You are using an outdated browser.
Please upgrade your browser
and improve your visit to our site.
Skip Navigation

Code Comfort

The Control Revolution
by Andrew L. Shapiro
(Perseus, 286 pp., $25)

Code and Other Laws of Cyberspace
by Lawrence Lessig
(Basic Books, 230 pp., $30)


I have a number of "bookmarks" on my Netscape program. Many of them come from Netscape itself, and I have not changed them. (I am not sure that I could.) These include Nextcard Internet Visa, Toshiba, FTD Flowers, CBS sportsline and ABC news (not CBS news and not ABC sports, and nothing from NBC), and Netscape Channel itself. My personal bookmarks--the ones that I have added myself--include (a site that can be used also as a research tool), The Washington Post (with regular updates of news, weather, and sports), (a site for renting or buying movies), and United Airlines (finding routes and times for a large number of carriers, along with the cheapest fares; United flights are listed first). A somewhat more exotic bookmark connects to the Social Science Research Network (SSRN), which contains over 16,000 abstracts on social science subjects and over 4,000 "downloadable" academic papers, that is, papers that you can print out and have, in full, just by clicking. Some of these papers were posted (published?) on SSRN within the last few days; some of them have already been downloaded thousands of times. (SSRN tells you exactly how many times.) I also have a bookmark for several sites involving Rhodesian Ridgeback dogs (I learned about the breed and found my new puppy via such a site); horse stables in Illinois (my daughter rides horses and we found out about her new stable this way); and a variety of federal regulatory agencies, including the Occupational Safety and Health Administration, the Food and Drug Administration, and the Environmental Protection Agency, whose sites give you an ability to read and to print out thousands of regulations, press releases, and proposals.

The rise of the Internet must count as one of the most astonishing technological developments of the century. This is not least because nothing remotely like it was anticipated by those who created it, and because this supposedly government-free zone was a creation of the national government. We are used to hearing tales of the unintended bad consequences of government action; but the Internet is an unintended good consequence of government action. By the Department of Defense, no less: beginning in the 1960s, the Advanced Research Project Agency (ARPA) of the Department of Defense created a new computer network, originally called the Arpanet, with the specific purpose of permitting computers to interact with one another, thus allowing defense researchers at various universities to share computing resources. To the government's great surprise, the computer researchers used the Arpanet mostly to send e-mail, not to exchange data.

Starting in 1972, hundreds and then thousands of early users began to discover e-mail as a new basis for communication. In the early 1970s, the government sought to sell off the Arpanet to the private sector, contacting AT&T to see if it wanted to take over the Arpanet; but the company declined, concluding that the technology was incompatible with the AT&T network. (So much for the foresight of the private sector.) Eventually the Arpanet expanded to multiple uses, operating under the auspices of the National Science Foundation. By the late 1980s, a number of new networks emerged, some far more advanced than the Arpanet, and the term "Internet" came to be used for the federally subsidized network consisting of many linked networks running the same protocols. In 1989 the Arpanet was transferred to regional networks throughout the country.

A key innovation came one year later. Having been intrigued by the design and the potential of the Internet, researchers at CERN, the European Laboratory for Particle Physics near Geneva, created the World Wide Web, a multimedia branch of the Internet. (It is signalled on computers by the ubiquitous "www" after "http://" and before domain names; "http://" stands for "hypertext transfer protocol," which was designed for global information transfer and part of the original design of the World Wide Web.) CERN researchers attempted to interest private companies in building the World Wide Web, but they declined. ("Too complicated": this was another case of hilariously inadequate private sector foresight.) Tim Berners-Lee, the lead researcher and web inventor, had to build it on his own.

The consequence of the World Wide Web was to link the 800 computer networks, and 160,000 computers, then attached to the Internet. Soon thereafter researchers at the University of Illinois at Champaign-Urbana invented a breakthrough browser called Mosaic, which would make it easy for ordinary people, and not just experts, to use the Internet.

