Wednesday, December 21, 2011

The COIN wars: a toe in the water

The gentlemen of On Violence are the latest to raise the whip over the carcass of that old glue-pot COIN, taking aim this week at what they've callled "the Chicago School of Counterinsurgency." The post is filled with generalizations, mischaracterizations, and the spurious received wisdom of second-rate popularizers and third-rate social scientists, but those errors are mostly peripheral to the main theme. This strange little essay is in fact a subtle confirmation of the rational choice theory the authors mean to criticize. The purportedly contentious assertion of what is in fact an entirely uncontroversial banality – that war is not a wholly rational, predictable phenomenon – gives the authors a platform to draw tendentious normative conclusions in purposeful contrast to the ostensible COIN orthodoxy. This serves to situate On Violence on the side "of the faction they assess to be the one most likely to win" the battle for influence and credibility in our new post-counterinsurgency era. It's almost 2012 – every blog needs to have something on record saying look, I told you they were doing it wrong!

Semantic carelessness is often the sign of poorly-formed ideas, and the post in question is rife with lexical errors. Beginning with the assertion that "the Chicago School [of economics] believes that humans always act rationally when it comes to money," the authors on several occasions use words in misleading or ambiguous ways. (Warfare in place of war is another frequent mistake.) The excerpted quotation is a simple mischaracterization of rational choice theory, which is based on the idea that people survey the choices available to them and select the one most likely to maximize gains. To assert that individuals "act rationally" is not to say that they always choose wisely, but rather that they base their choice on some expectation of benefit.

The On Violence guys clearly misunderstand this: their pronouncement that "rational investors frequently make irrational decisions, believing they are rational" is utterly nonsensical to an economist, for whom the actor's willful effort to maximize utility is precisely what defines "rationality." Of course, rationality does not imply perfect information or foreknowledge of consequences; rational actors can take decisions that turn out disastrously for them thanks to a failure to properly account for context or an imprecise evaluation of various costs and benefits. Richard Posner – who knows a fair bit about the dogmas and creeds of the Chicago school, and whose biography one might like to investigate before crowning David bloody Brooks as "our greatest living conservative commentator" – wrote an entire book about the calamitous conspiracy of rational acts that led to the recent financial crisis. (SPOILER ALERT! He does not conclude that "rational investors [made] irrational decisions, believing they [were] rational.")

On Violence further confuses the "rationality" issue by inexpertly applying the lessons of cognitive psychology (or at least Brooks's retelling of Daniel Kahnemann and Amos Tversky) to behavioral economics. According to OnV, Kahnemann and Tversky "proved that actual human behavior often deviates from the old, rational models, revealing flaws in the machinery of cognition." But about this, too, they're mistaken. The "old, rational models" say nothing at all about the machinery of cognition; they merely model the way cognition manifests as action. It's certainly true that emotion, unconscious bias, and other so-called cognitive "flaws" may interfere with the exercise of pure reason, but – as we've discussed above – that's not the same thing as irrationality. To put it rather more simply: the rational actor model may not be perfectly descriptive, but it is adequately predictive.

So who are these folks that populate "the Chicago School of Counterinsurgency"? Who accepts "the idea that in warfare—with death and subjugation on the line—mankind's rationality trumps his unconscious thoughts and emotions"? Who fails to consider how social context and personal psychology may influence the decisions of the enemy or the population held at risk? The only person named among the "military theorists [who] continue to ignore humanity's underlying irrationality" – by which the authors seem to mean the influence of emotion and other cognitive and psychological elements – is Andrew Exum, whose unsourced quote concerning "cold-blooded calculations about their self-interest" is reproduced out of context. Who are the rest?

What of the military theorist who wrote this?
[W]ar is not an exercise of the will directed at inanimate matter, as is the case with the mechanical arts, or at matter which is animate but passive and yielding, as is the case with the human mind and emotions in the fine arts. In war, the will is directed at an animate object that reacts. It must be obvious that the intellectual codification used in the arts and sciences is inappropriate to such an activity [as war]. At the same time it is clear that continual striving after laws analagous to those appropriate to the realm of inanimate matter was bound to lead to one mistake after another. (149)
Later in the same work, this theorist argued that
The effects of physical and psychological factors form an organic whole which, unlike a metal alloy, is inseparable by chemical processes. In formulating any rule concerning physical factors, the theorist must bear in mind the part that moral factors may play in it; otherwise he may be misled into making categorical statements that will be too timid and restricted, or else too sweeping and dogmatic. Even the most uninspired theories have involuntarily had to stray into the area of intangibles; for instance, one cannot explain the effects of a victory without taking psychological reactions into account. (184)
You'll surely have figured out by now that we're talking about Clausewitz. (The page numbers given above are from the Paret/Howard 1984 translation of On War.) The much-misunderstood "remarkable trinity" (pdf) introduced at the close of Book 1, Chapter 1 (89) is indeed composed of "(1) primordial violence, hatred, and enmity; (2) the play of chance and probability; and (3) war's element of subordination to rational policy." To repeat: the foremost theorist in the history of western military thought contends that emotion and chance make up two of the three vital considerations in any useful theory of war. Let's get straight to the point: any analyst who fails to consider the importance of moral factors in war – including the interplay of what On Violence misleadingly dichotomizes as "emotional reactions" and "rationality" – is a failure.

