Cyberspace's Legal Visionary

Lawrence Lessig on the fate of copyrights and computer networks in the digital future.

Few intellectuals have influenced the way people think about cyberspace as much as Lawrence Lessig. His first book, Code and Other Laws of Cyberspace (Basic Books, 1999), was described by one critic as "a direct assault on the libertarian perspective that informs much Internet policy debate." At the same time, by probing the complicated relationship between "East Coast code" -- that is, legislation -- and the "West Coast code" that creates the architecture of the Internet, the book changed the terms of the debate, influencing even the people Lessig was criticizing.

Last year's follow-up, The Future of Ideas: The Fate of the Commons in a Connected World (Random House), asks what allows innovation to take place online, examining such issues as Napster, software patents, and regulation of the electromagnetic spectrum. By explaining technology to the lawyers and law to the technologists, Lessig has deepened a lot of people's understanding of the Net. Few in the public policy community, for example, have given much thought to the different layers of cyberspace. By contrast, Lessig distinguishes the physical layer (the network's hardware and wires) from the logical layer (the protocols that determine who connects to what) and the content layer (the actual material delivered by the protocols over the wires). To maintain our ability to innovate online, he argues, one must recognize the different relationship public policy has with each layer of the system.

In the meantime, Lessig argues fiercely for preserving an online commons, a concept he distinguishes (but does not completely disentangle) from government control. In the wake of the socialist collapse, he writes, "the issue for us will not be which system of exclusive control -- the government or the market -- should govern a given resource. The question for us comes before: not whether the market or the state but, for any given resource, whether that resource should be controlled or free."

Lessig, 40, teaches law at Stanford University. He spoke with Associate Editor Jesse Walker in January.

Reason: What do you mean by "the commons"?

Lawrence Lessig: A resource is in the commons as long as access to it is not governed by someone's subjective decision. A free resource -- free as in zero price -- fits that definition. Even a resource that is not zero-priced could be a commons, so long as the terms on which it is offered are completely neutral.

In our culture, whenever we think about commons, we instantly affix the idea of tragedy. But logically, that can only be true if there is some rivalrousness about that property -- if my use of it interferes with your ability to use it the same way. Obviously, intellectual goods are not like that. My use of my poems doesn't stop your use of my poems. So if a commons is nonrivalrous, there can be no tragedy, because my consumption of it does not reduce the amount available to you.

That doesn't mean we don't have to worry about regulating access to those resources. The question is, How do you create enough incentive to produce that resource if everybody can take it without paying for it? That's where excludability becomes important. Once I give a bit of information to the world, it's hard for me to exclude anybody from getting access to that information.

Intellectual property deals with the problem of non-excludability by saying, We're going to give a government-backed monopoly right for a limited term to assure there is enough incentive for people to produce. But it shouldn't be expanded so broadly as to create a false protection for rivalrousness.

Reason: Another central concept in The Future of Ideas is the end-to-end principle.

Lessig: End-to-end is a design principle that network architects began to articulate around the early 1980s. The idea is that you should place intelligence at the edge of a network and keep the network itself simple.

This made the evolution of the network much more flexible, because it requires a minimum amount of coordination among network owners and users. The way I am trying to develop the concept, it also has very important political consequences: The end-to-end principle enables maximum liberty for users.

As with any principle, this isn't absolute. But if you have two solutions and one is compliant with end-to-end and the other is not, that's a good reason to pick the first.

Reason: You believe we're moving away from this principle now.

Lessig: In my first book I was quite pessimistic. It turns out I was not pessimistic enough.

The thing I'm most worried about is what happens as the network moves from an essentially common carriage-regulated medium to pipes that are unregulated and increasingly encouraged to discriminate for or against the content they serve.

That's a prediction of a negative consequence that will come about, and the way to test it is to see what things will resist it. One thing that would resist it is a strong competition among broadband providers, which has not happened. The other possibility would be a different kind of physical infrastructure which would compete with cable and telecom. Wireless could be a powerful competitor in providing Internet access. That would be a reason to be optimistic, if we saw a much greater embrace of that by the Federal Communications Commission.

Reason: Most of the concrete policy proposals in The Future of Ideas, especially in the areas of intellectual property and spectrum regulation, would loosen government controls. But not all of them. To what extent are you willing to use the government to enforce the idea of open architecture?

