For constitutionalists (as I am), this was a heady time. I had just graduated from law school in 1989, and in 1991 I began teaching at the University of Chicago. Chicago had a center devoted to the study of the emerging democracies in Central and Eastern Europe. I was a part of that center. Over the next five years I spent more hours on airplanes, and more mornings drinking bad coffee, than I care to remember.

Central and Eastern Europe were filled with Americans telling former communists how they should govern. The advice was endless and silly. Some of these visitors literally sold constitutions to the emerging constitutional republics; the balance had innumerable half-baked ideas about how the new nations should be governed. These Americans came from a nation where constitutionalism had worked, yet apparently they had no clue why.

The center's mission, however, was not to advise. We knew too little to guide. Our aim was to watch and gather data about the transitions and how they progressed. We wanted to understand the change, not direct it.

What we saw was striking, if understandable. Those first moments after communism's collapse were filled with antigovernmental passion - with a surge of anger directed against the state and state regulation. Leave us alone, the people seemed to say. Let the market and nongovernmental organizations - a new society - take the government's place. After generations of communism, this reaction was completely understandable. What compromise could there be with the instrument of your repression?

A certain American rhetoric supported much in this reaction - a rhetoric of libertarianism. Just let the market reign and keep the government out of the way, and freedom and prosperity will inevitably grow. Things will take care of themselves. There is no need, and can be no place, for extensive regulation by the state.

But things didn't take care of themselves. Markets didn't flourish. Governments were crippled, and crippled governments are no elixir for freedom. Power didn't disappear; it simply shifted from the state to mafiosi, themselves often created by the state. The need for traditional state functions - police, courts, schools, health care - didn't magically go away. Private interests didn't emerge to fill the need. Instead, needs were unmet. Security evaporated. A modern if plodding anarchy replaced the bland communism of the previous three generations: Neon lights flashed advertisements for Nike; pensioners were swindled out of their life savings by fraudulent stock deals; bankers were murdered in broad daylight on Moscow streets. One system of control had been replaced by another, but neither system was what Western libertarians would call freedom.

At just about the time this postcommunist euphoria was waning - in the mid-1990s - there emerged in the West another "new society," which was to many people just as exciting as the new societies promised in postcommunist Europe. This was cyberspace. First in universities and centers of research, and then within society in general, cyberspace became the new target of libertarian utopianism. Here freedom from the state would reign. If not in Moscow or Tblisi, then here in cyberspace would we find the ideal libertarian society.

The catalyst for this change was likewise unplanned. Born in a research project by the U.S. Department of Defense, cyberspace too arose from the displacement of a certain architecture of control. The tolled, single-purpose network of telephones was displaced by the untolled and multipurpose network of packet-switched data. And thus the old one-to-many architectures of publishing (television, radio, newspapers, books) were supplemented by a world in which everyone could be a publisher. People could communicate and associate in ways they never had before. The world then promised a kind of society that real space could never allow - freedom without anarchy, control without government, consensus without power. In the words of a manifesto by MIT research scientist David Clark that will define our generation: "We reject: kings, presidents and voting. We believe in: rough consensus and running code."

As in postcommunist Europe, first thoughts about cyberspace tie freedom to the disappearance of the state. But here the bond was even stronger than in postcommunist Europe. The claim now was that government could not regulate cyberspace, that cyberspace was essentially, and unavoidably, free. Governments could threaten, but behavior could not be controlled; laws could be passed, but they would be meaningless. There was no choice about which government to install; none could reign. Cyberspace would be a society of a very different sort. There would be definition and direction, but built from the bottom up, and never through the direction of a state. The society of this space would be a fully self-ordering entity, cleansed of governors and free from political hacks.

I taught in Central Europe during several summers in the early 1990s; I witnessed the transformation in attitudes about communism that I described earlier. And so I felt a bit of deja vu when in the spring of 1995 I began to teach the law of cyberspace, and saw in my students these very same postcommunist thoughts about freedom and government. Even at Yale - not known for libertarian passions - the students seemed drunk with what author James Boyle would later call the "libertarian gotcha": No government could survive without the Internet's riches, yet no government could control what went on there. Real-space governments would become as pathetic as the last communist regimes. It was the withering of the state that Marx had promised, jolted out of existence by trillions of gigabytes flashing across the ether of cyberspace. Cyberspace, the story went, could only be free. Freedom was its nature.