There is, of course, some irony to all of this, what with me here advocating Clausewitzian nuance and balance to a blog that has made a habit of gleefully slaying Clausewitzian straw men; claiming the Prussian's relevance to a form of war that the On Violence men have called "our style of warfare" as they scramble to distance themselves from it, ever in tune with the zeitgeist. I know, I know – try to keep it together. While we're on the subject of irony, how 'bout the OnV guys taking Pape at face value and citing his spurious conclusions as fact in an essay questioning the very legitimcy of the rational choice model?

Unsubstantiated assertion is unfortunately something of a theme in this post, and its conclusion is centered on the grand-daddy of them all: "people aren't rational when it comes to killing and death," an unequivocal pronouncement that is both demonstrably false (why then the persistence of instrumental violence?) and patently at odds with the authors' own previous acknowledgement that "war is about people, politics, and ideology."

The problem with On Violence's Gladwell-deep survey of behavioral economics is that bounded rationality cannot as yet meaningfully inform our models of human agency in conflict. We may recognize that rationality and utility maximization fail to perfectly explain all human behavior, but we have no better predictive model on which to base our efforts to influence the choices of others -- the most extreme of which is war. To paraphrase Milton Friedman, perfect realism may be impossible, but predictive accuracy is still the coin of the realm in social science. Until Kahnemann, Tversky, Thaler, Becker, et al can present a coherent, predictive theory of human choice that can be plausibly applied to economic and political behavior and which definitively falsifies the rational choice model – which is, let's remember, a model, not an attempt at descriptive realism – all this bleating about "humanity's underlying irrationality" is worse than useless: it distracts us from efforts to improve the valid models we have today and the policy prescriptions we may derive from them.

When boiled down to its essence, the message here is a simple if deeply controversial one. The authors of this essay mean to assert that the failure of rational choice theory to completely explain the whole of human behavior renders it useless as a guide to action. Because we cannot perfectly comprehend our interlocutor's decision calculus and cannot be sure his choices will be sensible to us, we must instead assume that calculus will be entirely insensate to our behaviors. If "people aren't rational when it comes to killing and death," then how can we possibly hope to exert our will in predictable ways through violence?

To accept this contention leads the thoughtful man down a dangerous path, at least so far as the anti-Clausewitzian OnV brothers are concerned: to the conclusion that predictable, willful influence is an impossibility, that the choices of an adversary or neutral cannot be shaped, and that our own alternatives are reduced to disengaged indifference to the other or his total annihilation. If we can't persuade, influence, or coerce in a predictable way that's consistent with human reason and our perceptions of causality, we are left to compel by destruction of the enemy's means to resist. This conclusion is perhaps closer to the truth than the chimera of calibrated influence offered by the many proponents of the indirect approach, but I'm quite sure it's not the one the On Violence men would like us to reach.

I'm quite conflicted about the whole thing, in spite of my stridency. The utility of force or its threat for any purpose short of compulsion is something about which I have less and less confidence every day, and I loathe the multivariate philosophies of international engagement that are founded on an unexamined faith in security through transformational change to our operating environment – whether the composition of polities or the hearts and minds of the people who inhabit them. But I see this On Violence post as reinforcing the very worst lessons of the recent COIN era: men can be changed, if only we understand what really drives them, and violence is at best a necessary complement to that transformative action. I fear COINdinistas and COINtras may both be missing the point, and if we throw out the hard-won lessons of counterinsurgency's history and present – that violence is essential to effecting needed change, however temporarily in the absence of legitimacy and consent of the governed – then this decade of "institutional adaptation" will have indeed been a waste.

UPDATE Thurs 22 Dec @ 1125: I see that Ex has responded to the original On Violence post here. I actually took out a fair bit of a draft that dealt with Kalyvas' work, so I'm glad to see that he touched on it.


  1. This looks more like a COIN debate than a debate about rational choice theory. I don't want to get dragged into a debate about COIN, because it's a quagmire, but I do want to make a comment on rational choice.

    Rational models and subjective models are not inherently in conflict. Rational choice is an observational lens, and it has a lot of uses-Dan Trombly wrote a post on Iran that pretty much exemplified how rational choice perspective should be used. However, there are limits to the model.