Hard as it now is to believe, the Internet began to become commercial only in the early 1990s, with legislation removing restrictions on commercial activity. It was during that time that government funding and oversight were almost entirely withdrawn. In 1995, the backbone of the national network was sold to a private consortium of corporations, and the government gave one company the exclusive right to register domain names. (You can now buy one for about $70.) Originally created by the government, the Internet is now largely free of federal supervision. To say the least, commercial activity has exploded. As late as 1996, educational sites dominated the list of the fifteen most often visited sites, and included no sites engaged in commerce; but almost all of the current "top fifteen" list engage in commercial activity. Moreover, the Internet, very much in use world-wide, is a distinctly American creation.

It is an understatement to say that the Internet has made life far more convenient for staggeringly many people. Though estimates are disputed, the sheer number of users grows rapidly every month. In 1990, there were 1.3 million total users; in 1992, 6.5 million; in 1994, 32 million; in 1996, 55 million; in 1997, 100 million; in 1998, 150 million; and at the present writing, over 200 million. More than 110 million Americans now use the Internet. In 1998, 99 million Windows-based personal computers were sold worldwide. In many families, stores have been to a greater or lesser degree replaced by websites; purchasing is increasingly a matter of pushing buttons. (In mid-1998, had 3.3 million customers; in mid-1999, it had nearly 11 million.) In some offices, telephones have become nearly obsolete, replaced by e-mail, which is quicker and more efficient. As every Internet parent knows, homework has been transformed; research projects often involve the World Wide Web rather than books. People can and do keep in touch with relatives and friends, with parents and children, in other states and other countries. On an amazing range of topics--Rhodesian Ridgebacks, horse stables in Illinois, recent environmental trends and regulations--information is astonishingly easy to get. Every home computer is now a world-class library.


It should be no surprise that libertarians and others have been celebrating the Internet as a government-free zone. This has been the overwhelming "first generation" response to the Internet. In 1996, John Perry Barlow produced a proclamation that he called a Declaration of the Independence of Cyberspace: "Governments of the Industrial World . . . I ask you of the past to leave us alone. You are not welcome among us. You have no sovereignty where we gather.... You have no moral right to rule us nor do you possess any methods of enforcement we have true reason to fear." But many people are uneasy about the emerging situation; and their best arguments point to possible harmful consequences for privacy, politics, and cultural life.

Andrew Shapiro and Lawrence Lessig have written impressive and worried books, rejecting libertarian homilies and suggesting, among other things, the need for a greater governmental role in control of the Internet. Shapiro and Lessig are democrats, arguing above all for more public consciousness about the dangers posed by the rise of the Internet; but they sharply diverge on whether the Internet offers more opportunity or more threat to individual control.

Shapiro's book is a mixed picture, with a heavy dose of optimism, even exhilaration. He celebrates what he calls "the control revolution," but he believes that it carries risk as well as promise. He hopes to devise ways for government to promote the promise while reducing the risk. What he urges is a kind of "balance" that would produce "a realistic compromise between personal liberty and individual obligation." Lessig's account is more learned, more ambitious, more innovative, and darker. He believes that the private power to create "code"--the architecture of the Internet--produces dangers to a number of important values, including privacy, free speech, sovereignty, and the open circulation of ideas. He thinks that courts and legislatures should do something about the emerging situation, but he doubts that they will. He thinks that we are, with respect to cyberspace, in a situation akin to that of Eastern Europe after the fall of Communism--exhilarated by new possibilities of individual freedom, but unprepared to manage, and even to perceive, the problems and the dangers that lie ahead. "It is the age of the ostrich," he says. Our excitement is that "of a teenager playing chicken, his car barreling down the highway, hands held far from the steering wheel. There are choices we could make; but we pretend that there is nothing we can do."

