What Causes War? An Introduction to Theories of International Conflict [2nd ed.] 0742566501, 9780742566507

Citation preview

What Causes War?

What Causes War? An Introduction to Theories of International Conflict Second Edition Greg Cashman

ROWMAN & LITTLEFIELD Lanham • Boulder • New York • Toronto • Plymouth, UK

Published by Rowman & Littlefield 4501 Forbes Boulevard, Suite 200, Lanham, Maryland 20706 www.rowman.com 10 Thornbury Road, Plymouth PL6 7PP, United Kingdom Copyright © 2014 by Rowman & Littlefield Previous edition copyright © 2004 by Lexington Books All rights reserved. No part of this book may be reproduced in any form or by any electronic or mechanical means, including information storage and retrieval systems, without written permission from the publisher, except by a reviewer who may quote passages in a review. British Library Cataloguing in Publication Information Available Library of Congress Cataloging-in-Publication Data Cashman, Greg. What causes war? : an introduction to theories of international conflict / Greg Cashman. — Second Edition. pages cm Includes bibliographical references and index. ISBN 978-0-7425-6650-7 (cloth : alk. paper) — ISBN 978-0-7425-6651-4 (pbk. : alk. paper) — ISBN 978-0-7425-6652-1 (electronic) 1. War—Causes. I. Title. U21.2.C37 2013 355.02'7—dc23 2013008136

TM The paper used in this publication meets the minimum requirements of American National Standard for Information Sciences Permanence of Paper for Printed Library Materials, ANSI/NISO Z39.48-1992.

Printed in the United States of America

To my students and colleagues at Salisbury University over a most enjoyable academic lifetime.

Contents

Preface and Acknowledgments

xi

1 Empirical Theory and the Causes of War Empirical Theory Levels of Analysis

1 1 11

2 The Individual Level of Analysis, Part I: Human Aggression Why Are Human Beings Aggressive? The Nature-Nurture Dispute Nature: Early Ethology Nature: Second-Wave Studies in Ethology and Primatology Nurture: The Basics Nurture: Cultural Evolution Nurture: Peaceful Societies Nurture: Social Learning Theory Nurture: Three Experiments in Learning Peace and War Conclusion

13 14 15 20 30 31 34 43 45 47

3 The Individual Level of Analysis, Part II: Psychological Explanations for War Do Individuals Make a Difference? The Role of Reason: The Rational Model The Role of Personality: Psychological Theories The Role of Emotion: The Cognitive Revolution The Role of Bias: Heuristic Theories The Role of Risk Bias: Prospect Theory The Role of Images and Beliefs The Role of Decision-Stage Psychology: The Rubicon Theory The Role of Misperception The Role of Stress Conclusion

49 50 51 52 62 66 76 84 97 98 108 112

4 The Substate Level of Analysis: Group Decision Making The Role of Rationality in Small Groups: The Rational Actor Model

115 115

vii

viii

Contents

The Role of Rationality: Expected Utility Theory Why Governments Don’t Use the Rational Actor Model The Role of Rationality: Bounded Rationality Theory The Role of Structure: The Organizational Process Model The Role of Structure: The Bureaucratic Politics Model The Role of Group Psychology: Groupthink Group Concurrence Beyond Groupthink Prescriptions for Better Small-Group Decision Making Conclusion

116 119 121 128 132 152 156 166 168

5 The State Level of Analysis, Part I: Political, Economic, and Demographic Factors Liberalism: Some Basics The Role of Regime Type (1): Democracy and the Monadic Peace The Role of Regime Type (2): Democratization The Role of Regime Type (3): Rogue States The Role of the Business Cycle: Good Times and Bad Times Does Size Matter? The Role of Power, Size, and Development The Role of Population (1): Lebensraum The Role of Population (2): Lateral Pressure Conclusion

169 170 172 180 185 190 192 194 196 198

6 The State Level of Analysis, Part II: Internal Conflicts, Nationalism, and War Weariness The Role of Internal Conflict (1): The Diversionary/Scapegoat Theory of War The Role of Internal Conflict (2): Contested Institutions The Role of Internal Conflict (3): “Kick ’Em While They’re Down” Wars The Role of Internal Conflict (4): Revolution The Role of Internal Conflict (5): Outside Intervention in Internal Conflicts The Role of Internal Conflict: Summing Up and Implications The Role of Nationalism The Role of War Weariness Conclusion

199 200 210 213 215 218 220 221 231 236

7 The Dyadic Level of Analysis, Part I: The Nature of Dyads—Really Bad Dyads and Pretty Good Dyads The Role of Contiguity and Territorial Disputes The Role of Shared Ethnicity The Role of Rivalry The Role of Regime Types in Dyads The Role of Dyadic Power Balances Conclusion

237 238 251 252 257 275 277

8 The Dyadic Level of Analysis, Part II: International Interactions The Role of Action and Reaction in Dyads (I): The Spiral Model The Role of Conciliatory Responses in the Spiral Model: Liberal and Conservative Views

279 280 283

Contents

Implications for Theory and Policy: The Security Dilemma The Spiral Model: A Dissenting Opinion The Role of Actions and Reactions (II): Arms Races The Steps-to-War Theory GRIT as a Solution to the Conflict Spirals Conclusion 9 The Dyadic Level of Analysis, Part III: Game Theory, Bargaining, and Deterrence Theory Game Theory Bargaining Theory Deterrence Theory Deterrence: Empirical Research Cognitive Psychology and Deterrence Nuclear Weapons and Deterrence Deterrence: Conclusion

ix

289 291 293 306 312 317 319 319 330 339 348 360 362 366

10 The International System Level of Analysis, Part I: Realism, Anarchy, and the Balance of Power International Anarchy Classical Realism Neorealism Defensive Realism Offense-Defense Realism Offensive Realism Neoclassical Realism Realism and the Roles of Polarity and Polarization Bipolarity and Multipolarity: The Theoretical Debate Polarization: Empirical Research Alliances: Empirical Research Polarity: Empirical Research Combined Effects: Polarity and Polarization Polarity and Offensive Realism: Mearsheimer’s Balanced and Unbalanced Systems Conclusion

371 373 374 377 378 380 385 386 388 390 394 395 401 403 404 405

11 The International System Level of Analysis, Part II: Power Dynamics, Cyclical Theories, and Historical-Structural Theories of War Precursor Theory 1: Status Discrepancy Precursor Theory 2: Power Transition Theory Copeland’s Dynamic Differentials Theory Historical-Cyclical Theories of War Gilpin’s Theory of Hegemonic War Modelski’s Leadership Long Cycle Theory Wallerstein’s World Systems Approach Doran’s Relative Power Cycle Theory Systemic Theories: Implications

407 407 411 424 427 427 431 440 446 454

x

Contents

12 Constructivism: A Digression The Constructivist Approach: Basic Principles Constructivist Theory and War The Iraq War: A Constructivist Analysis

461 461 465 474

13 Conclusion The Causes of War: Some Basics The Perfect Storm Different Paths to War? The Causes of Peace: Is War Becoming Obsolete?

477 477 479 486 488

Notes

491

Selected Bibliography

569

Index

589

About the Author

607

Preface and Acknowledgments

The first edition of What Causes War? was published in 1993. It was the first volume of its kind that attempted to summarize and synthesize for undergraduate students of international relations the rather unwieldy set of research efforts by social scientists to discover the causes of war. In the twenty years since the first edition, empirical research on the causes of war has proceeded geometrically. Many new and fruitful lines of inquiry have opened up. The bad news for students of international conflict is that there is now quite a bit more to chew on intellectually. The expansion from the original ten chapters to the present thirteen is an indication of this. The good news is that we know considerably more now than we did twenty years ago. I have always thought that the legendary cosmological story of the earth resting on turtles is useful in understanding the progress of empirical knowledge about the causes of war. According to one version of the story, a famous astronomer is giving a public lecture describing the fundamental architecture of the universe. He is confronted after the talk by an old woman who dismisses his version of cosmology by stating that the world lies comfortably on the back of a giant turtle. Hoping to undermine her argument, he asks what supports the turtle. The answer is that he rests on the back of another turtle. “And under that one?” “It’s turtles all the way down, sonny.” The social scientific endeavor to determine the causes of war—and this particular work—is fundamentally similar. As a scholar engaged in the field for more than three decades, I acknowledge that what we know at this moment in time is based on the scholarly work of a great many academic turtles currently at work, and that their work would not have been possible without the work of the great generation of post–World War II scholars in this country and abroad, and that their work was greatly informed by earlier generations of scholars. “And so it goes,” as Kurt Vonnegut tells us. In many ways this book belongs to the previous generation of academic turtles, without whose diligent work and intellectual curiosity this endeavor would not have been possible. Their names are scattered throughout the pages of this book. I am particularly grateful to the late Prof. J. David Singer and to Prof. Charles F. Doran, whose support for the first edition of this book had a considerable impact on my professional xi

xii

Preface and Acknowledgments

development. I would like to thank several colleagues who have shared their thoughts with me over the years: Art Gilbert at the University of Denver, whose nimble mind and dedication to teaching has always inspired me; my political science colleagues Len Robinson and Taehyun Nam, whose frequent hallway conversations are a source of joy; Beth Ragan and Jenny Ross, who have guided my thinking on the anthropology, archaeology, and ethnography of early warfare; Natalie Hopson, who encourages my ability to practice psychology without a license; and Eric Rittinger who has sharpened my understanding of constructivist thought. I am particularly grateful to a long list of international relations students, both graduate students and undergrads, who have had nice things to say about the first edition. I would also like to thank Salisbury University for sabbatical leave and released time to assist in writing this book. I am especially indebted to my editor at Rowman & Littlefield, Susan McEachern, who gently pushed me toward this project five years ago and who has always been generous in providing feedback as this edition began to take shape. Her guidance and support have made the entire enterprise a tremendous pleasure. I am also grateful to Carrie Broadwell-Tkach, Janice Braunstein, and the editorial staff at Rowman & Littlefield, who provided sorely needed assistance in putting these words between the covers of a book jacket. Finally, I am indebted as always to my wife, Linda, who contributes emotional support, pizza, and a large order of reality testing.

Chapter One

Empirical Theory and the Causes of War

What is called wisdom is concerned with primary causes. —Aristotle

This is a book about the causes of war. More specifically, it is a book about the causes of interstate war—large-scale organized violence between the armed forces of states. It should be understood that organized violence can take many forms, for instance gang wars, intercommunal wars, civil wars, wars of secession, and wars of national liberation, as well as interstate wars. To attempt to analyze the causes of all these forms of organized violence would be to force on them more uniformities and commonalties than they possess. These various forms of conflict are significantly different and their causes are also analytically distinct. This book will focus therefore only on interstate wars to the exclusion of other forms of warfare. Most of what follows is based on the assumption that, all things being equal (which, of course, almost never is the case), wars should be tenaciously avoided. Although wars have been with us for many centuries, the increasing ability of governments to mobilize their peoples for war, coupled with greatly enhanced technology of mass violence, have substantially increased the destructiveness of wars in the last one hundred years. The central imperative of our times is therefore to avoid wars of mass destruction; all other goals suffer in comparison. As Jacques Cousteau once mused, “Why protect fish if the planet is going to be destroyed?” 1 If the central imperative of our times is the avoidance of war, the primary dilemma of our times is how to achieve this. One theme of this book is that if we can understand the causes of war, we should be better able to prevent their occurrence. EMPIRICAL THEORY Since this is a book about theories of war, one good place to start is to figure out just what is meant by the term “theory.” We quite often hear our friends say, “I have a theory about why we lost the basketball game,” or “I have a theory about why Barack Obama was elected president.” In most cases, what they mean by theory is what we would call a hunch, an educated guess. In this book the term “theory” will refer to something more than a hunch.

1

2

Chapter 1

Social scientists who analyze war are concerned with two kinds of theory: normative theory and empirical theory. Normative theories deal with how things ought to be. They deal with ethics and morals and value judgments. They are concerned with questions about what is right and wrong, what behaviors are morally acceptable or unacceptable. Normative theories of war would deal with questions such as: Can wars be just (morally acceptable), and if so, under what conditions? What kinds of practices and techniques of fighting are acceptable in war and which are unethical or immoral? These questions are within the realm of political philosophers. Although such questions will inevitably arise in the course of this book, we will be addressing the second kind of theory, empirical theory. Empirical theories—also known as causal theories—deal not with how things ought to be but why things are the way they are. They are concerned with what causes certain behaviors and outcomes. The goal of empirical theory is the explanation of behavior—in this case, war. Although there are numerous methods of explanation, empirical theory—at least in political science—implies the use of the “scientific method” of inquiry. Discovering and verifying the cause (or causes) of political phenomena cannot be based on sheer hunch or intuition; it must be based instead on rigorous tests of proof. Social sciences such as political science have adapted the scientific method used in the “hard sciences” like physics, chemistry, and biology for purposes of research in political behavior. This methodology involves a step-by-step process, which is applied in an effort to discover the cause(s) of any particular phenomenon. Since many of the theories of war examined in this book have been developed and tested through the application of the scientific method—especially those in chapters 5–11—it might be useful to provide a brief introduction to social science research. The Scientific Method Theories in the social sciences are essentially explanations concerning the causes of human behavior. Theories of war consist of explanations about what causes war and why this causeand-effect relationship exists. Logically, theories may be constructed in three different ways: through induction, deduction, or a combination of the two. In induction the analyst constructs the theory on the basis of observation of the facts (or data), working from the specific to the general. As the investigator learns more about specific wars, and as hypotheses are tested, theories are constructed and refined. In deduction the theory is constructed on the basis of logical reasoning, usually prior to the investigation of the relevant facts. This probably involves deducing the theory of war from a more comprehensive, general theory about international relations or politics. In reality, theories are probably constructed by scholars working in both directions, from the bottom up through induction from the facts about specific wars and from the top down from more general theories and principles. How are theories tested? In the social sciences it is likely that no theory can ever be conclusively proven true. But theories can be proven false. Essentially, theories are tested by testing hypotheses that are derived deductively from them. If the hypotheses are proven to be incorrect, then either the deduction is wrong or the theory is flawed. If the hypotheses cannot be disconfirmed, then the theory is tentatively accepted. It continues to be accepted until

Empirical Theory and the Causes of War

3

disproved at a later time. Theories therefore tend to “expire” for two reasons: (a) they lack confirmation, or (b) they are replaced by better theories. A quick guide to the use of the scientific method follows. Step One: The Formulation of Empirical Definitions for Concepts Each theory of war identifies concepts or factors that are deemed important to the understanding of the cause of war. Certain lines of inquiry are specified as being more fruitful than others. The determination of which questions to ask and which factors are important or unimportant can be achieved inductively from studying the data or deductively from general principles, but each theory identifies certain concepts as more relevant to the causes of war than others. Concepts are terms or words that identify general classes of things or ideas. War itself is a concept; so are tanks, decision makers, diplomats, perceptions, arms races, mobilizations, power transitions, and alliances. Specific persons or things are not concepts; though presidents and wars are concepts, President Putin and World War II are not. These terms refer to specific persons and things rather than to a general class of phenomena. For purposes of research, concepts must be given operational definitions; they must be defined in terms of something that is directly observable and measurable. For concepts that are directly observable, such as tanks, diplomats, mobilizations, and wars, this poses few problems. Other concepts, however, are more abstract and are not directly observable. For instance, the concepts of power, status, deterrence, hegemony, and democracy are not directly visible and observable. The construction of operational definitions for these concepts is a little trickier. Concerning the concept of interstate war, it is usually desirable to make some kind of operational definition in order to be able to determine what is a war and what actions belong to categories of military action short of war—like border skirmishes, for instance. In collecting data on international wars for the Correlates of War (COW) project, J. David Singer and Melvin Small operationally defined interstate war as a conflict involving at least one member of the interstate system on each side and in which the battle-connected deaths of all combatants together surpass 1,000. This has become the standard operational definition of interstate war. Some theories may require that we know more than just whether a war was present at a particular time or not. It may be necessary to define the concept of war more specifically in terms of the war’s severity, its magnitude, or its intensity. Singer and Small devised simple indicators for each: severity is measured by the total number of battle-connected deaths suffered by all participants in the war combined; magnitude is measured by the combined number of months each nation spent at war; and intensity is indicated by the number of battle deaths per nation-month. 2 Some concepts are best defined through the use of more than one indicator. National capability (referred to crudely as “power”) is a good example. Conceptually, power is based on much more than just military strength. So when we devise an operational definition of national strength we will want to include indicators of several different properties of national

Chapter 1

4

power. We might construct an index of national strength that takes into account these properties: 1. Geographic size—as measured in square miles 2. Population size—as measured by numbers of citizens 3. Technological development—as measured by yearly iron and steel production and/or energy consumption 4. Military strength—as measured by the number of men and women in armed forces and/ or by annual defense spending 5. Political stability—as measured by the number of months since the last unconstitutional change of regime Likewise, the concept of democracy might be operationalized by constructing a scale of the degree of democracy present in a country based on such indicators as these: 1. The degree of press freedom—as measured by the number of “independent” newspapers per capita 2. The degree of freedom of opposition—as indicated by the number of political parties or the number of political prisoners in jail per capita 3. The degree of electoral freedom—as measured by the presence or absence of direct popular elections for the chief executive and the national legislature, the degree of regularity in national elections, the average number of candidates running per office, and the presence or absence of referenda, secret ballot procedures, and universal suffrage 4. The degree of individual freedom—as indicated by the presence of constitutional guarantees of individual civil and political rights such as freedom of speech, assembly, and suffrage and freedom from unlawful search and seizure 5. The absence of a role for the military in the political process—as indicated by presence or absence of military candidates for public office and the presence or absence of military action to negate election results It should be obvious that the scores for different states (countries) will differ considerably on these indicators. States vary on the amount of power they have, the degree of democracy they have, and the amount of war they have experienced. These concepts can therefore be called variables. Variables are things that vary—concepts that can take different values. The major goal of theory is to explain variation. For instance, why do some states experience more wars (or wars of greater severity) than other states? Without variation there is nothing to explain. If, for instance, all states were equally warlike and if the incidence of war was constant over time, there would be little need for scientific inquiry. Step Two: Construction of Hypotheses Hypotheses are unproved propositions; they are essentially guesses about the causal relationships between and among certain variables. In other words they are guesses that a particular outcome or behavior (a dependent variable) is determined or caused by a certain factor or set

Empirical Theory and the Causes of War

5

of factors (independent variables). Hypotheses may be arrived at inductively through the observation of events, facts, and data. Or they may be arrived at deductively, working backward from a general causal theory. Usually, hypotheses are constructed through a combination of deductive and inductive reasoning. Hypotheses may take several forms. They may, for instance, be universal (absolute) or they may be probabilistic. Let’s look at some examples. Drawing either inductively from our knowledge of the past or deductively from our understanding of democratic theory (or both), we may wish to hypothesize that there is a causal link between democratic states and peace— and conversely between nondemocratic states and war. Let us use this as an example to illustrate the construction of hypotheses. H1 places the hypothesis in universal form: H1 = All democracies are peaceful. We may readily admit, however, that this is not likely to be the case in reality, that there are exceptions to this “rule,” and we should therefore loosen the form of the hypothesis to admit these exceptions. In the social sciences in general we find very few universal truths, and we therefore tend to use probabilistic hypotheses to reflect this reality. A better hypothesis might therefore be this: H2 = Democracies tend to be peaceful. This hypothesis introduces the notion of probability into the relationship. Rather than stating that democracies are always peaceful, it suggests that democratic states are probably peaceful most of the time. H3 states the hypothesis in slightly different form. H3 = If a state is democratic, then there is a high probability that it will be peaceful. This reformulates the hypothesis into the classic “if . . . then” statement of the relationship between independent and dependent variables. H4 states a slightly different variation of the same proposition. H4 = The more democratic the state, the less war it experiences. This formulation incorporates the idea that neither democracy nor war are absolutes; both may be placed along a continuum in which some cases contain more or less of the particular quality. In other words, it includes the idea of variation. States may vary in the amount of democracy they experience; they may also vary in the amount of war they experience over time. The hypothesis suggests that variation in one variable (the independent variable, democracy) explains, accounts for, or causes the variation in the second variable (the dependent variable, war).

6

Chapter 1

Step Three: Gathering Data and Step Four: Testing Hypotheses Once hypotheses have been formulated, they must be tested against the evidence of the real world. This is the essence of the scientific method. We need to know whether our hypotheses are in fact correct or incorrect. Does the relationship between the two variables we have proposed exist in reality? Is there in fact an association between the two variables? All of this presumes that one has evidence from the real world against which one may test the hypothesis. Sometimes this requires an extensive effort to develop “data sets” concerning things like when wars have occurred, which states fought in them, how many casualties were suffered, the degree of democracy that existed in particular countries at particular times, the power capabilities of each state, alliances that have been formed, and so on. If one is lucky, other researchers will already have done this kind of legwork. Testing of hypotheses must be rigorous. That is, you must try through a variety of methods to prove that the relationship does not exist! It is incumbent on the researcher to try to prove that her own hypothesis is false. Hypotheses are not proved by scavenging through the historical record to find examples of supporting evidence. You can’t validate hypotheses by simply citing facts to support them; this is too easy. You must actively search for examples that would disprove the hypothesis. Only if your search for this contrary evidence is futile may you claim success. How might we go about testing our hypotheses concerning the postulated relationship between democracy and peace? If we start with the universal hypothesis that all democracies are peaceful, our immediate tasks are to determine which states are democracies and to determine what constitutes peace. These are matters of operational definition. Let us resolve them quickly, for argument’s sake, in a relatively simplistic way. Let’s stipulate that democracies are those states that have in the last twenty years an uninterrupted record of regular election of a national legislature in which candidates from two or more political parties have participated. Peace will be defined as the absence of war participation in the last twenty years, with war defined as the presence of military combat with another state that involves more than 1,000 combat-related deaths for the states involved. If the universal hypothesis is true we should find, on inspection of the data, that no democracy was involved in war and that all states that were involved in war were nondemocratic. Table 1.1 is a simple two-by-two table that depicts how we would expect the data to be dispersed if the universal hypothesis was correct. We have already conceded that this outcome is quite unlikely in the real world and that some type of probabilistic hypothesis would be more desirable. Let us look, therefore, at H4— the more democratic a state is, the less war it should experience. Once again our immediate task (in addition to the accumulation of the relevant data) is to determine operational definitions of democracy and war. Democracy and war are no longer dichotomous variables. That is,

Empirical Theory and the Causes of War

7

they are no longer two-dimensional variables whose values are either O (absent) or X (present). Now the two variables must be put in a form in which they can be given values along a numerical scale—either on an ordinal (rank order) scale or an interval scale in which the numbers represent real values. Let’s assume that through diligent spade work and good “horse sense” we can create a rough 1–9 ordinal scale from such indicators of democracy as freedom of the press, freedom of opposition, electoral freedom, and individual rights. And let’s assume that a good indicator of overall democracy would be the average of a country’s score on these four separate indicators. Let us select as our measure of war the number of wars a country has been involved in during the past twenty years. We have then an ordinal level measure of democracy and an interval level measure of war. We can now conduct several tests of varying sophistication to determine the validity of our hypothesis. We might start by simply arranging the data in a way that helps us to analyze it visually. A three-by-three table suffices for this task, but it does require some arbitrary judgments. So let’s arbitrarily decide that countries that score from 0 to 3.0 on our scale of democracy are nondemocratic states; those that score from 3.1 to 6.0 can be classified as mildly democratic; and those that score from 6.1 to 9.0 can be described as democratic. Let’s proceed similarly with our other variable, war. If the average number of wars that nations engaged in over the twenty-year period was one, let’s define peace as zero wars; one war as mildly warlike; and two wars and above as warlike. In order for our hypothesized relationship to not be proven wrong, the three- by-three table ought to look something like table 1.2. Simply arraying the data in two-by-two or three-by-three tables as we have done is good for a start, but tests of greater sophistication will also be required. Researchers will want to use a variety of statistical tests to assess the degree of association between the independent and dependent variable and to determine whether this association might exist on the basis of pure, random chance. We will also want to ascertain whether other variables that are not part of our theory actually have more of a causal effect than the ones we have specified. Perhaps it is better to stop at this point, however, before we get into hotter water than we presently find ourselves with regard to methodological problems. We have said that a variety of tests should be performed. There are certainly many ways of testing the same basic hypothesis. For instance, a logical assumption based on our theorized relationship between democracy and war is that any specific country’s war involvement will vary along with its change in the level of democracy. This suggests that for any given country, wars should be more prevalent during periods of time in which the government was nondemocratic than during periods in which it was democratic. States ought to become more peaceful once they become (more) democratic. And democratic states that revert back to authoritarianism should then become more highly involved in war. Furthermore, since democratic states

8

Chapter 1

are presumed to be relatively peaceful, war between two democratic states should be extremely rare or nonexistent. Examination of these related hypotheses ought to strengthen (or weaken) our confidence in the correctness of the original tests. While the previous discussion has focused on the possibility of quantitative research, no matter how rudimentary, social scientists also use qualitative methods to investigate the causes of war. This is done by using comparative case studies. One popular way to do this is through the method of process tracing. In this method cases of war are examined with a focus on variables or factors that the theoretical or empirical literature has identified as important. The research is at once historical, conceptual, and theoretical. It results in a detailed, finegrained analysis of how particular wars occur. Researchers typically attempt to trace the causal process by which initial conditions or root causes combine with more immediate and proximate factors to develop, over time, into war. They seek to identify and examine the causal chains that produce war and then compare the similarities and differences across the cases examined. Two warnings need to be issued at this point. First, while we have been using a single variable explanation of war merely for the sake of simplicity, multivariate explanations of war are likely to be much more powerful. Since social and political behaviors are extremely complex, they are almost never explainable through a single factor. Decades of research have led most analysts of international relations to reject monocausal explanations of war. For instance, international relations theorist J. David Singer has suggested that we ought to move away from the concept of “causality” since it has become associated with the search for a single cause of war; we should instead redirect our activities toward discovering “explanations”—a term that implies not only multiple causes of war but also a certain element of randomness or chance in their occurrence. 3 The second warning is that statistical correlations between independent and dependent variables do not automatically mean a causal relationship has been established. For instance, one may find an inverse statistical relationship between the quantity of cranial hair in Soviet leaders and propensity toward liberal reforms. (Lenin, Khrushchev, and Gorbachev were bald reformers; Stalin, Brezhnev, and Chernenko were hairy conservatives.) Or one might find a positive statistical relationship between global gauze production and the severity of wars in the international system. This doesn’t mean that baldness causes the political reform or that a booming first aid industry causes wars of great severity. Causal inferences may be drawn only if (1) there is a time lag between the independent and dependent variable—logically, the potential causal factor must precede the thing being caused; (2) other variables can be proven not to be related to the variation in the dependent variable; and (3) a theory can logically and plausibly explain the relationship that is found to exist. Finally, we should mention than social scientists looking for the causes of war are interested in finding both necessary conditions for war and sufficient conditions for war. A necessary condition means that without the presence of this condition (A), wars will not take place. This doesn’t mean that the presence of condition A guarantees war; when A is present war may or may not occur. But without this condition A, there will be no war. To complicate matters somewhat, it is possible in reality that what is necessary for war to occur is the joint existence of several conditions—both A and B perhaps.

Empirical Theory and the Causes of War

9

A sufficient condition is a condition that ensures an outcome. This means that if this condition (C) exists, war will take place, whether other conditions (M, N, O) are present or absent. Sufficiency may require the joint presence of two or more conditions (C and D), and other sufficient combinations may exist as well (Y and Z). Sufficient conditions for war are different than necessary conditions for war. While C may be sufficient to cause war, it may not be necessary; wars may occur without it. Necessary conditions refer to conditions that must be in place for war to occur. But necessary conditions may not, by themselves, be sufficient for war; it is possible that other conditions must be present as well for war to take place. The Holy Grail in war research is to find the conditions that are both necessary and sufficient for war, something that is much easier said than done. Theory Revisited If the hypothesis is confirmed repeatedly by different observers using different tests, then it achieves the status of a law or generalization. Laws are confirmed hypotheses that state a relationship between two variables. Laws may be universal or probabilistic, like the hypotheses on which they are based. However, laws only state that there is an association between two (or more) variables. They do not explain why that association exists. Theories are needed to provide this explanation. As political scientists are fond of saying, “The data never speak for themselves.” Without a theory there is no real explanation. Suppose we find a strong relationship between wars and the prior existence of arms races. What then? How are we to interpret this? What is it about arms races that may make war more likely? Likewise, suppose it turns out that there is indeed a relationship between democratic states and peace—how is this condition to be explained? This is the realm of theory. More than one theory can appear to explain a set of facts and relationships. We may find that several theories simultaneously contend with each other to explain the same set of facts. If we discover, for instance, a relationship between arms races and war, several theories might purport to explain how arms races “cause” war. Probably the reigning theory is that arms races induce war because they heighten the tension, suspicion, and mutual fear between the countries involved. This initiates or exacerbates a spiral of hostility between the arms racers. The hostility spiral escalates until greater and greater levels of conflict and violence are achieved, leading to an all-out war. In this theory, arms races are associated with wars, but they don’t directly cause them; they only compound and intensify other conditions that are more directly related to the cause of war. Alternatively, however, the relationship between arms races and war might be more direct. It is possible to argue that a tremendous buildup in arms leads to bureaucratic pressures to use those arms that have been accumulating. Military institutions and their political and industrial supporters may feel the need to justify the expensive purchases of weapons by regularly demonstrating that massive arms supplies are actually needed. The only real way to demonstrate this need is to engage in large-scale fighting. Military, political, and industrial elites initiate arms races in order to reap economic benefits and to increase their power and status within their respective institutions. They then use their influence to push the country into the actual use of military force in order to preserve their power and increase their economic gains.

10

Chapter 1

These two explanatory theories proceed along very different lines of reasoning and in fact operate at quite different levels of analysis. They also lead to quite different hypotheses and tests so that it is logically possible to determine which of these theories—if either—is correct. Eventually, these conflicts between contending theoretical explanations tend to get sorted out. We have said that theories tend to be replaced over time by better theories; let us now turn to the question of what makes one theory better than others. Evaluation and Comparison of Theories Not all theories are equally good. Since the reader will be faced in the balance of this work with a series of contending theories that purport to explain the occurrence of war, it is essential to have in mind some criteria for judging the relative worth of these theories. Some characteristics of a good theory are listed below. While these criteria are in no particular order, those which the author sees as the most important appear at the end of the list. 1. Good theories have well-defined concepts that can be operationalized. 2. Good theories are clear and precise. 3. Good theories are simple, or parsimonious; they explain the phenomena with as few variables and as little complexity as possible. (One should probably not be excessively picky about this particular criteria; the world itself is not particularly simple and parsimonious. Oversimplification in a theory may result in a loss of explanatory power.) 4. Good theories should be plausible. They should make intuitive sense and not challenge too vigorously our sense of the possible and probable. 5. Good theories must be logically consistent. 6. Good theories must be testable and verifiable (and therefore falsifiable.) 7. The best theories are usually those which have the greatest empirical evidence to support them. As we shall soon discover, the evidence concerning the validity of most theories of war is rather mixed and sometimes contradictory. The quality and quantity of confirming evidence constitutes a major tool in evaluating and comparing theories of war. 8. Good theories are usually able to explain “anomalies” in other theories—gaps, or things that are unexplainable or are poorly explained in other theories. 9. The more general the theory, the better. Good theories explain more than previous theories; they have greater range and wider applicability. The goal of theory construction is to create a general theory of war that would be applicable to interstate conflicts in all geographic regions of the world across the historical lifespan of states. Such a theory would have as few cultural, geographic, or temporal boundaries as possible. 10. Good theories frequently build bridges to other theories. They provide a way of linking several theories across different levels of analysis. As the reader approaches the final chapters in this book, it will become evident that phenomena at many levels will play important roles in war causation. Single-factor theories of war and single-level theories of war have all proven inadequate to the task. Therefore, a truly comprehensive theory of war must take into consideration factors at multiple levels of analysis.

Empirical Theory and the Causes of War

11

“Islands of Theory” International relations theories have been for the most part middle range theories rather than grand theories of a more comprehensive nature that attempt to explain a wide range of phenomena. Most theories of war address a limited range of behaviors at a single level of analysis with as few variables as possible. For instance, middle-range theories might attempt to explain the relationships between alliances and war, deterrence failure and war, power transitions and war, territorial disputes and war, and so on. Though the connections between these middle-range theories are at present only faintly sketched in, the assumption made by most theorists is that eventually the accumulation of midrange theories at different levels of analysis will lead to the construction of theories of greater and greater sophistication and complexity. Nevertheless, Snyder and Diesing’s famous analogy remains apt: In our teaching and research, we are like travelers in a houseboat, shuttling back and forth between separate “islands” of theory, whose relatedness consists only in their being commonly in the great “ocean” of “international behavior.” Some theorists take up permanent residence on one island or other, others continue to shuttle, but few attempt to build bridges, perhaps because the islands seem too far apart. 4

Prediction Finally, a word about prediction is necessary. One of the goals of theorizing about war is the ability to predict with some degree of certainty when and where wars will occur. The development of theories about the causes of war may be helpful in this regard, but theory is not always necessary for prediction. We do not have to know why the sun has risen every morning in the east for the past several millennia to predict that it will also rise in the east tomorrow morning at pretty much the same time that it did today. Nor do we need to know what causes the ebb and flow of tides to be able to predict their occurrences with any accuracy. It is enough to have identified the relationships and patterns without being able to explain why the relationship exists. On the other hand, unique events cannot be predicted—only regular, patterned events. In fact, if all events are unique, they have unique causes and there is little use trying to explain and predict classes of events like wars in a general way. One of the chief assumptions made by political scientists is that events are not unique and political phenomena do not occur randomly; instead, recurring patterns and trends can be identified. Definite similarities can be discovered in the behavior of states. If these assumptions are false, it would be possible only to explain the causes of each individual war, and the cause of each war would be essentially different. LEVELS OF ANALYSIS The causes of war may be said to exist at several levels of analysis. We will examine theories of war at five levels: the individual level, the substate/small group level, the state, the dyadic level of interaction between two states, and the international system. These levels of analysis can be viewed as essentially levels of aggregation. Each level is made up of larger and larger

12

Chapter 1

units than the preceding level. Small groups are aggregations of individuals, states are aggregations of many groups, dyads are made of two states, and international systems are made up of the combined interactions of many states. At each level, different types of theories purport to explain the causes of war. At the individual level, the basic contention is that wars are due to the nature of humans or to the specific nature of certain individual leaders who take their states into war. At the substate/ small-group level, the argument is that individuals are rarely responsible for decisions to go to war and that these decisions are instead made by relatively small groups of officials within national governments. If one wishes to understand the cause of war, one needs to understand the processes by which these small groups come to their decisions. At the level of the state, the basic premise is that there is something about the nature of particular states that causes them to be more warlike or aggressive than those that lack these attributes. At the dyadic interaction level, the focus is not on the nature of states themselves but on (a) the nature of pairs of states (dyads), that is, their mutual or shared characteristics, and (b) the interaction between these pairs of states. Here the focus is primarily on patterns of interaction that escalate in intensity and hostility and lead to war. Finally, at the level of the international system, war is deemed to be a product of some aspect of the structure of the international system itself—the balance of power within the system, the hierarchical structure of status and prestige in the system, or cycles of economic growth and decline imbedded in the structure of the international system. We will proceed in this book from the individual level on through the international system level in search of the causes of war. Chapters 2 and 3 deal with the individual level of analysis, chapter 4 with government decision making by small groups, chapters 5 and 6 with the attributes of states, chapters 7, 8, and 9 with dyadic interaction between pairs of states, and chapters 10 and 11 focus on the international system. Chapter 12 will deal briefly with the constructivist approach to international relations, an approach that is not easily confined to a specific level of analysis. Chapter 13 will bring together some of the insights from each of these levels of analysis.

Chapter Two

The Individual Level of Analysis, Part I Human Aggression

We used to wonder where war lived, what it was that made it so vile. And now we realize that we know where it lives, that it is inside ourselves. —Albert Camus We have met the enemy and he is us. —Pogo (Walt Kelly)

We frequently hear comments such as “There will always be wars because men are just aggressive animals.” Or “As long as there are lunatics like Hitler or Saddam Hussein running governments, there is going to be aggression.” These views reduce the cause of war to either the nature of humans in general or the nature of particular individuals. While these statements discern the causes of war within individual humans, they illustrate two very different types of theories. In fact, they point to a level of analysis distinction within the individual level of analysis. 1 Those who believe that the fundamental cause of war is that humans are naturally aggressive take the position that all people are the same. National leaders who make the decisions to go to war are no different from the masses: they share with all people the same aggressive traits that characterize humans as a species. This collective characteristic of human aggression affects the process of war at the macro level of collective action. On the other hand, those who believe that the root cause of war must be found in the personal, psychological characteristics of particular national leaders argue that people are not all alike. The individual makes a difference. It matters whether Germany is ruled by Kaiser Wilhelm II, Adolf Hitler, or Angela Merkel. Aggression is seen as an individual characteristic rather than a collective characteristic, and its effect on war is felt at the microlevel of decision makers who hold in their hands the ability to choose war or peace. Let us develop the two ideas separately. Aggression as a common characteristic of the human race is the subject of chapter 2. Individual and psychological sources of war will be explored in chapter 3.

13

14

Chapter 2

WHY ARE HUMAN BEINGS AGGRESSIVE? THE NATURE-NURTURE DISPUTE Philosophers and theologians through the ages have sought to explain the aggressiveness of humans by explaining the nature of humankind. In his classic of political philosophy, Leviathan, the seventeenth-century English philosopher Thomas Hobbes described living conditions in the “state of nature” (that is, in primitive societies before the creation of government) as “a war of everyman against everyman.” This constant conflict stemmed, according to Hobbes, from human nature. People were self-seeking, selfish, and greedy; they were only concerned with satisfying their own desires. Personal gain and glory were their basic motivations. St. Augustine of Hippo (354–430) also noticed humans’ great capacity for mayhem and slaughter. This propensity for evil seemed to Augustine to require a theological explanation: original sin. Humans’ aggressive nature was related directly to the fall from grace in the Garden of Eden. Early psychologists like William James noticed that combat and war seemed to satisfy deep-rooted needs of individuals and societies, needs that were presumed to be inherent in all humans. This aggressive drive could not be suppressed, but it might be redirected and diverted toward more peaceful activities that involved similar challenges and exertions. Thus, James suggested the need to create a “moral equivalent of war.” 2 Youths might be conscripted to plant trees and build roads or dams rather than to kill the young men of other societies. Such programs would inoculate them with the same “social vitamins” as war without causing the same destruction of life and property. Sigmund Freud also believed the aggressive behavior of humans stemmed from deepseated unconscious drives. Freud suggested that an explanation for such aggressiveness might be related to the existence of the life instinct (eros), which seeks to preserve and unite, and the death instinct (thanatos). 3 The death instinct presumably has as its goal the removal of all tension, stimulation, and excitement in the individual. This death instinct is centered inward, and the logical outcome of its potent hold is suicide—aggression directed toward the self. But these drives do not exist in isolation; they interact with each other and modify each other. People live because the life instinct counters the death instinct and channels the drive away from the self and toward others. Overt aggression is thus the result of internal aggressive drives being redirected at others. Freud argued that not only must aggression be released in some way or another but that humans gain a certain amount of satisfaction from its release. In other words, humans need to satisfy these aggressive drives, though not necessarily by means of overt aggression. In more recent times, a major debate has been fought in academic circles and in the popular literature about the source of aggressiveness at the macro level. The debate centers on the issue of whether humans’ propensity for intramural violence can be attributed primarily to some universally inherent (and probably genetic) trait or to the specific culture and environment in which some groups of humans are raised. The debate is often characterized as the “nature-nurture” debate.

The Individual Level of Analysis, Part I

15

NATURE: EARLY ETHOLOGY Ethology, the study of animal behavior, emerged as a new discipline in the 1930s with the work of the Dutch biologist Nikolaas Tinbergen and the Austrian biologist Konrad Lorenz. The publication of Lorenz’s On Aggression in 1966 brought its theories widespread attention. 4 And writer Robert Ardrey’s popular summarizations (African Genesis, The Territorial Imperative, The Social Contract) of the work of Lorenz and other ethologists added to the awareness of these new ideas. 5 The inquiry of the first wave of these ethologists led to some vivid arguments. Their major thesis was that humans are a product of two million years of biological evolution. Lionel Tiger and Robin Fox (I’m not making this up) argued that humans remain “Upper Paleolithic hunters, fine-honed machines for the efficient pursuit of game.” We are biologically or genetically “wired for hunting—for the emotions, the excitements, the curiosities, the regularities, the fears and the social relations that were needed in the hunting way of life.” 6 The most extreme formulation of this position was put forward by anatomist Raymond Dart. Dart argued that man is a direct descendant of a killer ape, Australopithecus africanus. On the basis of Dart’s investigation of the fossil remains of africanus, he concluded that this particular ape was not only carnivorous but cannibalistic as well—an instinctual killer who may have killed for the enjoyment of it. 7 The existence of a large number of hominid remains that had incurred significant fracturing and puncturing was taken by Dart to be evidence of widespread interpersonal violence among africanus. It appears now that Dart was mistaken. The fossil evidence has been reexamined by others and the consensus now appears to be that the skull fracturing that Dart took to be evidence of human violence was the result of leopard predation. 8 For ethologists the concept of aggression refers only to intraspecific aggression—fighting between members of the same species. When two species fight (as when one species kills another for food), aggression is not involved. Perhaps the best example of aggression is seen when animals defend their territory against another of the same species. Ethologists initially saw aggression as an instinct or an innate drive that once helped to ensure the survival of the individual and the species. 9 As such, it was passed down from generation to generation as part of our hereditary makeup. Aggression is believed to have several species-preserving functions: (1) It keeps a balance in the territory between the needed resources on the one hand and the number of individuals to be supported on the other. (2) It aids in defense of the young. (3) It contributes to the survival of the fittest through the process of sexual selection. (4) It contributes to the establishment of stable social relations through the creation of dominant-subordinate systems such as the well-known “pecking order.” The problem is, of course, that the presence of such a drive in the modern age—with its weapons of mass destruction—could be extremely counterproductive. An interesting feature of this intraspecific aggression, however, is that it is generally not aimed at killing or extermination. Early ethologists pointed out that intraspecific aggression among animals usually does not result in death for the loser. On the other hand, the behavior of people appears quite different. Apart from rats, which also engage in “clan” wars and murder members of their own species, Lorenz identified humans as the only species that routinely slaughters its own kind.

16

Chapter 2

In reality, Lorenz was clearly wrong about this. We now know that several species also occasionally kill their own kind. For instance, E. O. Wilson has described the notoriously aggressive behavior of ant colonies toward one another and of “colony warfare” both within and between species. Colonies of the common pavement ant defend their territories with pitched battles conducted by masses of workers. Murder and cannibalism among the vertebrates also seem to be more prevalent than previously believed. Lions sometimes kill other lions, and there are reports of the killing and cannibalism of cubs after one of their protector males had died and the territory was invaded by other prides. Indeed, humans may not even be the most aggressive species—the hyena is often granted this distinction now. 10 In sum, intraspecific killing not only seems to be the norm, but it appears to be one of the primary reasons for mortality in the animal world. Modern accounts do not paint a rosy picture. At any rate, how did the early ethologists account for what appeared to be a critical difference between humans and most other animals? The answer for Lorenz was that intraspecific aggression among most animals is highly ritualized. The combatants do battle within the strictures of routine and symbolic patterns. As the course of the battle demonstrates the relative prowess of the contestants, the inferior competitor will engage in appeasement gestures or releasing signals, which will signal deference and submission, thus inhibiting further violence and preventing the combat from continuing to the death. The most frequently cited example of such inhibiting mechanisms is that of the wolf’s baring his neck during combat, an act that would presumably leave him vulnerable to be killed, but instead acts as a kind of surrender signal to the opponent, who then ends the contest. Humans apparently lack such inhibiting mechanisms. Why? The answer given by ethologists was that in the early stages of his evolution they had no need for them. Unlike the sabertoothed tiger and other predators, humans were unable to kill others quickly; without claws and fangs they simply did not have the tools for it. The sheer difficulty of hand-to-hand killing meant that most humans would give up the effort long before death resulted. If the physical difficulty was surmountable, presumably the aggressor would also be dissuaded by the anguished pleading of his opponent. However, assisted by the development of a greatly enlarged brain, humans later devised tools and weapons that could be used to slay their enemies—even at great distances—thereby reducing both the physical and emotional constraints on killing. By this time, however, it was too late for humans to develop those inhibiting gestures that other species had possessed for millennia. Instead of these “instinctual” mechanisms passed down through genetic imprinting, humans have been forced to rely on other measures to inhibit killing: morality, religion, ethics, and cultural and legal prohibitions. It would be an understatement to say that these have proven to be ineffective. The result is that while aggressive instincts were once species preserving, they no longer serve that function—just the opposite. Humans’ aggressive instincts, combined with the lack of instinctual inhibitory mechanisms and the ability to develop weapons of long-range destruction, mean constant conflict and death. Lorenz summarizes the links between aggression and human development: It is more than probable that the destructive intensity of the aggression drive . . . is the consequence of a process of intra-specific selection which worked on our forefathers for roughly forty thousand years, that is, throughout the Early Stone Age. When man had reached the stage of having weap-

The Individual Level of Analysis, Part I

17

ons, clothing, and social organization, so overcoming the dangers of starving, freezing and being eaten by wild animals, and these dangers ceased to be the essential factors influencing selection, an evil intra-specific selection must have set in. The factor influencing selection was now the wars waged between hostile neighboring tribes. 11

Thus, from the perspective of the early ethologists, wars provide an outlet for the aggressive tendencies that are inherent in human beings. Indeed, Lorenz saw aggression as a drive that must seek release; in other words, humans have an innate “need” for aggression. Some have referred to this conception of aggression as the drive-discharge model; aggression is seen as a drive that seeks release or discharge, thus impelling humans toward aggressive activities. Others call it the hydraulic model, analogous to the pressure created by water that is held in check by a dam. 12 In other words, a kind of energy accumulates in the instinct center of the animal, generating pressure for its discharge. Aggression is in this sense spontaneous. Its source is internal to the organism, not external. There was considerable dispute among ethologists (and others) about how these aggressive actions were set off. The question centered around the stimuli needed to elicit such a response. Lorenz and psychiatrist Anthony Storr suggested that although the physical mechanism of aggression was inborn, it was still usually triggered from the external environment. But they also argued that the fact that aggression needed an outside stimulus to set it off did not imply that humans could escape the need to behave aggressively. Lorenz maintained that the longer the aggressive energy was dammed up, the lower the threshold value of the stimulus needed to elicit the aggressive response. Indeed, he speculated that following an extensive period of damming, aggression could be performed without the presence of an external stimulus. 13 Storr agreed, suggesting that when no external stimulus exists that was capable of stimulating aggression, humans would actually seek out such a stimulus! 14 Others, such as psychologist J. P. Scott, while agreeing that aggression is rooted in a physiological process and that it needs to be activated by external stimuli, argued that aggression doesn’t really ever need to be manifested. Since the release of aggression must be stimulated by an outside trigger, if that trigger is missing, aggression will not occur. 15 If Scott’s more optimistic view is correct, humans are not doomed by its genetic inheritance; violence is avoidable. A final idea dear to the heart of ethologists is territoriality and the relationship between territory and aggression. Ethologists cite considerable evidence that almost all vertebrate animals are territorial; they use aggression to gain and defend territory. 16 This is certainly true of chimpanzees, humans’ closest relative on the evolutionary tree. The argument from this perspective is that humans’ genetic inheritance has provided them with the same territorial instincts as their lower relations. (We will have more to say about territoriality later.) Critique of Ethology Lorenz and his followers have been criticized both for their methods and for the validity of their results. The substantive evidence seems to be extremely weak and based primarily on extrapolating disputed evidence from the behavior of certain animal species to human beings. A few specific criticisms should give the reader some of the flavor of the debate.

18

Chapter 2

1. The drive-discharge model. This part of the original ethologists’ position is now widely discredited. If there really is an accumulation of aggressive “energy,” we should be able to find some physical evidence of this—presumably in the brain. However, there is no neurophysiological evidence for the existence of any kind of energy accumulation in the brain that impels a spontaneous discharge. Unlike the internal changes (the change in blood sugar, for instance) that accompany the onset of hunger, no internal changes develop that precede the onset of aggression. 17 This is not to say that aggression itself is without neurological effects. Scientists have identified certain centers within the brain associated with different kinds of aggressive responses. (The hypothalamus seems to be the center for the emotion of anger, for instance.) Aggression has also been experimentally manipulated in subjects through hormone treatment and brain stimulation. But the brain centers most involved in aggression respond primarily to signals given after some external stimulus is interpreted, and endocrine levels in the body are typically affected in response as well. In other words, the physiological manifestations of aggression are brought about by the individual’s response to perceptions of the external environment rather than from within. 18 2. Is aggression really an instinct? Since individuals begin to learn from their environments at very early ages, it is very difficult to scientifically ascertain whether a particular behavior has resulted from a preexisting instinct or has been learned. Owing to this difficulty in distinguishing learning from instinct, many scholars have abandoned the term “instinct” altogether. 19 Some, such as Ashley Montagu, suggest that where Lorenz and the ethologists err is that humans are humans precisely because they have no instincts, at least beyond the reaction of infants to sudden loud noises or to the sudden withdrawal of support. 20 3. Humans are different from other animals. Lorenz’s primary failing would seem to be that he fails to recognize that humans are substantially different from other animals because of their enlarged brains. This makes them able to adapt and respond to their environment and to think. Behavior based on rational thought represents a distinct advantage over behavior based on instincts. This fundamental difference ultimately means that generalizations based on observations of the behavior of lower animals ought not to be assumed true of humans. Generally speaking, the more advanced the species, the less genetics determine behavior. 4. Monocausal explanation. Lorenz’s theory is incapable of explaining many aggressive behaviors because it relies on a single causal factor. It ignores other variables that may play a role in determining whether aggression takes place—such as the presence of frustration, the character of the sociopolitical environment, or the ability of humans to reason and to learn. 5. Territoriality. Although territoriality occurs only in higher animals (vertebrates and anthropods), some critics have argued that humans’ nearest relatives among the primates (savannah monkeys, chimpanzees, bonobos, and gorillas) do not show much territorial behavior—either collectively or individually. 21 However, these arguments were generally made prior to Jane Goodall’s famous discoveries at Gombe and subsequent confirming evidence of chimpanzee territoriality from other sites. Recent observations have indeed confirmed the territorial nature of many of primate species. The crucial point, however, is that we know virtually nothing about the territorial behavior of preliterate humans, but anthropologists note that the institutions of territory and private property vary greatly in modern humans. Scott cites as an example that in none of the Eskimo

The Individual Level of Analysis, Part I

19

societies except the Aleut was there anything such as defense of territory. 22 To the extent that territoriality exists, it may have a cultural basis instead of a biological one. Anthropologists generally argue that territoriality among contemporary hunter-gatherers appears to depend upon the situation and environment. 23 It is more likely among those dependent on stationary food sources, especially those areas having dense, predictable resources and relatively large populations, and less likely among those who follow migratory game animals (or who live in sparsely populated areas). The same was probably true with early humans. But even if band societies in Lower Paleolithic times were territorial, this would not necessarily have promoted warfare. Contemporary bands that are territorial are much more likely to fight over murder, witchcraft, and wife stealing than over territory-related issues like trespass. Many anthropologists argue that warfare over territory only appears with the Agricultural Revolution. 24 6. Methodology. Lorenz doesn’t form operational hypotheses that can be empirically tested. Instead, he relies on reasoning through the use of metaphor, analogy, and cross-species extrapolation. For instance, Lorenz argues that the cause of aggression in birds and fish is essentially the same as the causes of animal aggression, just as the cause of animal aggression is essentially the cause of human aggression. In addition, the cause of interindividual aggression is presumed to be the same as that of intergroup aggression and of interstate war. This is simply sloppy science. One should not use observations of the behavior of one species to explain the behavior of other species. Neither should one use observations of individuals to explain the collective behavior of groups. As Barbara Ehrenreich has said, “Wars are not barroom brawls writ large, but, as social theorist Robin Fox puts it, ‘complicated, orchestrated, highly organized’ collective undertakings that cannot be explained by any individual impulse.” 25 Thus, Lorenz compounds problems of interspecies analogies with level of analysis problems. Any attempt to apply the theory of instinctual aggression to the behavior of states in the international system should be treated with caution. In the final analysis, the substantive evidence for the original ethological theory of aggression is weak. There appears to be no single instigation to aggression even among lower animals; instead, there are several kinds of aggression, and patterns of aggression vary from species to species and even within species. Samuel Kim argues that as we climb the evolutionary ladder, aggression becomes less common, and when it does occur its causes are more complex, more numerous, less rigidly programmed by the genotype, and more influenced by ecological and experiential factors. 26 To the extent that violence occurs among primates, it appears to be situational rather than instinctual. Even Goodall argues that extreme violence is a rarity among chimpanzees: Because violent and brutal behavior is vivid and attention-catching, it is easy to get the impression that chimpanzees are more aggressive than they really are. In fact, peaceful interactions are far more frequent than aggressive ones; mild threats are more common than vigorous ones; threats per se are much more frequent than fights; and serious, wounding conflicts are rare compared to brief, relatively mild ones. Moreover, chimpanzees have a rich repertoire of behaviors that serve to maintain or restore social harmony and promote cohesion among community members. 27

20

Chapter 2

NATURE: SECOND-WAVE STUDIES IN ETHOLOGY AND PRIMATOLOGY While the works of Lorenz and other scholars of the first wave of ethological studies in the 1960s and 1970s focused primarily on the behavior of fish and birds, a second wave of scholarship by ethologists and primatologists (scientists who specialize in the study of primates) has taught us much more about the behavior of humans’ closest biological relatives— chimpanzees, bonobos, and gorillas. 28 Since chimpanzees are closer to humans physiologically and genetically than any other creature—in their DNA structures, the two species vary by just over 1 percent—evidence concerning the aggressive nature of chimps would seem to be crucial to the ethological argument. 29 Over several decades at Gombe National Park in Tanzania, Jane Goodall observed many behaviors that have shattered our preconceptions of humans’ evolutionary cousins. For instance, she discovered that chimpanzees are not only tool users but also tool makers. They also collect rocks for use as weapons and store and hide them for later use. 30 Perhaps the most startling revelations, however, deal with intercommunal violence. While violent conflict associated with establishing dominance among males within the community is quite ritualized and struggles to the death appear to be rare, rivals frequently incur severe physical punishment during dominance contests. Goodall and her associates also observed aggressive territoriality among chimps. Chimpanzee territories overlap considerably, and while chimps spend most of their time in the center of their territories, they frequently patrol their borders, though they tend to do this primarily in parties with many males—seeking safety in numbers. According to long-time observers, “territorial competition appears to be a pervasive feature of chimpanzee societies.” 31 Moreover, such territorial confrontations, while relatively infrequent, are not always ritualized and harmless. Goodall famously observed a four-year, one-sided slaughter involving rival communities. After a chimp community she was observing divided into two separate territorial groups, the adult males of the original community, working collectively in what appeared as almost premeditated and coordinated attacks into the smaller group’s territory, totally eliminated the members of the breakaway group one at a time over a four-year period. Each of the “secessionists” was brutally beaten to death by former friends. In some instances, one chimpanzee held the victim down while the others struck at it and bit it, in an attack that continued for ten to twenty minutes even though the victim was unable to fight back. Acts of cannibalism sometimes accompanied this violence. Some of the smaller group’s surviving females were reincorporated into the group and their territory was “annexed.” 32 Goodall remarked that “If they had had firearms and had been taught to use them, I suspect they would have used them to kill.” 33 What Goodall observed at Gombe turns out to have been not too different from what others were seeing elsewhere. Over a ten-year period, scientists at Kibale National Park in Uganda observed chimpanzees from the Ngogo community killing or fatally wounding eighteen members of neighboring communities, and they inferred three other possible violent deaths. As a result of this territorial violence, the community was able to annex part of the territory of its rivals. At Mahale in Tanzania the “K-Group” mysteriously lost all six of its adult males and one adolescent male between 1970 and 1982. Although they had not directly seen lethal

The Individual Level of Analysis, Part I

21

attacks, observers speculated that the males must have died as a result of attacks from the large neighboring “M-Group.” In the wake of the collapse of the K-Group community, most of its female members joined the M-Group. Neighboring groups absorbed the K-Group’s territory. Chimpanzees in the Tai National Forest in Ivory Coast have also been seen engaging in raiding behavior that involved direct violent attacks carried out by groups of five to six males. The attacks killed both adults (usually males) and infants. In fact, the killing of unprotected infants was somewhat more widespread than the killing of adults. In many of the cases of infanticide, cannibalism was reported. Chimps at all long-term sites engaged in activities to defend the group’s territory, and chimps at four of the five sites engaged in lethal attacks on neighboring groups. In sum, detailed data from five long-term sites indicate similar patterns of intergroup violence. Hostile interactions between neighboring groups are the norm. 34 It appears that chimpanzee violence is very sensitive to context, in particular to balances of power. Most attacks take place when a group of two or more chimps from one troop takes opportunistic advantage of finding a single individual from a rival troop traveling alone. Finally (and this is important), no one has yet observed two groups of chimps engaged in mutual (two-sided) and sustained combat with each other. 35 Battles between relatively equal groups from neighboring communities are normally avoided (though the 2012 Disney movie Chimpanzee purports to show exactly this). It turns out that chimpanzees are quite good at estimating the strength of the opposing side from the “pant-hoot” calls of each group, and they normally beat a hasty retreat if outnumbered. 36 When violent encounters between groups with many individuals do occur, the outcomes are usually indecisive and lack serious injury— unless one side is able to isolate an individual rival. Patrolling attacks in which a group of several males encounter an isolated individual from the rival group are the main source of fatalities. 37 It turns out that these observed behaviors eerily parallel the behavior of primitive human groups. Wilson and Wrangham note three tendencies common to chimpanzees and humans in hunter-gatherer societies regarding interactions with other groups: (1) they respond aggressively during encounters with members of other groups, (2) they avoid aggressive confrontation in battle lines of large numbers, and (3) they seek to take advantage of opportunities to use imbalances of power in which males can kill members of other groups. Most deaths are the result of asymmetric contests. Why such similarities exist is still a subject of great debate. Are they the product of evolution from a common human-chimp ancestor? (The two species split approximately 5 million years ago.) Or did such behaviors evolve independently in the two species for similar reasons? 38 Ethologists have also taken a look at a less well-known chimpanzee species: the bonobo. Bonobos (Pan troglodytes), sometimes erroneously called pygmy chimps, are a relatively rare chimp species living only in the rain forests of the Central African Republic. Though a separate and distinct species, they also share about 98 percent of their DNA with humans. Bonobo societies however, are quite unlike those of standard chimps (Pan paniscus). Females dominate the social system, and cooperation is enhanced by food sharing. Male bonobos are considerably less aggressive than their chimp counterparts and they do not form coalitions. Bonobos are also less territorial, and unlike other chimps, bonobos display no tendency toward

22

Chapter 2

collective or organized violence. Intense intermingling of neighboring groups occurs rather frequently without violence. 39 In sum, the level of violence among primate species varies greatly. Some are remarkably peaceful (like bonobos and stump-tail macaques), while others (like chimps, gorillas, and rhesus monkeys) fall toward the more violent end of the spectrum. Peacemaking among Primates While the potential for aggression may be part of our genetic heritage, so is the ability to reconcile the effects of aggression and restore peace. Frans de Waal, a zoologist and ethologist whose primary interest is the behavior of primates, has argued that the ability of a species to reconcile after episodes of conflict is an important part of its genetic heritage. 40 “Making up” after fights through some kind of peacemaking or reconciliation behavior is widespread among primates—including chimpanzees, bonobos, mountain gorillas, golden monkeys, and stump-tailed monkeys. It has been observed in twenty-seven primate species, though significant variation exists in reconciliation activity as well as in aggression. 41 Each primate species has its own standard repertoire of peacemaking behaviors: “Golden monkeys do it with mutual hand-holding, chimpanzees with a kiss on the mouth, bonobos with sex, and tonkeana macaques with clasping and lipsmacking . . . Many have evolved gestures, facial expressions, and calls specifically for this purpose, whereas other species lack such behavior.” 42 De Waal’s argument is that aggression and peacemaking are both indispensable tools in the behavioral repertoire of social species like primates. Indeed, complex societies would come undone without the ability to mitigate the effects of conflict. Why? First of all, primates, like humans, are social animals who live in communities. For many primates, leaving the community is not an option; for protection (against other species and against members of the same species in neighboring communities) they must live with and depend upon others. Primates need to get along with the other members of their community and cannot afford to be alienated. Second, primates are feisty; competition and low-level fighting are endemic in many primate species. De Waal reports that the typical rhesus monkey community experiences eighteen acts of aggression per monkey every ten hours of observation. 43 And aggression is hardly ever a “one and done” event, but part of a series of ongoing encounters. Moreover, the constant search for allies in the struggle for male dominance means that today’s foe may be needed as an ally tomorrow. 44 Third, primates are intelligent; they have long memories (including memories of violence committed against them) and they can look forward and plan ahead. In communities where acts of even small-scale aggression are endemic, the unrelenting stress of constant squabbles might well overwhelm the inhabitants in the absence of behaviors designed to facilitate “making up” after fighting. 45 Consequently, according to de Waal, peacemaking behavior evolved in conjunction with aggression. Peacemaking was adaptive for primates since it increased group cohesiveness, thereby aiding the ability to defend the community against outsiders. This may be applicable to humans as well. As de Waal argues, “In view of the close-knit community life and strong interdependency among [currently existing] hunter-gatherers, we can speculate that the capacity to find alternatives to overt aggression, and to restore the social fabric after disruption, must have been of critical value in human evolution.” 46 If chimpanzees, bonobos, rhesus

The Individual Level of Analysis, Part I

23

monkeys, and stump-tailed monkeys routinely make up after bouts of fighting, similar behavior by their human “cousins” quite probably has the same origin. 47 The similarity of peacemaking behaviors among higher primates suggests a common evolutionary origin and a genetic basis for human behavior. According to de Waal: The fact the monkeys, apes, and humans engage in reconciliation behavior means that it is probably over thirty million years old, preceding the evolutionary divergence of these primates. The alternative explanation, that this behavior appeared independently in each species, is highly “uneconomical,” for it requires as many theories as there are species. 48

The fact that humans are genetically wired for peacemaking as well as for aggression is very good news, but it is a limited sort of good news. Chimpanzees from different territorial troops rarely reconcile. The peacemaking behaviors we have been discussing all refer to intragroup personal conflict between individuals, and this is significantly different from conflicts between states composed of different groups of humans. 49 Nature: Sociobiology One interesting response to the popularity of ethology was the creation of a new field of study: sociobiology. Seen variously as a subfield of ethology, biology, or of evolutionary theory, the term was coined in 1946 by John Paul Scott. However, sociobiology officially entered the world with the publication in 1975 of Edward O. Wilson’s Sociobiology: The New Synthesis. 50 This new approach was labeled by its author as an effort to place all of the social sciences within a biological framework constructed not just upon ethological studies of animal behavior but on the study of evolution, genetics, population biology, psychology, and anthropology as well. Whereas evolutionary biologists have applied evolutionary theory to explain the development of physical characteristics (like the development of tails or opposable thumbs), sociobiologists attempt to apply it to the social behavior of animals and humans. Sociobiology is most controversial when explaining human behavior. Although Wilson argues that human behavior has been programmed to a substantial degree by natural selection, he doesn’t claim that genetics are the sole cause of behavior. Instead, his theory recognizes the interaction of genes with the cultural environment. Nevertheless, Wilson emphasizes the genetic determinants of human behavior and human culture. He argues that some human behaviors are at least partly inherited; they have evolved over time and have been affected by natural selection. Thus, if one wants to understand social behavior, one must analyze it in terms of evolutionary considerations. The controversial nature of sociobiology has led many of its adherents to label themselves as belonging to the related fields of evolutionary psychology or behavioral ecology. How do we know that human behavior has a genetic basis? First, the social cultures of humans and chimpanzees, humans’ closest relatives, are similar. But second, says Wilson, human behavior is also distinct from the behavior of our evolutionary relatives in ways that can only be accounted for by a unique set of human genes. Virtually every human culture known shows certain distinct characteristics; wherever they are, humans act similarly. For example, every human culture would seem to have these characteristics in common: athletic

24

Chapter 2

sports, community organization, cooperative labor, division of labor, education, ethics, etiquette, funeral rites, gift giving, government, hospitality, inheritance rules, kinship nomenclature, language, law, marriage, penal sanctions, population policy, property rights, religious ritual, residence rules, sexual restrictions, status differentiation, and trade, among others. 51 This could not have happened randomly; it must be genetic. How else could so many human societies have developed the same patterns of behavior? The concept of adaptiveness provides a key to understanding how this has happened. Certain behavioral traits of human nature were “adaptive” during the time that human behavior evolved, and genes consequently spread through the population that predisposed their carriers to develop these traits. Adaptiveness means that if an individual displays these traits, he or she stands a better chance of having his or her genes represented in the next generation than if he or she did not. This advantage is called genetic fitness. Behavioral traits that helped individuals adapt to a particular environment would tend to be passed down to further generations in larger numbers (they will be “selected for”), while behavioral traits not well adapted to the particular environment would be unlikely to persist in the population. As Bradley Thayer says, “The essence of evolution by natural selection is that most behavioral characteristics of a species evolve because they help the species survive and reproduce.” 52 Wilson contends that the greatest part of genetic evolution occurred over five million years ago, long before the era of human civilization. There has been some evolution since then, but not enough to affect a large number of traits. In Sociobiology, Wilson advocated a particular theory of evolution referred to as either inclusive fitness or kin selection. Inclusive fitness was originally developed by William Hamilton in the 1960s to explain altruistic behavior in a way that was consistent with Darwin’s principle of natural selection. A vexing puzzle for evolutionary theory was how to explain the existence among humans (and other primates) of altruistic behavior—behavior in which an individual sacrifices his own interests, perhaps even his own life, for the good of others. From the viewpoint of the specific individual, the act seems detrimental to his own reproductive success. If he dies, this severely limits his ability to pass on his own genes to future generations. However, altruistic acts may promote the reproductive success of genetically related individuals and hence work to maximize the number of shared genes within future generations. As one scholar notes, it is “better for your own biological fitness to sacrifice your own life to save the lives of three of your children (each of whom has fifty percent of your genes) or, even better, seven of your grandchildren (each of whom has twenty-five percent of your genes).” 53 Natural selection favors organisms that behave in ways that maximize their inclusive fitness. Thus, social behaviors evolve because of the benefits they create not just for the genetic fitness of individuals but also their close relatives. Because an individual’s close kin constitute a reservoir of his or her own genes, they are worth caring for in an evolutionary sense. Thus, altruism is an adaptive trait that adds to the genetic fitness of kin groups. 54 Cooperation and altruism are innate traits, presumably because they add to genetic fitness. That is the good news. The bad news is that humans are also innately aggressive. This genetic/ biological disposition is necessarily reflected in human social and cultural institutions. Wilson sees organized warfare as endemic to every form of human society, from hunter-gatherers to

The Individual Level of Analysis, Part I

25

modern urban-industrial society. Like Lorenz, Wilson argues that humans’ aggressiveness has added to their genetic fitness through preservation of the territorial balance, defense of the young, mating behavior, and survival of the fittest. Because aggressiveness has added to genetic fitness, there is a fairly high probability that the trait will develop in a specific set of environments, but there is no certainty that the trait would develop in all environments. One should not expect, therefore, that all societies will be aggressive. 55 Evolutionary theory suggests that groups may go to war to increase inclusive fitness. Political scientist Bradley Thayer argues that this is logical for defensive reasons (genetic fitness is increased if groups can defend their territory, their resources, and their ability to survive and reproduce) and for offensive reasons (groups can increase their fitness by improving their material conditions by taking other groups’ resources). 56 So it is possible that intergroup aggression and warfare were behavioral adaptations to situations in which kinship groups confronted competition for resources. In addition to aggression and altruism, sociobiologists and other evolutionary theorists point to several other social behaviors that appear to be the result of kin selection. 57 First, dominance and the creation of dominance hierarchies within groups may reduce intragroup conflict and help defend against outsiders. To the extent that evolution has produced a need to belong to a dominance hierarchy, this helps to explain patriotism and the attachment of individuals to the state. Second, both xenophobia (the fear of strangers) and ethnocentrism (belief in the superiority of one’s own group and preference for its members) help explain the ubiquity of intergroup warfare and the scale of violence associated with it. All would appear to be beneficial to group survival, and all make some contribution to the ability or willingness of groups to use violence against others. Laboratory tests and field observations of both primates and humans show that strangers trigger automatic responses of heightened alarm, fear, suspicion, insecurity, and aggression. Azar Gat argues that ethnocentrism, xenophobia, patriotism, and nationalism all have their evolutionary roots in the existence of early kinship groups—a subject worth exploring in more depth. 58 It is of evolutionary significance that 99.5 percent of man’s evolutionary history—the period in which modern humans evolved to what they are today—was spent in the huntergatherer stage of existence; since the Late Paleolithic age human evolution has largely been affected by cultural evolution rather than biological evolution. It is also important, from the perspective of inclusive fitness, that during this period hunter-gatherer communities were essentially small kin groups. They also formed regional groups (fraternal interest groups of males) of about five hundred, which constituted a dense network of extended kinship. At an even higher level, they formed confederations of regional groups (dialect tribes). An important function of these regional groups was cooperation in warfare. 59 The closer the kinship connection between an individual hunter-gatherer and these larger groups, the greater the evolutionary payoff for protecting them and the more likely one was to fight for them and with them. Moreover, hunter-gatherers who were attached to these larger regional groups had an advantage over communities that lacked such links. After the Agricultural Revolution, shared cultural communities expanded, and the perception of kinship solidarity expanded with them so that individuals came to fight for “imagined communities.” Who is “us”—always a function of

Chapter 2

26

cultural perceptions—expanded over time. Really large communities of non-kin emerged only in the aftermath of the Agricultural Revolution with the creation of states. 60 Genes and the Environment It needs to be made clear that sociobiologists do not claim to have found a gene for war or aggressiveness. This is a likely impossibility. As one primatologist has put it, “In reality, each gene acts in conjunction with hundreds of others. So every behavior is likely to depend on a wide range of genetic factors.” 61 And as Wilson notes, to the degree that human aggression is seen as genetic, it is only “in the sense that its components have a high degree of heritability and are therefore subject to continuing evolution” and in the sense that aggressive responses of some species are “specialized, stereotyped, and highly predictable in the presence of certain very general stimuli.” 62 Having said that, it is also true that aggressive behavior can be (and has been) selectively bred in animals over several generations; one only has to think of dogs and cocks bred for fighting. 63 Even though aggression is seen as genetic, Wilson is reluctant to define aggression as an instinct. Neither does he see it as an inborn drive that builds up pressure until it bursts the dam of inhibition. Wilson maintains that there is no general instinct for aggression, only particular patterns of aggressive behavior that different species have found to be adaptive for their environment. Aggression is sort of like a “genetic contingency plan—a set of complex responses of the organism’s endocrine and nervous systems, programmed to be summoned up in time of stress.” 64 Out of a large variety of possible violent behaviors, humans exhibit only a small range. Particular forms of organized warfare are not inherited; no genes differentiate between head hunting or cannibalism, between dueling or genocide. Humans inherit a wide variety of possible behaviors. Which behavior particular human beings display depends on environmental and cultural differences. Each culture gives a specific form to aggression. In turn, the cultural evolution of aggression would seem to be guided by (1) the genetic predisposition toward learning some form of communal aggression, (2) the necessities imposed by the environment, and (3) the previous history of the group, which biases it toward the adoption of one cultural innovation as opposed to another. 65 Biology is responsible for the initial evolution of organized aggression, but whether this behavior is maintained will be determined more by a cultural process under the control of rational thought. As Gat puts it, aggression is “both innate and optional,” though close to the surface and easily triggered. 66 In other words, even though warfare may have a genetic basis, the evolution of warfare can be reversed—an example being the now peaceful but previously warlike Maoris of New Zealand. 67 Nevertheless, as far as Wilson is concerned, humans are “predisposed to slide into deep, irrational hostility under certain definable conditions.” 68 The bottom line seems to be the following: Our brains do appear to be programmed to the following extent: we are inclined to partition other people into friends and aliens. . . . we tend to fear deeply the actions of strangers and to solve conflict by aggression. These learning rules are most likely to have evolved during the past hundreds of thousands of years of human evolution and, thus, to have conferred a biological advantage on those who conformed to them with the greatest fidelity. 69

The Individual Level of Analysis, Part I

27

Critiques of Sociobiology Biologists, geneticists, and some ethologists have criticized sociobiology for an overemphasis on the genetic basis of human behavior. 70 Here we will focus on the criticisms of sociobiology generated by anthropologists, and thus our discussion of the critiques of that discipline will serve as an introduction to the nurture side of the debate about the causes of human aggression. Wilson says much with which anthropologists can agree; indeed, there seems to be substantial common ground. Most anthropologists agree that much of human behavior has a genetic basis but this is different from saying, as Wilson does, that such behavior is genetically determined. Although Wilson recognizes in turn the importance of culture, environment, and learning in explaining aggression, he generally tends to give more weight to the genetic influences than anthropologists believe they merit. Critics also attack the historical basis of Wilson’s reasoning. Wilson argues that the “tendency under certain circumstances to indulge in warfare against competing groups may well be in our genes, having been advantageous to our Neolithic ancestors.” 71 Montagu disputes Wilson’s implication that warfare first appeared among Neolithic communities, saying there is no unambiguous evidence of such behavior. Additionally, many hunter-gatherer people never achieved a Neolithic stage of development, nor did they develop warfare. Is it more reasonable to assume that they did not inherit the aggressive genes Wilson talks about, or is it more logical that warfare was absent from these groups for reasons having to do with culture and environment rather than genetics? Montagu opts for the second explanation; hunter-gatherer peoples did not engage in warfare because they had no environmental or cultural reason to. Alternatively, if Wilson is correct and evidence of warfare does first appear among Neolithic communities, then one ought to examine the particular social and environmental conditions of such communities that led to the arrival of warfare. 72 Critics contend that Wilson has failed to build a convincing case for his general proposition that social behavior is determined to a large extent by heredity. His evidence for this is that humans share several similar modes of social behavior with their nearest animal relatives: the size of adult groupings is in the range of ten to one hundred, not smaller; there is a relatively long period of social training of the young; they share the institution of play; and so on. Likewise, all human groups share similar social behaviors, such as athletic sports, division of labor, education, funeral rites, gift giving, marriage, status differentiation, and so forth. The reaction of anthropologists to this is simply that common environmental influences on human societies are likely to produce common forms of social behavior. Faced with similar problems and similar tasks, various human groups throughout the world constructed similar institutions to resolve these problems and solve these tasks. Since Wilson can provide no genetic evidence for the claims he makes, the claims of the anthropologists are just as likely to be true. 73 Along these same lines, the evolutionary approach of the sociobiologists tends to explain social behavioral traits as the product of evolutionary adaptation. Human aggression and warfare exist because they were once adaptive. Critics contend that evolutionary theorists have fallen into a kind of “functionalist trap”: If the behavior in question is functional, that is, if it can be interpreted as aiding genetic fitness, then it must be due to inclusive fitness. 74 There is no logical reason that this must be true.

