Politics, the hardest of sciences

So, apropos of Lion and Ben's comments over the last few weeks about the nature of economics - how it isn't a science, really, how the entire discipline lacks much predictive value, and how those who know economics really really well still miss the really obvious stuff around them. This, to me, is inarguable, and it's part of the reason that this crisis is keeping me up at night (this, too). Economists got us into this mess; can we trust them to get us out of it?

I got to thinking, though, and it seems to me like this criticism is pretty easily leveled at political science as well. As is noted here, all the predictive powers of political science combined failed to see Obama coming, even four years ago. As Jennifer Hochschild puts it:

I come to this question after catching up over Christmas vacation on a lot of journal reading, almost none of which would lead one to predict that voters in Iowa etc. would support a black man with the middle name of Hussein for president, especially over a Clinton and/or a war hero. So have we political scientists screwed up again in our ability to understand and analyze contemporary politics?

As I was reading that, I initially thought to myself "Well, anyone who looked at this election from a systemic-level would have expected a Democratic victory in November. The economy was bad, and the party who controls the White House after two terms usually doesn't get a third." But Obama is not an insert-name-here-style Democrat. He's no Mondale, Kerry, Gore, or Clinton. And the point still stands: no one predicted that Iowans would be ready to vote for a black man. Except, I guess, for the Obama campaign.

Just so you don't think I'm picking on American political scientists, here's another example. International relations theorists have long prided themselves on their ability to understand, on a very basic level, the behaviour of states in the international system. No matter what their theoretical persuasion, they all consider a state's ability to defend itself through force as one of it's key characteristics, one that ensures the survival of the state. This, as I learned in my sole class on Monday, is what makes Japan so difficult for IR theorists to understand. Yes, Japan has a sordid history that makes Article 9 of its constitution understandable, but IR theorists (for the most part) have long felt that history is an unnecessary dimension for explanatory or predictive models. It makes abstraction difficult, which in turn makes explaining and predicting difficult.

Anyway, Japan has renounced war as a means of doing much of anything, is willingly ensconced in a series of really disadvantageous security guarantees with the United States (by this I mean they pay America a great deal and benefit America a lot, far more than they help themselves), and yet is surrounded by countries that are either a) bat-shit crazy (North Korea) or b) terrified of it (China, South Korea, most of Asia really). No one would have predicted this; it simply does not correspond easily with any existing theoretical paradigms in IR.

So what's my point? Just that political scientists, like economists, are given privileged status as interpreters and oracles, even as they repeatedly demonstrate how inadequate their tool-kits are to understand the things they study. My question is, how do we fix this? As an archaeologist, I've always felt that political scientists and economists are overly obsessed with a very limited spectrum of human behavior: namely, modern nation-states and modern marketplaces. Humans have been interacting with each other for millions of years, in millions of different ways - life didn't start in 1648 or 1776 - and real predictive abilities may develop only once we've worked backwards a bit. However, I'm biased. What does everyone else think?