Let us begin with Shapiro's more optimistic account. Shapiro thinks that the Internet magnifies the power of both citizens and consumers, fortifying both democracy and free markets. He notes that Slobodan Milosevic's efforts to close down an independent radio station were roundly defeated when the station re-routed its programming to the Internet, where it was available, in digital form, to computer users in Serbia and throughout the world; and so pro-democracy activists in Serbia have referred to their movement as the "Internet Revolution." The Internet makes it far more difficult for tyrants to close off the channels of political dissent, while at the same time allowing the international community to learn about government abuses and to speak to actual and prospective victims.

Shapiro emphasizes that, for many of us, the costs of communicating with centers of power, private and public, have been dramatically decreased. If a car company, a university, a state government--not to mention a manufacturer of computers or word-processing programs--has made a mistake, it will hear about that mistake from hundreds and perhaps thousands of people, sometimes on the same day. Since citizens can communicate so easily with government, democracy is a central beneficiary. And this marks a dramatic change from circumstances at the Constitution's founding, when it was thought to be impossible for the public to have continuing communication with its governors. Amartya Sen has observed that no famine has ever occurred in a democratic country with a free press; the power of publicity has been sufficient to avert famines, which are partly a product of social responses, not scarcity alone. If Sen's finding is taken as a dramatic illustration of a broader point--about how freedom of information can avert human suffering--then the Internet promises to produce incalculable benefits.

For consumers, certainly, the future appears truly limitless. To say that it is possible to use the Internet to buy cars, airline tickets, stocks, horses, and hotel rooms--to notice that people can act as their own travel agent or stockbroker--is to offer much less than the tip of the iceberg. (If you are shopping for a car, you can easily find crash test results not only from the United States government, and not only from private insurance companies, but also from Europe and Australia.) Shapiro emphasizes as well the new "ability of individuals to keep the media on guard," both by talking "back to those cultural gatekeepers who have traditionally assumed that their interaction with the public was a one-way street" and by giving individuals the remarkable power "to become publishers ourselves" by posting documents and thus making them available to millions of readers. Matt Drudge is just the most notorious example. Thus the effect of the Internet is to allow its users to "shake the foundations of the fourth estate and the culture business writ large."

Shapiro also stresses another point: our increasing ability to "personalize" our experience, by selecting information "packages" of our own choosing. (Nicholas Negroponte has called this "The Daily Me.") An increasing number of news personalization services now allow people to choose topics of particular interest, by including some topics--sports and entertainment, say--while excluding others. The Wall Street Journal provides something of this sort, and The Jerusalem Post has experimented with a reader-customized newspaper that can be printed on home computers and looks like an ordinary newspaper. This general trend toward personalization is part of a process by which intermediaries (such as general-interest newspapers and magazines) are increasingly replaced by more direct transactions between purchasers and providers. Consider, for example, the extraordinary recent growth of on-line investment, by which individuals (often uninformed and foolhardy) can buy and sell their own stocks almost instantaneously.

This is a rosy picture, but Shapiro is disturbed by two emerging problems. The first comes from large enterprises: the efforts by the "anxious state" to control speech and creativity, and also by big companies, above all Microsoft, to push users in their preferred directions. A number of countries, including Singapore and China, have restricted access to politically unacceptable websites, and the United States responded to the astonishing level of pornography on the Internet with the patently unconstitutional Communications Decency Act, regulating "indecent" and "patently offensive" speech. Shapiro wants to insulate the "control revolution" from public regulation of this sort.

But he is troubled by private power, too. Shapiro thinks that the "slogan of cyber liberation doesn't seem to hold up. `Where do you want to go today'? asks Microsoft; but in reality the company's guiding principle seems to be, Where do you want to go today within the Microsoft universe?" A part of the problem, Shapiro contends, is that "Microsoft has leveraged its position in the software-licensing process to force manufacturers into using Windows rather than competing operating systems." (This, it should be noted, is not the antitrust problem faced by Microsoft; the Department of Justice complains that Microsoft has forced people to adopt its Internet browser, and used Windows to do that, because most people want to have Windows, and Microsoft's Internet browser is bundled along with Windows.)

