Monday, April 11, 2011

The Logic of Violence in Civil War: Libya edition

Tonight a friend asked me "if everyone hates Qaddafi so badly, why isn't there more rebel activity in the west? Why isn't there more sabotage and insurgency and all that?" My first thought was something like, hell, you've got me. What the hell do I know about Libya? And that's true enough: I don't know much of anything about the specifics of the civil war, of ethnic cleavages or tribal rivalry, or really even about the geographic patterns of violence there beyond what's been widely reported in the media. But I do know a very little bit about the general subject of insurgency and irregular war, and the literature has something to say about this.

For one thing, popular interpretations of civil war and rebellion tend to overstate the power of ideology as a motivating factor for combatants. In the present example, this manifests as the belief that widespread loathing of an authoritarian leader should logically translate into broad-based, geographically indifferent rebellion. A melange of motivations is evident in anecdotal reporting from the front (none better than what runs under the byline of prolific multi-platform star C.J. Chivers), though we'd be well-served to keep in mind that people aren't always fighting for the reasons they say they are. These accounts don't really answer our question, in any event: if the rebels of Brega and Benghazi are fighting for democracy, the Qu'ran, or anything but Qaddafi, then surely some in the west of the country would take up arms in the service of one of those ends, right? Here's Stathis Kalyvas (p. 46) on the subject:
An extensive body of research shows that combatants are usually motivated to fight not by ideology or hate or fear but by peer pressure and processes involving regard for their comrades, respect for their leaders, concern for their own reputation with both, and an urge to contribute to the success of the group -- in short, what is known as "primary group cohesion."*
Chivers' account of one rebel straggler, too, puts paid to ideological explanations; Sgt. Abdullah Insaiti, a 13-year veteran of Qaddafi's army, defected to the opposition with his entire unit in February.
His unit, he said, had been scattered under fire in the fighting in recent weeks. He said he believed that eight of his friends had died, but offered that the number was probably much higher than that. Some, he said, had been blown apart in the shelling they had been subjected to out in the desert, where the forces loyal to Col. Muammar el-Qaddafi have been pounding the rebels with all manner of fire. 
Asked if he knew where his unit was now, Mr. Insaiti gave a perplexed shrug.
I'd be curious to see how many of this man's unit return to combat versus blending back into the considerably less dangerous routines of civilian life, but the tale is an instructive one either way: surely an entire platoon doesn't defect over ideology.

But regime security forces are a bit of a special case, to be fair -- this still doesn't really speak to the question of the west's relative quiet. But Kalyvas does, at least in a general sense. To boil a 400-page book down into a few sentences, let's put it like this: The Logic of Violence in Civil War argues that combatants are most likely to employ selective violence in areas where they are approaching, but have not secured, hegemony. To put it even more simply: the warring parties are most likely to kill people not in those places that either they or the opposing faction firmly control, but rather in those places where killing people might have a meaningful impact on defections and the supply of information. Here it is in the original Social-Science-ese (p. 13):
The theory bridges the meso- and microlevels and predicts the likelihood of violence as a function of control. On the one hand, political actors do not need to use violence where they already enjoy high levels of control and cannot use selective violence where they have no control whatsoever; having no access to information, they may use indiscriminate violence, but it will be counterproductive. Instead, they want to use selective violence in contested areas, where they have incomplete control. On the other hand, individuals want to denounce only where it is safe for them to do so; this is the case where their victims have no access to the rival political actor and, therefore, lack the option of counterdenunciation. In turn, this option is related to control: the higher the level of control for one actor, the lower the presence of the rival one, and, hence, of the option of counterdenunciation. The prediction is that violence is most likely to occur where one actor is near hegemonic, not where this actor is in full control or is being contested. Violence, in other words, is most likely where the organizational demand for information meets its individual supply. 
Now, if you're still awake: what does this mean for Libya? Well, Kalyvas doesn't directly address the question of sabotage and guerilla activity, but rather selective violence against neutral or undecided populations. We can, however, draw logical conclusions from his work, and the most meaningful of these -- widely known to students of insurgency but perhaps counterintuitive -- is that collaboration follows control. Conceptually inseparable from the belief that ideology drives rebellion is the commonly-held view that regime strength corresponds geographically (and demographically) with regime popularity; in point of fact, civilians are most likely to peaceably accede to the wishes of the government in areas where the regime's military power and physical control are at their peak.

All of which is just a long way of expressing a pretty simple hypothesis: it's entirely possible that few Libyans are rising up outside of the east and a few small pockets of rebel control simply because government control of the rest of the country just seems incontestable or insurmountable. (Kalyvas, p. 93: "The absence of alternatives often produces collaboration irrespective of the level of popular satisfaction or lack thereof," which come to think of it could've pretty much been the whole post all by itself.) As I mentioned to my friend, there may be other valid explanations related to ethnic or tribal politics, geographic history, and so on, and these would certainly be more effectively articulated by an area expert. But I'm not that guy, so this is what I've got.

*Why this should be so is a complicated question that's largely beyond the scope of what we're talking about here, but Kalyvas alludes to a number of biases that impact scholarly work on the subject. Notably: "There is a clear epistemic bias, at least in the sociological and historical traditions, in favor of the assumption that all (or most) participants in conflicts are motivated by ideological concerns. Because 'urban' scholars tend to be primarily motivated by ideology themselves, the often assign unambiguous ideological motives to participants, even if this is not the case" (p. 44).


  1. I'm all for reducing four-hundred page boooks (in non-standard small types) to a few sentences, so a few, very brief thoughts:

    1) While Kalyvas may account for why there has been no uprising outside of the East, it's important, perhaps to recognize that in "Logic," his (ostensible) dependent variable is not violence or uprising, but the selectivity or randomness of violence.