Lessig: Not unless necessary.

Reason: When do you think it becomes necessary?

Lessig: For example, when the physical layer doesn't have sufficient competition to assure that any particular actor can't behave strategically, from a competitive sense, to corrupt the core of the Internet. Then I think there's a role for the government to compensate for that lack of competition, either through open-access regimes, which have not proved successful, or through some kind of minimal rules about what it means to deliver Internet protocol packets. Companies could build the systems however they want, they wouldn't have to let anybody share their lines, but they'd have to deliver Internet protocol packets in a neutral way.

Reason: Of the economic interests you discuss, shouldn't the owners of the physical level have the strongest property claim? Wires are rivalrous and excludable, and the companies invested a great deal of money laying them.

Lessig: That's a legitimate argument. It's not simple, though, because they laid the wires in the context of protective regulation. The original monopoly deals for cable have been relaxed, of course. And if we're not talking about the Internet, I really have no concern with what cable does on their wires. They can block as much as they want; they can charge whatever they want. What I'm concerned about is what happens when cable adopts the Internet and then develops it in a way that no longer respects the end-to-end protocol. It's polluting the general standard that otherwise was producing a lot of innovation around the Internet.

Reason: How useful is it to think of intellectual property as property?

Lessig: If you're a lawyer, it's OK to think of intellectual property as property, because we're trained to use the word property in a careful way. We don't think of it as an absolute, perpetual right that can't be trumped by anybody. We understand property rights are constantly limited by public-use exceptions and needs, and in that context we understand intellectual property to be a very particular, peculiar kind of property -- the only property constitutionally required to be for limited terms. It's clearly established for a public purpose and is not a natural right.

The real problem is when people use it in the ordinary sense of the term property, which is "a thing that I have that nobody can take, forever, unless I give it to you." By thinking of it as property, we have no resistance to the idea of certain great companies controlling "their" intellectual property forever. But if we instead use terms like monopoly to describe the control that companies like Disney have over art objects like Mickey Mouse, it's harder to run naturally to the idea that you ought to have your monopoly right forever.

Another problem is the increasing ability of owners of intellectual property to control the actual use of that property. Before the network, if you bought a book, the First Sale doctrine made it impossible for the copyright owner to control what you did with it. Copyright law would not interfere with my ability to give you the book or copy a chapter or read the book a thousand times. All those things are completely within my control, partly because the law guarantees it, but also because the book producer couldn't do anything about it even if he wanted.

As you move to the Internet, though, lots more control is possible. If you look in the permissions for an Adobe eBook, it has the power to control how many times you can print certain sections of the book, whether you can use technology to read the book aloud, whether you can cut and paste sections of the book. All these controls would not have existed without digital technology, because the cost of regulating those uses would have been too high.

The convergence of this technology with rhetoric about property makes it seem increasingly natural that companies should be able to control the use of their "property" just like it seems natural that I should be able to control the use of my car. If I say you can use my car once every month, that means you can use it only once every month; you don't have the right to use it twice a month. But that's never been what intellectual property laws were about.

And now, if digital content has a built-in copy protection system, you aren't allowed to interfere with it, even if the content isn't protected by copyright laws. I have bought a number of eBooks, including Aristotle's Politics. Aristotle's Politics, of course, was never copyrighted, but the Adobe eBook reader forbids me from printing any pages of the book because the permissions have been set to disable any printing. If I try to interfere with those permissions -- if I write a bit of code to disable the limitations that forbid me from printing Aristotle's Politics from my Adobe eBook -- that would be circumventing an access technology, which under the Digital Millennium Copyright Act is a crime.

Reason: What do you think of the argument that the code itself is protected speech?

Lessig: I think code is protected speech, although under standard First Amendment doctrine, regulation of code would be analyzed as content-neutral regulation of speech, as the Second Circuit did in the 2600 case.

This was a challenge brought by 2600, the magazine, which had been prosecuted for posting links to technology to circumvent the encryption that protects content on DVDs. There were basically two claims here. One was that the Digital Millennium Copyright Act was overbroad because it prohibited uses that were perfectly legal. Second was a much more direct challenge to the idea that you could prosecute a magazine for linking to content that might be deemed illegal.