Another part of the problem is that Microsoft has tried "to capitalize on its built-in audience and become the ultimate information middleman." Thus users have been steered to particular places by means of a channel bar, a feature that users see whenever they turn on their Windows computer. "The channel bar does away with any pretense about Microsoft being a neutral player . . . . It features thirty commercial logos that the user can click to go straight to the websites of" particular companies. This means that Microsoft is directing users "toward the content it wants them to see." (Shapiro does not note that the channel bar was a bust with consumers, and has now been eliminated--points emphasized in Judge Thomas Jackson's recent findings of fact in the Microsoft case.)

In the same category of private manipulation of tastes, Shapiro thinks, can be found's much-criticized and now apparently abandoned practice of selling prominent positions on its website to book publishers, without informing customers that some prominent positions have been purchased. Shapiro is generally concerned about "push technology," by which companies send a "constant stream of content to a computer screen." And quite apart from the effects of public and private power, Shapiro is troubled by "oversteering," by which he means efforts by individual users to limit their own horizons. This is the dark side of personalization: the creation, by each individual, of a communications package that creates insulation from anything new, different, or jarring.

Hence the emerging problem of "freedom from speech." On a conventional speech-corner, or in a general interest newspaper, we may be confronted by ideas, people, and issues that we might have preferred to avoid. This kind of confrontation is a private and public good; it exposes people to genuine problems and can force them to rethink their view of the world. In cyberspace, by contrast, people may enjoy a perfect power of avoidance (recall "the Daily Me") and this can be a serious problem for a democracy. A possible result of this technologically facilitated narcissism is to reduce "shared communal experience" and to produce a situation in which people "share good times with others online who enjoy the same narrow passions as we do."

If members of the Christian Coalition shun mainstream news and entertainment, Shapiro warns, they will end "with a pretty skewed view of the world," and they will also have a hard time getting along with non-members. Similar problems arise from the recent increase in "push-button politics," which may undermine deliberation and promote a system of insufficiently reflective government-by-referendum. The ultimate risk is excessive populism, in which government acts less in accordance with the perceived public good than with momentary pressures as registered on the Internet. Shapiro is also exercised by the loss of privacy to the monitoring and retentive capacities of the computer, and particularly by situations in which people are asked to give up their privacy in return for interactive services; he thinks that people will likely be willing to sell their privacy too cheaply.

Shapiro offers several suggestions for government action, designed to respond to dangers in the control revolution. He likes Europe's privacy directive, which requires data collectors to get explicit permission from people before gathering sensitive personal data. He favors a "privacy safety net," in the form of a "minimal level of personal information that cannot be taken from us" even if people are willing to barter it away "with the click of a mouse." Shapiro also insists that Internet gatekeepers, like television broadcasters, should have obligations to the public as "trustees of the greater good." As an example, Shapiro proposes requiring all gatekeepers to give their audiences easy access, through an icon visible on all viewscreens, to "PublicNet," a publicly subsidized online public forum where the views of community groups, nonprofits, and individuals could be heard. One of the purposes of PublicNet would be to increase the likelihood that people would be exposed to ideas that they might otherwise not encounter, thus turning the Internet into something more like a street-corner or a general interest magazine. Shapiro also argues that government should support local gateways of online networks having "a firm commitment to providing space for citizen dialogue." Shapiro's points are illuminating, lucid, lively, and sensible, if sometimes a bit superficial, which is not so bad for readers unfamiliar with these important topics.


Shapiro wants government to monitor, and to correct, what companies do when they create "code." The whole subject of "code" is treated far more systematically in Lessig's similarly lucid but more penetrating and more dour presentation. But what is "code"? The term has an unsettling Blade Runner quality. Lessig means "code" to refer to architecture, understood broadly to include physical constraints on what can and cannot be done. The real world has a code or architecture in this sense, some of it natural (gravity, darkness at night, mountains, oceans), some of it humanly created (streets, parks, speed bumps).