    2) Remember, Kalyvas has also stressed that civil wars are messy and motivations ascribed to actors might not be accurate, or complete. A lot of people are doing a lot of things for a lot of reasons when established order breaks down; to reduce them into a simple "for/against" civil war might be simplistic, reductionist and/or a fallacy of composition.


  2. ADTS -- Agree with both of your points. I tried to be clear about the fact that Kalyvas' work can't be overlaid directly on this problem, and I thought I addressed your first point with this:

    Well, Kalyvas doesn't directly address the question of sabotage and guerilla activity, but rather selective violence against neutral or undecided populations.

    (I now notice that I excised a sentence that made this point in more detail. Oh well.)

    As for the second point: again, I completely agree. For example, Kalyvas may be useful to the question of "assuming many people in Tripoli reject the legitimacy of the regime, why is there not more anti-government activity there?" But we certainly can't even begin to address more complicated questions about the military strategy (such as it is) of the opposition through reference to his work.

    The act of applying whole studies of political violence to one narrow, circumstantial question (particularly one the study was not expressly designed to address) is fraught, inadvisable, and probably a giant waste of time. I had hoped to be sufficiently honest and humble about the limitations of this approach.

  3. Gulliver:

    I'll look over and (possibly - if appropriate) respond to your points a little later. But for right now, I wonder if the following could be apropos:


  4. ADTS -- Interesting links. The paper (perhaps unsurprisingly) strikes me as more directly relevant than the blog post, if only because it's focused more broadly on the factors that influence the outbreak (or not) of urban insurgency. The blog post is more narrowly concerned with the distinction between unconstrained and constrained responses by counterinsurgent/regime security forces, which doesn't really help us to understand why such an insurgency hasn't ignited in Tripoli, for example. (I think we can stipulate that Qaddafi's security forces are on the "unconstrained" end of the spectrum when it comes to choosing between "urban annihilation" and "coercive governance.")

    There's certainly food for thought here, but I don't think Staniland's gives a satisfying response to the original question. Assuming Libyan military forces are mostly concentrated against the oppposition in Misrata, Ajdabiya, Brega, Benghazi, and so on, we can assume that regime security forces are NOT engaging in an "urban annihilation" approach (in Tripoli and elsehwere) to prevent anti-regime sentiment escalating from social mobilization into insurgency. Staniland argues that the shift from peace to war has more to do with state interests and actions than simple state capability. So why isn't it happening? Is it because there's no anti-regime social movement that's sufficiently organized or evolved to make that jump? Are we back to assuming that the non-existence of a meaningful rebellion in the west is evidence that people there just don't mind Qaddafi all that much? This strains credulity, at least for me.

  5. Furthermore, at least on this narrow question, Staniland's work seems to nibble around the edge of Kalyvas, to augment LoViCW rather than contradict it. He says that interest and approach is more important than raw capability, which, ok, fine. But at the end of the day, aren't we still talking about the exertion of control? He's not saying "control isn't important," only that regimes will not always exercise control to the fullest extent possible (for a variety of reasons). And, well, duh: In those areas where the state DOESN'T exercise its capacity for control to the fullest extent possible, insurgencies are more likely to be inflamed. And if I'm Kalyvas, I'm looking at that and thinking QEF, mofos.

  6. Gulliver:

    I'd say this is a case of, to be Rumsfeldian, known unknowns.

    "Assuming Libyan military forces are mostly concentrated against the oppposition in Misrata, Ajdabiya, Brega, Benghazi, and so on, we can assume that regime security forces are NOT engaging in an "urban annihilation" approach (in Tripoli and elsehwere) to prevent anti-regime sentiment escalating from social mobilization into insurgency."

    That is, as you know, an assumption. I don't know the disposition of Libyan loyalist forces. Moreover, the amount of force - number of people; type of people; amount of equipment; type of equipment - necessary to secure Tripoli is unclear to me. It may be different qualitatively and quantitatively than that required to fight de facto (rather than potential) rebels in "in Misrata, Ajdabiya, Brega, Benghazi, and so on."

    As for the blog post, it's interesting and possibly worthwhile to note that some of the Southeast Asia references were to peaceful protests (student activists, etc.) that were never "cracked down on." I don't know the initial composition of the rebels in Benghazi, etc. And to the extent that the Southeast Asia cases refer to protests in which the regime failed to crack down Hama- or Tienanmen-style, they're often referring to the inability of the military or security forces to do as they're told, so to speak. Again, I don't know the composition, disposition or configuration of the security forces at Khaddafi's disposal either now, or at the beginning of the revolt, and what levers he had, used or lacked and failed to use over them (the security forces, that is).

    There does seem to be a tautological element you're hinting at. How do we know there's no state control - because there's insurgency. Why is there insurgency - because there's no state control. I think this leads into a bigger point, which is that I'm less down on Staniland's piece(s) than you are. First, state capacity has, IIRC, been roughly measured thus far, e.g., per capita GDP. Second, as Staniland rightly quotes numerous times, urban insurgency happens, even though - as the opening note from Kalyvas indicates - many claim that it cannot and/or does not. Third, Staniland brings in the idea of mobilization structures and the like - I don't know what such structures are, were or might have been in Benghazi, or Tripoli, but I can see where it would be valuable to find out.

    I think the primary issue is that we're trying to explain the outbreak and geographic spread of revolt, and Staniland is trying to explain prolonged urban insurgency. Once more, the conceptual muddle rears its ugly head, perhaps, not so much temporally (revolt is fast while insurgency is prolonged?) as geographically (revolt and insurgency can be anywhere?).