28

Chapter 2

Finally, there is a level of analysis problem. Sociobiologists have not put forward specific empirical links between individual genetic predispositions (like aggression or xenophobia or altruism) and group behaviors (like interstate warfare). Therefore, “importing sociobiology into international relations theory does not help illuminate the fundamental issue of international politics—the behavior and interactions of human aggregates, whether they be states, corporations, international organizations, or ethnic groups.” 75 Some Biological Bases for War Before we move on, we should briefly mention some biological attributes of human beings that are related to their capacity for aggression and war. No one would deny that human violence has biological components, several of which are worth thinking about. (There are many other biological factors, but this short list will give you a taste of the possibilities.) First, natural selection has endowed humans with hands and shoulders ideally suited for combat. While chimpanzees have a finger-thumb arrangement ideal for grabbing branches and swinging through trees, if the ability to grip a club or to throw things (rocks, spears, hand grenades) is what you want, then the opposable thumb arrangement and the human shoulder socket is what you’re looking for. 76 Second, humans inherited a binocular vision from our primate predecessors that allows a broad horizontal field of view as well as precise depth perception that permits us to locate a target and estimate its distance from us. As one anthropologist notes, “Such visual acuity as our ancestors had, paired with a projectile weapon, facilitate accurate targeting. Eye-hand coordination and an accompanying weapon is a lethal combination.” 77 Third, the development of the cerebral cortex and an enlarged brain has interesting implications for violence and war. The human brain empowers humans to conceptualize and create artificial realities. The logical consequence of this is that real threats are not required for human aggression and violence to take place, only the conceptualization that something might threaten us. Humans are the only animal (as far as we know) who fights for concepts: liberty, justice, god, revenge, ideology, credibility, or “making the world safe for democracy.” And, of course, humans’ creative intelligence also gives them the ability to devise sophisticated weapons of mass destruction. 78 Fourth, the amygdala, a neural structure in the human brain, triggers a “fight-or-flight” response to threats, danger, and other stresses. The amygdala works with the hypothalamus, the pituitary and adrenal glands to secrete stress hormones into the bloodstream—cortisol, glucocorticoids, and adrenaline. Thus, the human body gets an “adrenaline rush” from a rapid increase of sugar and adrenaline, providing the muscles with an energy boost and at the same time decreasing the ability of the body to recognize pain—all useful mechanisms for combat or flight. 79 Fifth, the dorsal striatum, in the subcortical region of the brain, apparently is activated when we block threats or when we injure or kill an attacker. In effect, it elicits a neuropsychological response of satisfaction in humans when these acts are committed. Otterbein argues that the activation of the dorsal striatum does not create war or homicide, but it does help to explain why armies always have willing recruits, why combatants will engage in risky actions,

The Individual Level of Analysis, Part I

29

why some people enjoy killing, why we want to take revenge, and why we enjoy bullies getting what they deserve. 80 Sixth, there is the issue of serotonin. 81 Serotonin is a neurotransmitter, a messenger molecule used by brain cells in all species to carry signals. Numerous studies now show that a biochemical anomaly in the human brain (as well as in monkeys) leads to a particular kind of violent behavior. A deficient gene (a mutation on the X chromosome) affects the production of the enzyme monoamine oxidase A (MAOA), which in turn effects the breakdown of serotonin. 82 The mutation results in low levels of serotonin, which leads to an increase in violent, impulsive behavior. Men convicted of premeditated murder have normal levels of serotonin. (Essentially, serotonin comes into play during “fight-or-flight” situations.) A widely published study showed that numerous males in a Dutch family—going back well over 120 years—exhibited aberrant behavior. They became enraged, aggressive, and violent without provocation. They were quick to pick fights; many were involved in acts of exhibitionism, arson, and attempted rape. Essentially, they were unable to restrain impulsive behavior. Why? DNA analysis indicated that several generations of males in the family had received this deficient gene. (The women were not affected; since women have two X chromosomes, a mutation on a single X chromosome will not lead to a problem if the second chromosome carries a normal copy of the gene.) What does all this boil down to? The genetic component of at least one type of violence is clear. Impulsive violence is directly linked to a gene that controls a chemical in the brain. This is not the sociobiologist’s holy grail of a single gene for aggression, but it is the most direct link scientists have found thus far. (By the way, a strain of aggressive rhesus monkeys with low serotonin has in fact been selectively bred.) Finally, we would be remiss if we failed to acknowledge that the biggest genetic and biological marker for violence and war is gender. Of all the behaviors studied by social scientists, aggression is the one that shows the most consistent gender differences. Most acts of violence and violent crime are committed by males. And this appears to be universally true across time and cultures. 83 (It’s also true of chimpanzee societies.) As two prominent scholars of violence and sex note, “there is no known human society in which the level of lethal violence among women even begins to approach that among men.” 84 To many observers, this commonality across cultures suggests biological causes rather than cultural ones. 85 While it is true that many gender roles are socially constructed and that this (in part) explains the differences in male-female aggression and violence, most evolutionary biologists believe that there are profound differences between the sexes that are due to genetics rather than culture. 86 This takes us to the largely misunderstood issue of testosterone. Though the average adult male has about twenty times as much testosterone as the average woman, the difference in behavior noted above is probably not due to testosterone levels. Only a small amount of testosterone is necessary for male aggressiveness, but higher levels than normal do not affect the frequency of aggression. The presence of testosterone has only a “permissive” effect. Testosterone levels may affect the intensity of aggression, but only minimally, and it does not cause aggression. 87 As Joshua Goldstein notes in his War and Gender, We would be lucky to find that war and sexism were biologically determined, 100 percent. We could find the hormone or neurotransmitter that inhibits these unfortunate behaviors, then add it to

Chapter 2

30

the water supply like fluoride. (Instant peace. Just add water.) Unfortunately, real biology is a lot more complicated and less deterministic. 88

Hormones, brains, the environment, and human behavior interact in complex patterns that are not entirely understood, though the relationships are clearly reciprocal. Goldstein ultimately concludes that “the genetic basis of aggression is neither very specific (a ‘gene for war’) nor very closely related to gender.” 89 Ethology, Sociobiology, and War? One general comment would seem to be in order now concerning the theories of human aggression discussed above. If war is seen as being derived from an innate aggressiveness that is a part of the nature of humankind, then warfare should be a relatively constant state of affairs. Yet we know that war and aggression are not constant in time or space. Why are some states peaceful? And why are states that are peaceful at some times aggressive at others? Theories that focus on the general aggressive nature of humans are unable to provide the answers to these questions. To the extent that ethology, sociobiology, and evolutionary theory accurately explain human aggression, and this is far from clear, they are still inadequate to the needs of an empirical theory of interstate war. Since they are unable to address the variability of war, they are essentially dead ends. A look at the nurture side of the debate should provide greater insight regarding the variation in human aggression. NURTURE: THE BASICS Different historical periods have witnessed different amounts of war, and within these periods different states have had varying war experiences. Not only has the frequency of war varied, but the goals of war, the rules for conducting warfare, and the reasons and justifications for it have varied from time to time and place to place. 90 All of this suggests that war is not as culturally acceptable in some times and places as in others. War exists within a general political culture, which is an important factor in determining whether conflicts are pursued through warfare or through less violent methods. The flip side of the nature argument is therefore the argument of “nurture”: aggression is culturally constructed, not biologically determined. Or as Bronislaw Malinowski said many decades ago, “Human beings fight not because they are biologically impelled but because they are culturally induced.” 91 People learn aggression from their cultural environment. Just as aggression is learned, peaceful, cooperative attitudes toward conflict resolution may also be learned. Aggression is not universal; rather, it is culturally specific and culturally derived. Those on the nurture side of the dispute, primarily behavioral psychologists, ethnographers, and anthropologists, argue that since people vary greatly in their behavior of aggression, different cultures are a good place to look for an explanation of the differences. There are six types of evidence for the argument that warfare is primarily a learned behavior. (1) In the past, hunter-gatherer societies were largely peaceful up until the Agricultural Revolution, making warfare a relatively new phenomenon. (2) Peaceful societies exist today, further demonstrating the absence of innate aggression and dispelling the myth that all

The Individual Level of Analysis, Part I

31

people are aggressive. (3) Globally, the degree of violence and aggression varies tremendously from culture to culture and from state to state, and this must be due to cultural and environmental factors rather than biological ones. (4) Even in states with violent cultures, some individuals are peaceful. (5) Studies of primate behavior show the importance of culture and learning on behavior. (6) The experimental evidence is fairly clear that aggression is greatly influenced by learning. Aggression can be taught; it can also be modified, reduced, and even eliminated by learning. We’ll explore several of these arguments below. NURTURE: CULTURAL EVOLUTION Anthropologists frequently argue that early humans were peaceful animals with a nonaggressive nature. Indeed, the primacy of cooperative behavior was essential to the social processes of hunting and gathering, to the development of tools, and even to the invention of speech. Scholars on the nurture side of the debate generally argue that there is little evidence of either intragroup or intergroup hostilities in early humans before the development of agriculturalpastoral communities. Such aggressive behavior would have endangered the whole population and would—in Wilson’s terms—not have promoted adaptiveness. 92 For anthropologists, an important key to aggression is the fundamental change in the social and cultural environment that confronted humans as we progressed from the nomadic huntergatherer stage of development to that of settled agricultural or pastoral existence. Box 2.1, Archaeological Timeline, provides a handy reference to the students unfamiliar with the time periods and terminology attached to them. In many ways the Agricultural Revolution (also known as the Neolithic Revolution) constituted a fundamental tipping point in the history of human violence. In agricultural or herding societies, land became a valuable possession; for the first time land was owned by individuals or groups and it required protection from other individuals or groups. For instance, paleoanthropologist Richard Leakey argues that as soon as people commit themselves to agricultural food production as against nomadic food gathering . . . they commit themselves to defending the land they farm. To run away in the face of hostility is to face certain loss: a year’s labour may be invested in the fields, and that cannot be given up easily. . . . As well as land that requires defending, agriculturalists tend to acquire property, both personal and communal, that needs to be guarded. 93

Leakey therefore suggests that the Agricultural Revolution constituted a major social, cultural, economic, and political change, which was followed by a substantial increase in military encounters between neighboring groups. The Agricultural Revolution led directly to the creation of towns, villages, and cities. These new communities depended on new forms of social organization quite different from those of nomadic bands. Principally, they required the creation of large-scale enterprises (construction of temples, palaces, walls, irrigation ditches, and canals), which in turn required centralized control. As towns and cities developed more centralized political structures to organize and control essential community activities, they also acquired the organizational ability to create and maintain large armies. Leakey thus sees

Chapter 2

32

warfare as a response to the changing economic and social circumstances in which early humans found themselves after the Agricultural Revolution. 94 Box 2.1. Archaeological Timeline Paleolithic Age (Old Stone Age) 2.5 million–10,000 BP • Lower Paleolithic (2.5 million BP–100,000 BP): era of Homo habilis, first use of stone tools • Middle Paleolithic (200,000–30,000 BP): hunter-gatherer societies, appearance of Neanderthals • Upper (Late) Paleolithic (30,000–10,000 BP): age of fully modern humans, Homo sapiens sapiens; hunter-gatherers and start of complex hunter-gatherers; first appearance of fire and bow; cave art in Europe Mesolithic Age (Middle Stone Age): 12,500–10,000 BP • • • •

Hunter-gatherers and complex hunter-gatherers Development of technology by humans Domestication of dogs Age ends with introduction of agriculture Neolithic Age (New Stone Age): begins circa 10,000–8,000 BCE

• • • •

Domestication of plants and beginning of agriculture Animal husbandry Creation of villages and towns Creation of states within local state systems–3,000 BCE

The Neolithic Age ends with the beginning of metal tools. Bronze and Iron Ages follow. Source: Paul Bahn, ed. The Penguin Archaeology Guide (London: Penguin Books, 2001). Notes: All dates are approximate. Ages for different locations will differ. BP refers to before present era; more specifically, before the use of radiocarbon dating in 1950. BCE refers to before the common era. In his Parable of the Tribes, Andrew Bard Schmookler adds another piece to the puzzle of how the Agricultural Revolution led to increased warfare. 95 Agricultural societies eventually encountered limits to their growth posed by the existence of other communities. This typical Malthusian problem that agricultural communities faced could be resolved either by a more intensive use of the land or by expanding one’s current farming and grazing areas at the expense of one’s neighbors through the use of force. With no overarching power to prevent it (that is, under the condition of anarchy), the more highly organized cities pursued the latter

The Individual Level of Analysis, Part I

33

course. In response, peaceful societies had essentially four possible options: destruction by their neighbors, subjugation, withdrawal through migration, and imitation. Imitation proved to be the choice of most communities. In order to survive, agrarian societies were compelled to emulate their most aggressive rivals. 96 They built large communities through consolidation and aggregation; they constructed large-scale political organizations in order to efficiently mobilize their populations; they initiated taxation systems to make the wealth of society available to these governments; and they created military institutions to protect and extend their power. In reality, certain avenues of cultural evolution were closed off. More highly organized societies drove out the less highly organized; large societies drove out the small; the more warlike cultures drove out the more peaceful. Social evolution proceeded in only one direction—toward the creation of ever more powerful and militarized societies. A process of cultural selection seems to have been at work here, and the result was that powerful, militaristic political units spread throughout the world, and wars between these societies became endemic. Here, Schmookler draws on the work of Robert Carneiro, who argues that until late in the Neolithic period (i.e., prior to the Agricultural Revolution), the effect of armed conflict was to drive defeated groups away, thus keeping villages separate. Warfare’s effect was dispersive. After the Agricultural Revolution, the effect of war was aggregative; it led to the building of larger and larger political units. 97 As you might expect, there is room for dissent within the anthropologist camp. Keith Otterbein, one of the foremost anthropologists of war, has argued that, historically, the cultural evolution of war has probably developed along two different paths. 98 The first pathway is the one just described (more or less): early foraging/gathering groups became sedentary agriculturalists. These were initially peaceful cultures because, in his view, domestication of plants cannot exist in a region where there is warfare. Thus, intergroup violence was absent until more complex states developed thousands of years later, at which point warfare appeared on the scene. Otterbein argues that it is unlikely that there was warfare between early farming villages; further, it was unlikely that victorious villages incorporated losing villages into an ever expanding political unit or state. In his view there is no archaeological evidence for this and it is logically flawed: villages did not have the political structure to accommodate incorporation. There was no government to actually carry out the joining of multiple villages. Warring villages did not aggregate; they destroyed each other. Plus, they lacked the “idea” of a state. 99 However, warfare most probably occurred much earlier as well, under different circumstances. For Otterbein, large game hunting is the other path of cultural development that has led to warfare. He believes that groups of early large game hunters in the Upper Paleolithic period sometimes attacked competing groups of hunters, using modes of fighting based on ambushes and line attacks. (Cave paintings from this period provide some evidence for this.) Otterbein identifies three circumstances that would have given rise to warfare: (1) hunting, especially large-game hunting, which requires cooperation among many individuals; (2) weapons (the spear thrower, or atlatl, was developed over 40,000 years ago); and (3) the existence of fraternal interest groups composed of related males. In uncentralized societies with fraternal interest groups, these men are essentially the military organization of the society. And Otterbein argues that uncentralized societies of this type (which he calls Type A societies) are characterized by a high degree of internal violence, feuding, and warfare—

34

Chapter 2

which is largely accounted for by competition between these separate groups of related males. Indeed, Otterbein believes that the presence of fraternal interest groups is the best predictor of both feuding and warfare between culturally similar groups. 100 Cross-cultural studies also show a positive relationship between warfare and societies who subsist by hunting large game. 101 Otterbein believes that as large-game hunting increased, so did warfare. And when hunting decreased, as when mammoths and other large game were killed to extinction in North America, warfare declined. Note that both paths to warfare depend on cultural factors: social and political organizations developed by groups in response to environmental conditions at certain times and places. They have very little to do with human nature, biology, or genetics. Military historian Azar Gat provides another version of how warfare developed. He argues that, at least initially, the primary threat to agricultural communities was from nomadic huntergatherers. In this contest farmers had the advantage of numbers—their birth rates were roughly double that of hunter-gatherers, and they lived in larger groups. However, agriculturalists were stationary and their crops and animals were thus vulnerable; moreover, mobile huntergatherers were harder to retaliate against and had less to lose. Hunter-gatherer raids were mostly small-scale, hit-and-run acts of violent thievery. The result was probably that the first fortifications by sedentary agriculturalists were most likely built as protection against raids by hunter-gatherers rather than against rival agricultural communities. Jericho, an ancient town dating to the late ninth century BCE, is thought by some to be the first example of this. (Sedentary hunter-gatherers also built fortifications to protect their settlements.) As the number of hunter-gatherer groups gradually declined, conflict between agriculturalists increasingly took its place. 102 The primary reason for the expansion of farming communities was not land accumulated through armed aggression but simply the “demographic and ecological facts that colonizers created on the ground over generations and centuries” in a protracted process. 103 NURTURE: PEACEFUL SOCIETIES Despite the general historical trend toward warfare at the time of the Agricultural Revolution, peaceful societies have not only existed in the distant past but survived into more recent times. A great many of these are hunter-gatherer societies. Observation of these communities may provide some insight into the roots of aggression, largely because these are the contemporary groups living closest to conditions that early human communities faced. David Fabbro’s classic study of peaceful societies focused on societies that were deemed to be peaceful due to (a) the absence of wars on their territory, (b) the absence of external war involvement by the group, (c) the absence of civil war or internal collective violence, (d) the lack of a standing military-political organization, and (e) little or no interpersonal violence in the society. 104 Fabbro investigated seven societies that met these criteria: the Semai of Malaysia, the Siriono of Bolivia, the Mbuti pygmies of Zaire, the !Kung San Bushmen of the Kalahari, the Copper Eskimoes of northern Canada, the Hutterites of North America, and the islanders of Tristan da Cunha in the South Pacific. (Other peaceful societies might include the Zuni of the American Southwest, the Arapesh and the Fore of New Guinea, the Walbiri aborigines of Australia, the Tasaday of the Philippines, the Tahitians, the Lepchas of Sikkim, and many more.)

The Individual Level of Analysis, Part I

35

What are these peaceful societies like? The seven peaceful societies examined by Fabbro could all be classified as “egalitarian band societies,” which generally lack patterns of ranking and stratification, place no restrictions on the number of people capable of exercising power or occupying positions of prestige, and have economies where exchange is based on generalized reciprocity. 105 All are small, face-to-face communities—a major factor contributing to their open and egalitarian decision-making process. Though the first five are hunting and gathering societies, the last two have some agricultural base. However, they all produce little or no surplus, and what is produced is distributed equally. The lack of an economic surplus (that is, the production of economic goods beyond what is needed for subsistence) would seem to be important. If there is no surplus, the political authorities cannot commandeer it and use the wealth derived from it as a basis for carrying on coercive activities, including the creation of a military. 106 Fabbro concludes that peaceful societies are peaceful essentially because they lack some of the most important structural prerequisites for engaging in war: a coercive hierarchy and leadership and an economic surplus to support a nonproductive military organization. 107 It is important to notice that the chronic scarcity of resources, with which most of these societies are faced, is not a factor that contributes to violence; quite the contrary, it is a factor that encourages close cooperation. Many of these societies also develop cultural norms that discourage violence. The !Kung disparage physical combat as a means of settling disputes. Instead, the most admired characters in the !Kung folklore are those who deal with adversity through trickery and deception rather than through force. 108 It should be mentioned, of course, that not all primitive societies have been nonviolent. The variability in violence among traditional societies is fairly large; violence and even warfare do exist. But the central point, according to Gwynne Dyer, is that precivilized societies did not kill people much. Dyer notes that, of the hundreds of hunter-gatherer societies that modern people have encountered, almost all have had the same attitude toward war: “it is an important ritual, an exciting and dangerous game, and perhaps even an opportunity for self-expression, but it is not about power in any recognizable modern sense of the word, and it most certainly is not about slaughter.” 109 Neither is it about the conquest of territory. Dyer contends that there is scarcely one recorded example of such tribes participating in a death struggle with its neighbors because of population pressure or economic scarcity. “Although many were engaged in low-level warfare against their neighbors in their spare time, . . . nobody thought ‘winning’ was sufficiently important to put much thought into organizing warfare efficiently.” 110 This low-level tribal warfare was limited in nature and highly ritualized. The American Plains Indians’ practice of “counting coup”—whereby the adversary was not killed, but simply touched with a stick or hand—is a standard example. The fighting often stopped for the day after a single casualty was exacted, and there were deliberate steps taken to limit the destructiveness of the warfare. Individuals got killed, though only a few at a time, and the societies survived intact. 111 Dyer concludes that precivilized warfare was mostly a “rough male sport for underemployed hunters, with the kinds of damage-limiting rules that all competitive sports have.” On the other hand, once these people have “progressed” toward agriculture and herding, the

36

Chapter 2

warriors have more free time and they begin to acquire material interests to defend; the outcome is that war becomes more destructive. 112 Quincy Wright’s analysis of 633 primitive cultures offers supportive evidence that the collectors, lower hunters, and low-level farmers were the least warlike of these primitive peoples while the most advanced herding and farming societies were the most warlike. 113 The conclusion of many anthropologists, ethnographers, and historians is, therefore, that a tremendous increase in violence accompanied the transition of societies from the huntergatherer life to the more settled, agricultural existence and the concomitant rise of towns and cities. Changes in the social, cultural, economic, and political environment associated with the Agricultural Revolution brought about significant changes in the behavior of many human communities. The level of violence and warfare increased substantially. Nevertheless, it is clear that there are many nonaggressive individuals within these generally aggressive, modern communities. Most societies are filled with people who would find it very difficult to take another’s life, even in anger. Even if aggression is part of the human genetic makeup, it would not seem to be enough to cause most people to be able to kill another human being. Werner Levi, in his much quoted analysis of the causes of war, notes that there never seem to be enough “aggressive” men flocking to the recruiting stations during war, so that everywhere men are drafted to perform such services. Once drafted into the military, they need heavy doses of indoctrination to turn them into killers. It takes a great deal of conditioning to prepare them for face-to-face combat. Even so, in some armies more than half of the men who were supposed to fight did not pull the trigger. 114 They were willing to die for their countries, but they were not willing to kill for them. Cultural and environmental factors would seem to have been at least somewhat successful in suppressing whatever genetic impulses humans have toward aggression. Debating Peaceful Societies Past and Present No one’s argument gets a free pass. (After all, that’s what science is all about.) The view of primitive hunter-gatherers as peaceful people has been vigorously challenged by a new generation of anthropologists and the debate has intensified. According to Lawrence Keeley, early anthropologists and ethnographers may have gotten things seriously wrong. They tended to interview peoples who had already been settled and pacified by government authorities and who relied on their somewhat idealized memories of the past when they talked about warfare. Some anthropologists also interpreted these accounts through Western lenses based on romanticized versions of Rousseau’s “noble savage.” They tended to see instances of warfare among primitive peoples as ritualized (which they were) and not very risky or deadly (which they were not). They overlooked the violence by contemporary “peaceful” peoples, and they “pacified the past” by ignoring the evidence of stone-age warfare. 115 Keeley and other critics of “peaceful societies” make several telling arguments. 1. Peaceful societies may exist, but they are far less numerous than previously believed. Keeley reexamines the cross-cultural anthropological record of premodern societies and concludes that the claims concerning present-day peaceful societies are generally unfounded. Numerous surveys of current or recent tribal societies—foraging, pastoral, or agricultural communities—consistently indicate that between 87 and 90 percent are warlike. 116 Melvin

The Individual Level of Analysis, Part I

37

and Carol Ember, who compile and analyze the Human Resources Area Files (HRAF), a database of preagricultural and agrarian cultures across the globe, conclude that the vast majority had experienced some warfare unless the society had been pacified by an external power. 117 The peaceful exceptions—like the Eskimos of northwestern Greenland, the Mbuti Pygmies of central Africa, and the Semang of Malaysia—are either small, nomadic huntergatherers, isolated agriculturalists, or defeated refugees living under the protection of a modern state. Many are just barely viable economically, living in harsh environments and having little contact with other peoples. At least one observer notes with irony that extreme environments are necessary to sustain such peaceful societies. 118 Azar Gat has analyzed the ethnographic record for one very large group of hunter-gatherers from the recent past—the roughly 300,000 Aboriginal peoples who were the sole inhabitants of Australia and Tasmania until the European settlements began at the end of the eighteenth century. 119 All of these people were members of simple hunter-gatherer bands. There were no agriculturalists or pastoralists on the continent. What do we know about the amount of violence in these societies? In terms of technology, they lacked bows and arrows but had developed spears, clubs, stone knives, wooden shields, and of course the boomerang. Aboriginal rock art, dating to 8,000 BCE, depicts fighting scenes. Throughout the nineteenth and early twentieth centuries, as Europeans came into contact with these people, they observed their societies and interviewed their members, developing an extensive ethnographic record. That record uniformly indicates evidence of territoriality, lethal raiding practices, and ubiquitous violence. Indeed, violent deaths appeared to be the primary cause of mortality for these groups. The same pattern exists for societies of complex hunter-gatherers on the northwest coast of America. These societies were more sedentary groups who preserved and stored food, had higher population density and extended family groupings, and accumulated considerable private property. They might be considered “collectors” rather than foragers. There is evidence of large-scale cemeteries, fortifications, burned-out settlements, and art depicting warfare. As in Australia, fighting appeared to be ubiquitous. Gat concludes that these historical laboratories “clearly show, across a variety of native peoples . . . that hunter-gatherers, from the very simple to the more complex, fought among themselves. Deadly conflict, if not endemic, was ever to be expected . . . Killing in fighting was among the main causes of mortality.” 120 Warfare cannot be seen as a late cultural invention dating only to the Agricultural Revolution; interclan and intertribal fighting predated the Agricultural Revolution and the rise of states. 2. Many societies previously regarded as exemplars of peaceable communities are now regarded as having a more violent existence. For instance, the !Kung San of southwest Africa were actually quite warlike in their relations with neighboring peoples prior to European colonization. They also participated in the civil wars that hit Angola and Namibia in the 1970s and 1980s. The Semai of Malaysia became involved in the brutal anticommunist insurgencies in Malaysia in the 1950s, though they returned to their peaceful ways after the war. 121 3. The seeming peacefulness of some isolated hunter-gatherer groups is merely an artifact of the kind of fighting that they engage in rather than any true pacifism. These communities typically engage in small-scale feuds, vendettas, murders, ambushes, and raids—social phenomena small enough that they do not warrant inclusion in the modern definition of war.

38

Chapter 2

Critics argue that the concept of war should not be confined to armed conflicts between modern states. Throughout human history, warfare has taken place between political units of lesser scale: from band societies, to autonomous villages, to tribes, to chiefdoms. In many cases these are not politically unified societies, and therefore while armed combat may have been carried out against individuals or groups in other societies, “it was usually not organized on behalf of an entire society or even a major section thereof.” 122 However, while the societies did not engage in war in the modern sense, they certainly were not peaceful. Moreover, the impact of violence on hunter-gatherer societies is similar in scale to modern societies. While the total number of deaths may not be high by “civilized” standards, the frequency of war in primitive societies is about the same as in modern societies, and so is the percent of the population mobilized for warfare. 123 The level of violence per capita in primitive societies may actually be higher than in modern societies, depending upon the measurement used. Fighting between some preindustrial societies in New Guinea and the Amazon typically result in the loss of between 20 and 30 percent of the communities’ males. 124 Some primitive peoples—like the !Kung San—have homicide rates much greater than that of the United States, and in one village of Copper Eskimos, every adult male had been involved in an incident of homicide. 125 The world record is apparently held by the Waorani of the Ecuadorian Amazon, with more than 60 percent of all adult deaths over five generations due to feuding and warfare. 126 Thus, it would be simply wrong to see contemporary hunter-gatherer band societies as inherently peaceful. Keeley concludes that it is certainly true that pacifist societies seem to have existed at every level of social organization, but they are extremely rare and seem to require special circumstances . . . the idea that violent conflicts between groups is an inevitable consequence of being human or social life itself is simply wrong. Still the overwhelming majority of known societies have made war. Therefore, while it is not inevitable, war is universally common and usual. 127

Gat believes that the anthropological and ethnographic records of primitive peoples (huntergatherers and primitive horticulturalists) around the globe show a remarkably consistent and uniform pattern. Some aspects of the pattern are similar to intraspecific aggression that takes place between primates from neighboring communities (remember the chimpanzees). As in many animal communities in which intraspecific aggression is primarily at the expense of the young and the defenseless, among primitive hunter-gatherer groups and primitive agriculturalists fighting appears to be overwhelmingly asymmetrical. The most serious killing targets victims who can be caught in a helpless condition, with the attackers avoiding risks. Most killings are the result of raiding and ambush tactics in which the victims are surprised (usually at night or just before dawn) and caught relatively defenseless and incapable of harming the attackers. 128 In some of these surprise raids, the target groups were completely annihilated. Typical raiding behavior results in extremely high mortality rates—in the neighborhood of 15 percent of adults and 25 percent of adult males. 129 On the other hand, pitched battles between large groups of males are relatively rare and are generally ritualized, though rituals sometimes escalate into serious fighting. The pattern is found in Aboriginal Australia, Northwest American Coast Indians, Alaskan Eskimos, American Great Plains Indians, the Papuan New Guinea Highlands, and among the Yanomano of the Amazon rainforest.

The Individual Level of Analysis, Part I

39

As we mentioned previously, Ember and Ember’s research on premodern societies indicates that the vast majority experienced some kind of warfare between communities or larger units, though the amount of warfare varied substantially. Their cross-cultural study of war (defined as “socially organized armed combat between members of different territorial units”) reveals several interesting regularities. 130 They believe that the frequency of warfare can be accounted for by two things. First, nonchronic resource problems created by unpredictable natural disasters are a major factor. Even more important is the existence of the threat of such unexpected scarcity in the future. Second, the socialization of mistrust in a society is related to frequent warfare. This is probably connected to the first factor. A history of unpredictable natural disasters may predispose people to be less trusting, and they socialize their children accordingly. Thus an interrelated fear of nature, fear of the future, and fear of others might be the basis for intercommunal warfare. Ember and Ember conclude: Thus, it looks as if the people who go to war more or less constantly might be trying not to cover present or regularly recurring shortages, but to protect themselves against future disaster which they cannot predict; that is they seem to be trying to protect themselves ahead of time by taking resources from enemies. 131

4. Ancient Neolithic peoples were probably more warlike than depicted by supporters of the “peaceful society consensus.” With regard to ancient hunter-gatherer peoples prior to the Agricultural Revolution, the physical evidence is rather spotty. There is no written record and there are very few burial sites to investigate. Keeley argues, however, that the limited burial samples available indicate the existence of homicidal violence dating from the Late (Upper) Paleolithic Era—well before the Agricultural Revolution. For instance, the now famous burial sites in Gebel Sahaba in Nubia (present-day Sudan) overlooking the Nile and dating from 10,000 to 13,500 BCE show evidence of warfare: Over 40 percent of the skeletons of men, women, and children had imbedded chert projectile points and many were multiple wounds. 132 Bodies found in cemeteries in Europe and the Nile Valley from pre-Neolithic times also show evidence of violent death associated with projectile points or cranial trauma. They can’t all be hunting accidents. Archaeological evidence also indicates that fortifications were used by even small tribes and chiefdoms at the dawn of the agricultural age. Long before states were organized, the earliest villages were fortified by palisades and ditches. These were originally explained in nonmilitary terms, but the presence of projectile points in the fortifications suggests warfare as the major motivation for the structures. According to Keeley, “warfare is documented in the archaeological record of the past 10,000 years in every well-studied region.” 133 Long before the invention of states, organized collective warfare was carried out by bands, tribes, and chiefdoms. It did not take long for Keeley’s arguments to be countered by a much more nuanced interpretation of the historical and cross-cultural data. In Warless Societies and the Origin of War, Raymond Kelly provides an analysis of the social evolution of warfare that partially rebuilds the case for warless societies. 134 Kelly rightly argues that physical violence takes many forms and that warfare is not simply “more of the same.” It is true that violence exists in Fabbro’s group of peaceful societies; but the presence of homicides (even at fairly high rates)

40

Chapter 2

is not the same as war. Warless societies do exist and are, in fact, “not scarce in the world ethnographic sample”; they’re just not peaceful or nonviolent. 135 Using new sources and a reexamination of Fabbro’s original sources, Kelly (like Keeley) concludes that the so-called peaceful societies are not at all peaceful. Among Fabbro’s seven societies interpersonal violence abounds: “we find instances of harsh physical disciplining of children, violence between spouses, women fighting each other with digging sticks, and men committing homicide.” 136 But this is not war. The definition of war greatly affects the conclusion one reaches about the amount of warfare (both internal and external) found among primitive communities. Kelly defines war as an act of lethal violence that has all of the following characteristics: (1) it is collectively carried out; (2) it is collectively sanctioned by the community; (3) it is morally justified by the participants; (4) its participants are esteemed by members of their community; (5) the acts are organized, planned, and premeditated; (6) the acts serve an instrumental objective; and (7) the principle of social substitutability applies to the targets. 137 The last characteristic is the one that distinguishes a society’s collective punishment of one of its members—including capital punishment—from war and feuds. Unfortunately, many surveys use rather broad definitions of war that include collective acts of capital punishment and even single acts of homicide. Kelly argues that although capital punishment is found in all known human societies, war is not. All of the cross-cultural surveys widely cited in the literature find some societies in which “war” is rare or nonexistent, and that number would be considerably higher if the definition of war was narrowed to eliminate instances of capital punishment. 138 Otterbein also argues that the question of whether hunter-gatherer societies are warlike depends on whether the sample includes primarily hunter groups or gatherer groups. Hunters tend to be more prone to warfare, gatherers not so much. 139 For Kelly, what distinguishes war from other types of collective violence is the presence of the principle of social substitutability: the idea that any member of an enemy group is a legitimate target for revenge or retaliation. Thus, individuals who are innocent of any direct responsibility for previous deaths or crimes may be legitimate targets of retaliation by members of the aggrieved group. This is the logic behind feuds and war, but not capital punishment. 140 War employs lethal violence on behalf of one social group against another. The key to understanding the lack of any real war in Fabbro’s peaceful societies is that they all lack the principle of social substitutability. 141 Social organization makes a difference. Warless societies are “unsegmented” societies. They have a minimum number of basic social groups, such as families and local communities. The groups that do exist are autonomous; they are not subunits of larger organizations, and they are not combined into any higher-level organizations within a hierarchy. Extended family organizations do not exist. There are no fraternal interest groups so crucial to the development of warfare in Otterbein’s theory. Marriages are seen as contracts between individuals rather than between families, unlike the situation in segmented societies where marriage is seen as a group transaction. 142 Analysis of killings in unsegmented societies shows that retaliatory vengeance is directed only at the individual responsible, not at other members of his group. 143 Retaliation for homicidal acts does not lead to vendettas or feuding between groups or clans. On the other hand, the vast

The Individual Level of Analysis, Part I

41

majority of foraging societies that operate on the principle of group liability are segmented organizationally. 144 As it turns out, the majority of Fabbro’s warless societies are unsegmented. Moreover, when Kelly looks at cross-cultural data collected on foraging societies (societies that are largely dependent on hunting, fishing, trapping, and collecting wild plants and animals) he finds a very strong association between unsegmented organization and either the total lack of internal and external warfare or very low levels of warfare. 145 Conversely, over half of segmented foragers in Kelly’s sample experienced annual warfare. 146 Kelly finds that war is rare or nonexistent in five of the twenty-five foraging societies in a global representative sample, and all of these are unsegmented societies. 147 The fact that war is rare or nonexistent in contemporary unsegmented foraging societies suggests that preliterate human communities experienced much less warfare than those in the later period of recorded history. Kelly argues that war and society evolve together and the key transition point has to do with the evolution from unsegmented to segmented social organizations. (This is not unlike Otterbein’s argument, which also identifies particular types of socialpolitical organization as underlying war.) Essentially, war “originates as a transition from one form of retributive collective violence to another, that is, as a transition from capital punishment to blood feud” to war. 148 He believes that all societies of the early Upper Paleolithic (about 35,000 BP) era were unsegmented. In the absence of segmentally organized foragers, warfare should have been quite limited. Thus, the origin of war is relatively recent—after about 10,000 BP with the exception of the Nile Valley, where warfare arrived somewhat earlier. This is consistent with the archaeological record. 149 Prior to 35,000 BP what warfare that did exist would be limited to areas where unsegmented foragers faced “circumscribed environments”—where foraging societies shared with their neighbors an environment with subsistence resources that were dense, diverse, and reliable, thereby supporting greater population densities, but geographically circumscribed so that the inhabitants had no option to leave. Here, population growth or resource depletion might trigger conditions permissive of warfare. This was the situation for the modern Andaman Islanders and in the Nile Valley, where the Jebel Sahaba cemetery shows clear evidence of death from warfare. 150 Finally, we should permit archaeologists to weigh in on the subject. Keeley’s evidence comes primarily from primitive prestate Neolithic peoples; thus he leave open the question of whether people were peaceful before the Neolithic era. What does the physical evidence of the past tell us? Most archaeologists contend that prior to the Neolithic age and the Agricultural Revolution, the physical evidence of warfare shows a high degree of variation between regions, areas, and specific sites—varying from “virtually none to apparent massacres.” 151 Perhaps the most reasoned judgment comes from Jonathan Haas. 152 Haas reviews the archaeological evidence from three successive periods of history: the Paleolithic or “old stone age,” in which stone tools first appeared and in which societies were organized into hunter-gatherer bands; the Mesolithic or “middle stone age,” in which micro stone tools are developed and dogs are domesticated; and the Neolithic or “new stone age,” in which human societies lived in settled villages, developed farming and animal husbandry and pottery, and completed substantial construction projects, as part of the Agricultural Revolution. At the end of this period, humans begin to create the first complex, state-level societies.

42

Chapter 2

Haas’s analysis of the archaeological evidence is that the first signs of organized violence appear among hunter-gatherer bands in the Late Paleolithic era between 20,000 and 30,000 years ago. The evidence, almost entirely in the form of trauma in skeletal remains, has been found in Europe, Egypt, and the Americas. However, in his judgment, the number of cases is small and each case appears to be isolated and not part of a larger pattern, and therefore warfare was probably episodic and infrequent rather than ubiquitous. In the subsequent Mesolithic period (from 10,000 to 20,000 years ago) the evidence of warfare is clearer and more varied. From sites in Europe, Egypt, and the Americas, there is evidence of skeletal trauma from grave sites, as well as some rock art that may depict organized conflict. Nevertheless, the data are not sufficient to determine the intensity of conflict during this era, and the relative infrequency of skeletal remains, coupled with the absence of other kinds of markers (fortifications, defensive locations of communities, burned or pillaged communities, weapons, etc.), renders it impossible to conclude that warfare was endemic or ubiquitous in this preagricultural period. The Gebel Sahaba remains, with their extremely high level of violent death, appear to be an exception to the relatively low levels of traumatic death and injury found elsewhere in the Late Paleolithic and Mesolithic periods. Once again, organized violence appears episodic and ephemeral—in the sense that it does not appear to have had a major impact on the nature of societies themselves. What the evidence does suggest is that the societies involved in violence during this period may have had some degree of residential stability (and hence territoriality). These hunter-gatherer groups were not entirely nomadic. It is only in the Neolithic period that we have overwhelming and compelling evidence of warfare—indeed the evidence is found in the archaeological record of almost every Neolithic culture studied anywhere in the world, and the violence is bloody, intense, and in some cases prolonged. This violence, as we have seen, does not come out of the blue without historical antecedents. There is organized warfare before the Agricultural Revolution and the creation of states. Agriculture was certainly not a necessary condition for warfare, but it provided a much more receptive environment for it. Haas also argues that while this intense phase of violence is first seen with the initial transition to sedentary living and intensive production, it appears aimed at killing people rather than at capturing land or resources. The relatively bloody period is typically followed by more peaceful periods. (Once stable sources of food become available, there is no need for war until population growth and changing environmental parameters create economic stress on communities—leading to a kind of second wave of war to attain land and resources.) Indeed, in any given location, the record shows periods of violence followed by periods of peace in waves. For Haas, even the evidence for violence in the Neolithic period shows that conflict was not omnipresent. The majority of Neolithic sites show no evidence of warfare: they are not located in defense positions, they are not fortified, they show no signs of destruction, and neither their rock art nor their burial sites show evidence of organized violence. Regional surveys indicate long periods in which there are no manifestations of war in the archaeological record. Thus, war appears to be an exceptional circumstance, even in the age of the Agricultu-

The Individual Level of Analysis, Part I

43

ral Revolution. This changes, of course, with the rise of more complex, hierarchical, and centralized polities such as chiefdoms and states. The views presented here from the perspective of anthropologists, ethnographers, and archaeologists indicate a healthy debate about human violence and warfare. They agree that human violence has existed for millennia, but they disagree about the form it has taken, its extensiveness, and its historical origins. They all agree, however, that violence and warfare have varied depending on environmental and cultural circumstances. Violence is, at least in part, a learned activity, and it is subject to modification. Let us now turn to a very general theory of how violence might be learned or perhaps even unlearned. NURTURE: SOCIAL LEARNING THEORY Clearly, violent behavior varies greatly between individuals and groups. What might account for this? One answer is that the differences may be accounted for by different learning experiences. Behavioral psychologists have shown that aggression can be modified through the learning of peaceful or cooperative responses, and experimental research in the laboratory has indicated the power of conditioning to change the behavior of animals. For instance, Scott reports that male mice have been trained to be completely peaceful. 153 It is also apparent that many societies have “learned” to be nonaggressive. Aggression is almost totally absent in some cultures, even as a response to frustration. Even when aggression is present, vastly different patterns exist. (In some societies, such as the Eskimos, there is some individual aggressiveness, but there is no group warfare, while in Pueblo Indian societies individuals are not pugnacious but there is group warfare.) These circumstances tend to indicate that both individual aggressiveness and group aggressiveness must be learned, and learned separately. 154 Albert Bandura, a proponent of social learning theory, maintains that aggression is learned in large part from the social environment. 155 Aggression is greatly influenced by the socialization process that almost all youngsters encounter—in the home, among family members, with peers, in school, and in religious groups—as a natural part of growing up and becoming familiar with societal norms. (There is a sizable amount of information, for instance, that more aggressive individuals come from homes where corporal punishment is used and that criminals have been abused as children.) 156 The form that aggression takes, the situations in which it occurs, its frequency and intensity, and the targets against which it is demonstrated are largely determined by social experience. The socialization process is instrumental in determining the contexts in which aggression is permitted (if any) and the targets (if any) that are permissible for individuals occupying particular roles in society. Aggression, like any other behavior, can be learned through direct experience or by observing the behavior of others—especially role models. Individuals quickly learn to anticipate the probable consequences of various behaviors through personal experience, through observing the actions of others, from communication, and so on. Once a particular behavior (aggressive or nonaggressive) is adopted, it is maintained, modified, or eliminated by positive or negative reinforcement. Human behavior is controlled to a great extent by its consequences: Responses that cause unrewarding or punishing effects are discarded, while those that produce rewarding

44

Chapter 2

outcomes are retained and strengthened. If aggressive responses to environmental stimuli are met with approval by peers and by dominant elders, or if those who use such behavior are met with attention and their desires are fulfilled, then aggression is reinforced. On the other hand, if aggressive tactics are met with rejection, reprimands, a lack of approval or attention, or an inability to attain goals, then aggression as a response to environmental stimuli will be reduced. It might be interesting to spend a moment pondering mass culture in the United States as reflected in films and video games. If films do indeed mirror the predominant cultural attitudes and norms of society, and if children and young adults do indeed learn attitudes and behavioral norms from such films, what does this portend for the United States? In particular, it might be interesting to reflect on the image of heroes in American film. Who are our heroes and why? From John Wayne in the 1950s to Clint Eastwood and Charles Bronson in the 1960s and 1970s, to Sly Stallone and Arnold Schwarzenegger in the 1980s and 1990s, and on to the films of Mel Gibson, Bruce Willis, Wesley Snipes, and Chuck Norris, the male American hero has been a man of violence, a man who settled his disputes outside of the normal legal institutions, a man for whom compromise, diplomacy, mediation, arbitration, and adjudication were farces to be treated with derision or suspicion. The hero was not the man of peace who could settle disputes between his neighbors through logic and persuasion but the man of action who settled disputes through violence and bloody retribution. In such a way do we learn violence from our culture. A complicating factor is that individuals are subject to several learning environments. Each society—whether it is twentieth-century American or nineteenth-century Plains Indian or tenth-century Viking—develops a general culture within which its citizens are socialized and whose cultural norms they absorb. However, most cultures—especially the more complex modern cultures—have subcultures with competing values and norms. Additionally, each individual is confronted with somewhat different learning experiences in the home and in the workplace. The learning experiences at these different levels may be considerably at variance with each other. For instance, it may be that while the general culture condones and rewards violence, the family finds it repugnant and teaches cooperative, nonviolent practices. At another level, the political practitioner of international relations may find a different set of cultural norms at work in the relations among states in the international system. The behaviors for which one is rewarded in the national culture may not be the behaviors for which one is rewarded (or punished) in the international environment. If individual behavior is the product of cultural environment, then the behavior of our rulers is probably the product of several different environments. To take just one example, Russian President Vladimir Putin is the product of several somewhat contradictory environments. In his formative years, he was a law school student in St. Petersburg, and several decades later served on the staff of the liberal mayor of that city, Anatoli Sobchack, a former law professor of his. This environment was at least partially infused with liberal political and economic values and attitudes such as individual freedoms, pluralism, justice, civil rights, and the importance of democratic institutions. On the other hand, Putin spent most of his adult years in an entirely different culture as a midlevel functionary in the Soviet KGB, primarily in East Germany. Here the culture would have been dominated by secrecy, adherence to author-

The Individual Level of Analysis, Part I

45

ity, loyalty to the organization, suspicion of outsiders, security-mindedness, Cold War realpolitik methods, and the need to do “whatever it takes” to achieve the organization’s goals. In summary, social learning theory reminds us of the importance of culture as a source of violence. It admonishes us that if we seek to understand the cause of violence and aggression, we need to understand that individuals (including national leaders) are quite frequently the products of social and cultural environments that condone and even reward aggression and belittle peaceful cooperation. Ultimately, however, the implications of social learning theory for controlling aggression and violence are optimistic. If aggression is learned, it can be unlearned. If violence is based on cultural and environmental factors, these can be changed— albeit slowly. Cultural institutions are man-made and subject to human manipulation over time; they are dynamic rather than static. Just as certain institutions that were once thought to be “natural” in their time have now been largely eliminated in most cultures—slavery, for example—violence may also come to be condemned and eliminated in the future. NURTURE: THREE EXPERIMENTS IN LEARNING PEACE AND WAR As we have seen, each species has its own way of handling conflicts. For instance, bonobos are renowned for using sex as a way of ameliorating and preventing conflict, while chimpanzees are more likely to use violence as a response to conflict. However, within each species there is significant variation between groups, just as there are between different human communities. There are cultures of violence and cultures of peace. Social learning and environmental factors play a powerful role in creating and passing on these different types of behaviors. Three experiments (or examples)—one a classical experiment with primates in a controlled environment, the second a “natural” experiment involving baboons observed by researchers in the field, and the third, a “natural” or historical experiment involving humans—starkly illustrate the role of environment and social learning. In the first experiment, Frans de Waal placed captive juveniles from two different macaque monkey species together for five months in his primate facility at Emory University in Atlanta. One species, rhesus monkeys, are hierarchic, nasty, quarrelsome, intolerant, and violent. The other species, stump tail monkeys, are more egalitarian, easygoing, tolerant, and less aggressive. Stump tails reconcile with each other fairly easily after fights (by holding each other’s hips), but reconciliation is far less frequent among rhesus monkeys. Oddly, the social dynamics of the artificially created mixed group tended to reflect the behavior of the stump tails: rhesus monkeys learned peacemaking and reconciliation from the stump tails. (However, the results were not due to imitation: rhesus monkeys did not use stump tail methods of reconciliation; they used their own techniques, just more frequently.) Moreover, when the two groups were subsequently separated and the rhesus monkeys were returned to an all-rhesus group, the monkeys from the mixed group continued to show three times more reconciliation behavior after fights than the other rhesus monkeys, though their peaceful behavior did not spread to the others. 157 The second experiment was observed by Robert Sapolsky in the natural environment of Kenya. Sapolsky had been observing a group of savanna baboons that he called the Forest

46

Chapter 2

Troop. The group was very combative and not very social (or sociable). Savanna baboons have a reputation for their nastiness: They achieve their hierarchies through violent challenges and they fight dirty; male mortality among savanna baboons is due primarily to violent encounters. Females are good at reconciliation, but males are not. 158 Some of the more aggressive members of the troop invaded the territory of a neighboring group, called the Garbage Dump Troop—so named because it had the good fortune of squatting on land near a tourist lodge with plentiful garbage to scavenge. Daily combat broke out between the two groups over the dump’s tasty leftovers. Then chance intervened: A tuberculosis outbreak devastated the Garbage Dump baboons and most of the Forest Troop baboons who foraged at the dump. This left the remainder of the Forest Troop with fewer aggressive males and with more social males than might otherwise be normal. And it left the troop with a female-to-male ratio now decidedly in favor of the former. As a result, the troop developed a much looser hierarchy and experienced less frequent aggression and more affiliative behavior (primarily grooming). Even more interesting is that twenty years later, after the young males had left the troop and new males had arrived from elsewhere (all males were émigrés), the uniquely low aggression/high affiliation culture of the group persisted. New males adopted the behavioral styles of the Forest Troop’s residents; the culture was passed on to a second generation. This was clearly not due to genetic reasons but to the social transmission of culture—primarily by the way that females treated new males in the troop. 159 For Sapolsky, the behavior of the savanna baboons and of de Waal’s mixed monkey group confirms that “some primate species can make peace despite violent traits that seem built into their natures.” 160 The third experiment—involving the social evolution of humans—is gleaned from the historical record by Jared Diamond. 161 In December 1835 Maori invaders from New Zealand’s North Island slaughtered the local Moriori people on the Chatham Islands 500 miles to the east. Diamond describes the Moriori as “a small, isolated population of hunter-gatherers, equipped with only the simplest technology and weapons, entirely inexperienced at war, and lacking strong leadership or organization.” 162 They resolved disputes among themselves peacefully, and when the Maoris appeared, they decided not to fight. The Maoris, on the other hand, “came from a dense population of farmers chronically engaged in ferocious wars, equipped with more-advanced technology and weapons, and operating under strong leadership.” 163 In this classic confrontation between people of a peaceful culture and those of a warlike culture, the outcome was never in doubt. The Moriori were wiped out. This was not simply a confrontation of peoples with vastly different cultures with regard to the use of violence. In fact, the Morioris and the Maoris were essentially the same people. Both were descended from Polynesian farmers who had populated New Zealand around 1000 CE. A group of these Maoris then left to colonize the Chatham Islands, becoming the Morioris. As Diamond explains, the two groups evolved culturally in quite different directions, largely as a result of the environments in which they found themselves. Since the colder climate of the Chathams could not sustain the type of agriculture the Morioris were used to in New Zealand, they reverted to a hunter-gatherer existence. The lack of crop surpluses meant that they “could not support and feed nonhunting craft specialists, armies, bureaucrats, and chiefs.” 164 The political structures they created were relatively less complex. The small size of

The Individual Level of Analysis, Part I

47

the Chathams meant that it could only support about 2,000 people. And with no other islands nearby to colonize, “the Moriori had to remain in the Chathams, and to learn how to get along with each other. They did so by renouncing war, and they reduced potential conflicts from overpopulation by castrating some male infants. The result was a small, unwarlike population with simple technology and weapons, and without strong leadership or organization.” 165 The Maori, on the other hand, with the advantages of New Zealand’s warmer climate, became densely settled agriculturalists who developed crop surpluses that supported political chiefs and a part-time warrior cohort. They built ceremonial buildings and forts and developed the tools of warfare. They also were apparently engaged in chronic and ferocious warfare with their neighbors. The Maori and Moriori developed distinct cultures over time in response to different environmental factors. What can we deduce from these experiments? In both primate and in human societies, behavior is socially transmitted. Primates, including humans, have the capacity for both violence and peacemaking. The degree to which these behaviors manifest themselves varies considerably across groups, and they change over time due to environmental factors and learning. CONCLUSION What reasonable conclusions might we come to at this point about the contribution of biology, evolution, genetics, and learning to the collective human behavior called war? First, aggression is probably not an innate drive triggered by an internal state in humans and may not be an instinctual response to certain external conditions. Violence and warfare are too rare for this to be the case. Nevertheless, few would dispute that the human potential for aggression and violence are part of our evolutionary inheritance. They are tools in our behavioral repertoire, available under certain conditions. Second, not only is aggression part of the genetic inheritance of humans, but so is the ability for peacemaking. Third, the capability for aggression is modified by human intelligence, environmental conditions, and social learning. As international relations scholar John Vasquez has put it, “humans may not be ‘hard-wired’ for aggression, but they may be ‘soft-wired’ to favor certain behaviors under certain conditions, including a proclivity to react with violence over the issue of territory.” 166 Fourth, there is much about the institution of war that has been socially constructed over the ages. It is largely a social institution developed in certain places, at certain times, and under certain conditions— but not at other times or places and under other conditions. To the extent that it is a product of evolution, it is primarily the product of social, cultural, and political evolution. Fifth, there is no doubt that violence has always been with us. The myth of the peaceful hunter-gatherer is certainly a myth. Violence and organized warfare (on relatively small scale) existed before the Agricultural Revolution and the creation of states. However, it is also clear that the record of this “precivilized” violence shows tremendous variation. And there is a serious debate over how much of the violence constitutes actual war instead of violence on a smaller scale. At any rate, there is no doubt that after the Agricultural Revolution the amount of warfare and its destructive magnitude increased rather dramatically. But again, there is

48

Chapter 2

great variation geographically and temporally. These diverse patterns show the impact of culture and environment. Finally, we should note that the nature-nurture debate is fundamentally misleading. (If you have been paying attention, you have probably already figured this out.) The answer to what causes violence and aggression—biology and genetics or culture and environment and learning—is, of course, all of the above. The interaction between biological and environmental factors is reciprocal and complex. Almost all scholars agree that, as Sapolsky says, “The action of genes is completely intertwined with the environment in which they function; in a sense, it is pointless to even discuss what gene X does, and we should consider instead only what gene X does in environment Y.” 167 Before we go on to the more individualistic explanations of war, two final comments about theories of the “natural” aggressiveness of mankind are in order. First, if aggression has a genetic or instinctual basis, if it is indeed part of “human nature,” then efforts to eliminate war are almost certainly doomed to failure. Logically, in order to eliminate war we would need to (1) change the nature of humans, or (2) somehow place humans’ violent nature under severe and necessarily artificial restraints, or (3) provide outlets for humans’ innate aggression that are morally and culturally more acceptable and physically less damaging. We do not presently know how to do the first, and even if we did, it might not be a good idea to begin tampering with humankind through some sort of radical genetic engineering. The second and third options have been more or less continually applied throughout the ages without significant success in behavior modification. Second, if war is derived from an innate aggressiveness that is part of the common human nature, then how do we explain peace? Do people somehow “rebel” against their own nature and become peaceful? As we have stated, war and aggression are constant neither in time nor in space. Some states are significantly more peaceful than others, and even the most warlike countries are not constantly engaged in organized warfare. Theories that attempt to explain war by reference to the reputed aggressive nature of humans are able to give us little insight into the fairly substantial differences in the behavior of states. Since they are unable to deal with the tremendous variability in the behavior of states, theories that reduce the causes of war to innate aggression are ultimately unsatisfying. Attempts at explaining the variability of interstate warfare by means of a reputed constant—our aggressive nature—are logically doomed to failure. We are much more likely to be successful in developing a theory of war if we focus, as the nurturists have done, on the factors that explain the differences in the behaviors of individuals, groups, and states. One explanation for the varied aggressiveness of states is that such variation is due to the individual and personal psychological characteristics of the states’ leaders. The variation in the aggressiveness of states is due to the variation in the personal psychological natures of their leaders. We explore this possibility in the next chapter.

Chapter Three

The Individual Level of Analysis, Part II Psychological Explanations for War

The common folk do not go to war of their own accord, but are driven to it by the madness of kings. —Sir Thomas More Wars begin in the minds of men. —UNESCO Preamble

Thus far we have been operating under the assumption that humans are aggressive due to things that they have more or less in common: predispositions for aggression that are part of human beings’ biological inheritance or common cultural responses to our environment—our common “nature.” It is quite likely that what is more important are the differences between people rather than their similarities. It should be apparent that not all people have the same nature: Some are clearly more prone to violent behaviors than others. There are great differences in the psychological makeup of individuals, differences that are important to the understanding of conflict. In sum, human nature is a variable, not a constant. 1 Consider for a moment the war between India and Pakistan in 1971. The forces of West Pakistan had invaded East Pakistan to overturn the results of an election that would have placed an East Pakistani political leader at the head of the state. India was burdened by over ten million refugees from East Pakistan and harbored longstanding grudges against Pakistan. Finally, Prime Minister Indira Gandhi ordered Indian forces to penetrate five miles into Pakistani territory and issued an ultimatum ordering the Pakistani ruler, Yahya Khan, to withdraw from East Pakistan. Yahya Khan chose war instead, a fatal mistake since his army was quickly and decisively defeated, with East Pakistan gaining independence as Bangladesh. Why did this occur? Consider John Stoessinger’s assessment of Yahya Khan’s reaction to the Indian ultimatum; he maintains that the ultimatum was a hard blow under any circumstances, but for a man with Yahya Khan’s fragile masculine ego, such an ultimatum from a woman was psychologically unacceptable. Thus, even though he knew that the Indian forces outnumbered his own by a ratio of five to one, the president of Pakistan ordered a massive air strike. 2 49

50

Chapter 3

The defense of Yahya Khan’s male ego from intimidation by a female rival seemed to have been a decisive factor in the decision to go to war! Stoessinger’s judgment may or may not be valid, but it opens the door to the link between individual attributes and war. This chapter investigates the ways in which the individual leader influences the decision about whether to go to war. We will first discuss the circumstances in which individual leadership most affects decision making and then the various ways that individuals are influenced themselves: through personality, cognitive biases, emotion, and more. DO INDIVIDUALS MAKE A DIFFERENCE? It is quite appropriate that we should seek the causes of war in the individual makeup of national leaders who are in a position to choose between war and peace. The basic assumption at this level of analysis is that individuals do make a difference. It matters that Vladimir Putin (or Mikhail Gorbachev or Boris Yeltsin) sits in the Kremlin instead of Josef Stalin; it makes a difference whether Barack Obama sits in the Oval Office instead of George W. Bush. It matters, presumably, because in most cases wars are precipitated by the decisions of individual leaders and their advisers. One would be hard-pressed to find examples of war that occurred without a command decision from the highest level of government authority. Thus, if we want to know what caused the outbreak of war, we need to understand the individuals who were responsible for those decisions. On the other hand, we must be careful not to completely reduce the causes of every war to the psychological makeup of individual leaders. It is clear that the ability of any individual leader alone to determine war or peace is constrained by a great number of important factors: by the international and domestic environments, by the role of governmental bureaucracies in policy formulation and implementation, by formal and informal decision-making processes, and so on. It is also clear that there are some situations in which these constraints are less powerful and in which individual leaders will be able to make a significant impact on national policy. In such situations individual level characteristics may be decisive. Under what circumstances might we expect individual leaders to be able to rise above the normal organizational constraints? The obvious answer is when only the top executives and their advisers make the decision. The higher one is in the bureaucratic hierarchy, the fewer organizational and role constraints there are to restrict the influence of one’s personality on the decision. Perhaps the decision will be one in which the highest decision maker has sole responsibility. The fewer the number of individuals involved, the more we will be able to focus on individual and personality factors instead of larger institutional factors. But when do we see decision making at the highest levels? 3 Decisions are made by a small number of top level executives under a number of circumstances: 1. When formal, constitutional procedures (or informal, situational factors) demand it, owing to the type of decision involved or the structure of the situation. 2. When the top leader is constitutionally permitted great personal latitude and discretion to make decisions or where power is highly concentrated in the hands of a single leader (as in a dictatorship).

The Individual Level of Analysis, Part II

51

3. When the leader has a high degree of interest in the decision. 4. When there is only one institution responsible for the decision, permitting a few officials in that one institution to make policy unencumbered by opposition from other organized bureaucratic forces. Furthermore, the personal characteristics of these top leaders should have greater impact on policy: 1. When the decision is a nonroutine, fluid, or unanticipated situation—one for which no standard operating routines have been developed that would restrict the latitude of decision makers—as in a crisis. 2. When the information concerning the situation is either extremely low or when top leaders are inundated with an information “overload”; under these conditions the situation is likely to be highly ambiguous, allowing individuals to define the situation themselves and to make decisions in accord with their own predispositions. 3. When the top decision maker has little experience or training in foreign affairs, thus forcing the leader to rely more on his or her natural problem-solving predispositions. 4. When the situation is accompanied by great stress. Assuming that many of these conditions are met, we may begin to examine some personal characteristics that may be crucial determinants of whether leaders will choose to take their countries to war. THE ROLE OF REASON: THE RATIONAL MODEL In the best of all possible worlds governments would pursue peaceful policies rather than aggressive ones, their leaders would be enlightened and humane rather than incompetent or venal, and reason and good intentions would prevail over stupidity and evil. In this world of philosopher-kings and well-intentioned governments, we would expect policy to be made in a reasoned, calculated manner. Political scientists generally define rationality as a process in which certain steps are taken to ensure that the conclusion arrived at is the one that has the best chance of accomplishing the goals desired. Several political science models—expected utility theory, rational choice theory, and the rational actor model—all incorporate this idea of procedural rationality. Procedural rationality would include the following steps: 1. Identify and define the problem. 2. Identify and rank goals. Often, policy makers will have more than one goal. If so, they will have to be ranked in order of importance, since it is quite likely that no policy option can achieve all goals and that the choice of any particular policy might mean that some goals will have to be sacrificed in order to attain certain other goals. Rationality requires that decision makers make conscious trade-offs of the things they value. 3. Search for information. (This process is more or less continuous and may in fact start at step 1. It is included arbitrarily as step 3.) 4. Identify many alternative means/options for achieving the goal(s).

Chapter 3

52

5. Analyze each possible alternative. At this stage, decision makers must estimate and weigh the probable costs and benefits of each alternative, examine the possible consequences of using each option, calculate the effect of each option on other goals, determine the relative effectiveness of each option for achieving the goals, and estimate the probability of each potential outcome. 6. Choose the option that is best able to achieve your goals. In other words, select the optimal strategy—the one most likely to maximize your interests. 7. Implement the decision; that is, put it into action. 8. Monitor and evaluate the policy decision. Is it a success or a failure? Does it have defects? Is it achieving the desired results? 9. Terminate/alter/continue the policy as determined by your evaluation of it in step 8. This last step recognizes that decisions are not normally one-and-done deals; they are more often part of a chain of iterated decisions over time. Procedural rationality would be the ideal influence on individuals’ decision making regarding war, and sometimes this rationality does work as intended. But these are rather stringent standards, and as we will see in this chapter and the next, the real world of decision making departs from the ideal model considerably. THE ROLE OF PERSONALITY: PSYCHOLOGICAL THEORIES Psychological Needs or Motives Psychologists have identified a variety of psychological needs that motivate the actions of individuals, some of which are relevant to politics. The need for self-affection or love, the need for self-esteem or dignity, and the need for self-actualization or fulfillment are often identified, for instance, as well as the needs for security, power, and dominance. All individuals have similar needs; however, the relative importance of these needs varies. While some individuals seem to be dominated by the need for self-esteem, others might be dominated by the need for power. Abraham Maslow hypothesizes a hierarchy of needs. 4 Listed in their order of assumed priority they are as follows: 1. 2. 3. 4. 5.

Physical (biological)—food, water, air, sex Safety—the assurance of survival, security Affection and belongingness—love Esteem—for self-esteem and the respect of others Self-actualization or self-development

According to Maslow, these needs are both universal and innate; all humans have these needs. Each set of goals in turn monopolizes the individual’s consciousness. When the first set of needs is fairly well satisfied, the next “higher” need emerges to dominate the conscious life and serve as a central motivator of human behavior. The “higher” needs may not become activated until the “lower” needs are reasonably satisfied. For instance, if both psychological and security needs are fairly well satisfied, affection needs will become important for the

The Individual Level of Analysis, Part II

53

individual. Thus, which needs are important to the particular individual depends on his or her prior pattern of need gratification. 5 Particularly important is Maslow’s depiction of self-actualizing individuals—those who have achieved satisfaction of their physical needs and their psychological needs for security, belongingness, and self-esteem. This physical and psychological security makes it possible for these individuals to have trust in their environments. Presumably, individuals with high selfesteem are not only more trusting of others but also more opposed to the use of force. However, their confidence in their own abilities would probably lead them to accept greater risks than others. On the other hand, individuals with low self-esteem have been depicted as being anxious, hostile, uncooperative, tough bargainers, paranoid, and nationalistic and as having a propensity toward the use of military force. (Many would put Wilhelm II and Stalin in this category.) The basic presumption, at least for some scholars, has been that aggressive behavior is the result of the need to compensate (and indeed to overcompensate) for anxiety caused by low self-esteem. A corresponding assumption is that people with high self-esteem no longer have an overwhelming urge to see themselves favorably; they’ve satisfied this need. Presumably, then, high self-esteem confers immunity from ego threats and greatly reduces a major reason for aggressive behavior. This does seem to be the case for many such people, but reality (as usual) is somewhat more complex. Studies of various kinds of people who manifest high levels of aggression—including narcissists (people with grandiose views of themselves), psychopaths (who tend to have truly astonishing egocentricity and a sense of entitlement to go along with their criminal tendencies), and adolescent bullies—indicate that none of these people suffer from low self-esteem. Some psychologists now argue that violence is associated with a small subset of people with high self-esteem—those whose high self-esteem is uncertain and unstable. When these people are confronted with threats to their egos inflicted by others, they tend to react aggressively. 6 On the other hand, those whose high self-esteem is stable tend to be the least violent and aggressive. Lloyd Etheredge’s famous study in political psychology shows that people with high self-esteem differ in their attitudes toward the use of force. Etheredge studied male U.S. foreign service officers, military officers, and domestic affairs specialists and their responses to international crisis scenarios. He found that those measuring high on self-esteem generally tended to oppose the use of force, but those with high scores on self-esteem coupled with high ambition were the most dangerous, showing a strong tendency to advocate military solutions. 7 Whether you happen to believe that Maslow’s interpretation is correct or not, most political psychologists argue that three fundamental sets of motives are readily recognizable among political actors: (1) People whose primary motivation is power desire to influence or to make an impact on other individuals or countries or on the world at large. (2) Individuals whose primary motive is achievement are driven to accomplish major deeds and attain high standards of excellence. They want to make a mark on the historical record. (3) Individuals whose primary motive is affiliation are concerned about establishing and maintaining positive personal relationships. They wish to occupy a seat of power that places them in the limelight. 8 Students of politics are probably all familiar with leaders who seem to have a tremendous need for power. According to Winter, power orientation has two sides—one good, one bad. 9

54

Chapter 3

Power orientation, coupled with self-control and a sense of responsibility, may have positive effects. However, if power orientation is not restrained by a sense of responsibility or selfcontrol, the results can be rather nasty. These power-oriented people tend to dominate others, to be argumentative, to be paranoid, and to have very little humanitarian or moral concern, and they may be impulsive and prone to taking extreme risks in the pursuit of prestige. Studies also find that high need for power is associated with a leader’s suppression of an open decision process. 10 The individual’s need for power is also linked to a tendency toward exploitative, conflictual, and aggressive behavior and a disposition toward the advocacy of war or the use of force in foreign policy decisions. 11 Power-oriented persons are frequently believed to be compensating for deprivations experienced during childhood, where their needs for security, love, achievement, and self-worth were not met. 12 Unfortunately, such individuals also tend to desire positions of leadership that permit them to exert control over others; indeed, this may be the defining characteristic of professional politicians! Harold Lasswell suggested many years ago that the primary motivation for political activity is emotional insecurity or low selfesteem, conditions that are compensated for by a drive for power. 13 There is even some evidence that the greater the top leader’s need for power, the more aggressive his government’s foreign policies. 14 Winter’s analysis of the link between power motivation and foreign policy actions by U.S. presidents shows that although the causal linkage is complex, presidents with power motivation were associated with involvement in wars. 15 Margaret Hermann’s analysis of forty-five world leaders also indicates that power motivation is related to high levels of hostility expressed toward other countries. 16 On the other hand, those individuals who are dominated more by need for affiliation and need for achievement tend to lean toward more cooperative interactions with others. Winter and Stewart’s study of American presidents indicates that presidents with higher affiliation and achievement needs (as opposed to power needs) were less likely to engage in war and more likely to support arms control. 17 And Terhune’s international relations simulations suggest that individuals who are achievement oriented pursue cooperative strategies at first, hoping their opponents will also be cooperative. 18 Winter’s analysis of Saddam Hussein’s motivations, based on content analysis of his interviews with reporters between 1974 and 1991, indicate a profile characterized by high power motivation, above average affiliation motives, and low achievement motivations. As we will see shortly, the dominance of power motivation corresponds with assessments of Saddam Hussein by others who describe him as having an unbounded drive for power and prestige, the use of aggression as an instrument of power, and a paranoid fear of enemies. Each of these traits is statistically linked to power motivation, especially when the subject lacks a sense of responsibility. 19 Personality Traits While human beings exhibit a wide variety of personality traits, some have special relevance for the topic of war. Dogmatism: One personality type that students of international conflict might wish to watch out for is what Milton Rokeach has dubbed the dogmatic personality. 20 Individuals with dogmatic personalities are rather closed-minded; they find it difficult to accept and use new

The Individual Level of Analysis, Part II

55

information that contradicts their beliefs. They do not tolerate ambiguous information very well; they are unlikely to examine the full range of alternatives available; and they have a tendency to rely on stereotypes. They are generally suspicious, have high levels of anxiety, and are likely to perceive conspiracies. They are also predisposed to condone the use of force. 21 Authoritarianism: One of the most frequently discussed set of traits belongs to individuals who may be described as authoritarian personalities. A famous study of T. W. Adorno and his associates identified a complex of characteristics that typified such a personality and then developed a scale through which they could ascertain (by the use of questionnaires) whether a particular individual possessed these traits. 22 Although Adorno named his scale the F-scale for fascism, those who score high on this scale tend to have beliefs that could categorize them as members of either the extreme right (fascists) or the extreme left. The traits involved include a preoccupation with virility and strength, a tendency to dominate subordinates, deference to superiors, the need to perceive the world in a highly structured way, discomfort with disorder, a preference for clear-cut choices, rigidity, and the use of stereotypes. The effect that such a personality would have on the ability of an individual to make rational decisions is obvious. In addition, authoritarians also tend to be highly nationalistic and ethnocentric—characteristics that are both highly associated with support for war and aggression. 23 Dominance: Students of politics are also familiar with individuals who possess domineering personalities: Lyndon Johnson, Richard Nixon, and Henry Kissinger come immediately to mind. Individuals with such personality characteristics are probably plentiful on the political scene; political recruitment patterns and self-selection combine to elevate those with high dominance characteristics to high positions. Winter generally concludes that the dominance trait is associated with a disposition toward war and the use of force. 24 Two separate studies of American presidents and their foreign policy advisers indicate that individuals who possessed the personality trait of dominance were usually much more likely to advocate the use of threats and military force and oppose conciliatory moves than those who scored lower on dominance. Indeed, based on knowledge of the individual’s personality, the authors in both studies were able to accurately predict 77 percent of the time whether that person would advocate the use of force or not. 25 In other words, the personality trait of dominance seemed to have been generalized from the individual’s everyday life to the realm of policy. Domineering leaders tend to relate to other countries in much the same way as they relate to other individuals. This is an important finding. It suggests that decisions on the use of force at the national level are determined at least in part by the subconscious personality traits. Extroversion/Introversion: One analyst has discovered that in disagreements concerning U.S. policy toward the Soviet bloc, the more extroverted personalities were much more likely to advocate cooperative policies than those who were more introverted. Since personality factors interact, predictions of behavior based on a combination of traits are preferable. The evidence indicates that the combination of high dominance and introversion factors would seem to create a particularly undesirable mix. Here is Lloyd Etheredge’s analysis of such individuals, whom he labels “Block (Excluding) Leaders,” and among whose ranks we find

Chapter 3

56

John Foster Dulles, Woodrow Wilson, Herbert Hoover, Charles Evans Hughes, Henry Stimson, Dean Acheson, and Cordell Hull: The Bloc Leaders tend to divide the world, in their thought, between the moral values they think it ought to exhibit and the forces opposed to this vision. They tend to have a strong, almost Manichaean, moral component to their views. They tend to be described as stubborn and tenacious. They seek to reshape the world in accordance with their personal vision, and their foreign policies are often characterized by the tenaciousness with which they advance one central idea. 26

Risk Orientation: This personality trait also deserves mention. The individual’s propensity for taking risks would seem to be a trait that may be of cardinal importance regarding decisions for war or peace. In such situations some decision makers are relatively more risk acceptant, while others are more risk averse. Given the same evaluation of the costs and benefits to be derived from war, some decision makers may be willing to take the risk for war given a particular probability of success (say, 50 percent), while other decision makers may require a greater probability of success (say, 75 percent). This individual difference may play an important role in the decision to go to war. 27 Byman and Pollack hypothesize that risktolerant leaders are more likely to cause wars, and they offer Hitler, Napoleon Bonaparte, and Saddam Hussein as historical examples. 28 We’ll have more to say about this later. Combinations: From her psychological study of foreign leaders, Margaret Hermann has constructed six personality orientations each of which is made up of different combinations of various motives, traits, and cognitions. Each orientation is associated with certain foreign policy behaviors. Probably the one that should cause us to stand up and take notice is the orientation that she refers to as “expansionist.” This type of leader is interested in gaining control over more territory, resources, and people. The psychological components that make up this complex are power motivation, nationalism, a belief in one’s ability to control events, self-confidence, distrust of others, and task motivation. 29 A separate study by Hermann places Saddam Hussein in the expansionist orientation category. 30 Stephen Dyson has analyzed British Prime Minister Tony Blair’s personality based on a content analysis of his responses to Parliamentary Questions on foreign policy. He argues that Blair’s support for British use of force against Iraq in 2003 can be accounted for by the combination of his high need for power and his high belief in his ability to control events. These attributes led him to tightly control the decision-making process in the hands of a small group and to permit very little open debate on the issue. Meanwhile, his low conceptual complexity led him to see the Iraq issue in a fairly simplistic black-and-white way. 31 Personality Complexes We can now advance from looking at mere personality characteristics or traits to actually personality complexes. The term personality refers to more than just an accumulation of discrete individual traits; it indicates a more complex and integrated whole. For political psychologists, personality “implies a patterned relationship that links belief systems, value systems, attitudes, leadership styles and other personality features.” 32 The elements of personality would include one’s knowledge; one’s affect or feelings (including anxiety, aggression, hostility, shame, and guilt); drives for power, achievement, and affiliation; one’s evaluation of

The Individual Level of Analysis, Part II

57

reality; interpersonal relationships (including capacity for empathy); one’s identity (and ambivalence about identity); one’s characteristic ego defenses; and cognitive/intellectual style. Particular personality types tend to be linked with particular belief systems and leadership styles. 33 Political psychologist Jerrold Post identifies three personalities that are of particular interest to political psychologists: narcissists, obsessive-compulsives, and paranoid personalities. He argues that the first two are “disproportionately represented among political leaders.” The third, those with paranoid personalities, are uncommon, but when present can create “catastrophic consequences for international relations.” 34 During times of stress, individuals with each of these patterns of personality organization can essentially cross the line into personality disorders, which become “psychologically disabling.” Post argues that the presence of a fullblown personality disorder is not consistent with sustained office-holding; such individuals cannot last long—especially in a democracy—but they may be found in closed societies. 35 Let’s take a quick look at two of these personality complexes—paranoid personalities and narcissism. As we look at these personality complexes, it is important to note that political leaders (like all of us) possess a considerable array of psychological characteristics that are unlikely to fit one pure type. Instead, what is important is the predominance of one style over another. 36 Paranoid Personality: Political leaders with a paranoid personality complex are identified by a host of interrelated characteristics. 37 They are incredibly suspicious and mistrustful of others; they are hypervigilant of threats to themselves and see themselves as surrounded by enemies. The world they perceive is one of complex plots, “hidden motives and special meanings.” They are also hostile, stubborn, unwilling to compromise, and defensive. Great effort is expended to create a near-perfect self-image as a compensation for feelings of inferiority, and paranoids are therefore hypersensitive and easily slighted. Perceived threats to their self-images trigger a projection mechanism that is likely to bring about an aggressive counterattack. They also have a need for self-reliance as a way of defending against the dangerous world they perceive. (And since self-reliance is related to power, there is a strong drive for dominance that is part of the paranoid personality.) Leaders with paranoid personalities have a cognitive style that is highly rigid. They believe that they know what the world looks like, and it is full of external dangers and threats. They are internally driven to search for information that confirms this view, and they tend to systematically exclude information to the contrary. Paranoids are prone to internally motivated biases and to disregard information at variance with their belief system. The paranoid creates an elaborate system of beliefs and theories to justify the projection of threat onto others. This process biases them in favor of worst case analyses. (The result, of course, is a highly distorted view of the world.) They need to maintain a set of internally consistent beliefs—all of which are centered around an image of an enemy. The enemy is viewed as totally evil and threatening without any shades of gray. When confronted with new situations, paranoids tend to rely on the use of historical analogies as a heuristic shortcut to understanding. They compare current situations to things that have happened to them before. Their diabolical view of the enemy means that they prefer the use of force over negotiation and they lean toward preemptive strategies in a crisis.

