Editor's Note: Roger D. Petersen is an Associate Professor of Political Science at the Massachusetts Institute of Technology. He has taught at MIT since 2001 and was recently named the Arthur and Ruth Sloan Professor of Political Science. Dr. Petersen studies comparative politics with a special focus on conflict and violence, mainly in Eastern Europe, but also in Colombia. He has written three books: Resistance and Rebellion: Lessons from Eastern Europe (Cambridge University Press, 2001); Understanding Ethnic Violence: Fear, Hatred, Resentment in Twentieth Century Eastern Europe (Cambridge University Press, 2002); and Western Intervention in the Balkans: The Strategic Use of Emotion in Conflict (Cambridge University Press, 2011). He is co-author, together with Jon Lindsay of Varieties of Insurgency and Counterinsurgency in Iraq, 2003-2009, US Naval War College, 2012.
Octavian Manea:Why do you talk about variety of insurgencies? Should we see Baghdad, Anbar, and Basra as different insurgencies?
Roger D. Petersen: Different countries and different regions possess certain qualities that form the potential “building blocks” for sustained insurgency. Whether that potential is realized or not depends upon the resources of government and the counterinsurgent and the way those resources are used. In Iraq, those building blocks included remnants of the former Baathist regime such as its bureaucracy and various security forces, tribal groups, clans, ethnic identities and religious and linguistic cleavages. When the combination of “building blocks,” resources, and strategies differ among regions, then those regions are essentially different types of insurgencies. I think the combination and interaction of these elements were very different in 4 different regions of Iraq: Kurdistan and the north, Basra and the south, Baghdad, and Anbar.
OM: In your research you pointed out to a spectrum of conceivable individual roles in an insurgency. What is the methodology behind this typology?
RDP: This methodology comes from my 2001 book (Resistance and Rebellion: Lessons from Eastern Europe) which focused on Lithuanian resistance to Soviets in the 1940’s. Insurgency is a complex phenomenon, especially in how violent organization and networks are created and sustained, and the methodology of that book involved breaking down this complexity into component parts and then building back up into a coherent whole. At the base of this process is the way individuals position themselves relative to the dramatic and violent events of insurgency. Most people may wish to remain neutral and just take care of their families but events push significant numbers of individuals into roles of unarmed support of insurgents, or local armed position of a militia, membership in a mobile non-local organization, or equivalent positions in support of the government. Furthermore, individuals may move back and forth along this spectrum of roles. If one is skeptical of broad and vague theories at a high level of aggregation, as I am, then you need to get down and observe dynamics at a basic level. Observing movement along this spectrum of roles is one way to do that.
OM: What are causal mechanisms? Why are the causal mechanisms important for a social scientist trying to understand an insurgent setting?
RDP: There are different understandings of what defines a causal mechanism among social scientists. My own definition is that a mechanism is a specific causal pattern that explains individual action over a wide range of settings. A mechanism must be specific and causal, on the one hand, but general and able to apply to a wide range of cases. For example, the “tyranny of sunk costs” is a mechanism. There is a specific causal logic—previous heavy investment produces continuation of an action that is no longer optimal. And the mechanism is general in that it can apply over a wide range of settings. Tyranny of sunk costs can apply to car ownership—it might be best to get rid of a problematic car but I may be less likely to do so if I just put some money into fixing the transmission, and also to a bad marriage—maybe my marriage is hopeless but I just paid a lot of money to a marriage counselor so I keep going on.
Given the spectrum concept above, with individuals moving back and forth along seven positions, the use of a causal mechanism approach is natural and crucial. The method seeks to understand which causal mechanisms push and pull individuals from one position to another. For example, which causal mechanisms—which small grained causal forces—pull individuals out of neutrality and into unarmed support for insurgents? Which mechanisms pull them the other way into unarmed support for the government? Which mechanisms pull them into armed participation?
OM: What role do causal mechanisms play in shaping the decisions of the local population?
RDP: I think a relatively small number of causal mechanisms consistently have effects at specific points on the spectrum. Emotions such as resentment push individuals into unarmed support. Community-based norms of reciprocity often pull individuals into local militias. Rational calculations operate between other nodes. All insurgencies are different, but most witness a similar basic set of mechanisms at play. Most or all insurgencies see a combination of rationality, social norms, emotions, and psychological mechanisms such as the “tyranny of sunk costs” at play. An individual causal mechanism approach, applied to insurgency, forces the analyst to think through just how this repertoire of mechanisms is producing general outcomes. The approach forces the analyst to start at a fundamental level—specifying the individual level forces at work—and then build back up to the general outcomes.
OM: In his 2009 IISS speech General McChrystal inferred that his main target audience is formed from rational people driven by extremely practical things:
“villagers are supremely rational and practical people: they make the decision on who they will support, based upon who can protect them and provide for them what they need. If a villager lives in a remote area where the government or security forces cannot protect them from coercion or harm from insurgents, he will not support the government – it would be illogical. Similarly, if the government cannot provide him with rule of law, the basic ability to adjudicate requirements legally, or just enough services to allow him to pursue a likelihood, it is difficult for him to make a rational decision to support the government.”