Lessig believes that code is the great blindness of cyberspace libertarians, who believe that the Internet is a realm of liberty, and that the basic task of government is to stay out. What they miss, he argues, is how code, which could take any number of forms, can serve either to promote or to constrain freedom on the Internet. Lessig's basic submission is that in cyberspace, code is law. It regulates what can and cannot be done, and often it does so invisibly. Code might protect or compromise privacy; it might destroy the rights of authors or defend them extraordinarily well; it might steer people to websites favored by certain companies; it might block sites that are politically incorrect; it might facilitate or prevent free access by children. Code might do any number of things.

The idea becomes clearer if we attend to Lessig's four-fold classification of the ways of regulating behavior. First, regulators might use the law; legal bans on heroin use and child pornography are examples. Second, the market may do the regulating, by increasing the price of behavior that we seek to discourage, as with cigarette taxes. Third, regulators might try to change social norms and attitudes; consider educational campaigns involving drug use and safe sex. And the last and least recognized approach involves "code." If people are speeding as they approach toll booths, regulators might create speed bumps; if people are trespassing on government property, regulators might build special fences; if people are stealing car radios, regulators might ensure that car radios will work only in their original vehicles.

Lessig thinks that code has been frequently neglected as an instrument of social control. He attributes the neglect to the fact that in the real world, much of code seems natural, not designed by anyone, and even immutable. In cyberspace, however, code is emphatically designed rather than natural--people produce code--and it sets the basic governing rules of interactions. This is because the architecture of the program--which can take any number of forms--will determine what can happen and who can do what to whom.

Against the cyberspace libertarians, Lessig insists that cyberspace is not anything in particular; what it is depends on what is (chosen as) its code. It is possible to design a system in which communication may be anonymous; it is possible to design a system in which e-mail is kept forever, forwarded to the FBI, or immediately destroyed; it is possible to design a system that keeps track of all of the user's purchases and visits, thus enabling sellers, and others, to know what the user likes to buy and to see (hence the recent emergence of "web bugs," recording activity at sites and reporting details to advertisers, without notifying the computer's user); it is possible to design a system that automatically includes certain bookmarks, or that draws the user's attention to certain sites. It is easy to collect data about purchasers' preferences without informing people that this is being done.

We could even imagine a system--it is on the horizon--in which code is designed to give copyright holders perfect dominion over their creations. In such a system, any "use" at all would have to be authorized by the author. Lessig concludes that it is the law-like nature of code that makes cyberspace into something other than a realm of freedom, a governance-free zone. His fundamental claim is that it is important to promote existing values, in particular constitutional values, by "translating" them into the distinctive context of cyberspace, so as to ensure that they are maintained there as well as elsewhere. (Lessig is hardly an unambiguous fan of governmental regulation. He notes, for example, that government might regulate code precisely in order to make the Internet more regulable. In a post-publication vindication of one of Lessig's fears, the FBI recently requested the Internet Engineering Task Force--the body entrusted with` keeping the protocols on the Internet--to change the protocols to make wiretapping easier.)

Lessig's general concern is that the power to create and alter code creates a range of dangers for longstanding social values. More particularly, he thinks that this power creates problems for privacy, free speech, and the free circulation of ideas. Objecting that outcomes are now being reached without our thinking about them, he wants the public to choose how to handle the resulting problems. But he does offer his own views. Thus he objects to the "monitoring of today" on the ground that with modern technology, it is extremely easy to collect a wide range of facts that people might reasonably seek to keep secret. "Data collection is the dominant activity of commercial web sites. Some 92% of them collect personal data from web users which they then aggregate, sort and use."

One possible result of this process of collecting--it is actually a variety of surveillance--is that web sites will attempt to engage in manipulation, based on what they have come to learn about you. Thus Lessig wants government to promote code that will increase individual power to retain personal information, through "far greater deployment of encryption code"; and he favors code that will allow individuals to decide what data can be collected by others, through a system in which the "user sets her preferences once--specifies how she would negotiate privacy, what she's willing to give up." On Lessig's plan, that set of preferences would interact with diverse sites, so as to disclose personal information only on the user's conditions.

