As Cass Sunstein, a professor of law at the University of Chicago, saw himself being skewered on various Web sites discussing his recent book, "Republic.com," he had the odd satisfaction of watching some of the book's themes unfold before his eyes. On the conservative Web site "FreeRepublic.com," the discussion began by referring relatively mildly to Mr. Sunstein's book about the political consequences of the Internet as "thinly veiled liberal." But as the discussion picked up steam, the rhetoric of the respondents, who insisted that they had not and would not read the book itself, became more heated. Eventually, they were referring to Mr. Sunstein as "a nazi" and a "pointy headed socialist windbag."
The discussion illustrated the phenomenon that Mr. Sunstein and various social scientists have called "group polarization" in which like-minded people in an isolated group reinforce one another's views, which then harden into more extreme positions. Even one of his critics on the site acknowledged the shift. "Amazingly enough," he wrote, "it looks like Sunstein has polarized this group into unanimous agreement about him." An expletive followed.
To Mr. Sunstein, such polarization is just one of the negative political effects of the Internet, which allows people to filter out unwanted information, tailor their own news and congregate at specialized Web sites that closely reflect their own views. A "shared culture," which results partly from exposure to a wide range of opinion, is important for a functioning democracy, he argues. But as the role of newspapers and television news diminishes, he wrote, "and the customization of our communications universe increases, society is in danger of fragmenting, shared communities in danger of dissolving."
This pessimistic assessment is a sign of just how sharply scholarly thinking about the Web has shifted. In its first years, the Internet was seen euphorically as one of history's greatest engines of democracy, a kind of national town hall meeting in which everyone got to speak. As an early guru of cyberspace, Dave Clark of M.I.T., put it in 1992: "We reject kings, presidents and voting. We believe in: rough consensus and running code."
Now, with the examples of business and government control offered by the explosion of Web commerce, the merger of America Online and Time-Warner, the Microsoft antitrust case and the litigation over Napster, that is no longer the case.
Andrew Shapiro, a guest lecturer at Yale Law School and the author of "The Control Revolution," said that the early euphoria over cyberspace had been replaced "by a kind of 'technorealism,' a second generation of Internet books" that are much more critical.
An example is the 1999 book "Code" by Lawrence Lessig, a law professor at Stanford University, who argues that the enormous amount of personal information people reveal when they shop online, browse Web sites or call up information offers extraordinary opportunities for both governments and businesses to control their lives. "Left to itself," he wrote, "cyberspace will become a perfect tool of control."
Mr. Sunstein's assessment is somewhat different from Mr. Lessig's, though still negative. "His is closer to Orwell's '1984'; mine is more like 'Brave New World,' " Mr. Sunstein explained. If to Mr. Lessig he danger is government or corporate control, to Mr. Sunstein it is a world of seemingly infinite choice, where citizens are transformed into consumers and a common political life is eroded.
Both agree, however, that society must begin to make more conscious choices about what it wants the Internet to be. Mr. Lessig's main point in "Code" is that the Internet does not have a "nature." The world we think of as "cyberspace," he said, is an environment created by the architecture of the computer code that gave birth to the World Wide Web.
Mr. Lessig's point is that because the Internet is based on "open source" computer protocols that allow anyone to tap into it, it has a specific character that can be, and is, modified all the time. Internet providers can write software to allow users maximum privacy or to track and restrict their movements to an extraordinary degree. The software engineers, as Percy Bysshe Shelley said of poets, are the unacknowledged legislators of our time. We must, Mr. Lessig said, acknowledge this reality and try to shape it.
"We can build, or architect, or code cyberspace to protect values that we believe are fundamental, or we can build, or architect, or code cyberspace to allow those values to disappear," he writes.
Mr. Shapiro describes himself as more optimistic than Mr. Lessig or Mr. Sunstein. "I came to see more potential in the Internet empowering individuals, but we are all 'technorealists' in that we see personalization and social fragmentation as features of the Net."
