The Paradox of Polarization
NOVEMBER 20, 2019
It is now commonplace to bemoan the growing political polarization in the United States, and to lay the blame at the feet of social media. There are good reasons for this attribution: social media plays host to many contentious and unproductive disagreements, and the design features of these services seem to undermine constructive debate. But the popular notion that social media produces ideological echo chambers, which in turn foster greater polarization, may not get to the root of the problem.
Social scientists have long known that deliberating groups tend toward polarization, and that like-minded individuals tend to corroborate and reinforce one another's views. Undoubtedly, this process has continued on social media. But breaking down the walls of the echo chamber may make matters worse: exposure to bots from the other side seems to increase polarization, and the act of selecting one's own news sources predicts less extreme views than reading from randomly selected sources. This may be partially explained by the tendency, highlighted by Jonathan Haidt and Tobias Rose-Stockwell in a recent Atlantic article, for moral grandstanding. The dynamics of social media encourage ideological expressions that simultaneously rally the in-group and, in their extremity, alienate the out-group. As the writer Scott Alexander has noted, this blinds us to areas of potential agreement because the most polarizing voices dominate the discussion.
Another consequence of the constant connection afforded by social media is that we learn, in detail, the political views of our fellow citizens. While this might sound like a recipe for the fertile cross-pollination of ideas (and it sometimes can be), the inevitable disagreement coupled with the one-dimensional representation we get of one another comes with a cost. It exacerbates the effects of epistemic spillover, the tendency to form holistic (often negative) opinions of others' overall judgment on the basis of their political views. Moreover, recent work has shown that we tend to overestimate how negatively the out-group views the collective behavior of our in-group. In a context of continuous exposure to the out-group, this translates into a pervasive sense of hostility and judgment.
Many of use have learned the value of occasional or even extended disconnection as a remedy for some of these problems, but it seems unlikely that we'll return to a world of blissful ignorance about the views of our fellow citizens. The challenge, then, is to create online contexts in which constructive discussion and debate can occur. Rose-Stockwell, Tristan Harris, and others have proposed a number of potential design improvements to existing platforms, including hiding like counts (which Instagram may soon experiment with) and adding friction to encourage more thoughtful engagement. Other platforms are attempting to build in an ethos of thoughtfulness and civility from the outset. Letter.wiki, for example, matches pairs of thinkers in long-form public letter-writing on substantive topics. Taking a cue from this trenchant white-paper on modern political fragmentation, the platform strives to be a home for memetic mediation, the peaceful search for common ground across ideological lines.
As the coiners of the memetic mediator term write, bridging ideological divides depends on "the mental agility, empathy, and wisdom needed to shift between a multitude of perspectives." In other words, it depends not only on systemic factors (like the right platform) but psychological ones. OpenMind, an educational platform, is attempting to spread this psychological toolkit far and wide. But as Haidt (an OpenMind founder) cautions, "as tribal primates, human beings are unsuited for life in large, diverse secular democracies, unless you get certain settings finely adjusted to make possible the development of stable political life." These settings are political, psychological, and, increasingly, technological. True deescalation will require progress on all three fronts.