In response to my post from yesterday about brokered conventions and dark horses, Jonathan Bernstein makes the case that I’m underestimating the possibility of a factional bolt if the (highly unlikely) deadlock came to pass:

I’m going to stick up for my argument a bit. The thing is: Matt’s objections are also objections to getting to a deadlocked convention in the first place. So if they really reach Tampa with no resolution, then that means that the party didn’t unify around one candidate during the primaries and caucuses, and they didn’t work something out during the two months between Utah and Tampa, and they didn’t work out anything in the first days of the convention. If all that happens, then either the rules and norms of the game are somehow getting in the way of cooperation (contrary to what I and some others believe, which is that the rules and norms of the game facilitate cooperation), or that there really is some sort of serious schism either predating the nomination battle or caused by it.

So basically, if we grant the implausible premise of a deadlocked convention…I’m going to say that all bets are off, and lots of chaotic outcomes are very possible. Including the possibility that some of the actors involved may not behave very rationally at all. For two reasons: one is the emotions of the moment, but another is that a lot of the delegates are probably not very sophisticated or experienced political actors at all.

You should take anything Bernstein says very seriously (I certainly do), and the point he makes here is a good one. I agree that the possibility of chaos increases dramatically if a delegate deadlock from the primary season is not settled prior to the 1st ballot vote at the convention. That’s definitely a signal (beyond simply the voters indecision) that the party itself is seriously split on the nomination decision. And I think it’s also a really good point that emotion can lead to irrational behavior among delegates, and that this can precipitate a bolt. Hell, that’s basically the story of Charleston, 1860: the southerners worked themselves into a frenzy in the convention hall after they lost the platform fight over Dred and the territories, and then walked out basically on impulse. Once they had bolted, however, most of them sat around Charleston either sulking or giving fiery speeches, but mostly wondering when they would be invited back to join the convention. Emotional and irrational indeed.

So I would say Jon’s argument convinces me that a bolt is at least plausible under a deadlock. But I still think there are serious structural forces working against it, even if we did get to the 1st ballot undecided. First, I think there’s a big difference between a party being closely divided and one being deeply divided. The Democrats in 1860 were deeply divided, but not really closely divided: the schism over slavery was (obviously) about to come to physical sectional conflict, but within the Democratic convention (where at the time delegates were assigned by state population proportion, regardless of party strength within the state), it wasn’t really closely: the northern view basically predominated. In the GOP right now, we kind of have the reverse: a deadlock would show a very closely divided party — perhaps split almost exactly 50/50 between Santorum and Romney — but it doesn’t seem to me to be a deep division. It’s more like Hillary and Obama. Sure, each faction has its preferred candidate, but it’s not like either faction would consider moving their support to the Democrats given the nomination of their non-preferred candidate. In 1860, that was exactly the situation.

Second, I don’t think a split aggregate judgement of the voters necessarily implies a schism among delegates, or among the competing candidates. The voters aren’t purposely choosing schism; they are only acting collectively by aggregation, over a series of months. The delegates, on the other hand, would have to take a dramatic step while everyone was watching in order to create a bolt. It would not simply be a ratification of the voter indecision; it would itself be a wholly new decision. Now, part of this is mitigated by Jonathan’s point: if the party can’t get this together between the end of the voting and the start of the convention, that’s evidence that something is wrong. But it’s still not the positive step of fracturing the party. That seems to me to be another purposeful action, at another level of seriousness, in and of itself. And I don’t think it would occur until after many ballots had been taken. In fact, I think the possibility of a dark horse — itself a radical longshot and a completely untenable idea in practice — is probably more likely than a bolt. But maybe I’m just agreeing with Bernstein now: this is so highly unlikely at two levels (first the deadlock, second the inability to solve the deadlock prior to, or subsequent to,  the first ballot), that once those two unlikely scenarios have come to pass, a bolt becomes an actual possibility.