The court of appeals basically rejected both claims. In the context of the more substantial challenge -- the challenge to the regulation being overbroad -- it applied intermediate scrutiny under the First Amendment, which meant content-neutral regulation was appropriate. It also said there was sufficient government interest that justified the regulation, in spite of the effect on speech. I think that's where the court went wrong. I think the regulation was overbroad.

Reason: That seems stricter than the standard in physical space. Paladin Press and Loompanics are allowed to publish descriptions of ways to break the law that are far more serious than de-encrypting DVDs. Yet that's protected and the code is not.

Lessig: I don't think there's a good explanation for the difference.

The view behind the Digital Millennium Copyright Act is that if you don't perfectly protect the world from bad code, it would be the end of intellectual property. If one copy of a work is unencrypted and put on the Web, there'd be no further demand for that product anywhere. As a factual matter, that's just false. Napster didn't result in the end of the CD market. In fact, many people argue that it increased the sales of CDs.

More fundamentally, it's just skewed in its way of dealing with technologies that can be used for both good and bad purposes. There is a technology called the gun, which a number of children every day are killed with. But Smith & Wesson doesn't shake in its boots waiting for the FBI to come knocking at its door, because we have adopted the view that this technology has both good and bad uses, and rather than prosecuting the technology we'll prosecute the bad users. But in the context of copyright, if the technology has good and bad uses, the good ones don't matter.

I was giving a talk at the New America Foundation. Jack Valenti from the Motion Picture Association of America stood up and talked about how awful it was that Stanford allowed a Morpheus server to exist on its network, and what did I think of this obvious technology to enable stealing? I told him I had a Morpheus server running on my machine at work, which was delivering my own content, both audio and text versions of lectures I've given. When I first made this available, the Stanford network police came in and pulled the plug on my server over the weekend because they thought I was engaging in "illegal acts." But I wasn't engaging in illegal acts. It's completely legal for me to distribute content that I have the copyright over, and this technology makes it very easy to do that.

The idea that you would assume that all uses are illegal is an overreaction to what I think is a legitimate problem. I'm not in favor of copyright theft, I'm just opposed to shutting down all technologies merely because copyright theft may occur on them.

Reason: Do we need software patents?

Lessig: I think of patents as a form of regulation, a government-granted monopoly that should only be granted to the extent that it's doing some public good. That indicates to me that when you start changing patent policy to encourage patents on software, or on business methods, you should be answering the question, Do we have good reason to believe the good that this system will produce outweighs the well-understood costs? The answer is no, we don't. At best the data are mixed, and at worse they provide a pretty good reason to believe that at least with software and business-method patents, there is no substantial increase in the innovation that could outweigh the costs of regulation. Where are the Republicans when you need them?

That doesn't mean I don't believe that no patents can justify themselves. There are plenty of contexts where we have reason to believe that they do good -- pharmaceuticals are a perfect example.

Reason: One of your more controversial claims in Code was that the Communications Decency Act, which directly censored "indecent" speech online, was actually less offensive to free speech than privately adopted filtering software.

Lessig: Between no law and good law narrowly tailored to a legitimate state interest in protecting kids from porn, I would favor narrow good law. If we had narrow regulation, you would destroy the market for this bad censorware, and you would have a much better free speech environment than you have right now.

Censorware blocks all sorts of content on the Web, including content critical of censorware. It's awful not just because it's censoring more than porn but because it censors in a way that there is no opportunity to question the censorship. It's a secret list of damned books.

Reason: The software companies may want to keep their lists hidden, but in effect they're only secret by government fiat. The Future of Ideas describes how people were sued for reverse-engineering a censorware program to discover what was being blocked.

Lessig: There was a lawsuit, and copyright law was used to hide the list. But that's the existing world of censorware. I would much prefer a world where each time you wanted to add on an additional layer of mandatory filtering, we had an opportunity to question it, rather than a world where it develops privately and unreviewably through these technologies of censorship.

A parallel that's a bit more intuitive is spam. We have no effective regulation of spam right now. Congress has been talking about lots of bad regulation, which would either ban spam or impose all sorts of stupid requirements that say whenever you send an e-mail you have to make sure it complies with the rules of the recipient institution. But there's also what I think would be perfectly good regulation of spam -- essentially, a labeling requirement.

