Recently, in conversation with a friend about the concept of the Overton Window I proposed that the Overton window was not of constant width, and eventually got her to agree with me. (key insight – do you agree it can be shrunk? Then why not widened?)
The next question is whether the two sides of the window can be moved independently. Because recently, on some issues, it has been dragged really far to the right. We are arguing about whether TORTURE is okay because the window has been dragged too damn far over (hint: torture is not okay). And I would like to see left left side dragged open some more and we could add some up and down for the libertarians as well. Can we open one side without closing off the other, simply by widening the window, or must we necessarily end up silencing some people? How could either goal be accomplished? I’ve read people claiming that we need more loonies on the left who scream about stuff, but that just doesn’t seem to be very effective – I live in lefty shoutytown, USA, and most of the shouty members are completely ignored in the political process. Thus, the answer is more subtle than that and it would be good to understand how.
Perhaps we could model idea flow within a population using epidemiology models. But this would make television equivalent to germ warfare, and the internet more like germ warfare between populations living on many separate islands, and I’m not sure that it’s possible for a person to become 10% more infected than they were previously. Anyhow, we don’t want to watch individual thought virii travel through a population, we instead want to watch aggregate bundles of correlated virii travel through a population and see how if affects thing like party identity and the like. Because there are very few single-issue voters, so in order to change someone’s party identity, we have to have significant shift on multiple issues. But I remain unconvinced that virus models are the right way of doing things here, and am open to better ideas.
Fluid flow? Distributed agents playing Nomic? Complicated LRC circuits? We need something to model the fact that peoples’ minds change slowly – they may get infected with a meme, but it won’t necessarily change their mind overnight. The most interesting thing in this vein that I have seen recently about this was a posting to ArXiv where the authors discuss a machine-learning approach to determining how and why countries go to war.