Is it FM 3-24 and the whole contemporary Western COIN discourse too narrow, too much focused on rational, cost/benefit models of decision-making? Is it too restrictive when making this inventory of driving motivations or causal mechanisms?
RDP: I think there is a tendency in the Western academic analysis to focus on rational theories. Those theories are straightforward. But they also might be too straightforward, too simple. In Iraq, the coalition did not plan on the emotion of resentment stemming from a status reversal affecting Sunni calculations in the beginning stages of the conflict. We did not understand the revenge norms that exist in some of the places. We did not fully understand the social norms that helped to produce the tribal militias in Anbar province. We did not understand the psychological mechanisms underlying the Sunni view of the new world they were living in.
OM: What other causal mechanisms (different than the rational cost/benefit calculations) could be responsible for moving the people from one role to the other in an insurgent ecosystem/setting?
RDP: To elaborate on previous comments, in my previous work on violence in Eastern Europe I found that status reversal and the accompanying emotion of resentment was one of the most ubiquitous mechanisms. When individuals had a clear idea of group membership and where that group was located in a hierarchy, then situation when top groups find themselves below bottom groups almost always moved individuals toward support of violence.
In Iraq, the Sunni had positions of control in bureaucracy, military, the government was Sunni, so they were naturally on top going back to the British days. With democracy this is not going to work anymore, and you could predict a reaction against that.
In my previous work, I found that individuals joined local armed groups through social norms—they often got involved les through political ideology and more because a family member pulled them in. If you have a society like Iraq where you have family, clan, tribal structures with accompanying powerful social norms, then you can expect rapid mobilization.
I think that identifiable psychological mechanisms play powerful roles in a different ways in different insurgencies, but these two non-rational mechanisms—the emotion of resentment in an ethnically hierarchical society and the social norm of reciprocity in a society with strong communities—are absolutely essential in understanding the evolution of insurgency in states like Iraq.
I would also emphasize that rational calculation is still perhaps the driving force in many insurgencies, but reliance on rationality alone is usually a mistake.
OM: Would you point out that for a counterinsurgent what matters most is to understand these motivational and causal mechanisms and try to influence them?
RDP: I think counterinsurgents need to understand that there are not only good guys and bad guys in the population, but also gradations. Counterinsurgents need to think about how their strategies move people up and down over these gradations. I hope my research might provide a manageable and intuitive framework for practitioners to productively consider the consequences of their strategies.
OM: How do you influence these causal mechanisms?
RDP: This is a big question, and of course a critical one. Jon Lindsay and I, in a recent study commissioned by the Center on Irregular Warfare and Armed Groups at the Naval War College, try to break down how four different strategies possess inherently differently logics on how to affect and trigger different causal mechanisms at different locations on the spectrum of roles. No doubt the coalition, implementing this range of strategies influenced how these causal mechanisms operated. Our present work tries to understand just how those strategies did that.
OM: In your research you emphasized that the Iraqi surge correlates with the decline in violence but also with the ethnic homogenization of Baghdad. What are the implications for the whole Afghan surge debate? Because at the time, the Afghan debate was very much influenced by some assumptions about why and what worked during the Iraqi surge.
RDP: There was a lot going on before and during the surge. There were deals made with locals in Anbar, ethnic separation in Baghdad, and a myriad of things going on in the south of Iraq. The surge brought both changes in the number of troops but also in their tactics. In Anbar, SIGACTS were declining before more boots were on the ground. There will continue to be different efforts to unpack how these phenomena interacted.
In Afghanistan you have more localized social structures, constant flipping, you don’t have a central government that can provide public goods to the local people, you don’t have social capital like in Iraq where you could create these game-changing local and regional deals as easily as in Iraq. Afghanistan is much more decentralized and the communication in social organizations is different. I don’t think that what we saw in Anbar province is likely to happen in Afghanistan. The Sunnis in Anbar were able to see what is happening in Bagdad with the Shia militias in the Sunni neighborhood and they knew that at a certain time that they needed to make a bargain. There is a different dynamic in Afghanistan, although I would again look at the group hierarchy (especially with the Pashtuns) and local social norms as critical forces breeding powerful mechanisms.
OM: What are the implications of emphasizing the role of +/-2 actors?
RDP: We are very good at find, fix, finish, exploit, analyze. The coalition was doing 300 raids a month or more at one point in Iraq. But if the basis of the insurgency is in the local social structures then those killed and captured are likely to be replaced. You can get rid of the -3, the mobile insurgents (the Al Qaeda guys), but if the local people are behaving normal during the day and different during the night, it is difficult to root out an insurgency. In my previous work on late 1940’s Lithuania, the Soviets had an enormous advantage in power but they couldn’t destroy the locally rooted networks. In some localities they had to deport whole villages in Siberia.
OM: Historically, the COIN discourse was, and still is, about winning “hearts and minds” at the grass-root level. Is this the right frame, the right angle to conceptualize the problem and what a counterinsurgent needs to understand is what causal mechanisms/forces move one or the other?
RDP: I’d like to think that my approach will provide a more sophisticated way to think about “Hearts and Minds.” There are perhaps eight or ten important mechanisms that are usually lumped together in the “Hearts and Minds” concept. In my own work, I try to specify just how, why, and when those mechanisms come to actually affect the roles played by individuals during insurgency.