• Choose a category
• All
• Classic-Fiction

By Root 763 0

in their rightness. Gamblers will marshal evidence to explain away their losses as unlucky flukes, but ignore similar evidence about their wins. Consumers who have just made a major purchase will seek out evidence supporting their purchase while eschewing evidence against it. In short, we like to fill our minds with information that confirms what we already believe; this information in turn doubles down our already existing support of what we think or dismissal of what we disbelieve.

It is in this regard that the Internet and confirmation bias are conspiring to erode what remains of reasonable political discourse in this country. The mechanics of this dissolution are simple: first, the sheer quantity of media has reached the point where even the most assiduous news fan can consume an entire day’s reading by simply ingesting only those tidbits that support his or her own views; and second, the network of political blogs, through a feedback loop among bloggers and readers, has evolved into a machine that supplies the reader exactly this prefiltered information. A vivid illustration of this latter point was made in 2005 by two researchers, Lada Adamic of HP Labs and Natalie Glance of Intelliseek, who had analyzed all the posts of the top twenty blogs of both political camps in the four months leading up to the 2004 election. In particular, they tracked how many times each blog linked to others on the list. What they found was that some 85 percent of the blogs’ links were to other blogs on their own side, with almost no blog showing any particular respect for any blog on the other side. When they graphically represented the total citations, with lines representing five or more cross-citations between each node, it looked like this—

FIG. 5-4—ADAMIC AND GLANCE’S POLITICAL BLOG CLOUDS

—with the nodes 1 through 20 representing the left-wing blogs, 21 through 40 the right-wing blogs. When only connections of more than twenty-five posts are shown, communion between left and right disappears entirely. This represents no conspiracy among the bloggers, of course; they simply self-select what they read and post, and their self-selection tends, as is natural, toward that which supports their preexisting conclusions. Ideas of a feather flock together, and it has always been thus.

But the problem, once again, is speed. The Internet allows the like-minded to find one another so quickly, and with so little exposure to other points of view. Indeed, in this regard, the forward march of search technology threatens to balkanize our politics even further: the ability to ever more agilely find what we are looking for, while excluding the rest, is exactly what a citizen does not need in making his or her political choices. The phenomenon is parallel to Chris Anderson’s “Long Tail” idea as it pertains to commerce: the combination of nearly infinite inventory, on the one hand, and a national or even global customer base, on the other, means that Internet-based businesses can both satisfy consumers’ most obscure tastes and make a profit doing so. For example, in the case of video stores, the limitations of geography and the cost of rent force Blockbuster stores to limit their selection only to broadly popular titles, whereas a delivery service like Netflix can stock at least one copy of literally every DVD ever released because their marginal cost of stocking each title is so low. What this means for culture, Anderson hopes, is a future where everything stays in print, always available, always profitable. “Abundant, cheap distribution,” he writes, “means abundant, cheap, and unlimited variety.”

In the realm of political discourse, and indeed of narrative in general, I fear we have fallen into a far less salutary situation. One might call it the Long Tail of Truth: given any trend that one wants to identify in the world—about the popularity of a buzzword or a band, the mendacity of a politician or a pundit, the rise or fall of any fashion—on the Internet one can readily convince oneself that the trend exists, as long as one runs the right targeted