58

Chapter 3

Since paranoids tend to see threats where they do not exist, tend to have delusions of complex plots against them, and tend to be extremely vigilant in scanning the environment for threats emanating from diabolical opponents, one would think that the primary risk of war with such leaders is that they engage in preventive or preemptive wars against presumed enemies who are not really a threat to their countries. This is logical and it may actually have occurred. One political psychologist, Raymond Birt, has argued that Josef Stalin—a leader who heads almost everyone’s list as the classic poster child for paranoid leaders—was actually under-prepared for Hitler’s invasion of the Soviet Union in June 1941 because of his paranoia. 38 This sounds quite counterintuitive, but it makes sense. Germany posed a real threat to Russia, not a delusional one, and at one level Stalin certainly understood this. But there were still delusional elements in his analysis of German behavior. It is well known that in the days before the German attack, Stalin and other Soviet leaders had abundant information that the attack was about to occur—all of which was dismissed by Stalin as being fabrications or part of a calculated German plot to provoke the Russians into a preemptive attack; thus, he gave orders for Soviet forces to avoid provocation. Even after the German attack began, he was hesitant to permit counterattacks by Soviet forces. Stalin is reported to have gone into a state of near catatonia, retreating to his dacha, drinking heavily, not communicating to his staff or issuing orders, seemingly engulfed by debilitating depression. In other words, he had a psychological breakdown. Hitler’s attack was a direct, personal insult to his narcissistic ego and this resulted in the collapse of his self-esteem. He assumed that his days as leader were numbered and the other members of his Politburo would arrest him. In this case, Stalin’s typical hypervigilance to threats was overcome by the rigid beliefs and attitudes that he had constructed regarding Hitler. According to Birt, Stalin believed he understood Hitler and even identified with him and admired him. Therefore, Hitler would not attack him. Other reconstructions of Stalin’s thinking are possible. Stalin might have believed he had figured out Hitler’s modus operandi: he always issued warnings or ultimatums before he invaded other countries (and no ultimatum had been issued to Russia), and he tried to provoke others to use force in order to justify his actions as self-defense. Or Stalin might have believed that Hitler was a realpolitik statesman like former Chancellor Bismarck who knew better than to put Germany into a two-front war. 39 Regardless of what Stalin thought, the problem is that paranoids have rigid belief systems and adhere to these beliefs very stubbornly to protect their infallible self-image, even in the face of clear and unambiguous information to the contrary. Another aspect of Stalin’s paranoia was that the threat that he was primarily fixated on— the supposed threat from Trotskyites and other internal opponents—came from within rather than from abroad. While these threats to Stalin’s rule had little basis in reality, the result was that he ruthlessly and relentlessly purged the party organization, the intelligence agencies, and the military officer corps—thus vastly weakening the state’s ability to protect itself against the real threat from Germany. Narcissism: Narcissism is a highly complex personality construct made up of several factors. 40 These factors include (1) extreme egotism and extreme self-absorption—including a sense of self-entitlement and superiority that verge on grandiosity, a preoccupation with fantasies of unrealistic goals, the need for continual attention and admiration, and physical vanity;

The Individual Level of Analysis, Part II

59

(2) a disposition to exploit and manipulate others, coupled with an inability to empathize with others; (3) an absence of conscience in the pursuit of one’s goals; (4) a reveling in leadership and authority roles; and (5) a hypersensitivity to the criticism of others. This external grandiosity constitutes a massive compensation for feelings of inferiority and/or uncertainty about identity: The massive hunger for acclaim is a reaction to threatened self-esteem. Narcissists are vulnerable—sensitive to slights and criticism—and they suffer psychological injury easily. This inhibits their ability to do reality testing. When placed in leadership positions, they are prone to surround themselves with toadies who will tell them what they want to hear. Moreover, their grandiosity leads to an overestimation of their skills and ability to control events, impairing their judgment. Strong relationships have been found between narcissism and hostility, aggression, and the need for power. Several psychologists have concluded that Iraq’s Saddam Hussein had a narcissistic personality. 41 In fact, Jerrold Post, who has constructed personality profiles of world leaders for the CIA, has concluded that Hussein’s personality was that of malignant narcissism, a particularly dangerous personality disorder. Saddam saw himself as a great historical figure—a world leader of the status of Nasser, Mao, or Castro. This identity was linked to dreams of glory and a messianic vision to rid the Arab world of Western influence and to unite it under a single ruler—himself. He is described as having a paranoid outlook on the world, prone to seeing the United States, Iran, and Israel behind plots against his leadership and justifying his aggression as required by threats from his enemies. (This is not to say that Hussein lacked enemies, just that he tended to ignore his own role in creating these enemies and that he generally exaggerated the threats.) He was consumed by a drive for unlimited power—a drive that was unconstrained by conscience or by a concern for the suffering of others. But these dreams of glory, feelings of specialness, and messianic ambition (as well as his acts of aggression) hid underlying feelings of self-doubt and insecurity. The combination of messianic ambition for unlimited power, absence of conscience, unconstrained aggression, and a paranoid outlook constitute the central constellation of his personality. While these traits do not describe someone with a pleasant personality, Saddam Hussein did not suffer from a psychotic disorder. He was capable of seeing the difference between fantasy and reality; he was rational rather than impulsive, and he was capable of deliberation and even patience. He was even capable of reversing course and backing down, but was likely to back down only if he could keep his honor and prestige intact. However, in Post’s analysis, “While he is psychologically in touch with reality, he is often politically out of touch with reality. Saddam’s worldview is narrow and distorted, and he has scant experience out of the Arab world . . . Moreover, he is surrounded by sycophants who are cowed by Saddam’s wellfounded reputation for brutality and are afraid to contradict him.” 42 Thus, he is a calculating political leader who was constantly subject to miscalculations. While he was indeed a ruthless political calculator, he did not see himself as a martyr; his chief priority was his own political survival. While not irrational, “he can be extremely dangerous and will stop at nothing if he is backed into a corner. If he believes his survival as a world-class political actor is threatened, Saddam can respond with unrestrained aggression, using whatever weapons and resources are at his disposal.” 43

60

Chapter 3

There is little doubt that Hussein’s personality played a role in the three wars Iraq was engaged in during his time as leader—the 1980–88 Iran-Iraq War, the 1990–91 Gulf War, and the 2003 Iraq War. In each he made grave miscalculations. In the first two he engaged in aggression against neighbor states, Iran and Kuwait, which eventually led to retaliation that left Iraq prostrate. In the third war his unwillingness to provide complete transparency about Iraq’s lack of WMD played a role in permitting Bush administration officials to perceive that Iraq constituted a threat to U.S. interests that required a military solution. 44 Psychohistory Anyone who has taken three final examinations on the same day is aware that there is a rather fine line between sanity and mental illness. The stresses and strains of high political office would seem to place government officials quite near the line more often than either they or we would desire. Jerome Frank maintains that as many as seventy-five chiefs of state in the last four centuries have suffered from severe mental distress while in power. 45 Hitler, Wilson, and Stalin have all been described by their biographers as having had severe psychological problems. Since these biographies combine history and psychology, they are frequently given the name “psychohistory.” Let us take just one example of how social scientists have applied psychological insights to explain a leader’s decision for war. Several scholars have looked at Kaiser Wilhelm II’s decisive role in starting World War I, especially historians John Rohl and Robert Waite and political scientist Richard Ned Lebow. 46 The Kaiser’s problematic personality was well known to his contemporaries. British Foreign Minister Grey declared his belief that the Kaiser was “not quite sane,” that he appeared somewhat like “a battleship with steam up and screws going, but with no rudder,” and that someday he just might “cause a catastrophe.” 47 Those who knew him described him as childlike, impulsive, mercurial, and unstable. He vastly overestimated his own abilities, had grandiose visions of himself, was unable to take criticism, and appeared incapable of learning from advice or experience. It is also speculated that he suffered from sexual confusion. These debilitating traits were most likely formed deep in childhood. The Kaiser was marked by conspicuous birth deformities, lacked a nurturing family life, and was saddled with the burden of living up to high expectations. As a result, he depended heavily on subconscious ego-defense mechanisms that protected his insecure self-image. Grandiose self-concepts, assertions of power, and “compensatory belligerency” were psychological defenses against feelings of inadequacy and perhaps also fears of homosexuality. He was clearly emotionally unstable. Freud had in fact diagnosed his problem as “neurasthenia,” a neurosis that was a manifestation of his inability to cope with emotional conflicts and feelings of inferiority. (Wilson had the same diagnosis.) 48 Medical authorities had informed his family that while he was not clinically insane, neither would he be entirely normal. He would be subject to sudden outbursts of anger and would not always be capable of making rational decisions. 49 In the lead-up to World War I, the Kaiser had initiated several diplomatic confrontations over Morocco, where he had been forced to back down by the French and the British. Henceforth, the Kaiser was committed to a policy of “compensatory belligerency” as a way of demonstrating his masculinity, toughness, and decisiveness. The July Crisis of 1914 would become a test of his manhood; he declared to the German industrialist Krupp, “This time I

The Individual Level of Analysis, Part II

61

won’t cave in.” 50 He encouraged his Austrian allies to take a tough stance against Serbia (and its supporter Russia) in response to the assassination of Archduke Franz Ferdinand on June 28, 1914. By the end of July, the crisis was spiraling out of control. The Kaiser’s illusion that Germany could fight a small, localized war in the Balkans and that the British would be neutral in this affair was shattered by telegrams received on the night of July 29–30. Both telegrams were misinterpreted by Wilhelm. The telegram from Russia announcing that Russia had begun mobilization against Austria five days earlier was interpreted as meaning the Russians had mobilized against Germany as well. The telegram from Germany’s envoy in London that Britain would not remain neutral if France was attacked created a paranoid reaction: the Kaiser now believed that Britain, France, and Russia were engaged in a conspiracy to keep Germany negotiating while the Entente powers mobilized to attack. Lebow argues that the Kaiser suffered an acute “anxiety reaction” to this news. According to Lebow, Wilhelm’s paranoid response was perhaps indicative of his need to resort to more extreme defense mechanisms to cope with the free-floating anxiety triggered by the breakdown of his former defenses. No longer able to deny the probability of Russian, French, and British intervention, yet unable to admit just how grievously he had miscalculated, Wilhelm chose instead to escape from his own aggressiveness and its consequences by portraying Germany and himself as hapless victims of the aggressiveness of other powers. Paranoid delusions of persecution are typically triggered by environmental or inter-personal stress, although they tend to occur only in persons who have formerly maintained an unstable psychological balance by resorting to denial or other defense mechanisms. The Kaiser was such a person. 51

Confronted by reports of Russian mobilizations, the Kaiser convinced himself that his choices had now been outpaced by events and that there were no options other than to mobilize German forces for war. He bore no responsibility for what would happen next. Although psychobiographies play an important role in alerting us to the importance of individual personalities in domestic and international politics, one must be careful in assessing their value. The main thing that separates psychohistory from the work of clinical psychologists and therapists is that the latter have the advantage of actually working with and interviewing their subjects; historians work one or two steps removed from this. Politicians are notoriously secretive about their personal lives, and none of the psychohistorians mentioned was able to actually interview his or her subject. Instead, they had to rely on already published biographical material, speeches, letters, diaries, and interviews with relatives and associates. Second, one often wonders about the objectivity of these studies. Biographers are usually not drawn to their subjects because they are neutral toward them; they are usually attracted to them as heroes or repulsed by them as villains. Finally, psychohistorians may be guilty of the sin of reductionism, reducing (in the most extreme form) the nation’s foreign policy to the toilet training of the president. One must be careful not to focus entirely on early childhood conflicts and related pathologies as a cause of later behavior without concern for the social and political environments of the time.

62

Chapter 3

THE ROLE OF EMOTION: THE COGNITIVE REVOLUTION Dramatic recent advances in cognitive psychology, biological anthropology, evolutionary psychology, and neuroscience (especially through the use of brain imaging techniques) have converged to fundamentally change the way we look at how humans process and evaluate information and make decisions. 52 Political scientists have traditionally believed that emotions were separate from and detrimental to the process of rational decision-making. Neither appears to be entirely true. Before we start, we need to define our terminology. This is not easy since (can you believe it?) scholars have found it difficult to agree on definitions. At any rate let us define emotions as being subjective human experiences that have physiological components; they are “inner states that individuals describe to others as feelings, and those feelings may be associated with biological, cognitive, and behavioral states and changes.” 53 Emotions thus involve thoughts (cognition), biological sensations, motivations, and an internal sense of experience. 54 A short list of emotions that political leaders might have to deal with would include fear, anxiety, awe, admiration, love, hate, anger, regret, panic, desperation, pride, and shame. Stimuli from the environment bring about physiological changes in humans, including the production of emotions such as fear. The limbic system—which is common to all mammals— is the part of the brain that regulates basic biological mechanisms. It includes the thalamus (which receives information from the outside world through sensory organs) and the amygdala. The latter is a walnut-sized organ on either side of the brain that constitutes man’s emotional center, where threats are recognized and where fear and anxiety are produced; it also serves as a filter for memory. Political psychologist Rose McDermott describes how the amygdala works: The amygdala . . . constantly scans incoming information for signs of an emotional emergency. If there is none, the amygdala recognizes that there is time for more advanced processing through the prefrontal cortex, which is responsible for more abstract, rational thought. But if the brain decides that the information indicates a threat, it shuts down the higher mental processes through the release of catecholamine neurotransmitters, which also turn on the action of the emotional brain to react quickly and efficiently to threats. This processing hierarchy evolved to help humans survive fight-or-flight kinds of challenges. 55

From an evolutionary perspective, emotions are adaptive and functional. The system is designed to give priority to potential threats, which produce the emotion of fear. Fear is linked to physiological changes such as increased adrenaline and a more rapid heartbeat, which prepare animals for fight or flight. Additionally, the emotional processing of information is an integral part of rational thought; indeed, it precedes higher-level cognitive processing and is much quicker. That is why it is given preference in situations of threat. Speed matters. Emotion and cognitive rationality are not processes that are independent of each other; they are intertwined. Social scientists point to the findings of prominent neuroscientist Antonio Damasio and his colleagues—particularly to their famous patient “Elliot.” As a result of surgery for a brain tumor Elliot suffered neurological damage to his ventromedial prefrontal cortex—a part of the brain that controls emotion. Elliot was clearly intelligent and retained his memory, language,

The Individual Level of Analysis, Part II

63

and cognitive skills; however, he was “emotionally flat” and lost his ability to “feel.” His practical reasoning ability was severely impaired, and he seemed unable to make socially appropriate decisions. The inability of people with similar brain damage to make good social decisions probably results from their inability to tap into emotions and to envision future consequences. 56 People lacking such contact with emotions “do not care about themselves or about others, and they neither try to avoid making mistakes nor are they capable of learning from their mistakes. They ‘know’ but cannot ‘feel’ and, consequently, they make mistakes.” 57 Damasio has theorized that information received through our senses creates emotions that provide clues—good/bad, pleasurable/painful—which then assist us in making future decisions in an efficient manner. 58 Emotion and rationality are not distinct systems in which rationality leads to good decisions and emotions lead to bad decisions; the relationship is much more complex. As Kahneman notes, “an inability to be guided by a ‘healthy fear’ of bad consequences is a disastrous flaw.” 59 In the estimation of Jonathan Mercer, “Emotion is necessary to rationality and intrinsic to choice.” And “people who are incapable of using feelings are irrational, and those who ignore feelings and base cooperation on observable facts cannot trust.” 60 Trust is important in facilitating cooperation, as we will see when we discuss game theory in chapter 9. One way to think of this is that the brain has two interconnected operating systems for decision making, one based on emotion and one based on reason. The emotion-based system—which Daniel Kahneman refers to as system 1—can be described as intuitive, preconscious, automatic, fast, effortless, unreflective, and slow to change. It is an ancient product of our evolution as humans. The reason-based or cognitive system—which Kahneman refers to as system 2—is rule-governed, conscious, slow, effortful, reflective, and flexible. It is sometimes thought of as the more “developed” part of our brain. Kahneman suggests that most human decisions are made through the first system and in comparative terms, the first system always trumps the second one, though system 2 can “manually override” system 1 if you want it to. 61 However, a word of caution is in order about these terms: Kahneman notes that systems 1 and 2 are somewhat “fictitious,” in the sense that “there is no one part of the brain that either of the systems would call home.” 62 No bright line divides cognition from emotion. Let’s add two further pieces of evidence. First, there is the oft-quoted fact that while the brain can absorb approximately eleven million pieces of information a second, it can only process about forty pieces consciously; the unconscious brain handles the rest. Second, there is the frequently repeated experiment that indicates that brain signals associated with the movement of a person’s limbs actually occur about a half second before one is conscious of deciding to move. 63 One reporter has suggested that all of this suggests “that the conscious mind is like a monkey riding a tiger of subconscious decisions and actions in progress, frantically making up stories about being in control . . . the conscious brain is only playing catch-up to what the unconscious brain is already doing.” 64 Most mental operations are, in reality, not conscious and deeply thoughtful, they are instead largely automatic and subconscious (they’re part of system 1). As Robert Jervis notes, much of the cognitive process is “beyond the reach of conscious thought.” 65 Another facet of emotion’s role in decision making is shown by the now famous “ultimatum game” first developed by economists. In this game there are two players, A and B. A

64

Chapter 3

controls a certain set of resources and has exclusive power to decide how these resources will be divided between himself and B. In the game, A makes a one-time offer to B to divide the resources: he may give B anywhere from 1 percent of the resources to 100 percent. B’s only choice is to accept the offer or reject it. If B accepts, the resource is split according to the deal; if B chooses not to accept, neither player gets any of the resource. At this point the game is over. It is a one-and-done game that will not be repeated. Logically, B should accept any split offered by A, since something—even a paltry 1%—is better than nothing. And just as logically, A should offer as little as possible, knowing that B will likely accept any deal A offers. In practice, however, results repeatedly show that B is likely to reject offers of less than 20%. Why? B players are apparently insulted and humiliated by low offers and deem them to be unfair. Their response to low-ball offers is normally quick and “intuitive” and clearly appears to be an emotional response to the situation. 66 All of this should push us toward the understanding that if one wants to understand how people—including government leaders—make decisions, then we need to understand the role that emotion plays. (Paul Slovic calls this the affect heuristic.) People make decisions by consulting their emotions; we tend to let our likes and dislikes determine our beliefs. 67 Here is a short list of other ways in which emotions may influence decision makers. First, emotions are crucial to the perception of threat—an important causal factor in war. As we have already seen, the limbic system plays a role in detecting threats. Threat perception happens at a much faster rate than the cognitive processing of potential threats. Given the high cost of failure to detect threats, evolutionary forces have set our “default position” at a rather low threshold. But as Neta Crawford notes, human emotions are not the only consideration: threats are both cognitively (consciously) processed by the human brain and socially constructed. That is, what constitutes a threat is determined partly by our relationships with others in the environment. 68 Second, “background emotion” in the form of one’s general mood can have an effect on memory, attention, perceptions, and beliefs. 69 According to Kahneman, when we are in a good mood, we are more intuitive, more creative, but also less vigilant and more gullible and prone to error. We are more reliant on the automatic functioning of system 1; a good mood biologically signals that everything’s okay and we’re safe. 70 When we are in a bad mood, we lose touch with our intuitions; we become more vigilant, more suspicious, and more analytical; we have activated system 2. 71 Third, while emotions may aid decision making at some levels, extreme emotions can interfere with rational choice. Emotions unrelated to the outcome of the decision may be present as individuals make decisions. During the choice process individuals may experience anxiety, compassion, shame, pride, awe, admiration, anger, regret, panic, desperation, love, and hate—any of which might influence choice. 72 Specific emotions are problematic and may be directly related to the outcome of the decision-making process; chief among these are fear and anger, which are closely related at the biological level. Fear and anger might overwhelm cognitive processes and bias the judgment of decision makers. This has interesting implications in international relations for the use of threats—as states frequently use threats to deter unwanted actions by others. Here is what Neta Crawford has to say about the use of threats:

The Individual Level of Analysis, Part II

65

Threats that produce fear instigate perverse cognitive effects . . . fearful actors tend to make less careful decisions that rely on cognitive heuristics rather than integrative complexity. In addition, the already fearful tend to gather information that is threat focused, thereby intensifying their fear in a feedback loop. Fearful actors also tend to emphasize the short term rather than the long term, even if this short-term bias is counterproductive. The careful weighing of costs, risks, and benefits . . . will become much more difficult if decisionmakers are fearful. Further, when threatened, decisionmakers may reasonably consider the threatener an enemy, slipping into cognitive schemas and scripts that attribute malevolent intentions to the other, rather than carefully evaluating the circumstances that may have motivated the other’s behavior. 73

Moreover, prior experience of fear affects later emotional and cognitive reactions to stimuli from the environment, as the brain learns to be fearful. Thinking about a prior experience of fear may cause decision makers to feel more fearful in the present. 74 Under these circumstances, it is quite likely that the response of national leaders to threats is an emotional response of resistance and bellicosity. The direct implication is that the use of deterrent threats to induce caution and restraint by the target is likely to backfire. Emotional states of high anxiety generally lead to counterproductive decision processes: reduced information gathering and processing, and the avoidance of information that might make one feel worse. In extreme cases, anxiety may lead to a complete shutdown of one’s decisional processes. This was apparently the case for Stalin in the immediate aftermath of Hitler’s attack on the Soviet Union in June 1941. 75 (Although, certainly, the emotional state of anxiety was not the sole reason for Stalin’s psychological collapse; personality issues were clearly involved as well.) On the other hand, moderate fear and anxiety might actually stimulate people to work hard to solve difficult problems; like a moderate amount of stress, it “concentrates the mind.” Fourth, emotion can affect the perception and calculation of risks. People in negative moods tend to overestimate the chance that a negative event will occur and underestimate the chance of a positive event. Conversely, people in positive moods tend to overestimate the chances of good things happening and underestimate the chances of bad things. 76 A positive mood state actually increases aversion to risk, especially in situations in which a significant loss is highly likely. (Happy people have more to lose, including the positive mood itself.) 77 As Vertzberger explains, although moods are rarely decisive, they may indirectly impact one’s assessment of risk by influencing judgment. “When the risk is ambiguous or low and known, people in a positive mood are more likely to take the risk because their mood produces optimism about the outcome. But when the risk is high, they are more likely to be conservative and risk averse to protect their mood.” 78 Fifth, previous emotional relationships between groups or countries may influence how we assess the intentions of others. In a relationship between countries A and B that has been characterized by fear, hatred, and antipathy, leaders of both states are more likely to negatively evaluate the intentions and motives of others than if the relationship had been warm and fuzzy—especially if the actions were ambiguous. 79 The negative evaluation is not entirely cognitive and rational; it is partly emotional. (This is one of the reasons why relations between pairs of states called enduring rivals are so prone to conflict and violence, as we will see in chapter 7.)

66

Chapter 3

Let us now explore some other aspects of the cognitive revolution in political psychology and return to Kahneman’s ideas about system 1 and system 2. THE ROLE OF BIAS: HEURISTIC THEORIES Cognitive psychology is the subdiscipline of psychology that deals with how our minds are hardwired to function: how mental processes allow us to think, to perceive, to remember, to process information, and to solve problems. It is especially important to our understanding, therefore, of how political leaders make decisions for war and peace. In the last forty years we have witnessed a cognitive revolution, as research has led to a fundamental reassessment of how our minds work. The results have critical implications for assumptions that national decision makers (like people in general) act “rationally.” 80 Indeed, Janice Gross Stein has argued that the deviations from rationality uncovered by cognitive psychologists and neuroscientists “are so pervasive and so systematic that it is a mistake to consider rational models of choice as empirically valid in foreign policy analysis.” 81 Policy makers live in a difficult cognitive environment. They are constantly confronted with complex problems derived from an equally complex global environment. Luckily (perhaps), the brain operates to simplify this complexity for policy makers, and these complexityreducing methods are not random; they operate in patterned and predictable ways. As John Steinbruner contends, the human mind “constantly struggles to impose clear, coherent meaning on events.” 82 In order to enhance the efficient processing of information, many of the mental operations performed by humans are actually automatic and subconscious. 83 In sum, here are some things that cognitive psychology tells us: 1. Our brains have a preference for simplicity. 2. We have a tendency to use heuristic shortcuts—such as reasoning by analogy—to help us solve difficult problems. 3. Humans are inherently poor estimators of probabilities. 4. Our attitudes toward risk are not necessarily balanced or neutral; they are biased toward risk aversion. 5. We have a tendency toward cognitive consistency and cognitive stability; we resist inconsistent images and try to maintain our current beliefs and images against significant change. 6. We are predisposed to make causal inferences to explain what has happened and why, and these inferences tend to be biased in predictable ways. 7. The images and beliefs that we hold in our minds tend to be organized and internally structured around certain strong central beliefs. 8. Images and beliefs do not necessarily correspond to reality; our perceptions of the world are subject to bias. The question is simply the degree to which our images are accurate reflections of the real world. We will explore most of these in more detail in the following sections.

The Individual Level of Analysis, Part II

67

We should note at the outset that many of the tendencies and systematic biases identified above are derived from what Kahneman refers to as our brain’s system 1 components. Kahneman’s conception of system 1 and system 2 are summarized in table 3.1. Biases in Information Processing How are political leaders to cope with the complexity of making decisions in the international arena? Cognitive psychologists have identified several heuristics or “rules of thumb” that policy makers use as convenient shortcuts in processing information, in assessing probabilities, and in forming conclusions. These shortcuts represent systematic biases (i.e., they are not random) that help explain why people deviate from standards of rationality. Importantly, these biases are not consciously chosen; we use them simply because our brain’s system 1 components are wired to work this way. For example, people tend to interpret new information in terms of what is most easily available to them from their own mental repertoire. Thus, we tend to predict the chances of something happening based on the ease with which we can think of examples of it. Instances of things that are unusual, dramatic, salient, or based on our own personal experiences are

68

Chapter 3

more readily available to us, and we tend to think therefore that these types of events are more frequent than they really are. This is called the availability heuristic. 84 Not only are potential events judged on the basis of how available they are to our thinking, they are also judged on the basis of the emotional charge they generate. (Consider the emotional impact of the vivid events of the September 11 attacks on the United States and the subsequent anthrax letter attacks.) As a result, we tend to distort the frequency of these or similar events and the probability that they will happen in the future. For instance, frightening thoughts and images of potential threats like a terrorist-induced smallpox epidemic or an Iraqi attack on the United States with nuclear weapons—both highly improbable events when judged objectively—can precipitate intense reactions because images of these two events are both easily available to us and are highly charged emotionally. As Kahneman says, “Terrorism speaks directly to system 1.” 85 A related bias is referred to as imaginability: We tend to “call up” information that is plausible without regard for real-world probabilities. For instance, whether we deem a particular event to be likely depends on the ease with which we can think of examples of it in our own recollection. The upshot of this is that the real risks involved in a particular scenario may be drastically underestimated if potential pitfalls are simply difficult to conceive of—as using commercial airliners as guided missiles was before September 2001. Conversely, the risks of a particular event may be vastly overrated simply because it is easy for us to imagine. 86 In general, we are intuitively very poor estimators of probability. The limited things that we know and remember have a “disproportionate impact on our forecasts.” 87 Several biases concern how we explain to ourselves the behavior of others. The first is called the egocentric bias: We tend to exaggerate the chances that the actions of others are due to our own previous actions and behaviors. And we also tend to overestimate the extent to which we are the targets of the actions of others. For example, officials in the Bush administration tended to think that to the extent Iraq had WMDs, they were intended for use against the United States. In reality, while Saddam Hussein did not possess WMDs in 2003, he attempted to maintained the fiction that these weapons still existed, not to deter the U.S. but to deter potential aggression by Iran—a motive that U.S. leaders failed to consider as a possibility. We also tend to vastly overestimate the degree to which the behavior of others is planned, coordinated, orchestrated, and centralized because we consider our own behavior to be planned. In reality, the behavior of others is frequently the result of coincidence, accident, or ad hoc, spur-of-the-moment reactions—or sheer stupidity. 88 As a result, we tend to see conspiracies where they do not exist. Another bias is called the fundamental attribution bias, which determines whether we attribute the actions of others to their inner dispositions (their desires, goals, ideology, psychological traits, personalities) or to situational factors (constraints produced by the domestic or international environments). 89 Political leaders are likely to interpret the negative behavior of rival states as due to the malevolent or aggressive intentions of its leaders; but the same state’s positive acts are likely to be attributed to “circumstances beyond their control.” As D’Anieri says, the result of this bias is that “actors are more likely to mistake a peaceful state for an aggressive one than vice versa.” 90

The Individual Level of Analysis, Part II

69

The fundamental attribution bias is reinforced by what is known as the halo effect. The halo effect refers to the tendency to either like or dislike everything about an individual, group, or country. If we know a positive or negative fact about the person or country, we tend to view other attributes of that entity in a similar way—even though we lack any information about these other attributes. This simplifies our images of others and exaggerates the consistency of the image. As Daniel Kahneman says, “Good people do only good things, and bad people are all bad.” 91 Friends are seen in a positive light and get the benefit of the doubt when we evaluate their intentions; we attribute their positive behaviors to dispositional factors and we attribute their negative behaviors to the situation. The opposite applies to those we dislike: their bad behavior is attributed to dispositional factors and their good behavior merely to the situation. 92 Another source of bias comes from the fact that system 1 is predisposed to search for causal explanations for things. Our mind automatically attempts to construct and impose causal explanations for phenomena in our environment. This is largely due to our need for coherence, and system 1 is pretty good at finding a coherent causal explanation to explain various fragments of knowledge we possess. This is done quickly, intuitively, and without conscious direction. Kahneman asks us to look at this sentence: After spending a day exploring beautiful sights in the crowded streets of New York, Jane discovered that her wallet was missing. 93

Without even asking ourselves the question “What happened to Jane’s wallet?” most of us will automatically conclude that her wallet was liberated by a pickpocket. Our mind has causal intuitions that kick in, even without any real evidence, that allow us to make any kind of causal explanation. As Kahneman says, “System 1 is prone to jump to conclusions.” 94 System 1 may get these causal relationships correct most of the time, but the notion that our minds constantly make these important causal connections—even in crisis situations of high complexity, low information, when errors could be catastrophic—suggests the need to closely monitor our own reactions and to consciously engage in more rigorous analytical thinking. The mind’s search for causal explanations is linked to its desire for coherence, which is another form of bias. As system 1 constantly monitors our environment, forming impressions as it goes, it links together in our mind things that appear connected. This aspect of our cognitive processes is called associative memory. System 1’s intuitive response to external stimuli sets off a cascade of associated ideas in a process called associative activation. This happens instantaneously, and very little of it is conscious; the mind simply imposes coherence on events. Coherence is important. We are strongly predisposed to believe that something is true if it is linked by logic or by association to other beliefs we hold (or from a source we trust). Kahneman suggests a frightening corollary of this process: “Anything that makes it easier for the associative machine to run smoothly will also bias beliefs. A reliable way to make people believe falsehoods is frequent repetition, because familiarity is not easily distinguished from truth.” 95 Repetition of ideas (whether true or false) pushes us toward belief simply because repetition fosters cognitive ease and a sense of familiarity. This is called the mere exposure effect. A related problem exists. Consistency of information matters much more than

70

Chapter 3

completeness: Knowing relatively little about something actually makes it easier to organize everything we know into a coherent pattern. 96 All of these tendencies and biases are exacerbated by several other characteristics of system 1. First, since system 1 operates automatically and cannot really be turned off, it is difficult to prevent errors associated with its intuitive analysis of the environment. Second, system 1 is pretty sure that its intuitive judgments are right. The result is that people tend to be overconfident and place too much faith in their intuitions (see George W. Bush about this). As Kahneman says, “System 1 is not prone to doubt. It suppresses ambiguity and spontaneously constructs stories that are as coherent as possible. Unless the message is immediately negated, the associations that it evokes will spread as if the message were true.” 97 System 2 is in charge of doubt—but only if it is activated. An essential hallmark of system 1 is that it is “gullible and biased to believe.” 98 This happens in a number of ways. We tend to look for (and uncritically accept) information that supports our existing beliefs. This leads to a kind of confirmation bias. In addition, Kahneman argues that “system 1 understands sentences by trying to make them true, and the selective activation of compatible thoughts produces a family of systematic errors that make us gullible and prone to believe too strongly whatever we believe.” 99 Without the active intervention of system 2 to process and analyze doubts, “we will automatically process the information available to us as if it were true.” 100 All of the above suggests that when political leaders meet in their oval offices or their “war rooms” they would be well-served by understanding that, as they evaluate their environments for threats and risks, their assessments are unlikely to be free from error and bias. “Lessons of History,” Historical Analogies, and Schema As we have seen, one of the principles of cognitive dynamics is that the human brain attempts to impose simplicity on the complexity of the world as a way of helping us cope with our environments. One way to simplify situations is for policy makers to match them to previous situations from which they can apply the “lessons” that have been learned. Thus, we tend not to see current events as unique but as just new versions of well-known and well-understood past events. This is just another shortcut that our brain is hardwired to perform. Leaders frequently use past events as cues or guides to aid thinking about present problems. Jervis notes that some events, like wars and revolutions, make such an impression on individuals that it takes equally dramatic developments to nudge them aside. The result is that images of past events hover like Banquo’s ghost over our attempts to understand the present. 101 Historical analogies and simple images (such as seeing the countries of Southeast Asia as “falling dominoes” or Iraq as a “rogue state”) provide an internal anchor around which ambiguous information can be structured and given meaning. They provide the individual with an elegant method for reducing the uncertainty inherent in complex situations. 102 Reasoning by analogy may provide a shortcut to understanding, but it is also fraught with danger. As Janice Gross Stein notes, political leaders “unconsciously strip the nuances, context and subtleties out of the problems they face in order to build simple frames. When they look to the past to learn about the future, political leaders tend to draw simple one-to-one analogies without qualifying conditions.” 103 Political scientists have been pretty critical of

The Individual Level of Analysis, Part II

71

analogy users. Ernest May’s Lessons of the Past finds that the misuse of historical analogies is practically endemic in the making of U.S. foreign policy, and Snyder and Diesing’s examination of international crises from 1898 to 1973 actually finds no examples in which analogies were able to produce “a correct interpretation of a message” for statesmen involved in a crisis. 104 Historical lessons are most important for those who experienced them firsthand in their formative years and for those whose adult lives and careers were significantly affected by the original events. Jervis mentions that three of the British foreign secretaries involved in the policy of appeasing Hitler in the 1930s—Hoare, Simon, and Halifax—had made their political reputations by appeasing Indian demands for greater self-government. The lessons learned in one environment were then inappropriately transferred to another. 105 Lessons also tend to have generational effects. American leaders who came of age in the 1930s and 1940s were generally guided by the Munich analogy; those who came of age in the 1960s were mindful of the “lessons of Vietnam” (however they came to interpret them); those who will come of age in the early years of the twenty-first century are likely to be influenced by “the lessons of Iraq” and the “lessons of September 11.” 106 Lessons learned early in one’s political career, especially lessons learned from crucial early successes or failures, are naturally important aspects of one’s overall image. President Reagan’s first political job and his first political success came as president of the Screen Actors’ Guild, where he was involved in a fight to prevent groups aligned with the Communist Party from taking over the union. His image of communists as devious and expansionistic was formed from this early experience. 107 Stalin spent the formative years of his life as an underground, conspiratorial political operator in the Caucasus, organizing bank robberies and other activities for the Bolsheviks. It was certainly in this cauldron of revolution and banditry that he learned the art of political intrigue, the importance of ruthlessness in pursuit of political survival and advancement, but also the importance of hiding his intentions, reducing the power of his rivals, organizing for action, and the necessity of a very fined-tuned antenna for the detection of potential enemies. 108 An additional consideration is that these historical lessons are not only important at the individual level; they can become institutionalized within government bureaucracies. Once this has occurred, they may form the basis of future planning, becoming permanent features of standard operating procedures, creating preferred frameworks for viewing events or preferred options for dealing with contingencies. 109 The Munich analogy—of Hitler’s expansion in Europe—is the classic example of an allpurpose historical analogy, used as a guide by American leaders in virtually every crisis from Korea to Vietnam to the Persian Gulf. The general axiom that appeasement of aggressors should always be rejected (and that the proper response to aggression should be immediate and forceful resistance) came to be applied indiscriminately, in a way that ruled out compromise through diplomacy. The image of Hitler’s bullying of Europe was etched in the memory of those on both sides of the Atlantic who lived through the 1930s and 1940s. Truman and his advisers saw communist activities in Greece, Turkey, and Iran immediately after World War II as analogous to Hitler’s step-by-step aggression in Europe in the 1930s. American diplomacy in the 1940s and early 1950s was aimed at preventing the recurrence of World War II.

72

Chapter 3

Moreover, Truman saw the North Korean attack on South Korea in 1950 in the light of analogies from the 1930s. Truman wrote in his memoirs of his thoughts on the plane from Missouri to Washington prior to his meeting with his advisers on the Korean crisis: I had time to think aboard the plane. In my generation this was not the first occasion when the strong had attacked the weak. I recalled some earlier instances: Manchuria, Ethiopia, Austria. I remembered how each time that the democracies failed to act it had encouraged the aggressors to keep going ahead. Communism was acting in Korea just as Hitler, Mussolini, and the Japanese had acted ten, fifteen, and twenty years earlier. . . . If this were allowed to go unchallenged it would mean a third world war, just as similar incidents had brought on a second world war. 110

President George H. W. Bush was a veteran of World War II for whom the memories of Nazi aggression were part of the living past. (He was the last U.S. president of this generation.) It is not surprising that he saw Iraqi President Saddam Hussein’s brutal 1990–91 annexation of Kuwait as Nazi-like aggression and Hussein himself as Hitler. In fact, Bush was quick to make these connections and to argue that Iraqi aggression “should not stand.” Iraq’s aggression should be met forcefully, by war if necessary, lest Saddam Hussein’s success lead him to attack Saudi Arabia or other Persian Gulf states, thus controlling the world’s most important resource. Critics were quick to point out that Saddam Hussein was not Hitler, that Iraq was not Germany, and that the threat to American interests was nowhere as great as they had been in 1941. Nevertheless, what mattered was the association made between the two situations in the president’s mind and the strategies that these associations triggered. Lessons of the past also played a major role in the Kennedy administration’s 1961 decision to make a major commitment to prevent the fall of South Vietnam to communism—arguably the most important U.S. decision regarding that war. 111 Ernest May recounts how the Kennedy administration rejected one set of lessons—those from the Korean War—and accepted another set from the Philippines and Malaysia. The lessons of Korea, as seen less than a decade later, were twofold: the United States should never again fight a land war in Asia, and the American people were unlikely to support a prolonged, limited war. In retrospect, this should have been a powerful analogy, and in fact it was—for the U.S. military. Kennedy and his civilian advisers, Robert Kennedy, Dean Rusk, and Robert McNamara, preferred, however, to apply the lessons learned in Ramon Magsaysay’s fight against the communist Hukbalahap rebels in the Philippines and by the British in their fight against insurgents in Malaya—both examples of the successful use of specialized, small-scale counterinsurgency operations. 112 May’s point is that leaders often apply analogies superficially, indiscriminately, and inappropriately. In fact, Gen. Lyman Lemnitzer, the chairman of the Joint Chiefs of Staff, listed five rather impressive reasons why the Vietnamese and Malayan situations were not comparable—reasons which, in retrospect, are worth listing, even though they were thoroughly ignored at the time: 1. Malayan borders were far more controllable. 2. The racial characteristics of the Chinese insurgents in Malaya made identification and segregation a simple matter compared with the situation in Vietnam.

The Individual Level of Analysis, Part II

73

3. The scarcity of food in Malaya versus the relative plenty in South Vietnam made the denial of food to the guerrillas a far more important and readily usable weapon in Malaya. 4. Most important, in Malaya the British were in actual command. 5. Finally, it took the British nearly twelve years to defeat an insurgency that was less strong than the one in South Vietnam. 113 The Malayan counterinsurgency analogy continues to crop up in American foreign policy circles, most recently as a frame through which to view U.S. strategy for nation-building attempts in Iraq and Afghanistan. Since the lessons of the last war play such an important role in creating images of international relations, leaders frequently believe that the next war will have the same causes as the last one and they prepare accordingly. Appeasement in France and Britain in the 1930s was at least in part based on the notion that the previous war (World War I) could have been avoided by conciliatory diplomacy. The Soviet demands on Finland in 1939 for territory near Leningrad in the Baltic were based in part on lessons from the past, both general and specific. One of the most powerful lessons of history for Russian leaders has been that security depends on the creation of buffer states. This is a lesson learned from years of invasion—by Genghis Khan’s Asian hordes, by the Swedes, Lithuanians, Teutonic knights, the Poles, the French, the Germans, and others. Each invasion taught that security depended on the ability to create friendly buffer zones to keep hostile neighbors at arm’s length, to create a set of more defensible borders than the Russian plains allowed, and, most important, to ensure that when war did break out, it could be fought outside the Russian homeland rather than on its soil. More specifically, incidents in the Russian civil war (1918–21) in the Baltic had given the Soviets an inflated perception of the importance of that region. The White leader, Yudenich, and the British had both been active in the Gulf of Finland area, an area the tsars had been able to protect with their naval base at Porkkala, in what was in 1939 independent Finland. The Finns, not having had the same historical experiences as the Russians, failed to understand the defensive motivation of the Soviets and believed instead that their demands were aimed at destroying the Finnish state. 114 Needless to say, the matter failed to be settled through negotiation. In the ensuing Russo-Finnish War the Russians were taught certain military lessons about fighting winter wars by the Finns, and the Finns were taught certain political lessons about accommodating the security interests of Great Power neighbors. The decision of George W. Bush and his advisers—especially Vice President Dick Cheney, Secretary of Defense Donald Rumsfeld, and National Security Adviser Condoleezza Rice—drew on lessons presumed to have been learned from both the earlier Gulf War and the events of September 11. 115 The combined lessons of September 11 were that rogue states with weapons of mass destruction and ties to terrorist groups were an extreme threat, that if even a small chance existed that a rogue state like Iraq might furnish terrorists with a WMD, the chance ought to be treated as a virtual certainty, and that time was not on our side. One of the presumed lessons of the first Gulf War (and the postwar inspections) was that fairly advanced nuclear weapons programs could be successfully hidden from the world community. The

74

Chapter 3

application of these lessons to Iraq in 2002–3 predisposed Bush administration officials toward the preventive use of force to prevent a second catastrophic attack on the United States. Yuen Foong Khong’s examination of the Johnson administration’s Vietnam decisions in 1965 is probably the gold standard for studies of the use of historical analogies. He asserts that analogies help decision makers perform six crucial tasks when faced with a policy problem (like whether to use force or not). Analogies (1) help define the situation, (2) help assess the stakes, and (3) provide potential policy prescriptions. These first three tasks are strongly connected, with the answer to number one largely controlling the content of the second and the third. The last three tasks help leaders to assess the potential options by (4) predicting the chances of success, (5) evaluating their morality, and (6) warning about potential dangers associated with the options. 116 In this way, Khong argues, analogies actually play a causal role in policy making, making a major impact on the selection and rejection of possible policy options. 117 What sets Khong’s study off from other examinations of the use of historical analogies is that he explicitly connects analogies to cognitive psychology’s schema theory. Schema theory starts from the basic assumption that human beings have limited computational capacities; we are therefore frequently referred to as having “bounded rationality” or being “cognitive misers.” To cope with the massive information flow that confronts us on a daily basis we rely on simplifying mechanisms to help us code, interpret, explain, simplify, store, and recall this information and to generally make sense out of our environment. Schemas help us do this. A schema is a “pocket” of information about particular concepts or things, based on our previous experiences. We may, for instance, have pockets into which we deposit our beliefs about nuclear weapons and nuclear deterrence, about dictators and rogue states, about Iran, about WWII and the Munich crisis, and about terrorism. These pockets contain interrelated and interdependent bits of knowledge, ideas, and beliefs that are structured somewhat hierarchically, with some general core beliefs connected to more peripheral specific beliefs. While each pocket of information contains information that is hierarchically arranged (top to bottom), a particular schema may or may not be linked to any other specific schema horizontally. So, an individual may have numerous schemas related to foreign affairs, but these schemas may be “atomized” and unrelated; thus, the individual may not have a coherent belief system about foreign affairs. 118 Indeed, inconsistent beliefs and pieces of information may be stored in different schemas. Analogies are functionally and structurally very similar to schemas. 119 When confronted with new and ambiguous information, decision makers code it in terms of a schema found in their memory banks. Analogies, like schemas, allow individuals to make inferences about missing information. We interpret ambiguous information in a way that is consistent with the analogy in a top-down way. We tend to “make the data fit the schema instead of the other way round.” 120 For instance, we may have a schema devoted to the concept of dictators in rogue states. In this schema are various attributes of dictators: they’re males, they repress their populations, they hold phony elections, they erect statues and portraits of themselves throughout their countries, and they are aggressive toward their neighbors. When we meet a new member of this species “dictator” we may be able to determine that indeed all the attributes are present

The Individual Level of Analysis, Part II

75

except for the last one: we do not yet know if he is aggressive toward his neighbors. However, the presence of the schema or analogy permits us to fill in the blanks. This is what cognitive psychologists refer to as the default value of schemas. For missing or unambiguous information, we simply fill in the rest of the picture from our more general understanding of the phenomenon (much like the halo effect). For example, neither Truman nor anyone else in June 1950 knew what the result would be if the United States chose not to intervene to resist North Korean aggression. To fill in this hole in their knowledge, the Munich analogy provided Truman and his advisers with the default value: Appeasement would not work; successful aggression would increase the appetite of the aggressors for further gains, and we would soon have to fight a larger war to counter the next act of aggression. 121 It should be emphasized that policy makers will possess a repertoire of analogies (or schemas) that may be called up when they face novel situations that require explanation and analysis. Several dueling analogies were at play in the discussions in the Johnson administration in 1965 concerning Vietnam. The old standby Munich analogy (and other analogies related to fascist aggression in the 1930s) was much in play and widely accepted; it was never challenged in the administration’s early discussions. The Malaya analogy, suggesting the possibility that Western states could defeat guerrilla insurgents in a nonconventional war, was also present. But the dominant analogy had to do with the lessons of Korea. The Korean War analogy was especially important to President Johnson and Secretary of State Dean Rusk. Its dominance in part was due to the fact that it was able to perform all six of the tasks singled out by Khong. First, it defined the situation in Vietnam as one of external (North Vietnamese) aggression against the South—just as the Korean conflict was defined as North Korean aggression against South Korea. Second, it assessed the stakes as vital to the U.S., since permitting North Korea to succeed would erode U.S. credibility and lead to communist aggression against other states in the region, raising the potential of even more serious wars in the future. Third, the analogy suggested that the preferred policy response should be the use of military force to stop the aggression. More specifically the analogy led to the decision in February 1965 to use a graduated aerial bombing campaign (Operation Rolling Thunder) and then in July to send 100,000 U.S. ground combat forces. Fourth, the chances of success were seen as good, though—like Korea—not without substantial costs. Fifth, the use of force in response to aggression was seen as morally correct. Sixth, the primary danger of the proposed American use of force might be that if the U.S. pushed too hard, China could intervene—as it had in the Korean War. Thus, any acceptable policy had to avoid the more provocatively aggressive options on the table. 122 The primary skeptic in the administration, Undersecretary of State George Ball, asserted that the Korean analogy was inappropriate for Vietnam, and he offered the earlier French experience in Vietnam and its defeat at Dien Bien Phu as a more appropriate analogy. Looking at the situation in Vietnam through this lens would define the situation as an internal conflict and a war of independence rather than a war of aggression by one sovereign state against another. It would define the stakes as medium rather than high. It would categorize the morality of military intervention as dubious, and it would rate the chance of success as poor. (The French had lost to Ho Chi Minh’s Vietminh after a five-year effort, even though they had

76

Chapter 3

250,000 forces still in the field.) Although it would rate the danger of Chinese intervention as low, it would place the danger of domestic U.S. opposition to the policy as high. (The French pullout was largely in response to lack of public support for the war.) Most important, it would imply that the best course of action for the U.S. would be to cut its losses by withdrawal. Ultimately, Johnson and his advisers were not interested in the lessons of Dien Bien Phu; those lessons were associated with failure, and “policymakers were uninterested in learning from losers.” 123 Khong uncovers some interesting findings concerning the use of analogies by American presidents and their advisers. Most analogies at play—the Korean analogy, the Malayan analogy, the Dien Bien Phu analogy—were disputed during the policy discussions as being inappropriate and flawed. In spite of rather vivid dissection of the analogies, no one changed his mind. Policy makers remained attached to their favorite analogies. Schema theorists refer to this as the perseverance effect, and it has been thoroughly confirmed by cognitive psychologists as an effect of schemas. 124 Given that several potential analogies might be available to explain a particular situation, what determines which analogy is chosen? Cognitive psychologists have noticed a number of principles at work, which have been verified by Khong’s study of the Vietnam decisions. First, individuals use those analogies that are most easily recalled; this is the availability heuristic. Recent events are examples of phenomena that are readily available. Second, surface commonalities are important; people tend to use analogies that are superficially similar to (or representative of) the event in question—the representativeness heuristic. The Korean analogy was both easily available (the attack on Korea in 1950 was only a decade and a half distant and clearly remembered by all), and it certainly had surface commonalities to the situation in Vietnam—it was in the same geographic region, it was a Cold War struggle between ideologies, it involved a country divided north and south, with an aggressive communist north attempting to unify the south by force, with the backing of the two major states in the communist bloc, China and the USSR. 125 Virtually all of this research on historical analogies leads to the inescapable conclusion that senior government leaders make poor use of historical analogies. However, this is not because they are ignorant of history; after all, LBJ’s advisers were extremely bright men steeped in world history and experienced at making tough foreign policy decisions. They weren’t babes in the woods. Something more systematic is at work here, having to do with the very nature by which the human mind works. THE ROLE OF RISK BIAS: PROSPECT THEORY The evidence from a vast amount of research in cognitive psychology indicates that people systematically violate almost all of the basic principles found in rational decision-making models. This research has resulted in an alternative model of decision making, called prospect theory, pioneered by Daniel Kahneman and Amos Tversky. 126 At the base of this theory is the fact that certain identifiable biases exist in how we treat risk. Prospect theory is based on several principles that we will outline here. 127 These should be treated as operating characteristics of system 1. 128

The Individual Level of Analysis, Part II

77

1. Evaluation of choices is relative to a neutral reference point. People tend to think in terms of gains and losses rather than in terms of their net assets, and therefore they evaluate their choices in terms of deviations from a reference point. The reference point is usually the current status quo, but it might be an aspirational goal or an expectation level. The selection of this reference point is highly subjective and will vary from individual to individual. At any rate, when decision makers think about potential outcomes, they tend to rate them in terms of relative gains or relative losses from the reference point. Jack Levy states that this reference dependence is the central analytic assumption of prospect theory. 129 2. Losses loom larger than gains. People tend to value what they have more than comparable things they do not yet have. In other words, the perceived harm of relinquishing what one has is seen as greater than the benefit of acquiring an additional gain. As one analyst puts it, “losing ten dollars hurts more than finding ten dollars gratifies.” 130 This overvaluation of current possessions is called the endowment effect. 131 Furthermore, the perceived value of a good is proportional to the length of time it has been possessed and the effort and resources used to acquire it. Since losses loom larger than gains, a general tendency of choice is toward loss aversion. The endowment effect and the tendency toward loss aversion combine to produce a bias toward status quo choices. 132 According to Kahneman, loss aversion is clearly due to the emotional impact produced in system 1, and it presumably has an evolutionary history: “Organisms that treat threats as more urgent than opportunities have a better chance to survive and reproduce.” 133 3. Losses and gains are asymmetrical in terms of risk orientation. Individuals treat gains differently than losses: We tend to be risk-averse in our selection of options if we believe ourselves to be facing gains, and we tend to be risk-acceptant toward our options if we believe that we are facing losses—even if standard rational choice models would indicate a preference for caution. As Levy puts it, “A state might be willing, for example, to fight to defend the same territory that it would not have been willing to fight to acquire.” 134 Because of the role of reference points in defining gains and losses, and because a change in identifying the reference point can result in a change of preference, identifying the reference point—the framing of a choice problem—is of critical importance. It matters considerably whether one frames an issue in terms of gains or losses, and this can be a fairly subjective judgment—especially when “the situation involves a succession of choices and where there is ambiguity regarding the status quo.” 135 Since framing is subjective, it is also prone to manipulation. 136 Advisers to presidents and prime ministers are adept at getting their bosses to frame issues in a particular way. This is important because, as Jervis says, “People will choose the risky alternative when the choice is framed in terms of avoiding losses when, in the exact same case, they would take the less risky course of action if the frame of reference is the possibility of improving the situation.” 137 4. An important consideration in prospect theory is how individuals accommodate gains and losses—that is, how quickly and completely they come to accept gains or losses. (Accommodation involves re-setting or “re-normalizing” the reference point.) As we might expect, people accommodate to gains more rapidly than to losses—this is called the instant endowment effect. 138 On the other hand, decision makers are unlikely to accept losses and are likely

78

Chapter 3

to take risks to recover them. Instead of “cutting their losses” leaders are constantly tempted to regain sunken costs. The French and American ventures in Vietnam and the Soviet and American ventures in Afghanistan are cases in point. 139 One implication of the biases associated with accommodation is that after a change in the status quo, both sides are likely to engage in more risk-taking than one might expect based purely on rational calculation, due to the way they frame the situation. Both sides perceive the status quo is on their side and are willing to take risks to preserve it. If A makes a gain at the expense of B, A quickly readjusts its reference point to accept the gain (remember the instant endowment effect). Meanwhile, B is unlikely to accommodate to the change and sees the old situation (pre-loss) as the legitimate status quo. (The endowment effect doesn’t apply to losses.) Thus both find themselves protecting their version of the status quo. Both will frame the situation as one in which they are faced with possible losses and will be inclined toward more risk-seeking. As Jervis points out, a central implication of prospect theory is that war is more likely when decision makers on each side believe that they are defending the status quo and therefore believe that they will suffer significant losses if they do not fight to defend it. Wars are presumed to be more commonly motivated by the desire to prevent perceived losses than to attain future gains. Fear, not desire for expansion, is the most significant motivator. Presumably, the tendency toward loss aversion is a force for international stability since states will not take risks to increase their positions relative to the status quo, but stability only exists if all relevant states identify with the same status quo. If leaders have references points above the status quo, then maintaining the status quo will be rated as a loss. 140 Prospect theory tells us that people are not very good in dealing with probabilities. Several tendencies appear. At the extremes, we tend to treat highly probable events as being completely certain, and things deemed to be highly unlikely are treated as being impossible and are thus ignored. We tend to overweigh outcomes that appear to be certain compared with those that are merely probable. This is the certainty effect. 141 This is important because when leaders perceive certain losses, prospect theory predicts they are likely to engage in high-risk behavior to avoid these losses. Faced with a choice between a certain loss and a gamble that could reduce the loss (or make it worse), we tend to choose the riskier alternative. As Levy and Thompson explain, “People are so eager to avoid certain losses, or ‘dead losses’, that they take enormous risks that often result in even greater losses, even though the expected value of the gamble may be considered worse than the value of the certain loss.” 142 Leaders who accept even limited defeats internationally are likely to be punished politically at home. Risking an even greater loss in order to prevent or overturn an initial loss might be irrational from the standpoint of national interests, but it might be rational for the individual politician to whom even a small loss might mean personal disaster. Jervis suggests that “domestic public opinion operates according to prospect theory in finding even small losses so painful that it prefers high risks to accepting them and will punish any leader who permits them to occur.” 143 A related implication of prospect theory has to do with diversionary wars. Leaders don’t always define gains and losses in terms of the state’s international influence or security; they also define them in terms of the domestic political situation. As Levy notes, the temptation to

The Individual Level of Analysis, Part II

79

undertake diversionary actions abroad might be “enhanced by risk-acceptant attitudes in the domain of losses created by a deteriorating domestic situation.” 144 Levy suggests that a combination of internal political insecurity and a relative decline in a state’s external relations can be conducive to extreme risk-taking. This was apparently the situation facing Jimmy Carter when he made the decision to attempt to rescue American diplomats held hostage in Iran. 145 Another tendency is that low-probability/low-risk events are either ignored altogether or they are overweighted compared with events with high or medium probabilities. 146 That is, events that are not deemed to be very likely are paradoxically given more importance than they deserve, especially if coupled with vivid images of impending doom. This is called the pseudocertainty effect: An event that is uncertain is weighted as if it were certain. It is easy to see how the pseudocertainty effect can lead to analysis based on worst-case-scenario thinking. 147 This almost perfectly describes the post–September 11 environment in the Bush administration. Vice President Dick Cheney had convinced many members of the administration that if very low probability events (such as a rogue state like Iraq developing nuclear weapons or other WMDs and then transferring them to terrorist groups like al Qaeda, which would then use them against the United States) could have a highly negative impact (destruction of a major American city), then the probability ought to be treated as a certainty. 148 In sum, prospect theory provides a probabilistic explanation of decision choice: leaders will show a bias toward risky options under certain circumstances—the perception that they are facing losses. The theory does not mean that unprovoked aggression never occurs, nor does it deny that sometimes leaders take actions to change the status quo. As Jervis notes, Actors will take some risks to improve their situations even if they are risk-averse for gains. But, the theory argues, actors should be much more willing to run risk when they believe that failing to do so will result in certain losses. Wars will then frequently be triggered by fear of loss; fighting in order to act on opportunity will be relatively rare. 149

Applications of Prospect Theory There have been numerous applications of prospect theory, mostly in the form of case studies or comparative case studies. We will mention a few so that you get the flavor of the argument. Audrey McInerney uses prospect theory to explain Soviet policy in the Middle East in the months preceding the Six Day War in 1967. 150 While Soviet policy in the area was generally risk-averse, by 1966 a major change had taken place. The neo-Baathist coup had brought a radically left-wing government to power in Syria that was much more aligned with Soviet ideological interests than any other government in the area—a gain for the Soviets that they quickly incorporated into the new status quo. However, that government was threatened on two fronts: it had serious internal political unrest, and the continued cross-border raids against Israel threatened the possibility of a war with Israel that would almost certainly mean an Israeli victory and the demise of the neo-Baathist regime. The risk of doing nothing was that the Soviet client regime in Syria would fall. This would certainly be framed as a loss. In order to prevent this loss, the Soviets undertook extremely risky behavior. They promoted a “war scare” by transmitting false intelligence information to Egypt and Syria that Israel was planning an invasion of Syria. The intent was to get Egypt to take strong action to back Syria and

80

Chapter 3

thereby deter any possible Israeli attack. The Soviets did not desire war, but they did desire to shore up the shaky government in Damascus—even if that meant the risk of war. Ultimately, Soviet risk-taking backfired: it brought about the very war they had hoped to prevent. The result was that two Soviet clients—Egypt and Syria—were both defeated, although they escaped with their regimes intact. McInerney notes that Soviet leadership was divided on these risky actions toward Syria. There was one faction—led by Prime Minister Kosygin and Defense Minister Gromyko—that was more cautious. This faction accepted an older status quo in which the Soviet Union was happy with a less radical Syrian regime that was less close to Russia; they therefore saw no need to undertake risks to achieve further gains. However, the Kremlin majority framed the situation within the realm of losses, not gains. and were therefore willing to accept significant risks. 151 Charles Sullivan examines the Soviet intervention in Afghanistan in 1979 to protect the local communist government there from Islamic mujahideen rebels. The Soviet leadership was clearly unwilling to send military forces in the spring of 1979 in response to requests from the recently installed communist government. However, before the year was out they had changed their minds. What accounts for the policy shift? Sullivan argues that it was not that the situation on the ground in Afghanistan had changed; the Soviets were still convinced that the communist government could prevail over the insurgent mujahideen forces. What had changed was the internal political situation. The communist movement in Afghanistan (the PDPA) had always been split between the moderate Parcham faction (headed by Babrak Karmal) and the more radical Khalq faction. The Khalq had come to dominate the regime in Kabul, and its most radical leader, Prime Minister Hafizullah Amin, had precedence over the more moderate President Nur Mohammad Taraki. The Soviet leadership apparently pushed Taraki to overthrow Amin, but the process backfired, and in October 1979 Taraki was killed by Amin. Soviet leaders saw Amin as radical, unpredictable, unreliable, and uncontrollable— sort of like an Afghan version of Egyptian President Anwar Sadat. Many Soviets believed that Amin, like Sadat, was attempting to reorient Afghan policy toward the West, moving out of the Soviet sphere of influence. This change in the situation brought about a change in frames. In the spring, Soviet leaders had looked on the Afghan situation as a gain (a recently installed, pro-Soviet communist party regime) and were thus reluctant to support a clearly risky policy of military intervention. With the unpredictable Amin now in control, a small group of important policy makers (a troika of KGB Chief Yuri Andropov, Defense Minister Dmitry Ustinov, and Foreign Minister Andrei Gromyko) now saw the USSR in the domain of potential losses—with the result being that the top leadership was now much more risk-acceptant. The troika was able to convince Party Secretary Brezhnev to go along, and the decision to intervene was essentially made over the heads of the other Politburo members. 152 Janice Gross Stein’s analysis of the 1973 Yom Kippur War in the Middle East focuses on the framing of the situation by Egyptian President Anwar Sadat. Even though Israel won the Sinai peninsula and the Gaza strip from Egypt in the 1967 Six Day War, Sadat never accommodated himself to this changed territorial situation; the loss was never accepted or normalized. Thus he was highly motivated to engage in a risky war with Israel to return the Sinai to

The Individual Level of Analysis, Part II

81

Egypt. A war for the Sinai would not be a war to acquire something new but to ensure that Egypt would not lose something it already possessed. Israeli leaders—unaware of the powerful effects of framing for losses—did not sufficiently understand the intense motivation on the part of Egyptian leaders and of the risks they were willing to take. They assumed that Egyptian leaders would be “rational” and refrain from war against a state that had clearly shown its military superiority and its willingness to defend itself in the past. 153 Let’s look at one final application. Jeffrey Taliaferro melds prospect theory to defensive realism (which we will address in a later chapter) to create what he calls a balance of risks theory. He is concerned with explaining the behavior of great powers in geopolitically peripheral areas—especially their propensity to engage in risky behaviors and then to maintain these behaviors even in the face of evidence that they are not working. He is also aware that there are examples of leaders deciding not to engage in risky behavior or to disengage from risky policies. He derives several hypotheses from prospect theory and defensive realism and then examines them in several historical cases. Taliaferro suggests that in “ill-structured and complex settings” like those typically found in international affairs, leaders (both individuals and small groups) are “more likely to assess outcomes and contingencies in terms of deviations from an expectation level, not a neutral reference point.” 154 Senior leaders’ perceptions of the relative power trends between their country and that of rivals have an impact on these expectation levels. He hypothesizes as follows: 1. If officials foresee a decline in relative power or status over time, they are more likely to adopt a more favorable international environment as their expectation level. Conversely, if they anticipate a relative increase in power and status, they are more likely to adopt the status quo as the expectation level. 2. Officials’ ability to revise their expectation level in response to adverse outcomes will be directly proportional to the length of time they adhere to a particular level. 3. Officials are more likely to initiate or persevere in risk-acceptant strategies in the periphery to avoid perceived losses. 4. Officials will likely continue and even escalate their commitment to risk-acceptant but failing intervention strategies in the periphery. 155 Taliaferro examines several pairs of decisions made by three great powers. For each country a decision accepting risk is compared with a situation in which risky options were rejected: Germany in the 1905 Morocco Crisis (and the 1911 Agadir Crisis); Japanese leaders’ war decisions of 1940–41 (and the decisions in the Changkufeng and Nomohan incidents in 1938 and 1939), and United States in the Korean War in 1950 (and again in 1951). In each set of cases the senior leaders’ framing of events changed over time and their policies reflected these changes. Let’s take just one example. The North Korean attack on South Korea in June 1950 challenged both the global status quo and the Truman administration’s post-WWII policy in Korea, which had treated the area as strategically nonessential. Faced with a temporary window of vulnerability at the global level and losses to U.S. power and reputation in Asia, Truman and his advisers set their expectation level at the territorial status quo prior to the war’s outbreak (two Korean states

82

Chapter 3

divided at the 38th parallel) and chose a high-risk strategy of military intervention to prevent losses to that expectation level. Once the North Korean army had been thrust backward, the possibility of a successful rollback of communism occurred, tempting Truman and his advisers to approve of U.S.-UN operations north of the 38th parallel to oust the communist government and establish a unified Korea controlled by the south. In other words, they were tempted to secure gains over and above the original expectation level. Senior American leaders adopted a new expectation level—a unified Korea—and undertook risky actions to achieve this new goal, sending troops north of the 38th parallel, in spite of Chinese warnings that it would intervene. This is viewed as an anomaly by Taliaferro; prospect theory predicts leaders are less likely to take risky actions when they believe they are facing potential gains, as American leaders clearly did. Once U.S.-UN forces neared the Yalu River (the border separating North Korea from China), things became dicey. In spite of increased warning signals, U.S. troops continued to advance northward. They escalated their commitment to a high-risk strategy. Anything short of a complete conquest and unification of Korea under a pro-Western government would be seen as a loss. And then disaster struck. A strong Chinese-North Korean counterattack pushed U.S.-UN forces rapidly back. At this point Truman and his advisers abandoned any thought of total conquest and once again recalibrated the expectation level, setting it back to the prewar status quo, and pursued a much more limited war whose goal was to restore South Korean sovereignty within its old borders. In this case it was relatively easy to reset expectations downward since rollback had only been the expectation level for a few months. In the heady days of September through November, officials would have seen the failure to create a unified Korea as a loss; by January 1951, after the communist counterattack, they saw it simply as forgoing a possible gain. Once again, consistent with prospect theory, risky actions (continuing the war) were taken to prevent or reverse losses to an expectation level. Taliaferro finds that most of the decisions in his set of cases were consistent with prospect theory. The decision that does not fit—the American attempt to make gains beyond the 38th parallel—is an anomaly but can be explained. 156 It must be remembered that prospect theory only points out probabilistic tendencies; it will not apply all the time. Decisions to attempt risky actions in pursuit of gains are not expected to disappear; they are only expected to occur less frequently than risky actions to prevent losses. Taliaferro is able to conclude that when great powers declined to intervene or decided to deescalate policies, it was largely due to “the unwillingness of leaders to pursue risky strategies to secure gains over and above their expectation levels.” 157 Personality and Risk Taking Prospect theory argues that propensity for risk taking is the result of a bias in the way humans deal cognitively with particular situations: when we frame situations to involve the prospect of losses, we are more inclined toward taking risks to prevent these losses or to reverse them. The situational context is not the only determinant of risk, however. Kowert and Hermann argue that individual leaders are likely to differ significantly in their tolerance for risks. They note that roughly a third of the subjects in prospect theory experi-

The Individual Level of Analysis, Part II

83

ments don’t actually behave in the way predicted by the framing effect. 158 Research indicates that personality, age, race, gender, education, income, and profession all have an impact on people’s attitudes toward risk. 159 The profession of politics would appear to be one profession that attracts risk-acceptant personalities. Political psychologists would emphasize personality as an important determinant of attitudes toward risk taking. Yaacov Vertzberger, who has studied risk taking in international relations, says there are three motives for risk taking: (1) deliberative/reasoned judgment that that the risky choice is better than the alternatives, (2) socially driven risks due to peer pressure, and (3) dispositional risk taking caused by personality attributes or core beliefs. He concludes that dispositional forces are the most powerful motivating forces of the three, though one’s propensity to take risks will be strongest when all three motives push individual choice in the same direction. Vertzberger’s analysis of decisions for military interventions finds that the psychological disposition of leaders is more important than situational frames: individual leaders will “consistently prefer risk-prone behavior over risk-averse behavior (or vice-versa), irrespective of situational contexts.” 160 While research on this is still in its infancy, it is probably correct to say that some kinds of people (and some political leaders) are simply risk-averse and others are risk-acceptant; it’s part of their psychological makeup. From the available research, Vertzberger concludes that “in general, risk-takers are persistent, confidant, outgoing, aggressive, domineering, manipulative, opportunistic in dealing with others, and needful of achievement.” 161 As you have probably noticed, quite a few of these characteristics match the narcissist personality. The foundation of a narcissist’s propensity for risk taking is his sense of omnipotence and feeling of invulnerability, but it can also be triggered by psychic need to compensate for injuries to his self-esteem. Kowert and Hermann use two separate instruments to assess personality differences and two other tests to measure risk propensity. They find a number of strong relationships between personality and risk taking. First, while conscientious and deliberate people are likely to be cautious and risk-averse, those with low scores on conscientiousness, deliberation, and judging tend to ignore risks. (In other studies, these types of people are shown to be hasty, impulsive, careless, and impatient.) Second, the trait of openness is positively associated with taking risks. People who are “open in thinking” may be imaginative and mischievous. Those who are “open in deed” may be adventurous and open to new experiences, consciously seeking out risky behaviors. Third, people who score high on the anxiety subscale (fear and tendency to worry) and the self-consciousness subscale (measuring shyness and inhibition) are risk-averse. Finally, as we would all expect, risk is related to gender. Males were much more likely to indicate a preference for risk than females. In sum, several types of people show a disposition for risk taking: some seek out excitement and risky experiences; others are simply impulsive and careless and tend to ignore risks. Kowert and Hermann’s experimental analysis of the relationship between personality and prospect theory indicates that both personality and context are important and that individuals respond in quite different ways to situations of risk. Some of the subjects in their study behaved just as prospect theory would predict. The behavior of others was constant, regardless of the risk or how it was framed; personality was the main determinant. A third group did respond to risks differently, depending on how the problems were framed, but not as prospect

84

Chapter 3

theory would predict: “Open and intuitive sensation seekers are willing to take risks primarily when they have something to gain (in another setting, such individuals might be considered gamblers). More agreeable, altruistic, feeling individuals prefer to avoid risks, but especially when facing a loss (and presumably when even greater harm could come to someone because of the risk).” 162 Personality may help explain how different leaders frame similar situations in different ways. Sensation seekers may frame situations differently than agreeable altruists or those who are anxious. 163 Emotion is also important. Emotions such as fear, anger, regret, panic, desperation, pride, and greed may greatly influence one’s subjective feelings about gains or losses and will therefore affect an individual’s framing of the situation. 164 Thus, while prospect theory is clearly helpful in explaining some situations in which leaders choose extremely risky actions (including war), in other situations prospect theory is trumped by personality attributes. 165 Risk taking due to personality may interact with cultural factors as well. Levy and Thompson suggest that the types of leaders who rise to the top of the political food chain in highly centralized or authoritarian regimes tend to be risk takers, otherwise they would not have been able to rise to the top and stay there. On the other hand, risk-averse leaders may be more likely in democratic systems where an important ingredient of success is compromise and the art of not alienating various constituencies. Likewise, leaders who have attained top-level office after long careers in government bureaucracies, like George H. W. Bush, are likely to have been socialized into adopting a cautious decisionmaking style. 166 THE ROLE OF IMAGES AND BELIEFS Not only do individuals have different psychological makeups, they also have different images and perceptions of the world. Images are “pictures” in an individual’s mind about objects, events, people, states, and policies. They are organized representations of the social and political environment in which we live. Images contain not only our knowledge about these things but also our evaluations of them—good, bad, neutral—and our attitudes toward them. An image is, of necessity, a simplification of reality. We keep in our mind only certain images of the events, politics, or people we think about. A great deal of information is lost. 167 These separate images are organized into a more or less coherent and integrated whole— a belief system or worldview. Ole Holsti explains that “our beliefs provide us with a more or less coherent code by which we organize and make sense out of what would otherwise be a confusing array of signals picked up from the environment by our senses.” 168 A significant part of this general belief system has to do with politics. We all presume that our images and perceptions of the world—its events, its states, and its leaders—accurately correspond to reality. Unfortunately, this is infrequently the case. Prior beliefs function as an anchor around which we organize incoming information. A leader’s current images and beliefs act as a filter through which she assesses the nature of the opponent, the perception of threats, the utility of force, and so on. This is another one of the cognitive shortcuts that is part of how the brain operates. However, these prior beliefs and images are also a source of bias that may seriously hamper our ability to create a realistic

The Individual Level of Analysis, Part II

85

picture of our surroundings. As two scholars put it, “Perception . . . is not a passive process of receiving information, but an active process of constructing reality.” 169 For a variety of reasons, which we will explore in this section, our images of the world around us may be seriously distorted. This is important because political leaders act on their individual images and perceptions of the world rather than on objective reality. For all practical purposes, the image is the reality. Two pioneers of international relations, Harold and Margaret Sprout, long ago made the important distinction between the psychological milieu (the world as perceived by the decision maker) and the operational milieu (the world as it really is, and the world in which politics must be carried out). They argued that decision makers act on their knowledge of the former rather than the latter. 170 We can only hope that the images and perceptions used by national policy makers are more accurate than not, but we know that this is not always the case. Image Content: Operational Codes One of the most important areas of investigation deals with the content of a person’s image of the world or belief system. Although several concepts have been used to describe the content of the belief system, the one most widely used is the operational code. Alexander George defines an operational code as a “particularly significant portion of the actor’s entire set of beliefs about political life.” 171 In other words, it is that subset of a person’s beliefs that have to do with international politics. (Although there are certainly subtle differences in definition involved, belief systems and operational codes overlap with what might be called an ideology—a coherent set of political beliefs.) Following an earlier work by Nathan Leites on the belief system of the early Bolshevik leaders in the USSR, 172 George developed a framework for the operational code that consisted of ten questions about politics. The answers to these questions would delineate the crucial aspects of a person’s political belief system. Five questions were philosophical and five were instrumental, dealing with political tactics. Philosophical Questions P-1. What is the “essential” nature of political life? Is the political universe essentially one of harmony or conflict? What is the fundamental character of one’s political opponents? P-2. What are the prospects for the eventual realization of one’s fundamental political values and aspirations? Can one be optimistic or pessimistic on this matter? P-3. Is the future predictable? In what sense and to what extent? P-4. How much “control” or “mastery” can one have over historical development? What is the role of the individual in “moving” or “shaping” history in the desired direction? P-5. What is the role of “chance” in human affairs and in historical developments? Instrumental Questions I-1. What is the best approach for selecting goals or objectives for political action? (For instance, on the basis of purely unilateral national interest or on the basis of multilateral considerations involving self-restraint?)

86

Chapter 3

I-2. How are goals pursued most effectively? (For instance, through force or diplomacy? Unilaterally or multilaterally? Through threats or promises of reward?) I-3. How are the risks of political action calculated, controlled, and accepted? (Under what conditions should one accept risks?) Is slow escalation of one’s actions or a fait accompli the best way to control risks? I-4. What is the best “timing” of action to advance one’s interest? (What, for instance, is the utility of preemption or surprise? Should force be used simultaneously with negotiation? Should one wait until military parity has been achieved before one pushes demands upon a rival?) I-5. What are the utility and role of different means for advancing one’s interests? George’s view of the operational code was that the ten sets of beliefs were linked in a hierarchical way. Beliefs in P-1 concerning the nature of the political universe and the source of conflict were deemed to be “master beliefs”; the other beliefs were dependent on the answers to these primary philosophical questions. Several political scientists have used George’s framework in their research efforts to investigate the operational codes of important national leaders or foreign policy advisers. Speeches, autobiographical material, transcripts of meetings and interviews, books, and articles by the subject are studied, using a method called content analysis to identify the political beliefs of the subject. Once this has been done, attempts are often made to determine whether the state’s actual policies have reflected the operational code of the individual policy maker. It is not argued that the policy maker’s operational code can be mechanically applied to the situation so that we will be able to reliably predict a state’s policy from its chief policy maker’s belief system. Rather, operational codes are seen as just one of many factors that play a role in determining policy. Sometimes they are extremely important, sometimes not. 173 What we do know is that decision makers may have a tendency to fall back on their belief systems and that the existence of belief systems may narrow the range of alternatives that decision makers may consider during the decision process. 174 International relations specialists have produced operational codes for many world leaders—including John Foster Dulles, the Soviet leaders (from the Old Bolsheviks to Khrushchev to Mikhail Gorbachev), Mao Zedong, Yitzhak Rabin, Shimon Peres, Lyndon Johnson, Jimmy Carter, George H. W. Bush, Bill Clinton, and George W. Bush. Many of the recent studies have used computer programs, such as the Verbs in Context System (VICS) to perform content analysis of public speeches or other documents. 175 It is worth contemplating the possibility that certain kinds of operational codes may be dysfunctional. For sake of argument, let us suppose that the operational code of a government leader consists of the following set of beliefs: P-1. The world is an inherently dangerous and conflict-ridden place. Our enemies are numerous, evil, and fundamentally opposed to our values. The only thing they understand is force. P-1. Peace is ephemeral and is created only by strength.