But why was never made clear. That cyberspace was a place governments could not control was an idea I never quite got. The word itself speaks not of freedom but of control. Its etymology reaches beyond a novel by William Gibson (Neuromancer, published in 1984) to the world of cybernetics, the study of control at a distance. Cybernetics had a vision of perfect regulation. Its very motivation was finding a better way to direct. Thus, it was doubly odd to see this celebration of noncontrol over architectures born from the very ideal of control.

As I said, I am a constitutionalist. I teach and write about constitutional law. I believe that these first thoughts about government and cyberspace are just as misguided as the first thoughts about government after communism. Liberty in cyberspace will not come from the absence of the state. Liberty there, as anywhere, will come from a state of a certain kind. We build a world in which freedom can flourish not by removing from society any self-conscious control; we build a world in which freedom can flourish by setting it in a place where a particular kind of self-conscious control survives. We build liberty, as our founders did, by setting society upon a certain constitution.

But by "constitution" I don't mean a legal text. Unlike my compatriots in Eastern Europe, I am not trying to sell a document that our framers wrote in 1787. Rather, as the British understand when they speak of their constitution, I mean an architecture - not just a legal text, but a way of life that structures and constrains social and legal power, to the end of protecting fundamental values - principles and ideals that reach beyond the compromises of ordinary politics.

Constitutions in this sense are built; they are not found. Foundations are laid; they don't magically appear. Just as the founders of our nation learned from the anarchy that followed the revolution (remember: our first constitution, the Articles of Confederation, was a miserable failure of do-nothingness), so too are we beginning to see in cyberspace that this building is not the work of an invisible hand. There is no reason to believe that the grounding for liberty in cyberspace will simply emerge. In fact, quite the opposite is the case. As our framers learned, and as the Russians saw, we have reason to believe that cyberspace, left to itself, will not fulfill the promise of freedom. Left to itself, cyberspace will become a perfect tool of control.

Control. Not necessarily control by government, and not necessarily control for some evil, fascist end. The invisible hand of cyberspace is building an architecture that's quite the opposite of what it was at cyberspace's birth. The invisible hand, through commerce, is constructing an architecture that perfects control - an architecture that makes possible highly efficient regulation. As science-fiction author Vernor Vinge warned in 1996 at the Conference on Computers, Freedom and Privacy, a distributed architecture of regulatory control; as author Tom Maddox added, an axis between commerce and the state.

My book, Code, is about that change, and about how we might prevent it. When we see the path that cyberspace is on, we see that much of the "liberty" present at cyberspace's founding will vanish in its future. Values that we now consider fundamental will not necessarily remain. Freedoms that were foundational will slowly disappear.

If the original cyberspace is to survive, and if values that we knew in that world are to remain, we must understand how this change happens, and what we can do in response. Cyberspace presents something new for those people who think about regulation and freedom. It demands a new understanding of how regulation works and what regulates life there. It compels us to look beyond the traditional lawyer's scope - beyond laws, regulations and norms. It requires an account of a newly salient regulator.

That regulator is the obscurity in my book's title - Code. In real space we recognize how laws regulate through constitutions, statutes and other legal codes. In cyberspace we must understand how code regulates - how the software and hardware that make cyberspace what it is regulate cyberspace as it is. As William Mitchell put it in City of Bits, this code is cyberspace's "law." Code is law.

This code presents the greatest threat to liberal or libertarian ideals, as well as their greatest promise. We can build, or architect, or code cyberspace to protect values we believe are fundamental - or we can build, or architect, or code cyberspace to allow those values to disappear. There is no middle ground. There is no choice that does not include some kind of building. Code is never found; it is only made, and only made by us. As Mark Stefik wrote in Internet Dreams, "Different versions of [cyberspace] support different kinds of dreams. We choose, wisely or not."

