Mechanical man?

I have a new article out in New Scientist, in case anyone might be interested. It explores the idea that perhaps we humans aren't the uniquely conscious and rational animals we think we are, but act on a much more instinctual and mechanical basis, our actions often being determined by stimuli in our environment. I've touched a little on the work of Alex Pentland, whose experiments I mentioned briefly in an earlier post, and also the experimental work of psychologists John Bargh of Yale University and Ap Dijksterhuis of the University of Amsterdam.

To read a little of the original research, I suggest looking at this nice paper of Pentland, which explores two interesting ideas. First, that traditional behavioral science makes a big mistake by looking first to our rational thinking and verbal communications to explain what we do. He suggests that the baseline assumption should instead be that we, like other animals, often act for reasons that we're not really aware of, and communicate with other through instinctual and non-verbal means. In experiments, he shows that most of what we do (especially in our routines) can be accounted for in this simpler way. Second, he also argues -- and this I think is really interesting -- that our intelligence doesn't actually reside at the individual level, but at the level of the group. In other words, it's not the cleverness in our individual heads that makes people so capable, but the delicate and effective non-verbal communications that bind us into cohesive groups with agile collective behavior.

Winning by repeating -- Mark Buchanan

I'm going to have some more to say on the issue of the polarization of opinions, which I wrote about a little last week. Cass Sunstein emailed me with some insightful comments I'd like to share, but I first have to read and digest a couple papers he sent me.

Meanwhile, check out some new work that brings out in a disturbing way how the apparent consensus of a group can be strongly influenced by the loudest members. Suppose you go around and talk to the students in some class, listen to their opinions, and then later try to give an impression of the distribution of opinions within the group. Clearly you do better in terms of accuracy if you hear from lots of people, and so sample the group in a reasonably fair way.

But psychologist Kimberlee Weaver of Virginia Tech and colleagues found that real people don't form their views this way. Rather, if we encounter one person, repeatedly, who voices the same opinion many times, we tend to weight that opinion more strongly, even though it is obviously just one person's opinion. They found that if one person voiced an opinion, say, three times, it ended up being counted almost as much as if the opinion were voiced separately by three different people. The reason our minds do this, the researchers suspect, is that "the more often an opinion has been encountered in the past, the more accessible it is in memory and the more familiar it seems when it is encountered again." You can read the full paper here.

The conclusion is worrying: "repetition of the same opinion gives rise to the impression that the opinion is widely shared, even if all the repetitions come from the same single communicator."

This psychological propensity clearly feeds into the problem of "pluralistic ignorance," which I touched on in one of my New York Times essays. Unfortunately, our instincts for assessing what "most people think" can easily be led astray, especially by a powerful (and often less than responsible) media.