Thread by @KevinZollman

Today in my Network Epistemology class we discussed the well known phenomena "Pluralistic Ignorance." The idea of pluralistic ignorance is that people may privately believe one thing, but publicly profess another. (Privately they may hate the regime, but publicly praise it.)

The Anderson story "The Emperors New Clothes" is often used as a motivating example. Each person privately believes the Emperor is naked but is afraid to say it because they worry they may be revealing themselves as uncultured.

en.wikipedia.org/wiki/The_Emper…

We started by reading this paper, which seeks to clarify the concept and asks the question: is pluralistic ignorance always the result of individual irrationality? Or could rational groups come to be pluralistically ignorant?

link.springer.com/article/10.100…

This is an important question because pluralistic ignorance requires that one treat oneself as unusual. "I know that I say 'P' but really believe 'not-P.' Everyone else says 'P' so they must really believe P."

It also connects with one main theme of the course, does group irrationality always require that the members be irrational? Or could rational individual constitute irrational groups?

The class was concerned that the definition of pluralistic ignorance left out something that many authors thought are critical: social sanction. It's not enough that there be ignorance, but that it must be maintained in a certain why by actions of others.

The next paper modeled the emergence of pluralistic ignorance in a social network. They show that pluralistic ignorance can emerge when "true believers" are well connected to one another, and capable of acting as crystal from which the ignorance grows.

journals.uchicago.edu/doi/abs/10.108…

The paper is interesting in that it represents a very different modeling framework from economics models -- it doesn't start with a utility function or anything like it. (Although it's not necessarily inconsistent with utility maximization.)

The paper's conclusions are broadly plausible, that pluralistic ignorance must start with a set of true believers who use social sanctions to "quiet" rebellion around them. The fear of sanction causes that group to grow to eventually take over the whole population.

The model does contain one odd assumption, that those who are believers and who don't see dissent, opt not to punish rare instances of dissent. This struck us an odd assumption, but it doesn't really drive the results.

The model doesn't consider another plausible way that pluralistic ignorance could emerge: that the underlying beliefs change over time. So while most people may have believed that homosexuality is a sin 100 years ago, private beliefs changed.

But people may have been afraid to give voice to that change out of pluralistic ignorance. People were unaware that others hand changed, too.

So, while Centola et al show one way for pluralistic ignorance to emerge, it's not the only one.

Finally, I wanted to draw a connection between pluralistic ignorance and another network phenomena: the friendship paradox.

The friendship paradox is that it is possible for most people to have fewer friends than their friends do. This has the ring of "everyone is above average" to it, but in fact it's completely coherent.

As an extreme example, assume that everyone is friends with one person: Bob. Everyone (except Bob) has no other friends. Bob will have N-1 friends. The N-1 other people will have fewer friends than their one friend.

Of course, that's a silly illustration but it scales up.

This paper illustrates how this can relate to pluralistic ignorance. It can be the case that for any property, most people say that "most of my friends have that property" despite the property being relatively rare.

journals.plos.org/plosone/articl…

If that property is a belief, say, that "the regime is good." and if having that belief correlates with being socially well connected, than most people will think that most people believe the regime is good. Even when that belief is relatively rare.

This process is distinct from other processes, where punishment is required to get pluralistic ignorance off the ground. Here, all it requires is converting the relatively well connected. Or, alternatively, increasing the social connections of those who are true believers.

Pluralistic ignorance is a really important topic that applies to a large swath of social changes (and failed social changes). We only scratched the surface in class, but already we saw how interesting the problem is.