Lord, I was born a ramblin' man

3 hours panels/meetings, page count = 112

Greetings from Chicago! It's been a long day, but fun. Most of my meetings today ended up being career-related; tomorrow is about my dissertation. I'd forgotten what a madhouse the APSA convention is. We take up the meeting rooms of two hotels, and at least the one I'm staying in is swarming mid-career professionals on the make: lining up jobs, discussing book contracts, trading departmental gossip, visiting old friends. It's a good, crazy energy. Berkeley's department is among the larger ones, and we have a really good placement record, so there have been a number of, "oh yeah, that guy" encounters in the hall. It's nice to come back to this world and see familiar faces.

It's funny, I'm here this year as a "tourist," that is, I'm not presenting anything, just meeting people and going to panels (in that order), but I'm programmed down to the quarter hour. My obsessive travel planning paid off, because I have all my meetings entered into my phone, including locations, with reminders 15 minutes before, so I just show up somewhere, meet someone, start talking, the phone buzzes with the reminder, I check my next location, shuffle off to that, meet someone, start talking: lather, rinse, repeat.

There are literally dozens of panels in each timeslot, so earlier in the summer, I narrowed it down to 2 or 3 each in a few different timeslots, and then built my meetings around the gaps in the schedules. I'm going to one panel each day; today's was about historical causation, which is a huge part of my dissertation. I argue that policy choices made about the security forces 100 years ago had an impact on regime dynamics 50 years later, and even reverberate today; in a way, my project is all about historical causation. All three papers dealt in different ways with the importance of context, with how to assess which causal factors are most important, and with how far back in time to go to explain a given phenomenon. This is a familiar problem: you can ask, "what caused the Soviet Union fall?", and there are some proximate causes, like economic decline, but then there are things that caused that, like changes in the global economy, but then there are things that caused that, and so on. "Infinite regress," it's called. How do you decide where to stop, how far back is far back enough? One way to do so is to talk about, as I do in my dissertation, a "critical juncture": before X period, a lot of different countries were in a similar situation (weak states, politically unstable, chronic civil wars); after X period (1880-1910) when Y things happened (state building, political bargains around electoral reform, security force configuration) they looked very different overall than before X (relatively stronger states, greater political stability, absence of civil wars), but with A and B distinctive patterns related to the Y things that happened during X period, the critical juncture (regimes with politicized security forces are susceptible to insurrection, regimes with militarized security forces are susceptible to coup). That's definitely a step forward, and one that my dissertation committee co-chairs at Berkeley used to great effect in their main scholarly work. The question becomes refining that critical juncture framework to make it even more effective. So on this panel, one paper looked at antecedent conditions: how do we think about the time before the critical juncture, is everything that happened then equally important, or are there "critical antecedents" that are not only more important than other antecedents, but also similar to antecedents of other critical junctures that other scholars identify? Because my argument is driven by what happens in the wake of the critical juncture, I haven't been very explicit in distinguishing which antecedent conditions are most important, even though I have a whole chapter planned about the period before the critical juncture. So that paper will encourage me to think more systematically about which factors going on before the critical juncture are most relevant for my argument. Good!

Another paper looked at one of the current methodological buzzwords, "causal mechanisms," and considered it in relation to historical context. Causal mechanisms are all the rage because they get us closer to showing exactly how one thing causes another in the infintely complex social world. I say that the type of security-force configuration causes regimes to be susceptible to certain types of armed challenges rather than others. How does it do that? Well, by making arms available to some people and not others. When local police forces are under the control of politicians (one type of configuration), they encourage insurrection by providing an incentive to rebels, either because they're fighting back against repression (as when politicians use local police to quash their political enemies) or because they want to take advantage of the police's corruption and induce them to switch sides (as apparently happened quite a bit during La Violencia). In this case, the causal mechanism is the presence of armed men at the local level who are sanctioned and armed by the state, and thus given a degree of legitimacy, but who are basically available to the highest bidder, which is usually, but not always, the politicians who control their hiring and firing (sometimes it's rebels who take advantage of their corruption to induce them to switch sides). What this paper on the panel today argued was the need to set causal mechanisms in a particular context. The same mechanism, in two different contexts, may generate two different outcomes. One mechanism in the study of social movements is "certification," in which the state or another recognized entity (Amnesty International, for example), certifies that a given movement is operating according to agreed-upon standards, and therefore is worthy of support, funding, being listened to, etc. In a democratic context, certification by the state can help an NGO flourish because it's been given a seal of approval and can attract funding and support. In an authoritarian context, certification by the state could cause potential supporters of an NGO to run screaming in the other direction, because they think they'll be spied on or that the organization is a puppet of the government. So I need to ask myself, are there different types of local contexts in which the availability of police forces at a local level would generate different outcomes in terms of insurrection? The answer is probably yes, and that difference is probably about the level of electoral competition. That really helps me process my advisor's comment that security-force configuration and electoral competitiveness are not alternative hypotheses in explaining conflict, but it's in their interaction that the story lies. That interaction may be about the context that electoral competitiveness sets. Cool.

Finally (whew!), the third paper on the panel was about standards for assessing which antecedent causes out of the many that exist are most relevant for explaining the outcome of interest. The authors advocated using a certain type of logic to make those choices, the logic of set theory (think diagrams with partailly overlapping circles and shaded areas where the circles intersect). Using this tool, it's not always the most recent causes that are the most important. So if you have two different causes for the same outcome that you're trying to prioritize, you look at the logical relationship between each of them. I like the idea, but I don't think I really got the examples they used, so I'll have to check out the paper itself when it becomes available. It'll be worth thinking about security-force configuration and electoral competitiveness in this light.

Geez! I'm sure I won't have this much to say tomorrow, but clearly it's been a good trip so far. Weather's supposed to be nice again in Chicago - although I'll be spending most of it indoors! Cathy gets in tomorrow night, though, and my folks are in town from Milwaukee, so we'll have Saturday and Sunday to hang out with them.