With respect to freedom of speech, Lessig favors a system that would protect children from inappropriate sites by means of an identification system, one that would require sites to block access to "any self-identified minor." But Lessig's more fundamental concerns mirror Shapiro's: he is worried that private code-makers will push people to their preferred sites, and he is also worried that individuals will engage in excessive filtering. "We individually may want to avoid issues of poverty or inequality, and so we might prefer to tune those facts out of our universe. But from the standpoint of society, it would be terrible if citizens could simply tune out problems that were not theirs."

Lessig is especially critical of PICS, the World Wide Web Consortium's "Platform for Internet Content Selection," an innovation that has generally been greeted with great enthusiasm. PICS allows filtering of content; it would permit users, or employers, or others, to "personalize" the Internet by filtering out content that is deemed inappropriate. Lessig believes that this system is objectionable especially because if "you can't see what content there is, you can't know what's being blocked." Ordinary people are hardly in a position to use PICS to filter out sites, and they must therefore rely on one of a small number of companies able and willing to do this work for them, even though people have little information about what exactly is being blocked. Lessig notes, correctly, that the Constitution applies only to government--to "the state"--and hence private actors, such as Microsoft and others principally responsible for the creation of code, can do whatever they wish without worrying about the first amendment; but he supplies what he sees as a proper "translation" of the First Amendment, one that would subject the design of code to the First Amendment, on the grounds that the power to create code, though nominally private, is government-like in its effects.

One of Lessig's most interesting discussions involves "intellectual property"--the law governing the property rights not to "things" but to information, as in the law of copyright. Under current law, you cannot copy what other people write without their permission; but you can make "fair use" of their writing, in reproducing small portions. (Hence the quotations from Shapiro and Lessig in this review do not violate the copyright laws.) Most people are afraid that the property rights of owners of intellectual property will be lost in cyberspace. Can't anyone now copy anything, basically for free? Lessig thinks that this is all wrong. He believes that the real danger is the opposite--that owners of intellectual property will have perfect control over what is distributed, blocking any use, however small. "The lesson in the future will be that copyright is protected far too well. The problem will not be copy-right but copy-duty--the duty of owners of protected property to make that property accessible."

The reason for this anxiety is code, which can create innovative methods for monitoring access and for blocking any reproduction at all, even small excerpts. The name of this innovation is "trusted systems," a technology that can prevent access entirely, thus giving authors protection of their works through "a privatized alternative to law." This is a problem insofar as it prevents the free circulation of ideas, which is an interest protected by existing copyright law insofar as it authorizes people to make "fair use" of copyrighted material, whatever the owner may wish. For cyberspace, Lessig urges that the law should allow individuals "fair use" notwithstanding code innovations that would prevent it. He claims that code should not be allowed to prevent an "intellectual commons"; he wants the law to prevent authors from having perfect control over who has access to their writings.

Lessig has a more general theme, which he leaves mostly in the background. Most of his book deals with code that is "closed," in the sense that its content is commercial property and in that way secret, and cannot be altered by users. Microsoft's code is closed in that sense. But "open code," now with a small market share but a genuine competitor to Microsoft, is very different. It keeps "its core code public . . . Anyone can take it, and use it as they wish. Anyone can take it, and come to understand how it works." Both private and public regulation are greatly diminished to the extent that code is open. Lessig expresses a great deal of enthusiasm for open code. While it is not clear what he wants government to do about it, it is clear that he would not be unhappy if government were to help it to succeed in the marketplace.

