Infotopia: How Many Minds Produce Knowledge by Cass Sunstein (Oxford University Press, 288 pages, $25.00)
The Wealth of Networks: How Social Production Transforms Markets and Freedom by Yochai Benkler (Yale University Press, 520 pages, $40.00)
Internet utopianism can seem so 1998. The future was silicon in the late Clinton years, when government was flatlining in petty scandal and technology stocks seemed to rise exponentially. Not only was anything possible: If you believed the mavens of Wired magazine and assorted other cyber-prophets, pretty much anything was inevitable. Soon, they assured us, people would spend more time in virtual communities than in “meatspace.” Politics would be transformed by the universal pamphleteering of Netizens. Oh, and some of us would go all the way and upload our consciousness into mainframes to live forever as data. The new world might or might not be brave, but it was certainly weird.
The future has not quite lived up to its billing, which is what Cass Sunstein would have expected. Sunstein, a polymath law professor at the University of Chicago, responded to some of these wide-eyed forecasts in his 2002 book, Republic.com, in which he argued that the Internet polarized politics and fragmented cultural life by creating echo chambers of the loud and likeminded: partisan media, howling blogs, and news selected to reinforce solipsism and narcissism -- “the Daily Me.” In Infotopia he has followed up that skeptical broadside with a survey of the evidence on how information technology affects political debate and institutional decision making. The result is a vivid, readable, and informative work of empiricist skepticism -- a show-me-the-money guide to what soars and what stumbles from the stable of Internet dreams.
Yochai Benkler, who teaches law at Yale, has written a very different book. The Wealth of Networks is Internet utopianism for grown-ups. Benkler's sprawling argument sums up years of work on the economics, sociology, and politics of information technology. He is interested in the world that exists mostly for what it shows about what might be -- a charitable definition of the utopian temper. His book is one part introduction to a vast and rapidly growing field of technology, economics, and law, and one part an object lesson in the place of utopian hopes in mature liberal thought. Benkler makes a strong case that we disown utopianism at our peril. In making that argument, he develops a liberal and realistic version of utopian thinking that avoids some of the hazards of the approach.
The scope of Sunstein's book is narrower than its sweeping title might suggest. His theme is twofold: the fact that valuable information is dispersed among many different people and the problem of how to combine it to make it useful. His starting point is the Jury Theorem, a mathematical proof that originated with the 18th-century philosopher Condorcet. The theorem envisions a group, such as a jury, trying to answer a question. Condorcet showed that if each member of the group is more likely to have the right answer than a wrong one -- even only very slightly more likely -- than in a group vote, the majority is quite likely to reach the right answer. As the size of the group grows, the likelihood of a right answer approaches a mathematical certainty. The same results hold even if some members are likely to be wrong, so long as the rest are a little more likely to be right. In short, groups are much smarter than their members, even their very smart members. A vote by any group of people not basically ignorant or confused is a mathematically reliable machine for generating right answers.
So, why don't we vote on everything? For one thing, lots of people turn out to be ignorant or confused about lots of things. And, dismayingly, the Jury Theorem works in reverse. If individuals are overall more likely to be wrong than right, as the size of the group grows a wrong answer from a majority vote approaches mathematical certainty. The challenge, then, is to find a way of bringing together dispersed information that sorts good information from bad, instead of just amplifying it as voting does.
There are two basic strategies for gathering and sifting dispersed information, and both have inspired some Internet utopianism. The first is markets: Markets “set” prices and production levels by taking account of the preferences of hundreds of millions of otherwise unconnected individuals, expressed purely in decisions to pay or not to pay a particular price. This is better than, say, voting on production levels, not just because voting would be complex, but also because market signals are particularly reliable expressions of preferences precisely because money is at stake. It's cheap to say that you'd like to see more Lamborghinis produced, as many 19-year-old male voters likely would if asked. It's another thing to pony up the purchase price.
“Prediction markets” attempt to apply these virtues of markets to gathering other kinds of information. Participants in prediction markets bet on questions as various as whether (and when) there will be another terrorist attack in the United States, who will be the next president, and (in an internal market organized among Microsoft employees) when a new product will be ready for launch. The rules of the markets vary, but the basic idea is that participants win big if they are right and (in most cases) lose something if they are wrong. Moreover, they can generally choose the size of their bet, depending how likely they think they are to be right. Markets can therefore elicit and amplify good information and silence (by punishing) bad information. Results for presidential elections have been closer to the actual result than most polls, and Microsoft has accurately reset launch dates by months as internal prediction markets revealed that announced targets were unrealistic. But, of course, markets are subject to speculative bubbles and frenzies: Imagine setting anti-terrorism policy on the public-policy equivalent of the 1998 NASDAQ.