The Individual Level of Analysis, Part II

87

P-2. The ability to prevail over one’s opponents is ultimately likely because our opponents are internally weak, their political and economic systems are deficient, and they understand the superiority of our system. I-2. Reliance on treaties, international law, international organizations, and international courts is unlikely to be productive. I-2. When bargaining takes place it should always be from strength, and one should always put forward one’s maximum terms and wait for the opponent to make the first concession. I-2. Military force is an effective tool of foreign policy, and threats, ultimatums, and brinkmanship tactics are effective in getting one’s opponents to back down and to change their policies. I-3. Risky policies are part of the territory and must be accepted if one is to achieve one’s national interests. I-3. First use (preemptive and preventive use of force) is a good way to eliminate risks posed by opponents. This set of views approximates what might be referred to as a hawkish or hardline operational code. One can imagine the relative ease with which leaders with this operational code might become involved in an endless series of conflicts. Indeed, as we will see in chapter 8, states whose leaders follow these prescriptions tend to be involved in numerous wars and militarized conflicts. Let us now move from the hypothetical to the real world. Henry Kissinger has provided a fertile field for analysts. Having been a respected professor of international relations with numerous publications prior to his service as national security adviser and secretary of state in the Nixon and Ford administrations, there were plentiful paper trails from which to map out his operational code; in fact, there was a veritable gold mine of writings waiting to be subjected to content analysis. Here was a man who developed an extremely coherent approach to international politics (based on the realist perspective) before he entered public service. Would this coherent worldview, or operational code, have a significant impact on U.S. policy once Kissinger attained a position of authority? One study of Kissinger’s bargaining behavior in the negotiations to end the conflict in Vietnam indicates a very close relationship between Kissinger’s belief system and his strategy and tactics in this particular negotiation. In fact, the author states that Kissinger’s operational code was the most influential variable affecting the sequence of American actions during the spring and summer of 1972. Kissinger’s goals and behaviors during the negotiations seemed to be a logical extension of his general beliefs, formulated many years prior to public service. 176 On the other hand, another study of U.S. policy toward the Soviet Union and China during the Kissinger years concluded that Kissinger’s images of the USSR and China were only indirectly related to American foreign policy behavior. 177 Let us finish this section by looking at the operational code of top-level officials in the George W. Bush administration. It is frequently argued that one of the most important causal factors behind the Bush administration’s decision to invade Iraq was the neoconservative worldview of powerful officials such as Deputy Secretary of Defense Paul Wolfowitz, Undersecretary of Defense for Policy Douglas Feith, and I. Lewis Libby, Vice President Cheney’s chief of staff, and others. It is also asserted that the president himself was motivated by this

88

Chapter 3

worldview. An approximation of the neoconservative operational code is presented in box 3.1. You should make your own judgments about whether someone with this operational code would have been likely to use military force in Iraq in 2003. Box 3.1. The Neoconservative Operational Code A. Philosophical Questions 1. Q: What is the essential nature of politics? Harmony or conflict? What is the fundamental character of one’s opponents? A: It’s a Hobbesian world in which conflict is endemic. The nature of at least some enemies—such as rogue states and global terrorist organizations—is that they are evil and they hate America and its democratic institutions. A world of democratic states would be much more harmonious. 2. Q: What are the prospects for the eventual realization of one’s fundamental values and goals? (Optimistic or pessimistic?) A: Optimistic. The U.S. has an opportunity to restructure the world so as to increase democracy and freedom. American values will become universally accepted; the American model of democratic capitalism has no serious rival in the post–Cold War world. U.S. leadership will produce bandwagoning by others. 3. Q: Is the future predictable? In what sense and to what extent? A: Yes, at least in the long term. Democracy and freedom will spread throughout the world, though there may be short-term setbacks. 4. Q: How much control can political leaders have over historical development? What is the role of the individual in shaping history? What is the role of chance in human affairs? A: Leaders with the courage to pursue large goals can make a difference.

B. Instrumental Questions 1. Q: What is the best approach for selecting goals or objectives? (For instance, on the basis of purely unilateral national interest or on the basis of multilateral considerations involving self-restraint?) A: American national interests should have clear priority over other goals. 2. Q: How are goals pursued most effectively? (For instance, through force or diplomacy? Unilaterally or multilaterally? Through threats or promises and rewards?) A: Unilateralism is preferable to multilateralism. Threats of force, especially those made by states with unrivaled military power, are effective. However, deterrence, defense, and containment may not work against rogue states and terrorists; preventive force may be necessary.Countries are more likely to be taken advantage of if they rely on rewards. Appeasement doesn’t work.

The Individual Level of Analysis, Part II

89

3. Q: How are risks calculated, controlled, and accepted? (For instance, through slow escalation of one’s actions or through fait accompli?) A: The lesson of Vietnam is that slow escalation is unlikely to be effective. The lesson of 9/11 is that the risk of inaction is greater than the risk of taking military action. 4. Q: What is the best timing of action to advance one’s interests? (What, for instance, is the utility of preemption, prevention, or surprise? Should force be used simultaneously with negotiations? Should one only negotiate from strength?) A: In the current era timing is extremely crucial. Time is not on our side. The U.S. cannot afford to wait until threats are manifest. Preemption and prevention may be preferable to defense. One should always negotiate from strength. 5. Q: What are the utility and role of different means for advancing one’s interests? A: Military means are more effective than liberal critics believe them to be— especially in the hands of a dominant hegemonic power. Source: Greg Cashman and Leonard C. Robinson, An Introduction to the Causes of War: Patterns of Interstate Conflict from World War I to Iraq (Lanham, MD: Rowman & Littlefield, 2007), 334. The Structure and Dynamics of Images: Why Images Resist Change Although the content of images and belief systems varies enormously among different individuals, the structure of these beliefs is much more regular. Certain definite patterns can be discerned in the way incoming information is dealt with, in the way images are formed and maintained or changed, and in the relationships among the different parts of the belief system. One of the most interesting issues has to do with the process by which images are either changed (or preserved) as a result of new information. Our images constantly shift and are reappraised as new information is received, but as cognitive psychologists tell us, most of us engage in selective perception. While we readily see (and mentally “register”) those things we want to see or expect to see, we ignore many things that do not fit very well with our existing images of the world. This is especially true if the information is ambiguous. 178 To help us cope with the daily information that bombards us in a complex environment, we use previous images as a cognitive filter to reduce complexity, to screen out inconsistent or irrelevant information, and to make sense of the world. However, when the new information clearly doesn’t fit with the images already held, but instead challenges one’s present images, a number of factors inhibit image change. Information that is contrary to our current image of the world must cross a higher threshold to gain admittance into our worldview compared to information already consistent with our current image. How images change is not entirely understood, but one important factor is the complexity of an individual’s image structure. A simple image contains less information and that information is much more internally consistent—it all sums up in the same direction. Simple images tend to be very rigid and difficult to change; they may even be “closed.” A complex image, on the other hand, has both more information and more diversity of information, even inconsis-

90

Chapter 3

tencies. Because such an image is so eclectic and diverse, its holder is more open to the possibility of change. Cognitive complexity is generally related to the ability to adapt to novel or ambiguous situations—a good thing. Political psychologists who have explored the related concept of integrative complexity (measured from the public statements or communications of leaders) have found that high integrative complexity is associated with crises that are resolved peacefully, while lower levels of integrative complexity are associated with crises that end in war. Lower integrative complexity has also been associated with three “bungled” escalations widely identified as resulting from the “groupthink” process—the U.S. decision to cross the 38th parallel in the Korean War, the Bay of Pigs crisis, and the Johnson administration’s escalation of the war in Vietnam. 179 Let’s consider another factor that promotes resistance to change. One of the early conclusions of cognitive psychology is that individuals strive for cognitive consistency. We try to reduce the inconsistencies between our various beliefs. Humans are “consistency seekers.” A discrepancy between inconsistent parts of our image causes cognitive dissonance. 180 We don’t tolerate cognitive dissonance very well, and we deal with it in one of two ways: by modifying our original image or (more likely) by striving to retain our original image. Striving to retain the original image is especially likely if central, core values are challenged by the new information. Consistency theory argues that beliefs are hierarchical in nature, with strong core beliefs at the top that are particularly resistant to change and associated intermediate and peripheral beliefs that are more malleable. When change occurs, it tends to be incremental and it takes place at the margins among peripheral beliefs. In the rare event core beliefs are changed, this presumably sets off a cascade of related changes among many associated peripheral beliefs. Many techniques exist for retaining the original image in the face of such discordant information: (1) We may simply ignore or reject the new information; (2) we may discredit the source of the new information; (3) we may twist or distort the information or reinterpret it in a way that conforms to our present image; (4) we may search for alternative information that does conform to our present image of the world; or (5) we may simply treat it as “the exception which proves the rule.” 181 We may also participate in premature cognitive closure, shutting down the search for new information as soon as our preexisting beliefs have gained adequate support. In all these cases, our bias toward consistency impairs our ability to accurately perceive the world, to make realistic judgments, and to actually learn from our own actions and the actions of others. The pressure toward cognitive consistency also helps to explain the tendency for beliefs— even those that are wildly inaccurate—to persist over time. Oddly enough, continued exposure to contradictory (and correct) information actually strengthens our inaccurate beliefs. Consider public opinion in the United States concerning whether Iraq possessed weapons of mass destruction at the time of the U.S. invasion in 2003. In the early days of the war, large majorities of Americans (in some instances well over 80 percent of those polled) indicated they were certain or fairly certain of this. In July of 2006, two years after the Iraq Survey Group had issued its authoritative conclusion that there were no WMDs in Iraq, a Harris Poll showed that fully 50 percent of Americans still believed that Iraq had WMDs when U.S.

The Individual Level of Analysis, Part II

91

forces invaded that country. 182 Similarly, in September of 2002, polling data indicated that 51 percent of Americans believed Saddam Hussein was directly responsible for the September 11 attacks, a belief totally devoid of any substantive support. In 2003, as the United States went to war against Iraq, this figure actually rose to 53 percent. Though it has gradually declined since that point, in 2008 the number was still an astounding 28 percent of Americans. Meanwhile, a 2007 Pew Global Attitudes Project poll taken in the Islamic world indicated that the number of people who did not believe Arabs carried out the September 11 attacks was rising—to 59 percent of Turks and Egyptians, 65 percent of Indonesians, 53 percent of Jordanians, and 41 percent of Pakistanis. Many thought the U.S. or Israel was responsible. 183 Why do mistaken beliefs persist in spite of contrary evidence? First, because of their very nature, negative impressions are quicker to form but also more resistant to change. (This has to do with how system 1 processes information.) 184 Second, recently reported studies indicate that the presentation of information and arguments to rebut or debunk misinformation may paradoxically contribute to the resiliency of the misinformation. Let’s say, for instance, that a newspaper reports says, “The U.S. government has learned that it is not true that Saddam Hussein had weapons of mass destruction.” Part of the problem is that denials frequently require repeating the inaccurate information, but this “bad” information is more likely to be retained on second hearing than the “good” information meant to debunk it. The “bad” information already exists in our images and belief systems and is therefore more accessible. One of the brain’s subconscious rules is that we tend to rely on things that are accessible to our memories and that things that are easily recalled—in this case the bad information—are deemed to be true. The facts or arguments that are asserted first have a structural advantage over later denials. 185 Two warnings should be issued at this point. First, in spite of all we have said so far, most individuals are able to perceive reality correctly in many cases. 186 Not everything is distorted. And second, cognitive consistency is not always irrational. It may be quite logical to view incoming information as being consistent with one’s current image. It has to be evaluated in some way, and given the complexity and uncertainty of much information, we may quite logically evaluate it in a manner consistent with our current image of the world, especially if it accords with our past experience. Although the tendency toward cognitive consistency is not always irrational, its presence does indicate a systematic bias. The original image is given preference, and new information is made to fit preexisting hypotheses. Since the various images in the belief system are highly interrelated, any major adjustment of beliefs (especially core beliefs) is likely to set off a chain reaction that would in turn place a severe burden on one’s information-processing system— sort of like overloading our circuits. Stability of images is therefore preferred. 187 As cognitive psychology has developed over the years, a more complex picture of belief change has begun to emerge called schema theory. According to schema theory, a person’s beliefs and images may not be structured as hierarchically as originally thought. Instead, beliefs are more fragmented, more horizontally organized, less inter-connected, and less consistent. 188 Beliefs are organized into somewhat autonomous schemas—pockets of knowledge about different aspects of one’s world. When new information about the world is presented, people will then amend individual schema by adding subtypes or exceptions to the rule.

92

Chapter 3

People will still resist change to core beliefs, but schemas cannot resist change if the new information is unambiguous and obviously relevant. The bad news is that this kind of information—unambiguous and irrefutable—is unlikely to be found in the realm of international politics. 189 The good news is that, if individuals have beliefs and images that are organized less tightly and less hierarchically (as in relatively autonomous schema), they are more likely to change their beliefs over time than those whose cognitive structures are more coherently and hierarchically organized in the way described by consistency theory. 190 What determines the type of cognitive structure a political leader is likely to have? One’s level of expertise and role are both important. Research indicates that experts (including experts on international affairs) are likely to develop rather coherent belief systems, while those with only limited knowledge or experience tend to hold more fragmented and inconsistent beliefs. 191 Steinbruner has connected typical cognitive processes to governmental roles. 192 Midlevel officials and advisers (many of whom have spent time in academia) tend to be “theoretical thinkers” with internally consistent beliefs organized around a single value. They have strongly deductive thought processes and are able to act with great confidence in interpreting current situations. The models here are Henry Kissinger and Zbigniew Brzezinski, national security advisers to presidents Nixon and Carter. On the other hand, high-level officials at the top of the decision-making chart tend to be “uncommitted thinkers.” These leaders tend to be generalists rather than experts. They sit at the center of political cross-pressures and information; their beliefs tend to be unstable, and, “for many of the problems which face them, they quite literally do not know what to think.” The typical pattern is vacillation: “Since his own experience does not commit him to a particular belief pattern, he will adopt several competing patterns, not at once, but in sequence.” 193 One thinks here of Franklin Roosevelt, Jimmy Carter, Lyndon Johnson, and Bill Clinton. More recent evidence from the schema perspective indicates that since experts have information organized along more dimensions, they can actually incorporate much more inconsistent data as exceptions to their schemas. But these inconsistent data have a relatively small impact on their schemas. Thus, “deviant cases must confront a greater wealth of evidence stored under well-developed schemas,” creating a situation in which, paradoxically, “informed people can better use contradictory information but are less likely to be influenced by it.” 194 Philip Tetlock finds that political experts actually do very poorly in predicting or forecasting events. In spite of their extreme confidence in their predictions, they performed about as well as monkeys throwing darts would in predicting the probability of events. This was especially true for experts with a “hedgehog” approach to thinking who work deductively from basic principles, versus experts with a “fox” approach who work inductively from many facts. 195 The Resistance of Images to Change: Some Examples Let’s take an example of an international situation that might challenge a national leader’s image of the world. Soviet General Secretary Gorbachev’s foreign policies in the late 1980s— making major concessions to the United States on arms control, withdrawing Soviet troops from Afghanistan, unilateral reduction of troops in Eastern Europe, renouncing the Brezhnev doctrine (which in turn permitted democratic revolutions in Eastern Europe), reducing Soviet forces in Mongolia and along the Soviet-Chinese border, pressuring Vietnam to remove its

The Individual Level of Analysis, Part II

93

forces from Kampuchea, and virtually abandoning Marxist regimes in places like Mozambique—created a situation of severe cognitive dissonance for hardline “Cold Warriors” in the United States. Instead of changing their images of the Soviet Union, many American policy analysts in the late 1980s resorted to various techniques to preserve their already existing cognitive structures. Gorbachev’s policies were at first discredited as mere public relations ploys cynically designed to change the West’s image of the USSR. (The special problem of international politics is, after all, that since deception is common, actions by other states that do not conform to one’s image can be easily categorized as deliberate attempts to deceive.) Alternatively, Gorbachev’s policies were initially ignored as relatively meaningless gestures, or they were reinterpreted so as to be consistent with Cold War images. For instance, they were seen as simply a temporary retreat brought on by a combination of Western strength and firmness and by the horrendous condition of the Soviet economy rather than as concessions brought about by any real change in Soviet thinking (an example of the fundamental attribution bias). And Gorbachev himself was seen as an exception to his predecessors, and one not likely to last very long given the opposition to his policies from more hardline Soviet elites. 196 As a result of these image-preserving techniques, the changes associated with Gorbachev’s “New Thinking” foreign policy caught virtually everyone in Washington by surprise and after they happened, U.S. leaders were rather slow to assess the real importance of Gorbachev’s revolution in foreign policy, squandering opportunities for greater cooperation. The American Central Intelligence Agency (CIA) was at the heart of the problem, and at the center of the action in the CIA was Robert Gates, the CIA’s chief Soviet expert from 1979 to 1982. He became its deputy director of intelligence from 1982 to 1986 as Gorbachev first entered office and took over as deputy director of the CIA from 1986 to 1989, as Gorbachev’s transformational policies reached their peak. He then became deputy national security adviser from 1989 to 1991, as the Soviet Union descended into chaos. 197 Recently declassified CIA documents show that after Gorbachev’s famous announcement to the UN in December 1988 that he would unilaterally cut 500,000 troops from the Soviet army and withdraw six tank divisions from Eastern Europe, the CIA’s analysis of Gorbachev’s foreign policy was that it was “in the Leninist tradition: it calls for weakening the main enemy—the United States—by exploiting ‘contradictions’ between it and other centers of capitalist power.” Later—after Soviet withdrawal from naval installations in Vietnam, the Indian Ocean, and the Mediterranean Sea and its abandonment of pro-Soviet partners in several Third World conflicts—a CIA report incredibly described “Moscow’s intention to expand its role as a global actor.” 198 Washington Post reporter David Kaiser illustrates how personal and political dynamics added to the problem: Secretary of State George Shultz, who had many private conversations with Gorbachev, knew that Gorbachev and his policies were in fact quite different from the norm; however, he was apparently unwilling to voice his views about Gorbachev to other members of the national security team because he was afraid that Gates and Defense Secretary Caspar Weinberger “would accuse him of being hopelessly naïve.” 199 One of the classic studies of images has to do with a very similar set of circumstances. Ole Holsti undertook to identify through content analysis the belief system of Eisenhower’s secretary of state, John Foster Dulles. 200 In particular, Holsti was interested in Dulles’ image of the

94

Chapter 3

Soviet Union and whether he was capable of changing his image if new events and information challenged it. Significantly, Holsti found that Dulles’s general negative evaluation of the USSR was not subject to change—even when he perceived reduced Soviet hostility. Positive Soviet actions, such as the signing of the Austrian State Treaty in 1955 (which removed the Red Army from Austria) and the reduction of the size of the Soviet army, did not change Dulles’s general evaluation of the Soviet Union. Instead, these cooperative moves were credited to Soviet internal weaknesses: the result of necessity rather than good will. What is important about this is that if cooperative Soviet actions were unable to change the secretary’s basic image of the Soviet Union, then how could the Soviets reliably demonstrate to American officials their good intentions? What could the Soviets have done to indicate to Secretary Dulles their good will? The implication is that nothing short of the dismantling of their own system would have done it. (These cognitive dynamics were probably also in play among many members of the George W. Bush administration in its analysis of Iraq in the run-up to the war in 2003.) Kissinger has referred to this kind of image structure as the inherent bad faith model. 201 An individual who holds this type of image structure will be able to explain away almost any sincere change in behavior by the opponent; there is virtually nothing the opponent can do therefore to change the original image. This kind of image structure clearly makes learning questionable. The problem is bad enough if leaders in one country have closed images of this nature; imagine what would happen if leaders in both countries have similar image structures. How Images Might Be Changed Since there seems to be a bias in favor of maintaining one’s present image, what does it take to bring about a change? It appears that a change in images comes about more easily if new information descends all at once rather than piece by piece over extended periods of time. 202 Small bits of information coming at irregular intervals stretched out over time are easily discounted, assimilated, or otherwise coped with. On the other hand, spectacular events (often of a profoundly disturbing character) and dramatic pieces of conflicting information that strike all at once virtually demand that the individual take action to deal with the discrepancy. The occurrence of spectacular events, however, may not be enough to induce a major change in image. Real change may require both spectacular events and the accumulation of less impressive, long-term developments that challenge the image. 203 Some parts of a belief system are easier to change than others. According to both cognitive consistency theory and schema theory, beliefs that have short-run consequences rather than long-run consequences are most likely to be altered. Fundamentally, the less-important, peripheral parts of one’s belief system are more likely to change than its central or core parts and they are most likely to change first. As Stein says, “People will also make the smallest possible change; they will change their beliefs incrementally, generate a large number of exceptions and special cases, and make superficial alterations rather than rethink their fundamental assumptions.” 204 When central beliefs change, they initiate a corresponding cascade of changes to related peripheral beliefs. Several studies of the operational codes of world leaders indicate that in fact beliefs do change over time, sometimes over a period of just a year or two. What is most interesting,

The Individual Level of Analysis, Part II

95

however, is that contrary to prevailing hypotheses, philosophical beliefs (basic assumptions) are more likely to change than instrumental (tactical) beliefs. And when instrumental beliefs do change, the changes tend not to be very large. In Renshon’s study of change in the operational code of George W. Bush over four separate time periods, the events of September 11 brought about reversal of several philosophical beliefs. His view of the nature of the political universe (P-1) and of the ability to realize political values (P-2) were both substantially affected in a negative direction, and his view of the predictability of the future (P-3) also declined somewhat. However, Bush’s instrumental beliefs did not change accordingly; they remained stable—not at all what consistency theory would predict. 205 In fact, Renshon argues that very few studies find wholesale changes in a leader’s operational code. While some beliefs may change, especially as a result of traumatic events, they do not trigger a very comprehensive overhaul of the belief system. The degree to which one’s beliefs are resistant to change is also related to personality. For instance, leaders with dogmatic personalities or authoritarian personalities—people who tend to have rather simple, black/white images of the world—are less likely to change their views, as what they believe is likely to be protected by ego-defense mechanisms. On the other hand “uncommitted thinkers” are much more likely to be able to change their belief systems. A distinction frequently made by political psychologists is between “hedgehogs” and “foxes” (a classification originating with the political philosopher Isaiah Berlin). Hedgehogs “know one big thing” and foxes “know many little things.” The one big thing that hedgehogs know is a central idea or perhaps an ideology that is a grand, overarching scheme that can be (and frequently is) applied to virtually everything. Hedgehogs are deductive thinkers who reason from the top down, from one big theory to details that are predicted by, or are consistent with, the theory. Foxes are inductive pragmatists who reason from the bottom up—from a large accumulation of facts and evidence to larger theories. They tend to be skeptical of overarching theories that explain everything, and they’re more likely to change images to correspond to new information. 206 As you have probably guessed, George W. Bush is a hedgehog rather than a fox. This clearly played a role in his approach to the decision for war in Iraq in 2003. Cashman and Robinson’s review of the relevant literature sums it up like this: Bush appears to be an “ideological thinker.” . . . He works more or less deductively from moral, religious, and political principles rather than inductively from empirical facts. He is a fundamentalist rather than a rationalist; his approach to information and ascertaining “the truth” does not lend itself to nuanced, complex thinking. In Isaiah Berlin’s famous comparison of the hedgehog and the fox . . . President Bush is clearly a hedgehog. He is not a seeker of facts or opinions. His decision style is to make decisions quickly without detailed study and investigation. . . . He is predisposed to believe things on faith and principle without examining the evidence and to hold these beliefs with absolute certainty. He is a “true believer” rather than a pragmatist. 207

With these personality characteristics, once a person becomes committed to a big idea—for example, “Saddam Hussein has weapons of mass destruction that constitute a threat to the United States, and this threat requires preemptive action by the U.S.”—the big idea drives

96

Chapter 3

everything else and is relatively impervious to change. Table 3.2 summarizes the factors that aid or inhibit image change. We have already seen that several of President Bush’s fundamental assumptions about international politics changed as a result of the traumatic events of September 11. President Jimmy Carter’s conversion to a more hardline, anti-Soviet world view in the wake of the 1979 Soviet intervention in Afghanistan provides another interesting example of (temporary) image change. Although the Soviet intervention certainly qualifies as a spectacular event, it came in the wake of other international (and domestic) events that probably had a cumulative impact on the president. Soviet activity in Angola and Ethiopia, the Soviet arms buildup, continued human rights violations by the Soviets, the toppling of the Shah of Iran by anti-American forces led by Ayatollah Khomeini, domestic opposition to Carter’s SALT II treaty with the Soviets, and the rising prominence within his foreign policy inner circle of National Security Adviser Zbigniew Brzezinski probably all had an effect on the president. Carter’s image of the world was changed rather dramatically. His vision was no longer that of a benign and liberal world where states could get along together through reason and diplomacy and law, but of a world where states harbored aggressive intentions against each other and where one’s opponents could not always be counted on to abide by laws or listen to reason—a world where force must frequently be employed instead of diplomacy. A sophisticated analysis of Carter’s operational code before and after the Soviet intervention in Afghanistan (using the computerized Verbs in Context System) confirms a change in his fundamental assumptions about the world. The results showed a significant decrease in his cooperative view of the world (P-1) and a decrease in his optimistic view of his ability to realize his goals (P-2). 208 While the natural resistance of images to change probably contributes to the stable, incremental nature of foreign policies, Jervis argues that a major change in image by a national leader always brings about a change in policy. 209 This may or may not be true, but the Carter presidency provides a good example. After Carter’s “conversion” he swiftly initiated a new set of policies that reflected a more hawkish approach: he withdrew SALT II from Senate consid-

The Individual Level of Analysis, Part II

97

eration, placed a grain embargo on the Soviets, began to restore a military aid to El Salvador, expanded covert support for mujahideen forces in Afghanistan, and increased the defense budget. Evoked Sets Another important influence on our interpretation of reality is the phenomenon of “evoked sets.” The things that we are concerned about when information is received play a role in determining how we will interpret information by creating predispositions to notice certain things and to neglect others. 210 While most of our correct inferences about the world are strongly influenced by our expectations, seeing what we expect to see can create some consequential distortions in our evaluations of the world. A perfect example can be found in the July crisis that preceded the outbreak of World War I. (You will remember this situation from our discussion of Kaiser Wilhelm II’s psychological breakdown.) British Foreign Secretary Grey sent a note to the German government warning of dire consequences if war were to begin. Kaiser Wilhelm II’s evaluation of this note was conditioned by the fact that he had just received information of the Russian military mobilization. The timing of these two messages predisposed him to view the British letter as part of a joint British-Russian plot against Germany. The British information was interpreted in light of the just-received Russian information. In this instance, the Kaiser’s “evoked set” played a part in his (mis)perception of a combined Russian-British threat to Germany in 1914. 211 Evoked sets certainly played a role in the U.S.S. Vincennes’s shooting down of an Iranian airliner in 1988. The crew’s expectations were affected by three situations. First, U.S. intelligence—based on intercepts of Iranian communications—had predicted a strike against an American ship in the Persian Gulf. Second, just minutes before the Iranian plane took off there was an attack: Iranian boats fired on a U.S. helicopter and U.S. ships then skirmished with a flotilla of Iranian vessels. Third, recent military reports indicated that F-14s were now stationed at Bandar Abbas airbase. 212 Thus, the American crew expected an attack; they expected aircraft in the area to be hostile; and they were predisposed to see Iranian F-14s flying out of Bandar Abbas. In the light of the ambiguous information available about the plane on the radar screen, the crew’s evoked set helped define (erroneously) the plane as a military F-14 rather than the civilian airbus that it was. THE ROLE OF DECISION-STAGE PSYCHOLOGY: THE RUBICON THEORY Several theorists, Geoffrey Blainey most prominently, suggest that overconfidence on the eve of war is a primary causal factor in the decision for war. Dominic Johnson and Dominic Tierney have furthered this thinking with their so-called Rubicon theory, which helps explain why leaders on both sides seem to be overconfident just prior to the outbreak of war. Based on a series of psychological experiments, they identified a sequence of stages in the decisionmaking process, each of which activates a different mindset or cognitive orientation. 213 The first stage of decision making is a deliberative/predecision stage. At this stage, leaders do not believe that war is inevitable. Once leaders believe that war is inevitable, their mindsets change to the implemental/postdecision stage. The difference is so crucial that the dividing

98

Chapter 3

line is seen as crossing a psychological Rubicon. On the far bank, the deliberative mindset approximates the rational actor model outlined early in this chapter. Leaders are reflective, are open-minded, search diligently for answers, and weigh the expected utilities of various alternatives. On the near bank, the implemental mindset takes over, and the cognitive process undergoes interesting changes that trigger a number of processing biases. The focus shifts to carrying out the chosen course—war. At this point leaders are characterized by a kind of tunnel vision: they focus on achievement of goals and avoid being distracted by alternatives or doubts. They resist reconsideration of alternative courses of action. They are less receptive to incoming information and search instead for information that supports their choice. They process information in a more biased way. When beset by cognitive dissonance, they subconsciously try to fit contradictory information into their preexisting beliefs. They are vulnerable to self-serving evaluations and to the illusion that they can control events. Most important, they are overconfident about the war—about the probability of victory, the benefits of said victory, and the relatively low costs of the enterprise. All of the cognitive aspects of the implemental stage push leaders in the direction of overconfidence. Overconfidence may then lead to war in two ways: a direct decision for war or indirectly through the pursuit of more risky war plans and tactics. The Rubicon theory shares some similarity with the concept of bolstering—the idea that once a decision has been made, those responsible for it engage in subconscious efforts to “bolster” and support the notion that the decision will work, as a way of decreasing stress and decreasing threats to one’s ego that might result from a disastrous policy choice. The Rubicon theory is slightly different in that it suggests that the crucial point is not the point at which the decision is actually made, but the point at which war appears to be inevitable, no matter what one does. Johnson and Tierney test the theory by examining the leaders in France, Russia, Germany, Austria-Hungary, and the United States in the days before the outbreak of World War I. They attempt to determine the point at which leaders decided that war was inevitable, whether this perception precipitated a change from a deliberative mode of decision making to implemental mode, and whether the result of this was overconfidence. They claim the case of WWI is consistent with the Rubicon theory, and the theory has potential utility for explaining other wars. THE ROLE OF MISPERCEPTION Misperceptions occur when an individual’s perceptions of the world do not correspond to reality. As you may have already realized, misperceptions by national leaders abound in international relations. This is in part natural. Policy makers cannot ultimately know what is really going on in much of the external environment. International politics are rarely experienced directly. Instead, national leaders learn about them through secondhand reports: through the press, from emails from diplomatic stations around the world, through briefings by advisers, or through the television screen, as reported by CNN or Al Jazeera. Additionally, as we have already seen, our understanding of external events is subject to the misinterpretation that is provided by our preexisting images and belief systems. Our perceptual screens may bias and distort information received from the environment.

The Individual Level of Analysis, Part II

99

We have already provided some evidence that many foreign policy decisions are, in fact, nonrational. Misperceptions may be a key component in such nonrational decisions. 214 In fact, misperceptions by national leaders have frequently been cited as the immediate cause of war. World War I, Vietnam, Korea, the Middle East Wars of 1967 and 1973, the first Gulf War, and the Iraq War of 2003 are typically seen as examples. Cashman and Robinson look at seven cases of interstate war in the last 100 years and find that “in almost no case did the leaders in the initiating country, operating under what we have called the fog of prewar, accurately perceive the situation in which they found themselves.” 215 In cases in which leaders would not have chosen war if they had had a more accurate understanding of real-world conditions, we can say that misperceptions had a causal effect on the choice for war. But misperceptions are notoriously difficult to define. True misperceptions are distorted images of current real-world situations, whereas mere erroneous judgments are errors in assessing probabilities or predicting future contingencies; the difference is pretty thin. 216 Keeping that in mind, let us move forward. A number of misperceptions are typically related to war: misperceptions of the opponent and his intentions, misperceptions of the opponent’s military capabilities and of the relative balance of power, misperceptions of the opponent’s willingness to give in to one’s demands, misperception of the risks involved in pursuing one’s policies, misperceptions of the intentions and capabilities of third countries, and misperceptions of the inevitability of war and of its eventual outcome. Each of these misperceptions is associated with a particular causal path to war. Let us take them one at a time. 1. Misperceiving the opponent as having more hostile intentions and as undertaking more hostile activities than is actually the case (and, conversely, failure to perceive that opponents view one’s own activities as threatening to them). The exaggeration of an opponent’s intentions is probably one of the most common misperceptions. It is due primarily to the combined effects of trying to deduce the adversary’s intentions from his military capabilities and to the related tendency to do this on the basis of worst case analysis. These problems are compounded by the opponent’s attempts to conceal his intentions through strategic deception. The pre–World War I international environment is the classic case of the overestimation of opponents’ intentions, and President Theodore Roosevelt’s personal notes are the classic statement of the predicament. TR wrote in 1904 that the Kaiser sincerely believes that the English are planning to attack him and smash his fleet, and perhaps join with France in a war to the death against him. As a matter of fact, the English harbour no such intentions, but are themselves in a condition of panic terror lest the Kaiser secretly intend to form an alliance against them with France or Russia, or both, to destroy their fleet and blot out the British Empire from the map! It is as funny a case as I have ever seen of mutual distrust and fear bringing two peoples to the verge of war. 217

John Stoessinger sums up this type of situation nicely: “When a leader on the brink of war believes that his adversary will strike him, the chances of war are fairly high. When both leaders share this perception about each other’s intent, war becomes a virtual certainty.” 218

100

Chapter 3

Overperception of threat was a major problem in the crisis that initiated World War I. Robert North’s content analysis of documents from the 1914 crisis demonstrates that the perception by German leaders of the intentions of the Triple Entente was significantly more hostile than was warranted by an objective analysis of the situation. 219 The Germans therefore reciprocated the hostility that they believed, erroneously, they were receiving from their opponents. Overperception of threat was also a major problem for senior members of the George W. Bush administration regarding Iraq in 2001–3. Their perception of the threat to the United States and of Iraq’s possession of WMD capability—chemical, biological, and nuclear weapons—turned out to be erroneous. Jack Levy sees two paths to war that may be generated by an exaggerated perception of an opponent’s intentions. The first path is a direct one: a preemptive strike may be launched against the state that is perceived to harbor hostile intentions. The second path is an indirect one in which the perceiving state increases its military capabilities to compensate for the hostile intentions that it perceives to be held by the rival state, and the rival responds accordingly, initiating a spiral of ever-increasing hostility, which ends eventually in war. 220 It should be noted that occasionally the opposite perceptual fault is present: the opponent is perceived as being less hostile than is really the case. During the appeasement of Hitler in the 1930s many political leaders in the west assumed that the German führer shared their objectives of limited political goals and a peaceful Europe. Richard Ned Lebow’s explanation is that this misperception can be attributed to the projection of their own national values (and national self-images) onto Hitler’s Germany. 221 Knowing little about other countries, we may tend to see them and their leaders as being similar to ourselves. The consequences of underestimating the opponent’s hostile intentions are, of course, a false sense of security and a lack of defensive preparation. Saddam Hussein’s perceptions prior to the American invasion in 2003 (and also the attack of the coalition forces in 1991) provide another example of under-perception of threat. Based on postwar interviews with Iraqi officials and on Iraqi documents, Duelfer and Dyson uncover a lengthy list of misperceptions held by Iraqi leaders (and also by American officials). 222 In the run-up to the 2003 war, Saddam and other Iraqi leaders (1) underestimated U.S. hostility and its commitment to use force against Iraq; (2) misread the effects that the end of the Cold War and the September 11 attack would have on U.S. policy on Iraq, which gave the U.S. both the incentive and the unfettered ability to take action in the Persian Gulf; (3) overestimated U.S.-Iraqi shared interests in fighting Islamic terrorism and balancing against Iran; (4) overestimated the omniscience of U.S intelligence agencies, assuming that the Americans were well aware that Iraqi WMDs no longer existed; (5) misread the United Nations and its ability to constrain U.S. action; and (6) used a faulty analogy—U.S.-Libyan relations—to evaluate how the U.S. might treat Iraq. The cumulative effect of these misperceptions was that Saddam was unaware of the serious jeopardy he had put himself in. 2. Inaccurate perception of military capabilities and the relative balance of power between one’s self and one’s opponents—in particular, the perception that the opponent is weaker than is the case.

The Individual Level of Analysis, Part II

101

Geoffrey Blainey suggests that wars begin when leaders in different states have different perceptions about their relative strength. Conversely, wars cease because those same leaders come to share a similar perception about their relative strengths and weaknesses. 223 In a sense, the actual combat of war gives each nation involved a crash course in reality testing—the purpose of this course being to determine whose initial perceptions were correct. Lebow discovers that in all five “brinkmanship crises” in the twentieth century that ended in war, leaders in the initiating countries grossly misjudged the military balance and were sure of victory if the crisis were to result in war. For instance, in the Middle East War of 1967 Nasser apparently became “intoxicated” with the cornucopia of weapons and men he saw during his inspections of Egyptian positions in the Sinai. 224 The Pakistani decision to pursue full-scale combat with India over the Kashmir in 1965 was due in part to the sense of overconfidence that Pakistani leaders gained from the brief border skirmish in the Rann of Kutch earlier in the year. 225 And prior to the Russo-Japanese War, the Russian perception that Japan would not risk war was based on the misperception of extreme Russian military superiority—a perception that rested in part on racial stereotypes. 226 All of the above are examples in which misperception of an opponent’s capabilities created a feeling of overconfidence that heavily influenced the decision to initiate war. Clearly such misperceptions are important; after all, states rarely initiate wars they don’t expect to win! In cases where the perception of the military balance is erroneous (as revealed by the state’s subsequent defeat in the war it initiated), we can be fairly confident in stating that misperception was a direct cause of the war. In fact, Levy has argued that the perception of a military advantage is probably a necessary condition for war, though it is probably not sufficient by itself, since leaders usually require that the war not only be won but be won without prohibitive expense. 227 A different type of overconfidence was involved in the George W. Bush administration’s decision to attack Iraq in 2003. While in many ways the perception of the military balance (highly in favor of the U.S. and its allies) was correct, the civilian leadership’s perceptions of the postwar balance was fatally flawed. It was assumed that once the Iraqi army had been defeated, there was little chance of pro-Saddam insurgency or anti-American opposition forces, let alone a full-blown civil war that American occupation forces would have to deal with. 228 An overestimation of the opponent’s military power may also lead to war, though in a different way. Lebow finds that one of the most important perceptions of threat involves the perception of an impending dramatic shift in the balance of power in the adversary’s favor. This was a factor in roughly half the thirteen brinkmanship crises in Lebow’s study. 229 Misperception of future shifts in power to one’s disadvantage create a motive for preventive war—a war fought now, while the chance of victory is perceived to be good, rather than a war fought later, when the chance of success is perceived to be worse. Perceptions of immediate changes in the balance of power are also important—especially changes in short-term tactical advantage that would give one side an immediate military advantage. Such perceptions of immediate, tactical changes in the military situation create a motive for preemptive war—war to defend against a perceived imminent attack by an opponent. While not all such perceptions are false, history provides adequate examples of errone-

102

Chapter 3

ous perceptions of tactical advantage. With regard to the outbreak of World War I, Tsar Nicholas II’s order to mobilize Russian troops on July 30, 1914, was based on the erroneous perception that Germany had undertaken secret military preparations against Russia—preparations that would give her a decisive edge if not quickly countered by Russian action. In reality, no such secret preparations had occurred; Germany did not order premobilization until July 31, in response to the Russian mobilization. French mobilization was apparently based on similar misperception of a secret German call-up of “tens of thousands” of reserves. 230 With regard to the Six Day War, we shall perhaps never know whether the Egyptians and Syrians planned to attack Israel in the summer of 1967; over four decades later, the matter is controversial. Israeli leaders, however, were certain that the attack was imminent. The preceding discussion has probably already alerted the reader to the powerful multiplier effects of misperceptions. An overperception of an opponent’s hostile intentions coupled with a misperception of his relative military weakness would certainly be a potent combination. Unfortunately, misperceptions are rarely “only children”; they tend to be twins or triplets. Since perceptions of strength are linked both to perceptions of intentions and to perceptions of risk, there are two corollaries to the basic misperception of relative strength: 2a. The inaccurate belief that the opponent will give in to your threats and ultimatums rather than go to war. 2b. The misperception of the degree of risk one faces in initiating a conflict. The defining characteristic of the brinkmanship crises that Lebow investigates is that the initiator fully expects the adversary to back down rather than resort to arms. In each crisis that he investigates these initial perceptions were proved to be inaccurate, and the initiator either had to back down or face actual hostilities. His findings suggest that the presence of a vulnerable commitment by the opponent is not a precondition for brinkmanship crises; what counts is the perception by the initiator that a vulnerable commitment exists—a perception that was frequently erroneous. 231 Misperception thus appears to be a major cause of war in brinkmanship crises. A good example is the misperception of Indian leaders in the 1962 crisis with China in the Himalayas. Indian leaders believed that China would back down when confronted with India’s aggressive “forward policy” of manning military posts in the disputed regions—even though China had military superiority on both fronts involved. This perception by Indian political leaders was abetted by a sycophantic military that was unwilling to challenge what they believed to be an erroneous image of Chinese behavior. 232 Not only do different individuals have different propensities for risk taking, but they also perceive different degrees of risk in the same or similar situations. Jervis maintains that (unlike what many historians have suggested) Hitler was not reckless in his attempts to dominate Europe in the 1930s but merely certain the other side would back down. 233 It was not that Hitler was more willing to take risks than others, but that he believed the risks of war were slight. Perhaps the classic case of misperception of risk was the American decision to try to unify Korea after the initial task of driving back the North Korean attack on the south had succeeded. Both Chinese strength and Chinese willingness to defend North Korea were severely underrated by the American side—in particular by General MacArthur and his staff. Part of

The Individual Level of Analysis, Part II

103

the problem was bureaucratic. Army intelligence played down the evaluation of Chinese strength both to avoid demoralizing the South Korean army and to avoid the wrath of MacArthur, whose position on Chinese strength was well known. MacArthur’s staff (like the Indian military two decades later) had learned to be properly submissive to their leader. Consequently, MacArthur was unaware of the anxiety of some of his generals because his closest subordinates had insulated him from it. 234 De Rivera suggests that the basis for this was that MacArthur had a psychological need to be free from disagreement and be surrounded by people supportive of his own views, a need created by his own lack of self-esteem. 235 To some extent the misperceptions about Chinese capabilities and intentions were also based on wishful thinking by American leaders. Though warnings abounded that the official interpretation of Chinese strength and intentions was inaccurate, U.S. leaders persisted in their adherence to the image of Chinese weakness and bluff. With Truman’s popularity plunging like a pelican looking for a quick meal and with his administration under attack from the Republicans for being soft on communism, the creation of a unified and democratic Korea would certainly have been the answer to a host of political problems for Truman and Secretary of State Acheson. On the other hand, failure to push ahead with unification would be seen as appeasement. According to Lebow, Truman and Acheson needed a victory in North Korea; there was no acceptable alternative. In the face of information that such a victory might be complicated by Chinese entry into the war, American leaders responded by bolstering (looking for evidence to reinforce one’s belief) and with selective attention to information in order to dismiss Chinese threats as a bluff. 236 Thus, they were able to conclude that the risks of pushing northward across the 38th parallel were relatively small. 3. Perception that war is inevitable. There are two versions of this perception. The general phenomenon of war may be seen as an inevitable feature of the landscape of international relations, or specific wars may be perceived as inevitable at specific times. Each perception has an impact on the willingness of leaders to opt for war. Evan Luard has remarked on the importance of the perception of the general acceptability of war as a feature of international relations. He believes that ultimately decisions for war depend on leaders’ beliefs about war: “about its usefulness, legitimacy or morality, about its value in enhancing national prestige, upholding national honor, or asserting national will; above all about its normality as a constant feature in the behavior of states.” 237 Luard traces beliefs about the normality of war in the international system from 1400 to the late twentieth century. He argues that the main change is not in the degree to which war in general has been regarded as legitimate (this hasn’t changed too much) but in the type of war that is seen as legitimate in successive periods. 238 Recently some scholars have made the argument that, at least among the advanced industrial democracies of the North, the spread of peaceful norms since the end of World War I has meant that war no longer is seen as serving a necessary political, social, or economic function and that offensive war is not acceptable among these states. 239 The bottom line, however, is that while attitudes toward outright aggression may have gradually become more negative, war is still seen as an acceptable instrument of national policy under certain circumstances. And to the extent that war is perceived as a normal and

104

Chapter 3

acceptable feature of international relations, leaders will continue to include it in their repertoire of options for dealing with other states. The extent to which political leaders perceive that a particular war is inevitable at a particular point in time may be an even more important factor in the decision to go to war. Whether this actually constitutes a misperception is difficult to say; after all, the war actually does take place. Leaders’ forecasts or predictions about future realities are substantially unlike their perceptions of current realities; the latter can more easily be seen as accurate or inaccurate than the former. 240 Nevertheless, we shall consider such predictions to be misperceptions since they are commonly treated as such in much of the writing about war. World War I appears to some authors as the classic example of a situation in which leaders on all sides perceived war as inevitable. 241 All of the European crises that immediately preceded it led to international conferences of leaders or foreign ministers. But these attempts to settle the Austro-Serbian crisis by means of an international conference never really got off the ground. We can speculate that it was probably seen by many as a useless waste of time, given the perception that war was inevitable. The perception that war is inevitable, combined with the perception that the present time will be more advantageous militarily than a future time, creates an especially dangerous set of circumstances. Certainly this was the case in the July 1914 crisis. German and Austrian leaders felt that the summer of 1914 might be the last opportunity to win such a war, which would certainly come sooner or later. 242 Russian rearmament would be completed by 1916–17, creating a much more dangerous environment. Ironically, many British and French leaders, the French general staff in particular, also believed it would be best to fight Germany sooner rather than later. Similar perceptions affected French decision making in 1870 prior to the war with Prussia and Japanese military leaders in 1903–4, who believed they had a temporary strategic advantage vis-à-vis Russia that would be quickly lost. 243 4. Perception that the war will be relatively inexpensive and short. Generally decision makers must believe that war will be winnable, short, and not prohibitively expensive. 244 On the other hand, if war is seen as being ruinous, political leaders will be less likely to plunge their nations into it. The Bush administration’s perception in 2003 that a “shock and awe” war in Iraq would be short and relatively inexpensive with little postwar nation-building was certainly a part of the reason why the decision for war was so compelling. The same was true with World War I, for several reasons. First, there was the power of historical analogy. The shortness of the last major wars in Europe—the Franco-Prussian War and the Austro-Prussian War—led twentieth-century statesmen to expect more of the same. Second, there was the belief that no great power could financially sustain a long war, especially given the commercial and financial interdependence of European states. Finally, there was the factor of military strategy. European militaries emphasized the offensive side of warfare, believing offensive strategies and technologies to have a decisive advantage over the defense. 245 The presumption was therefore that someone’s offense would quickly overcome an opponent placed on the defensive, thus precluding a long war. This brings us to an important corollary misperception. 4a. The perception that offensive weapons and strategies are superior to defensive weapons and strategies. As we will see in chapter 10, Stephen Van Evera claims that even though

The Individual Level of Analysis, Part II

105

defenses are usually superior to offenses (as they were in World War I), the perception that the offense has an advantage is a crucial factor in pushing leaders toward the decision for war. 246 It places a hair trigger on crisis interactions; it leads to the belief that conquest is easy and defense is difficult; and it leads to opportunistic wars, preemptive wars, and preventive wars. 5. Misperceptions of the intentions (and capabilities) of third states. Blainey points out the importance of perceptions (or misperceptions) concerning the behavior of third states: who will enter the war on whose side and who will remain on the sidelines, who will honor their alliance commitments and who will weasel out. 247 The correct perception of third-party intentions can be enormously beneficial. American leaders correctly perceived that no European states would intervene on behalf of Mexico in 1846. European leaders correctly perceived the American Civil War would prevent the United States from aiding Mexico against their intervention in the 1860s. And Japanese leaders accurately perceived that no great power would intervene on China’s side in its 1937 invasion of that country. Misperceptions are just as plentiful, however. The typical misperception (perhaps a kind of wishful thinking) is that one’s potential enemies will remain neutral while one’s allies will remain faithful to their commitments. Erroneous perceptions of this sort directly affect leaders’ cost-benefit analysis of the desirability of war by contributing to military overconfidence. For instance, Hitler’s perception that Poland would receive no outside assistance was rudely corrected by the British and French declaration of war in 1939. In 1914 German and Austrian leaders believed the war against Serbia could be kept localized without outside intervention by other great powers. The Kaiser’s initial perception that British involvement was unlikely seemed to have been a crucial factor in German calculations. While the decision of the North Korean and Soviet leadership to attack South Korea was probably based (erroneously) on the perception that the United States would not intervene, the Truman administration’s decision to send UN troops into North Korea was based on the (equally erroneous) perception that China would remain neutral in spite of her protestations to the contrary. 248 In the weeks preceding his 1990 attack on Kuwait, Saddam Hussein almost certainly perceived little risk that the United States would intervene. He was probably aided in this misperception by U.S. leaders. Just as the initial North Korean attack on South Korea may have been affected by the absence of a formal U.S. military commitment to aid South Korea (as defined by Secretary Acheson’s “defense perimeter”), Saddam Hussein’s decision was probably affected by statements by the U.S. State Department that the United States had no formal commitment to defend Kuwait. Additionally, at a hastily called meeting between Saddam Hussein and American Ambassador April Glaspie in Baghdad just days before the attack (as Iraqi forces were massing on the Kuwaiti border), the American envoy, adhering to standard policy positions, assured the Iraqi leader of the American desire for good relations and declared that the United States took no position on the dispute between Iraq and Kuwait. 249 The perceptions that war will be economically manageable and militarily winnable and that there will be no third-party surprises all create a sense of optimism that Blainey believes is a key to the cause of war. He argues that it is exceedingly doubtful whether there has been a war since 1700 in which the initial hopes about the impending war were low on both sides. He

Chapter 3

106

concludes that “optimism is a vital prelude to war. Anything which increases optimism is a cause of war. Anything which dampens that optimism is a cause of peace.” 250 Why Misperception? Two Answers Given that misperceptions are plentiful and harmful, why do they occur so often? Social scientists generally point out two categories of bias that might affect perceptions—unmotivated and motivated biases. Unmotivated biases generally derive from the normal cognitive mechanisms by which our brain works to process information. This type of bias is pervasive and largely unavoidable, as we have discussed. The cognitive approach used by Jervis and others “emphasizes the ways in which human cognitive limitations distort decision making by gross simplifications in problem representation and information processing.” 251 An alternative approach is offered by Irving Janis and Leon Mann, who argue that the primary source of perceptual distortion is motivational. 252 Their basic assumption is that we are all emotional beings (rather than rational calculators) with strong psychological needs to construct and maintain positive images of ourselves and our environment. In the midst of a crisis, if there appear to be no alternatives that promise success, the psychological toll becomes too great for leaders. Unable to face the discouraging truth about their options, they retreat to a psychological state of defensive avoidance—in which they refuse to submit their plans to reality testing. Janis and Mann identify three forms of defensive avoidance—procrastination, shifting the responsibility, and bolstering. These are all methods of coping with stress, but each leads to some kind of perceptual distortion. 253 Lebow summarizes the difference between the cognitive approach of Jervis and the motivational approach of Janis and Mann: For Jervis, the starting point is the human need to develop simple rules for processing information in order to make sense of an extraordinarily complex and uncertain environment. Janis and Mann take as their fundamental assumption the human desire to avoid fear, shame, and guilt. Jervis describes cognitive consistency as the most important organizing principle of cognition. Janis and Mann contend that aversion of psychological stress is the most important drive affecting cognition. Whereas Jervis concludes that expectations condition our interpretation of events and our receptivity to information, Janis and Mann argue for the important of preferences. For Jervis, we see what we expect to see, for Janis and Mann, what we want to see. 254

Lebow’s analysis of international crises leads him to the conclusion that the motivational approach provides the better explanation for misperceptions. His study of brinkmanship crises indicates that crisis initiators uniformly (and erroneously) expected the adversary to back down rather than resort to war. Lebow found, however, that an “objective” good opportunity for aggression existed in only a third of the cases. In each case, however, there was a perception that an opportunity existed and there were also strong needs on the part of the initiator to pursue aggressive foreign policies. Lebow suggests that decision makers are more responsive to internal imperatives than to external developments. Aggression would seem to be less a function of opportunity than it is a function of need. 255

The Individual Level of Analysis, Part II

107

Linking Individual Characteristics: Beliefs, Personality, and Leadership Let’s look briefly at a recent study that examines several of the individual-level factors we have been describing and links them to war. In his Why Leaders Choose War, Jonathan Renshon attempts to assess the impact of several individual level factors that move leaders to decide in favor of a particular kind of war—preventive war, undertaken to counter a perceived serious future threat. 256 His study is a qualitative assessment of three cases of preventive war (Britain’s role in the Suez War in 1956, Israel’s attack on Iraq’s nuclear reactor in 1981, and the U.S. attack on Iraq in 2003) and two cases in which leaders contemplated preventive war and decided against it (the Truman and Eisenhower administrations in the immediate postWWII period and India and Pakistan in the post–1971 War period). Renshon hypothesizes that several factors might account for a country’s willingness to use preventive war: (1) declining power relative to an adversary—or at least the perception of decline, (2) the perception that there is only a short window of time in which to act, (3) the belief the situation gives an advantage to the side that strikes first, (4) the belief that war is inevitable, (5) an inherent bad faith image of the enemy, and (6) a black-and-white worldview that includes making stark moral contrasts between good and evil. While all of the factors appear important, Renshon concludes that “leaders—their psychology, their leadership styles, and their judgments—play key roles.” 257 What does Renshon’s examination of the cases reveal? First, choosing preventive war places a premium on strong leadership. In the three cases of war, Renshon’s judgment is that without the strong prowar stances by Anthony Eden in 1956, Menachem Begin in 1981, and George W. Bush in 2003, war would have been less likely. Decisions for preventive war require leaders with a high degree of self-confidence—in their interpretation of the threat and about the morality and utility of the use of force in response. But leadership strength was just as important for those administrations that chose not to use preventive force. Second, the inherent bad faith image of the enemy appears to be crucial. The perception that the opponent is implacably hostile, is unpredictable, and cannot be trusted is pretty much the foundation for a decision for preventive war. But because it is also present in the two cases where war was not selected, what appears to be important is that the bad faith images are particularly rigid, intense, and personalized. 258 The presence of an inherent bad faith image is linked to two other phenomena—the idea that war is inevitable (the stronger the image of the enemy as evil and threatening, the more likely war will be seen as inevitable, at least in the long run) and the use of historical analogies that compare the opponent to figures such as Hitler or Mussolini. All of these components are “heavily dependent on individual psychology.” 259 The bad faith image is closely connected to another psychological variable: black-andwhite thinking. Indeed, the cases where all of these phenomena were present were the three cases of war. Clearly, a Manichaean world view that relies on rigidly categorizing states as good or evil means that once states are coded as evil, they are seen as more threatening. This also has an enabling effect on leaders: “Hard decisions are often surrounded by doubt and uncertainty, but a leader’s view that they are on the ‘right’ side makes hard decision much easier.” 260 Black-and-white thinking also makes it easier for leaders to use rather crude analogies—seeing one’s opponent as a Hitler—as a lens for analyzing the problem.

108

Chapter 3

Third, situations in which leaders choose preventive war or opt to forgo it both involve substantial risk, and therefore propensity for risk (and prospect theory) needs to be taken into account. “Any discussion of preventive war is inherently framed in terms of losses. By the time leaders have decided to consider preventive action, they have already framed a given issue in terms of something bad that may happen in the future, that is, they are in the domain of losses.” 261 Since this is true both for those leaders who choose war and those who forgo it, the decision involves individual psychology, which affects whether something is framed in terms of gains or losses, how a reference point is calculated, whether individuals are predisposed toward taking risks, along with numerous other variables. 262 THE ROLE OF STRESS As if it weren’t enough that we have to worry about leaders whose personality disorders interfere with their ability to make rational decisions, we must also be aware that even normal individuals have difficulties making rational decisions under conditions of stress. Those who occupy political offices—presidents, prime ministers, foreign ministers, national security advisers, counterterrorism advisers, and so on—encounter stress on an almost daily basis. Stress is physiologically induced by the combination of threats, sleeplessness, anxiety, time pressure, frustration, information overload, and the inner conflicts that risky decisions entail. 263 International relations scholars assume that stress is at its height during times of crises. Stress induces a cluster of hormonal and metabolic changes in the human body. Stress reactions are brought on by neuromodulator chemicals called catecholamines (which include dopamine and epinephrine). Our bodies’ automatic reaction to stress makes several things happen very quickly: Adrenaline is secreted; stored carbohydrates begin flooding into the blood with sugar, mobilizing the body’s energy reserves and alleviating the effects of muscular fatigue; the blood becomes more coagulable; the heart beats faster; and breathing patterns are altered. All of this combines to produce an “adrenaline high”—feelings of heightened abilities and of mastery and optimism, all of which might lead to miscalculations. 264 The presence of these chemicals in the bloodstream makes us feel more disorganized and distracted and more reliant on habitual responses. These neurochemicals “shut down the prefrontal cortex, which houses higher-level cognitive functioning, and encourages primary reliance on the amygdala, or the emotional center, to control behavior.” 265 In other words, they give priority to what Kahneman calls system 1. All of these changes prepare men and women for some type of extraordinary physical exertion—for the classic alternatives of “fight or flight.” However, since in modern times the circumstances that stimulate these bodily changes don’t normally lead to the actual expenditures of physical energy, the individual is often left in a frustrated, disturbed, and debilitated state. These ancient mechanisms that evolved to prepare us for stressful encounters may now actually detract from our ability to deal with stressful situations. 266 Let’s return to the example of the Vincennes. In July 1988 the U.S.S. Vincennes, an Aegis cruiser sent to help guard U.S.-flagged tankers in the Persian Gulf during the Iran-Iraq War, erroneously shot down a commercial airliner flying from Iran to Dubai. It had been a stressful day for the captain and crew. U.S. ships in the Gulf had just finished a skirmish with Iranian

The Individual Level of Analysis, Part II

109

boats, sinking two, when radar reported a plane taking off from Bandar Abbas airfield in Iran. Information regarding this plane from radar and electronic intelligence sources was ambiguous, and the captain and crew had only a scant few minutes to determine what to do before the plane would be within range to launch an air-to-sea torpedo at the Vincennes. In an attempt to determine whether the plane might be a commercial airliner, one crew member was given the task of checking the airliner guide for flights from Bandar Abbas; the crewman hurriedly thumbed through the guide and missed the flight (which was seventeen minutes late). Another crewman incorrectly recalled the plane’s attitude from the screen, leading all concerned to believe the plane to be descending toward the Vincennes rather than ascending. Independent psychologists who reviewed the incident concluded that the mistakes were due to a combination of stress, information overload, and a breakdown in communication among the Vincennes staff in the Combat Information Center. 267 Incidents such as these sometimes provide the spark from which wars explode, either because they are used by governments as an excuse for initiating a war that has been long desired or because the incident genuinely arouses a passion for retribution that can seemingly be met in no other way. States respond in anger; this anger is reciprocated; an escalating spiral of violence and counterviolence leads to war. In this particular case, war between the United States and Iran was avoided, but states are not always this fortunate. Experimental simulations and the study of individuals under real conditions of stress both conclude that stress has some rather debilitating effects on the ability of individuals to react rationally to their environment. Table 3.3 provides a short list of the possible effects of stress. 268 Ole Holsti sums up very nicely our knowledge of its nasty effects: The conclusion is sobering: men rarely perform at their best under intense stress. The most probable casualties of high stress are the very abilities which distinguish men from other species: to establish logical links between present action and future goals; to create novel responses to new circumstances; to communicate complex ideas; to deal with abstractions; to perceive not only blacks and whites, but also the many shades of grey which fall in between; to distinguish valid analogies from false ones, and the sense of nonsense; and, perhaps most important of all, to enter into the frames of reference of others. With respect to these precious attributes, the law of supply and demand seems to operate in a perverse manner; as crisis increases the need for them, it also appears to diminish the supply. 269

This is not to say that all stress is bad. Moderate levels of stress are actually beneficial for the average person: they concentrate the mind. Even high levels of stress may increase performance for elementary tasks, though only for a short time. The negative effects of stress generally come at the extreme ends of the spectrum. Low levels of stress fail to focus one’s energies on the problem, and high and protracted stress may impair cognitive functions— especially memory and information processing. Stress especially degrades our ability to deal with complex circumstances marked with considerable uncertainty—exactly the types of scenarios we associate with international crises. Thus, the relationship between stress and the decision-making quality is curvilinear rather than linear and looks like an inverted U. 270 The effects of stress can also be compounded by physical factors such as age, health, and fatigue. But since national leaders tend to arrive in office in their more mature years (this is

110

Chapter 3

being charitable: some arrive well past their prime), the news is not all that wonderful. Increased age is associated with an increased susceptibility to illness and fatigue and with a decreased ability to cope with stress. The only saving grace may be that leaders with past experience under stress may perform better. At any rate, while it would be nice to be able to create a stress-free environment for our political leaders (and make sure they get plenty of rest), it would appear that stress, unfortunately, is just one of those things that come with the territory. Stress also varies according to the individual. While stress is debilitating for some, others seem to thrive on stressful situations. Jerrold Post notes that what threatens one leader may invigorate another. For Israeli prime minister Menachem Begin, external threats from Arab states acted as a tonic, but internal Israeli political criticism was quite stressful. 271 Prime Minister Anthony Eden suffered a severe case of exhaustion at the height of the 1956 Suez Crisis due to stress (and possibly an addiction to amphetamines). And the chief of staff of the Israeli Defense Forces, Maj. Gen. Yitzhak Rabin, collapsed at the height of the 1967 Middle East crisis and needed thirty-six hours of rest. Uri Bar Joseph and Rose McDermott examined the effects of stress on five Israeli political and military leaders in command positions on the crucial second day of the 1973 Yom Kippur War. Two of the four performed poorly (including Defense Minister Moshe Dayan), while two others performed relatively well (including Prime Minister Golda Meir). The authors conclude

The Individual Level of Analysis, Part II

111

that two factors account for their differences. The first variable was a sense of personal accountability or responsibility for the problem. Those who failed to properly prepare for the war performed poorly once it actually began, while those who had taken necessary preparations for the possibility of war performed well. The second variable was the presence (or absence) of a positive social support system. Leaders who were surrounded by supportive staff and others capable of providing them with personal psychological reassurance performed well; leaders who lacked a close team of supporters or who felt they had lost the trust and support of their comrades performed poorly. 272 There is little doubt that personality makes a difference in how leaders respond to stress. As Post says, “Under stress, individuals become more like themselves.” 273 The typical defensive patterns and coping mechanisms used in the past become intensified and exaggerated. Anxious people become more anxious and their cognitive abilities become seriously degraded. Leaders with compulsive personality characteristics (and government cubicles are filled with these types) react with increased attention to detail. Their compulsive characteristics become virtually disabling: They become overwhelmed by detail, unable to see the big picture, and indecisive, seemingly paralyzed by doubt. In the end, the decision is most likely to be the result of frantic impulse rather than careful weighing of evidence and viewpoints. Leaders with narcissistic personality characteristics (also highly represented in the councils of world leaders) follow a different path to dysfunction. The mild narcissistic personality features of some leaders may bloom into full-blown character disorders. Their extreme selfcenteredness and egocentricity mean that they are not likely to admit ignorance or accept constructive criticism. By itself this is problematic, but narcissists are also likely to have surrounded themselves by toadies who are well-practiced in telling the leader what they think he wants to hear instead of supplying him with the unvarnished truth. Thus they tend to be out of touch with political reality. The effect of stress on leaders with paranoid characteristics is that their negative characteristics—their suspiciousness, their stubbornness and rigidity, their unwillingness to compromise, their defensiveness—are all intensified and become dysfunctional. Their heightened suspiciousness leads to the construction of worst-case scenarios and the conclusion that compromise is impossible and that preemptive force is necessary. 274 Those leaders whose personality profile includes risk-acceptance are generally more comfortable with stress, and high-risk situations may actually bring out the best in them, whereas risk-averse leaders may find their abilities much more impaired. But even in risk-acceptant personalities, there is a level of stress that will adversely affect judgment; it’s just that the threshold is at a much higher level. 275 While most evidence—from experimental studies or qualitative case studies—suggest the negative effects of stress on decision making, a few studies point in the other direction. Brecher and Wilkenfeld’s study of international crises provides some limited evidence that high levels of stress evoke greater attentiveness by decision makers to the actions of others, leading to appropriate responses. Their study indirectly infers the level of stress felt by decision makers based on two variables: (1) a ten-point scale judging the level of threat to the state and (2) a ten-point scale ranking the differences in power between the two states involved— with unequal power (small power/superpower) representing the greatest threat, though cases involving two superpower adversaries are also ranked at the most dangerous level. They were

112

Chapter 3

interested in the degree to which the actions of adversaries matched each other’s behavior: Did a crisis with a nonviolent trigger event lead to a nonviolent response? Did a crisis triggered by a violent act lead to a violent response? Their data show that actions were most likely to be appropriately matched under conditions of high stress, while conditions of low stress were more likely to lead to unmatched responses (nonviolent triggers leading to violent responses). 276 Finally, Vertzberger’s analysis of high-risk foreign military intervention decisions indicates that stress actually had a rather limited impact on decision making. 277 CONCLUSION This chapter has examined those factors that might explain the differences in the behavior of individuals regarding war: different psychological needs, different personality traits, cognitive biases inherent in the way the brain processes information, differences in willingness to take risks, different perceptions (and misperceptions) of the environment and of one’s opponents, different images of the world and operational codes, and differences in ability to change or adjust present images. Several points should be made in summary. First, virtually all the theories and concepts discussed in this chapter undermine normal assumptions about the ability of government leaders to make rational decisions. They support the notion that decisions for war or the escalation of conflict are likely to result from nonrational, subconscious, or emotional processes rather than purely logical calculation. Errors might result from the need for cognitive consistency and the failure to incorporate new data into old beliefs, from cognitive limitations, dysfunctional operational codes or schemas, motivated or emotional bias, reliance on inappropriate analogies, the framing of reference points and options, feelings of low self-esteem, stress brought on by fear and anxiety, and so on. Determining which of these departures from rationality is most responsible for a particular policy error is difficult. 278 Second, it seems likely that many of these individual-level variables are highly interrelated. If, for the sake of argument, we identify a major cause of war to be an overestimation of an opponent’s hostility, coupled with a predisposition (in spite of tremendous risks) to confront that opponent with hostile actions in the hope of forcing him to back down, then several types of variables at the individual level—cognitive factors, personality variables, and emotional variables—might combine to produce such a situation. We have seen how psychological needs might foster misperceptions concerning the risks inherent in a crisis situation and about the relative military capabilities of the states involved. It is also likely that operational codes will influence the leader’s perceptions of the opponent and of the strategies most likely to be effective in dealing with that opponent. Leaders with a hawkish operational code are more likely to perceive opponents as aggressors who will back down in the face of bullying tactics. They therefore are predisposed to follow hardline, brinkmanship-type tactics against others. Actions by rivals may trigger analogies and historical lessons that serve to heighten the calculation of threat and the need to use force to counter it. Leaders whose personalities are characterized by risk acceptance or by dominance or power motivation will be inclined to pursue tough, high-risk tactics vis-à-vis their international rivals. Likewise, leaders with narcissistic or paranoid personalities or with active ego defense

The Individual Level of Analysis, Part II

113

mechanisms are prone to discount warnings of risk in their pursuit of foreign policy. Finally, a leader’s inability to change his or her incorrect image of opponents is likely to be influenced by personality variables and psychological needs as well as cognitive variables such as image structure, use of heuristic short-cuts like analogies, and the content of one’s operational code. Third, we have a lot of evidence from case studies (though some of it is fairly unsystematic) that misperceptions have frequently accompanied decisions by national leaders to go to war. Determining a leader’s perceptions (and her psychological state) in time of crisis is fraught with danger. Any such analysis should be viewed with caution. But the evidence from an extremely large number of cases indicating that decision-making elites have been the victims of misperceptions is virtually overwhelming. What we know from these cases is that in a sizable number of instances, elite misperceptions appear to have played a crucial role in the decision to go to war. What we don’t know, and perhaps can’t know, is why these misperceptions occurred. What would have happened if they had not? Would war have been prevented? How widespread are such misperceptions across the universe of cases of interstate war? Neither can we at this time be precise about the exact relationships between misperceptions and other individual-level variables. Finally, none of the ideas we have encountered here add up to a fully developed theory of war. We do not as yet have a psychological or cognitive theory of war. Individuals matter, but how, when, and why they matter is not very easily or very parsimoniously incorporated into theories of war. Nevertheless, these individual-level variables are often important enough that if they are omitted from our explanations, we will fail to understand the reason for war. Perhaps Herbert Kelman has put this in perspective best. He has argued that there can be no really autonomous psychological theory of war and international relations, only a general theory of international relations in which psychology plays a role. 279 While it would seem that psychological and cognitive variables may not always play the determining role in whether war breaks out, they are certainly quite important. As we will see in later chapters, psychological and cognitive factors are also highly interrelated with variables at other levels of analysis. This increases their importance as elements of a theory of war. If national decisions for war really are affected by the factors we have identified in this chapter, what are the implications for solving the problems of war? If psychological factors really are that important, one implication might be that selection for high office ought to be conditional upon passing a rigorous psychological examination. And frequent psychological checkups for the political elite ought to be instituted just as often as physical checkups. Unfortunately, political candidates are about as likely to submit to observation and testing by professional psychiatrists as they are to refuse campaign contributions from wealthy admirers. In the final analysis, states are probably best served simply by developing procedural checks on decision making so that single leaders—whether or not they suffer from psychological or emotional dysfunctions—cannot on their own make the momentous decision of war and peace. Let us switch gears for a moment. We have assumed thus far that the key element in the cause of war is the role of the individual national leader. However, this may be an oversimplification. Decision making is not always an individual process; frequently it is a collective political process. Government decisions are typically made by small groups of people. If this

114

Chapter 3

is the case, then the role of the individual may be less important than how the members of the decision-making group interact with each other to determine government policy. We must move on therefore to explore explanations of war at the next level of analysis—the small group.

Chapter Four

The Substate Level of Analysis Group Decision Making

Madness is the exception in individuals but the rule in groups. —Nietzsche The process was the author of the policy. —Under Secretary of State George Ball, 1962

In this chapter we explore the causes of international conflict that may be found at the substate level rather than the individual or state level. The basic premise here is that, within modern governments decisions are frequently made by cabinets, politburos, juntas, or committees rather than by presidents, prime ministers, or generals acting alone. Policy making is very much a group activity. Thus, our search for the causes of war leads us to an examination of the methods by which small groups of government leaders make decisions about war and peace. The central assumption at this level of analysis is that the structures and processes of decision making, rather than the characteristics and perceptions of the individuals involved, have the decisive effect on government policies and behaviors. THE ROLE OF RATIONALITY IN SMALL GROUPS: THE RATIONAL ACTOR MODEL As noted in the previous chapter, political scientists have tended to assume that governments (like individuals) make decision in a rational manner. The rational actor model (RAM) includes the following steps. 1. Identify and define the problem. 2. Identify and rank goals. Often, policy makers will have more than one goal; this introduces the problem of value complexity. If multiple goals exist, they will have to be ranked (prioritized) in order of their importance. This is necessary because it is quite likely that no policy option can achieve all goals and that the choice of any particular policy might mean that some goals will have to be sacrificed in order to attain others. Rationality requires that decision makers make conscious value trade-offs. 1 3. Gather information continuously. 115

116

Chapter 4

4. Identify many alternative means/options for achieving the goals. 5. Analyze each possible alternative. This step requires several types of evaluations. Decision makers must estimate and weigh the likely costs and benefits of each alternative, examine the potential consequences of using each option, calculate the effect of each option on other goals, determine the relative effectiveness of each option for achieving the goals, and estimate the probable success of each alternative. The calculation of probabilities is made necessary by the fact that decisions—especially in the realm of choices about peace and war—take place under conditions of tremendous uncertainty, due to incomplete information about the intentions and capabilities of others and their potential behaviors. The result is that it is not possible to make an absolutely accurate assessment of the consequences of enacting a particular policy option. Thus, all outcomes have to be addressed as probabilities. This is why we have to refer to expected utility of a particular option. We expect an option to yield a result of such-and-such a nature, but we cannot be certain. 2 6. Choose the alternative best able to achieve your objectives. In other words, select your optimal strategy—the one most likely to maximize your interests. Under conditions of uncertainty, leaders should choose the option that has the greatest value when both the payoffs and the probabilities of obtaining them are considered. 3 The option selected should be the one that promises, on average, to deliver the highest expected utility (payoff). Like all models, the rational actor model requires certain simplifying assumptions. Proponents of the RAM assume, among other things, that governments can be thought of as single (unitary) rational actors—that the decision-making process within government proceeds as if there was a single, dominant leader who controls the outcome. It also assumes that decision makers are generally able to place a concrete and meaningful value on each outcome so that two or more outcomes can be compared. It further assumes that decision makers can calculate the probability that a desired action will lead to success. Bruce Bueno de Mesquita argues that although actors are assumed to choose those options that will lead to the “best feasible outcome for them defined by their personal values and preferences,” the interests that they value may or may not include enhancing the national interest. The world is full of leaders who govern for their own benefit at the cost of the collective interests of their subjects. 4 A final assumption is that if a rational decision process is adhered to, then unwanted and unnecessary wars can be avoided. As we will see in chapter 9, many rational choice theorists agree with James Fearon that if all relevant participants have access to perfect information, then wars should not occur; there should always be a potential bargain that is better for all states than the expected utility of war. Potential initiators and target states alike should back down when confronted with an outcome (war) that presents high costs and for which there is a better alternative. 5 THE ROLE OF RATIONALITY: EXPECTED UTILITY THEORY We should not discard the possibility that the primary cause of war is simply the calculation by government leaders that war was the best option for achieving the goals and interests of the state under certain conditions—that by going to war the country will be better off than if it had

The Substate Level of Analysis

117

not gone to war. From this viewpoint, wars are intentional and rational; they result from careful cost-benefit analyses of alternative strategies. 6 This is essentially the argument of rational choice theorists or, more properly, expected utility theorists. Leaders are assumed to be goal-oriented: they are guided by the desire to maximize the net benefits they expect from their foreign policy choices. In other words, a “policy maker will never choose an action that is expected to produce less value—or utility—than some alternative policy.” 7 Leaders will decide in favor of war only if the expected utility of doing so is greater than the expected utility of not going to war. The decision for war depends on an estimation of the costs and benefits of war. Specifically, “the size of the expected gains or losses depends on: (a) the relative strengths of the attacker and defender; (b) the value the attacker places on changing the defender’s policies, relative to the possible changes in policies that the attacker may be forced to accept if it loses; (c) the relative strengths and interests of all other states that might intervene in the war.” 8 Bruce Bueno de Mesquita has formulated and reformulated various expected utility models over the years and has used them to test various hypotheses about decision for war. In The War Trap, he postulates that a necessary (but not sufficient) condition for war is that the gain in expected utility for winning the war should be greater than the expected loss in a defeat. Testing his model against war data from the nineteenth and twentieth centuries, he confirms that decisions for war conform to what we would expect if such decisions are the products of rational calculation. Specifically, he found strong support for the hypothesis that measures of utility correlate highly with both escalation of disputes and the outbreak of war. It appears that states will not go to war if the expected gains are smaller than the expected losses. Bueno de Mesquita concludes, “Almost all initiators of international conflicts appear to have a reasonable expectation of success in their conflicts, while most opponents did not.” However, in the final analysis, “most situations involving a nation with a positive expected utility and its potential foe result in no overt conflict.” 9 In other words, wars are relatively rare events to which alternatives are usually found. Publication of The War Trap elicited many criticisms regarding how the model was formulated and tested, leading to further refinements. In War and Reason, Bueno de Mesquita and Lalman use a two-state (dyadic) model of decision making for war called the international interaction game (IIG). 10 A sequence of interactions between a potential initiator (A) and the potential target (B) may lead to eight possible outcomes, depending how the two sides react to the policies of the other. At each decision point A and B choose the option that would appear to generate the greatest expected utility, assuming complete and perfect information. The domestic politics variant of the model incorporates the idea that state leaders are subjected to domestic pressures and they understand that there may be important domestic political costs of using force. Using data on over 700 interactions between pairs of European states (dyads) from 1816 to 1970, they find support for the general idea that states attempt to maximize their subjective expected utility in making choices for war. Bennett and Stam retested results from War and Reason, using a larger database that includes data on militarized international disputes (MIDs) for all countries from 1816 to 1984. 11 For each dyad year they created a predicted outcome based on Bueno de Mesquita and Lalman’s IIG model, then compared the predicted outcome with the real historical outcome.

