Dan Lindley

APSA Remarks on

"Is War Rational? The Extent of Miscalculation, Misperception, and Accident as Causes of War."

August 1, 2001

Welcome to a roundtable to discuss the question: "Is war mostly the result of rational, deliberate policies by states and leaders or is war mostly the result of miscalculation, misperception, or accident?" The issue of whether war is pursued for rational Clausewitzian purposes of state interest and strategic necessity, or whether war is more often the result of miscalculation, misperception, and accident is a fundamental schism in the causes of war literature. Even though this is a broad brush, most causes of war fall fairly clearly into one camps or another. For example:

Which of these two groups of causes has more explanatory power? This is a large, difficult, but also very important question. Strides to answer the question would bolster or undercut the assumption of rationality that is the bedrock for the R/C tradition and most quantitative scholars as well. The tools developed to tackle this question would also help assess a number of discrete arguments and debates, three of which I'll mention soon: Betts' critique of Van Evera's Causes of War, Sagan's argument that the Japanese calculated rationally prior to Pearl Harbor, and the arguments of Van Evera, Levy, and Sagan on WWI - with Van Evera arguing for a cult of the offensive, Sagan arguing for strategic necessity, and Levy arguing for a combination of calculation and miscalculation.

I think it is somewhat remarkable that there are few efforts to grapple with this question of rational vs. non-rational causes (the rational deterrence debate - George Downs' piece in particular being a notable exception). ⁽¹⁾ Quantitative researchers can not cope with miscalculation and misperception, and usually dismiss it in introductions to their projects by saying it's important, can best be examined with case studies, and thereafter ignoring the issue. Rationalists obviously assume rationality, though work such as Fearon's on uncertainty and incomplete or private information is quite nuanced and raises a tricky coding question: at what point is calculation based on incomplete or false information actually miscalculation? ⁽²⁾ And those case study researchers who are proponents of miscalculation and misperception often do not deal adequately with rationalist counterarguments. The prime problem in here is that not enough attention is paid to the issue of where the equilibrium would be in the absence of miscalculation and misperception. Did miscalculation and misperception make the crucial difference in going to war? The purpose of this panel is to call attention to these issues, and to spur research in this area.

Let's begin to make this more concrete. Stephen Van Evera came out swinging as a proponent of miscalculation and misperception in his book Causes of War. In response, Richard Betts swung back in his lengthy IS critique arguing that Van Evera underestimated the extent to which wars are fought for deliberate political reasons, in part due to a bias in favor of the prevention of war. Although Betts would be among the first to say that strategy is hard to formulate and execute, and that surprise attacks often succeed due to bungling, a theme in this review and his piece in Foreign Affairs about impartial intervention is that the search for manipulable causes of war leads many (liberals) to underestimate the extent to which war is fought for strategic and deliberate reasons. The core of the Betts-Van Evera debate is that Betts thinks that the causes of war are mostly Clausewitzian, and Van Evera does not. Who is right?

So, the issue is: how can one sort out the question behind this debate and figure out the extent to which war is rational and deliberate or based on miscalculation and misperception?

If I were looking into this question, and I plan to, I'd begin with a statistical overview to get a sense of how big the miscalculation problem might be. As rough indicators, I would examine how many war initiators lost wars, how many states started wars against more powerful states (and won or lost), how many states initially won their wars, but got balanced against and ended up losing, and so forth. ⁽³⁾ When Bueno de Mesquita wrote War Trap, initiators won 42 of 58 interstate wars, ⁽⁴⁾ suggesting on some level that correct calculation is more frequent than miscalculation. (sidenote: it would be interesting to see how this varies by scale and fatalities, as initiators seem to have lost many of the biggest wars: WWI, WWII, Korea...)

From there, I would proceed to case studies, and use two main criteria for assessing them: First, did states and leaders accomplish their goals with war? If not, was war at least a reasonable bet, was there a reasonable theory of victory? (ie accepts narrow rationality postulate and sees if war furthered those goals) Second, were these goals those that could reasonably be defined as Clausewitzian? What were possible alternative means and goals? If not the goals were not Clausewitzian, and/or the means seemed inefficient, what drove or colored the move toward war?