    If you look crisis situations where there isn't much time for actors to make decisions, cognitive biases can play a larger role. Cultural and symbolic perspectives are helpful for understanding the context behind political statements. Examining social status, pride, ideology, and fear can assist in understanding motivations. There are also phenomena such as groupthink, crowd dynamics, and bureaucratic politics (that can explain state or institutional irrationality). A practitioner of war or statecraft might be guided by an intuitive or relational approach where, while perhaps not violating rational model tenets, a subjective perspective may have more to offer in explaining what is going on.

    So when it comes to war, the point you (or CvC) made about emotion and chance is very important. People can forget this with both rational choice and subjective models. With rational models, you run the risk of overemphasizing the materialistic elements of war, like how McNamara over-relied on metrics about material strength. With subjective models, you get things like EBO, thinking that you can achieve "cognitive collapse" and such things easily. I think most discussions of war tend to have a rationalistic or technocratic bias, although the intellectual ferment of the COIN debate seems to have moved conversation a little further from that.

  2. Here is the link to the Exum quote:

    I don't really think it is taken out of context, but I suppose that is in the eye of the beholder.

  3. Keith -- I'm not saying the post misrepresented what he was saying, only that it was quite literally stripped from its context.

    Students of COIN can't win for losing, apparently: if you push a fraudlent love-n-hugs doctrine based on gratitude theory, you're an out-of-touch loon, but if you push back against gratitude theory and try to explain that pacification is about something rather more complex, you're criticized for a failure to understand the nuances of human motivation.

    Let's just be clear: Ex's point was made in an article where he pushed back against the oversimplified, nonsense narrative explanation of how counterinsurgency works. Considering Michael C.'s comment yesterday @ 1249, it seems obvious to me that his post is largely about defending the legitimacy of softer forms of COIN in the face of control-based theories.

    I'm not sure you and Mike Few understand quite what it is you're supporting when you commend him for this.

  4. Michael C's comment yesterday confused me. I read his post as an argument against a key COIN assumption--that we can reliably influence behavior--not as a defense of something else.

    I guess that proves that I read what I wanted to read.

    I don't really understand what you're defending or advocating here. A deeper understanding of what motivates humans? Sure, I'm all for that, but let's not experiment in a situation that gets Americans killed.

    Is this, really, the lesson we drawing from 5 years of counterinsurgency: "that violence is essential to effecting needed change, however temporarily in the absence of legitimacy and consent of the governed." That makes me think that COIN was really, really not worth it.

  5. What I'm advocating is that we say and write things that are true, and that we ought to defend those things with evidence and a firm grasp on reality. That's pretty much it.

    I think the reason his comment confused you is that you're not appreciating its context: this post was part of a campaign to push back against what the OnV guys think of as a trend towards emphasis on the control aspects of counterinsurgency (vice the consent angle). Essentially, they're working to show that the "I don't care if the population likes me, as long as they do what I require" attitude is flawed. (It's not, at least not when it's a third-party counterinsurgent who holds it.)

    Your frustration (and Mike's) with our national policy decision to wage counterinsurgencies is totally understandable, and I share it (particularly in the case of Afghanistan). But that's not the issue we're talking about here. I was disappointed with Mike's recent article in Foreign Policy for similar reasons: the debate about COIN-as-policy is a separate one from the conversation about best-practices-in-COIN. I appreciate that your view (shared with COL Gentile, Mike, and others) is that having the latter conversation encourages bad conclusions in the first debate, but I don't think that's predetermined (and I don't think that gives us the right to punt on the whole thing).

    Should we consider COIN as an operational approach of first resort, something around which to base our foreign policy? No. Can we afford to ignore the study of what drives conflict and how it may best be controlled or suppressed simply because we'd prefer not to be involved in certain types of conflicts? I'd say the answer to that question ought to be no, too.

  6. And I think it ought to go without saying, but I'm not advocating "experiment[ing] in a situation that gets Americans killed." I don't think anyone is advocating that, and I don't think it's helpful to suggest that people are willing to sacrifice the lives of others in order to validate pet theories.

    Military forces are assigned a mission. They have to do what they can to figure out how to be successful, and their institution needs to support them in that effort (by writing doctrine, developing TTPs, organizing/training/equipping as appropriate, etc.). As military analysts, we ought to do what we can to contribute to that effort. As citizens and strategic thinkers, we can (and probably ought to) work on a parallel track to influence our government to make better strategic decisions (and thus to employ our forces in a more effective way, a way that's more attuned to the realities of what force can and cannot achieve).

