December 12, 2008
The spread of false information and rumors poses growing risks to society and the economy.
That was the message delivered by Harvard Law School Professor Cass Sunstein '78 in a major lecture—titled “He Said THAT?? She Did WHAT?? On False Rumors and Free Speech”—marking his appointment as the Felix Frankfurter Professor of Law at HLS.
Introducing Sunstein before the talk, Dean Elena Kagan ’86 described him as “the world’s pre-eminent legal scholar,” one who “challenges our assumptions and changes the way we think about legal issues.” Suggesting that Frankfurter was a forbearer of what she called “Sunsteinian Minimalism,” Kagan noted that the Justice, who was sometimes accused of being too leftist, brought to the Supreme Court a strong belief in judicial restraint. “He believed more in the institutions of democracy than in the courts,” said Kagan. “He also insisted … on respect for Federalism, for the decisions of state governments.”
Opening his lecture, Sunstein declared that one of his goals was “to drive a wedge between the ‘Marketplace of Ideas’ and ‘Truth.’” Identifying truth specifically with factual accuracy, he outlined three mechanisms by which false rumors gain traction in that marketplace and become widely held beliefs.
The first of these mechanisms he termed “Biased Assimilation.” Belief formation is a motivated process, said Sunstein, and when an individual is presented with a statement of unknown accuracy, he or she will respond to it with preconceived notions of what is correct. If the statement is in agreement with the individual’s bias, the individual finds validation and holds the belief even more strongly. If the statement disagrees with what he or she already believes to be true, studies have shown that generally the person will not change opinions, but rather will still find validation in his or her own belief, rejecting the statement as questionable for one rationalization or another.
“What supports your prior convictions will be new and fresh…and have credibility,” said Sunstein. “What goes the other way will seem preposterous and easily dismissed.” Once a false rumor has thus been accepted, a correction then becomes difficult to accept. The first reaction, he said, is to ask “Why is my belief being denied?” and to answer the question by stating, “It wouldn’t be denied unless there was some foundation for it.”
“The very denial of the false rumor focuses people’s attention on it,” said Sunstein, “and the focus of their attention tends to leave a kind of cognitive echo in the mind, thinking ‘It’s probably so.’”
Building upon the first mechanism, Sunstein went on to describe the second, what he called “Informational Cascades”—the process by which awareness of the aggregate opinions and beliefs of others can largely control an individual’s taste and belief formation. Through this process, he explained, “early adopters” can have an undue influence over the public’s general acceptance or rejection of a statement of fact. He described a recent experiment in which nine different websites offered free downloads of songs by unknown bands. On eight of the nine websites, however, users were able to see the download preferences of the “previous users” in the worlds created those who were running the test. Overwhelmingly, usage patterns of the users aligned with the preferences of the previous users, and indeed, the users’ musical tastes were overwhelmingly shaped by the opinions with which they happened to be presented.
In cases of political belief-formation, Sunstein said, the emotional resonance of the first mechanism can help to jumpstart these cascades, with the viewpoints of others then overriding that which one might otherwise choose to believe on his or her own. Thus, many citizens opposed to the presidency of Barack Obama were prone to falsely believing that he was a Muslim, while many of his supporters were open to the mistaken idea that Sarah Palin would think that the continent of Africa was a country.
Social Influence Information Exchange
Sunstein’s described the third mechanism by which false rumors are propagated as “group polarization.” He cited studies in which individuals, deliberating on a given political subject with like-minded groups, tended to shift toward greater extremism. Again, in these experiments the social interaction overrode the individual’s process of objective consideration. He proposed three standard explanations for this behavior:
- Corroboration: Most people start out tentative about facts, simply because they lack information. After affirmation from others, however, they become more confident.
- Exchange of information: In any group with an antecedent position, most evidence will tend to support that position.
- Concern for reputation: People want to be held in esteem, and thus don’t want to be seen as more skeptical than the rest of the group. So, they adopt the strength of conviction of their peers.
Focusing on false rumor propagation, Sunstein voiced two concerns unaddressed by these explanations. First, people tend to be unaware of the bias of the groups in which they are participants. Second, individuals discount the importance of ideologically minded people to willfully mislead. As he explained, “It’s underestimated the extent to which, with respect to certain rumors, there’s a self-interested or ideologically-motivated mover who is starting the information [process].”
Connecting these behavioral observations to issues of freedom of speech, Sunstein discussed certain Supreme Court decisions. Using the example of a case centered on a newspaper’s publication of the name of a rape victim, he noted the Court’s reliance on the argument that, if a fact is already in the public domain, then wide publication of that fact should always be protected. But this sort of publication can cause irreparable damage, he said, which might prompt a more nuanced application of law.
Raising a more recent phenomenon—YouTube—Sunstein warned of the dangers of turning every citizen into “their own Truman Show,” in which the minutiae of everyday life is broadcast to the world. “A life is not an incident or an event, but a series of them,” he explained, a fact which is lost when incidents are broadcast over the Internet or other media, without context. “Sometimes the isolated segment or event will have a kind of defining character, in a way that will be extremely destructive, not only to the individual involved, but also to people trying to make rational judgments about the relevant person.”
Sunstein quoted Felix Frankfurter as saying, “Freedom of the press is not an end in itself, but a means to the end of achieving a free society.” After offering some examples in which uninhibited press freedom leads to the destruction of other freedoms, he proposed a reconsideration of the idea of the ‘chilling effect’”:
“Many First Amendment questions in this domain are resolved by reference to the ‘chilling effect’ concern. Indeed, it has become quite clear that references to the ‘chilling effect’ have had a very serious ‘chilling effect’ on engagement with the constitutional question …The question shouldn’t be whether there’s a chilling effect and how to avoid it, but how to achieve the optimal chilling effect.”
“Zero chilling effect, in light of the mechanisms just described, would be profoundly destructive to a host of relevant variables.”