118

Chapter 4

When they examined only politically relevant dyads (pairs of countries that are contiguous or pairs in which at least one state is a major power) they found that the predicted outcomes were significantly correlated with real state behavior. Specifically, the mutual use of force (short of war) was approximately 2.2 times more likely to occur in dyads in which the IIG’s predicted outcome is war than in dyads predicted to maintain the status quo. When the IIG predicted war, there was less negotiation, more abject capitulation, more use of force, and more war. When war was predicted but did not occur, there was a lot of “backing down” by targets, perhaps in anticipation that war would occur if they continued. When control variables were added, the strength of the relationship was weakened but not eliminated. The most recent version of Bueno de Mesquita’s model, found in The Logic of Political Survival, incorporates a “theory of the selectorate.” 12 Whether a country is democratic or nondemocratic, its leaders are motivated primarily by the desire to survive in office, and this is deemed to be a critical consideration when leaders decide questions of war and peace. In all states, there is a subset of residents with the necessary characteristics and qualities required to choose the state’s political leaders; this group is termed the selectorate. In a democracy the selectorate is assumed to be approximately equal to the number of voters. In nondemocratic countries the selectorate is a much smaller group. (And, of course, to be eligible for leadership, or for membership in the winning coalition, one must belong to the selectorate.) Capturing office and keeping it depend on the support of a winning political coalition—a subset of the selectorate of sufficient size to effectively place in power the government’s leaders, by whatever means—from military coup to mass elections. Threats to a leader’s position may come from the rise of domestic challengers, from revolutions at home, or from threats from a foreign adversary. To stave off these challenges and remain in power, leaders provide benefits to those members of the selectorate who are part of the winning coalition. These benefits can take the form of private goods—government jobs, cash payments, tax benefits, government contracts, and so on. Or they may take the form of public goods—government policies from which all people benefit, not just members of the winning coalition—such as physical security, national health care, or subsidized prices for gasoline and vodka. The model posits a substantial difference between democratic systems and nondemocratic systems. Democratic systems are characterized by the existence of a large selectorate and therefore the necessity for a relatively large winning coalition that calls for wider payoffs in the form of public goods. Nondemocracies have much smaller selectorates and smaller winning coalition sizes, making it possible to use private goods to buy off the elite members of the winning coalition. The differences appear to have important implications for war, which Bueno de Mesquita and his colleagues are able to demonstrate empirically. We will investigate this in a later chapter. None of this is meant to suggest that wars that are begun “rationally” are always won by their initiators. This is frequently not the case. And since so many initiators actually end up losing, their choice appears, in retrospect, to be irrational. Why is this the case? The easy answer is that once war begins, the initiator loses control over what happens next. Another answer is simply that it might be rational to fight a war even though the probability of winning is pretty low. States choose the option with the highest expected utility. What this might mean

The Substate Level of Analysis

119

is that the war option—even with a very low probability of success—may have a higher expected utility than the certainty of being beaten and occupied by an opponent. The war option has a low probability of success, but it may appear better than other potential options. 13 WHY GOVERNMENTS DON’T USE THE RATIONAL ACTOR MODEL In the last chapter we asked why individuals don’t make rational decisions; in this chapter we ask why government organizations don’t make rational decisions. Obviously, we all hope our governments apply something close to the rational actor model to their foreign policy problems. Unfortunately, in the real world it often becomes impossible for governments to use the RAM in making their decisions. The best methods are not always used and the best results are not always obtained. Why aren’t decisions made through a strictly rational process? The following is a short list of impediments to the RAM, stemming from factors at both the individual and the substate level of analysis. 1. As we have seen, psychological factors interfere with the ability of government leaders to engage in rational problem solving. To the extent that these factors exist, decisions will be made in order to meet the subconscious needs of political leaders rather than to address legitimate national security needs. 2. Analytical processes are affected by unmotivated biases—normal cognitive heuristics or shortcuts such as the use of reasoning by analogy, reliance on the fundamental attribution bias, the use of stereotypes, risk avoidance, and so forth. 3. We also know that misperception, if not exactly rampant among policy makers, is at least pervasive. Attempts to resolve foreign policy problems require accurate images of the situation—a condition not always met. 4. Tough policy choices are frequently accompanied by hefty amounts of stress. Stresses that can impair rationality are often abetted by lack of sleep and physical infirmities among the political elite. None of this is conducive to good decision making. 5. Decision makers may not always have the quantity or quality of information required to make a perfectly rational decision. National security policies are frequently made in an environment of uncertainty; information may be false, missing, contradictory, or ambiguous. The most important information—the intentions of others—is probably unknowable from normal sources of intelligence. Alternatively, information may be so overwhelming in quantity that analysts are simply overloaded in their ability to make sense of the product of the intelligence services. This creates a signal-to-noise ratio problem. This is the idea that when policy makers confront important issues of national security, a huge amount of noise (irrelevant and potentially misleading information) frequently overwhelms signals (useful, relevant information). 6. The amount of time available for making the decision may be limited, thus impairing the ability of policy makers to develop and analyze options. Even if adequate time is available, leaders may choose not to take it. The use of time entails costs, and a speedy decision may be seen as preferable to a better decision that takes longer. 7. Evaluation of policy options requires that decision makers peer into the future and predict the likely outcomes of various alternatives. Most of us probably have enough

120

Chapter 4

trouble trying to figure out the past and the present; trying to imagine the future entails more insight than we can muster. If you hope government experts might be better at this than others, you would be wrong. Philip Tetlock’s research shows that being an expert in the field doesn’t help in making these predictions; “expert judgment” is just as fallible, if not more so, than predictions by “normal folk.” 14 8. Identifying all the goals the state might pursue and then prioritizing them may be beyond rational assessment. 15 The selection of goals frequently means choosing the lesser of several evils. Often the pursuit of one goal can only be achieved at the expense of others: These value trade-offs are notoriously difficult to make, and leaders tend to avoid making them. 9. Identifying all the alternatives and subjecting them to a cost-benefit analysis is an incredibly daunting task. For practical reasons, such as lack of time and lack of human insight, the search for and analysis of options is likely to be limited. A special problem in foreign policy decisions is the perceived need for secrecy. In an attempt to prevent leaks of sensitive information, top decision makers frequently choose to restrict the number of people consulted, thus severely limiting the range of alternatives as well as the variety of analysis. McGeorge Bundy suggests that this was a typical problem for American decision makers in the Cold War age, but it also bedeviled Soviet leaders. Khrushchev, for instance, apparently consulted only one diplomat (Gromyko) for his analysis of the decision to put missiles in Cuba in 1962. 16 Snyder and Diesing’s study of decision making in sixteen international crises discovered that most alternatives were either ignored or quickly eliminated as unfeasible by preliminary scanning, and only a few options were really seriously considered. 17 In an extremely interesting study of the decision-making processes of the Truman, Eisenhower, and Kennedy administrations (concerning Korea, Vietnam, and the Cuban Missile Crisis), Paul Anderson finds that, although decision makers did consider a surprisingly large number of alternatives, very few of the alternatives were actually mutually inconsistent. A large number of alternatives, once proposed, were simply ignored without analysis; they essentially “died for lack of a seconding motion.” Additionally, when incompatible alternatives were identified, they were usually in reaction to a previous suggestion with which a member of the group disagreed, and consisted of “don’t do X” rather than a separate course of action. Perhaps most important, Anderson also found a tendency to search for goals after alternatives had been identified and approved. 18 10. A final impediment to the functioning of the RAM deals with the actual choice of the preferred policy. If several individuals were to go through our rational process model with regard to a particular policy problem, each would probably select a different optimal solution. It is quite possible that one individual’s “rationality” is not the same as another’s. Different individuals or subgroups within the decision-making body will rank goals differently, will prefer different outcomes, and will support different policy options. Here we return to the central assumption of the small group level of analysis: governmental decision making is usually a group sport rather than an individual activity. There will inevita-

The Substate Level of Analysis

121

bly be disagreements between cabinet ministers about the best possible solution to the problem and the best policy to pursue. There may be even more fundamental differences over the correct goals to pursue, leaving aside the methods by which these goals are to be met. The bottom line is that decisions require bargaining and compromise between the members of the group. Policy making is ultimately a political process rather than a cognitive, rational process. And, to the extent that decisions are political, the probability of poor (that is to say, nonrational) decisions increases. In conclusion, we must be skeptical of the ability of governments to make good, rational, reasonable decisions. Lousy decisions get made, nonrational decision-making processes abound, and every once in a while governments will bring themselves closer to war because the decision process failed and the policy factory turned out a real lemon. THE ROLE OF RATIONALITY: BOUNDED RATIONALITY THEORY If government decision makers are not rational in the strict sense that would be recognized by rational choice theorists, in what sense are they rational, if at all? To answer this question, Herbert Simon has coined the term bounded rationality. 19 The concept of bounded rationality recognizes the inherent limitations that we have just discussed, regarding limited access to information, limited cognitive ability to process information, limited attention, limited ability to predict probabilities, limited time, and so on. In Simon’s words, bounded rationality describes a person with limited computational capacity, and who searches very selectively through large realms of possibilities in order to discover what alternatives of action are available, and what the consequences of each of these alternatives are. The search is incomplete, often inadequate, based on uncertain information and partial ignorance, and usually terminated with the discovery of satisfactory, not optimal, courses of action. 20

In sum, bounded rationality describes a process or procedure for making decisions that is not necessarily optimal, but is as good as can be expected given human limitations and real-world circumstances. This kind of problem-solving is not irrational but “imperfectly rational.” Several concepts help to explain how real-world decision makers operate in this world of bounded rationality. Decision makers use satisficing and sequential search, incremental policies, cybernetic approaches and rely on standard operating routines, and they use poliheuristic approaches, all of which we will explore next. Bounded Rationality: Satisficing and Sequential Search Studies of corporate decision making by Herbert Simon, James March, and Richard Cyert indicate that the number of alternatives that are considered is frequently limited by the twin processes of satisficing and sequential search. 21 Decision makers don’t really lay out all the options and assess the pros and cons of each. Instead, the possible alternatives are examined one-by-one in a sequential manner (sequential search) until an option is discovered that meets some minimal standard of acceptability (satisficing). The options are not ranked from best to

122

Chapter 4

worst; they are simply categorized as acceptable or unacceptable. The first option discovered that fits into the “acceptable” category is the alternative that is chosen. 22 (It was this insight that earned Simon the Nobel Prize in economics in 1978.) It is quite likely that a better option could be found if the search were to continue and other options analyzed; indeed, the optimal solution may be the very next alternative. But the decision makers will never know because the selection process ends with the first acceptable solution. Why are such processes employed? Simon and March claim that satisficing and sequential search are devices designed to simplify and expedite the decision process. Executives realize that time is short and that long decision processes are costly. Furthermore, the perfect solution may never be found. Since it is difficult if not impossible to compare the value of two outcomes, there is no rational process by which the best outcome can be ascertained. 23 Under these circumstances, it is better to seek a solution that is merely acceptable, rather than engage in a drawn-out process that may become a wild goose chase. We have some indications that government policy making has in fact followed these practices. Anderson’s study of the Korean War, Vietnam War, and the Cuban Missile Crisis concludes that American decision makers did not consider each alternative (or each subset of alternatives) before making a final decision; they considered alternatives sequentially, making a yes-no decision about each in turn. 24 Bounded Rationality: Incrementalism Many decades ago David Braybrooke and Charles Lindblom made a major impact on political thinking with their analysis of how political decisions were made with their theory of incrementalism. 25 For many of the reasons already cited, decision makers rarely use the RAM. Instead, they use a stripped-down, simplified, incremental process. In this process, major changes and comprehensive reforms are left off the table, and only those options that differ in small degrees from the present policy are fully scrutinized. The result is that most policies vary only marginally (incrementally) from previous policies. This is shown schematically in figure 4.1. The policy options that are likely to be given the most serious consideration will usually be those that differ from the present policy (M) in only marginal ways. Policy options A and Z, and even G and T, are unlikely to be evaluated. Instead, policy makers will most likely look closely at only K, L, N, and O. These options will result in policies not too much different from the present policy. Why is this method used rather than the RAM? First, this method is less complex. Decision makers only need to analyze the differences between the old policy and a few new ones. Presumably, they know—roughly—how the present policy functions, what its results are, to what degree it is successful, and what its defects are. In other words, they understand the consequences of the present policy. Comparing the present policy with only marginally different proposals makes it possible for analysts to be more confident in their predictions about how the new policy might work, as the consequences are likely to be only marginally different as well. In one sense incremental decision makers are acting rationally; they acknowledge their understanding of the problem is limited and they act to limit the damage that low understanding might have on policy making.

The Substate Level of Analysis

123

Allowing only small policy changes brings a second benefit: major mistakes can be avoided—at least in principle. By making incremental choices, policy makers can be reasonably sure they are not choosing a policy that will be a serious blunder. Policy makers recognize that unanticipated consequences will always turn up and that these are more likely when the policy chosen is dramatically different from the present policy. The selection of A or Z might lead to a monumental disaster! It is also true that the selection of A or Z might turn out to be the best choice that could possibly be made. The problem is that one can’t be absolutely sure, since adopting such a policy would constitute a major change. Braybrooke and Lindblom imply that policy makers are motivated more by avoiding policy disasters than by achieving policy successes. As a result, incrementalism can be seen as a “damage limitation” approach to policy making or as a “muddling through” strategy. Government policies are continuously made and remade by successive marginal adjustments in an endless series of trial-and-error experiments. The third reason for the popularity of incrementalism has to do with the nature of politics. The choosing of alternatives by groups requires that accommodation, compromise, and coalition building take place. It is not likely that too many coalitions will be built around alternatives (like A or Z) that require radical changes in present policy; neither are these options likely to be selected as compromise choices. It is much easier to build broad-based support for options that resemble the current policy; just about everyone can agree that L or N represents a good compromise position between keeping the present policy and making a major change. In other words, incremental decisions are politically easy. Choosing non-incremental policies usually requires overturning a previous incremental decision based on some carefully crafted package of political compromises. Non-incremental choices are therefore politically difficult. The result is that policy makers are not necessarily seeking the best choice; they are instead pursuing the safest choice. This may not be such a bad thing. After all, we certainly prefer that our governments do not rush hastily into conflict through rash judgments. Especially when faced with a decision that might result in war, political leaders ought to be wary of choosing options that might have unforeseen or unintended consequences; so perhaps they ought to stick to the safe path and proceed incrementally. On the other hand, Baybrooke and Lindblom argue that policy areas such as wars and revolutions fall outside of the types of decisions that are made incrementally. By their very nature they entail changes that are quite large and important; thus, they are not susceptible to incremental decision making.

Figure 4.1. The Incremental Model of Policy Choice

124

Chapter 4

Incrementalism and Vietnam The American involvement in Vietnam is frequently depicted as having come about through incremental decisions by several successive administrations. In these analyses, incremental decision making is seen at best as inappropriate and at worst as the source of disaster. Leslie Gelb and Richard Betts, in their iconoclastic The Irony of Vietnam: The System Worked, 26 argue that decisions by each administration concerning the American involvement in Vietnam were highly incremental. U.S. involvement increased step by step as each decision deepened American commitment and increased the ante of money, material, and manpower. The most important decisions included the Truman administration’s decision to support with military assistance the French effort to retain Vietnam as a part of the French colonial empire against the nationalist Vietminh; the Eisenhower administration’s decision, after the end of French rule, to support the newly created South Vietnamese government against the rebellious Vietcong; the Kennedy administration’s decision to vastly increase the number of U.S. military advisers to South Vietnam; and the Johnson administration’s decisions to begin an aerial bombing campaign against North Vietnam, to send U.S. combat troops to fight in Vietnam, and then to steadily increase the numbers of these troops. Gelb and Betts describe American strategy as “doing what was minimally necessary not to lose” until 1965 and then “doing the maximum feasible to win within certain domestic and international constraints” after that. 27 Over the course of time the floor on the minimum necessary was raised by conditions in Vietnam; later, domestic and international constraints were increasingly discarded to stave off defeat. The result was an incremental escalation of the war effort. What was responsible for the incremental nature of these decisions? First, U.S. presidents understood that incrementalism would preserve their flexibility by keeping their options open. A major jump in U.S. commitment would reduce the number of future choices available to the president, whereas incremental policies would give the president something to do next time. Second, incremental choices were also compromise choices. Especially for Johnson, who desired consensus from his advisers, compromises were politically pragmatic; they represented a way to avoid political defections and to minimize the intensity of dissent. Each president chose middle-ground solutions between positions of more hawkish and dovish advisers. Incremental, middle-of-the-road policies kept both extremes on board: “The Right was to be given escalation. The Left was to be given occasional peace overtures. The middle would not be asked to pay for the war.” 28 The middle ground was politically safe and smart. A small step that maintained momentum toward the final goal gave each president the best chance to gather more political support before going forward. Third, incremental choices were pragmatic, a way of playing it safe. They gave the appearance of minimizing potential mistakes. Incremental escalation represented the middle of the road between the high-risk alternatives of losing the war or doing what was really necessary to win. Gelb and Betts believe that the incremental process worked: the series of decisions on Vietnam were rational rather than irrational. Political leaders were not fooling themselves; this was not a case in which American leaders believed that each small step would be the last one necessary to gain victory. 29 Gelb and Betts argue that American leaders were not overly

The Substate Level of Analysis

125

optimistic and were not deluded by reports of progress—as some have suggested. The United States did not march into a quagmire because its leaders were unaware that there was any muck in the area; they knew it was a swamp. The escalation was not a blind slide down a “slippery slope”; instead, American escalation was a rational response to the progressive escalation of the minimum price of keeping our commitment. In other words, American decision makers understood that they were operating in a policy area of high uncertainty and low understanding, and in those situations it was quite rational to use an incremental decision style. If one takes as a given the U.S. commitment to prevent a communist victory in South Vietnam (and this is a big given), then the incremental policies designed to attain this goal may be seen as logical and appropriate responses to changing national and international conditions. Bounded Rationality: Poliheuristic Theory of Decision Making Poliheuristic theory (PH) is a relatively new perspective on decision making developed by Alex Mintz and others. 30 As we have seen, in a world of bounded rationality, decision makers use shortcuts (heuristics) to help them deal with all the problems inherent in international relations—time constraints, incomplete information, cognitive limitations on information processing, and high levels of complexity. Poliheuristic (literally, many shortcuts) theory argues that policies are made in a two-stage process. The process is deemed to be applicable to decisions by single individuals or by groups. In stage 1 the menu of potential options faced by decision makers is cut down to size by using a rather simple heuristic: All policy options are first subjected to scrutiny on a single dimension. At this stage, all options are rejected that are unacceptable to decision makers based on that one critical dimension (or perhaps a small number). Usually this dimension is domestic politics, though it could be something else. The decision calculus at this stage is noncompensatory. In other words, the negative evaluation of a particular option with regards to one criteria (typically, domestic politics) cannot be compensated for by high scores or significant benefits on other criteria (economic, diplomatic, or military/strategic factors). Any option that falls below some “cutoff” level on the political dimension is screened out as not feasible and is automatically dropped from further consideration. The remaining options have been “selected in” through a satisficing process: They have met a minimum standard of acceptability at stage 1 and their respective merits are analyzed by a different and more rigorous process in stage 2. In stage 2, a choice is made from the remaining alternatives in an attempt to maximize the benefits and minimize the risks, using a much more analytical process. This process resembles (more or less) the rational actor model/expected utility theory approach to making decisions. In this stage options are assessed in a compensatory way: An option’s negative evaluations on one dimension can be offset by positive evaluations on others. Since most government decisions in the realm of international politics are not “one and done” but sequential in nature, PH theory argues that each decision in the sequence is a two-stage poliheuristic decision. 31 Mintz has proposed that two somewhat different types of decision strategies may be used in stage 2. 32 Decision makers may use an expected utility calculus that balances the costs and benefits of each option against the others. Alternatively, they may apply lexicographic choice,

126

Chapter 4

in which the choice is made in favor of the option that maximizes utility on the dimension regarded as most vital (which may, once again, be the political dimension). For instance, at stage 2 of the process in 1954 President Eisenhower decided not to use U.S. airpower to assist the French fortress at Dien Bien Phu. Even though U.S. airstrikes looked superior on the military/strategic dimension, on the more important political dimension the best option was to forgo direct U.S. military intervention in the French war in Indochina. 33 At any rate, supporters of PH suggest that both cognitive and rational choice models are needed to provide a complete picture of the decision-making process. Stage 1 operates on principles consistent with bounded rationality; stage 2 comes closer to approximating principles of rational choice. Steven Redd’s study of the Clinton administration’s decision to use force in Kosovo suggests that even at stage 2, noncompensatory political considerations are important factors in determining the choice among options. 34 Under great pressure to do something effective to stop the ethnic cleansing by Serb government forces in Kosovo against ethnic Albanians, President Clinton determined in stage 1 that four options—(1) doing nothing, (2) “containment” of the Milosevic government through sanctions and diplomatic pressure, (3) allowing the UN to take the lead, and (4) arming the Albanian Kosovar rebels—all should be rejected outright, largely for political reasons. The options that survived the first cut were to use either air power or ground forces to remove Serbian forces from Kosovo. According to Redd, the air power option was chosen primarily on political grounds. Although its probability of success was lower than the use of ground forces, the political costs of using ground troops could well exceed what was acceptable; too many troops would have to be sent and too many soldiers would die. Public opinion was strongly opposed to this option, and its price tag would be too high. Poliheuristic theory clearly sees domestic politics as “the essence of decision.” 35 The fundamental priority of political leaders is to avoid major domestic political losses; they are highly unlikely to select policy alternatives that will harm them politically. For politicians, political losses are indeed noncompensatory; that is why they must be avoided. This axiom applies to leaders of democracies but also to leaders of nondemocratic states, who must also be concerned about their political survival. No leader wants to take a direct hit on his or her political power or chances of remaining in office. Notice that the priority of political survival is also a major principle incorporated into expected utility theory by Bruce Bueno de Mesquita. It is, of course, also consistent with the findings of prospect theory that people tend to be loss-averse; they are more interested in preventing losses than acquiring gains. PH theory proposes that if a decision to use military force will have unacceptable domestic political results, it will be rejected early in the process, even though it might lead to substantial diplomatic, military, or economic benefits. The flip side of this is that if military options are deemed to have positive results politically, they may be accepted even though their cost/ benefit ratio may be considered low on military and other dimensions. Indeed, it is possible that the military option may be the only option that passes the step 1 test. 36 Supporters of the theory have argued that PH accurately describes the decision-making process in quite a few cases: the Bush administration’s decision to use force against Iraq in 1990–91 37 and its decision not to continue the 1991 Gulf War once Iraqi troops had been

The Substate Level of Analysis

127

pushed out of Kuwait; 38 the Eisenhower administration’s decision not to use force to rescue French forces at Dien Bien Phu; 39 U.S. diversionary uses of force; 40 Pakistan’s decision to test nuclear weapons; 41 Chinese choices in foreign policy crises between 1950 and 1996; 42 the Clinton administration’s decisions on the use of force in Kosovo; 43 and U.S.-China and U.S.Iraq bilateral relations. 44 This last study of U.S. bilateral interactions with China and with Iraq shows some interesting results. The authors note that domestic politics may not be the only noncompensatory dimension for policy makers. They hypothesize that in dyadic crises situations, empathy with one’s opponent—especially with their evaluations of the options we might propose—is an important factor helping to resolve crises and prevent escalation. Empathy at stage 1 may be part of on side’s noncompensatory criteria: “leaders eliminate any option that is not minimally acceptable to the opponent.” 45 Later, at stage 2, the acceptability of an option to the opponent becomes one factor in the cost-benefit analysis by which options are compared with each other. The use of empathy as a screening criteria at stage 1 should be especially conducive to crisis resolution and escalation prevention. And, of course, the salutary effects of empathy should be more effective if both sides use empathy as a criteria at a given stage, rather than just one. The authors test their hypotheses on U.S.-Chinese relations in the aftermath of the Tiananmen Square massacre and on U.S.-Iraqi relations during the 1990–91 Persian Gulf crisis. Using contemporary memoirs, official documents, media reports, and historical accounts, they are able to trace the decision processes for both sides during the crises. At stage 1 in the Tiananmen crisis, President George H. W. Bush’s empathy for Deng Xiaoping’s domestic circumstances ruled out the use of full-scale sanctions, as it would hurt reformers like Deng and work to the advantage of his hardline internal opponents. But for domestic political reasons at home in the U.S., Bush could not do nothing. What was left was the imposition of moderate and measured sanctions of some sort. Thus Bush employed two noncompensatory dimensions to screen out his alternatives: concern for U.S. domestic political losses and the feasibility of options from the Chinese standpoint. For his part, Deng’s empathy with Bush’s domestic situation meant that his reaction to (mild) U.S. sanctions would not be overly harsh or provocative, though for his own domestic purposes, he could not turn off the repression overnight and give in to American demands. In the end, the two leaders were able to keep U.S.-Chinese relations on a positive footing despite the setback of the Tiananmen Square incident. The second set of dyadic interactions—between the Bush administration and Saddam Hussein in 1990–91—serves as a kind of reverse image. Both sides seemed well attuned to their own domestic political imperatives, but neither side showed any empathy for the other side’s domestic constraints—either at stage 1 or stage 2. To some degree, the two leaders made conscious efforts to prevent the other from saving face. For both, high political costs screened out moderate policies and pushed both toward more extreme options involving force.

128

Chapter 4

THE ROLE OF STRUCTURE: THE ORGANIZATIONAL PROCESS MODEL An important contribution to our understanding of government decision making was put forward by Graham Allison in his famous analysis of the Cuban missile crisis, Essence of Decision. 46 Allison offers three competing frameworks from which to view American decision making during the crisis. The first framework was our old friend the rational actor model; the second he called the organizational process model (OPM); and the third—and the one that Allison claimed fit best—was termed the governmental politics model. Shortly after their creation, the latter two models came to be seen as complementary and were combined to create the bureaucratic politics model (BPM). 47 Nevertheless, because some of the insights of the OPM are analytically distinct from those of the governmental politics model, let us take a brief look at the organizational process model before investigating its more famous sibling. 48 Allison’s organizational process model—which draws on organizational theories of large firms—sees governments as conglomerates of somewhat autonomous, “semi-feudal, loosely allied organizations.” The OPM focuses on the typical shortcuts used by large organizations in decision making. For instance, problems are typically “factored” or split into different segments and parceled out to subunits with specific roles and missions that deal with only a particular aspect of the problem. Coordination by top leaders is sporadic, and subunits attempt to deal with their problems in isolation from other subunits, devising solutions and then implementing them in a relatively independent manner. Governmental subunits function according to standard operating procedures (SOPs)— routines that are devised to deal with immediate problems. This repertoire of contingency plans provides the options available to an organization. If a crisis arises for which there is no contingency plan, decision makers make do with the contingency plans they do have available, even if they were not designed for the current crisis. For instance, in 2001 the Defense Department had no contingency plan for a military invasion of Afghanistan; therefore Operation Enduring Freedom was based on a paramilitary plan developed by the CIA that used special forces instead of placing a large number of “boots on the ground.” 49 When President Kennedy’s “Ex-Com” was summoned to devise tactics to remove Soviet missiles from Cuba, the Pentagon called up its “organizational memory” and responded with a plan to overthrow Castro left over from 1961. No contingency plan existed for destroying Cuban missiles because no such missiles had previously existed, but there was an invasion plan and its organizational sponsors were still in office. 50 As Allison notes, “the fact that the fixed programs . . . exhaust the range of buttons that leaders can push is not always perceived by these leaders.” 51 Thus, existing SOPs help determine which routines are actually identified as policy options, which may be actually chosen, and how the options—once chosen—are put into effect. The imperatives of organizational routines may therefore severely limit flexibility in decision making and decision implementation. In these cases, not surprisingly, decision making is characterized by sequential search, satisficing, and incrementalism, as well as by the lack of flexibility and imagination. Several general points should be made. First, the OPM essentially sees organizations attempting to cope with complexity in a world of bounded rationality. 52 Second, Allison sees organizational routines as rigidly maladaptive devices that may prevent decision makers from

The Substate Level of Analysis

129

effectively responding to their environments. 53 Third, the focus of the OPM is not on the moment of actual decision by government leaders but on how organizational SOPs constrain the range of options prior to the decision and how they affect the implementation of those decisions after the choices have been made. 54 The effects of these processes are relevant for our discussion of the causes of war. First, because the search for alternatives is limited to those options available in some organization’s repertoire of contingency plans, the choices are probably narrower than necessary. Options that might prevent conflict or slow down escalation toward war may be missed because they are not included in any organization’s package of routines. Second, organizations that operate according to the OPM are subject to bureaucratic inertia and are slow to respond to major changes in the environment. The OPM allows for incremental “patches” to be placed on failing policies. Rather than undertaking a wholesale reassessment of the policy and devising major changes, a cybernetic damage-limitation process is followed. As John Steinbruner explains, organizations attempt to reduce complexity. They do not monitor all aspects of the present policy; this would be impossible. Instead, a few simple, critical factors are monitored and there is an attempt to keep these variables within a tolerable range. If these variables for any reason exceed the acceptable limits, corrective procedures and adjustments are triggered. At this point decision makers proceed routinely through SOPs designed to restore the critical factor to within the range of tolerance. (Organizations are satisficers, not maximizers.) The policy then persists until the next instance of trouble. Such cybernetic procedures lead to an incremental tinkering with policy, but not to a total reassessment. 55 Third, SOPs determine how policies, once decided upon, are actually put into practice. Policy makers who are unaware of the procedures by which the policy will be implemented may be surprised to see that the actions being taken are significantly different from those they originally contemplated. Allison’s classic example concerns President Kennedy’s decision to blockade Cuba. The U.S. Navy implemented the blockade according to standard procedures, which placed the U.S. fleet as far as possible from Cuban jets, but far closer to Soviet ships approaching Cuba than was desirable if Kremlin leaders were to have a chance to deliberate without haste. Although the implementation of the blockade decision was monitored by political decision makers and the Navy was ordered to move the blockade closer to Cuban shores, Allison argues that in fact the Navy did not follow the president’s orders and intercepted the first Soviet ship along the original line. President Kennedy was forced to compensate for Navy intransigence by letting several Soviet ships through the blockade in order to increase the Kremlin’s decision time. 56 Quite a few analysts note that propositions derived from the OPM do not correspond very well to the events of the Cuban missile crisis itself. As Allison and Zelikow note in their revised edition, even the range of military options was fairly large, and contrary to expectations of the OPM, the options were continually refined and amended during the decisionmaking process based on new information and on urging by various actors. Even the air strike option had a degree of flexibility. Moreover, once the decision was made to impose a naval blockade, President Kennedy was able to change the plan on the fly. Decisions, once made, were even capable of being overridden (for instance, Kennedy’s decision not to retaliate

130

Chapter 4

militarily after the U-2 was shot down). Some of the most dangerous moments came, not from runaway routines, but from breakdowns in the normal routines. 57 For instance, Soviet forces on Cuba shot down the U-2 without asking directly for permission from Moscow—which was supposed to be the procedure. The Organizational Process Model and Implications for War Levy and Thompson suggest that the OPM can lead to war through several pathways: (1) Reliance on organizational routines (SOPs) may prevent leaders from finding better, more flexible solutions to complex problems that arise in crises. Moreover, SOPs may take on a life of their own that deny leaders the ability to make necessary corrections. The prime example here is Germany’s pre–World War I Schlieffen Plan, which we will discuss soon. (2) The factoring of problems may mean that military strategies are devised without consideration of political consequences and that political leaders may make decisions while lacking important information about military policy. For instance, most Japanese civilian leaders gave the goahead in late 1941 for an attack on multiple U.S., British, and Dutch targets in the Pacific without being informed by the military that the attack centered on Pearl Harbor. Many civilian leaders hoped that an attack would buy Japan time and force the U.S. to negotiate a modest mid-war settlement; however, since very few knew the exact target, there was no attempt to assess how the U.S. might respond to a surprise attack on Pearl Harbor. 58 (3) Government decisions may be implemented in the field in ways political leaders did not envision—as in the case of the U.S. naval blockade of Cuba. 59 Let’s look at how some of these problems played a role in World War I. The Organizational Process Model and World War I Jack Levy’s review of the pre–World War I crisis suggests that organizational routines— specifically, the war mobilization plans of the great powers—were a major cause (though not the only cause) of that war. 60 The rigidity of these routines made modifying mobilization plans extremely difficult. The German response to the crisis in the Balkans between Austria and Serbia (and Serbia’s protector, Russia) was to implement the Schlieffen Plan. The plan called for the invasion of France (Russia’s ally) through neutral Belgium and its central rail hub of Liege on the third day of mobilization. Germany could turn its attention to the Russian front in the east after France had been defeated in the west. While this made sense militarily, it was politically disastrous. It would greatly increase the probability that Britain would enter the war, even though German political leaders devoutly hoped to keep Britain neutral. The Germans were unable to change the plan to one that mobilized troops for an offensive in the east , thus ensuring both French and British participation against Germany and a two-front war. The Austrians were likewise unable to change their plans so that they could accept a political compromise and temporarily punish the Serbs by occupying Belgrade (the Serbian capital); Austrian troops were not positioned to march on Belgrade because their mobilization plan called for them to prepare for war against the Russians. Austria thus turned down the diplomatic compromise that might have kept the conflict localized. Similarly, Russia found itself unable to carry out a partial mobilization in the south against Austria alone; its only plan

The Substate Level of Analysis

131

called for a more provocative complete mobilization against both Austria and Germany, and modifications could not be made on the spur of the moment without disastrous military consequences. 61 In addition to the rigidity of SOPs, several other factors at the substate level contributed to the outbreak of World War I. First, some organizations and individuals had a vested interest in the standard mobilization policy because they devised it and were responsible for it. Because of their investment, they resisted any change in the plan, a tendency consistent with the bureaucratic politics model. For instance, Moltke’s resistance to the Kaiser’s last minute attempts to alter the Schlieffen Plan are well known. Second, the rigidity of mobilization plans in Germany was abetted by the fact that German war plans were constructed without consultation with political authorities and in disregard of political considerations—illustrating the tendency of large organizations to “factor” problems. Thus, political leaders were largely ignorant of the mobilization plans and their consequences. While mobilization in reality meant preparation for an immediate war, the political leaders were more likely to see it merely as a tool of coercive diplomacy that might have a salutary deterrent effect on opponents. They were unaware that once begun, military logic required that mobilizations be carried forward toward their intended conclusion—the initiation of war. In this respect it is interesting to reflect on the effect that the deployment of 425,000 U.S. active duty and reserve forces in the Persian Gulf had on the ability of the United States to choose a course other than war in January 1991. It was widely believed to be impossible—for political, economic, cultural, and logistical reasons—to keep such a large force at the front for an extended period of time while permitting economic sanctions to force the Iraqi military out of Kuwait. The mobilization created a “use them or retreat” situation, which was extremely difficult for civilian opponents of the use of force to challenge. This thorny problem repeated itself in the 2002–3 run-up to the Iraq War. Third, the fact that militaries possess mobilization plans places them in a good internal bargaining position vis-à-vis those who oppose mobilization but who have no alternative plans to offer. Militaries own several built-in bargaining advantages over other institutions on issues of national security: they are perceived as having a legitimate place in the debate; they control sizable resources germane to the problem; they are likely to have the ability to mobilize public support on behalf of their preferred solutions; and civilian officials are usually dependent on military sources of information. 62 Finally, organizational interests may have been responsible for the creation of the offensive military doctrines upon which the mobilization plans of Germany and other states were based. Levy argues that the offensive doctrines were probably not based on rational strategic calculations, since the technology of warfare at the time actually favored the defense over the offense. Instead, the organizational interests of the militaries dictated an offensive doctrine. Offensive strategies would increase the size and resources of the military (offensive forces require greater personnel and weaponry than defensive); offensive doctrines would enhance military morale and prestige (who wants to stand behind barricades when you could be in on the charge?); and offensive doctrines would require a large standing army rather than a system of reservists.

132

Chapter 4

THE ROLE OF STRUCTURE: THE BUREAUCRATIC POLITICS MODEL The model that Professor Allison believes best explains how governments really make decisions is his model III, the bureaucratic politics model (BPM). The BPM assumes that governments are not single, rationally calculating units. Instead, they are made up of organizations and individual actors who hold differing opinions about government policy options and who compete with each other to influence decisions. 63 Decision making in the governmental arena involves considerable conflict between disparate players who have different responsibilities, resources, and sources of information; who have different perceptions and operational codes; who see different sides of each issue; and who have different stakes in the outcome. Most important, governmental actors have different interests and different goals and therefore prefer different solutions to policy questions. It is assumed that no individual or organization has preponderant power. The president, if he is involved, is merely one participant among many. Although his influence may be strong, he is far from omnipotent. His preferences are not always the ones chosen, and even if he were to exercise his authority, his decisions are not always binding. They can be reversed, ignored, or emasculated through the mischief of those in charge of implementation. The BPM implies a kind of “bureaucratic captivity” of the president (or of the chief executive in other countries), who is seen as heavily reliant on the bureaucracy for information, for identification and definition of problems, for identification and analysis of alternatives, for the advocacy of solutions, and for policy implementation. How do we know if bureaucratic politics exists? Preston and ‘t Hart operationally define bureaucratic politics as consisting (in various degrees) of the following characteristics: (1) the presence of multiple bureaucratic actors; (2) actors who have conflicting interests and are involved in games of conflict and cooperation with each other; (3) power relationships that are diffuse and vary according to context; (4) interaction characterized by “pulling and hauling” or bargaining between clusters of actors; (5) decisions reached through bargaining, coalition formation, and compromise; and (6) decisions that are sensitive to “slippage” both in time and in content. 64 One of the most important propositions of the BPM is encapsulated in the famous phrase “where you stand depends on where you sit” (this is sometimes referred to as “Miles’ Law”). This means that the stand a particular participant takes on an issue depends on what organization he or she represents. Players see most issues through the parochial institutional perspective of their own organizations. This is not a new idea, but one that has been borrowed from role theory, a basic staple in sociology departments for years. It is assumed that all organizations—the foreign ministry, the defense ministry, the intelligence agencies, and so forth— have their own well-defined interests. For instance, all organizations have certain institutional goals that are important to them: more missions, greater autonomy from outside interference, greater influence within the government, greater capabilities, resources, and personnel, and, of course, a larger budget. In addition, each organization tends to equate the national interest with their own institutional interests. 65 It’s also possible that a desire to demonstrate loyalty to the group causes players to be more competitive in bargaining with representatives of other groups, thus increasing intergroup conflict. 66

The Substate Level of Analysis

133

The reasons for this parochialism are primarily sociological. Leaders within an organization share a similar set of values and assumptions because individuals select (and are selected by) organizations whose values are compatible with their own. Additionally, the process of institutional socialization, which takes place once an individual joins an organization, helps to inoculate the individual with the official perspective of his or her institution. A certain amount of attitudinal conformity comes with employment in large organizations. Role expectations connected to organizational membership encourage the individual to modify her attitudes and behaviors to accord with the perceived requirements of the position. Presumably, all those occupying a particular role will act similarly. Hence, we would expect government actions and policies to remain fairly stable, regardless of changes in personnel. It is assumed that a new role player’s position on basic issues will tend to be the same as his or her predecessor’s. 67 The most famous story concerning the effect of roles on behavior is that of Henry II of England and Thomas Beckett. As Chancellor, Beckett was Henry’s most loyal and aggressive supporter. Henry reasoned that if he made Beckett the archbishop of Canterbury, he would have an ally who would help him tame the power of the church—an ally who would assert the prerogatives of the king over the religious realm. But once Beckett became the archbishop, he became a defender of religious authority and Henry’s mortal enemy. The role had transformed the man. A similar phenomenon occurred in American politics. As George Shultz’s second-incommand at the Office of Management and Budget (OMB), Casper Weinberger had been called “Cap the Knife” for his willingness to slash federal programs in order to keep the United States within budgetary constraints. It was assumed that when he was named secretary of defense by President Reagan, the Pentagon would, at last, have a boss who would ride herd over its runaway budgets. Instead, Weinberger became an extremely effective supporter of military spending and played no small part in helping the United States pile up the largest budget deficits in history. Of course, individuals do have the capacity to transcend their roles. Some actors may actually remake the role, thus turning the proposition of role theory—that the role makes the individual—on its head. A strong player may redefine the interests of the organization he or she leads. Changes in policy may therefore be initiated by role transformations. Politics, Conflict, and the BPM Because individuals and organizations have different goals and interests and because they differ over the best methods of implementing them, policy making involves conflict and competition. Political elites struggle among themselves to exert power over the making of policy. The occasion for a decision mobilizes a wide variety of political forces (both inside and outside the government) to join in the political struggle. This is essentially a pluralist model of policy making: Many political actors and organizations are presumed to be involved in making the decision. 68 Policy options are linked to institutional interests. Options are not likely to be put forward nor supported unless they reflect some institutional preference. 69 But because foreign policies usually require the support (or at least the neutrality) of several different organizations, and because it is unlikely that a single, powerful individual can simply decide for everyone, players must negotiate with each other or risk deadlock. 70 The issue is not so much which

134

Chapter 4

position is best, but how to reconcile all the conflicting views over what constitutes the best policy. Decisions are the result of bargaining, compromise, logrolling, “pulling and hauling,” coalition building, competition, and conflict among individuals positioned hierarchically in roles (as not all actors are equal in rank). The actual decision itself will depend not just on the reasons or rationales used to support a course of action, but also on the relative power and skill of proponents and opponents. In this game, bargaining advantages are based on possession of formal authority, control over resources and information, personal persuasiveness, access to other powerful players, ability to mobilize external support, and control over policy implementation. (While the BPM assumes players are relative equals, it also posits that relative power, whether formal or informal, matters in determining the outcome of policy conflicts.) Presumably, bureaucratic politics results in compromise. This compromise could be based on an option constructed to contain some aspects preferred by each of the major contenders or on a brokered solution that splits the difference. Or it could represent the lowest common denominator—a solution that is the least intolerable for all sides. It is quite possible, in fact, that the solution chosen would actually have been no one’s first choice, and therefore the solution might please none of the players nor be very effective in achieving their preferred goals. This phenomenon is called “Arrow’s paradox” after its discoverer, Kenneth Arrow, another Nobel Prize–winning economist. 71 By their very nature, political compromises usually turn out to be suboptimal in achieving objectives. 72 Aside from compromise, there is another possibility that is not often identified by theorists: A majority coalition may be able to force its will on the minority. 73 Inducements (which may take the form of compromises) may be given to coalition partners to attain their support, but the result is victory for one side rather than compromise. Snyder and Diesing find that this outcome is more frequent than compromise and that decision making is decidedly less pluralistic than the BPM would predict. Domestic coalitions were able to exclude certain points of view and participants from the decision process, thus limiting the number of options placed under consideration. 74 A final proposition of the BPM is that there is considerable slippage between decisions and implementation. How the decision is carried out depends on standard operating procedures and on the political and organizational interests of those in charge of implementation. The latter may carry out the policy in ways totally unanticipated by the policy makers; they may consciously choose to make minor alterations in the policy or even totally subvert the policy with major changes; or they may simply refuse to carry out the orders at all. Players who are opposed to the policy imposed by the winning coalition may attempt to sabotage its implementation. Sometimes this has a direct effect on situations of war and peace. For instance, Maurice Paléologue, the French Ambassador to Russia, ignored orders from his government to pressure the Czarist government against taking risky action during the July 1914 crisis. Instead, he urged mobilization of Russian forces, pledging unqualified French backing! 75 As it turned out, Russian mobilization was a key factor leading to German mobilization and World War I. Such insubordination is carried out not only by major individual players but also by midlevel officials. President Roosevelt’s 1940 export orders permitted licensing of low-grade gasoline and crude oil exports to Japan, but hardliners in the State

The Substate Level of Analysis

135

Department who were in charge of implementation turned this into an all-out embargo— making the Japanese economic plight more desperate and increasing Japanese hostility toward the United States. 76 And at the very apex of the Cuban missile crisis in 1962, when tensions were at their highest levels, a local Soviet commander on the island gave orders that led to the shooting down of an American U-2 spy plane in violation of standing orders from Moscow not to fire on U.S. aircraft without permission from Moscow. 77 Fortunately, this last example of insubordination did not trigger hostilities that might have ignited World War III. Consequences of Using the BPM If the BPM is an accurate reflection of how governments make decisions, then there would seem to be several important consequences. 1. Domestic interests can be expected to predominate over national and international interests. Political leaders rise and fall depending primarily on whether they satisfy domestic needs, and certain powerful institutions (the White House staff in the United States or the Kremlin staff in Russia, for example) are very much attuned to the domestic political climate. They see their function as protecting the power of top office holders from domestic political repercussions of policy. 2. Policies will be fragmented. States seldom have a single, integrated foreign policy; they are more likely to have several competing “mini policies” made by different organizations or tailored together by different coalitions, which change over time. Policies may be uncoordinated and only loosely connected; they may even be incoherent and contradictory, with different agencies pursuing different goals. For instance, while President Reagan’s National Security Council staff and the White House were negotiating with moderate forces in Iran for the release of U.S. hostages in Lebanon, the State Department was pressuring U.S. allies to refrain from negotiating with hostage takers. And in 1960 the State Department and the CIA apparently even gave assistance to opposing armies in Laos! 78 The fact that policy tends to be fragmented and incoherent also means that mixed signals will be sent to other countries, creating the potential for misperceptions. 3. Middle-of-the-road, incrementalist policies will prevail. Political allies must be kept on board and temporary opponents appeased, and this requires moderation and compromise. Incremental choices avoid offending major political constituencies. Furthermore, if government policy is significantly influenced by standard operating procedures, as the BPM implies, then the government’s behavior at time t will probably be only incrementally different at t + 1. Bureaucratic elites are stifled by organizational inertia: Policies at rest tend to stay at rest. Decisive and innovative policies are discouraged and incremental ones are preferred because of the preference of career officials to “go along to get along.” The competition for political power and for influence over policy does not necessarily lead to risk taking. 79 As the saying about U.S. State Department personnel goes: “There are no old, bold foreign service officers.” 4. Few decisions will be either decisive or final. 80 Government decisions are likely to be the result of unstable, internally inconsistent compromises. Actors who have been de-

136

Chapter 4

feated in early rounds of policy struggles will try to win support from other players in an attempt to reverse the policy in succeeding rounds. What one political coalition decides is a reasonable solution to a conflict may be abandoned or overturned by the next, especially if the continuation of the conflict serves the vital interests of some group. This may help to explain the existence of protracted conflicts and enmities in international relations. 81 5. Decisions are likely to be nonrational and non-optimal. Policy makers do not seek the best solution to a problem, but the one that satisfies the most important and powerful political actors on this issue. Alternatively, decisions may represent the lowest common denominator. Before we proceed, several important questions need to be answered concerning the BPM. First, is it an accurate reflection of the way decisions are made in the U.S. government? Second, is it useful as a general model of decision-making? That is, can it be accurately applied to countries other than the United States? Third, does it help us to understand why governments make decisions for war? Fourth, is the BPM a good theory? Does the BPM Accurately Depict Decision Making in the U.S.? Let us look at two essential propositions of the BPM that have generated significant controversy: (1) Do organizational interests really determine the preferences of players? (2) Is the chief executive really an equal player? 82 1. Do organizational interests really determine the preferences of players? Although organizational loyalty is important, players’ personal experiences matter too. Their personalities matter; and they have their own personal and political interests that may be different from those of their organizations. Personal interests might involve winning the respect of colleagues, advancing an ideological agenda, personal advancement to the top of the political food chain, and “a place in history.” Allison and Zelikow acknowledge this in their revised edition of The Essence of Decision: Model III is not meant to be interpreted to mean that where you stand “is always determined by” where you sit, but rather where you stand “is substantially affected by” where you sit. 83 Likewise, it would be much too simplistic to argue that organizational roles translate into policy preferences in a deterministic way. There are several reasons for this. First, roles are not always very distinct, and they do not exist in detail specific enough to render preferences automatic; for instance, occupying the office of the national security adviser in the United States should probably offer no particular organizational clue as to whether the occupant would be a hawk or a dove. Second, players often occupy several roles simultaneously that may tug them in different directions. Top-level advisers must be advocates for the interests of their departments, but they also have a role to offer their unvarnished best advice to the chief executive concerning the national interest. Sometimes occupying a role can simply be reduced to nothing more than “a broad duty to act so as to be able to justify oneself afterwards.” 84 Third, personality and beliefs and ideology matter. Dick Cheney would have supported hawkish policies whether he was George W. Bush’s vice president, his secretary of state, or his UN ambassador. Roles don’t always trump individual idiosyncrasies.

The Substate Level of Analysis

137

Allison’s own analysis of the Cuban missile crisis indicates that institutional perspectives had little relationship to the policy positions taken by major players. For instance, the secretary of defense initially saw no major threat to U.S. security from the missiles in Cuba and opposed a full military option. Several other players had no bureaucratic role to defend; they were “wise men” trundled in from the cold—like former Secretary of State Dean Acheson. And some “ExCom” members, such as Robert Kennedy and Ted Sorenson, were more loyal to the president than to any bureaucratic fiefdom. As Stephen Krasner notes, it is hard to know where some players sit because “sometimes they are not sitting anywhere.” 85 As you might have suspected by now, research yields only mixed support for the proposition of “where you stand depends on where you sit.” Support comes from the often-quoted study by Robert Axelrod of interdepartmental groups in the American government. Axelrod found that the participants in these groups could accurately and reliably predict the stand of other group members on any issue on the basis of agency affiliation, independent of the personality of the particular individual involved. 86 Individuals didn’t matter; organizational affiliation did. The results of Andrew Semmel’s study of U.S. State Department officers’ attitudes toward multilateral diplomacy are also consistent with Allison’s proposition. However, Semmel’s study suggests that players’ stands depend less on their affiliation with the larger bureaucratic entity that employs them than on the immediate position within a departmental subunit. Large organizations like the State Department are best seen as several interacting subcultures rather than as a single monolithic culture. 87 Sociologists and social psychologists would not be surprised at Semmel’s findings. Competition and conflict occur within groups as well as between groups. We would hardly be stunned to find that large organizations contain fiercely individualistic subgroups and subcultures with their own interests. For instance, interservice military rivalry between the Army, Navy, and Air Force is well understood in the United States, and it has been an important political force in other countries as well. The complex and rather nasty interservice rivalries in Japan in the 1930s and 1940s played an important role in fostering an aggressive Japanese foreign policy. 88 Intragroup and intergroup conflict are interrelated. 89 Competition between subgroups within the same organization will generally be increased by competition between organizations. Edward Rhodes looked at the effects of subcultures within the military and the potential for rivalries between subunits within the U.S. Navy. He hypothesized that if the BPM operated as it should, then the chief of naval operations (CNO) should always show favoritism toward his home field—naval aviation, surface fleet, submarines—regarding decisions about budget matters and military force structures. But he found the CNO’s home affiliation to be a poor predictor of CNO policy decisions—a finding completely at odds with the BPM. 90 Anderson’s study of three foreign policy decisions from three U.S. administrations indicates that major foreign policy actors were almost as likely (43.8 percent of the time) to suggest alternatives outside of their institutional/organizational domains —such as a military staff member suggesting diplomatic action—as they were to suggest alternatives relevant to their own institutions (56.2 percent). 91 Similarly, Graham Shepard’s analysis of top players in the national security bureaucracy of the United States from 1969 to 1984 indicates that at least

138

Chapter 4

when it came to secretaries of state and defense, it was difficult to determine the effect of their roles on positions they took on issues concerning the use of force. 92 A popular assumption (and one tested by Shepard) is that players who represent military organizations are more likely to advocate the use of force than other actors. Another is that the military will have a unified position on the use of force. Neither of these propositions is consistently true. Anecdotal evidence abounds. General Ridgeway, the Army chief of staff, successfully led the political coalition that opposed direct U.S. military action to rescue the French war effort in Vietnam at the time of the Battle of Dien Bien Phu in 1954. Meanwhile, top civilian officials such as Vice President Nixon and Secretary of State John Foster Dulles joined Admiral Radford, the chairman of the Joint Chiefs of Staff, in supporting the proposal. 93 In the Vietnam discussions in 1966 and 1967, Secretary of Defense McNamara advocated diplomacy while Secretary of State Rusk supported the military’s policy of escalation. 94 Likewise, we know now that the Soviet military leadership (including Marshal Nikolai Ogarkov, the chief of staff, and his principal deputy, Marshal Sergei Akhromeyev) opposed intervention in Afghanistan, but was essentially overruled by a coalition of powerful civilian leaders led by General Secretary Brezhnev and Defense Minister Ustinov. 95 And it was Madeleine Albright, President Clinton’s UN ambassador and later secretary of state, who pushed for the administration to intervene militarily in Bosnia and Kosovo. She once famously challenged Joint Chiefs of Staff Chairman Colin Powell, “What’s the point of you saving this superb military, Colin, if we can’t use it?” 96 Ever since Samuel Huntington’s groundbreaking study, The Soldier and the State, political scientists have supported the contention that military officers are generally more timid or conservative than their civilian counterparts about recommending the use of military force. 97 Soldiers, Statesmen, and Cold War Crises, Richard Betts’ landmark analysis of the influence of American military advisers on decisions to use force in Cold War crises, concludes that military advisers were neither more nor less aggressive than their civilian counterparts on intervention decisions. In fact, the advice of the Joint Chiefs of Staff was similar to that of the civilian advisers more than half the time. In addition, there was frequent disagreement among military advisers; the armed forces were usually divided on recommendations to commit U.S. forces to action. (The Army chiefs were the most cautious, and Navy chiefs were the most aggressive.) Betts discovers that the Joint Chiefs were more influential when they opposed intervention than when they supported it. Presidents and civilian advisers were not persuaded to use force by the military, but they could be persuaded against it if the military thought the use of force unwise. 98 However, once civilian leaders have opted for war, military leaders generally become strong supporters of escalation of wars under way. Other analysts come to very similar conclusions. Snyder and Diesing, in their study of decision making in sixteen international crises in the nineteenth and twentieth centuries, found that military representatives supported accommodation as often as or more often than they supported a firm stand. Their preference, however, was usually based on estimates of military preparedness rather than personal bias. Role influences, although generally less important than personal values, were seen most often among military players. The military aside, however, they conclude that the stances taken by policy makers are based primarily on personal values and cognitive mindsets rather than on role or bureaucratic positions. 99

The Substate Level of Analysis

139

It should be noted that this view of the conservative military has been recently challenged. Todd Sechser contends that the argument is largely based on case studies of civil-military relations in countries (mostly democratic) in which the military is under tight civilian control. 100 He performs a statistical analysis of militarized international disputes that strongly suggests that countries characterized by weak civilian control over the military or by outright military government tend to initiate MIDs (and MIDs involving the use of force) at substantially higher rates than states with strong civilian control over the military. He suggests that whether militaries are cautious or not depends considerably on the degree of civilian control in the political system. The studies mentioned above provide considerable empirical evidence to undermine the BPM’s assertion that where one sits determines where one stands. Non-organizational personal baggage—culture, life experiences, values, personality, beliefs—do matter. Even BPM supporters such as Morton Halperin and Arnold Kanter recognize that some actors are less prone to represent narrow, parochial institutional perspectives than others. They make a distinction between “organizational participants,” whose stands can be predicted with high reliability from their organizational affiliation, and “players”—those for whom institutional membership is not a good predictor of policy stands. They hypothesize that the higher one’s formal position, the less likely one is to be affected by parochial institutional interests. 101 2. Is the chief executive (president or prime minister) really an equal player who bargains with other players to determine government actions? Many of the critiques of the BPM are centered on the role of the president. 102 Remember that Allison sees the president as only one of many central decision makers. However, if the political system has a chief decision maker—such as the American or Russian president, the British prime minister, or the German chancellor—then it is possible that the real locus of decision is the individual, not the group. Even though each of these top leaders has a council of advisers of some sort, the decision process may in fact revolve around the individual dispositions of the top leader. In that case, the bargaining, coalition building, and logrolling among the members of the leadership group may have a relatively insignificant effect on policy selection. Critics of the BPM argue that the fact that cabinet members are hired (and may be fired) by the president means that they cannot be entirely independent of him. As Amos Perlmutter puts it: “How powerful is a group which can be dismissed at the President’s whim?” 103 If a secretary of defense as powerful as Robert McNamara can be forced out once he begins to oppose the president’s policy on Vietnam, then it is fairly clear where the balance of power lies. Cabinet heads are tied to the president every bit as much as they are tied to their departments. In fact, one critic suggests that Allison’s famous dictum ought to be rephrased as “Where you stand depends upon where the President stands.” 104 The president not only appoints the heads of bureaucracies, he also sets the rules of the game, determining which actors will participate in which policy decisions and who will have access to him. The search for, and evaluation of, options may be greatly affected by presidential preference. 105 Allison’s own analysis of the Cuban missile crisis decision determined that many of the options (like “do nothing” and “make a diplomatic protest”) were immediately foreclosed because the president (for domestic political reasons) was not interested in them. 106 Even in the central debate between proponents of an air strike and a naval quarantine, some supporters

Chapter 4

140

of the former understood that making arguments for the air attack was probably an exercise in futility, given the well-known support of the Kennedy brothers and Secretary McNamara for the quarantine. 107 Bendor and Hammond argue that during the Cuban missile crisis, “the most obvious manifestations of politics—coalition formation, bargaining, logrolling, splitting the difference, leaks to the press—were largely absent.” 108 All of the players knew that the decision was the president’s to make; there would not be a committee decision. Welch agrees: Never did one player assent to X only on condition of receiving Y as a quid pro quo. This was because authority was not evenly distributed among the members of the group; the president held all of it. The “pulling and hauling” that went on took the form of normal debate in which players argued for and against various options; the name of the game was persuasion, and the only player who had to be persuaded was the president. 109

Finally, critics argue that it is inaccurate to depict the president as the captive of a bureaucratic consensus. Even a total negative consensus against the president may not dissuade him from pursuing the option he prefers. In the missile crisis when the ExCom did agree on a decision (to bomb a SAM base after a U-2 spy plane had been shot down), the president reversed the decision on his own when a U-2 was shot down! The old story about President Lincoln and his cabinet is worth noting here. Lincoln once put a policy proposal up for vote in his cabinet and they all voted against the president’s proposal; the president’s vote was the sole vote in favor. Lincoln famously announced, “The ayes have it!” Critics admit that on certain occasions bureaucratic interests are decisive in the formulation of policy; some policy options never get presented owing to bureaucratic imperatives, and frequently the president fails to seek out options other than those presented to him by the bureaucracies. But this all depends on presidential interest. When presidents are uninterested, when they fail to assert control, or when they delegate authority, they take themselves out of the equation and reduce their role to that of merely being first among equals; but they have the power to rise above this if they wish. The ability of bureaucracies to independently establish policies on their own is a function of presidential inattention, a fact that is acknowledged even by advocates of BPM. 110 When Does the BPM Apply? Most political scientists acknowledge that decision making follows different models at different times. Other models of decision making may more accurately describe the policy process, depending on the issues. Preston and ‘t Hart argue that perhaps the best way to look at this is to accept that the phenomena associated with the BPM may be present to a greater or lesser degree. That is, bureaucratic politics exists in government settings as a continuum—from low intensity to high intensity. This means that bureaucratic politics is itself a variable, and there are different types of bureaucratic politics, each with different potential results. Thus, while some BPM arrangements can have positive, functional results, some may produce dysfunctional decision processes. 111 Bureaucratic politics should be seen as a contingent phenomenon that varies across situations, policy domains, political structures, and social interaction patterns. 112 The decision-

The Substate Level of Analysis

141

making literature suggests that the BPM is most likely to be in operation when the following conditions are met: 1. When the number of actors and organizations involved is fairly large. In order for a decision to be made by the bureaucratic politics process, a minimum number of three individuals or organizations probably need to be involved. 113 Essentially, BPM processes require committee decision making of some sort or at least a central decision maker who must consult with a circle of advisers. The more individuals or groups involved in the decision, the more likely the BPM will be used. “Intermestic issues,” which cross boundaries of domestic and foreign policy issues, are particularly susceptible to bureaucratic politics—international economic issues being a prime example. 114 2. When the group that decides the issue is fairly heterogeneous and is socially, culturally, and institutionally lacking in cohesion. For reasons we will examine later, the less intragroup cohesion there is, the more conflict one can expect. 3. When members of the group are relatively equal in power. Bureaucratic politics doesn’t require complete equality of participants, but it doesn’t come into play when one member exercises decisively more power than others. Bureaucratic politics is most likely if the top leader is either absent from the process or is part of a truly collegial leadership group. If there is a strong leader who is particularly invested in the issue, then she is in a position to push the decision toward her own preferred solution. 115 4. When the primary loyalty of the decision makers is to their own institution rather than to the decision unit. 116 This ensures institutional competition instead of consensus. 5. When there is sufficient time for members of various institutions to organize their attempts to influence the decision process. The longer the time available for decisions, the greater the possibility that interested actors and organizations will be able to insinuate themselves into the policy-making process and mobilize support for their positions. Decisions with short time frames are not conducive to bureaucratic politics. Therefore, true crises probably are not handled through the BPM. 6. When the decision-making process is open. Bureaucratic politics, though it does not require popular participation and democratic institutions, does require a degree of openness to input from a variety of institutional actors. Do these conditions eliminate issues of peace and war from consideration through the BPM? Probably not. Not all decisions on war and peace are made in the midst of a crisis that limits the amount of time available. The Soviet decisions on Czechoslovakia and Afghanistan and the American decision to use force against Iraq in 1991 and 2003, though they certainly operated under some time constraints, could in no way be called quick decisions. Neither could U.S. decisions on Vietnam. Likewise, issues of war and peace are so important that actors and institutions that might normally play only limited roles in foreign policy issues could legitimately participate. Also, while the chief decision maker would certainly be keenly involved in such an issue, he or she would not be likely to want to carry the ball alone on this one. On issues of such momentous import, top decision makers want to have a lot of political padding if they stumble; this normally requires the support of a wide spectrum of political elites.

142

Chapter 4

Is the BPM Applicable to Governments Other Than the United States? Since the BPM has been developed largely by studying U.S. foreign policy and its use has been confined almost exclusively to case studies of U.S. government decisions, one has to question its generalizability to other states. On this score, Snyder and Diesing see the BPM as actually less relevant to the United States, where the president holds the ultimate power of decision in foreign policy, than to other regimes where responsibility for foreign policy decision making is shared. 117 For example, it has been argued that the BPM is applicable to British-style parliamentary systems. The same kind of parochial concerns motivate policy makers in these systems as in presidential systems—the major difference in Westminster systems being the concentration of authority in the cabinet, where “ministry is set against ministry.” 118 It would seem that the BPM would be applicable in many regimes where there was a collective leadership body of elites, including even the former Soviet Union with its ruling Politburo. In fact, the BPM does have striking similarities to Kremlinological studies of Soviet decision making. Several models used by specialists to describe the decision-making system in the former Soviet Union focus on the conflict among elites representing institutional interest groups. 119 One Kremlinologist, Jiri Valenta, argues that while the BPM must be modified somewhat to take into consideration the distinctive features of the Soviet system, the approach is nevertheless useful in explaining Soviet foreign policy decisions. Soviet foreign policy actions . . . result from a process of political interaction (“pulling and hauling”) among several actors—in this case, the senior decision makers and the heads of several bureaucratic organizations, the members of the Politburo, and the bureaucratic elites at the Central Committee level. Bureaucratic politics is seen as based upon and reflecting the division of labor and responsibility for various areas of policy among the Politburo members. 120

What made the Soviet system essentially a bureaucratic politics system is that decisions in the post-Stalinist era were made by a collective body—the Politburo. From Stalin’s death until the Gorbachev era, Soviet leaders rarely possessed enough power to make decisions individually on foreign policy matters. Compared with the American president, the general secretary of the Communist Party of the Soviet Union (CPSU) probably enjoyed more limited power in making decisions. His power within his own “cabinet”—if we wish to see the Politburo in these terms—was certainly less dominant. He was typically primus inter pares (first among equals), while the president dominates his cabinet. According to Dennis Ross, the ending of the use of terror against the party elite after Stalin’s death and the “palace coup” in which the Politburo ousted Khrushchev in 1964 institutionalized a collective leadership based on multiple power centers. 121 Khrushchev’s demise symbolized the fact that the general secretary could not run roughshod over the interests of major political players and over the institutional prerogatives of essential organizations. If the general secretary could be sacked by a majority faction within the Politburo, he could not be considered an ultimate authority; he must rule through his ability to create a consensus or, at least, a majority coalition. He was forced to become a broker of the interests of various factions within the Politburo rather than an initiator. This was especially true of the Brezhnev period.

The Substate Level of Analysis

143

Creating a consensus for policy required at least the acquiescence, if not the approval, of a majority of senior leaders within the Politburo, and perhaps a wide body of supporters within the larger Central Committee as well. This frequently compelled such tactics as internal bargaining, compromise, and logrolling. It also involved the mobilization of pressure groups, the changing and rearranging of personnel who might make up the decision team, attempts to manage information, the use of the press to try to influence the debate, and the persuading of uncommitted leaders—all techniques familiar to those who use the BPM to study the American political system. The most important factor in this process was coalition maintenance. This was in large part compelled by the key ideological myth of party unity, which denied the possibility of conflict among the party elite. The system was characterized therefore by a willingness of political actors to compromise, by a lowest common denominator approach to policy making, and by incrementalism—all the marks of a bureaucratic politics system. This “muddling through” style also meant that Soviet decision makers were likely to avoid rash decisions. Soviet elites were extremely cautious anyway, but they were especially sensitive to the high political costs of failure. Policy failure constituted the paramount basis on which individual members of the elite could be successfully challenged and relieved of their positions by their political enemies. However, Ross concludes that even though the system was structured to avoid risks internally, the Soviets might still run risks externally, especially if their internal positions depended on it. Is the BPM Useful in Explaining the Cause of War? It would seem that a BPM explanation of the cause of war would require several things. (1) The decision for war is made in a situation where numerous individuals, organizations, and governmental institutions that have differing interests are competing to have their versions of governmental policy adopted. (2) The decision for war is the result of either bargaining or compromise or power struggle between these various political factions in which a balance of power within the government favors the initiation of war. (3) The decision for war is seen by one or more of the groups as promoting its organizational or political interests, or the decision to forgo war is seen as detrimental to its interests. Levy and Thompson suggest three causal paths by which bureaucratic politics may lead to war. 122 (1) The first path is fairly direct: The organizational interests of the military may tip the internal political balance in favor of war. (2) A second path is more indirect: Militarysecurity organizations (and others) push for policies that lead to larger military budgets, larger armies, navies, and air forces, more advanced weapons, offensive-oriented military doctrines, and more hawkish and confrontational foreign policies. These policies may indirectly lead to war through the reactions they bring about in other countries, creating conflict spirals that end in war. (3) The third path is the politicization of intelligence, which may skew the internal debate in favor of military force or war. For instance, in the discussions leading to the Bay of Pigs Crisis, the CIA deliberately overestimated the chances that the invasion would trigger a domestic anti-Castro uprising in Cuba. 123 There is also widespread speculation that various intelligence factions skewed intelligence about the presence of WMD in Iraq so as to increase estimates of the threat. 124

144

Chapter 4

Several other paths might be postulated. It is possible that a coalition led by civilian leaders effectively neutralizes military opposition and pushes for war for domestic and/or international political reasons. This is very similar to the situation under which Soviet leaders made their fateful decision to invade Afghanistan in December 1979. A very small circle managed to keep other officials at arms’ length and control Afghan policy by themselves. This troika consisted of KGB head Yuri Andropov, Foreign Minister Andrei Gromyko, and Defense Minister Dmitry Ustinov, and also included Boris Ponomarev, head of the CPSU Central Committee’s International Department. The troika secured Brezhnev’s support for the intervention and forced the Politburo to sign off on the decision—in spite of serious misgivings by senior military officials led by Chief of General Staff Nikolai Ogarkov, who was essentially told to butt out. 125 This path dovetails with an argument made by John Vasquez in his “steps to war” theory. Vasquez emphasizes that an important step in the path to war is the coming to power of hardline factions who dominate the political coalition. Hardliners may be in power as a result of the last war, which has been seen as both victorious and necessary. Or they may come to power because the previous “accommodationist” coalition has been ineffective in dealing with the hardline policies of rival states. Once in power, this political coalition of like-minded leaders are predisposed by their beliefs, worldviews, and operational codes to be nationalistic, distrustful of others, uncompromising toward rival states, militaristic and supportive of the use of threats and military force, and prone to risk taking in foreign affairs. A government in which such a coalition is dominant is more likely to choose military force and war as a policy option or to stumble into war because of the reactions of other states to its aggressive policies. War is unlikely to happen unless hardliners are in power in at least one of the states in a dispute; the probability of war is highest when both states are led by hardliners. 126 Let us concentrate briefly on the first path, which is the one most often referred to by international relations scholars. A primary assumption is that the military is the group most likely to see its organizational interests advanced by going to war. They’re also likely to have beliefs that Vasquez would characterize as hardline. One of the clearest examples of this is probably the Japanese military’s support for wars in Manchuria, China, and Southeast Asia. In this case the military had extreme influence in the government: not only did it participate forcefully in the making of policy, but it was also capable of autonomous acts of its own to embroil the state in war in spite of contrary civilian policy. 127 While examples like the one just mentioned can be cited, we have seen that often the military advocates against war. First, it is certainly not in the military’s interest to go to war for the sake of going to war. Losing—especially losing big—is not in the military’s interest and it can be counted on to argue against war if it is not adequately prepared. For instance, the German military generally opposed Hitler’s push for war in Europe during the 1930s. Second, the military does not always speak with one voice. Even in Japan, the army and navy were almost completely at odds about the identity of Japan’s chief threat, the goals of military action, and the preferred targets and strategies to follow. Another classic example of divided military interests is the situation in pre–World War I Germany. While Chief of Staff Moltke argued for a preventive war against Russia in 1912, Admiral Tirpitz opposed it because the German fleet was not yet ready. 128

The Substate Level of Analysis

145

Cartelized Systems with Logrolling Coalitions Finally, we should mention a distant cousin of the BPM—Jack Snyder’s logrolling coalition model. In Myths of Empire, Snyder poses the question of why some great powers engage in reckless and aggressive policies of overexpansion. One similarity among Wilhelmine Germany, modern Japan, and Russia/USSR is that they were all “late industrializers”—that is, they began their crash courses in the industrial revolution much later than other great powers. This was a social-economic-political phenomenon that worked to create a particular kind of polity—a cartelized political system. In a cartelized system, the government is not very strongly centralized and it is dominated by powerful interest groups or cartels, each of which has narrow, concentrated interests. The elite governing coalition, dominated by these narrow interest groups, can only function through logrolling tactics. Instead of thwarting each other’s interests, each group must agree to support the preferred policies of the other groups in return for reciprocal support for its own interests. The resulting collection of policies might be entirely contrary to the state’s national interests, but they generate substantial benefits for the participating elites. Ultimately, this is the only kind of policy that would allow the coalition to function, since no group could implement policies on its own. The prime historical example of this was pre–World War I Germany, whose government was based on the famous coalition of “Iron and Rye.” The primary partners in the coalition were the Junkers, the landed Prussian aristocracy whose interests were in agriculture (rye), and the military and heavy industry (iron). Each group had an interest in some form of national expansion. Logrolling created a set of foreign policies in which the Junkers obtained a high grain tariff that raised the price of grain but also antagonized Russia (a major exporter of grain); the Navy, supported by the Navy League and the Pan German League, won a large modern fleet that frightened the British; and the Army got a high degree of autonomy and an offensive war plan (the Schlieffen Plan) that centered on initiating an offensive attack on France regardless of the location of the military threat. Germany’s aggressive foreign policies were justified by the propagation of “strategic myths” that induced the public to support them, even though they were extremely risky and ultimately their costs far exceeded the gains. Moreover, the myths created to support these nationalistic, imperialistic, and militaristic policies created a kind of ideological blowback that trapped officials in their own rhetoric. Snyder’s theory is probably most consistent with the second of the three causal paths to war identified previously for the BPM: the aggressive and militaristic policies of the coalition lead to war not directly, but indirectly because other states see these policies as threatening and react in a way that escalates the overall degree of threat until something sparks a war. As others note, Snyder’s theory is not a theory of war, per se. Whether war occurs or not depends on the reaction of other states to the coalition’s aggressive policies. Other states decided that Germany’s policies were over-expansionistic. 129 Bureaucratic Politics and Soviet Involvement in War Perhaps the most interesting study that uses the BPM to explain war is Jiri Valenta’s Soviet Intervention in Czechoslovakia, 1968: Anatomy of a Decision. 130 When the Czech government under Alexander Dubcek began to implement a variety of liberal economic, political, and

146

Chapter 4

social reforms in 1968, the Soviet Union was faced with a serious foreign policy crisis. Although virtually all Soviet leaders saw the Czech reforms as a threat, there was division about how to respond to it. Two rather informal coalitions began to form early in the crisis— pro-interventionists and anti-interventionists—with some members uncommitted. The major institutions of the Soviet system—the Communist Party, the Soviet military, and Soviet state bureaucracy—were internally split. The case clearly illustrates the fact the sub-units of large organizations have their own perception of the national interest and their own interests that serve to motivate them. The pro-interventionists were motivated by several interests: 1. Party bureaucrats in the non-Russian Republics such as Pyotr Shelest, a Politburo member and first secretary for the Ukrainian Republic, were concerned about the spillover of reformist ideas from Eastern Europe into the nearby Soviet Republics. Shelest’s concerns were echoed by P. M. Masherov, the first secretary of the Belorussian Republic who was a candidate Politburo member. 2. Central Committee bureaucrats charged with ideological supervision and indoctrination were concerned about the spread of revisionist ideas from Czechoslovakia and their ability to contain dissidents within the Soviet Union. These men included A. Pel’she, a Politburo member and head of the Party Control Commission; P. N. Demichev, a candidate Politburo member and Central Committee Secretary responsible for the Ideological Committee; and S. P. Trapeznikov, the head of the Central Committee Department of Science and Education. Party officials in large cities, such as Moscow Party Secretary Grishin, who had to deal with large numbers of vocal dissidents in the urban intellectual and literary communities, also perceived the need to stifle Czech reformist experiments. 3. The KGB, led by Yuri Andropov, and the Department of Main Political Administration (in charge of ideological and political supervision of the army and led by General Yepishev) perceived that the new political winds in Czechoslovakia were a threat to the morale and discipline of Eastern Europe forces in the Warsaw Pact. Soviet generals responsible for Warsaw Pact forces, including the Pact’s commander-in-chief, General I. Jakubovsky, certainly perceived the Czech reforms as a threat to the organizational mission of their forces. The KGB had additional reasons for wanting to reverse the tide of reform in Czechoslovakia; Czech officials had dismissed many of their most trusted men from positions of responsibility in the Czech Ministry of the Interior. The KGB’s organizational mission inside Czechoslovakia was therefore at stake. A similar problem probably confronted the military’s Chief Intelligence Directorate (GRU), as GRU collaborators within the Czech Army were also being dismissed. 4. East German and Polish party leaders Walter Ulbricht and Wladyslaw Gomulka, respectively, feared the spread of liberal reforms to their countries and attempted to influence the Soviet decision in favor of intervention. The anti-intervention coalition was guided by the following interests:

The Substate Level of Analysis

147

1. Mikhail Suslov, the Politburo’s leading ideologist and the man responsible for the coordination of Soviet policies in the international communist movement, emerged as the spokesman for the anti-interventionists. Suslov and Boris Ponomarev, the Central Committee Secretary for the International Department, were concerned that a Soviet military intervention would undermine their organizational mission—maintenance of good ties with Communist parties in the West and progressive forces in the Third World. Intervention would also jeopardize the World Communist Conference scheduled for November 1968 and organized by Suslov. Finally, the use of force would jeopardize the strategy of a rapprochement with West Germany and a united front with West European social democratic parties. 2. Bureaucrats in the Foreign Ministry and the Central Committee International Department who had responsibility for Soviet relations with the West clearly felt intervention would adversely affect Soviet interests. 3. Premier Anatoly Kosygin, presumably the number two man in the Politburo and responsible for government diplomacy at the time, feared that an intervention would jeopardize the desired goals of concluding a nonproliferation treaty and an early beginning of SALT negotiations. Kosygin’s personal power and prestige were also at stake because of the similarity of some of the criticized Czech economic reforms to those advocated by him for the Soviet Union. 4. Dubcek supporters in Eastern Europe, such as Hungarian leader Janos Kadar, Yugoslavia’s Tito, and Rumania’s Ceausescu, feared that Soviet intervention against Czech reform would threaten their own reforms. A third group was the lone fence-sitter: 1. General Secretary Brezhnev is the only major player identified by Valenta as an uncommitted thinker, one who sees many sides of the issue, though there were assuredly others as well. Brezhnev wavered between the two coalitions until the very end, acting as a broker between the two factions but also trying to identify himself with the winning coalition. The position of Brezhnev was crucial in this case, not just because he was the general secretary but because the Soviet leadership were divided on the proper policy to follow. Neither of the coalitions had a strong enough majority to force its will on the other. Thus, a change in the position of any major uncommitted player might tip the balance. Negotiations between the Soviets and the Czechs took place at Cierna and Bratislava in late July and early August. The immediate result was a policy of compromise—both a compromise among members of the Soviet leadership and a compromise between Soviet and Czech leaders. The Czechs provided assurances of their loyalty to the Warsaw Pact and COMECON, agreed to control their news media more effectively, promised to prevent the creation of political parties, and agreed to purge certain leaders from high office. For their part the Soviets agreed to withdraw all troops from Czech territory (they had been there on Warsaw Pact maneuvers) and to approve the September Congress of the Czech party. Neverthe-