  7. I think it is a mistake to separate the two questions. And I'm just going to leave it at that because I lack eloquence to make the point stronger with more words. (something about how experience should say something about the possibility of effective COIN, assuming best practices are studied to make COIN effective)

  8. Not sure how I got involved in all this, but as to how people make decisions in combat, here's what I directly observed,

    "In How to Win Friends and Influence People, Dale Carnegie reminds us that ―when dealing with people, remember you are not dealing with creatures of logic, but with creatures of emotion, creatures bristling with prejudice, and motivated by pride and vanity.‖26 Even in a protracted insurgency, people make decisions based off the perceived expected costs and benefits from both the government and the insurgency.27 Sometimes, these decisions are rational when based strictly on thoughts (minds). Other times, the decisions are foolish when one base decisions from an emotional response (heart). Sometimes, this perception is driven from a combination of one‘s heart and mind. It can be influenced or coerced by a host of options ranging from physical violence to financial gain. At the confluence of the heart and mind, small wars appear as a contest of wills or a battle for the soul.28"

  9. cont...

    "We learned that relationships are essential to counterinsurgency operations for intelligence and collection value; however, one must remember that the mission comes first and the intent is to influence, coerce, or persuade the subject in order to force him/her to agree to your mission. It is quite naive to believe that you are attempting to win anyone‘s heart and mind during a violent insurgency."

  10. You see Gulliver, I've dealt with over 22 suicide bombers including the families. When I left the town, AQI started recruiting 12 year old girls to blow themselves up. So, please don't patronize me on rational choice.

  11. But since you brought it up Gulliver, I'll tell you the same thing I told the other Pentagon pricks when I was an MNC-I operational planner and a CJSOTF-AP operational planner in 2005, "you don't know what you are talking about." And, that's the point of my foreign policy article. I haven't just done the work at the tactical levels in the Thunder Runs and the Surge. In 2005, I was looking at this whole thing at a Corps level. There are other ways.

  12. Mike -- "How you got involved in all this" is that you've recently written on the subject, and commented approvingly on the On Violence post. I didn't think it unfair to try to summarize and respond to what I understood to be your argument; if you feel like I've unfairly represented you, I apologize.

    I've read your work on Zaganiyah, and there's obviously no substitute for first-hand experience and lessons learned. But I think we may be talking past one another a little bit. You cite suicide bombers as a sort of definitive disproof of rational choice theory, but there are a number of rationalist, instrumental explanations for apparently irrational acts –- Pape's being one of them. (To your specific example: from AQI's perspective, isn't it perfectly rational to use 12-year old girls as bombers rather than sparing able-bodied fighters?)

    It seems to me that you're falling into the same trap as On Violence: you've confused "rational" with "sensible," or worse yet, "sensible to us." If you consider me having pointed out this distinction to be "patronizing," then I apologize, but I'm quite sure that's not how it was intended.

    As to whether or not I know what I'm talking about, well, I suppose I have to leave that up to the reader. I don't disagree with your contention that "there are other ways," and I'm afraid you've misunderstood me if you take this post as general advocacy for the widespread utility or advisability of COIN operational methods. (I'm planning to deal with your FP article under a separate heading in the near future, but I wonder if I'm misreading your comment above as saying that its main theme is that I don't know what I'm talking about...?)

  13. Gulliver,

    Please stop reading into things. I never commented positively or negatively on Cummings piece. In fact, the only comment that I've made about it is that the advanced behavioral economists are trying to MERGE rational choice theory with emotional intelligence.

    Which, if you reread my article, is summarized by hearts, minds, and souls.

  14. I quite honestly, I don't care if you read, "deal," review, or critique my paper.

  15. I'm not sure what I've done to warrant your ire, but when you write things like that, it suggests moral certitude -- not a commitment to constructive dialogue and the improvement of our collective understanding. I fear it reflects a troubling recent inclination to teach others the truth, rather than an eagerness to learn.

  16. Gulliver,

    You still don't get it. I'm off to teach high school students. The immediacy that you are reacting to is me saying, "here is what I learned that I want to share with you before I go."

    You can use it as you like. You've never heard me ask you to 1. buy MY book, 2. go to MY website, or buy MY brand. If anything, I shared other books that helped me from other authors.

    That is all. And with that, Merry Christmas.

  17. Mike, I'm really having a difficult time understanding what you're getting at. I'm not asking anyone to buy anything. I'm not asking for anyone to promote my career. I'm trying to have a conversation with like-minded people that leads us to a better understanding of reality, and that hopefully allows us to offer up better ideas to policymakers and opinion leaders.

    Why can't you do this as a high school teacher?

    Why can't we do this in a way that allows us to remain open to the influence of other people's good ideas?

    Why can't we let the ideas speak for themselves, instead of speculating about malign motives and disparaging the credibility of others?

    I suspect I'm being used as a proxy for someone or something else in this instance. I'm still not even sure what it is you think we disagree about.

  18. Gulliver,

    You still don't get it.

    I'm done with war.