Because we don't have that regulation, there are all these private solutions to the problem. Some of them are these vigilantes, like the Realtime Blackhole List, who develop lists of e-mail servers that they don't think live up to the appropriate spam policy and then block e-mails coming from those servers, which creates all sorts of costs for providers who now have to decide whether they're going to cave into the spam policy that the Blackhole List wants to impose or if they're going to live up to their own policies. Or they encourage the development of these highly complicated technologies like on Outlook for filtering a bunch of junk folders on the basis of routines that are themselves never public.

Both of those would be rendered unnecessary by a rule that effectively said you needed to label your unsolicited commercial e-mail as spam. If people had to do that, it would be very easy for those who don't want to receive it to chuck it or have the server itself filter it out.

Reason: I get a fair amount of spam that is clearly illegal already under the rules governing fraud. How regulable is this realm to begin with?

Lessig: Here again is a lack of creativity on the part of the government. My proposal says label at one end and use a kind of bounty system for enforcement. Take all this energy devoted now to things like the Realtime Blackhole List and turn it to the creative process of tracking down spammers. If you did that, the problem would disappear very quickly.

Reason: Isn't there a definitional issue, too? Some people have a broader notion of spam than you do, and that's one reason why you have these different private alternatives. I get unsolicited non-commercial e-mail all the time -- one man's press release is another man's spam.

Lessig: There will always be definitional problems. There is a constitutional problem with how you'd be able to define the required labeling. But again -- and this may be my bias as a lawyer -- I think it's much better to have that argument in the open, where a court has to weigh it against public values, than to have it operating secretly in systems I have no way to question. If the Realtime Blackhole List decides it doesn't like the e-mail servers at Harvard and Stanford, you're stuck. Your mail is just chucked off the part of the Internet that subscribes to that service. That's a pretty coercive system to force everybody to comply to their rules, and I just don't presume that their rules are the best rules here.

Reason: How would you describe your politics?

Lessig: Fundamentally mixed. You could say confused. I was born a Randian, and I have strong views about the goodness of free markets. There's certain parts of my politics that are more liberal than conservative. I think of myself as a liberal.

Reason: When you were interviewed on Slashdot, you sometimes seemed angry at that site's community of open-source programmers and advocates. You said they were "political slugs," "pathetically apolitical," and addicted to "irrelevant bickerings." What were you getting at?

Lessig: In a sense I'm trying to shame them into doing something. These people, more than anybody, understand the values built into the Internet. But they are among the least politically active segment of our community, out of libertarianism or just a deep skepticism about the use of government. They just don't want to pay any attention to it, don't have any respect for what government does.

If they were to become more politically active, however distasteful that may be to them, that could begin to put a check on what is achievable by others who have no hesitation about being politically active.

Reason: Some of them would argue that they're active in the form of civil disobedience.

Lessig: In a world where civil disobedience was treated with toleration, that might be a good strategy. But we're in a world where disobedience is treated with felony convictions. The idea that you are going to get lots of civil disobedience against the Digital Millennium Copyright Act is just crazy. You're going to get lots of prosecutions and people going away to jail. The cost of disobedience has become too high, and I'm not sure it's a viable strategy anymore.

There's some basic cultural differences here. Many of the people who have great ideas in the Slashdot context about the way to run the world -- if you put them in Washington, they just don't fit.

Reason: Aren't you facing the same problem? Some of the proposals in Future are guaranteed to be political non-starters. The chances that Congress is going to adopt five-year renewable copyright terms in this political context are zero.

Lessig: I am not writing because I think it's likely that policy makers will sit down and figure this stuff out right now. I'm more writing about what I think is true, and hoping that eventually a group of people who have the time to think through it will try to do something about it.

Over 15 or 20 years, the movement that Reagan is associated with got the world to think about policy things differently, through many small chips at taken-for-granted assumptions about the world. So eventually it can happen.

Reason: How did a lawyer get interested in technology?

Lessig: My first work in constitutional law was in Eastern Europe. There I learned that constitutional law is about trying to set up structures that embed certain values within a political system. Once I started thinking about constitutional law like that, it was a tiny step to see that that's exactly what the architecture of cyberspace does: It's a set of structures embedding a set of values. To the extent that we like those values, we ought to be defending the architecture of cyberspace. To the extent that we're skeptical about those values, we should be asking whether the architecture is justified or not. Either way, the architecture is analogous to the Constitution.

So it was a simple step. Also, the conferences had better coffee.