148

Chapter 4

less, only seventeen days after the final meeting, Czechoslovakia was subjected to a military intervention. Why did this reversal occur? The action revolved around trying to create enough political converts that a winning coalition could be created in favor of overturning the Cierna-Bratislava compromise and launching a military intervention. The normal techniques of persuasion were attempted, including the use of the press by various factions to try to mobilize support. The KGB and the Soviet ambassador to Czechoslovakia, S. V. Chervonenko, also attempted to distort information and analysis to lend credence to their preferred course of action. The KGB’s primary motivation was to reverse the continuing purge of Soviet agents by the Czech government. Chervonenko’s motives were more personal. Having been the Soviet ambassador to China when the Sino-Soviet split erupted, he did not desire to be tagged with the loss of another socialist state on his watch. He advised his colleagues in Moscow that although the situation in Prague was getting worse and that a “second Hungary” was possible, the Dubcek faction was a minority within the Czech Politburo, lacked support among the masses, and could easily be replaced by “healthy elements” if the Soviets intervened. According to Valenta, the modus vivendi with the Czechs began to come apart when segments of the Soviet military elite who were dissatisfied with the compromise began to press the political leadership for a reversal. We have mentioned that Warsaw Pact commander Jakubovsky viewed the Czech reforms as undermining discipline in East European forces. Other Soviet military leaders, especially General Pavlovsky of the newly restored Ground Forces Command, were dissatisfied with the absence of Soviet armed forces from Czech soil. Given the growing loss of confidence in Czech forces, forward deployment of Soviet ground forces on Czech soil was no doubt seen as essential to carry out Soviet military doctrine in Europe. A military operation against Czechoslovakia would, of course, coincidentally improve the role and prestige of the Ground Forces Command. This is not to say that the Warsaw Pact generals were all in favor of intervention. General Kazakov, chief of staff of the Warsaw Pact forces and a participant in the Soviet intervention in Hungary in 1956, was apparently very skeptical of Soviet intervention in Czechoslovakia. He was unexpectedly replaced after the Cierna-Bratislava conferences, however, by General Shtemenko, described by Valenta as a “Ground Forces lobbyist.” The military and their allies had another argument that was becoming increasingly more relevant—logistics. Not only were regular units involved in the Soviet troop buildup and maneuvers in Czechoslovakia, but thousands of reservists had been called up and thousands of motor vehicles had been requisitioned from the civilian sector in eastern Russia. A shortage of civilian manpower and trucks was beginning to have a detrimental effect on the 1968 harvest, and this effect could only worsen. The Soviets would have to either move quickly or dismantle the whole military effort. The implementation of organizational routines for military maneuvers had a substantial effect on the options available to Soviet decision makers. At this point Ukrainian Party Secretary Shelest mounted a pro-intervention campaign. We have already mentioned Shelest’s concern over Czech reforms infecting Ukraine. Shelest was probably also motivated by considerations of his own political position—after all, he had been on the losing side at Cierna and Bratislava; his prestige and position within the Politburo were at stake. Simultaneously, renewed pressure was applied by party bureaucrats responsible for

The Substate Level of Analysis

149

ideological issues and concerned about the failure of the Czechs to reinstate censorship of the media. Polish and East German leaders also opened up a new political offensive aimed at overturning the accords. Gomulka (and to some extent Ulbricht) felt threatened internally by the compromise with Czech reformers; their domestic opponents would draw strength from Soviet willingness to permit reform in Prague. Thus, renewed pressure was placed on wavering Politburo and Central Committee members to overturn the Czech compromise. Finally, anti-Dubcek forces in Czechoslovakia mounted a desperate last-ditch attempt to save their political skins. Their communications to Moscow signaled to the Soviets that the political situation was deteriorating in Prague. Intelligence reports, skewed by the personal and institutional interest of the senders, permitted the Soviets to believe that a military operation would involve low risks and would have a high probability of success. The combined pressure of the military, the KGB, the regional secretaries in the USSR’s western republics, party bureaucrats dealing with ideological affairs, and their allies in East Germany, Poland, and Czechoslovakia turned the tide in the internal debate. Those who had sat on the fence in the past, and even some who had previously supported diplomatic compromise, now gave their support to military intervention. Perhaps there was a fear of being in the minority on an issue of such momentous importance. Forced resignations could be the penalty for being on the wrong side on major policy questions. Non-interventionists like Suslov, Kosygin, and Ponomarev were already being criticized in reports to the party organization and in the press for their failure to understand the dangers of the Czech reforms. Savvy political observers could tell that the wind was now blowing in a different direction. Brezhnev himself now supported intervention. Domestic political considerations were probably paramount in Brezhnev’s analysis of the situation according to Valenta. The fact that the Cierna-Bratislava accords had not been well received by certain political and military elites certainly did not enhance Brezhnev’s political position. The general secretary came to see Soviet intervention as required both by Soviet national interests and by his own political interests. 131 The Czechoslovakian case shows some similarities with the process by which the Carter administration finally decided in favor of military intervention in Iran in 1980 to rescue American hostages. During the hostage crisis, the pro-intervention faction (National Security Adviser Zbigniew Brzezinski, Secretary of Defense Harold Brown, CIA Director Stansfield Turner, Chairman of the JCS Gen. David Jones) and the anti-intervention faction (Secretary of State Vance and his deputy Warren Christopher) were essentially stalemated. A faction of “presidential supporters” (President Jimmy Carter, Vice President Walter Mondale, Presidential aide Hamilton Jordan, and Press Secretary Jody Powell) were at least initially fence-sitters. Carter, like Brezhnev, was a prototypical “uncommitted thinker.” The decision for intervention was made possible by the growing concern of the fence-sitters that the president had to take action due to the changing mood of public opinion and the growing domestic pressures connected to the upcoming presidential election. 132 Bureaucratic Politics and the 2003 Iraq War Did bureaucratic politics characterize the Bush administration’s deliberations over the need for war against Iraq? The picture is mixed. Coalitions certainly formed on both sides of the

150

Chapter 4

issue. One coalition, based in the State Department and led by Secretary of State Colin Powell and his deputy Richard Armitage, supported containment of Saddam Hussein through coercive diplomacy pursued primarily through UN sanctions. The pro-war coalition included Vice President Cheney and his chief of staff Scooter Libby, Secretary of Defense Donald Rumsfeld and his deputy Paul Wolfowitz, and leaders of military services. National Security Adviser Condoleezza Rice attempted to manage the process of deliberations but ultimately sided with the pro-war group. To the extent that “pulling and hauling” among the players created compromise, it was only with regard to the path by which the U.S. would proceed to war. Powell was able to force the hawks to agree that “the road to Baghdad” had to run through the UN Security Council. Pulling and hauling resulted on agreement on the series of steps the U.S. must take to arrive at war. The issue of whether the war would take place was not an issue for compromise: the hawks won, and Powell and the State Department had to stand down. The anti-war faction was essentially marginalized, but in large part this was because the president himself was clearly in the war party. 133 Martin Smith argues that where the Bush administration officials stood on the war “was not, in the main, a function of where they sat in the bureaucracy.” 134 (However, the generally hawkish Armitage, once his office switched from the Pentagon to the State Department, heartily supported the anti-war positions of his boss and friend, General Powell.) Moreover, he finds little evidence of serious advocacy and counter-advocacy; in short, little pulling and hauling took place. 135 Indeed, one of the curious things about the process is that at no time did Rice orchestrate a grand debate on all the potential options for dealing with Iraq. Very early in the process it became clear the president preferred war and the real issue became “how do we get to there from here.” BPM as a Theory At this point we ought to state that it is not entirely clear whether the BPM is a theory or something less. In their reformulation of the BPM, Allison and Halperin label their combination of the organizational process model and the governmental process model a “paradigm” rather than a theory, making more modest claims for it. 136 One critic argues that in fact, as paradigms, Allison’s models are “pretheoretical”; they have no explanatory or predictive power. And none of Allison’s models is a “fully specified causal model relating dependent and independent variables.” 137 Keeping this in mind, let us examine the theoretical assets and liabilities of the BPM. What does the BPM attempt to explain and how is this explanation achieved? What is the dependent variable, and which independent variables are supposed to “cause” it? The dependent variable that the theory purports to explain is government “action” or policies. It is not a theory of war per se. The independent variables are things such as the players, their positions and roles, their motivations, their organizational and political interests, the bargaining process, and the regularized procedures through which the policy is made. The relationship between these variables and their relative importance is unclear, and probably varies from case to case. Making these variables operational, not to mention measuring them, presents the researcher with considerable difficulties. As we have seen in Valenta’s explanation of the invasion of

The Substate Level of Analysis

151

Czechoslovakia, the BPM lends itself to lengthy qualitative assessment rather than quantitative analysis. This is all pretty messy. The BPM is an extremely complex and nonparsimonious theory— like almost all decision-making theories. The result is that BPM explanations tend to be rather complex as well. Such explanations require a substantial narrative, much like a historical explanation of the decision process, which focuses on those concepts or variables identified in the model. In its extreme form, “It is possible to include so many variables that the theory does not explain even one case very well . . . In general, a model that includes everything explains nothing. If it does not simplify, it cannot explain.” 138 This presents several problems. A BPM explanation requires access to a quantity and quality of data that most researchers are unlikely to find readily available. After all, the most reliable information would be notes of cabinet meetings, politburo meetings, National Security Council meetings, or interviews with participants. Objectivity is also a problem. As one critic notes, “given the often ambiguous data they have to work with, bureaucratic politics analysts run the danger of imposing their theory on the data, rather than testing their theory on the basis of the data.” 139 If one is looking for evidence of bureaucratic politics at work, one is likely to find it. Another problem with the BPM as a theory is that it has generated a paucity of specific hypotheses that can be tested to assess the validity of the theory itself. With the exception of the crucial central hypothesis that “where you stand depends on where you sit,” one is hard pressed to find many others. And as we have stated, this crucial hypothesis seems to be incorrect as often as it is valid. This is crucial because some theorists have argued that the BPM creates no real “value added” unless its supporters can demonstrate a strong and systematic relationship between officials’ preferences and their organizational base. 140 The difficulty of testing specific hypotheses is related to a more general problem. What constitutes proof of the existence of bureaucratic politics? And what evidence is necessary to prove that the decision in question was the result of the bureaucratic politics process? The developers of the model have left us without clear answers to these questions. Given the BPM’s lack of parsimony, its failure to produce many testable hypotheses, the lack of evidence for its central proposition, the relative lack of important BPM-based research done in the forty years since its initiation, and its general inability to progress beyond the status of a paradigm and advance a real theory, many theorists have more or less concluded that the BPM is a dead end. In fact, one retrospective analysis concludes that the BPM has simply failed to produce any systematic knowledge about the explanatory status of the model. 141 This is not to say that the BPM is without merit; it is just extremely difficult to employ. And while the kind of evidence used to provide a BPM explanation of war is not of the statistical or correlational sort with which political scientists feel comfortable, there is certainly more than one way of assessing the validity of empirical theories. 142 It is virtually impossible at this point to determine the extent to which bureaucratic politics has played a role in war initiation. But even if we find that it is useful in explaining only a relatively small percent of cases of war, the BPM still provides the theorist with several important insights on the

152

Chapter 4

causes of war that should not be discarded. And in some cases bureaucratic politics may in fact provide a more satisfying explanation than rival theories. THE ROLE OF GROUP PSYCHOLOGY: GROUPTHINK The final theory we will examine at the small-group level of analysis was developed by Irving Janis, a social psychologist who is interested in international affairs. Janis deliberately chose the term “groupthink” for its Orwellian sound. He describes groupthink as a deterioration of critical thinking, mental efficiency, reality testing, and moral judgment that results “when the group members’ striving for unanimity override their motivation to realistically appraise alternative courses of action.” 143 The decision-making group seeks conformity, harmony, and consensus at the expense of sound policy making. Following is a list of the dominant characteristics of the groupthink syndrome: 1. Members of the group consider loyalty to the group to be the most important objective. 2. Group members seek to foster and maintain consensus, harmony, and unity. 3. Group loyalty requires each member to avoid raising controversial questions, challenging weak arguments made by other members, or criticizing the opinion of the majority. Personal doubts are voluntarily suppressed. As a result, the consensus that seemingly appears is actually an illusion. 4. Dissent is seen as disloyalty to the group. 5. Nonconformists are excluded from the group and some group members act as “mindguards” to pressure possible nonconformists to withhold their objections or mute their criticisms. 6. Group members hold the conviction that the policy positions of the group are moral. 7. Group members hold “hard-headed” attitudes toward out-groups, believing for instance that the opponent is “fiendishly evil” (but also weak and stupid). Stereotyped thinking about outgroups abounds. 8. The attitude of the group is generally characterized by over-optimism, a sense of false security, and invincibility. There is a belief that a group made up of such good and intelligent individuals can do no wrong. As you can imagine, such a situation might lead to serious distortions in the group’s ability to perform rational problem solving. For instance, Janis identifies the following flaws in decision making that may result from groupthink: 1. There is little or no attempt to gain information from experts, seriously impairing the information-gathering process and casting doubt on the objective nature of the search. 2. There is a selective bias toward facts and judgments. 3. Discussions of the group are limited to a few alternative courses of action. 4. There is an incomplete survey of objectives. 5. The group fails to reexamine its preferred solution in order to assess its risks and flaws. The group coalesces around an alternative that is uncritically accepted.

The Substate Level of Analysis

153

6. Assumptions shared by group members are never examined; misconceptions are never corrected. 7. The group neglects to examine fully those courses of action that were initially evaluated as unsatisfactory. 8. Little time is spent examining how the plan might fail; contingency plans are rarely developed. 9. There is a lack of vigilance and sensitivity to the threat of failure. 10. Past decisions that have ended in failure are rationalized. Whereas Allison starts from the proposition that governmental decisions made by groups have a character all their own owing to the political nature of actors representing different institutional interests, Janis starts from the proposition that decisions made by groups are different than decisions made by individuals due to the social nature of the decision-making process. Specifically, small groups—under certain circumstances—have a tendency to seek conformity. Sociologists have long known of the strong pressures for conformity that exist within social groups, and the greater the cohesiveness of the group, the more pressure for conformity. Pressures for conformity come in part from the simple desire to get along with one’s coworkers. Additionally, many individuals fear that if they voice dissent too often, they will lose their effectiveness or their chance for further promotion. Pressures for conformity also arise due to the need for social comparison. Individuals strive to determine whether their opinions (about the best policy to pursue, for instance) are correct. When objective means are unavailable, we evaluate our opinions by comparing them with the opinions of others. If the opinions of the other members of the group coalesce around an opinion different from our own, there is tremendous pressure to write off our opinions as faulty and accept those of our peers. Most important for Janis, highly cohesive in-groups constitute a source of security for their members, which serves to reduce anxiety and to heighten self-esteem. This mechanism is especially needed and relied on in time of stress, because stress increases self-doubt and insecurity. It should come as no surprise that in-group solidarity increases when clashes arise with out-groups. Janis has been careful to point out that not all decision-making groups are subject to the groupthink syndrome. Groupthink can be avoided and often is. We need to know, therefore, what conditions lead to its presence. Janis points out several antecedent conditions that give rise to groupthink; some have to do with the nature of the group and others with the nature of the situation. It might be speculated that the psychological attributes of the individuals within the group would be a factor. For instance, individuals with strong affiliative needs or individuals who are most fearful of disapproval and rejection might be particularly susceptible to groupthink. Although this may be true, Janis contends that all decision makers, even those with high self-esteem, may be vulnerable to groupthink under certain conditions. 144 In Janis’s formulation, the most important antecedent condition for groupthink is the presence of a cohesive in-group—a group whose members are socially compatible, who get along well, who respect and admire each other, who are loyal to each other, and who value the group’s compatibility. Indeed “Janis’s Law” is that the more amicability and esprit de corps

154

Chapter 4

among members of the group, the greater is the danger that independent, critical thought will be replaced by groupthink. But while cohesiveness is a necessary condition for the appearance of groupthink, it is not a sufficient condition. Indeed, group cohesiveness may actually lead to better decisions if individuals feel secure enough in the confines of the group to voice critical viewpoints. On the other hand, noncohesive groups make poor decisions also—though not because of groupthink. Too much conflict between group members may turn the decision process into a power struggle. Thus, cohesiveness is a two-edged sword; too little cohesiveness and the group degenerates into endless squabbling, too much and groupthink takes over. Group cohesiveness leads to groupthink only when other conditions are met. First, a group that is insulated from the input of others is more susceptible to groupthink than one that is open. Second, a group that has directive leadership—that is, leaders who lack impartiality and strongly advocate certain policies—is more likely to succumb to groupthink than those with impartial leadership. Directive leadership makes it difficult for subordinates to challenge the leader’s ideas, thereby inhibiting critical analysis.. Third, a group that lacks norms of methodical decision making is more likely to develop groupthink than one that adheres to more rigorous procedures. Fourth, a group whose members share similar social and cultural backgrounds is more prone to groupthink than a socially and culturally diverse body. Finally, certain provocative situational factors also contribute to the presence of groupthink. Typically, groupthink develops when group members are under a high degree of stress from an external threat. Additionally, group members may suffer from low self-esteem brought about by a recent policy failure, by the recognition that the decision they face may be beyond their competence, or by the fact that they face a difficult moral dilemma. Group members find psychological security and mutual support in the company of like-minded colleagues who agree among themselves that the policies they have devised to deal with a crisis will work. The cohesive interaction of the group preserves each member’s self-esteem. Without the illusion of unanimity brought about by groupthink, the sense of group unity would be lost, gnawing doubts would start to grow, confidence in the group’s problem-solving capacity would shrink, and soon the full emotional impact of all the internal and external sources of stress generated by making a difficult decision would be aroused. 145

Essentially, groupthink behavior acts as the functional equivalent of the defensive avoidance practiced by individuals faced with stressful situations in which failure is a distinct possibility. 146 When a group is moderately or highly cohesive, the more of these antecedent conditions that are present, then the greater the chance that groupthink will occur, leading to a faulty decision. And, of course, the more frequently a group displays the symptoms of groupthink, the worse will be the quality of decision-making (on average). Groupthink Cases: What’s the Evidence? Having outlined his theory, Janis then proceeds to illustrate it with several examples from American foreign policy. He identifies the Kennedy administration’s decision to carry out the

The Substate Level of Analysis

155

Bay of Pigs invasion of Cuba in 1961 as the classic case of groupthink—one that resulted in “a perfect failure.” He also builds convincing arguments that groupthink played a major role in the faulty decisions surrounding three other cases: the Truman administration’s decision to send United Nations troops into North Korea despite Chinese warnings of intervention, the lack of preparation by the U.S. military in Pearl Harbor prior to the Japanese attack in December 1941, and the escalation of the war in Vietnam. (Janis’s 1982 revised edition also identifies the Watergate cover-up by the Nixon administration as a groupthink fiasco, though this case is in the realm of domestic politics rather than foreign policy.) In each case he identifies the symptoms of groupthink within the decision-making unit, the antecedent conditions that might have given rise to groupthink, the resulting flaws in the decision-making process, and the errors in policy that flowed from them. Janis balances these failures by citing how the Truman and Kennedy administrations were able to avoid the pitfalls of groupthink and produced good results during the planning for the Marshall Plan and the Cuban missile crisis. These cases illustrate what Janis calls the vigilant appraisal model, his version of the rational actor model. The latter case is especially significant because the very same decision makers who participated in the Bay of Pigs fiasco learned from their mistakes and took conscious steps to avoid the problem-solving errors of the previous year, illustrating that it is possible for cohesive in-groups to avoid groupthink. Several researchers have re-analyzed Janis’s original case studies and added further examples; others have used laboratory experiments to test groupthink hypotheses. What do we have to show for forty years of research on groupthink? Since this is social science, you already know the answer to the question: The results are mixed. 147 McCauley took a second look at Janis’s original cases, though with one twist. 148 He splits the Cuban missile crisis into two separate cases: (1) an initial decision that a tough response was required—a decision that involved considerable concurrence-seeking behavior that looks like groupthink, and (2) the decision for a blockade—a decision that conforms to Janis’s vigilant appraisal model. McCauley generally agrees with Janis’s identification of the five original cases as groupthink, and he adds the sixth case of the missile crisis (A) decision. Other researchers, led by Philip Tetlock, also looked at Janis’s seven original cases (five groupthink and two vigilant appraisal cases) and three other “candidate cases” proposed by Janis: the Chamberlain government’s appeasement of Germany, the Ford administration’s Mayaguez rescue, and the Carter administration’s attempted rescue of American hostages in Iran. 149 They looked at the views of various authors concerning these cases and used a Q-Sort technique to identify and classify patterns of defective decision making in each case and to compare it with an ideal pattern for groupthink or vigilant appraisal. Like McCauley, the findings generally confirmed Janis’s classification of the original seven cases. They concluded that the British policy of appeasement also fit the groupthink mold, but both the Mayaguez and the Iranian hostage crisis should be classified as non-groupthink decisions. Their statistical tests supported the importance of several structural and procedural faults of the group as antecedents of groupthink; however, they found no support for the importance of a provocative situational context as an antecedent for groupthink. Nor did they find any support for group cohesion—the presumed necessary antecedent condition—as being causally important in bringing about groupthink.

Chapter 4

156

Other retrospective case studies have argued for a groupthink explanation for the Thatcher government’s response to the Falklands Islands crisis, for the Reagan administration’s IranContra fiasco, for the Carter administration’s attempted rescue of hostages in Iran, and for the George H. W. Bush administration’s handling of the Gulf crisis in 1990–91. 150 This last decision is generally considered to be quite a success, not at all a fiasco, and it admonishes us to remember that not all decisions arrived at through the groupthink process lead to disastrous results. 151 Laboratory experiments of groupthink have produced some useful counter-evidence. First, experiments indicate either weak or no support for a relationship between group cohesion and groupthink symptoms. In fact, some studies show that members of noncohesive groups actually engage in more self-censorship. 152 Group cohesion—Janis’s presumed necessary condition—appears completely unnecessary for groupthink. 153 This finding is also supported by case studies. For instance, Yetiv points out that while esprit de corps contributed to high cohesiveness within President George H. W. Bush’s inner “Gang of Eight,” the bonds of friendship also allowed various decision makers (Bush, Cheney, and Baker) to be completely blunt with each other without fearing it would undermine their further interactions—normally a good thing. 154 The rather uniform drubbing of the cohesiveness-leads-to-groupthink argument actually has an easy explanation. As ‘t Hart notes, the effect of group cohesiveness is that it “increases the power of group norms, and these may or may not favor unanimity” or concurrence seeking. 155 The group norms, especially in groups at a higher stage of development (as opposed to newly created, ad hoc groups, for instance), may in fact foster greater critical thinking, and members may feel secure enough socially and politically to challenge each other’s zanier schemes. Thus norms and the level of group development emerge as important factors. Second, experiments consistently support the notion that directive leadership is linked to groupthink and defective decision making. 156 When strong leaders let other members of the group know early on how they stand on the issues, compliance from lower-status group members becomes more likely and the search for alternatives is short-circuited. But, of course, using directive leadership as an antecedent condition of groupthink creates the problem of circularity, since directive leadership itself is a symptom of defective decision making. Overall, while several additional cases of groupthink fiascoes have been tentatively identified, and many assessments by case studies and laboratory experiments have supported some elements of the groupthink model, we must agree with one set of researchers that there is little support for the complete groupthink model as originally documented by Janis. As Fuller and Aldag conclude, “to our knowledge, no study of groupthink has fully tested the model, and in no study were all results consistent with the model.” 157 GROUP CONCURRENCE BEYOND GROUPTHINK Risky Shifts and Choice Shifts Several scholars have identified decision-making concepts that are compatible with groupthink and might be incorporated to strengthen the theory, while others may be useful in

The Substate Level of Analysis

157

providing alternative explanations for concurrence seeking in group situations. For instance, Janis himself notes a tendency for groupthink decisions to be “hardheaded” toward outgroups. He also sees a propensity for group members to take stances of “virility.” 158 Others have referred to this phenomenon as the “hairy chest syndrome.” Richard Barnet, for instance, gives this picture of American policy makers: One of the first lessons a national security manager learns . . . is that toughness is the most highly prized virtue. The man who is ready to recommend using violence against foreigners, even where he is overruled, does not damage his reputation for prudence, soundness, or imagination, but the man who recommends putting an issue to the U.N., seeking negotiations, or, horror of horror, “doing nothing,” quickly becomes known as “soft.” 159

The hairy chest syndrome is somewhat similar to the concept of the risky shift developed by social psychologists. In the early 1960s researchers accumulated a sizable amount of evidence that indicated that while individuals in their own problem-solving experiences would choose more conservative, risk-averse solutions, when they were asked to make decisions as part of a group, they tended to support much riskier solutions to the same problems. 160 Compared with decisions by individuals, the making of decisions in a group context has a distinct impact on the nature of the decision. This initial conclusion has recently been modified as a result of further research. The shifts are now seen more appropriately as merely “choice shifts” that may proceed in either direction—toward more risky or more cautious solutions. 161 The evidence now suggests the effect of making decisions in groups is to create a tendency toward group polarization. The group process enhances and magnifies whichever point of view—risk-acceptant or risk-averse—is initially dominant within the group. 162 The effect of the group is to make the solution more extreme (in either direction) than the one preferred by separate individuals, since group members are prone to reinforce each other’s extreme positions. To the extent that a shift to a riskier position takes place within the group, several overlapping explanations have been put forward. 163 First, a risky shift could be attributed to the psychological bolstering and peer pressure that is part of the groupthink syndrome. This is consistent with Janis’s view that group members tend to seek and maintain group cohesion, and one way to do this is to support the prevailing view—even if that view is extreme—rather than to challenge it. Second, a risky shift might be attributed to the recognition that group decisions relieve individuals of direct personal responsibility for risky courses of action; the risks are shared, making individual acceptance of them easier. Third, strong, confident leaders who are risk-acceptant may, through the process of group interaction, be able to pull along more reluctant, undecided members. Fourth, the shift may simply represent an intensification or strengthening of the initial predispositions of individuals through their association with the group. The makeup of the group may have an effect on the direction of the choice shift. Andy Semmel and Dean Minix carried out a number of experiments using simulations of different national security crisis scenarios to study the shift in choices between individual and group recommendations. They used three different sets of groups: university students, U.S. Army officers, and ROTC cadets. In the experiments, all of the Army officer groups shifted to riskier

158

Chapter 4

recommendations, compared with what they, on the average, had chosen initially as individuals—almost always recommending the use or threat of force. Almost all of the ROTC groups shifted to more extreme options as well. But most of the student groups shifted to a milder set of preferences, generally preferring negotiations. 164 Thus, the direction the shift took depended on the composition of the group and its norms. Initial beliefs of the group members laid the foundation for the option chosen, but the group discussion amplified the initial choices. Minix’s analysis of the transcripts of the group discussion indicated that groups that made risky shifts tended to use historical analogies in their analysis of the problem, tended to focus on one option from the start, and tended to be dragged in the direction of the most extreme member of the group, who also tended to hold an informal leadership role within the group—all patterns one would expect to find in groupthink. 165 One clear implication is that higher-quality decisions are more likely when the group is composed of a heterogeneous mix of individuals recruited from different organizational subunits, a conclusion with which Janis would concur. 166 Although the risky shift is not part of Janis’s groupthink theory, it does share some interesting similarities with it. It is possible that the need for group security, the consensus seeking, and the general deterioration of decision-making skills that Janis describes as groupthink may lead to a shift by the group to riskier actions than the same individuals would ordinarily choose if deciding by themselves. New Wave Groupthink Theorizing More recent scholarship in the groupthink tradition has expanded and clarified the theory and has put forward some alternative mechanisms by which concurrence seeking may take place in governments. Here are a few of the insights produced by this new wave of groupthink scholars. Prospect Polarization Glen Whyte argues for a prospect polarization theory as an alternative explanation for how excessive concurrence seeking by groups creates foreign policy fiascoes. Prospect polarization combines insights from group polarization and from prospect theory. 167 He argues that decisions that lead to foreign policy fiascoes “are most naturally framed, whether appropriately or not, as a choice between two or more unattractive options.” 168 In other words decision makers see themselves in the realm of losses. One option clearly entails a negative deviation from the reference point: There will be a certain loss unless action is taken. Another option holds out the possibility that the adverse situation might be avoided, but at the risk that an even worse situation might be created. In this situation, prospect theory suggests that individuals will prefer risk-laden options in order to avoid losses. Whyte argues that this tendency toward risk taking is actually exacerbated in decisions made by groups. Social psychology research has demonstrated that group discussion creates pressures for uniformity in group members. Specifically, there is a tendency for members to adopt the position of the majority. Whyte’s argument here is both logically compelling and politically disturbing. If we assume a five-person group selected randomly, four members (80 percent)

The Substate Level of Analysis

159

are likely to be risk-seeking in the domain of losses. (This is what the prospect theory research indicates.) Mathematically, 94 percent of all five-person groups will therefore contain a majority who prefer a risky option at the outset of group discussion. Thus, when choice shift (group polarization) takes place within the group, it is likely to be in the direction of risky options. 169 The effect of group discussion and interaction is to amplify the likely majority preference for risky action that already exists. The proximate cause of foreign policy fiascoes is not groupthink, but the initial framing of the situation as within the realm of losses. That, coupled with group polarization, creates a push toward high levels of risk-acceptance with the resultant high probability of policy failure. Whyte argues that groups are just as subject to the framing effects of prospect theory as individuals, and that in fact choices made by groups should be more consistent with prospect theory than choices made by single individuals. 170 Compliance and Manipulation McCauley has extended and clarified groupthink theory with regard to the nature of group agreement and the mechanism by which concurrence takes place. 171 He makes a crucial distinction between two types of agreement: (1) Internalization refers to the private acceptance of members of the dominant view. This is essentially what Janis had in mind in his discussion of concurrence brought about by group amicability. However, internalization can also come about because a group member is simply persuaded by another’s arguments. (2) Compliance refers to the mere public agreement with the dominant view but without private acceptance. Here, agreement is most likely produced by considerations of status and power within the group rather than desire for amicability. While McCauley acknowledges the internalization process, he argues that mere compliance more accurately describes the group dynamics in the Bay of Pigs crisis and the Johnson administration’s handling of Vietnam escalation. Others point out that group members can manipulate small-group decisions for political and personal reasons. 172 This can be done (1) by changing the group’s procedures, for instance through agenda setting or through the framing of issues in a particular way; (2) by altering the decision-making structure through the deliberate inclusion or exclusion of certain individuals, or (3) by the use of personal pressure on group members to change their positions. From the outside this looks a lot like mind-guarding and direct pressure on dissenters that Janis described, but it is more complex. The authors describe competitive bureaucratic politics that results in outcomes similar to groupthink—group concurrence. However, in this case, concurrence is not the result of a group social-psychological process brought about by the need to relieve anxiety created by stress; it is induced by the political manipulation of one (or more) group members. Paul ‘t Hart’s Integrated Model of Groupthink The contributions of Paul ‘t Hart to the reformulation of groupthink theory are well known. 173 ‘t Hart has updated Janis’s theory to take into account much of the recent literature of social psychology and public administration. He has broadened (and loosened) the concept of groupthink to prevent it from being confined to cases of highly cohesive groups. For ‘t Hart, there appear to be a limited number of general structural/situational settings that are required for

160

Chapter 4

groupthink to take place: (1) a small group exists as a key decision-making body, (2) the group must be relatively informal or unstructured, and (3) a certain amount of stress due to internal or external threats must be present, causing anxiety and uncertainty for group members. 174 Most new-wave groupthink theorists would agree that stress is a necessary condition for groupthink, implying that it is likely to be limited to the most important types of policy decisions. Like McCauley, ‘t Hart has suggested some important distinctions. First, he distinguishes between different modes or outcomes of groupthink: (1) Groupthink I is driven by a pessimistic view of potential policy outcomes. This is consistent with Janis’s original formulation, in which stress induced by threatening situations and the potential for failure induces members of the group with low self-esteem to engage in collective avoidance of the risks as a way of coping with anxiety. (2) Groupthink II is driven by an optimistic view of the situation. Here, a more high-confidence group engages in a kind of collective over-optimism; the chosen policy is an adventuristic response to the perception of opportunities to achieve a major policy or career success—often involving the pursuit of a pet project of one of the members. This path to groupthink is most likely during an administration’s early “honeymoon” period—the Bay of Pigs disaster, for instance. ‘t Hart also distinguishes between two types of pressure felt by group members: horizontal pressures from colleagues within the group who have similar status, and vertical pressure or influence that stem from status differences (formal or informal) between group members. Group leaders (who can impose vertical pressure) can be either formal leaders or de facto leaders recognized by other group members. 175 ‘t Hart identifies several causal pathways to groupthink—each of which is deemed to be in itself a sufficient cause of groupthink. 176 The first pathway is group cohesiveness—the path emphasized by Janis. Here, a highly cohesive unit is driven to excessive concurrence-seeking in order to mitigate the anxiety of the stressful situation. Consistent with other researchers, ‘t Hart finds that strong affective ties between group members (high group social cohesion) are not a necessary condition for groupthink. He believes that governmental decision-making groups rarely display the type of amicability and esprit de corps Janis posited. 177 The most important pathway to groupthink for ‘t Hart is simply anticipatory compliance. 178 Here, external or internal stress creates a tendency for group members to search for a strong group leader, especially “if there are no tightly fixed decision-making procedures, or when the formal group leader does not display sufficient leadership in the given situation.” 179 In this pathway, groupthink occurs “because low-status members anticipate thoughts, wishes, or commands from leader figures, and adapt their own thinking and action accordingly.” 180 Once leaders are identified, members engage in concurrence seeking, depending on cues from the leader. Members concur with the leader even in advance of the leader pronouncing his views. Anticipatory compliance appears to ‘t Hart to be the process at work in the Reagan administration’s insidious dealings in the nefarious “arms-for-hostages” deal in the Iran-Contra affair. 181 While the extent of President Reagan’s personal involvement in the decision process is murky, it appears that most high-level members of the decision group anticipated what the president probably wished them to do.

The Substate Level of Analysis

161

The construction of a groupthink theory that does not rely on group cohesion is important. Most governmental groups lack social amicability and tend to look more like conflict-ridden groups described by the bureaucratic politics model. But ‘t Hart notes that while the overall general context of governmental groups may be conflictual, centered around the promotion of individual or organizational goals, intergroup conflict may paradoxically increase intragroup cohesiveness. Factions or subgroups may become more cohesive than the larger groups of which they are a part, thus becoming highly prone to groupthink. 182 ‘t Hart notes that in typical BPM settings one can identify several processes or mechanisms that can be associated with groupthink tendencies. Let’s mention just two of the most important: (1) Groups may manufacture consensus by reducing the number of actors in the decision process who oppose the dominant policy options. (2) Leaders may attempt to dominate the decisional process and impose “directive leadership,” creating pressures for consensus, especially through anticipatory compliance. Thus, for ‘t Hart, BPM and groupthink models are not mutually exclusive. Two other new-wave groupthink theorists, Stern and Sundelius, add that typical BPM considerations—such as the desire to avoid making enemies, the desire to be on the winning side of policy arguments, the need to remain effective and to “save one’s political capital,” and the desire to remain a member of the decision-making group—all lead in the direction of conformity dynamics and possibly even a full-blown groupthink. 183 Group decision processes that work according to the BPM may end up creating many of the symptoms of groupthink. Indeed, Cashman and Robinson find in their case studies of war that neither groupthink nor the BPM model ever appear in their ideal or pristine form; the decision process frequently tends to look like a mix of the two. 184 For instance, many groupthink symptoms are apparent in the George W. Bush administration’s decision to go to war in Iraq in 2003. But the president’s advisers were anything but cohesive. Secretary of State Colin Powell was barely on speaking terms with Secretary of Defense Rumsfeld or Vice President Cheney, and Rumsfeld exuded little but disdain for National Security Adviser Condoleezza Rice. Two rather heated political coalitions existed— with the president, the office of the vice president, the secretary of defense, and the Joint Chiefs of Staff all lined up in favor of a military option on the one hand and the State Department (supported by DCI Georg Tenet) arguing against it. One could argue that in terms of ‘t Hart’s analysis two important things happened. (1) The two factions became more internally cohesive—especially the State Department. (2) The BPM process of silencing opposition forces left the game pretty much in the hands of the pro-war faction, with Powell and Tenet sitting on the sidelines and offering reluctant public compliance rather than a real internalization of the dominant viewpoint. 185 Diligent use of BPM techniques resulted in a groupthink-like final stage. This is pretty much the analysis of Dina Badie. 186 She argues that post 9-11 stress (and pressure to produce an effective policy response to Al Qaeda’s attack), combined with promotional leadership techniques and intergroup conflict, created the condition of groupthink. She notes however, that administration hawks and skeptics became victims of groupthink through different pathways. Administration hawks immediately internalized the concept of the Bush Doctrine and the need for a preventive strike against Iraq. Horizontal pressures among like-

162

Chapter 4

minded hawks created group pressures for conformity much as Janis originally speculated—as a need for cohesion after traumatic events (for which they bore little responsibility). Administration skeptics (Powell, Armitage, Tenet) faced vertical pressure to conform, as group leaders—Bush, Cheney, and Rumsfeld—engaged in promotional leadership techniques, including mind-guarding, to influence other group members. The combined vertical pressure and mind-guarding appear to have driven Powell and Tenet toward conformity. Badie argues that the skeptics were not excluded from the group, but were pressured to conform by the group’s leaders (Bush personally requested Powell’s support). Badie argues that Powell eventually internalized the new policy and began to focus on the task of gaining allies for the U.S. intervention. To some extent as well, he succumbed to anticipatory compliance once it was clear the direction the group leader (Bush) was intent on taking. However, Tenet’s acceptance of the policy was somewhat different. His desire to retain influence in the core group, even though he doubted its policies, led him to engage in self-censorship in the face of pressure from the hawks. He appears never to have really internalized the new policy, but he did produce the “actionable intelligence” the administration required. Newgroup Syndrome This brings us to a final revision of groupthink theory, called newgroup syndrome (NGS). Eric Stern and Bengt Sundelius have suggested that newly formed groups will experience particularly strong pressures for conformity. 187 Why should new groups be subject to extreme pressures for conformity? One reason is “the shadow of the future.” If group members perceive that they will be members of the group for an extended period of time, they will realize that they must get along with the same people and must try not to antagonize them; cooperative behavior will be beneficial in the long run. After members indicate their willingness to be good team players, they will develop a certain amount of “idiosyncrasy credit” and be allowed to get away with a certain amount of dissent by group leaders or dominant group members. On the other hand, if the shadow of the future is short and the group is likely to be only a onetime, short-lived thing, then there are fewer inhibitions or costs connected to expressing dissent. 188 New groups that are in the early stages (called forming stage) of their development lack a common subculture, an agreed-upon set of roles for members (though a rudimentary pecking order may have developed), and a set of well-developed procedural norms for making decisions. Indeed, uncertainty over group norms may be hampered by the fact that group members all bring external baggage with them from their previous experiences. As Stern says, “This vacuum creates uncertainty among members who are likely to be anxious, tentative, dependent, and therefore particularly inclined to take direction from a leader or other assertive group members within the group.” 189 Members may even fear exclusion from the group or being used as a scapegoat for group bungling. All of this creates incentives for members to avoid the expression of dissent. This is consistent with what social psychologists have been telling us about groups for years: All sorts of groups rather rapidly develop informal pecking orders—much like the chimps we discussed in chapter 2. These emerging hierarchies have a strong effect on communication and interaction with the group. As Stern and Sundelius say, “High status members

The Substate Level of Analysis

163

tend to set the norms to which others conform and may feel free to deviate from emergent or established norms to a greater extent than the lower-status members.” 190 Severe asymmetries of status and power are conducive to conformity. In early stages of group development lowerstatus members will be especially wary of the need not to cross higher-status members, making them prone to newgroup syndrome. 191 While NGS can take place within the confines of a moderately cohesive group, it may also take on a more heightened form in politically conflictual environments such as one might find in groups operating under the bureaucratic politics model. Not all new groups will develop NGS. For Stern and Sundelius, the crucial factor appears to be whether group leaders can quickly intervene to establish roles and norms that encourage rational processes and vigilant problem solving. 192 Stern finds that NGS provides an alternative explanation for the Bay of Pigs disaster. The decision group was ad hoc and relatively informal, and it was created in the early weeks of the Kennedy administration to deal with a plan devised by the outgoing Eisenhower administration. Contrary to the myth of Camelot, many of Kennedy’s advisers were new to each other and to the president himself. As a new group, it lacked a set of group norms and procedures— something exacerbated by Kennedy’s own rather loose, informal style. Its members also lacked much idiosyncrasy credit. These social and political factors were a much more powerful motivator of group conformity than the need for a highly cohesive group to allay stress in the face of strong doubts about the chances of success. In fact, Stern argues that to the extent that groupthink played a role in the Bay of Pigs, ‘t Hart’s second type of groupthink— collective over-optimism—was much more to blame than collective pessimism and defensive avoidance. 193 This was not a case in which the group was motivated by stress induced by the realization that the option chosen by group leader was likely to fail. The reluctance of group members to challenge the rather hare-brained schemes of the CIA for the invasion is now understandable. “Wanting to give their new colleagues the benefit of the doubt and to avoid unnecessary confrontation, the New Frontiersmen guarded their political capital, and wholeheartedly or reluctantly gave consent to the CIA plan.” 194 In sum, the primary difference between the Kennedy administration’s decisions at the Bay of Pigs and the subsequent Cuban missile crisis is that in 1961 the decision group was a new group, but by 1962 it had evolved and developed into a more highly performing group. Groupthink as a Theory Groupthink has some of the same theoretical defects as the bureaucratic politics model. It requires a great deal of the sort of information that is difficult to come by—such as minutes of cabinet or national security council meetings—making replicability a problem. It also requires a rather long, fine-grained narrative, though this narrative can be more precisely organized than in the BPM, thanks to Janis’s precision in laying out the theory. And like bureaucratic politics, groupthink is a fairly complex theory, though the relationships between the variables are more clear-cut than in the BPM. On the positive side, Janis has taken great care to specify how the theory can be tested (and falsified) against real-world conditions and to specify the conditions under which groupthink should arise—again in a way that is observable.

164

Chapter 4

The theory charts a clear causal path between the presence of certain antecedent conditions (including one necessary condition, a cohesive in-group), the independent variable (the presence of observable groupthink symptoms), and the dependent variable (the observable symptoms of defective decision making). The links between antecedent conditions, independent variables, and dependent variables are stated in a probabilistic, if . . . then, manner. If a group is modestly or highly cohesive, then the more of the antecedent conditions that are present, the greater the chance that the group will experience groupthink. The more numerous the symptoms of groupthink, then the more likely that the decision will be defective. The more defects in the decision-making process, the lower the probability that the policy will be a success. Case studies have been developed to confirm the existence of groupthink in foreign policy decisions of the American government, though these in no way indicate how widespread the groupthink phenomenon may be. Herek, Janis, and Huth have studied the relationship between the quality of decisionmaking procedures and foreign policy outcomes. Their investigation of American decision making in nineteen Cold War crises reveals that high-quality decision procedures are associated with better outcomes, while defective decision procedures are associated with outcomes that have more adverse effects on American interests and are more likely to increase the level of international conflict. 195 Thus, there appears to be a direct link between the quality of information processing by groups (Janis’s symptoms of defective decision making) and outcomes. A second study by Schafer and Crichlow addressed another link in Janis’s chain. 196 Using the same nineteen crises as Herek et al. and the same source materials, they investigated the connection between the presence of Janis’s antecedent conditions of groupthink and symptoms of defective decision making, which they refer to as information processing errors. (Their list of antecedent conditions does not include the presumed necessary condition, group cohesion.) They find some antecedent conditions to be common, while others are rare. For instance, group insulation was found in only one case (the decision to send U.S. ground forces to Vietnam); on the other hand, closed-mindedness was found in twelve cases, group homogeneity in eleven cases, and high stress in nine cases. Four of the ten antecedent conditions were related to defective decision making in a statistically significant way: lack of a tradition of impartial leadership, lack of methodical procedures for decision making, overestimation of the group, and closed-mindedness. A fifth, pressure for uniformity, was nearly significant. An index of these five variables was a good predictor of the number of information processing errors a group was likely to make. The findings indicate that some of Janis’s antecedent conditions are less strongly related to defective decision making than originally thought. Situational factors such as time, stress, and recent policy failures seemed to have a fairly low impact, as did the presence of homogeneous groups. Nevertheless, taking the two studies together, it appears that Janis’s links in the causal chain from antecedent conditions to policy fiascoes can at least partially be validated. Janis recognizes the problems associated with a single-factor theory of behavior and is therefore careful to note that other variables that are not part of the groupthink syndrome can cause defective decision making. As he points out, “blunders have all sorts of causes—some, like informational overload, being magnified by groupthink; others, like sheer incompetence

The Substate Level of Analysis

165

or ignorance, having nothing at all to do with groupthink.” 197 Janis sees the groupthink syndrome as likely to be a contributing cause that augments the influence of other sources of error, though it can sometimes be the primary cause. 198 Does Groupthink Have Direct Relevance to the Causes of War? We must note that Janis is not trying to discover the cause of war, only the cause of policy fiascoes. Of his original cases, one deals with a decision for the use of force short of war (Bay of Pigs), one with the lack of preparation in defense of an attack (Pearl Harbor), and two with decisions to escalate wars already in progress (Korea and Vietnam). The only other academic study that uses a groupthink explanation for war is Steve Yetiv’s analysis of the Bush administration’s decision to reverse Iraq’s invasion of Kuwait in 1990 by force if necessary. 199 It is also useful to note that virtually all of the foreign policy cases attributed to groupthink involve decisions made over a lengthy period of time, rather than in the classic crisis situation involving limited time for a decision. Groupthink is not a theory of war, but a theory of decision making which, as ‘t Hart says “can at best, be only a partial theory, i.e., relevant to only a limited category of decision making events.” 200 That limited domain includes nonroutine situations dealt with at the top of a governmental hierarchy, like crisis situations. Though groupthink is possible in these situations, it is apparently fairly rare, and when it does occur, it hardly ever will provide a complete explanation for why a particular decision was made. To the extent that government decisions to go to war are arrived at through a groupthink process, we could say that the process itself was at least in part a cause of the war. The nature of social interactions within the group was probably responsible for a decision-making process that deviated significantly from rational problem solving. A better (and presumably more peaceful or less risky) solution could have been devised through a more rational process. Especially important in this regard would seem to be the interrelated patterns of lack of critical analysis, lack of vigilance concerning risk and error, and feelings of over-optimism. Comparison of Groupthink and the Bureaucratic Politics Model Both groupthink and the bureaucratic politics model are empirical theories. Neither groupthink theorists nor BPM theorists advocate that policies should be made through the groupthink or BPM method; they only state that policies are in fact made that way, whether we like it or not. They both depict policy making as nonrational and suggest that governments frequently fail to make the best decisions in international affairs. Both theories deal with decision making by small groups—politburos, cabinets, juntas, national security councils, interagency committees, and so on. Each suggests that a certain set of group dynamics has a negative effect on decision-making—though for different reasons. 201 Both recognize that conflicts and disagreements over policy are likely within groups, but in the groupthink syndrome policy makers try to avoid conflict by seeking group cohesion, while conflict is managed in the bureaucratic politics process by bargaining and other political maneuvers among players. 202 The major difference between the theories is that groupthink envisions a decision process in which group cohesiveness, unity, and harmony are paramount, while BPM theorists see

166

Chapter 4

group dissension, division, and conflict as predominant. Charles Hermann sees the pivotal difference as the fact that in groupthink the individuals attach their primary loyalty to the decision-making group itself, while in bureaucratic politics most of the players owe their primarily allegiances to the outside groups they represent. 203 This factor appears decisive in determining whether consensual or conflictual politics are pursued. Finally, neither theory is specifically a theory of war; both are general theories of decision making that can be applied to decisions about war. Neither claims even to be able to explain all decisions made by governments, let alone all decisions for war. While both theories are rather complex and difficult to apply, and while they are most likely to provide only subsidiary explanations for most cases of war, both may occasionally provide important insights into the initiation of particular wars. PRESCRIPTIONS FOR BETTER SMALL-GROUP DECISION MAKING If flawed and irrational government decision-making processes do in fact play a major role in causing war, what is the solution? Presumably, we should be able to find and implement better methods of decision making. Whether the problem is bureaucratic politics or groupthink, the solution might be to devise a policy-making process that approximates as closely as possible the rational actor model. 204 Janis offers several techniques of his own to deal with the problems associated with the groupthink syndrome. 1. The leader should assign each group member the role of critical evaluator and encourage all group members to air their objections and doubts. 2. Leaders should remain impartial and refrain from staking out their initial preferences so as not to influence other group members. 3. Several independent policy-planning and evaluation groups might be set up on each policy question. 4. The group should occasionally divide into two or more subgroups under different chairmen to reduce the possibility that the entire group will develop a concurrenceseeking norm. 5. Each group member should discuss the group’s deliberations with associates in his or her department or agency and report back their reactions. 6. Outside experts who are not core members should be invited to each meeting on a staggered basis and encouraged to challenge the views of the core group. 7. At least one member of the group should be assigned the role of a devil’s advocate at each meeting. 8. Group members should set aside a block of time to survey all warning signals from rival states and construct alternative scenarios of the rival’s intentions. An outsider might be enlisted as a “Cassandra’s advocate” to call attention to alarming possibilities that might otherwise be overlooked.

The Substate Level of Analysis

167

9. After a preliminary consensus is reached, the group should hold a “second chance” meeting at which members would be expected to express whatever residual doubts they might have and to generally rethink the entire issue. In this same vein, Gary Klein has suggested an interesting technique that he calls a premortem. He proposes that just before a group has committed itself to a policy solution, it should hold a session that would be kicked off by the following task: “Imagine that we are a year into the future. We implemented the plan as it now exists. The outcome was a disaster. Please take 5 to 10 minutes to write a brief history of that disaster.” 205 The primary virtue of this exercise is that it legitimizes suppressed doubts—a primary pathway to policy failures in groupthink— and it encourages supporters to search for potential threats that the group has thus far failed to identify. This would certainly be considered a good leadership technique. Alexander George has also put forward a highly developed decision-making scheme designed to address the problems that develop out of bureaucratic politics. 206 His approach, which he calls “multiple advocacy,” is a normative (prescriptive) theory of decision-making intended to give guidance to those practitioners who actually make government decisions. George realizes that in spite of its well-documented defects, the bureaucratic politics process also has positive components. After all, bureaucratic politics is pluralistic, and therefore it is a process in which different positions will be put forward and (presumably) taken into consideration. It therefore avoids the problem of artificial consensus that one finds in groupthink. Pluralistic procedures in which different groups compete to influence policy can potentially be quite healthy for the decision-making process, and a certain degree of conflict is beneficial in problem solving—if the conflict can be managed and resolved properly. Unfortunately, conflict is handled in an unstructured and unregulated manner. The policy options that are put forward in this process are quite limited. Unpopular viewpoints that do not represent the mainstream opinion of any particular agency or department are given little attention. Furthermore, the most powerful actors or coalitions—not necessarily those who argue for the best solution—win the policy battle. Institutional and personal interests make it difficult for detached political analysis and logic to prevail. George’s multiple advocacy approach calls for a balanced, open, and managed process of debate centered around an institutionalized system of advocates, each of whom would make the best case for a particular option, ensuring that a wide range of options would be addressed by the group. A top government official (probably the president’s adviser for national security affairs in the United States) would play the role of a custodian manager. This official would be an honest broker of ideas, a coordinator who would ensure fair competition. She would see that all the options were represented by an advocate and that all advocates had equal resources—such as influence, competence, information, analytical resources, and bargaining and communication skills. She would make sure that time was scheduled for adequate debate and discussion. She would coordinate the independent analysis of options and objectives and monitor the policy-making process for malfunctions. The manager would refrain from being either an advocate herself or an adviser to the leader or a spokesman for the administration. The leader would listen to the presentation of the options and to the debate that followed. The leader asks questions, evaluates alternatives, and then chooses among the options.

168

Chapter 4

If this sounds too good to be true, it probably is. Multiple advocacy is certainly not without its flaws. The process may create greater variety, ambiguity, and complexity than is necessary for a good decision. Confronted with all the argumentation the system would spew forth, the leader may be less able to decide which option is best than if he were given fewer alternatives to choose from. One critic points out, in the context of data overload, uncertainty and time constraints, multiple advocacy may in effect give all of the various viewpoints an aura of empirical respectability and allow a leader to choose whichever accords with his predisposition. 207

The practicality of these remedies is debatable. Both Janis and George recognize the rather hardy resistance of small groups to rational procedures and then recommend that the groups should try harder to be more rational. As Richard Ned Lebow points out, their prescriptions are based on the assumption that leaders will be willing to make a serious effort to structure the decision-making process so as to encourage and enhance critical thinking and dissent. But this may be totally unrealistic. While there are certainly exceptions, most leaders dislike criticism and dissent; it threatens their authority (or at least they think it does), it loosens their control over the decision process, and it may be interpreted by their opponents as a sign of weakness. Thus, leaders may be both psychologically and politically unwilling to accept even the soundest criticism. 208 CONCLUSION None of the theories of group decision making discussed here—expected utility/selectorate theory, bounded rationality, poliheuristic theory, the organizational process model, the bureaucratic politics model, groupthink, and newgroup syndrome—provide a clear-cut answer to the question of why wars occur. But they do offer insight into decision making in the real world. Governments are not automatons that inexorably grind out solutions to problems by identifying national goals, rigorously analyzing potential solutions, computing likely outcomes of choices, and choosing policies that will optimally maximize interests. Instead, they seek procedural shortcuts, are unduly influenced by group leaders, are subject to the organizational interests of participants, and make decisions through political processes (bargaining, logrolling, coalition building, etc.) rather than rational processes. Almost all of this leads in one direction: toward the proposition that decisions for war—like all governmental decisions—are rather less subject to “pure rational strategic choice” than we would like to think and are the product of nonrational processes that, at their extreme, may become dysfunctional processes. Wars may, in the end, be unnecessary—at least in the sense that they may be the result of faulty decision processes.

Chapter Five

The State Level of Analysis, Part I Political, Economic, and Demographic Factors

In Switzerland, they had brotherly love, and they had five hundred years of democracy and peace, and what did they produce? The cuckoo clock. —Orson Welles The strong do what they can, and the weak suffer what they must. —Thucydides Hard times breed hard lines. —Strobe Talbott

Since most international relations scholars see states as the primary actors in international politics, it should come as no surprise that many theorists of international conflict focus on the nature of the state itself, rather than individual leaders or decision-making small groups, as the primary determinant of war. The underlying assumption is that a certain national attribute (or some combination of attributes) influences how states behave. States with similar characteristics should act in similar ways. Personality variables and the psychological makeup of national leaders are relatively unimportant since attributes of the state itself compel decision makers to act in certain ways. 1 One of the most interesting discoveries of the research on war is that states are not equally violent; great variation in conflict activity exists among the states of the world. Some of the first data-based research on nineteenth- and twentieth-century warfare conducted by J. David Singer and Melvin Small concluded that “most of the war in the system has been accounted for by a small fraction of nations.” 2 Indeed, data from the Correlates of War Project (COW) indicated that of the 176 members of the interstate system from 1816 to 1980, 94 states (53.4 percent) never participated in interstate war. More recent explorations of the historical record by Zeev Maoz confirm the pattern of extreme inequality among states with regard to both participation in militarized international disputes (MIDs) and wars. 3 Surveying the period from 1816 to 1992, Maoz reports that nearly 57 percent of all states with twenty or more years of existence as sovereign members of the international system were not involved in a single war. On the other hand, the top 10 percent of states in the system accounted for 56.8 percent of 169

170

Chapter 5

all wars. States highly involved in warfare, which Maoz refers to as “fightaholics,” come in two varieties: (1) young states in the Middle East and East Asia and (2) major powers. What makes some states more warlike than others? As usual there is no lack of competition for answers. Maoz’s analysis of the empirical record provides some relatively straightforward answers. Most states that abstained from war were militarily and economically weak, ranked low on power-related status, achieved independence through an evolutionary process, and were politically democratic at independence. “Fightaholic” states that were disproportionately involved in wars were mirror images of pacifist states: They had high levels of military and economic capabilities, held status as either major powers or regional powers, were likely to have achieved independence through a revolutionary process, and were not democratic at birth. International relations theorists have generally emphasized five general factors that might have a bearing on this question: (1) the type of government the state has, (2) the type of economic system the state has, or certain economic factors that are present in the state, (3) certain demographic, cultural, physical, or geographic attributes of the state, (4) the degree of political instability that exists in the state, and (5) the state’s previous war involvement. The purpose of this chapter is to investigate some of these theories. LIBERALISM: SOME BASICS As most students of international politics know, the two most basic approaches to international relations are realism (or realpolitik) and liberalism (sometimes referred to as idealism). Since a major focus of this particular chapter is on the relationship between democracy and war, this is as good a place as any to provide a brief summary of the liberal approach to international relations. (We will include a similar summary of realist theory in a later chapter.) Liberalism is a theory of domestic politics and economics with implications for international politics. The intellectual roots of liberal thought can be traced to political philosophers like Adam Smith, John Locke, Jean-Jacques Rousseau, and Immanuel Kant—but also to more modern politicians like Woodrow Wilson. Liberalism is not a single, unified theory but a family of related theories. These theories are found at different levels of analysis, but primarily they are residents of the state level and dyadic level. The following is a short list of the essential elements of liberal thought that are relevant to international relations. First, the liberal view of human nature is optimistic; humankind is seen as basically good, rational, and peace-loving. The human condition is generally harmonious; people are able to get along with each other due to reason, mutual trust, and the recognition of shared interests. Second, the behavior of states in the international system is analogous to the relationships between individuals within human communities. Conflicts can be resolved between states through the use of law, the administration of good government, through democratic procedures and through negotiation, diplomacy, and mutual adjustment. Third, while it is true that international politics is anarchic due to the lack of global government, this does not mean that there is a constant condition of war. Cooperation between states is possible even under the condition of anarchy. International politics is seen as a mix of conflict and cooperation: Peace is normal; war is abnormal. The state of nature (without global government) is

The State Level of Analysis, Part I

171

seen more as a “troubled peace” than as a Hobbesian “war of all against all.” Liberals do not see international relations as a zero-sum game of competition, but as a nonzero-sum game where all sides can make mutual gains. They emphasize the cooperative side of politics rather than the conflictual side. Fourth, certain moral principles are universal and based on natural law: justice, civil rights, individual liberty, democratic government, self-determination, political and moral equality, and tolerance for diversity. (And based on these moral principles, wars can be judged to be just or unjust.) Liberals provide several answers to the question of what causes war (and what is conducive to peace). First, the nature of the international system (anarchy) is a problem. Thus, if the lack of an overarching global authority to keep the peace is the problem, liberals are likely to support international organizations, the creation of international regimes based on legitimate norms and international law, and ultimately world governance as a solution to the problem. While world government may not be immediately attainable, international organizations that provide collective security (like the United Nations) can help to deter aggression. Second, the nature of certain types of states is problematic. Wars are not due to the evil nature of human beings but to the existence of bad institutions that corrupt human behavior and foster conflict. While democracies are seen as peaceful, autocratic states are seen as more likely to pursue aggressive, militaristic, and warlike policies. Hence, decreasing the supply of autocratic governments and increasing the supply of democracies should help to increase the supply of peace. Third, the lack of self-determination for ethnic and religious communities is a source of conflict. The unjust denial of political and cultural rights for minority nations (including the right to create sovereign states) fosters internal conflicts that can spread throughout the region. On the other hand, liberals also emphasize factors that unite people rather than divide them according to parochial national loyalties. For instance, liberals have pushed for integration of smaller communities into larger ones—as in the case of the European Union—as a way of fostering peace. Fourth, the pursuit of “power politics”—the building of war machines, arms races, and the use of foreign policies that emphasize coercion—can also lead to war. Thus, militaries should be subjected to constraints, states should forgo the establishment of military-industrial complexes, leaders should refrain from coercive bargaining and other attempts to bully their rivals, and arms races should be reined in. Fifth, countries that trade with each other are unlikely to go to war with each other. Trade draws states into a web of economic interdependence that makes war both costly and unnecessary. Moreover, economic cooperation fosters habits of cooperation that can spill over onto other issues. On the other hand, protectionist policies that inhibit trade and penalize potential commercial partners limit the potential for cooperation and sow the seeds of conflict between states. Trade wars can lead to real wars. Thus, states should pursue free enterprise capitalism at home and free trade policies abroad. This short summary of liberal thought will have to do for now. Let us proceed to what many see as the key to the liberal view on war: the connection between democracy (or lack of it) and international conflict.

172

Chapter 5

THE ROLE OF REGIME TYPE (1): DEMOCRACY AND THE MONADIC PEACE In a world of simple explanations (in which we, unfortunately, do not live) states can be divided into two categories—good, peaceful states and bad, aggressive states. Who are the good guys and who are the bad guys? The most often heard argument is that democratic states are peaceful and authoritarian states are aggressive. This is, as we have seen, a liberal theory based on the assumption that humankind is basically peaceful, rational, and cooperative and that this desire for peace will be reflected in the policies of governments if they are democratic. Democratic governments represent the wishes of their peaceful citizens—or at least of the peaceful majority. Permitting prospective soldiers and their families the chance to share in the decision to go to war should greatly reduce its likelihood. Few people will vote for an unnecessary war that might mean the destruction of their property, the reduction of their standard of living, and the death of themselves or their loved ones. Moreover, institutional procedures constrain the ability of executive branch officials to make war without wider consultation and consent. In addition, there is presumably a cultural dimension inherent in the identity of democratic states to the extent that both leaders and the public generally see wars as abhorrent—unless they are fought for reasons grounded in strong moral principles. A narrow (and, I think, correct) reading of this theory deals with war initiation, but not war participation. The more democratic a government is, the less likely it will be to initiate a war. This is a form of the “monadic hypothesis,” which predicts the behavior of a single state (a monad) based on that state’s political system. (We will look at a second type of hypothesis about democracies—“the dyadic hypothesis” that democracies don’t fight each other—in a later chapter.) Strictly speaking, the monadic hypothesis does not predict that democracies should be less likely to be participants in war, only that they should be less likely to initiate war. They may be participants in war, but this may not be due to their own choice. Indeed, it might be that democracies are likely to be the targets of other states’ aggression if the democracies’ unwillingness to use force undermines their ability to deter the aggression of others. In nondemocratic countries, it is assumed that the leadership is bound neither by the popular will nor by constitutional restraints on central power. Leaders of autocratic governments will therefore be more likely to initiate hostilities. The existence of autocracies therefore endangers world peace. In short, wars occur because some states are bad. If the view of the liberals is correct, how is war prevented? Obviously, the long-term solution will have to be the creation of a world of democratic states. The more important question is how is this to be achieved? The answer comes in two varieties—active and passive. 4 American isolationists have been typical proponents of the passive policy. They argued that the United States should lead other countries along the enlightened path to democracy and peace through example. The United States should function as a beacon—a “shining city on the hill”—to all those who would follow. Since democracy is quite obviously the best system of government, reason dictates that eventually most states will become democratic and peace will be the result. Direct intervention is therefore unnecessary. Interventionists are proponents of the active policy. They believe that a state cannot be a force for peace by standing by while bad states attack good ones, eventually even one’s own

The State Level of Analysis, Part I

173

state might be attacked! Because of this risk, it might become necessary for democratic states to intervene in the affairs of authoritarian states to make them more democratic. As Edmund Burke, one of the founders of modern conservatism, remarked, “all that is necessary for the triumph of evil is that all good men do nothing.” 5 And Woodrow Wilson (a liberal) sent the Marines to Mexico to “teach them to elect good governments” and sent American forces to Europe to “fight for democracy,” while Presidents Reagan and George H. W. Bush (both conservatives) sent troops to Grenada and Panama to “restore” democracy. And President George W. Bush (a neoconservative) launched an invasion of Iraq that was, at least partly and retrospectively, designed to bring democratic government to the Arab world. The problem with this approach should be readily apparent. War is seen as the way to create peace. Furthermore, wars waged for such universal principles tend to become unlimited. As A. J. P. Taylor stated, “Bismarck fought ‘necessary’ wars and killed thousands; the idealists of the twentieth century fought ‘just’ wars and killed millions.” 6 War and Democracy: Empirical Evidence How does the monadic hypothesis fare against the world of historical fact? Is there any evidence that the liberal theory of war is valid? Are democracies more peaceful than autocracies? If ever there was a good theory mugged by a gang of facts, I suppose this is the one. However plausible it may seem, and however much our democratic values predispose us to root for this theory, there seems to be little evidence to support it. Quincy Wright was one of the first political scientists to systematically study war. In his mammoth work A Study of War, he built elaborate theoretical explanations for the proposition that democracies should be more peaceful than absolutist autocracies. The power of their central governments is limited by constitutional restrictions; dispersed by the principles of federalism and separation of power; restrained by widespread political participation, procedural deliberations, and freedom of criticism; and ultimately subjected to majority rule. Political accountability requires that government elites in democracies maintain public support for their continuance in office. Unpopular policies (such as war) will be avoided out of fear of electoral punishment. When Wright looked at the historical record, however, he was forced to conclude that democracies seem to be highly involved in war—and not just because they are forced to defend themselves against the attack by others. There just didn’t seem to be much difference in the war activity of different kinds of political systems. 7 The work of Small and Singer on the Correlates of War (COW) project yielded similar results. Investigating a sample of ninety-three wars occurring between 1861 and 1965, they discovered no substantial difference between democracies and nondemocracies in terms of either their war participation or war initiation. 8 The heavy war involvement of democracies was not due to their selection as defenseless victims by others; they seemed to be just as involved in the aggressive use of force as nondemocratic states. In the nineteen wars democracies participated in, they initiated (or sided with the initiator in) 58 percent of them. While democracies did participate in fewer wars than nondemocracies, they also constituted a relatively small group of states for most of that period. Thus, based on Small and Singer’s data, democracies would seem to have been somewhat overrepresented as war participants and war initiators.

174

Chapter 5

Russett and Monsen also found that the type of political system a state possesses has little effect on its war proneness; size seems to be a much more important predictor. Large polyarchies (representative democracies) have been involved in more wars than either small polyarchies or nonpolyarchies of any size. 9 The historical record indicates that, contrary to the logic of the monadic theory, democracies have been responsible for many types of war. As Michael Doyle points out, liberal or democratic states have been initiators of expansionist colonial or imperial wars, wars against weak liberal regimes, and even wars of annihilation against native populations. 10 As we have seen, however, nothing is completely cut-and-dried in conflict research. Contradictory evidence began to creep in around the edges almost from the beginning. Michael Haas grouped states into three categories: constitutional, authoritarian, and totalitarian. When these three types of government were run against foreign conflict data from the late 1950s, Haas found that authoritarian regimes exhibited the most foreign conflict behavior, constitutional governments the least, and totalitarian governments fell in between. None of the statistical links were very strong, however. 11 Moreover, Haas’s study suffered from certain limitations. First, it examined a very limited period of time (1955–60); second, it didn’t directly look at war involvement but at general conflict behavior (which included nonwar measures of conflict such as diplomatic protests, sanctions, etc.); and third, its statistical results were meager. Two other studies, by Wilkenfeld and Zinnes and by Salmore and Hermann, also indicated a difference in the conflict participation activities of democracies and autocracies. 12 As in the study by Haas, the statistical results were not very strong and war was lumped together with other kinds of conflict behavior. The findings of these three studies were consistent, but they were also consistently weak. Thus, the first wave of research on this topic generally failed to find a strong relationship between democratic government and peace. Most analysts sided with the original Small and Singer consensus that democracies were not noticeably more peaceful than other types of governments. Nevertheless, the ideological attractiveness of a theory linking democracy and peace propelled further research. The debate was revived in the 1980s when a study by R. J. Rummel found substantial support for the theory. 13 For the years 1976–80, Rummel found that the more libertarian or free the state, the less its foreign violence; the less free a state, the more its violence. This appeared to be true whether the dependent variable was a scale of foreign conflict intensity or war data only. When the traditional operationalization of free states was expanded from the presence of political rights and civil liberties to include economic freedom (free markets) as well, the relationship appeared even stronger. Other scholars begged to differ, noting several problems with Rummel’s research. In an analysis of the 1816–1980 period, Steve Chan, using COW data, expanded and retested the older Small and Singer study; once again no strong relationship between government type and war involvement was found. Nor were democracies less inclined to initiate wars or fight on the side of initiators. 14 Most important, however, Chan discovered a significant difference in war behavior after 1973. From 1816 to 1972 he found a positive relationship between democracy and war; democracy and war were highly correlated—just the opposite from what the theory predicts. But this relationship was reversed in the 1973–80 period, roughly the same period

The State Level of Analysis, Part I

175

covered by Rummel. Erich Weede’s investigation of the 1960–80 period produced somewhat similar results. Using three measures of war and two different measures of democracy, Weede found no strong statistical relationship between frequency of war involvement and the type of government; however, the 1960–74 period was different from the 1975–80 period. These results both suggest that Rummel’s use of the 1976–80 period may have been responsible for his results. The more recent time period appears to be atypical when compared to previous periods. 15 Morgan and Campbell’s analysis of militarized disputes between 1816 and 1976 adds further evidence. (Militarized disputes are conflicts that involve either the display of force, the threat of force, the use of force, or all-out war; they’re often referred to simply as MIDs.) Instead of focusing on democracies per se, they emphasize the presence (or absence) of structural constraints on the ability of a government to make decisions for war. Such political constraints include the accountability of the leader to a selection process, the presence of political competition, and institutional power sharing on decisions of war and peace. They find that the probability of war does decrease slightly as the constraints on government decision making increase, but the relationship is hardly significant. They conclude that structural constraints are not an important determinant of whether militarized disputes escalate to war. 16 Further studies also seemed to confirm the original Small-Singer consensus that the monadic proposition was invalid. Regardless of whether the dependent variable is war, intervention, or militarized dispute involvement, democracies are not particularly less conflict prone than nondemocracies. 17 For instance, an exhaustive study of militarized disputes and war from 1816 to 1976 by Zeev Maoz and Nasrin Abdolali found that when research focused on the activity of states as the unit of analysis, regime type was not a good predictor of the extent of a state’s involvement in conflict. They conclude that with few exceptions, free states “are neither more nor less conflict prone” than nonfree states. 18 And while one would expect that as a country’s government became more democratic over time, its involvement in conflict would decrease, this was not the case. Regime change seemed to have no significant impact. When Maoz and Abdolali look at the conflict behavior of pairs of states (dyads), they find only mixed support for the contention that democratic states are less likely to initiate conflict than other regimes. They argue that this is because, while democracies are unlikely to initiate disputes against other democracies, they are disproportionately likely to initiate disputes against autocracies and anocracies (mixed regimes). Autocracies are also unlikely to initiate disputes against other autocracies, but they disproportionately target states with different regime types. The only good news appears to be that while democracies were more conflict prone than expected (based on the number of democracies in the system) in the nineteenth century, this has been reversed in the twentieth century. In the later period, dyads with at least one democracy are underrepresented in both militarized disputes and wars. 19 In recent decades, intrepid research has been able to pump a certain amount of life back into the monadic hypothesis. For instance, Morgan and Schwebach studied dyadic disputes that did or did not escalate to war for the 1816–1976 period. They found that democracies or states with constraints on executive power were less likely to escalate disputes to war, regardless of the nature of the opponent’s regime type. A higher percent of disputes between nondemocracies escalate to war than those between democracies and nondemocracies. They argue

176

Chapter 5

that democracies would seem to fight less often than other types of states, but that this relationship is “exceedingly weak.” 20 The evidence for the general peaceful effects of democracy seem strongest for the last four decades of the twentieth century. For instance, Benoit uses more sophisticated techniques to perform a reappraisal of Weed’s earlier analysis of the 1960–80 period. Using two different measures of war involvement and two different measures of democracy, he determines that regardless of which measures are used, democracies fought fewer wars on average than nondemocracies. Specifically, democracies actually fought on average one less war over the twenty-year period than nondemocracies, a difference which he claims is both substantively and statistically significant. 21 One possible explanation for the declining war involvement of democracies is that when democracies were few (in the nineteenth and early twentieth centuries) they were most likely surrounded geographically by autocratic states, while in the post–WWII era, many democracies have other democracies as their neighbors. And, as we shall see, states are overwhelmingly inclined to fight their geographic neighbors, but democratic states are highly unlikely to fight each other. Gleditsch and Hegre make a similar assessment. Looking at COW data on wars and militarized interstate disputes (MIDs) as well as the Uppsala data on armed conflict (which included small conflicts with a minimum of only twenty-five dead in a year) they recognize that the war behavior of democratic states has been problematic. The COW data from 1816 to 1994 indicate that of the wars in which democracies were participants from the start, 73 percent were initiated by democracies. They have several responses to this. First, deciding which side actually initiates the war is notoriously difficult and the COW coding may be suspect on this issue. Second, in the post-1945 period, democracies would appear to initiate violence rarely, except in situations of protracted conflicts and long-running disputes. Finally, the six most violent interstate wars since 1816 (WWI, WWII, the Sino-Japanese War, the Korean War, and the Vietnam War) were all initiated by nondemocratic states. 22 They concede that the frequency of war involvement and MID involvement of democracies is not noticeably different from nondemocracies for most of the time period, but this changes during the Cold War era. Moreover, they argue that the total amount of war involvement by democracies is statistically inflated because democracies tend to join in alliances with each other. 23 A study by Sven Chojnacki of the years between 1946 and 2002 agrees that democracies are less involved in interstate wars in the post–WWII era, but that a more complete picture would have to include extrastate wars (between a state and a nonstate group outside the state’s territory), intrastate wars (between a government and an armed opposition group), substate wars (between two nonstate groups), as well as outside military interventions into any of the four types of wars. 24 His data show that democracies were involved in thirteen of the twentythree interstate wars in the period, though they initiated only four of these. When one looks at the total number of war years instead of simply the number of wars, democratic states were involved in interstate wars far less often than other regime types, but were almost equally involved in extrastate wars as autocracies and anocracies. (This figure is largely due to high French and British involvement in wars of decolonization in Asia and Africa.) But most important, Chojnacki finds that democracies intervened in ongoing conflicts somewhat more frequently than other types of regimes—especially in intrastate wars. 25 However, a relatively

The State Level of Analysis, Part I

177

small handful of democracies account for most of this activity, and (as we would expect) these are democratic great powers—the United States, France, and Britain. Jack Levy’s investigation of four cases—Israel’s 1956 Sinai War, Israel’s 1981 surgical strike on Iraq’s Osirak reactor, the U.S. intervention in the 1990–91 Persian Gulf War, and the U.S. invasion of Iraq in 2003—indicates rather clearly that democratic states engage in preventive war or the preventive use of military force against states perceived to pose a long-term threat due to an adverse shift in the balance of capabilities. 26 All four actions garnered substantial domestic support in the initiating democracies. Contrary to the expectations of liberal theory, American officials came to understand in 1991 and in 2003 that the best way to sell a war against Iraq to a potentially skeptical public was to frame the necessity of war in terms of preventive war logic—the U.S. might need to fight Iraq now to reduce a potential threat in the future if Iraq was to gain nuclear weapons. This is important because some theorists have argued that democratic institutions should especially preclude resort to this kind of war. Informed publics should not support such wars since they are wars of choice rather than necessity and therefore fail to meet ethical and legal standards for “just wars.” 27 A fifth case for Levy, the near war between the Clinton administration and North Korea in 1994, was averted not because the American public viewed a preventive strike against North Korea as immoral or contrary to American democratic values but because the war was likely to be very costly and extremely risky. Evidence has gradually begun to accumulate with regard to conflict short of war that democracies, especially since WWII, are less conflict-prone than autocracies. Using data on international crisis behavior (ICB) and democracy (polity II) David Rousseau and his colleagues discover that compared with nondemocratic states, democracies may be less likely to initiate crises, regardless of the regime type of the opponent. This aversion to crisis initiation may be related to the status quo orientation of democratic states. On the other hand, while democracies are less likely to initiate force against other democracies in a crisis, they are only marginally less likely to initiate force against other types of regimes. The dyadic proposition—democracies are unlikely to fight other democracies—seems to hold once democratic states are involved in a crisis. Under these circumstances, democracies are much more likely to escalate against nondemocracies than against other democracies. Overall the results were much more supportive of dyadic hypotheses than monadic hypotheses. 28 Other studies confirm that once involved in a crisis, democracies engage in escalation and violence, but tend to limit the severity of violence against other democracies. 29 Paul Huth looks at a different facet of the puzzle. His study of territorial disputes between 1950 and 1990 indicates that democracies are less likely to escalate territorial disputes regardless of the regime type of the adversary. 30 Huth finds that of forty-one territorial disputes between 1950 and 1990 that escalated to high levels of violence, only one (India v. Portugal over Goa) involved a challenger with a mature democracy. And of all the disputes in which democracies were the challengers, only two (10 percent) escalated to high levels. 31 Perhaps the most vigorous defense of the monadic proposition has been undertaken by Bruce Russett and John Oneal, who studied data on MIDs for “politically relevant” pairs of states from 1885 to 1992. 32 Their analysis goes something like this. Given the fact that any two democracies are very unlikely to fight each other, two democratic states are also logically

178

Chapter 5

much less likely to be involved in militarized disputes than either mixed dyads (democracyautocracy) or dual autocratic dyads. Now, suppose we have a mixed dyad containing a democracy and an autocracy. This mixed pair has a certain likelihood of engaging in a militarized dispute. If you then replace the autocracy in the pair with another democracy (creating a dual democratic pair), the probability of conflict drops from 73 percent above the baseline probability for all pairs to 41 percent below the baseline. On the other hand, if you start with an autocratic-autocratic dyad and change one of the states to a democracy (creating a mixed pair), the risk of a militarized dispute remains virtually unchanged. When a state becomes more democratic, its chance of conflict with another democracy drops dramatically, but its risk of conflict with an autocratic state remains pretty much the same. Thus, as a state becomes democratic, the risk of conflict will either go down (with other democracies) or remain the same (with autocracies), but will not increase. One cannot say this of a situation in which a democracy becomes autocratic. Russett and Oneal find no evidence of an “autocratic peace.” Autocracies seem to fight each other about as much as mixed dyads. 33 As a result, they conclude that “on average, democracies, as individual states, are more peaceful than autocracies.” 34 Russett and Oneal make it clear that the monadic effect of democracy is not as strong as the dyadic effect. 35 But notice that the logic put forward for the monadic proposition has to do with the effect of democracy on dyads, not the effect of democracy on the individual state per se. The reputed peacefulness of democracies “on average” is due almost entirely to their reduced conflict with each other. 36 Nevertheless, the previous consensus that the monadic hypothesis is invalid is not as clear as it once was. A Digression: Are Democracies Better at War? A new front has been opened on the monadic hypothesis in recent years. There appears to be growing evidence that democracies are much more likely to win the wars they participate in than are nondemocratic states. David Lake’s review of wars from 1816 to 1988 suggested that of the twenty-six wars fought between democracies and autocracies, democratic states have won 81 percent. 37 This finding was especially strong for wars that are initiated by democracies. A series of analyses by Reiter and Stam for the period 1816–1990 helped to solidify the argument. They report that democracies won 93 percent of wars they initiated, while autocracies won only 60 percent and mixed regimes only 58 percent. (Overall, initiators are more likely to win wars than their targets: Initiators—regardless of regime type—win about 65 percent of wars.) 38 According to Reiter and Stam, not only do democracies win most of the wars they initiate, they win almost two-thirds of the wars in which they are the targets. Whether they are initiators or targets, states are more likely to win as they become more democratic. Moreover, democracies tend to suffer fewer casualties in wars than other types of regimes. 39 All of this suggests that democracies are relatively careful about the wars they fight: they seem to fight those wars they can win and win at relatively low cost. Several arguments are put forward to explain these phenomena. First, the accountability of democratic leaders to their voting publics means that leaders understand the high political costs of initiating wars they can’t win (or win only at great cost) and are thus cautious with respect to war initiation. Second, the robust “marketplace of ideas” in democracies—particularly the presence of oppo-

The State Level of Analysis, Part I

179

sition political parties, interest groups, and independent media—means that leaders will have access to high-quality information about the choices of war and peace. Third, once involved in wars, democratic armies perform better because a democratic culture and strong public support for the political system produce soldiers with greater skills, individual initiative, and loyalty. As you can imagine, the “Democracies Are Smarter and Tougher” argument set off an avalanche of criticism. 40 Critics argue that there is little evidence that democratic leaders actually suffer political punishment for leading their states to defeats in crises or wars. They also argue that autocratic leaders are just as aware of the political costs of losing wars and may even suffer more severe punishments than just loss of office. Critics also point out that the marketplace of ideas certainly did not work as hypothesized to instill political caution in the run-up to the Bush administration’s war in Iraq. The democratic political process may actually operate inversely so that the fear of political punishment may actually push leaders toward war rather than away from it. For instance, it appears that the democratic process in the U.S. added to President Johnson’s incentives to undertake a risky escalation of the war in Vietnam. Critics also point out that democracies tend to be wealthier and more technologically advanced than autocracies and are more likely to come to each other’s defense; these factor are more likely to be responsible for the high success rate rather than democracy itself. There have also been disputes about the statistical analyses of Reiter and Stam. Alexander Downes reanalyzed Reiter and Stam’s data with two major changes: he added wars that ended in draws to the categories of wins and losses, since electoral punishment should apply to draws as well as to losses. He also distinguished states that join the war later from states that were targets. He finds that when these changes are made, democracies are not significantly more likely to win wars, whether they are initiators or targets. 41 The jury is still out on whether democracies are smarter and tougher than autocracies. Why Aren’t Democracies Peaceful? Why is it that democracies are not as peaceful as theorists suggest they should be? A number of possibilities exist. (1) It is possible that public opinion and legislative majorities in democracies may actually be supportive of war—especially if the opponent is seen as a nondemocratic regime. John Owen concludes from his case studies of American crises that “democratic structures were nearly as likely to drive states to war as to restrain them from it. Cabinets, legislatures, and publics were often more belligerent than the government heads they were supposed to constrain.” 42 (2) Even if the regime is in the hands of liberal leaders, the possibility exists that they will pursue war on behalf of liberal causes such as the promotion of democracy abroad (which, of course, may have strong public backing). (3) Inherent in the nature of pluralist democratic governments is the possibility that “illiberal” leaders who support war will come to power. (4) It is also possible that, contrary to democratic theory, executives may not fear electoral punishment as a consequence of war involvement. And while there may be no electoral punishment in autocracies, autocratic leaders may fear the consequences of a war gone badly even more than leaders in democracies; in autocratic states, the punishment may be exile or death rather than mere electoral defeat. (5) It is possible that executive branches can find ways to get around legislative constraints. And constraints may be

180

Chapter 5

weakened by the fact that legislators may fear electoral punishment themselves if they fail to back the executive leadership’s attempts to improve national security by taking strong offensive action. (6) The peaceful effects of democratic structure may only apply to mature, stable democracies and not to marginal democracies (of which there are quite a few) or states in the midst of a transition to democracy. Logically then, there would appear to be plenty of loopholes to explain the warlike behavior of democracies. THE ROLE OF REGIME TYPE (2): DEMOCRATIZATION Not all democracies are equally democratic. (Remember that democracy is a continuous variable; it exists in degrees.) Transitional democracies may have relatively little of the thing we call democracy, and many countries in the early stages of democracy never make it all the way to becoming stable, mature democracies. The process of becoming democratic is a fairly messy one, fraught with dangers—a lesson that certainly resonates with current-day Iraqis and Afghans. This is in fact the argument of Edward Mansfield and Jack Snyder, who propose that the process of building democracy—democratization—may actually increase the likelihood of war. 43 This thesis is not really a direct attack on the monadic hypothesis; it’s sort of a “friendly amendment,” suggesting only that a particular subcategory of democratic states is dangerous. Why should states undergoing democratization be prone to war? Mansfield and Snyder argue that when autocracies begin to democratize, political participation increases faster than the growth of democratic institutions—such as an independent judiciary and the rule of law, organized and competitive political parties, fair elections, and a professional news media. A large number of individuals and groups are now participants in the political process thanks to universal suffrage and elections, but the institutions and rules that exist to resolve conflicts between these political actors are weak or nonexistent. When autocratic institutions lose their legitimacy and effectiveness before new democratic ones can fully replace them, this creates a “free-for-all” situation in which a wide variety of groups vie with each other for political power. In addition, the political power of old elites is threatened by the potential transition: “they fear that democratization will end their ability to use state power to serve their own interests, and they fear being punished for crimes of repression and patronage that were business-as-usual in the autocratic system.” 44 If existing political institutions are too weak to guarantee declining elites a “soft landing,” they will use their own resources to recruit allies and manipulate the political process to ensure their survival. In many cases the old elite is, at its core, a military elite, and many of the interests threatened by democratization are military in nature. The newly empowered masses may be unwilling to bear the costs of subsidizing these old elites. In this situation, “what constitutes power and who will have it are up for grabs.” 45 Both endangered, downwardly mobile old elites and rising new political elites recognize the need to mobilize the public on their behalf. And the easiest path to this is through the use of nationalistic appeals and propaganda that demonize certain out-groups as enemies and rally mass support behind an attempt to protect the country from external perils. And at the same time the identification of national threats provides a justification for repression of political opponents. All sides tend to use the same strategy; there may even be a tendency for political elites to

The State Level of Analysis, Part I

181

“outbid” each other in making populist and nationalistic appeals. And certain interest groups—military elites, economic protectionists, and dominant ethnic groups—are especially attracted to this strategy. An additional dynamic is caused by the absence of strong democratic institutions that can mediate between various interests. Instead, certain narrow, concentrated interest groups— especially militarist, protectionist, and imperialist groups—engage in logrolling to trade favors and satisfy their parochial self-interests. Strong democratic institutions that might otherwise look out for the more diffuse interests of the state as a whole are unable to counteract the elite logrolling. As previously noted, the classic example is pre-WWI Germany, where the alliance of “iron and rye” meant that the military, the industrialists, and the feudal agricultural Junkers of Prussia each supported each other’s interests without regard for the general effect on Germany’s global and regional situation. For Mansfield and Snyder the type of state with the greatest risk of war is one where a kind of arrested development takes place. This is a mixed regime (an anocracy), partially authoritarian and partially democratic, that has made an incomplete transition from autocracy to democracy. The danger is not permanent. If the state’s transition stabilizes into a steady-state arrested development, the danger recedes. It is the transition itself, not the half-way situation between autocracy and democracy, that is the cause of the problem. And the probability for war is most likely to affect those transitional regimes that have weak state institutions at the onset of the transition and have elites whose political fortunes are in flux. How does any of this lead to war? Mansfield and Snyder distinguish two pathways leading from democratization to war. 46 In the first pathway, incomplete democratizers with strong administrative institutions and strong militaries but with weak or biased representative institutions, pursue nationalistic policies, including war, as a way of competing for public support. War is consciously chosen as a strategy to enhance or preserve the political standing of a governing elite. This is much like the diversionary war argument, which we will examine in the next chapter. Both posit that political elites who find their rule endangered by political opposition consciously gamble that using aggressive foreign policies will unite the public behind their rule in the face of a supposed threat from foreign states. 47 Mansfield and Snyder simply say this is more likely if states are in the throes of democratization. The second path leads to war through an indirect process. As part of a nationalistic strategy, governments are likely to pursue aggressive, adventuristic foreign policies designed to curry favor with the public. At home, these policies mean that the power of the military grows, and the state may even become a “garrison state.” Abroad, these policies may provoke others and embroil the transitional country in militarized disputes and crises with its neighbors. Indeed, several studies indicate that democratization promotes MIDs and conflicts short of war. 48 These disputes may then escalate to war. In such cases, democratizing states may be instigators of war, even though they do not necessarily fire the first shots. In sum, war may be either a direct result of populist attempts to curry favor with the public, or it may be merely the indirect result of risky foreign policy actions. Democratizing states may be the initiators or they may be the targets of military actions initiated by others, although Mansfield and Snyder argue that the democratizing state is most likely to be the initiator.