To give a brief example of these questions at work, let's look at the Japanese decision to bomb Pearl Harbor. Did the Japanese accomplish their goals? No, the Japanese did not accomplish their goals of making the US back away and/or coercing a deal with the U.S. At a basic and profound level, they miscalculated. But, was it a good bet? There was a slim (but, as Sagan shows, a fairly elaborate) chance that destroying Pearl Harbor would work, and the Japanese were very risk acceptant to take this chance. ⁽⁵⁾ Given that they thought they had no other choices, this may have been the only bet to make. So, in their shoes, this may have been a rational, calculated decision. And the goal of reducing dependence on oil is certainly strategically sound, and maybe necessary given increasing tensions with the U.S.. What about alternative means and ends, were these given full consideration? Here, Sagan presents what seems like smoking gun evidence: the Clean Slate debate in which the Japanese reconsidered 'anew' whether to withdraw from China to placate the US, whether they could successfully attack South for oil, and whether they could win a war with the U.S. A 'clean slate' debate is the epitome of rational decision-making.... except...that withdrawal from China was off the table for the military. Although Sagan doesn't quite put it this way, the clean slate debate was not clean, and once China was off the table, conflict with the U.S. was likely inevitable. Sagan suggests three possibilities to explain why the Japanese military wouldn't discuss China, in order of increasing 'Clausewitzianism': honor, aggressive intent, or his preferred interpretation: belief that war with the U.S. was inevitable with or without withdrawal from China. But there may have been a fourth reason, implying significant miscalculation and misperception: that the Japanese were infected with virulent hyper-militaristic/racist/nationalism. This argument is derived from writing of Japanese historian Saburo Ienaga, ⁽⁶⁾ and, if accurate, this is what took options off the table by coloring Japanese perceptions.

This raises the possibility that maybe it would be helpful to think of states and their leaders in the terms of psycho-analysis: if they have a neurosis, their views are colored and their choices somewhat biased or constrained and if they are psychotic, that represents a complete break from reality. Now, I know what most of you are thinking, how the heck do you code for that? That's obfuscating, and not clarifying. And I'm thinking, this sure isn't a pre-tenure exercise. Or as Jervis said about his work on systems effects: "graduate students: don't try this at home."

But the serious question remains, the difference between rational, Clausewitzian and miscalculation, misperception, and inadvertent causes of war is a fundamental dividing line in the literature, and is part and parcel of many specific debates from the rational deterrence debate, to causes of ethnic conflict ranging from ancient hatreds to resource wars, and to those mentioned here. Yet, have we made much progress resolving these debates? Do we have enough theorizing and observable implications to tell the difference between a Clausewitzian and non-Clausewitzian war? Is there enough data? If we had enough, people would be able to code wars similarly.

Lets take the example of a war for which there is perhaps the most amount of evidence and scholarship: WWI. Here, Van Evera argues that there is a cult of the offensive that created significant miscalculation and misperception about the offense that war seemed easy and cheap, and therefore became likely. On the other hand, Sagan ⁽⁷⁾ looks at WWI and finds that the offensive strategies of the day were rational responses to political objectives and alliance commitments, and that the instabilities caused by this strategy were not sufficient to tip the July crisis into war. For Sagan, much of WWI is politics by other means. For Van Evera, war is bureaucratic and organizational politics by other means.

Levy ⁽⁸⁾ makes a complex argument that states before WWI were not averse to war, and that many saw some advantages in limited war, but that no one "wanted, expected, or sought" a world war. On some level this is an obvious miscalculation and inadvertence argument, but the interlocking web of conflicting interests, war plans, domestic politics, etc. makes the war quite hard to code. In the end, Levy makes an 'exacerbated systemic pressures' argument in which systemic developments such as changes in power, and little Clausewitzian steps like testing, alliances, and so forth combine to make war all but inevitable.

So, who is right? And how and when will we know? What tools and data do we need to figure out who is right? That is the question which merits further research. And if these three experts on WWI can not agree on how to code this war, a war of great importance about which there is reams of data and scholarship, what does this say about the rationality assumption in the R/C and quantitative literatures? Is the rationality assumption questionable, or necessary because of these coding difficulties? But, if the rationality assumption is necessary, where do preference and utility functions come from?