The other basic model, much beloved of some reformers and political theorists, is deliberation: drawing out everyone's information through dialogue as a group presses toward a decision. Here Sunstein is a skeptic. In some experiments, deliberating groups do worse at solving problems than groups that vote blind. In others, they do worse than the best-informed members would do by themselves. Clearly the best information does not reliably get recognized. Groupthink seems to be the culprit. Sometimes people defer to charismatic or outspoken group members or to the perceived drift of the group, even to the point of withholding valuable information rather than mark themselves as dissenters. Sometimes the group emphasizes information everyone has -- redundant information -- which looms large simply because everyone knows it, and so eclipses important information that just one or two members have. All this is a reminder of why the secret ballot was such an important reform in 19th-century America: Social pressure can override individual judgment for reasons unrelated to the merits of the question. Sunstein spends time discussing blogs and other new forms of Internet-based political debate, and finds them analogous to formal deliberating groups: The groups sometimes elicit important information and reach good judgments, and sometimes produce bullies and mobs.
There is no panacea, so the trick is to build a better megabyte trap. For prediction markets, enlist people likely to have information about problems that involve both a lot of uncertainty and a lot of relevant, dispersed information. Avoid wildly speculative topics like major terrorist attacks. For deliberation, set up rules that encourage people to share all the information they have and try to neutralize social advantages from outside -- say, by assigning participants advocacy roles in the debate. Deliberate over technical problems, where the right answer will become clear when someone generates it, rather than over more speculative issues, where groupthink can override the merits. (Of course, this is why we have civil engineers for technical problems and encyclopedias for disputes of fact, which leaves a little unclear where deliberation should make its contribution.)
Yochai Benkler argues that information technology can change not just some of our decision making, but everything we do. His most innovative argument is about economic life. He claims that the capital requirements of producing industrial-age goods pressed production into centralized, hierarchical systems, whether governments or firms. The cheapness and power of information technology, however, mean that in areas such as film, music, publishing, and information processing, those vast capital concentrations are no longer necessary. Anyone with some gadgets and the right software can do something surprisingly close to what recently required a huge infrastructure in Hollywood, Manhattan, or Nashville.
In remixes, mash-ups, and online volunteer projects such as mapping the terrain of Mars from fragmented satellite images, people are becoming makers as well as consumers. Sometimes they produce idiosyncratic but vivid -- or tedious -- stuff. At other times they produce real value, as with open-source software, the codes maintained by a loose network of volunteers and part-timers who power much of the world's information processing.
“Everyone a Creator” seems to be Benkler's first slogan for an economy based on cheap and powerful information technology. The second is “Share Nicely.” Just as we can now make valuable -- or at least entertaining -- stuff in our spare time, we can share it almost costlessly with whomever might be interested or link it up with someone else's similarly quirky project. The new universe of free downloads is one instance of sharing, from Web-based music distribution centers to YouTube. A phenomenon that unites voluntary production with open access is Wikipedia, a free, open-source online encyclopedia, which at the time of writing has more than 1.3 million English-language articles written “collaboratively by its readers,” most with an interest or expertise they burn to share. Wikipedia is sometimes lampooned for here-and-there factual errors (the satirical Onion recently reported the Wikipedians' celebration of the 750th anniversary of American independence, which made the United States “212 years older than the Eiffel Tower, 347 years older than the earliest-known woolly-mammoth fossil, and a full 493 years older than the microwave oven”), but it is nonetheless both a useful resource and a terrific achievement. Such sharing is normal among family and friends, particularly when we have something we can hand over without losing much ourselves. What is new is both the quantity of stuff that costs nothing to share and access to billions of people who might be interested in it.
Benkler believes voluntary production, often tied to sharing, is emerging as a major mode of economic life, alongside markets and hierarchically organized production. He has enough examples to prove it in entertainment and some elite technological sectors. So, why should anyone else care? First, Benkler argues, because it is intrinsically good to be able to use more of your capabilities -- to create and share as well as make a living, or even to make a living partly by creating and sharing. Second, improved technologies for sharing valuable information could spur scientific research as they already have software production and culture. Schemes are afoot to “open source” large segments of basic scientific knowledge to facilitate research on neglected diseases and other problems that markets neglect because they mainly affect the world's poor. If they succeed, these models could reduce the cost of important biomedical, pharmaceutical, and other research without compromising the basic integrity of the patent system that fosters innovation. Third, the kinds of people who habitually revise, comment on, and add to their cultural setting may be more critical and engaged citizens, if competence, initiative, and collaboration in cultural life affect the way people engage in politics.
Benkler is no technological determinist. He argues that new forms of production can grow beyond their elite and eccentric ghettos, but that they need our help. Law and policy can facilitate new forms of production and sharing or they can get in the way, as the expansion and strengthening of copyright law has done in the last 15 years. And that, the argument goes, is precisely why we need thinkers who contemplate the outermost possibilities of new technologies, even if their forecasts sometimes go awry. Those forecasts aren't mere predictions, but proposals about what might be if people choose to pursue it. What the prophets of 1998 missed was that sometimes technology ensures only the opportunity for positive change. The rest is up to the political imagination, where Benkler's visionary enthusiasm and Sunstein's skeptical caution belong together. tap
Jedediah Purdy, the author of For Common Things and Being America, teaches law at Duke University and is currently an ethics fellow at Harvard's Kennedy School of Government.