182

Chapter 5

They argue that virtually all the great powers were involved in adventuristic or populist wars during the early stages of their experiments with democracy. France fought several wars in the wake of the 1789 revolution and its experiment with Napoleonic republicanism; Britain fought the Crimean War in the aftermath of its expansion of suffrage to the middle class in 1832; Germany fought several wars in the late nineteenth and early twentieth centuries as it pursued an on-again, off-again path toward democracy; Japan’s militarist policies on the Asian mainland followed hard on the heels of its first attempt at democracy in the 1920s; and the populist Mexican-American War followed the movement toward mass democracy in the U.S. associated with the presidency of Andrew Jackson. 49 Democratization and War: The Empirical Evidence Over the years, Mansfield and Snyder have continued to refine their analysis, both with statistical studies and with qualitative case studies of democratizing states involved in war. Mansfield and Snyder use various measures of democracy and regime change (transition) to examine the 1816–1992 time period. Regardless of the measures used, they discover that incomplete democratic transitions (from autocracy to anocracy) increase the probability of war involvement, but only if the state’s institutions are weak and fragmented. This is the most dangerous set of political conditions found in the study. These states are between eight and ten times more likely to be involved in an interstate war than states without transition. Transitions culminating in a coherent democracy do not generally produce war. Contrary to their earlier findings, transitions toward autocracy (autocratization) apparently do little to increase the risk of war. Only a particular kind of transition substantially increases the risk of war involvement. 50 Moreover, their results suggest that democratizing states are more likely to be the initiators of the wars they are involved in rather than the targets, and that democratizing states are more likely to initiate wars than are stable regimes or autocratizing states. 51 Their dyadic analysis of the data shows that the probability of interstate war increases considerably if either state in a given pair is undergoing an incomplete transition toward democracy when its political institutions are weak. The risk of war is greatest when the least democratic state in the pair is the one making the transition. Their database (from 1816 to 1992) identifies several dozen wars connected to democratic transitions. Since they believe democratizers are the most likely initiators in these conflicts, they identify ten countries who were involved in wars (some were involved in more than one) during the early stages of their transition. These countries and their wars are listed here: • France under Napoleon III (Roman Republic War, 1849; Crimean War, 1854; Piedmont War against Austria, 1859; Mexican intervention, 1861; Franco-Prussian War, 1870). Of course, the prototype case (which falls outside their statistical data set) was the French Revolution and the wars associated with this period of French history. • Prussia/Germany (First and Second Schleswig-Holstein Wars against Denmark, 1848–49 and 1862; Austro-Prussian War, 1866; Franco-Prussian War, 1870. The last is technically not a war of democratization because it occurs over a decade after the democratizing changes of Wilhelm’s constitution.)

The State Level of Analysis, Part I

• • • • • •

183

Chile (1879 War of the Pacific against Peru and Bolivia) Guatemala (Central American Wars of 1885 and 1906) Serbia (Balkan Wars of 1877, 1885, 1912, 1913, WWI) Thailand (1940–41 Franco-Thai War) Iraq (Palestine War, 1948) Argentina (Falklands War, 1982)

Two countries whose democratization process ended in complete democracy rather than arrested development also initiated wars during the process of democratization: • United States (Mexican War of 1846) • Turkey (Invasion of Cyprus 1974) Several wars that have occurred in the period after 1992 also appear to fit with their theory: • • • • •

Armenia (Nagorno-Karabakh War, 1992– ) Peru (1995 War with Ecuador) Eritrea (Ethiopia-Eritrea War, 1998–2000) Pakistan (Kargil War with India 1999) Serbia (Kosovo War, 1999)

The authors are careful to point out that democratization is only one of many causes of war. They find that 7 percent of all wars since 1816 are associated with an incomplete democratic transition. Thus democratization “is only one of many causes of war, but it is a potent one.” 52 The relative scarcity of “wars of democratization” is to be expected: Democratization is a relatively rare event in and of itself and is thus not likely to be a major cause of war in general. Mansfield and Snyder freely admit that they are not trying to create a general theory of war with broad applicability. But they believe that for some wars an adequate explanation cannot be fashioned without an understanding of the effects of democratization. Mansfield and Snyder’s findings, especially those of their original study, have been subject to considerable criticism. Studies by other researchers have failed to replicate Mansfield and Snyder’s findings about the dangers of democratization. Some have found no association at all between democratization and war; 53 others have found that democratization may actually reduce the probability that the state will initiate a militarized dispute. 54 Let’s look at several different empirical studies to try to nail down the validity of the democratization argument. Using somewhat different data and methodology, Ward and Gleditsch focus on the nature of political change in countries undergoing transitions. 55 Their results indicate that in general the process of democratizing actually reduces the probability of war involvement, regardless of whether the changes are small or large. Changes that lead to increased constraints on the power of the executive branch are especially helpful in reducing the risk of war involvement. However, they find that certain kinds of transitions are problematic. While smooth transitions toward democracy have a peaceful effect, rocky transitions that include setbacks and reversals along the way appear to increase the probability of war involvement. They also find that transitions toward autocracy increase the risks of war in-

184

Chapter 5

volvement. Thus, like Mansfield and Snyder they find that not all democratic transitions produce higher levels of risk, only some. But they differ somewhat on the type of transitions that are problematic. As part of their huge comparative assessment of sixteen hypotheses about war at three levels of analysis, Bennett and Stam investigate the effects of democratization in a way that draws on Ward and Gleditsch’s analysis. Their conclusion appears strongly counter to Mansfield and Snyder’s hypothesis. They find that democratization generally lowers the risk of disputes short of war (MIDs). In dyads composed of two democratizing states, the risk of MIDs escalating to war is reduced by 20–30 percent, compared with dyads that are becoming neither more democratic nor more autocratic. And doubly democratizing dyads have a 40–50 percent lower risk of MIDs than mutually autocratizing dyads. Moreover, they find that abrupt transitions, compared with more gradual changes, actually reduce the chances of conflict at all levels. Contrary to the expectations of many analysts, it does not appear that sudden regime transformations lead to transition states being the target of predatory actions by others or to diversionary wars against others. 56 In Triangulating Peace, Russett and Oneal look at the effect of major shifts in regime change (from autocratic to democratic or vice versa) on the probability that dyads will become involved in MIDs. They find that neither shifts toward or away from democracy have any impact on the probability of dispute involvement. Their data indicate that the level of democracy attained by members of a dyad is much more important than whether the level was the result of a recent shift. They argue that most instances of democratization and MIDs involved states that have not yet achieved complete democracy. Once they reach such an end result, their probability of conflict is sharply reduced. In general, while democratizing states are involved in disputes as Mansfield and Snyder say, this is offset by the many instances in which such states have remained peaceful. 57 Narang and Nelson reanalyze Mansfield and Snyder’s data analysis and their case studies and come to some startling conclusions. 58 They argue that the central issue is not whether incomplete democratizers with weak institutions participate in wars but whether they initiate wars. Their analysis concludes that for the entire 1816–1992 period there is only one such instance: Chile’s initiation of the 1879 War in the Pacific. And in this case Chile is defined as an incomplete democratizer by only one of the four indicators of regime change used by Mansfield and Snyder. If one uses a composite measure of democratization, there are zero cases of incomplete democratizers initiating wars in the time period. Moreover, Mansfield and Snyder’s data indicate that there are only two other democratizing countries with weak institutions even involved in wars, Peru in 1841 and the Ottoman Empire in 1877, 1911, 1912, and 1913. And after World War I, there are no instances of incomplete democratizers even participating in an external war, even though about 80 percent of countries going through incomplete democratization have done so in the period between 1945 and 1992. Narang and Nelson contend that the statistical relationship that Mansfield and Snyder find between democratization and war is due to the disproportionate effect of one state’s war experience—the Ottoman empire. In each of the wars—1877, 1911, 1912, and 1913—Ottoman Turkey was the victim of others who repeatedly (and successfully) detached portions of the empire from a state undergoing political upheavals at home. None of these wars were

The State Level of Analysis, Part I

185

initiated by the Ottoman empire for diversionary reasons. In an experiment, Narang and Nelson treat the three wars from 1911 to 1913 as a single war (a reasonable assumption); consequently, the statistical relationship between incomplete democratization and war disappears. They also argue that Mansfield and Snyder’s paradigmatic case—Wilhelmine Germany’s initiation of World War I—actually constitutes a poor fit with the theory’s expectations. Germany did go through a transitional process in the direction of democracy, but this occurred long before the outbreak of World War I. They argue that by any measure of regime transition, Germany had been stable for almost fifty years by 1914. Finally, they argue that because Mansfield and Snyder select their cases on the dependent variable (only those cases in which war does break out), they fail to report on an absolutely huge number of instances in which states make incomplete transitions to democracy and yet experience no external war at all. One criticism of the democratization theory is that the real culprit might just be regime instability of any kind. It may simply be that any type of political change and instability increases the likelihood of war. Instability may create weakness and invite attack, or it may create diversionary motives for war. The explanation does not depend on the process of democratization, merely on the process of change. Stable states—whether democratic or authoritarian or mixed regimes—tend to be more peaceful than unstable states. 59 A second alternative explanation of Mansfield and Snyder’s findings is that states undergoing a transition to democracy will experience conflict if its neighbors are nondemocratic. What is important is that the democratic transition increases the political distance between the state and its neighbors. If the democratizing state is surrounded by autocratic states, it is probably not reasonable to expect that its democratization will have a pacifying effect. Oneal and Russett believe that newly democratizing states are likely to experience conflict only if their neighbors are nondemocratic. The explanation is dyadic rather than monadic: It depends on the changing relationship of the two states, not on the nature of a particular state as being involved in a transition. 60 THE ROLE OF REGIME TYPE (3): ROGUE STATES Foreign affairs junkies have become accustomed to a new term to describe states deemed especially dangerous to American interests: rogue states. With the collapse of the Soviet Union as the primary threat to U.S. interests, attention has shifted to this small group of states that lack great-power status or capabilities but nonetheless represent a serious threat. Since the 1980s five states have consistently been lumped into this category: Cuba, Iran, North Korea, and, until recently, Iraq and Libya. President George W. Bush’s famous 2002 State of the Union address specifically named Iran, Iraq, and North Korea as constituting an “axis of evil.” Moreover, the administration’s 2002 National Security Strategy document identified rogue states as the primary threat to U.S. national security in the post–Cold War era. The existence of dangerous rogue states was certainly behind the assertion by the “Bush Doctrine” that the United States might need to use preventive force against these threats to American interests, and the invasion of Iraq constituted the most concrete example of this American response to the presumed threat from rogue states.

186

Chapter 5

Rogue status generally derives from two actions: sponsorship of international terrorism and attempts to possess weapons of mass destruction (WMD), specifically chemical, biological, or nuclear weapons. The external behaviors of these states thus run counter to accepted standards of behavior in the international community. While the first two criteria are based on a government’s policies, other, more implicit criteria relate to the nature of the state. Rogue states are assumed to be developing countries whose regimes are nondemocratic. Their leaders were also frequently referred to as unstable, unpredictable, and undeterrable. Caprioli and Trumbore have attempted to empirically test whether there is any truth to the allegation that “rogue states” (as defined by American officials) are more dangerous or prone to aggressive actions in their external affairs than other states. 61 To identify rogue states, they accepted the five “rhetorical rogues” (Cuba, Iran, Iraq, Libya, and North Korea) consistently named by American policy makers and analysts, then added other states that met the objective criteria of being sponsors of international terrorism and pursuing the possession of WMD. This created a list of nineteen other states over the period from 1980 to 2001. The list includes Israel, Sudan, Syria, Serbia, Pakistan, South Yemen, the Soviet Union, and several Eastern Bloc communist states, among others. The next task was to test whether rogue states were especially dangerous and conflict prone. They hypothesized that rogues should be (1) more likely to be involved in new militarized interstate disputes (MIDs), (2) more likely to be the initiator of MIDs, and (3) more likely to use force first in the context of a MID. Their statistical analysis indicates rogue states as a group are no more likely to be involved in MIDs in any given year, no more likely to initiate MIDS, and no more likely to be the first side to use force when disputes turn violent than are nonrogue states. There were some differences among rogues, however. North Korea and Iraq were most likely to be involved in MIDs, and only Iraq was more likely to be the first to use force when involved in MIDs. Since rogues were no more likely to initiate MIDs than nonrogues, their high level of participation in international disputes would appear to be due to the fact that other states direct threatening behaviors at them, perhaps because they are perceived as rogues. In sum, only Iraq seems to be very representative of the stereotype of the dangerous rogue state, and this probably has much more to do with the idiosyncratic nature of Iraq’s political leadership than its membership in a particular category of states. Overall, the concept of rogue states does not seem to be very helpful in determining which groups of states are more likely to be involved in international conflict. The Economic System: Capitalism and Imperialism John Hobson, the British economist, was intrigued by the problematic nature of certain types of states. As a believer in democracy, he was forced to address why democratic countries, especially his own Britain, engaged in imperialism—aggressive expansion aimed at the establishment of foreign colonies. His answer was that although Britain’s political system was democratic, its economic system was capitalist. 62 The problem, then, was to be found not in the nature of the state’s political system but in the nature of its economic system. For Hobson, imperialism resulted from a “maladjustment” in the capitalist system. Capitalist countries seemed to suffer chronically from the unhealthy syndrome of overproduction, unequal distribution of economic wealth, underconsumption, surplus capital, and periodic

The State Level of Analysis, Part I

187

depression. The wheels of production turned apace, but the vast majority of people—their wages having been held down by the capitalists’ desire for increased profits—lacked the ability to purchase goods. Supply exceeded demand and surplus capital existed in the hands of the factory owners. These problems could be remedied in several ways. The economic elite could choose to redistribute wealth through higher wages, or the government could redistribute income through its taxing and spending policies, thereby raising the standard of living and creating greater consumer demand. Instead, the capitalists chose to invest their surplus capital abroad, carve out new foreign markets in which to sell their surplus production, make use of cheaper foreign labor to further reduce production costs, and dominate foreign lands in hope of securing necessary raw materials. As governments found it difficult not to help the elites in these economic endeavors, the result was the policy of imperialism. 63 Hobson produced data to show that imperialism had been largely unprofitable for the economy in general. It entailed high costs and high risks, while generating low returns. In fact, it was bad business policy. If this was the case, why was imperialism the order of the day? Because, said Hobson, it benefited a few powerful and well-placed groups: ship builders, export industries, international bankers, investors, and arms merchants—those elites whom President Eisenhower, twenty years after Hobson’s death, would describe as the “militaryindustrial complex.” These elites were able to induce the government to pursue colonial policies to benefit the few. The policy of imperialism was essentially “a vast system of outdoor relief for the upper classes.” 64 The elites in capitalist states turned democracy into a sham. Minority interests dominated the general will. All of this begs the question of what imperialism and colonialism have to do with interstate war. The answer is nothing—if the world is made up of relatively few, well-behaved capitalist states, each with plentiful opportunities for overseas investment and trade. Otherwise, quite a lot. In a world of many capitalist countries imperialism means economic competition between rival states. Each state strives to gain exclusive control over markets, raw materials, sources of cheap labor, naval bases, trade routes, and investment opportunities. At some point these can only be gained at the expense of other capitalist states. Economic conflict eventually leads to military conflict. If this is the case, the logical solution is to change the economic system. Hobson argued that with the advent of socialism (achieved through an evolutionary and parliamentary process), imperialism could be ended and wars avoided. And as the economic system became more egalitarian, the political system would become more democratic. The ability of certain strategically placed economic elites to control government policy would be greatly curtailed and the need for foreign investment would decrease as the domestic market expanded. Once true democracy was achieved, reason and peaceful cooperation could be pursued. Vladimir Illich Ulyanov, better known as Lenin, embraced Hobson’s ideas and (drawing as well on Marxist writers such as Hilferding, Kautsky, and Bukharin) set out his own theory of international conflict in Imperialism: The Highest Stage of Capitalism in 1916. 65 While Hobson believed that imperialism was due to a maladjustment in the capitalist system and that it was possible for imperialist wars to be prevented, Lenin argued that imperialism was an inevitable by-product of the final stage of capitalist development.

188

Chapter 5

Lenin believed that the capitalist economic system must expand or die. Falling rates of profit required economic expansion, which sooner or later (through the process described earlier by Hobson) required war. But foreign economic expansion is not a matter of choice; it is a necessity. When the state pursues a policy of imperialism, it is only doing what it must to deal with the inevitable crises capitalism engenders. Lenin’s explanation for war was not too dissimilar from Hobson’s. The root cause lay in the crisis of the capitalist economy: surplus production and underconsumption. Lenin argued that the final years of capitalism as a stage of history would be characterized by “finance capitalism” or “monopoly capitalism.” In this situation the control of the economy becomes increasingly concentrated in the hands of a smaller and smaller group of owners. The most important of these corporate oligarchies are the financial and banking institutions; this paralleled Lenin’s belief that further economic growth would depend on the export of surplus capital (that is, foreign investment). As monopoly capitalism progresses, giant corporations begin to divide up the world. Owing to the “fusion of state and capital,” the government becomes the agent of the capitalist class. One thing leads to another and—given the simultaneous existence of many states in the late stages of capitalism—it eventually becomes necessary for the capitalist states to fight each other over increasingly limited overseas markets, materials, and investment opportunities. International economic relations become a zero-sum game—a game in which the winners win only what the losers lose. The world becomes like a pie of a fixed size; it contains only so many areas usable for colonial expansion. Once the pie has been initially divided among the competitors, a state that desires to expand (that is, to increase its portion of the pie) can do so only at the expense of another state’s portion. Since capitalist states develop at different rates of growth and from different starting points, some of them will inevitably be economically (and therefore militarily) superior to others. This is the law of uneven capitalist development. Economic competition naturally leads the more powerful capitalist states to exploit weaker capitalist states. The result, of course, is war. This conclusion is inevitable from Lenin’s point of view. Lenin’s solution was significantly different from that of Hobson. Peace could be secured only by the abolition of capitalist states and the creation of a world of socialist states. However, the transition from capitalism to socialism would not come about peacefully through elections; rather, socialist revolutions would be the agent of this change. However, as Kenneth Waltz has indicated, there is some confusion here about what actually brings about peace: Is it the destruction of capitalism or the destruction of states? 66 The two are linked in Marxist theory because socialism, the economic phase of history that replaces capitalism, will eventually develop into true communism—a stage of development in which the state “withers away.” As Michael Haas points out: “The most peaceful political system turns out to be the one with no visible government at all. There will be no elites to compel men to fight, and there will be no states to attack or defend.” 67 Critiques of Theories of Imperialism How well have the theories of Hobson and Lenin held up? Several observations about capitalism and war can be made.

The State Level of Analysis, Part I

189

First, the essence of the theory is that domestic underconsumption and the lure of higher profits through foreign investment cause imperialism (and thus war). But there exist several other ways to cure the economic problems, as Hobson points out: for instance, pumping up domestic demand by redistributing wealth so that the masses would have more purchasing power. Thus, political choices determine whether imperialism and war occur; they do not necessarily flow from capitalism. 68 Second, contrary to the contention of Lenin, capitalism benefits from a policy of peace. War upsets economic planning; destroys land, labor, and capital; makes profits uncertain; hinders trade; and uses up scarce resources in the process—all of which make it an unappealing choice for capitalists. Third, some capitalist states in the late nineteenth century did not produce any surplus capital (or exported little to their own colonies) and yet still pursued imperialist policies. As D. K. Fieldhouse points out, Hobson and Lenin were “entirely wrong in assuming that any large proportion of British overseas investment went to those underdeveloped parts of Africa and Asia which were annexed during the ‘imperialist’ grab after 1870.” 69 England invested half of its capital outside the colonies, primarily in the United States, and France consistently ranked only second or third in investment in her own colonies. 70 Most of the capital investment during the age of colonialism was in other capitalist states, not in colonies. If this is true, it would be illogical for the great powers to fight over them. Indeed, colonial competition did not usually end in war but in accommodation and the negotiation of spheres of influence. With the exception of the Boer War and the Russo-Japanese War, almost all colonial conflicts were settled with diplomacy. 71 Fourth, although many capitalist states were engaged in imperialism, some were not. How does one account for the peacefulness of capitalist Sweden and Switzerland? On the other hand, some important imperialist states were neither capitalist nor surplus producing, for example Japan and Russia. 72 We have a situation here where not all capitalist states were imperialistic and not all imperialist states were capitalistic. One should logically conclude from this that capitalism is neither a necessary nor sufficient condition for imperialism. Fifth, socialist states have themselves exhibited aggressive tendencies. The past seventyfive years have witnessed the Soviet invasion of Finland (1939), the Baltic countries (1940), Hungary (1956), Czechoslovakia (1968), and Afghanistan (1979); China has attacked Tibet (1956), India (1962), and Vietnam (1979); Vietnam has attacked Cambodia (1975); the Soviets and the Chinese have fought several border clashes; North Korea has invaded South Korea (1950); and the Ethiopians and the Somalis have fought over the Ogaden (1977). In the 1945–67 period, socialist systems constituted only 15 percent of countries but participated in fifteen—roughly 25 percent—of the sixty-one international conflicts. Sixth, it should be obvious that imperialism and war both occurred before the age of capitalism and that the feudal, agricultural status of many states, past and present, has not prevented them from pursuing aggressive policies of expansion. Imperialism is older than capitalism, a situation that prompts Waltz to reflect that we have a unique (and totally unsatisfying) situation here of a theory whose cause (capitalism) is much younger than its effects (imperialism and war). 73

190

Chapter 5

Seventh, there is no doubt that capitalist states have been highly involved in expansion and warfare in the nineteenth and twentieth centuries, but not necessarily because they were capitalist. Capitalism may not require economic expansion or growth in order to survive. What may be more likely, as William Appleman Williams, a founder of the revisionist school of American history, argues, is that political and economic leaders in capitalist countries believe that capitalist economic systems require growth and act accordingly. 74 This explanation takes us back to the individual level of analysis. THE ROLE OF THE BUSINESS CYCLE: GOOD TIMES AND BAD TIMES One commonly held belief has been that states tend to go to war during times of economic distress and depression. The explanation for this reputed phenomenon comes in several varieties. First, some theorists suggest that economic hard times create pressure on political leaders to expand the economy through searches for greater markets for their products and investments or for access to more productive resources—a process that eventually leads to war. Iraq’s 1990 attack on Kuwait in order to control its oil reserves and to relieve itself of its debt to Kuwait is only the most recent example of war based on such motives. Second, it is also argued that national leaders seek war in the belief that this will stimulate the economy through the creation of more products and jobs. In other words, war itself is believed to have a beneficial impact on the economy. Third, during economic hard times political elites may seek war as a method of diverting the attention of the public from internal woes. Severe economic conditions in Argentina certainly led to the decision to invade the Falkland/Malvinas Islands in 1982. (This last explanation, the scapegoat theory, will be taken up in the next chapter.) Finally, it is also possible that political leaders are more willing to take risks during periods of sudden economic distress. 75 It might be argued that industrial democracies should be especially affected by situations of economic distress. For instance, the vulnerability of American presidents to electoral punishment during times of economic adversity is well known. It is also understood that the president’s domestic popularity may be increased by forceful actions toward rival states. Ostrom and Job point out that in the post–World War II era, the international use of force by presidents has been especially common during lean economic times (as indicated by the “misery index,” which combines data on unemployment and inflation). 76 Research by Bruce Russett indicates that there is a relationship in the nineteenth and twentieth centuries between economic downturns and American participation in MIDs. The combination of poor economic performance with U.S. presidential election years is an even better predictor of involvement in international disputes. However, there does not appear to be a relationship between weak economic conditions and interstate war per se. 77 Several scholars argue the opposite point of view—that the depths of the economic cycle do not instigate international conflict but serve instead to constrain the pursuit of war. Blainey cites this as a major factor in preventing Austria from attempting to recapture Silesia in 1749 and for delaying the Japanese invasion of Korea in 1873. 78 In fact, several analysts suggest that it isn’t depression that leads to war, but economic recovery. The most famous argument is probably that of A. L. Macfie, who published in 1938 a study of the effects of the British

The State Level of Analysis, Part I

191

business cycle on twelve wars from 1850 to 1914. (The British themselves were only involved in three of these wars, but it is assumed that the fluctuations present in the British economy reflected a truly international business cycle, thus justifying the use of British economic statistics with the war behavior of several countries.) Matching annual statistics on employment against the onset of war, he concluded that wars were most likely when an economic recovery was in its later stages. 79 A recent study of global economic cycles (called “long cycles”) and war from 1495 to 1975 by Joshua Goldstein finds a strong and consistent correlation between the severity of war and economic upswings. 80 Although wars have occurred in roughly equal numbers throughout history in the upswing and downswing phases, the most severe wars have taken place in upswing phases. From 1495 until 1918 each peak in war severity occurred near the end of an upswing phase. Why should economic recoveries be related to war? It could be that the underlying cause of hostilities had been present for many years, but that governments practiced restraint during the period of economic distress. War was undertaken only when the economic upswing made it financially feasible. 81 In this explanation, economic upturns are not the actual cause of war, but a factor that enables wars to occur. The historical association of wars with economic upswings may therefore actually obscure the real causes of war, which might be found in the preceding period of economic decline. 82 A psychological explanation for the relationship between war and economic upturns is also frequently made. Indeed, both Macfie and Blainey, as well as Goldstein, suggest that economic recoveries are associated with a general mood of optimism, which is the real cause of war. Blainey argues, When trade is deteriorating and when unemployment is increasing the mood of governments tend to be cautious and apprehensive. Dwindling revenues and soaring claims for the state’s aid aggravate the mood. On the other hand, when prosperity is high—and this time is the most dangerous to peace—there comes a sense of mastery of the environment. 83

Blainey is describing here a collective national mood, a feeling of optimism and confidence that colors the judgment of both political leaders and common people. 84 Blainey believes such a feeling of optimism and mastery was evident on the eve of the Crimean War, the FrancoPrussian War, the Boer War, and others. 85 Dexter Perkins finds the American experience fits this general pattern. He contends that belligerent, pro-war feelings in the United States coincided with recoveries from economic downswings. The War of 1812 followed hard on the heels of a commercial upturn; the Mexican War occurred after the depression of 1837–1842; the Spanish-American War took place after the return to prosperity following the depression of 1893; World War I followed the economic decline of 1913–14; and World War II took place during the recovery from the Great Depression of the 1930s. 86 William Thompson looks at business cycle data for Great Britain, France, Germany, and the United States from the nineteenth and twentieth centuries in an effort to test the Macfie/ Blainey hypothesis (Macfie suggests war is associated specifically with the last phase of the recovery, Blainey with any phase of the recovery). Thompson finds that support for Macfie’s

192

Chapter 5

hypothesis is confined to Britain’s nineteenth-century colonial wars. Support for Blainey’s more general thesis is found only in the American experience—validating Perkins’s observations. With the exception of the Boxer Rebellion, all the American wars examined were begun during an upward phase of the business cycle. Britain and France had some warfare in every phase of the cycle. 87 Probably the most we can say about the business cycle is that it may play a role in the development of war, but its effects are less than clear. Some wars have broken out in hard times, others have occurred in good times. Neither economic weakness nor prosperity seem to prevent war. DOES SIZE MATTER? THE ROLE OF POWER, SIZE, AND DEVELOPMENT An argument quite frequently made by “realists” is that large, powerful states (regardless of the nature of their political or economic systems) tend to be perpetrators of war rather than small states. 88 This would seem to make intuitive sense. It is more likely that large states would pick on small states than vice versa. In a world of rational calculation, larger, more powerful states are more likely to win and this enters their leaders’ calculations of the costs and benefits of war and of the probable outcome. Also, larger states would seem to be more likely to get involved in conflicts simply because they are more involved in international affairs in general. They have more interests, are participants in more international organizations and alliances, have more international commitments, and have greater capacity to act in international affairs than smaller states. And they are more likely to feel a responsibility to take action (to rectify the international balance of power, for instance). Leaders of large states are more likely to hold national role conceptions that picture their states as responsible for protecting allies, defending the international status quo, promoting a particular ideology, and ensuring world or regional order. A sizable amount of empirical evidence tends to support this thesis. In their study of wars from 1815 to 1965 Singer and Small found that the larger, more powerful nations seemed to be the most warlike. Eleven nations—Britain, France, Russia, Turkey, China, Spain, Germany, Italy, the United States, Japan, and Austria-Hungary—accounted for 90 percent of all battle deaths and 60 percent of all nation-months of war. At least one of these eleven states was involved in 71 percent of the wars in the period studied. 89 While the great powers were clearly disproportionately involved in war, during the same period 77 of the 144 smaller powers were able to escape war entirely. 90 A post–World War II listing of states highly involved in war would not only include the five great-power members of the UN Security Council but also a conflict-ridden group of developing states from the Middle East and South Asia: India, Pakistan, Israel, Egypt, Iraq, and Syria. Stuart Bremer ranked members of the international system from 1620 to 1964 according to a composite index of demographic, economic, and military power. He discovered what appeared to be a strong linear relationship between power rank and war involvement. The states with the greatest capabilities were involved in the most wars and also initiated more wars than the less powerful states. For instance, the states occupying ranks 1–5 averaged war once every ten years while states ranked 41–45 and 46–50 averaged a war once in every one hundred

The State Level of Analysis, Part I

193

years. 91 Bremer predicted that at least one of the world’s ten strongest states would be involved in a war within the next decade. As other scholars have noted, this turned out to be a spectacularly successful prediction. Depending on whether you choose to start the prediction in 1979 or 1980, up to four of the top ten states on Bremer’s list were in fact involved in a war within ten years of the prediction. 92 Power is not just linked to involvement in war, but involvement in conflicts short of war as well. Michael Haas used UN dues (which are based on a state’s ability to pay) to divide countries into four categories based on wealth. The results indicate that the richest states rank highest in significant foreign conflict, and the amount of foreign conflict a state experiences drops consistently with its level of wealth. Since wealthy states are frequently also militarily strong states, this is a good test of the relationship of power and conflict. 93 Gochman and Maoz show that the most powerful states in the nineteenth and twentieth centuries are also the ones most involved in MIDs. 94 The previously mentioned survey of international conflict from 1816 to 1992 by Maoz provides strong confirmation of these relationships. Maoz identifies “fightaholics” who are disproportionately involved in MIDs or wars on a continuous basis over several historical periods. These conflict-prone recidivist states typically had high levels of military and economic capabilities and were deemed to be either major powers or regional powers in terms of status. States with no involvement in wars and limited involvement in MIDs looked like mirror images of these fightaholics. Maoz in fact concludes that “minor powerhood” is a necessary (but not sufficient) condition for war pacifism. 95 Some research has contradicted these findings. 96 R. J. Rummel’s Dimensionality of Nations (DON) project, which studied international affairs in the 1955–57 period, finds no relationship between the size and military capabilities of a state and its foreign conflict behavior. 97 Several other studies of the post–World War II era indicate that while larger, more developed states have more total acts of conflict behavior (this would include verbal as well as nonverbal action) than other countries, this is because they are generally more involved in international affairs. When the total number of actions is controlled for, larger, more developed states actually have a slightly higher percentage of cooperative actions. Also, the conflictual acts of developed states are slightly more likely to be verbal and less coercive than the conflict behavior of less developed states. 98 It seems, however, that the preponderance of evidence in this debate favors those who find an association between power capabilities and war. Much of the contrary evidence is from studies that examine an extremely limited period of time or examine the more general concept of “conflict behavior” rather than war per se. A large amount of the globe’s warlike behavior is accounted for by a small number of states. Great powers seem to be quite heavily involved, with the distinction that (at least in the post–World War II era) they rarely seem to be directly involved against each other. Of the roughly thirty-seven interstate wars from 1945 to 2007, only one (Korea) involved forces of the great powers arrayed against each other. 99 If we assume that larger, more developed (and therefore more powerful) states are indeed the most war prone, then how could the threat of war be lessened? There is no great movement calling for the creation of a world of small, weak states. And while political theorists from Plato to Rousseau to Dahl have trumpeted the peaceful virtues of small societies, until recently

194

Chapter 5

there has been little evidence of the willingness of modern states to participate in their own fragmentation into smaller societies. The recent peaceful disintegration of the USSR into fifteen independent states and the splitting in two of Czechoslovakia may presage an incipient international trend. However, one cannot be encouraged by the violence with which the fragmentation of the former Yugoslavia was accomplished. Finally, even if it were possible to create an international society of small states, the history of small city-states is not at all encouraging: witness the Peloponnesian Wars or the wars of the Italian city-states. The problem is that power, wealth, and development are relative concepts. Even in a world of small states, some will be relatively larger, wealthier, and stronger than others. The difference is based on mutual comparison. This discovery takes us to a different level of analysis— away from the nature of the state and toward the power disparities between them, or to the individual level of analysis and perceptions of threat based on comparisons of relative power. If it is the relative power of states that is at fault, or the perception of power disparity, then it would seem that war will last as long as people are organized into entities whose power disparity can become a cause of alarm to others. THE ROLE OF POPULATION (1): LEBENSRAUM One interesting state characteristic often discussed as a cause of war is that of overcrowding caused by population growth. This is the lebensraum theory of war. Lebensraum, or the state’s need for “living space,” was closely associated with the development of the discipline of geopolitics in Germany in the late nineteenth and early twentieth centuries. 100 German political geographers like Friedrich Ratzel saw states as living organisms with similar life cycles: they occupy space, they grow, they contract, and they eventually die. Ratzel and the Swedish geographer Rudolf Kjellen contended that states were involved in a continuous struggle for living space and survival, thus introducing concepts from social Darwinism into geopolitical theory. The chief proponent of these ideas in the interwar years was General Karl Haushofer, a professor of geopolitics at the University of Munich and a mentor of Rudolf Hess—who introduced his ideas to Adolf Hitler. 101 Many of these themes eventually found expression in Hitler’s Mein Kampf. The thrust of German geopolitics from Ratzel to Haushofer was that great powers needed to achieve population growth, needed to enlarge their boundaries to obtain lebensraum, and needed to achieve economic self-sufficiency. Japan’s successful imperial expansion in the 1930s was taken as proof of the desirability of following such geopolitical strategies. 102 The lebensraum theory had both empirical and normative aspects. It was the normative side that Hitler emphasized. If one believed that states either increased their populations and their territories or they died, then expansion became a matter of compelling national interest. Hitler believed the German or Aryan race must expand at the expense of Slavic peoples in Eastern Europe. As he stated in Mein Kampf, “nature has not reserved this soil [Europe] for the future possession of any particular nation or race; on the contrary, this soil exists for the people which possesses the force to take it.” 103 One does not have to view the lebensraum theory only in terms of its German, fascist, and racist background. More generally, the lebensraum theory of war simply states that if popula-

The State Level of Analysis, Part I

195

tion pressures within a country become extreme, the government will be faced with interrelated political and economic problems. There will be a greater demand for many kinds of resources, including land, water, energy, food, and government services to relieve the poverty and unemployment caused by rapid population increases. Failure to solve these problems may erode government legitimacy. Political leaders may seek to address these demands through the expansion of their national territory at the expense of their neighbors. Several different causal paths from population growth to war are possible. One path is the one generally described by lebensraum theorists. Population growth (especially growth relative to available space) leads to higher levels of population density and pressure on the land. This leads to increased demand for and use of resources, leading to resource depletion and perhaps environmental degradation. The state, in response, initiates an expansionary war to gain resources and territory. A second path might lead from population growth and resource depletion to increasing demands on the government. The government’s failure to satisfy popular demands leads to internal conflict, and the government responds to this with diversionary force against neighboring states. 104 Recent research on these theories has been mixed. Stuart Bremer and his associates in the Correlates of War project studied wars involving a sample of European states between 1816 and 1965 and found that neither population density nor changes in population density were associated with participation in war or initiation of war. 105 As a general rule, population growth does not seem to induce warlike behavior. Indeed, there were important instances in which international conflict was precipitated by states experiencing a relative decline in population. Quincy Wright relates that one of the causes of French bellicosity in the late nineteenth century was that its population was declining relative to that of Germany. 106 On the other hand, Nazli Choucri’s investigation of forty-five cases of international violence from 1954 to 1972 involving less developed countries (LDCs) found statistical relationships generally between violence and various measures of population distribution (population density, population movement, and population pressure) and measures of population composition (nature of age cohorts). 107 The discrepancy between Bremer’s study of the nineteenth and early twentieth centuries and Choucri in the second half of the twentieth centuries is somewhat logical. Prior to the twentieth century, few states were likely to have the rates of population growth that we see in the post–WWII period, and when they did, war was usually not necessary because potential colonial territory was still readily available. But in the post–WWII era few unclaimed areas are available for expansion, so expansion had to come at the expense of other established states. We might also expect that developed states would feel less necessity to take drastic action to deal with population pressures than LDCs; their governments have many more domestic tools to deal with the pressures generated by population growth. Moreover, we might also hypothesize that population pressure would probably not be likely to drive democratic states toward militaristic actions. As Gleditsch notes, “Since democracies rarely if ever fight one another and since they rarely experience civil war there is no reason to believe that they will suddenly start fighting over resource issues between themselves, or internally, any more than over other issues.” 108

196

Chapter 5

A recent study by Tir and Diehl looked at population data for all states in the international system between 1930 and 1989. 109 They were interested to see whether the rate of change in population (population growth) or population density had any effect on involvement in MIDs, initiation of MIDs, or escalation of MIDs to war. They found that while population density had no independent effect on MID involvement, population growth did have a significant but modest positive effect on MID involvement, especially when coupled with strong levels of military capability. (States with growing populations, but minuscule armies are unlikely to choose the use of force with their neighbors.) And, the relationship between population growth and dispute involvement was much stronger for minor powers with low levels of technology than for great powers. Finally, neither population growth nor density were linked to the initiation of militarized disputes. If anything, states with high population growth rates were less likely to initiate MIDs—a finding very much at odds with what the theory would predict. They are forced to conclude that states do not appear to engage in militarized conflicts for the purpose of gaining land to support their growing populations. 110 While population growth and declining access to resources continue to be seen as a potential threat to world peace, there is little empirical evidence of much interstate war precipitated by these factors in the last several centuries. Population growth can lead to internal political conflict through a variety of complex paths, but the link to external conflict is much less clear. THE ROLE OF POPULATION (2): LATERAL PRESSURE In Nations in Conflict, Nazli Choucri and Robert North developed a more complex version of the old lebensraum theory and attempted to evaluate its ability to explain the occurrence of World War I. 111 They contend that it is not population growth alone that is the root cause of war but the combination of population growth with technological growth. A growing population leads to increased demands for resources, while simultaneous technological growth means that resources are actually used at an increased rate. Resources are needed in greater variety and in larger quantities. These two simultaneous developments lead to rising demands. When existing domestic capabilities are insufficient to meet demand, new capabilities have to be developed. Demand leads therefore to lateral pressure—the expansion of activities beyond the state’s borders by private citizens, corporations, and governments. This can take a variety of forms: foreign investment, trade, acquisition of spheres of influence or colonies, the dispatch of troops to foreign areas, the establishment of military bases in foreign lands, and so on. A national policy of expansion becomes institutionalized. The state develops a stake in external growth, and lateral expansion is increasingly seen as a national interest that must be protected. This process of lateral expansion caused by population and technological growth leads to war when the external interests of two or more great powers intersect or come into conflict. The stronger a country’s lateral pressure, the greater the likelihood of intensification of competition, and thus arms races, crises, and war. Choucri and North argue that lateral pressure itself rarely triggers war. Indeed, if the only form that lateral pressure takes is trade, the result may be that countries develop closer relations with each other. Intersections are most likely to develop into war when relations are already hostile or when at least one of the parties per-

The State Level of Analysis, Part I

197

ceives the actions taken toward them by the other to be “dangerously competitive, threatening, coercive, menacing, or overtly violent.” 112 Choucri and North’s theory of lateral expansion may sound like those of Hobson and Lenin. But whereas all these theorists trace the cause of war backward to economic competition between the great powers, Hobson and Lenin pinpoint the presence of capitalist economic institutions as the cause of such competition, while Choucri and North argue that the type of economic system is irrelevant. What matters is the simultaneous presence of population and technological growth, regardless of the type of economy. Choucri and North subject their theory to statistical analysis using data for six great powers from 1870 to 1914—a period that begins at the apex of European colonial competition and ends with World War I. (Colonialism is just one route by which lateral pressure may lead to war, but it is the most logical route for this particular slice of time.) Their results are largely disappointing. No direct path from domestic growth to lateral pressure to colonial competition to war was found. 113 In another study, the authors use the lateral pressure theory to explain Japan’s foreign policies from 1878 to 1987. 114 Attempts to apply lateral pressure theory to the more recent actions of the United States, the former Soviet Union, and the People’s Republic of China were equally disappointing. There appears to be little relationship between measures of lateral pressure and conflict. 115 Choucri and North followed up on their initial attempts by developing country profiles based on lateral pressure theory. Alpha countries are states with large and growing populations, with advanced technologies, and with access to plentiful resources. Alphas are highlateral-pressure states. Beta countries have large populations relative to their territory, are advancing technologically, but have impeded access to resources. Rising demands in these states lead to pressure to expand through increased territory or trade. Gamma states have a limited resource base but a high degree of access to resources through an extensive trade network (present-day Britain and Japan). Choucri and North claim that high-lateral-pressure countries (Alphas, Betas, and Gammas) fight more wars per country than developing states (Epsilons, Zetas, and Etas), though most wars tend to be fought in the developing regions of the world rather than in the industrialized portion. High-technology, low-population countries (Delta states, such as Norway) seem to fight the fewest wars, and when they are involved in wars, it tends to be as victims, rather than as initiators. 116 Lateral pressure approaches to war have now been largely abandoned as unproductive. Choucri and North’s analysis of country profiles and other studies indicate much the same phenomenon. Strong, developed countries with growing needs are prone to international conflict. The implications of this are fairly pessimistic. As Choucri and North admit, a conscious attempt to change one’s profile as a method of reducing one’s chances of conflict is not likely to work. These attributes are highly resistant to change in the short term and difficult to alter even in the long term. Nevertheless, Choucri and North suggest that population management programs might be helpful in reducing these long-term pressures for expansion, if combined with a more even diffusion of technology and a more equitable access to global resources. 117 As we have seen, attempts to link political pressures stemming from resource scarcity to international violence have been largely unsuccessful. While it is possible that in the future resource scarcity (crops, fresh water, forests, fisheries) that has been engendered by popula-

198

Chapter 5

tion growth and by environmental degradation will be responsible for interstate conflict, there is no strong empirical evidence that such scarcities have been at the root of interstate wars in the last several centuries. However, resource scarcities may well be an important factor in causing internal strife within states. 118 CONCLUSION It’s hard to escape the conclusion that the state-level theories we have looked at thus far have done a relatively poor job in explaining the incidence of war. Neither the type of government a state has, nor its economic institutions, its economic well-being, its population growth rate, or its status as a rogue state seem to matter much. The only proposition that seems to be consistently supported is that the larger and more powerful a country is, the more likely it is to be involved in war. We will have to delay a definitive summation, however, because there are a few more bites to take out of this apple. There are several more theories that address the connection between state characteristics and war that we must examine in the next chapter. Perhaps they will turn out to be more promising. It is important to remember the differences between the operational milieu and the psychological milieu with regard to state level attributes. It may not be that capitalist states or states with resource pressures require expansion but rather that their leaders believe that they require expansion. Similarly, it may not be true that states with rapid population and technological growth pursue expansionist policies in a manner determined by these attributes. It may be more important that the leaders of these states perceive that they must pursue an expansionistic foreign policy because of their growth. The same logic applies to rogue states (powerful states believe that rogue states are dangerous and thereby treat them in ways that promote conflict) and the power status of states (leaders of powerful states believe that they have, for example, a policing role in the international system). However, instead of saying that state-level theories are totally negated or invalidated by individual level perceptual factors, it is probably best to see factors at these levels as linked. A comprehensive theory of war will require variables at several different levels of analysis. State-level variables such as size and power may be underlying conditions that cause war, but individual-level variables such as perceptions and national role conceptions operate as the mechanisms through which these underlying conditions lead to war.

Chapter Six

The State Level of Analysis, Part II Internal Conflicts, Nationalism, and War Weariness

Peace, like charity, begins at home. —Franklin D. Roosevelt Breathes there the man, with soul so dead, / Who never to himself hath said, / This is my own, my native land! —Sir Walter Scott

As we continue to look at characteristics of states that might make them prone to war, we turn next to the variable of internal political turmoil. One of the most persistent theories of war is that internal conflict and external conflict are causally connected. Perhaps the most widely cited articulation of this link is by historian Geoffrey Blainey in his The Causes of War. In the 125-year period he examined (1815–1939), Blainey discovered that at least thirty-one wars (representing just over 50 percent of all wars during the period) were immediately preceded by civil conflict in one of the belligerent countries. 1 This seems to be a clue worth following. There are some scholars who question whether there is any significant link at all between internal and external conflict. In fact, much of the early empirical literature cast doubt on this central proposition. Most famously, R. J. Rummel investigated data from eighty countries for the years 1955–57 to see whether states with high levels of domestic conflict (as indicated by assassinations, purges, strikes, riots, coups, demonstrations, etc.) also exhibited high levels of foreign conflict (including wars, sanctions, troop movements, expulsions, and verbal protests and threats). Using the technique of factor analysis, Rummel found three distinct dimensions of foreign conflict (war, diplomacy, and belligerency) and three different dimensions of domestic conflict (turmoil, revolution, and subversion). He further discovered that the various dimensions (or factors) of foreign conflict behavior were almost completely independent of the various domestic dimensions. That is, states with high scores on any of the three domestic conflict factors did not necessarily have high scores on any of the foreign conflict factors. Unrest at home was not necessarily linked to foreign conflict. 2 Rummel’s conclusion that internal and external conflict are unrelated was supported by similar factor analysis studies by several others. 3 199

200

Chapter 6

This absence of support in early data-based studies for an internal-external conflict link flies in the face of the very substantial reports of its existence in historical and political case studies. However, virtually all these early empirical studies indicating an absence of an internal-external conflict link are subject to strong criticism. The results were based on a narrow (and perhaps unrepresentative) time span; they used inappropriate statistical methods; they were formulated without a coherent theoretical framework to guide them; and the foreign conflict variable includes a wide variety of behaviors, only one of which is actually war. 4 Later studies using more sophisticated methodology and larger databases have shown much more support for the linkage between internal and external conflict. For instance, Geller’s analysis of a variety of internal and external conflict variables for thirty-six states for the years 1959–68 shows that states with high levels of internal instability are more likely to be involved in external conflict than states with lower levels of instability. 5 Mansfield and Snyder find a very strong and statistically significant relationship between the presence of civil wars and interstate wars (but not world wars and not extra-systemic or colonial wars) for the period between 1816 and 1992. 6 As part of their research they perform an analysis of the probability that a given dyad (a pair of countries) will be involved in an interstate war with each other. Their results indicate that if one country in the pair is undergoing civil war, this strongly increases the probability of war. 7 Another analysis by Houweling and Siccama demonstrated that civil wars and interstate wars cluster in both time and space, implying that the presence of one increases the probability of the other. 8 Cashman and Robinson’s case studies of seven wars in the last one hundred years (though based on an unscientific sample) found that in five of the six wars between equals, domestic political variables helped to explain the initiation and escalation of conflict. In these cases, one of the participant states had significant domestic political instability and/or ruling elites perceived themselves to be politically vulnerable. (The cases are Iran prior to the Iran-Iraq War, Pakistan prior to the 1971 war with India, Japan in the 1930s and 1940s, Ethiopia in 1998, and Germany and Austria in 1914.) 9 If we assume a relationship does in fact exist between internal conflict and external conflict, what theory might explain this relationship? Several possible causal pathways might explain it. Some of these are complementary and others are clearly incompatible: (1) the diversionary or scapegoat theory of war, (2) a variation of diversionary war theory based on the concept of contested institutions, (3) the “kick ’em while they’re down” theory, (4) the effect of revolutions, and (5) the internationalization of internal conflicts through external intervention. (A sixth phenomenon, democratization, is closely related to contested institutions and was described in the previous chapter.) THE ROLE OF INTERNAL CONFLICT (1): THE DIVERSIONARY/SCAPEGOAT THEORY OF WAR One explanation of the relationship between internal and external conflict has been called the scapegoat or diversionary war theory. The argument here is that when states are beset by increasing political opposition and civil strife or by deteriorating economic conditions, their leaders will seek to solve these internal woes by initiating conflict with an external foe. 10 War is undertaken in the belief that it will rally the masses around the flag in the face of a foreign

The State Level of Analysis, Part II

201

threat; the external foe becomes a scapegoat. Internal problems can be blamed (justly or unjustly, though the presumption is the latter) on the external opponent, and victory over the scapegoat is touted as essential to reverse the wretched internal situation. Perhaps the territorial occupation of the opponent can even provide access to needed resources that can actually alleviate real internal problems. Alternatively, the war may simply be used by the government to divert the attention of citizens from the internal situation or to achieve a foreign policy success as a way of increasing its domestic popularity. Whether the use of war actually alleviates the internal situation is, of course, another question. In terms of the theory, it really doesn’t matter. What is important for the theory is that political elites believe that precipitating a foreign conflict gives them a chance, however slim, to save their political skins. The diversionary theory has underpinnings in the sociological literature. The well-known conflict-cohesion theory associated with Georg Simmel and Lewis Coser argues that conflicts with out-groups increases the cohesion of the in-group—at least if the group perceives itself as a distinct entity, already has some minimal level of cohesion, believes the group’s preservation is important, and believes that the out-group indeed threatens the in-group as a whole. 11 And while extrapolating sociological principles from small groups to state-sized political entities is problematic, there is substantial evidence that a rally-round-the-flag effect does exist in many countries. For instance, the immediate effect of a state’s involvement in crises and wars is normally to increase support for the government and its leadership in public opinion polls. 12 While the theory is usually referred to as the diversionary war theory, the actual dependent variable may be something other than war. 13 Leaders may simply make threats, put on displays of force, or take militaristic actions short of war. They may feel that these actions are enough to divert the public’s attention and shore up support without intending war; after all, wars are costly and their results are unpredictable. But aggressive behaviors initiated as a diversion may lead to retaliatory acts by the target states that escalate inadvertently to war. General Galtieri’s government in Argentina most probably thought that it could seize the Falklands/Malvinas without having to fight Britain. Saddam Hussein almost certainly thought that he could create a fait accompli by seizing Kuwait without fighting a full-scale war with others to keep what had been seized. As Jack Levy notes, this means that the diversionary theory is not really a theory of war, but a (partial) theory of foreign policy behavior. Whether war occurs frequently depends on the interactions between states, not on the behavior of a single state. 14 The independent variable is also somewhat difficult to nail down. Actual internal conflict may not be a necessary condition for a diversionary war (or use of force short of war). Leaders may simply feel politically insecure or feel that they are vulnerable to political opponents, even in the absence of any real domestic turmoil—which might range from antigovernment strikes or demonstrations to armed rebellions. Political vulnerability may be the result of declining economic or social conditions or just the routine cycle of elections. 15 Diversionary war theorists frequently make reference to a number of systematic qualitative studies of war. For instance, Quincy Wright’s early analysis, A Study of War, is often quoted to the effect that “foreign war as a remedy for internal tension, revolution, or insurrection has been an accepted principle of government.” 16 Richard Rosecrance’s classic study of international instability within nine different European systems from 1740 to 1960 concluded that the

202

Chapter 6

domestic insecurity of political elites was one of the most important causes of major-power war. Elites in democratic and nondemocratic systems alike sought relief from internal troubles through war. 17 Richard Ned Lebow’s investigation of “brinkmanship crises” in the twentieth century also underscored the importance of domestic political uncertainty. Ten of the thirteen crises in his sample were initiated by leaders who perceived their rule to be vulnerable to domestic opponents; in four of these ten cases, the political system itself was weak and unstable. 18 Historians have frequently made scapegoat arguments for the French decisions for war in 1792 and in the Crimean War, for the Russian provocations that led to the Russo-Japanese War, and for the Austrian and German decisions for World War I. 19 Other examples might include U.S. military interventions in Grenada (1983) and Panama (1989), and Iraq’s invasion of Kuwait (1990). The modern poster-child for scapegoat wars has been the Argentine government’s decision to wrest the Falkland Islands from Britain by force of arms in 1982. Severe political and economic problems in both Argentina and Britain had led to increased political opposition to both the Galtieri and Thatcher governments, thus potentially providing the former with a strong incentive to take the Malvinas by force and the latter with an equally strong reason to reverse the situation through war. In the early 1980s, the Argentine economy was in dire straits. There were numerous bank collapses, triple digit inflation, massive unemployment, and a growing awareness of human rights abuses by the junta in the so-called dirty war against leftist dissidents. Middle-class support for the regime plummeted; political opposition parties united against the junta; there were calls for a return to democracy; and just days before the invasion there was a massive labor demonstration in Buenos Aires calling for bread, freedom, and work. Most analysts agree that the military action was meant by the Argentine junta to create a fait accompli to which the British would be unable and unwilling to respond militarily—a serious miscalculation. 20 However, even in this presumed air-tight case of diversionary war, analyses differ, and almost all are much more nuanced than a simple scapegoat thesis. Hastings and Jenkins argue the Argentine decision to invade was only partly aimed at quelling general public discontent. 21 Another important factor was General Galtieri’s promise to navy admiral Anaya—in exchange for Anaya’s support for his presidency—that Argentina would attempt to recover the Malvinas during Galtieri’s two-year term. Within weeks of Galtieri’s ascension to the presidency in December 1981, naval contingency plans for an invasion had been revised and a date set for sometime between July and October 1982. Thus, interservice bureaucratic politics, and the interests of the Argentine navy (which was set to receive the lion’s share of the glory for the Malvinas operation), were largely responsible for the initial decision. Hastings and Jenkins argue that “recovery of the ‘Malvinas’ would not stifle internal dissent, but at least it would unite the nation for a time. It would serve as a vindication of military rule and cleanse the reputation of the armed forces after the horrors of the dirty war.” 22 The severe political unrest in Argentina and the incident on South Georgia Island (in which a team of commercial scrapmetal foragers journeyed to the British island aided by the Argentine navy, established a presence, and raised the Argentine flag), and the subsequent dispatch of the British ice-patrol

The State Level of Analysis, Part II

203

vessel Endurance to the area several months later (in March) merely forced the government to push the invasion plan forward. Levy and Vakili also argue that the decision can be explained best by a combination of bureaucratic and domestic politics. 23 They argue that militarized bureaucratic-authoritarian regimes like the Argentine junta that are pushing free-market policies while suppressing internal dissent have difficulty establishing legitimacy due to their narrow base of support. But the real threat is not from lack of civilian support alone (as Hastings and Jenkins imply) but from internal disunity and conflicts within the regime itself. Levy and Vakili hypothesize that these regimes are most likely to turn to diversionary behavior when the following conditions exist: (1) a narrowing base of political support, (2) an economic crisis, and (3) intramilitary conflict within the regime connected to the absence of a unifying mission. Under these circumstances, military diversions can both unify the regime and increase its domestic political support. All of these conditions were met by Argentina in 1982. Like Hastings and Jenkins, Levy and Vakili agree that the domestic political and social context—an economic tailspin, middle-class unrest, a major labor union strike, calls for a return to democracy—were probably more important in the timing of the attack than on the initial decision for the invasion. Freedman and Gamba-Stonehouse argue that the decision was triggered more by the urgency of the dispute with Britain and misperceptions of British actions than the internal situation in Argentina. 24 They wisely note that there is rarely a single reason for war, and to the extent that a diversionary motive existed, it was probably secondary to other motives, among them sincere support for the idea of regaining national sovereignty over the Malvinas. 25 While the use of military force was clearly on the table, there was also a hope that a diplomatic solution might be found first. Ultimately, they believe the decision for war came because once the British had sent the Endurance to forcibly evacuate the Argentines from South Georgia, the junta could not afford to appear to be weak on the only issue they had any credibility on— foreign/security affairs. They probably overreacted out of concern that British military countermeasures would quickly create a fait accompli that would make it impossible to recover the Malvinas. Freedman and Gamba-Stonehouse argue that while the South Georgia crisis “had not been manufactured to deflect attention from the substantial economic and political discontent afflicting the country, publicly backing down to British demands would have further harmed the Junta’s already shaky popular standing.” 26 Finally, Richard Ned Lebow’s analysis also points to multiple, interlocking reasons for the Argentine decision. 27 First, British leaders failed to understand the gravity of the situation from the perspective of Argentina and therefore did not show resolve sufficient to deter Argentina from military action. Second, Galtieri and his colleagues engaged in defensive avoidance by resisting information suggesting that their policies would lead to war and by accepting any and all evidence that might bolster their belief in a successful outcome. They rationalized that they did not risk war because Britain lacked resolve, because it would be unable to retake the Falklands once it had been captured by Argentina, and because the U.S. would surely remain neutral. Third, Argentine leaders made use of a flawed analogy to buttress the decision for force. Their analogy was the Goa crisis in which India had forced Portugal out of the enclave of Goa—a use of force that was implicitly condoned by the international community as a proper exercise in national self-determination and the elimina-

204

Chapter 6

tion of colonialism. The British, however, saw the grab for the Falklands through the lens of a different analogy—Hitler’s aggression in the 1930s and 1940s. Finally, the domestic situation gave the junta further incentives to take military action. Given the domestic turmoil, the military could not afford to back down in the face of British intransigence; a retreat would lead to certain political disaster at home, while plunging ahead could produce significant gains with relatively little costs. While historians and political scientists who use case studies have made the Falklands Crisis the primary workhorse for the diversionary war theory, political scientists with a more statistical approach have sought to mine the experience of the United States. Quite a few studies have looked at the foreign policy behavior of the United States, and here too the results are mixed. Some studies question whether there is any empirical support at all for diversionary behavior, even by the United States, let alone other states. 28 And many of them use participation in militarized disputes (which might involve merely the issuing of threats rather than the actual use of force) instead of war as the dependent variable. For example, in the initial round of studies on the use of diversionary force, Russett found that for the period 1853–1976, elections and economic distress combined to explain U.S. participation in MIDs. Dispute participation was related to economic downturns (declining GDP per capita with a two-year lag), especially in presidential election years. However, the pattern did not hold for other industrial democracies like Britain and France. Moreover, nondemocracies were more likely to engage in disputes in times of economic expansion (increased GDP with a three-year time lag). 29 Russett reasoned that, in the absence of electoral pressures, nondemocracies were able to engage in international conflicts when their economies were better capable of supporting the effort. Studies by Ostrom and Job and by James and Oneal both point to the importance of domestic factors and declining levels of political support in accounting for U.S. use of force. They find that worsening economic conditions (measured by a misery index) and relatively high levels of presidential approval (the percent approving the president’s performance in office), coupled with declining popularity (the difference between his approval rating upon taking office and the current level), were related to the U.S. use of force in the post–World War II era. 30 Kevin Wang investigated the severity of responses made by the U.S. in thirtytwo international crises between 1954 and 1986. 31 He concluded that presidents were more likely to take severe actions under conditions of economic decline, in later stages in the electoral cycle, and with high levels of congressional support (as indicated by the number of seats in Congress held by the president’s party). On the other hand, the level of approval for the president by the public at large or by members of his own party appeared to have no impact. Numerous recent studies add support for the thesis that the U.S. uses external force for diversionary purposes. 32 The primary focus of these studies has been on the U.S., raising questions about whether the results could be generalized to other democracies or whether the U.S. case was perhaps special. And all attempted to explain involvement in militarized disputes or crises rather than war specifically. All use proxies for internal unrest—decline in GDP per capita, rising misery indexes, declining support for the president—rather than actual political unrest. Some look at behavior of the U.S. in crises or disputes already under way, making it difficult to draw

The State Level of Analysis, Part II

205

conclusions about the intention of political leaders to initiate new disputes for diversionary purposes. Many diversionary war scholars predict that the use of diversionary force by the U.S. and other democracies should follow the electoral calendar. They assume that leaders perceive greater vulnerability as elections approach and that the immediate run-up to elections is the precise point in time when leaders need a quick political boost that can be generated by a foreign crisis or the use of force abroad. Several studies have investigated links between American use of force and the presidential election cycle, but the results are decidedly mixed. 33 Morgan and Bickers believe the diversionary war theory needs to be reformulated. 34 They argue that incumbent elites, whether in a democracy or an autocracy, are more sensitive to declines in support from their own ruling coalition than from the public in general. Hence, the critical issue is whether there is a relationship between a decline in support for a regime from its own backers and the use of diversionary force. This is easy enough to do in a democracy like the U.S., but more difficult for other countries. In the U.S., we can look at yearly (or even monthly or weekly) polls aimed at gauging the partisan support for the president among members of his own political party. This is essentially what Morgan and Bickers do, using Gallup poll data from 1953 to 1976. Their dependent variable is U.S. engagement in militarized use of force—though they make no distinction between threats of force, displays of force, the use of force, and war. They find that lower levels of partisan approval of the president are associated with a higher probability of aggressive action by the U.S., completely independent of the level of support that the president might have among the public as a whole. No one argues that all wars are diversionary in nature: domestic vulnerability of leaders is not a necessary condition for war. Regimes whose leaders are not politically vulnerable initiate wars too: George W. Bush did not lack for high approval ratings in 2003. Neither is internal conflict a sufficient condition for the use of force. States frequently experience some kind of internal turmoil without resorting to any kind of external aggression. Elite vulnerability is probably a contributing factor in some wars, though the identity of this set of wars is open to question. The problem, as Levy notes, is this: “It is one thing to argue that the diversionary motivation contributed to the outbreak of certain wars, but quite another to evaluate its relative causal weight in the process leading to those wars; to assess the frequency with which diversionary action occurs; or to identify the internal and external conditions under which it is most likely to arise.” 35 So under what circumstances are states most likely to react to situations of (perceived) domestic unrest with diversionary uses of force? Let’s look at several types of domestic circumstances that make diversionary uses of force more attractive. The Severity of Domestic Unrest First, let’s examine the severity of domestic unrest. A straightforward linear hypothesis would posit that the greater the internal unrest, the greater the probability of diversionary force. As Jack Levy contends, during extremely high levels of internal conflict, elites might develop a “fortress mentality” and a propensity for high-stakes gambling. The stress that accompanies such internal strife may lead both to greater chances of misperception and to greater psycho-

206

Chapter 6

logical need to attain foreign policy successes, even if these entail enormous risks. 36 Facing almost certain losses if they do nothing, they might be willing to take monumental gambles to save their political skins. (This should sound familiar; the argument is consistent with the insights of prospect theory.) The possibility of losing a war doesn’t matter if the government expects to be replaced. 37 On the other hand (you knew this was coming), one might just as logically argue that once internal conflict becomes acute, it’s too late for diversionary action. Moreover, going to war is risky, given that the loyalties of the population and of the military are uncertain. And in a country with serious political fissures, an external war might lead to civil war or state collapse rather than to a rallying effect. Under these conditions, governments are more likely to focus on fighting the rebels than external foes—real or imaginary—as Blainey argues. 38 Finally, it might make sense that the relationship between internal and external conflict is curvilinear. Diversionary force is only likely in the middle ranges of internal unrest. When internal conflict is weak, governments have less incentive to react with risky international diversions, and they can either ignore the turmoil or use repression to quell it before it goes too far. When the domestic turmoil is severe, it is too dangerous to try diversionary war. Thus, it makes more sense to use diversionary forces at mid-levels of domestic conflict. 39 Empirical studies have been contradictory on this point. Russett’s analysis, using data for the years 1953–76, finds that democratic and nondemocratic states react differently to different levels of internal conflict. Democratic countries’ participation in MIDs is strongly and positively related to weak (nonviolent) domestic protest but not to violent rebellion. For nondemocratic countries, however, participation in MIDs was strongly related to both nonviolent protest and violent rebellion. 40 Two relatively recent studies provide some evidence that high levels of internal conflict are related to the initiation of external force. Graeme Davies analyzed data on pairs of states (dyads) from 1950 to 1982 to try to unravel the effects of violent and nonviolent strife, regime change, democracy, and contested institutions on a state’s decision to use military force or war. 41 He finds that violent strife increases the likelihood that states will initiate the use of force abroad, while nonviolent internal strife reduces that possibility. Presumably, states prefer to address nonviolent turmoil with repression or other domestic measures instead of scapegoating. A one standard deviation increase in violent strife increases the probability of initiating external force by 46 percent, while a similar increase in nonviolent strife decreases the likelihood of external use of force by 23 percent. 42 Christopher Gelpi looked at potential diversionary uses of force in the context of 180 international crises between 1948 and 1982. 43 He discovered that the effects of nonviolent protests and violent rebellions depend on the type of regime. While authoritarian states were unlikely to use diversionary force in crises in response to either type of unrest, democratic states were much more likely to respond by initiating diversionary force in reaction to both nonviolent protests and to instances of violent rebellion. The Role of Rivalries A second circumstance that may affect the use of diversionary force in response to domestic unrest is the presence (or absence) of longstanding rivalries with other states. 44 A particularly

The State Level of Analysis, Part II

207

interesting study by Mitchell and Prins suggests that states involved in enduring rivalries can more easily justify a resort to external force during periods of internal turmoil because the rival provides an easily identifiable scapegoat that might be legitimately blamed for internal problems. 45 Threats from rivals can be more easily and credibly exaggerated for domestic purposes. A rivalry allows political elites to “more easily manipulate foreign affairs to satisfy their own personal and/or political objectives.” 46 The presence of a rivalry creates an “opportunity-rich environment” that presents fewer problems for elites who wish to use diversionary policy. Some states—like the U.S., Israel, India, and Iran—live in international environments of almost constant “high opportunity” because they are embroiled in multiple ongoing rivalries. Mitchell and Prins examine dyadic data from 1960 to 2001. They use the consumer price index (CPI), which measures inflation, as their indirect indicator of domestic turmoil, and the initiation of militarized international disputes (MIDs) as their dependent variable. They find that rivals are much more likely to initiate the use of forceful actions when the inflation rate is high than are nonrival states. Economic weakness is dangerous for rival states but not for nonrivals. In opportunity-poor international environments, domestic turmoil (or at least high inflation rates) tends to discourage the initiation of MIDs. 47 They conclude that diversionary theory may apply best to states involved in enduring rivalries. The Opportunity for Internal Diversions A third circumstance that may affect the use of diversionary force with other states is the availability of internal targets. Given that most states live in neighborhoods in which there are few appropriate external targets for diversion—their neighbors are too strong or they have powerful allies or they are not enduring rivals—what are unpopular and politically vulnerable leaders to do? Tir and Jasinski suggest that since most states are multiethnic in composition, and thus have ethnic minorities who might effectively play the role of “enemy outsiders,” they are much more likely to use diversionary force against internal minorities than against external opponents. 48 (They hardly need the support of these groups, since they are probably not part of the “selectorate” anyway.) By using the tactic of domestic diversion, leaders not only minimize the risks and the costs involved in external warfare, they also neutralize whatever internal threat this group might pose. Painting unsympathetic minorities as disloyal and threatening by accusing them of collaboration with external foes is alarmingly easy; and exploiting internal interethnic tensions often works to rally ethnic in-groups around the leader. This is especially true if the minority is marginalized politically and economically and has a history of antagonism with the other groups that make up the country—like Chechens in Russia or Kurds in Turkey. (These are groups that Gurr refers to as “minorities at risk.” 49) In a study covering the years 1996–2002, Tir and Jasinski assessed whether political vulnerability (determined by low economic growth rates and an index of political unrest based on incidents of riots, strikes, and protests) was associated with the use of force against internal out-groups. They used data from the Minorities at Risk (MAR) project that records incidents of armed force by the government against minorities at risk. Tir and Jasinski find that their measures of leader vulnerability are associated with statistically significant increases in the likelihood of the government using force against minorities. Leaders at risk due to lagging