This is a large and impressive collection of ideas and suggestions; but Lessig doubts that anyone, in the current climate, is making any choices at all. He believes that some of these choices pose constitutional issues, but he thinks that courts are "stuck," partly because "they can't be seen to be creative." He thinks that the result is "to exclude" from constitutional scrutiny "the most important aspect of the cyberspace's law." Lessig also thinks that legislatures are disabled, because "we are weary with governments," and "profoundly skeptical of democratic processes." Thus government's self-described role, with respect to cyberspace, "is to simply get out of the way." This, Lessig contends, is a momentous mistake, not least because it disables the public from making important choices.


Shapiro and Lessig have written emphatically "second-generation" books on the Internet. No longer breathless over the possibilities, these writers are sharply critical of cyberspace's libertarian orthodoxy. The importance of their arguments lies in their shared concern about the lack of public discussion of the emerging problems, and the possible adverse effects not only of public power but also of private power over both consumption and citizenship--and particularly over the possible adverse effects of private manipulation, which is often invisible to consumers.

Yet the most general question, unfortunately, is not addressed in detail by either writer. Will the Internet's astonishingly chaotic, even anarchic, qualities, alongside antitrust law and a generally competitive market, ensure that market forces will suffice as a response to these legitimate anxieties? Consider the fact that many companies have responded to concerns about privacy with reliable and explicit privacy guarantees; and recall that Microsoft's channel bar was an utter failure with consumers. In any case, no company, not even Microsoft, exercises serious control over access to content.

Shapiro and Lessig also overstate, I think, the centrality of the Internet, currently or in the foreseeable future, to human experience. Notwithstanding its large and growing importance, the Internet remains only one of a wide range of sources of ideas and information. Books, television, newspapers, and magazines all still exist, and they are complementary and sometimes competitive goods. The fact that the Internet is part of a much bigger picture makes the authors' fears seem overdrawn.

Both Shapiro and Lessig appeal to some undeniably attractive ideals. There is a genuine risk to privacy on the Internet, which enables third parties--employers, companies, government--to obtain information that users would reasonably like to keep confidential. If people want to read left-wing magazines, or learn all about Star Trek, or study the political views of George W. Bush or Patrick Buchanan, or buy Led Zeppelin's CDs, they should be able to do so without letting the world know about it. To some extent, the Internet market is responding to this concern, with reliable privacy guarantees on many websites. There is a market for privacy, as for everything else; and sites that do not provide privacy are likely to lose business for that very reason. Still, to the extent that the response is only partial, legal reforms--such as the European Union's privacy directive or Lessig's more inventive solution (allowing individual users to set privacy limits)--would make a great deal of sense for the United States. For the same reason, it makes sense to consider rules that would prevent certain sites from making their materials available to children.

The question of intellectual property is more complicated. Lessig is worried about the free circulation of ideas, thinking that authors will use "trusted systems" so as to prevent what would ordinarily be fair use. He demonstrates pretty convincingly that, contrary to the conventional wisdom, the Internet may actually increase authorial control over intellectual property. But it is not plain that this is a serious danger. Are owners of intellectual property likely to use technological opportunities to maintain perfect control over the reproduction of their writing? To say the least, most authors like their work to get a certain degree of exposure; and surely the best prediction is that authors and owners will encourage a measure of spreading. It is quite reasonable to suggest, as Lessig does, that some version of the right to "fair use" should be applied in cyberspace. As a speculation about likely behavior by intellectual property owners, however, Lessig seems to me to exaggerate the threat to the free circulation of ideas.

The largest concerns involve democracy. Shapiro and Lessig write as "deliberative democrats"; they think that a well-functioning political order requires not only accountability but also a high degree of reflection and deliberation. One of their central goals is to restore reasoned and informed discussion to its proper place in democracy. They seek democratic deliberation about the social role of the Internet (Lessig's emphasis), and they want the Internet to promote such deliberation (Shapiro's emphasis). They are quite correct to raise concerns about entirely personalized, individuated communications packages: if each person was able to design his own preferred speech culture, too many people would simply hear echoes of their own voices.

