguity in our present political judgments about liberty and control. I our peoples are change integrity in their reaction to this picture of a system of control at once perfect, and yet invisible. Many would say of this system wonderful. All the better to trap the guilty, with little burden on the innocent. But there are many as well who would say of this system awful.That while professing our ideals of liberty and freedom from government, we would have established a system of control far more effective than any in history before. So the reply to all this is not necessarily to give up the technologies of control. The response is not to insist that Net95 be the perpetual architecture of the net. The response instead is to find a way to translate what is salient and important about present day liberties and constitutional democracy into this architecture of the net. The point is to be critical of the power of this sovereignthis emerging sovereignas we are properly critical of the power of any sovereign.What are these limits As government takes control or influences the architecture of the code of the net, at a minim um, we 15 Lessig The Laws of Cyberspace Draft April 3, 1998 must assure that government does not get a monopoly on these technologies of control. We must assure that the sorts of checks that we build into any constitutional democracy get built into regulation by this constitution the code. We must assure that the constraints of any constitutional democracy the limits on efficiency constituted by Bills of Rights, and systems of checks and balances get built into regulation by code.These limits are the bugs in the code of a constitutional democracy and as John Perry Barlow says, we must build these bugs into the code of cyberspace. We must build them in so that they, by their inefficiency, might recreate some of the protections we have long known. *** Cyberspace is regulated ? by laws, but not just by law. The code of cyberspace is one of these laws. We must come to see how this code is an emerging sovereign omnipresent, omnipotent, gentle, efficient, growing and that we must de velop against this sovereign the limits that we have developed against real space sovereigns.Sovereigns will always say real space as well as cyberspace that limits, and inefficiencies bugs are not necessary. But things move too quickly for such confidence. My fear is not just that against this sovereign, we have not yet developed a language of liberty. Nor that we havent the time to develop such language. But my fear is that we sustain the will the will of free societies for the past two centuries, to architect constitutions to protect freedom, efficiencies notwithstanding. 16