Although the country is divided politically, we actually agree on a lot of things: for example, that baseball and apple pie symbolize America, that we drive on the right side of the road, or that June is a nice month for weddings. A group of researchers has explored how we come to these agreements.
A computer model of consensus. I'm Bob Hirshon and this is Science Update.
In America, we shake hands when we meet somebody. In Japan, they bow. How do societies come to these agreements, without a law, declaration, or even a formal vote?
Luis Amaral and his colleagues at Northwestern University designed a computer model to find out. The model suggests that agreement spreads fastest when people seek out information from their neighbors, but don't trust any single neighbor completely.
To get an example of how this works, imagine how people select where to live. One solution to this problem would be to go and look at the scores of all the schools in the entire area. And look at all the police reports of crime. And then you select the best. But I personally don't know anyone who has made decisions about where to live that way.
Instead, he says people tend to ask others for personal recommendations, and go with the majority. Those people, in turn, are influenced by your opinions, until a consensus is reached.
Paradoxically, the model suggests that when people have stronger reasons to doubt their neighbors—for example, when an oppressive government censors speech—they seek out even more opinions, and information spreads even faster. That could explain how the harshest regimes often breed a united and well-organized resistance. I'm Bob Hirshon for AAAS, the Science Society.
Making Sense of the Research
Who decided that brides wear white dresses? (In China, they wear red.) That burping is impolite? (In some cultures, it's a sign of respect for the cook.) It's not like anyone ever passed a law on these issues. Yet somehow, over time, our culture comes to an agreement.
Amaral's model reduces this process to a fairly simple operation. The model is made up of imaginary people called "agents." You can think of the agents as points on a grid or a spider-web: every agent is directly connected to a bunch of other agents, who in turn are connected to other agents, and so on—just as every person is directly connected to friends and family members, who each have their own friends and family members.
Unlike people, who have complicated opinions on all sorts of things, agents can have only two states of mind, which are symbolized by the number 0 or 1. Think of 0 versus 1 as yes versus no, or white dress versus red dress, or so on.
The model starts with every agent being assigned a state of mind at random. Some start at 0, others at 1. These are their starting beliefs. But the agents know they don't have all the information they can get to come to a decision. So in Amaral's model, the agents start checking in with their neighbors: do they agree with me or not? If enough of them disagree, they might change their opinion.
This resembles the way people can come to a consensus about something. Suppose you felt that it wasn't important to have a cell phone. But then gradually, you start to notice that most of your friends have cell phones, and they use them to talk to each other. Eventually, you might update your opinion and get a cell phone yourself. You go from a 0 (cell phone non-owner) to a 1 (cell phone owner). On the other hand, if you move to a new area where nobody has cell phones, you might decide your cell phone is useless and get rid of it.
Using just this set of rules, Amaral's model would result in small pockets of people who all had the same beliefs—for example, a clump of zeroes here, a clump of ones there—but might not result in total consensus. To make his model more realistic, Amaral added a rule that agents could sometimes talk to other agents that were far away. For example, you might live in an area where people don't use cell phones, but you might hear from your friend across the country that cell phones are great. Maybe that friend convinces you to get a cell phone, which in turn might convince your friends at home to get them too.
It sounds complicated, but the basic message of Amaral's model is that if you don't believe you have perfect information, and you don't think any single other person has perfect information, consensus will eventually be reached. That's because everyone's trying to figure out what the majority of people think. And by doing that, they actually help shape majority opinion.
Now try and answer these questions:
- Describe the basic findings of Amaral's model.
- How can following a "majority rule" approach help create consensus?
- Why would consensus spread faster under an oppressive government, or in a situation where people felt they could not trust each other?
- How does introducing an element like television news change the model?
- How can Amaral's model help explain major political changes, such as the Civil Rights movement of the 1950's and 1960's?
The Amaral Research Group website at Northwestern University is devoted to its many lines of research.
Reynolds Engineering and Design is a private computer graphics firm that develops computer models of individual and group behavior, some of which can be seen on its website.