A functioning democratic order, not to mention a flourishing culture, requires that people hear views that are unfamiliar and challenging, and even views that are hard to take. To be sure, the Internet allows people to expand their horizons. It is now easy for Americans to read foreign newspapers if they want an external perspective on domestic events. And it is also important to stress that everyone is selective, to one or another degree, about information; there is nothing necessarily bad about a system in which companies attempt to offer you products, including informational products, similar to those that you have enjoyed before. Still, no one should doubt that the Internet can produce an excessive narrowing, simply because it allows personalized packages at very low cost.

Consider in this regard a fascinating phenomenon, one with important implications for the Internet and for Shapiro and Lessig. I refer to "group polarization," a phenomenon whereby group discussion typically leads individuals toward a more extreme point in the direction of that very group's initial tendency. A group of people tentatively disposed in favor of gun control is likely, after discussion, to think that gun control is a wonderful idea; a group of people who think affirmative action is probably harmful is likely, after discussion, to think affirmative action should be abolished; a group of people moderately disposed to be critical of President Clinton is likely, after discussion, to think that President Clinton is quite terrible.

Since the Internet leads to many discussions among like-minded people, it can fuel extremism (a point emphasized by Patricia Wallace in her fascinating book The Psychology of the Internet), particularly to the extent that people are devising their own communications packages, constituting themselves as on-line groups, and insulating themselves from diverse views. There is also reason to worry that the Internet may aggravate government's current tendency to be extremely responsive to short-term swings in public opinion. Shapiro proposes the establishment of a taxpayer-supported PublicNet, which would provide users with visible access to continuing public discussions from diverse perspectives. It is hardly a panacea--people may ignore, and be annoyed by, the icon--but it is a fine idea, likely to do little harm and possibly some good.

Yet all these problems should not be overstated; and I think that Shapiro and Lessig overstate them. Only crazy people spend all their time on the Internet. What most of us find there is just a part of the information that we receive. In any case, many websites have general-interest features. A newspaper's site is not likely to be much different from a "hard" newspaper, and many people use the web precisely in order to get access to a wide range of general interest magazines and newspapers. Far from narrowing horizons, the Internet often expands them.

Moreover, technology and market pressures have generally worked well to prevent the Internet itself from being a place for monopolistic control. Indeed, it is the sheer diversity and heterogeneity of sites that is overwhelming. For some purposes, Microsoft may perhaps be a monopolist; but even Microsoft cannot (and does not seek to) steer people toward politically preferred viewpoints and away from what it considers politically incorrect. Lessig's discussion of the "open code" movement is intriguing, but the very fact that "open code" exists, and is an emerging competitive force, tends to qualify his alarm about the law-like features of code. It is too early to have a full sense of the advantages and the disadvantages of open code. But its market success suggests that ordinary purchasing decisions will counteract some of the problems with private power over closed code.

Shapiro and Lessig also neglect an important emerging issue: the easy availability of remarkable communications technologies to some but not to others, in a way that threatens to exacerbate existing disparities in social opportunities. For many Americans, including many American children, the Internet has already become like some combination of telephones, televisions, and libraries--second-nature, a familiar part of life, a place where you go if you want to communicate with others or learn about products, Rhodesian Ridgebacks, political candidates, history, finance, anything at all. But tens of millions of Americans remain computer illiterates. And the same inequalities can be found among nations: about one-half of Internet users are American, and in many countries e-mail is unreliable and the Internet is barely used at all. I wish that Shapiro and Lessig had discussed the implications of these disparities, and the question of what governments might do to make the new opportunities more generally available.

The real value of these books lies less in their concrete recommendations than in their skepticism about libertarian platitudes, their insistence on the need for public attention to emerging problems, and their spirit of uncertainty and wariness about possible consequences of the Internet. The celebration of the increasingly remarkable technological possibilities continues, as well it should; but the celebration should be accompanied by close attention, from consumers and citizens alike, to concrete consequences for privacy, free speech, and the open circulation of ideas.

Cass R. Sunstein is a contributing editor at The New Republic.

By Cass R. Sunstein