Saturday, September 15, 2007

Why Self-Selection Is Hurting Us

I posted a slightly different version of this as a comment to this David Brin piece. Brin made two important points in the comments:

  1. The current calcification into political "sides" is much more rigid than it has been for a long time, if ever.

  2. The positions on both sides have mutated, with those on the right being almost unrecognizable to those of the right wing of 20 years ago.

Brin argues that differences in conservative vs. liberal personality acount for this. I'll buy that that may have some impact, but not enough to account for what's going on. There's a common mechanism acting on both sides.

Here's my hypothesis: The portions of the electorate that actually care about politics have unprecedented access to information about public affairs. The efficiency of information flow has increased manyfold in the last 15-20 years.

There is a fundamental change in the system dynamics of how groups organize and identify themselves when information flow is that efficient. Specifically, the efficiency allows the groups to self-select extremely rapidly. Members who might otherwise be sympathetic to a portion of the agenda of a group have any unorthodoxies that they espouse instantly exposed. These members are then either forced to hew to the orthodoxy or leave the group. Since those that leave the largest groups run the risk of being marginalized if they can't form their own group with their own orthodoxy, recantation of unorthodoxy becomes more likely.

Look what this does:

  1. Positions on issues harden rapidly into orthodoxies. Once they've hardened, they're almost impossible to change because self-selection crowds out any diversity of opinion.

  2. Positions on weird issues harden and become as important as positions on more mainstream issues. Again, this is a natural consequence of self-selection. As long as there's room on a group's agenda, any random event can trigger the rapid formation of a position.

This second point is, in my opinion, one of the main reasons why the hard right appears to have gone somewhat more insane than the hard left. When they got a lot of their agenda (e.g. anti-communism, free trade, reduction in union power, welfare reform) institutionalized in the 80's and early 90's, the lack of a new agenda allowed random stuff to float in and take over. The hard left's agenda didn't suffer quite as badly because it was a bit more stable when the system dynamics actually changed. If the left manages to accomplish some of their agenda, I predict that the stuff that fills the void will be equally nutso.

Note also that the "culture war" fits neatly into this hypothesis as well. Once you're hardened into sides, self-selection prevents any sort of moderation or compromise in your positions. If you can't compromise, you have to demonize the other guy--it's all you've got left.

Forget the manchurian scenario and your "neocon monsters." This systemic change is the true threat to the country, because it's the root cause.

It's also fiendishly difficult to change. You obviously can't reduce information efficiency. Everybody's going to know everything at the same time, allowing each group to marshall its forces and respond by developing a consensus opinion that won't budge once it's formed.

I think the answer must be to find ways of delivering more nuanced, better quality information. As you have pointed out before, the internet is notoriously bad at delivering good data, but the crap that is does deliver it delivers really fast. If you improve the quality of the data, presumably each group will produce more nuanced positions. In nuance lies hope for negotiation.

The other major thing you can change is the process through which groups are formed and identify themselves. If you have a way to slow down the self-selection process through some structural change, you'll get more diversity of opinion within a group. This is why I'm such a big fan of the disputation forum idea. Again, moderation of opinion leads to compromise. Compromise is good.

No comments: