One of the questions we are pursuing in the Information, Communication, Cognition course is how humans make sense of the world. We understand this question to concern the individual level: How do our senses work, and how does our brain compute their inputs? But as a social science major, I am also interested in the societal level: How do human communities make sense of the world? In particular, we can phrase this as the question how we bring together inputs from different members of any given group to come to a common understanding and, eventually, a decision.

A long-standing and wide-spread view is that deliberation is a particularly well-suited means of decision-making, an idea prominently elaborated by sociologist Jürgen Habermas, who has defined processes in which the "forceless force of the better argument" should prevail. The expected gains from deliberation are described by philosopher John Rawls, who writes that "the benefits from discussion lie in the fact that even representative legislators are limited in knowledge and the ability to reason. No one of them knows everything the others know, or can make all the same inferences that they can draw. Discussion is a way of combining information and enlarging the range of arguments."

But do these promises hold? An intriguing book by Cass R. Sunstein of the University of Chicago Law School suggests otherwise. Sunstein finds that deliberating groups often err in their conclusions - even if, and that is the astonishing point, knowledge about the correct solution is held by some of their members. He gives the example of NASA, where in the 2003 investigation of the explosion of the space shuttle Columbia it was found that knowledge that could have prevented the catastrophe was held by some members of the agency, but never uttered as people were "forced into a party line".

In Sunstein's terms, the problem is one of information cocoons: We - and even the highest leaders - find ourselves surrounded by opinions that confirm the the ones we already hold, while exposure to competing views is diminished. But how can such a situation come about, especially in an organization such as the NASA, which abound in expert knowledge? As it turns out, in deliberative processes, some people are more likely to speak than others, and some views are more likely to expressed.

Speaker's status is one powerful factor: The higher a group member's social status, the more likely he or she is to speak out. This is particularly true for statements opposing the majority view, which to voice might harm the status of the individual within the group. And the other way round, people are more likely to support a view held by a high-status member of their group. This bias goes so far that in some cases, high status even outweighs sound reasoning.

From an evolutionary perspective, this behaviour makes sense: Individuals who already have a high status must have gotten some decisions right in the past - otherwise they would not have gained their status. In the absence of counter-evidence, following them might well increase the chance of success (and thus be adaptive). But in today's highly complex environments, the decision-making tactics evolved in our ancestors can lead us badly astray.

How should we then go about coming to a common conclusion in a group? Sunstein - who points out some more troubles with deliberation, e.g. that it tends to polarize opinions - proposes prediction markets: If you let people place anonymous bets on particular outcomes on a question, many hazards of social influences can be avoided. Indeed, such markets have been proven to be uncannily accurate on many questions where knowledge is vastly dispersed (e.g. when predicting election outcomes or future Oscar winners). Still, it seems unlikely that these markets will soon replace deliberation and individual expert opinion, given the high democratic value we ascribe to the latter. But maybe you think otherwise - wanna bet?

Cass R. Sunstein: Infotopia. How Many Minds Produce Knowledge. Oxford University Press, 2006.