I agree with Cass Sunstein that “general interest intermediaries” are essential for a well-functioning democracy. But I am not convinced that the dangers of Internet personalization and group polarization are quite as real as he suggests. At the fringes, certainly, the Internet will encourage extremism and intolerance, as the militia and white supremacist movements demonstrate. But these groups existed before the Internet, and were perfectly capable of using “old” media (including face-to-face contacts and social networks, which have always been among their most effective techniques) to recruit new members and reinforce extremist views. In order to determine whether the Internet is corroding the public sphere and producing a more fragmented speech market, we must look beyond the actions of fringe groups and examine the media consumption practices of more typical citizens.

How do most people use the new media? Consider the most popular Internet sites: Yahoo.com, AOL.com, and MSN.com.1 All three reached more than 45 percent of the Internet audience in March 2001. Each site offers a search engine, links to various types of news stories, chat room/message board services, and a hefty amount of banner advertising. In many ways, these sites functionprecisely like the general interest intermediaries that Sunstein thinks are so important. That is, they provide unanticipated encounters as well as common experiences. Indeed, Sunstein’s description of newspapers and magazines could apply equally well to Yahoo, which is filled with articles, information, and services that most readers would not have selected in advance. While they may not read all of the articles, many users will likely read at least some of them, and browse some of the message boards. Moreover, because the reach of these general interest portal sites is so deep, they are likely to be capturing the attention of many of the same people who set up personalized profiles on Sonicnet.com and Broadcast.com. Here, too, there is little difference from the older media, where, for example, the same people who subscribed to specialized magazines also tended to read Newsweek or the New York Times. The dangers of personalization, then, appear to be somewhat overstated.

What about group polarization? Again, consider the typical user, who spends about ten hours with the Internet each month, with an average session of about thirty minutes. This amount of usage leaves plenty of time for watching television, reading, going to work, talking with friends and family, shopping, and wandering in public places. Any one of these activities holds the possibility of unanticipated encounters, alternative viewpoints, disagreements, and arguments. In a heterogeneous society, the judicious use of new communication technologies is simply not enough to escape different points of view, except perhaps for those marginal individuals who have already taken active steps to barricade themselves from society.

Even though I am less concerned than Sunstein about the dangers of unlimited personalization, fragmentation, and polarization, I still share his sense of caution about the Internet. I am mostly worried about what is going to happen to the “deliberative enclaves” that Sunstein describes. These public forums are a crucial part of any democracy, because they nurture arguments and viewpoints that tend to get ignored in larger pubic spheres. In my own research, I have examined the history of the African-American press, an example of a minority public sphere organized through mass media. By developing alternative interpretations of public events and nurturing arguments that would eventually be used in larger public forums, the black press has played an important role in the history of American democracy. But the black press has not fared well in the face of new technologies. The advent of television was much worse for black newspapers like the Chicago Defender than it was for “mainstream” papers like the New York Times. My fear is that Internet technology will hurt deliberative enclaves such as the black press in much the same way.

New communication technologies are worse for deliberative enclaves than for general interest intermediaries in two ways.

First, deliberative enclaves have a much more tenuous financial existence, because their alternative viewpoints make them less attractive to advertisers. With fewer revenue sources, these kinds of public forums find it harder to respond to the presence of new media. African-American newspapers have, for example, beenmuch slower to set up web sites; most of them, in fact, have yet to do so.

Second, new technologies often help to create closer links between the largest public spheres, at the expense of alternative media. Journalists for the major newspapers and television networks already share wire services, sources, and, increasingly, employers. The development of the Internet does not change this trend. After all, where do the stories on Yahoo come from? The problem is that as mainstream media become integrated with popular Internet portal sites the deliberative enclaves become further insulated from the larger speech community. When you can read Associated Press stories right off your AOL home page, how much energy are you going to spend searching out alternative media?

Because of the dangers that new media present to deliberative enclaves, I agree completely with Sunstein’s suggestion about links, hyperlinks, and public sidewalks. There is no reason why the hyperlinks on Yahoo have to be connected only to mainstream media sources. They could also provide links to alternative sites, like the recently established BlackPressUSA.com, a joint Web presence of African-American newspapers. If large portal sites and other general interest intermediaries decided to include these enclave sites in their repertoire of hyperlinks, the result would be a much more vibrant and inclusive public dialogue. For this to happen, of course, Internet content providers would need to begin thinking more about democracy. Sunstein’s article would be a good place for them to start.


1 Internet usage data are from studies published in April 2001 by two marketing firms: PC Data Online and Nielsen-Netratings, Inc. Both reports are published online, at the following URL addresses: http://nielsen-netratings.com, http://www.pcdataonline.com.