My argument is not for some top-down form of control; my claim is not that regulators must occupy Microsoft. A constitution envisions an environment. As Supreme Court Justice Oliver Wendell Holmes said in 1920, it "call[s] into life a being the development of which [cannot be] foreseen." Thus, to speak of a constitution is not to describe a 100-day plan. It is instead to identify the values that a space should guarantee. It is not to describe a "government"; it is not even to select (as if a single choice must be made) between bottom-up or top-down control. In speaking of a constitution in cyberspace we are simply asking, What values are protected there? What values will we build into the space to encourage certain forms of life?

There are two types of "values" here: substantive and structural. In the American tradition, we worried about the second first. The framers of the Constitution of 1787 (enacted without a Bill of Rights) were focused on structures of government. Their aim was to ensure that a particular government (the federal government) did not become too powerful. And so they built into its design checks on its power and limits on its reach over the states.

Opponents of that Constitution insisted that more checks were needed, that the Constitution needed to impose substantive limits on the government's power as well as structural limits. And thus the Bill of Rights was born. Ratified in 1791, the Bill of Rights promised that the federal government will not remove certain protections - of speech, privacy and due process. And it guarantees that the commitment to these substantive values will remain despite the passing fancy of normal government. These values were to be entrenched, or embedded, in our constitutional design; they can be changed, but only by changing the Constitution's design.

These two kinds of protection go together in our constitutional tradition. One would have been meaningless without the other. An unchecked structure easily could have overturned the substantive protections expressed in the Bill of Rights, and without substantive protections, even a balanced and reflective government could have violated values that our framers thought fundamental.

We face the same questions in constituting cyberspace, but we have approached them from an opposite direction. Already we are struggling with substance: Will cyberspace promise privacy or access? Will it preserve a space for free speech? Will it facilitate free and open trade? These are choices of substantive value.

But structure matters as well. What checks on arbitrary regulatory power can we build into the design of the space? What "checks and balances" are possible? How do we separate powers? How do we ensure that one regulator, or one government, doesn't become too powerful?

Theorists of cyberspace have been posing these questions since its birth. But as a culture, we are just beginning to get it. We are just beginning to see why the architecture of the space matters - in particular, why the ownership of that architecture matters. If the code of cyberspace is owned, it can be controlled; if it is not owned, control is much more difficult. The lack of ownership, the absence of property, the inability to direct how ideas will be used - the presence of a commons - is key to limiting, or checking, certain forms of governmental control.

One part of this question of ownership is at the core of the current debate between open- and closed-source software. In a way the American founders would have instinctively understood, free software or open-source software - or open code, to (cowardly) avoid taking sides - is itself a check on arbitrary power. A structural guarantee of constitutionalized liberty, open source functions as a type of separation of powers in the American constitutional tradition. It stands alongside substantive protections, like freedom of speech or of the press, but its stance is more fundamental. The first intuition of our founders was right: Structure builds substance. Guarantee the structural (a space in cyberspace for open code), and (much of) the substance will take care of itself.

There are four areas of controversy - intellectual property, privacy, free speech and sovereignty - and values within each that are now at risk. The interaction between law and code helps constitute these values.

But are we, meaning Americans, up to the challenge that these choices present? Given our present tradition in constitutional law and our present faith in representative government, are we able to respond collectively to the changes I describe?

My strong sense is that we are not. We are at a stage in our history when we urgently need to make fundamental choices about values, but we trust no government institution to make such choices. Courts cannot do it, because as a legal culture we don't want courts choosing among contested matters of values. Congress should not do it because, as a political culture, we so deeply question the products of ordinary government.

Change is possible. I don't doubt that revolutions remain in our future; the open-code movement is just such a revolution. But I fear that it is too easy for the government to dislodge these revolutions, and that too much will be at stake for it to allow the revolutionaries to succeed. Our government has already criminalized the core ethic of this movement, transforming the meaning of hacker into something quite alien to its original sense. This is only the beginning.

Things could be different. They are different elsewhere. But I don't see how they could be different for us just now. This no doubt is a simple confession of the limits of my own imagination. I would be grateful to be proven wrong. I would be grateful to watch as we relearn - as the citizens of the former communist republics are learning - how to escape our disabling ideas about the possibilities for governance.

Lawrence Lessig is Berkman Professor of Law at Harvard Law School. This article was excerpted from Code and Other Laws of Cyberspace (Basic Books, $30), to be published this month.