In the end, the issue of whether war tends to be Clausewitzian or not is both important and difficult to sort out. When it comes to coding, it may come down to judgement calls. Despite the fact that we are social scientists with much to be modest about, we are dealing with war and peace and we have to try to make the most accurate judgement calls about the causes of war. This effort is only likely to produce intersubjectively agreeable results with more theorizing, more research, and with more methodological pluralism. There is much wisdom and data on the extent of Clausewitzian causes in the quantitative, r/c, and case study streams, yet not enough effort to pull it all together, or even to explicitly reject arguments from other streams when making one's own arguments.

A final note on policy. Whether war is mostly Clausewitzian or not has obvious policy implications. If it is Clausewitzian, deterrence and transparency should work to deter or coerce short of war. If not, and prevention of war is the goal, then more complex and perhaps much more difficult remedies may be needed. For example, to reign in militarism and nationalism may require rewriting of constitutions and textbooks. These sorts of solutions may require military occupation.

One thing that I think drives Betts and some other realists to try to shake people by the shoulders and point out that war is often a deliberate policy choice is that they fear that the real reason that liberals argue that war is caused by miscalculation and misperception and accident is because liberals want to believe that war is fundamentally an aberration. If it is an aberration, it is easily prevented: tincture of light peacekeeping, elixir of transparency, a little truth and light and there will be hugs all around. Betts may be right about the Clausewitzian nature of war, but he could push his policy implications further. If war is Clausewitzian, it is relatively easy to deter. But if it is the result of miscalculation and misperception, reducing the likelihood of war is much harder, for it would require changing the nature of bureaucratic politics, and human psychology or at least doing enough social science to figure out how to constrain these sorts of often malignant influences.

INTRODUCE SPEAKERS

We'll go in the order listed in the program:

1. Jack Levy: Miscalculation and Misperception: Methodological Issues

2. Stephen Van Evera: Miscalculation and Misperception as Causes of War

3. Scott Sagan: Do Accidents Cause War?

4. Bradley Thayer: Is War Rational? Insight from Evolutionary Theory

We are panel 19-8, Saturday 3:30 - 5:15. I'll try to keep my remarks to 10-15 minutes, and if you keep yours to around 15 minutes, we'll have about half an hour to talk inter-alia and for questions from the audience.

1. "The Rational Deterrence Debate," World Politics, 1/89.

2. IO, Summer 1985.

3. Work by Colonel T. N. Dupuy add another wrinkle. His data on battles, not war, show that even superior defenders often lose against weaker attackers. If it turns out that outcomes in war are equally hard to predict in advance, then it will be hard to judge whether miscalculation occurred. In-depth case studies will help determine if decision-makers went to war at least trying to make well-informed and deliberate choices, even if the final outcome is partially explained by chance. A danger in the study of miscalculation is to argue that only losers miscalculate. Blainey falls into this trap, but it may not be fair to decision-makers if outcomes can not be well predicted in advance. See Numbers, Predictions, and War (Fairfax, VA: Hero Books, 1985), pp. 12-15, as well as Richard K. Betts, "Conventional Deterrence: Predictive Uncertainty and Policy Confidence," World Politics, Vol. 37, No. 2 (January 1985), and Geoffrey Blainey, The Causes of War (New York, N.Y.: The Free Press, 1973).

4. 1981, pp. 21-22.

5. Scott D. Sagan, "The Origins of the Pacific War," The Journal of Interdisciplinary History, Vol. 18, No. 4 (Spring 1988)

6. Saburo Ienaga, The Pacific War: 1931-1945, (New York, NY: Pantheon Books, 1978)

7. 1914 Revisited, IS Fall 1986

8. Jack S. Levy, "Preferences, Constraints, and Choices in July 1914" in Steven E. Miller, Sean M. Lynn-Jones, eds., Military Strategy and the Origins of the First World War, (Princeton, NJ: Princeton University Press, 1991)