Other legal scholars agree that fragmentation and polarization have increased with the Internet, but they do not necessarily see it as a problem. "I do not mourn the demise of the domination of the main outlets of news and information," said Peter Huber, a conservative legal scholar who is a fellow at the Manhattan Institute and the author of "Law and Disorder in Cyberspace: Abolish the F.C.C. and Let Common Law Rule the Telecosm." "It's true that the oracles of traditional authority, The New York Times, the network news and the universities have lost power. Just look at the declining market share of the major TV networks. But whether you regard that as good or bad depends on where you sit."
That doesn't mean he dismisses claims that new technology causes social fragmentation; he just feels that the individual empowerment of the Internet is well worth the price. "The Soviet Union had a 'shared culture' and one source of information, 'Pravda,' " he said. "I think it's impossible to judge what is the exact point at which you have the right mix of diversity and common culture."
Mr. Sunstein said he was not talking about limiting diversity but rather the insular way that most sites were structured. For example, he said, most political Web sites have links only to other like-minded sites. Although he stops short of calling for government intervention, he says, "We might want to consider the possibility of ways of requiring or encouraging sites to link to opposing viewpoints."
Until the early 1980's, the Federal Communications Commission required broadcasters to provide equal time to opposing viewpoints, a policy eliminated during the Reagan administration. When critics of Mr. Sunstein's book pointed out that his own site at the University of Chicago offered no such links, he responded by including the Web addresses of two well-known conservative colleagues.
What some political Web sites are already trying to do is figure out ways to encourage more intelligent deliberation rather than simply name-calling and insults.
"We are trying to design sites so that they promote diversity as well as a sense of community," said Scott Reents, the president of two political Web sites called E-ThePeople and Quorum.org that recently merged.
The software design of the sites, Mr. Reents said in support of Mr. Lessig's point, can shape discussion in important ways. For example, at Quorum.org readers are asked to give a thumbs up or thumbs down to a particular posting; that item's placement is determined by reader reaction. (The site tries to prevent people using multiple identities from voting more than once by requiring visitors to register.)
On other sites, a group of regular users rank the value of contributions, and the rankings then determines their place on the "bulletin board." How well that works, however, is an open question. When Mr. Sunstein tried to intervene in a discussion of his own book on a techie Web site called slashdot.org, his contribution was given a very low ranking. "I think maybe they didn't believe I was the author of the book," he said.
James Fishkin, a political scientist at the University of Texas, said that such efforts at Web democracy follow the model of debate in ancient Sparta called the Shout. "The idea of the Shout is that the candidate that got the loudest applause or shout would win," he said. "Unless we make special efforts to implement more ambitious democratic possibilities, the Internet, left to its own devices, is going to give us an impoverished form of democracy in the form of the Shout."
Mr. Fishkin is trying to follow the example of ancient Athens, whose assemblies consisted of several hundred citizens who, after being chosen by lot, would deliberate and vote. He has developed a technique called "deliberative polling" and would like to bring the idea to the Internet. "The idea is this," he said. "What would public opinion be like if people were motivated to behave more like ideal citizens, if they had access to a wealth of information and to competing arguments on a given issue?"
Over the last decade Mr. Fishkin has collected a random group of several hundred people and given them carefully prepared briefing documents on both sides of a given issue. Participants question panels of experts and discuss the issues in smaller groups with trained moderators so that no single person is allowed to dominate discussion. After their deliberation, they are then surveyed privately as in any opinion poll, but their views now reflect, it is hoped, careful deliberation. Texas actually used the method to help determine its energy policy, holding a series of deliberative polls between 1996 and 1998. "Because of it, there are now windmills all over the state of Texas," Mr. Fishkin says.
Mr. Fishkin is hoping to use the Internet to conduct "deliberative polling" on a much larger basis. To Mr. Lessig, deliberative polling is one of the few hopeful developments when it comes the democracy and the Web. "If Jim can transfer to cyberspace what he has done in real space, I think the Internet could be very different," he said.
Yet some view efforts to tame the Internet as doomed to failure. "I think it's a waste of time," said Mr. Huber. "All this talk about `links' and so forth is interesting intellectually, but by the time you try to implement it the technology will be 10 years ahead. When online video becomes as accessible as e-mail, the whole game will change again. And if you think there is fragmentation now, you ain't seen nothing yet."