208

Chapter 6

economies and political unrest use diversionary force against minorities at risk. It also appears “they do not select their targets randomly but pick precisely those groups that are the most disliked, marginalized, and discriminated against.” 50 A single dampening factor seems important; states with high levels of economic development are less likely to resort to internal diversion, regardless of regime type. To the extent that American leaders use external diversionary force to address their vulnerabilities, this is somewhat unique. Its strong economy allows it to placate discontent with domestic programs, and its superior global capabilities allow it to use diversionary force against numerous foreign targets. Most state lack these advantages. Regime Type A fourth circumstance that may affect the use of diversionary force is regime type. It is still an open question as to whether autocratic regimes or democratic regimes (or democratizing regimes) are more susceptible to scapegoating. On the one hand, it might be argued that autocratic regimes make the best candidates for scapegoating. Political incumbents in autocracies are just as fearful of losing power as their counterparts in democracies and they are also less constrained in their ability to go to war. It is even possible that autocracies have stronger incentives to use diversionary force due to the inherent lack of political legitimacy of those types of regimes. 51 On the other hand, since autocracies are only minimally constrained in the use of repression at home, they will undoubtedly prefer to use direct internal action against opponents; it is cheaper than war and more effective in countering domestic opposition. 52 Other theorists argue that it is democracies who should be most prone to diversionary force. Democratic regimes are more dependent on popular support and more sensitive to fluctuations in that support and therefore perhaps more inclined to use foreign adventures to affect the domestic political situation. It may also be that since democracies have greater political and constitutional difficulty using repression at home to deal with domestic conflict, they may be more likely to address it through external action. (It’s their only real alternative.) Alastair Smith argues that the political dynamics of democratic states introduces a consistent bias toward external intervention and against acting timidly in international disputes. 53 On the other hand, as the monadic democratic peace theory argues, democracies are presumably constrained in their foreign policy behavior by both normative and constitutional rules. As we have seen, many political scientists have noticed what they believe is a propensity for the United States to become involved in militarized international disputes during election years, especially if this coincides with a period of economic stagnation. But the U.S. experience may be an anomaly. Leeds and Davis examine diversionary use of force for eighteen advanced industrial democracies for the period 1952–80. They find no consistent statistical support for a meaningful relationship between domestic political vulnerability (economic decline), the election cycle, and external aggression. 54 So, who is more likely to use diversionary force—democracies or autocracies? Ross Miller tackles this question head on. He uses expanded global economic data sets and MID data from 1820 to 1992 to explore whether democratic states are more or less likely than autocratic ones to use force when faced with declining levels of internal support. 55 He uses three measures of internal unrest: economic growth rates, levels of rebellion, and levels of protest. His dependent

The State Level of Analysis, Part II

209

variable, based on MID data, is an ordinal scale of the use of force—coded 0 if an actor did nothing or initiated only threats of force or displays of force, coded 1 if the actor initiated minor military force, and coded 2 if war was initiated. His results suggest that democracies and autocracies behave differently under conditions of declining internal support. Miller finds that for autocratic states, while the results are somewhat complex, there is no relationship between levels of protest and rebellion and dispute escalation. 56 He argues this is probably due to the fact that autocracies do not need to engage in external diversions because they are able to repress internal unrest. Regarding economic indicators, he finds that while relatively poor economic growth rates are related to higher levels of force for all initiating states, only autocratic states engage in diversionary behavior to any significant degree. Moreover, they do this by joining disputes already in progress rather than by initiating new disputes. This suggests that diversionary behavior is limited by the absence of strategic opportunities in the global environment: without the existence of an ongoing dispute, potential diverters have difficulty locating a suitable external target. 57 On the other hand, ongoing disputes create an opportunity for diversionary activity—at least for autocratic regimes. Democratic states appear to be much better behaved. Whether the indicator of internal unrest is economic growth rates, levels of rebellion, or levels of protest, there appears to be no relationship between internal unrest and dispute escalation by democratic states. Miller’s study does not allow him to examine why democracies should appear less prone to use high levels of force in diversionary behavior than autocracies. His guess is that they have fewer strategic opportunities to do so. Mitchell and Prins also find that the effect of rivalry depends on regime type. Once again, democratic regimes are the “good guys”—less likely to use diversionary action in response to worsening economic conditions (rising inflation) even if they are involved in rivalries. On the other hand, nondemocracies in opportunity-rich environments tend to initiate MIDs as inflation mounts, and their diversionary behavior is directed especially at their rivals. 58 Davies also finds that while all types of regimes tend to engage in diversionary use of force in response to violent internal conflict, democracies are less likely to use diversionary force than autocracies. 59 Gelpi’s investigation of ongoing international crises from 1948 to 1982 suggests a contrary dynamic. Like Miller, he finds that authoritarian states are unlikely to respond with external force to situations of domestic protest or rebellion. On the other hand, democracies do not appear to be constrained in this regard. Democratic regimes are more likely to respond to increasing levels of protest or rebellion with high levels of external force. 60 Finally, Pickering and Kisangani believe that simply classifying states as democratic or autocratic masks important differences within the categories. 61 Using a sample of 140 countries for the years 1950–96, they divide countries into five subcategories: mature democracies, consolidating democracies, transitional states, consolidating autocracies, and mature autocracies. Though their results do not conform very well to the “selectorate” model that they use to guide their research, they do produce some interesting findings. First, diversionary behavior (in this case the use of military intervention) is not practiced by all types of democracies or all types of autocracies. It appears to be common among both established/mature democracies and consolidating autocracies but not among consolidating democracies, mature autocracies,

210

Chapter 6

or transitional states. The authors reason that mature democracies and consolidating autocracies are the regimes whose leadership is most vulnerable to elite and mass pressures. Second, economic problems do not appear to trigger military interventions by either democracies or autocracies. Indeed, it appears that leaders are more likely to use diversionary force when economic growth rates are rising rather than falling. Third, mature democracies use force as a response to both elite unrest (the presence of government crises and purges) and to mass unrest (general strikes, antigovernment demonstrations, and riots). For these mature democracies, the year before elections is the hot part of the electoral cycle. On the other hand, mature autocracies do not use diversionary force in response to either elite or mass unrest. Their grip on power is so great that they have little trouble dealing with internal political challenges. Consolidating autocracies occasionally use external force as a reaction to elite unrest, but not when it grows too high. They do use diversionary force in response to mass unrests, but only when it is severe. The Diversionary Hypothesis: A Summation All right, then. What does the empirical research tell us about the diversionary hypothesis? Recent attempts to assess the validity of the diversionary hypothesis have lent modest support to the proposition that elites tend to use diversionary measures in response to internal difficulties, but the results are somewhat mixed and hardly robust. Many studies use economic indicators, such as economic growth rates, inflation, unemployment (or a combination of the two—the “misery index”), rather than actual domestic turmoil. (One should probably be wary of the use of indirect indicators where direct indicators exist.) Regime type seems to make a difference, but the patterns are both complex and inconsistent so that at this point we do not have a clear handle on how regime type affects diversionary behavior. On the other hand, the external situation, especially the presence of enduring rivalries, seems an important factor in determining which states are able to find appropriate targets for attempts at diversion. Ultimately, it is not entirely clear whether the theory of diversionary wars—applicable to a wide variety of countries and representing a general pattern—has much basis in reality, and if so what kinds of countries are most likely to use diversionary force and under what conditions. The support for the theory is certainly not very robust. Diversionary war appears to be a distinct path to war, but one that is very little traveled. Moreover, it looks suspiciously as if the diversionary war theory has been downgraded to “the diversionary behavior theory.” This is a less than satisfactory development since most states suffer some kind of domestic turmoil (loosely defined) and also engage in a broad range of external disputes. THE ROLE OF INTERNAL CONFLICT (2): CONTESTED INSTITUTIONS Kurt Dassel has put forward a version of the diversionary war theory that focuses on the domestic political concerns of the military in countries undergoing civil strife. 62 Frequently this strife is caused by transitions toward democracy (and thus the theory also has links to Mansfield and Snyder’s theory of democratization and war). Dassel argues that a specific kind of internal conflict is most likely to produce external aggression—political transitions characterized by contested institutions. 63 This condition exists when political institutions are in flux

The State Level of Analysis, Part II

211

and powerful groups advocate different sets of rules and constitutional forms. As a result, various political actors “refuse to play politics according to the rules of the game.” 64 Not all political transitions create contested institutions; this is an extreme situation. For instance, revolutions typically create a period of contested institutions. The situation Dassel is describing is a lawless “Wild West” scenario. Powerful groups favor different basic constitutional rules, and in the absence of agreement many groups resort to extra-constitutional means to maintain and increase their political power. Dassel quotes Samuel Huntington to explain what happens in this political free-for-all: “Each group employs means which reflect its peculiar nature and capabilities. The wealthy bribe; students riot; workers strike; mobs demonstrate; and the military coup.” 65 Under these conditions, the military—by definition—has a substantial degree of autonomy and is able to ignore civilian commands. On the other hand, civilian groups recognize the increased utility of force and attempt to gain coercive capabilities of their own. They challenge the autonomy of the military and attempt to control and perhaps even eliminate it. Civilians may construct their own, alternative military organizations—revolutionary brigades, palace guards, and militias. This constitutes a fundamental threat to the interests of military institutions. Like all groups in this fluid political environment, the military attempts to maintain and expand its power and organizational interests. Its interests include maintaining or increasing its budget and its resources, its weapons, its personnel, its unity, and its autonomy from civilian control and protecting its values and culture. Ultimately, the military is prepared to use force to protect these interests. Dassel argues that the military prefers to use force at home against its domestic enemies because that is where the primary threat is. However, repression may (under certain conditions) divide the military against itself; therefore the military may choose instead to promote diversionary wars abroad as a way of protecting its interests at home. In a situation of contested institutions, the military essentially has three options. It may choose to do nothing, but this will normally fail to protect its interests and may even lead to a revolt by officers willing to take matters into their own hands. The preferred option is to use domestic repression against the military’s domestic opponents, but this may split the military and it requires the support of at least some civilian political leaders. The third method is diversionary war, but this may not lead to the elimination of the military’s domestic opponents. It does, however, temporarily create a nationalistic rally-round-the-flag effect, diminish internal opposition to the military, and lead to greater military autonomy from civilian control. Whether the military chooses domestic repression or external war depends on (1) the breadth of civilian support for internal repression and (2) the nature of cleavages within the military itself. If the military is relatively homogeneous (in terms of political ideology, ethnicity, regional or clan affiliation, religion, and economic class) and has few cleavages, it may use force internally without risking a civil war in which some factions of the military end up fighting other factions of the military. All it needs is the support of some civilian politicians—otherwise it would pit the military against the rest of society. If these conditions prevail, we should expect the military to forgo using force abroad. However, when civilian support for military repression is lacking, the military may have to use force abroad to protect its organizational interests. This, Dassel argues, is the best explanation for the decision of the Argentine military to attack the Falklands in 1982. The Radical Party, previously allied with the military and

212

Chapter 6

supportive of repression at home, had been cut out of politics by the military junta. It therefore allied with the opposition, leaving the military without the support of a major political party. A second possibility is that the military is heterogeneous and has overlapping and reinforcing cleavages. In this case, it is probably irrevocably split. Rather than a single military, there are opposing factions who see each other as the enemy. Under these conditions, the military will not attempt to provoke a foreign conflict but is more likely to become embroiled in internal conflict. (This was the situation in the former Yugoslav National Army, which broke along ethnic lines, with the old Yugoslav army essentially becoming a Serb army.) Finally, some militaries may be fairly heterogeneous, but with cleavages that are crosscutting. In this case, the military’s splits are not clean; factions may be split on some dimensions (say regional loyalty and ethnicity) but are united on another dimension (say religion). The military is fractured, but it is not so divided that it is impossible to take united action. The problem with heterogeneous militaries is that if the military uses force internally, this entails some groups oppressing others. It may set one military faction against another, creating the potential for civil war—which virtually all factions wish to avoid. Under these circumstances, the military can embark on domestic repression against its opponents only with very wide civilian support. Otherwise, it will choose diversionary war abroad. Pre–World War II Japan is perhaps the best example of this. 66 In the 1920s Japan entered into a transitional period referred to as the Taisho democracy. Western democratic institutions—political parties, parliamentary elections (and a cabinet based in part on these elections), a constitution, and a vigorous (though not impartial) press—were pasted onto a political system that retained a powerful emperor and court. With the ascension of a weak emperor— the ineffectual Yoshihito—all political institutions were embroiled in a contest over political power and rules. A central concern of many liberal politicians was how to curtail the power of the military in Japanese politics. The military itself understood this and was extremely active in preventing any attempt to undermine its well-established political clout. The military was highly autonomous and greatly politicized; in the 1920s and 1930s there were numerous political assassinations (even of prime ministers) and threatened coup attempts. However, the military was heterogeneous, with cleavages based on ideology, regional and clan affiliation, service branch, and officer age cohorts. Moreover, the Japanese military was broadly representative of Japan as a whole, and therefore any internal use of force by one military faction against a particular domestic opposition group threatened to divide the military against itself. This was, in fact, the lesson of the 1936 Tokyo Rebellion, which pitted several different contingents of military forces against each other in the heart of the capital city. The rebellion served as a wake-up call for the military, and fear of civil war was a primary factor in nudging the military toward external aggression. Moreover, the resort to internal repression (including a military coup in the name of the emperor himself) was adamantly opposed by the emperor and therefore impossible, as many officers were fanatically loyal to him. The one thing that the military could agree on which could maintain the power and organizational interests of the military was the use of force abroad. From this perspective, Japanese military aggression against Manchuria and China was largely a domestic political response by the Japanese military to protect its interests in the political environment of contested institutions. In the Manchurian crisis of 1931, the Shanghai incident of 1932 and the subsequent invasion of Jehol,

The State Level of Analysis, Part II

213

and the China incident of 1937, autonomous actions by Japanese military forces were very much out in front of official government policy emanating from Tokyo. Dassel does not argue that the military is always the primary instigator of the diversionary war; the dominant partner in Indonesia’s aggression against Malaysia in the 1960s was President Sukarno rather than the military. However, he argues that while the civilians in the coalition may influence how the military uses its force, they do not exercise control over whether the military will use it. Moreover, civilian governments are extremely unlikely to pursue aggression abroad without the approval of the military. Dassel and Reinhardt tested the contested institutions theory on data for 107 countries from 1827 to 1982. 67 They were interested in whether contested institutions or other domestic variables could account for violence in general (internal or external) but most specifically the initiation of external force or war (as indicated by events coded as 4 or 5 on the MID data). Their operational indicator for contested institutions was a version of the polity II data set’s “major abrupt polity change” variable, which included cases of significant changes in political arrangements that were the subject of struggle among domestic groups within a country. Their results confirmed that the existence of contested institutions increases the probability of both internal violence and international violence. Moreover, other domestic variables such as regime type, regime change, democratization, internal violent strife, and nonviolent strife are insignificant predictors of violence once contested institutions are included in the analysis. They conclude that domestic turmoil in general is insufficient to create diversionary violence; the turmoil must be of a nature that threatens the interests of the military and this only occurs in a situation of contested institutions. Other researchers have recently begun to try to replicate Dassel and Reinhardt’s findings. Davies investigates the initiation of the use of force or war (MID level 4 or 5) in his analysis of dyadic data from 1950 to 1982. 68 He finds a strong positive correlation between the existence of contested institutions and the initiation of armed force. But unlike Dassel and Reinhardt, he finds that violent domestic strife is related to external use of force even in the absence of contested institutions. THE ROLE OF INTERNAL CONFLICT (3): “KICK ’EM WHILE THEY’RE DOWN” WARS Let’s return to Blainey’s discovery of a historical connection between internal conflicts and subsequent wars. Blainey thinks that the diversionary explanation for this phenomenon is unproven and probably wrong. 69 He proposes that another theory is supported more abundantly by the facts and provides a more logical explanation of the relationship between internal and external conflict. In his list of thirty-one wars in which civil unrest preceded external conflict, war was rarely initiated by the strife-torn state. Instead, most of the wars were initiated by others, with the internally troubled state in the role of the victim. Blainey argues effectively that states do not generally start wars to quell their own internal revolts. Instead, wars occur because internal conflicts change the balance of power between states. For any pair of countries, if internal strife occurs in the stronger country in the pair, this lowers its margin of superiority and opens a window of opportunity that might tempt the other

214

Chapter 6

state to strike at an opportune moment. The existence of serious unrest in the rival state exposes its weaknesses and diverts its resources away from defensive efforts. For the potential attacker (most likely a contiguous neighbor), it lowers the expected costs of war and increases the expected benefits. 70 On the other hand, if the internal conflict occurs in the weaker state in the pair, peace is more likely to be preserved. The internal strife simply confirms the assessment of inferiority; the balance of power is not altered. 71 This could be described as the “kick ’em while they’re down” theory of war. (Blainey uses the term “death watch wars” to describe situations in which foreign leaders sit—rather like vultures in trees—waiting for monarchs in the target state to die and bring on political uncertainty.) Iraq’s decision to attack Iran during the domestic upheavals associated with that country’s revolution can certainly be placed in this category of opportunistic attacks. Under normal circumstances Iraq could not seriously contemplate winning a major military confrontation with Iran, given the latter’s considerably larger population, armed forces, and industrial might. The Iranian revolution temporarily changed this equation. 72 It should be noted here that this theory asserts that domestic conflict in state A provides the opportunity or an excuse for state B to attack, but it does not address the underlying cause of state B’s attack. Presumably, the underlying cause predates the internal conflict in the target and is unrelated to the internal conflict itself. In other words, this theory doesn’t help explain why B attacks, only why B attacks now. The kick-’em-while-they’re-down theory makes intuitive sense. Governments with internal conflicts aren’t likely to attack foreign states; instead, they attack the rebels within. If the disorder isn’t serious, the government doesn’t need to seek war with outsiders. If the situation is extreme, government leaders will be much more inclined to seek peaceful relations externally in order to devote their attention and their resources to internal problems. Crucially, serious internal dissent reduces the political reliability of the military as well as the internal political cohesion necessary to pursue an external war. Indeed, as Blainey notes, most states involved in war at the same time as they were beset by unrest at home have been anxious to seek external peace: Russia in 1905 and 1917, Germany at the end of World War I, and the United States during the later stages of the Vietnam War. 73 The kick-’em-while-they’re-down theory certainly is not an explanation that can be widely applied; only a small percentage of all wars fall into this category. After all, the number of states experiencing high levels of civil unrest in a given year is small and relatively few of these states are invaded by their neighbors. A condition of weakness brought on by internal strife that brings about a change in the balance of power is not a sufficient condition for war; other conditions are probably also necessary. For instance, it helps that the weakened state lives in a dangerous neighborhood containing a rival state that is normally weaker and has old grievances perceived to be amenable only to solution by force. This was certainly the case in the Iran-Iraq War. What we can say is that for the small number of cases that fit this description, the presence of internal conflict is a critical factor in bringing about the war.

The State Level of Analysis, Part II

215

THE ROLE OF INTERNAL CONFLICT (4): REVOLUTION A fascinating study by Zeev Maoz suggests another relationship between internal and external conflict: the link between war and two kinds of revolutionary changes—the birth of new states out of revolutionary struggles and the revolutionary transformation of older political systems. 74 Revolutions are violent, mass political conflicts over fundamental issues in a society. They are more than just a struggle for control of the government (and thus more than civil wars). They are attempts to tear down the institutions of the old regime and replace them with a totally different set of institutions. As you can imagine, then, real revolutions are relatively rare events, but they have significant consequences. States emerging from revolution or midlife transformations are likely to receive a chilly welcome from their peers, and political elites within these revolutionary states may perceive the international environment to be hostile. Political elites in the surrounding established states may perceive (often correctly) the goals and ambitions of these new states as a threat to themselves and to the current international order. Thus, the political transformation of states through revolutionary means creates mutual distrust between old and new states in the system, which can lead to violent conflict. Since this mistrust is mutual, the initiator of the war may be either the newly transformed state (as in revolutionary France’s declaration of war against Austria and Prussia in 1792) or members of the old order (as in Iraq’s attack on revolutionary Iran in 1980). Using data on militarized interstate disputes from 1816 to 1976, Maoz confirmed that both new and old revolutionary states were involved in a larger number of militarized disputes (uses of force, displays of force, and threats of force) than new or old states whose political development was more evolutionary. Revolutionary change within states appears to increase the likelihood that they will experience subsequent militarized conflicts—either as initiators or as victims. 75 Perhaps the most comprehensive study of the link between revolutions and war is Stephen Walt’s Revolution and War. 76 For Walt, revolutions cause several intersecting phenomena that combine to increase the probability of war. First, revolutions cause an abrupt shift in the distribution of power between states. The internal chaos, violence, and economic dislocations decrease (temporarily) the aggregate power of the state. However, just how much is never certain. Revolutions make it difficult to accurately measure the balance of power and therefore to predict the outcome of war. Alliance commitments and other international agreements are also thrown into jeopardy by the existence of a new government. Thus, revolutions increase uncertainty in the international system, making miscalculation and misperception more likely. Neighbors and rivals may try to measure whether the revolutionary state’s military capacity has been reduced by the chaos and disruption or rather enhanced by the new state’s ability to mobilize the population and the resources of the state. Second, revolutions alter assessments of the balance of threats. Walt argues that revolutions increase the level of threats on both sides. This is because the ability of each side to accurately assess the other’s intentions is reduced, in large part because the revolutionary state has no track record upon which other states can judge its intentions. The likely result is to rely on ideological filters and worst-case scenarios to exaggerate the threat posed by the other.

216

Chapter 6

External states are likely to see the revisionist ideas and goals of the revolutionary state as a threat to the prevailing international order. They are likely to perceive the revolutionary state as bent on subverting the regimes of its neighbors and on exporting its revolution to others. On the other hand, the revolutionary state is likely to perceive a threat of counter-revolution. Both exaggerate the other’s hostility, and both exaggerate their own vulnerability. Both are likely to commit actions that contribute to a spiral of suspicion, and both see war as an effective way to eliminate the threat. Somewhat paradoxically, both also tend to be optimistic. Revolutionaries tend to see their opponents as paper tigers whose own populations may revolt against them, and they may at times overestimate their own capabilities. Status quo states see revolutionary states as (at least temporarily) weak. Finally, changes in the balance of power and in the balance of threats create the possibility for security dilemmas and intensify security competition. 77 Third, the level of threat is affected by new estimates of the offense-defense balance. The general argument of offense-defense balance theory (which we will discuss in a later chapter) is that war is more likely when offense is easy; it is less likely when defense is easy. The military balance is not just a function of objective factors like military technology or geography. What governments believe also plays a role. Revolutionary states believe that victory will be easy because revolutionary ideology will rally the masses in the revolutionary state and undermine the governments of one’s enemies. Opponents believe victory will be easy because of the chaos of revolution. Both sides feel that they have the offensive advantage, but when offenses are presumed to be dominant, both sides simultaneously fear for their safety and see their opponents as vulnerable. This encourages wars of preemption or prevention. Striking first is more appealing. Fourth, revolutions damage the normal channels of communication, diplomacy, and intelligence just when states need them the most. Revolutionary governments lack trained diplomatic personnel. Moreover, former elite exiles may imply to their new host states that the revolutionary regime is weak and can be easily overthrown (this is what many Cuban, Nicaraguan, and Iraqi exiles told U.S. officials, and it is what many Iranian exiles told Iraqi government officials). Both phenomena lead to a greater than normal chance of misperception and miscalculation. The bottom line for Walt is that while revolutionary states can initiate wars, this is rare. They are more likely to be the targets of other states’ aggression. 78 (This accords with Blainey’s general views.) He argues that revolutionary states are rarely ready for war; instead, their awareness of their weakness causes them to act cautiously. Revolutionary states are seen simultaneously as a threat by others and as being temporarily weakened, thus making them an inviting target. Walt’s list of ten cases of revolution between 1789 and 1979 provides some evidence for this. If we confine our list of revolutions to cases characterized by (1) significant violence, excluding the relatively nonviolent “refolutions” in the former USSR and the former Soviet bloc from 1989 to 1991; and (2) mass revolutions “from below” rather than elite revolutions “from above” such as Ataturk’s Turkish revolution; and (3) exclude the American Revolution as being more properly classified as a war of independence or national liberation, then we get

The State Level of Analysis, Part II

217

the following list of nine revolutions: France (1789), Mexico (1910), Russia (1917), China (1949), Cuba (1959), Ethiopia (1974), Cambodia (1975), Iran (1979), and Nicaragua (1979). Six of the nine revolutions led to wars within five years; the other three were near-misses that resulted in militarized disputes short of full-scale war. Of the six cases in which revolution led to war, there are five cases in which the revolutionary state was the target rather than the aggressor: Russia suffers Allied intervention (1919) and an attack by Poland (1920), though one could make the argument that Russia was as much to blame for the war as Poland; Ethiopia is the victim of an invasion by Somalia (1978); Cambodia is the target of a Vietnamese invasion (1979); and Iran is the victim of Iraqi aggression (1980). There are two relatively clear examples of revolutionary states initiating aggression: France’s war against Austria and Prussia and their allies in the First Coalition (1792–1793) and China’s relatively reluctant entry into the Korean War (1950). China also invaded Tibet in 1950, but how one characterizes this use of force is somewhat problematic. The near-misses include three military actions initiated by the U.S.: the ill-fated Bay of Pigs invasion of Cuba in 1961; the limited military intervention in Mexico in 1914 and 1916; and support of the Contra rebels and mining of harbors in Nicaragua. 79 The classic case that illustrates Walt’s thesis is the Iran-Iraq War of 1980–88. Iran and Iraq have a long history of rivalry and hostility. 80 In the post–World War II period, Iran—with a much larger population and military than Iraq—was clearly the dominant power in the Persian Gulf region. Indeed, Iraq was forced to accept the somewhat humiliating terms of the 1975 Algiers Accord in order to forestall what might have been an all-out war between the two states. All of this changed with the revolution in Iran that overthrew the Shah in 1979 and brought Ayatollah Khomeini to power at the head of an Islamic state. Khomeini saw Islamic government as the only legitimate type of regime, and revolutionary Iran stood as a direct challenge to the secular rule of most states in the region. Moreover, Iran was bent on exporting its brand of Shi’a fundamentalism to neighboring states—especially those that had significant Shi’a populations, like Iraq, Saudi Arabia, Kuwait, and Bahrain. Indeed, to some extent, Iranian revolutionaries believed that success at home required expansion abroad. Khomeini himself declared, “We will export our revolution to the four corners of the world because our revolution is Islamic.” 81 Iraq was a special target of interest for Iranian revolutionaries because Shi’a were the largest ethnoreligious group in Iraq and Saddam Hussein’s largely Sunni government had actively repressed them over the years. Iranian leaders launched a rhetorical assault on Iraq and its Ba’athist party government, increased support for Shi’a groups in Iraq, and rejected Iraqi demands that Iran turn over to Iraq lands promised in the Algiers Accords. Iraqi leaders clearly perceived a threat from revolutionary Iran, which it saw as intent on breaking Iraq up into ethnoreligious fragments: a Shi’a south connected to Iran, a Sunni center, and a Kurdish north. The fact that the Islamic revolution in Iran threatened others as well led to a decision by Kuwait and Saudi Arabia to financially back Iraqi military operations against Iran. Saddam Hussein’s government also saw significant weakness in Iran and therefore a window of opportunity in which military action might be successful. The once-powerful Iranian military appeared to be in shambles: U.S military support had ended; the revolutionary government had decimated the old officer corps; deep budget cuts had reduced the military to

218

Chapter 6

about one half its previous size; the military lacked parts and training; the desertion rate was sky-high; and the loyalty of the armed forces was questionable, as illustrated by several military coup attempts. Moreover, coup attempts, assassinations, and fighting in the streets continued long after Khomeini came to power. Thus, Iraqi leaders perceived that the military balance had changed from one of decided Iraqi inferiority to one of at least parity; a window of opportunity had opened. Consistent with Walt’s theory, perceptions of the balance of power and the balance of threats changed dramatically as a result of the Iranian revolution; leaders in both Iran and Iraq felt vulnerable to the other; a spiral of mutual suspicion was launched; the aggressive ideology of the fundamentalist revolution in Iran meant that the balance between offensive and defensive forces favored the offense; misperceptions and miscalculations were plentiful; and a security crisis (especially for Iraq) enhanced the motivation for a preventive war, especially since a window of opportunity was perceived to exist due to a temporary Iranian weakness. THE ROLE OF INTERNAL CONFLICT (5): OUTSIDE INTERVENTION IN INTERNAL CONFLICTS When civil conflicts break out—typically between a government and one or more rebel groups—each side is likely to have external “friends” who might be willing to give assistance in order to affect the political outcome of the struggle. States may thus intervene militarily in the internal conflicts of others—rebellions, civil wars, secessionary wars, and revolutions. (They may also intervene indirectly by providing training and/or by sending military, economic, intelligence, and logistical assistance to one side in the contest.) Internal conflicts may create security threats to other states who may seek to support the government, reduce instability, protect endangered groups, protect economic interests, or prevent the coming to power of dangerous political groups. Alternatively, the internal conflict may provide an opportunity for rival states who wish to weaken or aid the overthrow of the government in power or perhaps even gain territory and resources at its expense. A persistent dynamic in civil wars is that governments under siege frequently call on allied states to assist them internally against the rebels. It is just as likely, however, that rebel groups have also forged strong ties with foreign governments to assist their revolt. Such ties between rebels and foreign governments were present in twenty-six of the thirty-one wars studied by Blainey. 82 Several interrelated factors might increase the potential for intervention. 83 First, those most likely to intervene are contiguous states and great powers. These are the states most likely to have the opportunity and willingness to intervene and most likely to feel themselves affected by a civil conflict in another state. And, of course, rebels in the internal conflict in one state often seek refuge (and staging areas) inside the territory of a neighboring state. Second, potential interveners may have some important transnational links to the target’s government or citizenry through ideological affinity, economic interdependence, or security alliances. The most important links are probably based on ethno-linguistic-cultural affinities, and these links are especially likely in contiguous states; for instance, in the long Lebanese civil war, Israel was allied with Maronite Christians factions, while Syria allied itself with various Muslim

The State Level of Analysis, Part II

219

groups, including Palestinians in Lebanon. Third, as discussed earlier, the existence of rivalries probably increases the chances for intervention. Fourth, intervention by one side’s friends leads to counterbalancing by the other side’s friends. This is what Vasquez calls “pure contagion.” 84 Once one state intervenes—whether a rival or not—the potential for interventions by others increases. Fifth, intervention may be motivated by external considerations of security or by domestic political considerations, including humanitarian considerations—especially if there is large-scale slaughter of innocents and the creation of refugees. Several of the factors listed above may interact with each other. A partial list of direct international military interventions will be familiar to many. • Russian Civil War (1918–21): intervention by Great Britain, France, Japan, and the United States • Spanish Civil War (1936–39): intervention by forces from Germany, Italy, USSR, as well as France, Ireland, Portugal, the United States, and others. Many were retired military who were formally called volunteers • Angolan Civil War (1975–91): intervention by Cuba, Zaire, and South Africa • Chad Civil War (1965–79): intervention by France to support the government and by Libya to support the rebels • South Vietnam Civil War (1955–75): intervention by the U.S. and North Vietnam and their supporters • Pakistani Civil War (1971): Indian intervention • Afghanistan Civil War (1979–89): Russian intervention • Lebanese Civil War (1975–90): interventions by Syria and Israel (and the PLO) • Zaire/Democratic Republic of Congo Civil War (1996–2004): interventions by Rwanda, Uganda, Angola, Zimbabwe, and Namibia A more nuanced data-based assessment of the internal-external conflict relationship has been carried out by Gleditsch, Salehyan, and Schultz. 85 Their study investigates whether serious internal conflict (civil wars) increases the probability that countries will engage in disputes involving the threat of force or use of force. They combine data on MIDs with three sets of data on civil wars covering the 1946–97 period. The data show that a substantial portion of all MIDs took place in the presence of a civil war in at least one of the participants. The percentage ranged from 19.3 of MIDs in the COW civil war set to 51.9 percent in the Uppsala civil conflict list. 86 They then set out to determine how the existence of a civil war affects the probability a MID will occur. A cross-tabulation shows that MIDs are about twice as likely to occur in dyad years with civil wars than in dyad years without civil wars. 87 A multivariate analysis using various control variables yields strong results as well. They find that the presence of a civil war increased the probability of a MID in a dyad by between 50 percent and 80 percent over the baseline predicted probability depending on the civil war data set used. Moreover, the increased probability of a MID due to the existence of a civil war was roughly equal to the decreased probability of a MID due to joint democracy. Thus, there is a “civil peace” due to the absence of civil war that is just as strong as the “democratic peace” due to joint democracy. While the focus of the research is on the link between civil wars and the initiation of militarized disputes rather than war per se, the authors report that when they use

220

Chapter 6

only MIDs involving the use of force or MIDs resulting in casualties, the results are essentially the same. 88 Most important for this discussion, the researchers scrupulously investigated all 431 MIDs whose onset occurred during a civil war and coded them as belonging to various categories: (1) direct intervention by an outside state to assist the rebels; (2) externalization by the state undergoing civil war in order to attack rebels in a neighboring state, coerce neighbors to end support of rebels, or to deter such support; (3) spillovers in which acts targeted against rebels unintentionally spill over onto another state; (4) potential opportunism in which an outside state initiates military action against the state undergoing civil war, but the act appears unrelated to the civil war; (5) potential diversion in which the state undergoing civil war initiates a MID that does not seem related to the civil war; and (6) counter-intervention in which the two states involved in the MID intervene on opposite sides in a third state’s civil war. Just under half of the MID dyads appear to be cases of opportunism (the kick-’em-whilethey’re-down phenomena) or diversion. On the other hand over 40 percent of the MID dyads represent direct interventions, externalizations, or spillovers—those categories in which the actions taken are intended to have a direct effect on the civil war itself. Gleditsch argues that the acts coded as interventions, externalizations, or spillovers could only happen if there was a civil war, while the acts coded as potential diversions or opportunisms might have happened in the absence of a civil war. Thus, it is most likely that the “surplus” of MIDs—above and beyond what one would normally expect in the absence of civil wars—is accounted for by MIDs that flow directly from the interest of states in affecting the outcome of the civil war. 89 They conclude that the increased risk of MIDs during civil wars is not due to opportunism or diversionary motives but an attempt to affect the outcome of the civil war itself. THE ROLE OF INTERNAL CONFLICT: SUMMING UP AND IMPLICATIONS While early empirical studies seemed to indicate that internal conflict is not very strongly related to external conflict, this appears to run counter to both common sense and historical example. While not all wars are preceded by internal unrest and not all domestic conflicts result in war, still there are enough real-world examples of these for the theorist to be able to conclude that there is something important here. The early discrepancy between the historical evidence and cross-national evidence compiled by political scientists is most likely due to certain methodological peculiarities of the latter’s research. While a relationship probably exists between internal and external conflict, it is obscured somewhat due to the fact that several different causal mechanisms link the independent variable to the dependent variable. Through the diversionary mechanism, states torn by certain levels and types of civil strife seek foreign conflicts to solve domestic problems; through the kick-’em-while-they’re-down mechanism, states experiencing serious internal weakness or conflict are attacked as easy prey by their opponents; through the internationalization of civil wars, rebels and governments alike may attain external allies and turn domestic conflicts into international wars; and finally, through the process by which revolutionary regimes are brought into existence, interstate conflict may be precipitated between the supporters of the old international order and the new revolutionary states. 90

The State Level of Analysis, Part II

221

The existence of a direct relationship between internal and external conflict implies that the world would be a much more peaceful place if states themselves were more peaceful places to live. International conflict could be substantially reduced if internal conflicts could be reduced or eliminated. This, of course, raises the question of just how internal conflicts might be diminished and whether the world community can bring about a reduction of internal conflict in turbulent states through policies and programs applied from the outside or whether this is a matter that can only be addressed internally. To some extent, U.S. foreign policy in the post–World War II era has been predicated on just such a presumed link between the reduction of internal conflict and prevention of external conflict. The general purpose of American foreign aid—whether Marshall Plan aid to Western Europe, Alliance for Progress aid to Latin America, Caribbean Basin aid, or assistance to Iraq and Afghanistan—has been to prevent internal disturbances in the receiving countries that might make those states fertile grounds for takeover by radical groups or attractive targets for aggression. The rationale for foreign aid programs has been that economic unrest leads to internal political unrest, which leads to external subversion or assault. There is, of course, a second set of assumptions as well: that economic aid will bring about economic growth and development, that such growth will result in stable social and political conditions, and that this will prevent the rise of strong, left-wing (or right-wing) radical groups. Quite a few social scientists, however, would argue that economic growth and the process of development— especially if the growth is rapid—bring about social and political instability rather than stability. Development aid might be instrumental in fostering the very conditions of instability that the aid was designed to prevent. Since the resolution of this debate is beyond the scope of this inquiry, let us agree to leave this question unresolved and move on to other theories of war at the nation-state level. THE ROLE OF NATIONALISM The concept of nationalism has been noticeably absent from mainstream empirical studies on the causes of war until recently. This is due to several reasons. First, the concept is difficult to define and even more difficult to measure. There is no international data base that measures the degree of nationalism in a particular political community at a given point of time. Second, nationalism is a very slippery concept, in large part because the term “nation” is itself conceptually fuzzy. It is most appropriately used to refer to a group of people who share a common identity, a common culture, a common heritage and history, a common language, a common religion, and a desire to govern themselves. Thus, the Kurds are a nation, as are Basques, the Dutch, Norwegians, Chechens, Russians, and Palestinians. However, the term is often used interchangeably (and, I would argue, inappropriately) with the term “state”—a political and legal entity, such as Iraq, the Russian Federation, Honduras, Spain, and Nigeria. Since the term “nation” appears to have two meanings, so does the term “nationalism.” One way to define it is the identification of a people with (and their loyalty to) their state. In this sense, nationalism may be seen as a collective attribute of a state and its people. This is what we might call type A (or state-centric) nationalism. On the other hand, nationalism may also be a characteristic of various substate communities (nations). In this sense it refers to the

222

Chapter 6

identification of a people with (and loyalty to) their nation. This is what might be called type B (or nation-centric) nationalism. Thus, we can talk about the loyalty of Iraqi citizens to the state of Iraq (type A nationalism), but we can also talk about the loyalty of substate groups within Iraq such as Iraqi Sunnis and Iraqi Kurds to their particular national communities (type B nationalism). In true nation-states like Japan or Finland, the population is fairly homogeneous and there is little distinction between nation and state. However, the vast majority of states in the international system are culturally heterogeneous, and the difference may have crucial implications for internal and external security. Type A (State-Centric) Nationalism and War Given the underdeveloped state of theorizing about the relationship between nationalism and war, the following statements should be taken as empirical propositions. Type A nationalism (fierce or intense loyalty to one’s state) is not a necessary condition for war. Wars existed long before the people within a state’s boundaries had a strong sense of identity with the state. Well into the eighteenth century wars were mostly fought by mercenaries whose motivation was not loyalty to a state, but monetary gain. Nor is type A nationalism a sufficient condition for war. It is quite possible for a people to have strong nationalist feelings and not engage in war. In states like Sweden, their avoidance of military action may even be an integral part of their national identity. Ultimately, type A nationalism is almost never the primary cause of war, but it may help to create an internal political environment conducive to war in several ways. First, intense feelings of nationalism are typically strongly ethnocentric, leading to the inflated evaluations of one’s own state and negatively distorted and stereotyped images of other states. These distorted images of self and others may produce miscalculations that play a role in decisions for war. Second, type A nationalism may be related to interstate grievances in which national pride is involved—issues of status and prestige and international role. To the extent that political elites and masses believe that other states have disparaged their country or that others see their country as racially or culturally inferior, nationalism may serve as a catalyst to rally a country toward war. Third, insecure political elites who perceive themselves to be vulnerable to internal political opposition forces may attempt to mobilize mass public support for themselves by using the language, themes, and symbols of nationalism to create greater unity and undercut their opponents. They may emphasize in-group/out-group differences to establish the basis for a diversionary or scapegoat war against a neighboring state. And we know that aggressive actions by governments do generally induce a rally-round-the-flag effect that elites are likely to manipulate. (However, as we have seen, vulnerable leaders are just as likely to target domestic minorities as a diversionary tactic.) Fourth, during wars, people are urged to love their own countries and hate the enemy. The intensity of type A nationalism is normally increased during wartime. This makes the resolution of the conflict more difficult, and nationalist feelings last well after the war’s end, poisoning future relations between states. Thus type A nationalism is both a factor that contributes to war and is also an effect of war.

The State Level of Analysis, Part II

223

Early twentieth-century Japan provides an example of how an intense, institutionalized type A nationalism may contribute to the propensity for war. 91 The enthronement of Emperor Hirohito in 1926 initiated a conscious effort by the state to construct a new “emperor cult” as the basis for a new nationalism, called kodo, or “imperial way.” The Education Ministry and other government agencies propagated an elaborate mythic history of the Japanese state that gave the emperor a divine origin and attributed racial, cultural, and moral purity to the Japanese people. Japanese nationalism incorporated desires for outward expansion coupled with a belief that Japan had a historic mission to expand into East Asia and lead a new Asian renaissance. One aspect of that new Japanese nationalism was a belief in the creative and constructive force of war and a mystical belief in an apocalyptic final victory. Since the emperor was a living god and the state was the incarnation of morality and justice, Japan only engaged in “just wars.” Wars became sacred struggles or divine missions to unify and purify the world, creating a new world order under the benevolent rule of the emperor. The new nationalism also incorporated the old bushido code of the samurai warriors of the past. The samurai code, like many honor codes, glorified loyalty, obedience to superiors, military discipline, and honorable death as the highest personal values, and it provided cultural support for a belligerent, zero-sum approach to interstate relations. After 1925, the national system of education became militarized, deliberately spreading the samurai ethic. Japanese nationalism was also linked to ideas of racial superiority of the Japanese people. The nationalism of the period had much in common with European fascism: deification of the national racial community and its embodiment in a mystical cult figure, militarism and the glorification of war, political dictatorship, the idea of national mission, emphasis on moral regeneration and national spirit, and propaganda about national essences. Nationalism did not cause Japanese aggression in the Far East, but it did contribute to a climate in which military force became the primary tool by which Japan pursued its international objectives. In the final analysis, nationalism (along with militarism and racism) was part of the belief system of the Japanese leadership and colored their views of the world and Japan’s role in it. Kodo was part and parcel of the perceptual lens through which the Japanese political and military elite viewed the world. Type B (Nation-Centric) Nationalism and War Type B nationalism refers to a group’s identification with, and loyalty to, the primary cultural group (nation) to which it belongs. Once again, let us start with some general empirical propositions that appear to be true, but that will require much more intensive investigation than has as yet been done. Type B nationalism frequently intermingles with type A nationalism—especially in homogeneous states like Japan. It is often difficult to determine which type of nationalism is the dominant factor behind aggression. For instance, when Somalia invaded Ethiopia’s Ogaden region, was the war driven by nationalistic fervor tied to the Somali state or by nationalistic fervor tied to Somali ethnic unity? In many states, however, various nations (again, in the sense of ethnic or cultural communities) coexist. Type B nationalism is clearly a contributing factor to wars of secession in which minority nations attempt to break away from the larger state—which is usually in the hands of other national communities. These wars of secession may then become international-

224

Chapter 6

ized—the Bangladesh War of 1971 and the various wars of Yugoslav secession in the 1990s, for example. Type B nationalism is also present in a category of war variously called wars of independence, wars of national liberation, or anticolonial wars. Once again, these internal wars against a foreign ruler may become internationalized. The Russo-Polish War (1919–21, in which the Poles were aided by the French) and the first phase of the war in Vietnam (the French-Indochina War, 1945–54) are examples. To the extent that Muslims in Kashmir wish to become independent from India, Pakistan has been willing to assist in this struggle (and perhaps to incorporate Indian-controlled Kashmir into the Pakistani state). Several wars have been fought between India and Pakistan over Kashmir. Type B nationalism is frequently linked to irredentist wars to regain “lost” territory— territory believed to belong to one’s own state (and/or nation) but in the hands of another. It is also an important factor in wars of national consolidation or national unification in which states attempt to gather together all the land in which members of the nation live. Examples include the Ogaden War (1977–78) in which Somalia attempted to annex the Ogaden region of Ethiopia, a land largely populated by ethnic Somalis. The Korean War (1950–53) and the early phases of World War II in Europe also had aspects of national consolidation. In addition, the desire for a “Greater Serbia” or for a “South Slav Confederation” centered around Serbia was seen as a threat to the Austro-Hungarian Empire and was one of the central problems in the period before World War I. In the post–World War II era, once Yugoslavia fragmented as a state, Serbia’s conflicts with Croatia and Bosnia-Herzegovina centered around an attempt by Serbia to incorporate territories of those newly independent states in which Serbs were a local majority. The policy of the Croatian state toward the ethnic Croat diaspora was similar. Paul Huth’s study of 129 territorial disputes between 1950 and 1990 reveals that 51 percent of the disputes involved either the protection of national minorities living in the disputed territory or issues of national unification. He finds that those territorial disputes that involved ethnic irredentism or national unification were most likely to escalate to high levels of political or military conflict. In the first situation—which he calls irredentism—minority peoples within the target state share a common culture or language with the dominant population in the challenger state (e.g., ethnic Somalis living across the border in Ethiopia, Muslims living across the Pakistani border in Indian administered Kashmir, ethnic Serbs living across the borders in Bosnia or Croatia). In the second situation—national unification—the general population living in both countries is essentially the same (Koreans living in North Korea and South Korea, Vietnamese living in North Vietnam and South Vietnam, Arabs living in Iraq and Kuwait). 92 One of the few theoretical pieces on the connection between nationalism and war is Stephen Van Evera’s classic article “Hypotheses on Nationalism and War.” 93 As the title suggests, Van Evera has not constructed a theory about how nationalism causes war; rather, he is engaged in one of the crucial early steps in theory building—the construction of testable empirical hypotheses. His hypotheses link nationalism to both internal war and interstate war. His focus is on stateless nations—ethno-religious communities that do not have a state of their own, but live (as minorities) within the boundaries of a state dominated by other national groups. Indeed, his narrow definition of nationalism is that it is a political movement whose members give their primary loyalty to their own ethnic community and who seek their own

The State Level of Analysis, Part II

225

independent state. His argument is that this particular type of nationalism provides the most danger of war. Van Evera hypothesizes that the existence of several sets of conditions makes war (internal or external) more likely. The first set of conditions are proximate (immediate) causes of war. First, the greater the number of nationalist movements without statehood, the greater the risk of war. Demands for statehood are not likely to be met because the nature of the desired change is both large and disruptive. Therefore, nationalist movements engender wars of secession that tend to become internationalized. The second condition involves how one treats a national diaspora—a national ethnic community outside the boundaries of its “natural” state. To the extent that nationalist movements (or states) adopt a policy of incorporating the diaspora into the state through annexation of the lands where this scattered community lives, war is more likely. (The alternatives are to simply accept the existence of a national diaspora or to rely on immigration to bring the diaspora into the state.) The third and fourth conditions have to do with whether the nationalist ideology respects the freedom of other nationalities (good) or claims the right to rule them (bad) and whether the nationalist ideology respects the rights of minorities (good) or seeks to oppress them (bad). Nationalisms that oppress minorities lead to wars through two paths: “(1) by provoking violent secessions by its captive nations; or (2) by spurring the homelands of these captive nations to move forcefully to free their oppressed co-nationals.” 94 Van Evera argues that if all four conditions are malign, a nationalism is bound to conflict with others. If the nationalism has no state, the risk of civil war arising from its struggle for national independence is increased; this also raises the risk of inter-state war, since civil war can widen to engulf nearby states. If, after achieving statehood, the nationalism seeks to incorporate a diaspora by force, oppresses minorities found in its claimed national territory, and seeks hegemony over nationalities lying beyond that territory, violence between the nationalism and its neighbors is inevitable. 95

Thus, there are two paths to war. In one path nationalism by stateless communities leads to wars of secession, which may then become internationalized. In a second path, nationalism causes conflict when a state’s leadership attempts to annex national diasporas located in other states or oppresses minorities at home, thereby inducing intervention by other states to “rescue” them. Other (background) conditions make it possible for these factors to have a malignant effect. Van Evera puts forward a number of hypotheses worth considering: 96 • Stateless nationalisms pose a greater risk of war if they have the strength to plausibly reach for freedom and the central state has the will to resist their attempt. • The more densely nationalities are intermingled, the greater the risk of war. • The greater the defensibility and legitimacy of borders, and the greater the correspondence between these political borders and communal boundaries, the smaller the risk of war. • The greater the past crimes committed by nationalities toward one another, the greater the risk of war. • The greater the coincidence of power and victimhood, the greater the risk of war.

Chapter 6

226

• The more divergent the beliefs of nationalities about their mutual history and their current conduct and character, the greater the risk of war. • The less legitimate the governments or leaders of nationalist movements, the greater their propensity to purvey mythical nationalist beliefs, hence the greater the risk of war. Two other hypotheses might be added. • Especially in multiethnic states, regime transitions (including democratization) can serve as a trigger to heighten the sense of ethno-religious nationalism, which then may increase the probability of internal (and perhaps external) war. Postwar Iraq must be seen in this light. • Power transitions between ethnic groups (due to differential rates of population growth) heighten type B nationalism and increase the risk of internal war. (This is a phenomenon which occurred in the decades preceding the conflicts in Lebanon, Bosnia, and Kosovo.) Miller’s State-to-Nation Imbalance Theory Benjamin Miller’s States, Nations, and the Great Powers represents one of the few books in the field to target nationalism as a primary cause of war and to construct a theory that explains how the causal relationships work. 97 Like Van Evera, he is interested in both internal war and interstate war; indeed, his theory points to a common source of both types of conflicts. His initial purpose is to explain why some regions of the world appear more prone to war than others. His answer combines concepts related to nation-centric nationalism, issues of territory, revisionist states, and the weak state–strong state dichotomy from the comparative politics literature. He argues, in a nutshell, that regions with a greater state-to-nation imbalance are more prone to war. The state-to-nation balance refers to the degree of congruence between the division of a region into territorial states and the national aspirations and political identifications of the region’s peoples. This balance also refers to the prevalence of strong versus weak states in the region. There is a state-to-nation imbalance when there is a lack of congruence between states and national identifications and at least some of the states in the region are weak states. 98

Thus, there are actually two dimensions to the regional state-to-nation balance. The first is the degree of state-to-nation congruence and the second is the extent of state strength in the region. Although the state-to-nation imbalance really works at the regional subsystem level, it includes many of the concepts that we have touched on in our state-level analysis of nationalism. And Miller actually prefers to think of states and regions as inhabiting one level of analysis, which he simply refers to as the regional/domestic level. And as we’ll see, the character of the state (and its pattern of nationalism) helps to determine the character of the region. Nevertheless, his hypotheses are essentially about regions; regional war or peace is his dependent variable. The first dimension of the regional balance is the degree of congruence between states and nations in the region. High congruence means that peoples in the region strongly identify with the existing states and accept their boundaries as legitimate. Low congruence suggests the opposite. Incongruence may present itself in one of two guises: too few states or too many

The State Level of Analysis, Part II

227

states. The “too few state” problem exists if state-seeking nations are unable to rule themselves. In this case states are internally incongruent, which typically leads to internal secessionist struggles. The “too many states” problem exists if one national group is spread across several states. Here there is external incongruity, in which nationalist groups may seek to unite all the lands of this people into a single state, thereby threatening the sovereignty and territorial integrity of some states in the region. For instance, pan-Arab nationalism and pan-Islamic nationalism in the Middle East reject the legitimacy of individual states and seek a united Arab (or Islamic) community. The degree of regional state-to-nation balance can be measured by combining (1) the proportion of states in the region which contain more than one national group and (2) the proportion of states in the region in which the majority ethnic group lives in substantial numbers in neighboring states, either as a majority or a minority. The higher the combined proportion, the greater the incongruity in the region. 99 Both types of incongruity between states and nations create incoherent states (incoherent states are both incongruent and weak), which in turn creates pressures by (and temptation for) dissatisfied groups for either secession or irredentism. Leaders of incongruent states may also use diversionary wars to buttress their own hold of power. The second dimension of the regional balance is the prevalence of strong or weak states. A strong state controls the means of violence within its territory and possesses an effective set of institutions able to raise economic revenues, mobilize manpower for military service, maintain an effective bureaucracy, and provide fundamental services such as regular water supply, electricity, education, health care, and a transportation network. A weak state lacks these abilities. The degree of state strength affects the ability of states to wage wars and to fend off challenges to their own internal integrity. Wars occur in regions with strong states and weak states; as long as the region is characterized by nation-to-state incongruence, it will be prone to war. The only difference is the type of war. All that is required for conflict is that some states in the region be weak, and this is almost always the case. Based on these two dimensions, we can find four types of states. 100 Strong-congruent states are status quo–oriented states. Regions dominated by these states—such as Europe and North America—will be balanced and should be relatively peaceful. These states generally have no nationalist claims on the territory of others and no internal nationalist groups seek to secede. All other categories are imbalanced and are more prone to war. Strong-incongruent states tend to be revisionist states who take advantage of their weaker neighbors and who seek to create a larger state to incorporate their national diasporas. Three patterns of conflict are associated with these types of states: 101 1. Majority-Majority: Two or more states share the same ethnic majority—like North and South Korea, China and Taiwan, or Iraq and Kuwait. This situation tends to give rise to pan-nationalist groups that push for national unification, labeling the other state as illegitimate. Iraq’s invasion of Kuwait, North Korea’s invasion of South Korea, and Chinese conflict with Taiwan can be seen in this context.

228

Chapter 6

2. Majority-Minority: The majority group in one state has co-ethnics present as a minority in a neighboring state. This pattern leads to irredentist wars, and attempts to create a “greater state.” German aggression against its Central European neighbors, Serbian aggression against Croatia and Bosnia, and Pakistani aggression in Kashmir are examples. As we will see in chapter 7, both majority-majority dyads and majority-minority dyads are among the most dangerous of all pairs of states. 3. Minority-Minority: An ethno-national group is a minority in two or more contiguous states—like the Kurds in Turkey, Iraq, and Iran. Here secessionist attempts may be abetted by states wishing to weaken their neighbors. (Both Iran and Iraq attempted to support secessionist movements among the Kurds in the rival state in the 1970s.) Alternatively, regional states may cooperate to quash these nationalist groups—as Turkey, Iran, and Iraq periodically do. As you can see, regions containing strong-incongruent states are imbalanced and prone to several types of interstate wars. There is also another problem: Strong states who are themselves internally incongruent may attempt to alleviate their own internal problems by resorting to diversionary wars against their neighbors. Miller’s primary example of this is Syria, where the minority Alawi faction dominates the Ba’ath Party government. In the period after the radical Ba’athist takeover in 1966, Syria’s militant policies toward Jordan, Saudi Arabia, and Israel ultimately led to the 1967 Six Day War. 102 One might also add Iraq, where Saddam Hussein’s Ba’ath Party government was also led by a minority (Sunni Arabs) and where scapegoating was partially responsible for Iraq’s invasion of Kuwait in 1990. Weak-incongruent states are incoherent states, and the most extreme of these are so-called failed states. Such states pose two serious problems: they tend to contain ethno-nationalist groups desiring to secede (especially if dominant groups are intolerant) and the state’s weakness both makes secession attempts likely and invites intervention by neighbors—either out of greed or fear. Regions including one or more of these states are imbalanced and prone to intrastate war and to interstate war. Examples are the old Austro-Hungarian Empire, Somalia, Sudan, Afghanistan, Jordan, Lebanon, Yemen, Bosnia, and Zaire/Democratic Republic of Congo. Weak-congruent states are a lesser problem. Although the state is not fragmented, it does not fully control its territory. It tends to have borders subject to disagreement. Regions with these states are prone to border wars over territorial issues. The best examples appear to be from nineteenth-century South America. For Miller, the state-to-nation imbalance within a region is the underlying cause of wars— both intrastate wars and interstate wars. The imbalance creates vital substantive issues that divide states and that worsen security dilemmas and power rivalries (two concepts we will discuss in the next two chapters). However, Miller argues that anxieties about power and security are unlikely to lead to war in the absence of a state-to-nation imbalance. 103 The stateto-nation imbalance is a necessary condition for war, but not sufficient. (After all, the underlying cause persists over extended periods of time; it varies only slowly.) Other, proximate causes determine when and how the war breaks out. Miller sees the main proximate causes as being derived from realist theories. These realist factors are only likely to be relevant if a high

The State Level of Analysis, Part II

229

degree of regional imbalance already exists. 104 Miller identifies three realist-related proximate causes of regional wars. 105 First, the state-to-nation imbalance intensifies the security dilemma. State insecurity is increased in internally incongruent weak states because ethno-nationalist challenges to the legitimacy and boundaries of weak states are endemic in the region. This leads states to take measures to increase their security (like increasing their arms), thus decreasing the security of their neighbors. This has destabilizing effects. An offense-defense military balance in favor of the offense makes this security situation even more dangerous. Civil wars or wars of secession in the region threaten neighboring states and the conflicts tend to expand as other states intervene. Miller calls wars arising from the security dilemma “wars of strategic vulnerability.” Second, the state-to-nation imbalance produces substantive motives for states to use interstate war to maximize power (and profit) at the expense of weaker neighbors. The initiators are likely to be strong revisionist states that are externally incongruent rather than internally incongruent. They may initiate regional hegemonic wars, limited wars of expansion, wars to reclaim lost territory. or preventive wars. Miller refers to these wars collectively as “wars of opportunity.” Third, the state-to-nation imbalance produces diversionary or scapegoat wars. Strong states that are internally incongruent and face domestic ethno-nationalist challenges are expected to externalize their domestic conflicts and initiate scapegoat wars. However, this is clearly not a primary motive and plays a relatively minor role in the wars Miller discusses. Miller calls these “wars of domestic vulnerability.” These are all related to two other realist factors—proximity and territory. Contiguity (proximity) is related to issues that arise out of nation-to-state imbalances, such as disputes over territory and populations among contiguous states. We have already seen how state-to-nation imbalances create territorial issues—probably the most dangerous type of issue over which states might disagree. As we will see later, some theorists argue the most wars are between contiguous neighbors because these states are likely to fight over disputed territories claimed by both. Miller argues that territorial issues are much more likely to escalate to war if they involve state-to-nation imbalances. 106 This is because those territorial issues tied to nationalism and ethnicity tend to be less divisible. Nationalist aspirations raise strong passions, making deterrence by stronger states problematic, as less powerful states might initiate violence regardless of their weakness. They also threaten the territorial integrity of contiguous states and raise concerns about dismemberment and occupation. Great-power activity in the region can mitigate or intensify the region’s proneness to war, but they cannot change it. Given a region with a state-to-nation imbalance, great-power cooperation or hegemony is likely to lead to a “cold peace,” while great-power competition or disengagement serves to promote a “cold war” that is punctuated by periods of “hot wars.” 107 As Miller says, “State-to-nation imbalance tells us which regions are inherently war-prone; great power involvement tells us when this inherent war proneness is likely to erupt into war.” 108 The nature of great-power involvement helps explain why the war-proneness of a region varies over time. Great-power involvement is, of course, a system-level variable, and Miller’s theory combines system-level factors with domestic/regional factors to explain varia-

230

Chapter 6

tion in conflict. Most regional wars are initiated by regional states rather than the great powers. Miller derives several hypotheses from his theory and provides some evidence in support of their validity. First, he argues that if the theory is correct, then more wars should be caused by state-to-nation issues than by other kinds of issues. According to Miller’s coding, such issues were associated with 62 percent of the wars during the Cold War period. Moreover, the end of the Cold War had little effect on this; 60 percent of the wars in the post–Cold War period were associated with state-to-nation issues as well. 109 He also finds support in the work of K. J. Holsti’s qualitative analysis of the issues associated with wars over several hundred years. Miller believes Holsti’s combined categories “national liberation/state creation, national unification/consolidation, and secession/state creation” jointly fit his concept of state-to-nation issues. Holsti sees these combined issues as a source of over 50 percent of the wars in the 1815–1914 period and in the 1945–89 period. They are less important in the 1918–41 period and of course in Holsti’s two earlier periods, 1648–1713 and 1715–1814. 110 In those earlier ages, nationalism was a relatively impotent force; indeed, in many areas of the world it barely existed in the minds of men. Thus, Miller’s theory is clearly time-bound. It may be appropriate for the last two centuries—the “age of nationalism” is typically dated from the French Revolution in 1789—but it is not relevant for previous eras. Miller’s second hypothesis is that regions with higher levels of state-to-nation imbalance should be more prone to both internal and external wars and most of these wars should be explained by state-to-nation issues. He finds that the regions that are most unbalanced are Africa, the Middle East, East Asia, and South Asia. The most balanced regions are Western Europe and South America. He finds that, as suspected, the four most unbalanced regions had the most wars in the 1945–91 period and that the majority of conflicts were state-to-nation. Appropriately, the regions with a low imbalance—Western Europe and South America—had fewer wars. 111 (Remember, Miller is counting both internal wars and interstate wars.) Finally, the distribution of internal war versus interstate wars in a region should depend on the type of state-to-nation congruence. In regions where both internal incongruence and external incongruence are high, internal wars and interstate wars should both be plentiful; this is the case in the Middle East and in the post–Cold War Balkans. Low levels of both types of incongruence should produce internal peace and a lack of interstate wars; this is the case in South American in the twentieth century, where those wars that do exist tend not to spread and neighbors tend not to intervene. In regions with high external incongruence and low internal incongruence, there should be domestic peace but proneness to interstate war; this was the pattern in Europe in the 1930s. In regions with low external incongruence and high internal incongruence, internal wars should abound, but interstate wars should be rarer; this is the pattern in postcolonial Africa, post-Soviet republics, and some Asian states. Thus, Miller claims to have found at least partial support for this hypothesis. 112 To help confirm his hypotheses, Miller also develops systematic case studies of the Middle East since 1945, the Balkans since 1830, nineteenth- and twentieth-century South America, and Western Europe since 1945. He finds that the vast majority of his predictions are validated.

The State Level of Analysis, Part II

231

THE ROLE OF WAR WEARINESS In the ninth volume of Arnold Toynbee’s A Study of History the British historian claimed to have identified a cycle of peace and war. 113 The one-hundred-year cycle, repeated over the centuries, consisted of the following sequence: A general war was followed by a period of peace, then by a cluster of small wars, a second period of peace, and finally by another general war. Toynbee’s tentative theoretical explanation was that war made a deep psychological impression on those who lived through it, and these men and women were hesitant to have their children’s lives disrupted by similar experiences. A whole generation of leaders whose lives had been molded during wartime were determined to keep the peace as long as they ruled. Eventually, however, a new generation would come to power. Never having experienced firsthand the horrors of war, they were much more inclined than their elders to test the waters of combat. The result would be a series of small wars—larger wars presumably still being avoided due to the aversion to such things handed down to them by the previous generation. The period of peace following these small wars would finally be shattered by another great war—a war that would come about only when the last memories of the first great war had been erased by the death of the wartime generation. Then, of course, the cycle would start over again. Toynbee’s one-hundred-year cycle was almost a perfect fit when matched with the period between the Napoleonic wars of the early nineteenth century and World War I. The explanation given by Toynbee for this cycle of war and peace is generally referred to as the war weariness theory. It predicts that nations that have recently experienced a long and costly war should be the most peaceful—at least in the short term. Conversely, states that have endured long periods of peace are probably most likely to experience war in the near future. As Blainey observes, the theory ironically implies that we should beware Sweden and the Canary Islands. 114 More specifically, why and how is war weariness supposed to lead to peace? To some extent the war weariness theory draws on psychological arguments. It maintains that political leaders who have directly experienced the devastation of war are deeply touched by the experience—presumably at a conscious level, but perhaps unconsciously as well. The experience of war results in a strong aversion to war, and this war weariness affects these leaders’ personalities, their operational codes, their images of the world, and their value preferences. This explanation operates at the individual level of analysis. The war weariness theory also operates at the state level of analysis. Here the theory implies that the common experience of a severe and destructive war has made an impression on the collective psyche of the nation and affected the national character, or political culture. War weariness is not a personal characteristic as much as it is an attribute of the state. A complete theory of war weariness would have to specify the link between the war weariness of society as a whole (or certain groups within society) and political decision makers. The presence of popular aversion to war implies that, at least in a democracy, people will make known to the government their predisposition for peace and that government policy will reflect the wishes of the people. Even in authoritarian countries this collective support for peace cannot be discounted entirely; dictators, too, must take into consideration such widely

232

Chapter 6

held sentiments. Consequently, other state-level attributes may not matter greatly; states with similar histories of war experience should act similarly in the future. 115 War weariness has relevance for war initiation but not necessarily for war involvement. It is concerned with the effect of a costly previous war on a country’s desire to initiate a new war. However, if a nation is attacked, war weariness is not likely to prevent a state from participating in a war if it has been attacked. One might in fact hypothesize that states that have recently experienced war are more likely to be attacked than other states. Opponents may sense the spirit of war weariness in such a country and see this as a sign of weakness. Similarly, a state that has been physically, economically, and militarily weakened by a costly previous war might be seen as an easy target by other states. War weariness shares much of the conceptual framework of disease and immunity. As Richardson has suggested, war resembles a disease. One possible cure for this disease is, ironically, a potent dose of war itself. The actual experience of war constitutes a kind of immunization against future war. Unfortunately, like a tetanus shot, the effects of the inoculation don’t last and the state’s immunity fades. 116 If the war weariness theory is true, wars are caused by states whose immunity to war— generated by the devastation and death of a previous conflict—has begun to lose its potency over the course of time. The implications of this theory are fairly pessimistic. The good news is that wars can be prevented. The bad news is that the best way to prevent them in the future is to fight them in the present—and even this won’t prevent war indefinitely. War weariness theory has some major weaknesses, however. First, it does a poor job of explaining some conflicts that cluster close in time. World War II occurred just two decades after the end of World War I, and leaders of the states of Europe were all part of the generation whose lives were molded by what had been the most horrible war in history. Surely, if any generation were to be war weary, this would have been the one. 117 It also does not square very well with the experience of the United States in the post–World War II era, where very little peace seems to have interrupted an almost endless series of “small wars” and interventions: Korea, Vietnam, Grenada, Panama, Afghanistan, and Iraq. Certainly, our confidence in the war weariness theory must be reduced by such glaring examples of events that run counter to the theory. Second, war weariness theorizes that any involvement in war will cause participants to refrain from war. But in reality whether a state wins or loses a war will have an important effect on its future policy. In most cases, winning states experience a lesser degree of devastation and suffering than losers; thus, they should be less war weary and more willing to engage in future aggression. From the perspective of behavioral learning theory, individual and state practices that are successful (such as winning a war) receive positive reinforcement and are therefore likely to be repeated as behavioral responses to certain environmental situations. On the other hand, unsuccessful behaviors (losses in wars) constitute punishment or negative reinforcement and tend not to be repeated. And on a practical level, a winning state’s physical capabilities might be enhanced as the result of victory. Victory might raise levels of nationalistic fervor, create a climate of optimism about war, solidify the power of hawks, or reinforce cultural norms of militarism and aggression. 118 All are good reasons why previous winners might not be subject to war weariness.

The State Level of Analysis, Part II

233

It could be argued that a defeat in a long and destructive war would be the situation most likely to bring about the war weariness syndrome. As we have suggested, it can usually be assumed that the greatest devastation of the war will be borne by the loser. The greater the devastation and loss of life and the harsher the war, the greater the resultant war weariness and (presumably) the greater the immunity to future war initiation. This argument is not foolproof, however. Losing sides might be motivated by the desire for revenge or to regain lost territories, peoples, and resources. It is frequently asserted, for instance, that Germany’s desire for revenge following its defeat in World War I was a major cause of its aggression in World War II. John Vasquez adds another consideration: Even if a state loses a war, if the consensus of opinion is that the war was worth the cost, then leaders with a hardline approach to foreign policy might still be able to dominate the political landscape in that country. 119 Again, Germany after World War I is an example. If it is equally plausible that A (a costly defeat) causes B (peace) and that A also causes C (initiation of future wars), then neither proposition is likely to be true as a general rule. Theoretical arguments that war could logically lead to periods of peace as well as to future war are equally plausible and, of course, incompatible. Herein lies perhaps the greatest weakness of war weariness as a general theory of war and peace. Previous wars may be negatively contagious (inhibitory) or positively contagious (contributory) to future wars; if so, the negative and positive effects will tend to cancel each other out. Third, regardless of whether a country has been defeated or victorious in a previous war, it might be argued that the mere experience of war is a factor enhancing the possibility of a more warlike rather than a more peaceful future. For instance, Karsten argues that war accustoms individuals to military attitudes and values, which are in turn spread by returning veterans. 120 Not all veterans have militaristic attitudes, but many frequently return home with new or reinforced attitudes toward military virtues and the use of force. 121 A kind of “reverse war weariness” may in fact occur, as the military way of life is glorified. Veterans organizations, whose ranks would be swelled by the war, may become important institutional factors influencing governments toward more aggressive policies. Wars also lead to the creation of large military establishments with equipment, personnel, bases, budgets, bureaucracies, and general staffs, not to mention the various industrial manufacturers involved in military production. All of these may be difficult to reduce in size and political power after the end of the war. In other words, wars lead to the establishment of “military-industrial complexes” that may become powerful lobbyists for continued military activity. 122 Fourth, even if it is true that wars are followed by sustained periods of peace, war weariness may not be the only or best explanation for this. It may be that a decisive end to the war has resolved all outstanding issues, removing the political causes for future wars. It may be that a state’s resources have been so depleted that it is physically unable to pursue war. Similarly, a decisive victory by one side may have created a regional or global distribution of power so lopsided that those who have grievances are nonetheless deterred from redressing these grievances through force. 123 Therefore, an empirical link between war and subsequent periods of prolonged peace does not necessarily confirm the hypothesis.

234

Chapter 6

War Weariness: Empirical Studies Empirical studies of the war weariness hypotheses, though somewhat mixed, have not been very kind to the theory. Some studies find virtually no support for the war weariness hypotheses. In his analysis of great-powers wars from 1816 to 1965, David Garnham finds no relationship between the cost of war and the elapsed time until the next war, and he determines that victorious major powers are no more likely than losers to initiate a future war. He also investigates the classic proposition of Immanuel Kant that democratic states ought to be particularly susceptible to war weariness phenomenon but finds no evidence that war weariness has constrained the behavior of the major democracies of the nineteenth and twentieth centuries—Britain, France, and the United States. 124 Levy and Morgan study the involvement of great powers in wars between 1500 and 1975. They discover many more instances of nations reengaging in wars in a relatively short period of time than should be expected. In fact, out of 115 cases of wars that were begun after a previous great-power war, ninety-one wars were begun within ten years of a previous greatpower conflict. Sixteen were begun between ten and twenty years later; five were begun in the next ten years, two in the next decade, and only one in the fifth decade following a greatpower war—almost exactly the opposite of what the war weariness hypothesis would predict. 125 When Levy and Morgan shift their attention to the hypothesis that the more serious the war, the greater the length of time between wars, the results are similarly disappointing. Using duration of war, the number of participating countries, battle deaths, and a battle death/war duration ratio as indicators of the seriousness of war, they find only weak correlations between these measures of the independent variable and the dependent variable (elapsed time until the next war). Neither are they able to confirm that inhibition to war is brought on by a series of wars instead of a single previous war. In short, they are neither able to confirm any of the war weariness hypotheses nor find any distinctive or consistent patterns with regard to the effect of previous wars on subsequent war participation. 126 It should be noted, however, that Levy and Morgan investigate the propensity of war-weary states merely to become involved in subsequent wars rather than to initiate them, and initiating war is the true focus of war weariness theory. Other studies have produced modest support for the idea that war weariness affects states whose last brush with interstate war resulted in a defeat. In a preliminary analysis of wars from 1816 to 1965, Singer and Small conclude that neither initiators nor defenders were very likely to initiate war within a decade, though winners were far more likely to initiate subsequent wars than losers. Both results are consistent with the war weariness theory. 127 A follow-up study by the same authors reinforces the conclusion that war winners are much more likely to get involved in subsequent wars than the losers of previous wars: “It is not so much that war begets war, since very few nations initiate wars in the decades following prior war experience, but that victorious war begets war.” 128 Perhaps the lack of evidence for the war weariness hypothesis is because for winners of previous wars (roughly half of all war participants) the negative effects of wars are offset by the fact that they were victorious. A later study by Singer and Cusack, using war data from 1816 to 1965, focused on war participation rather than war initiation. Their general conclusion is that prior war experience

The State Level of Analysis, Part II

235

seems to have had little effect on the propensity of states to become involved in subsequent wars. Winners of previous wars have a slight tendency toward early reentry into war, but so do defeated states. In fact, the average time interval to the next war is even shorter for defeated states. Although the difference between defeated states and winning states is not statistically significant, the finding points toward the revenge motive and away from the war weariness theory. 129 However, the authors also discover that defeated states that have fought costly wars with high fatalities do seem to be inhibited from participating in wars too soon after their defeat. The combination of defeat and high cost would seem to be much more important than either of these factors alone in explaining why states returned to war so soon after their experiences in previous war. 130 Finally, a more recent study by psychologist John Nevin focuses specifically on the differences between winners of previous wars and losers, though he does not evaluate the magnitude of costs for losers or winners. 131 Nevin uses Singer and Small’s data from the COW project on interstate wars from 1816 to 1989 and also Levy’s data on great-power wars from 1495 to 1815 to look specifically at war initiation. And he considers only states with a minimum of two wars, so that the actions of states in war Z can be considered in light of the outcomes of previous wars W, X, and Y. Nevin’s analysis produces some interesting results. First, states that win previous wars are more likely to initiate subsequent wars than states that lost their previous wars. The larger the number of previous wars that a state has won, the greater is the probability that it will initiate another war. The probability of initiating a war is consistently lower for losers than for winners for a period of twenty years; at that point the effect is wiped out and losers experience an abrupt increase in war initiation. Second, winning and losing have an effect on the time interval between the last war and the onset of the next one. War initiation occurs sooner for states that have won their previous war, compared with losers of the previous war. Third, the slow-down effect of a previous loss appears to be offset if the loser has a history of one or more wins prior to its latest loss. Consistent with what the psychological literature would predict, a previous history of rewards (winning wars) leads to the persistence of the behavior (initiating further wars) in spite of recent setbacks. Finally, Nevin takes notice of an interesting (but dangerous) feedback cycle. We know from other studies that wars tend to be won by the initiators. 132 Thus, initiators tend to win, and winners tend to initiate (again). This suggests that, at least in the short run, winning creates a class of international bullies and repeat offenders that are in some way addicted to war as a solution to problems. In sum, we must conclude that the evidence supporting the war weariness theory is considerably less than compelling. 133 But we also have some clues as to why the effect of war weariness appears to be negligible: Winning wars appears to create a positive environment in which starting a subsequent war seems more acceptable. To the extent that war weariness exists, it has a greater effect on war losers than on winners. But even for losers, the effect is not particularly long-lasting, and it is certainly not multigenerational.

236

Chapter 6

CONCLUSION While many state-level factors appear to have little relevance on the possibility of war, a few appear to have a potent effect. In addition to the finding (reported in the previous chapter) that large, powerful states are more likely to be involved in war, we also find some connection between political instability and war. There does seem to be an association between internal conflict and interstate war, though there are clearly multiple paths from the former to the latter. Nationalism, especially nation-centered nationalism, appears to increase the risks of certain kinds of wars, and it also appears to be connected to the internal instability of states. Both internal incongruence and external incongruence of states and nations increase the likelihood of conflict, but we have yet to home in empirically on the strength of these factors and the conditions under which they operate. None of these potent state-level factors would appear to be either a necessary condition for war or a sufficient condition for war. It is probably best to consider these attributes as risk factors. If states possess one or more of the above characteristics, the risk of war is increased compared to states that lack these characteristics. In the next chapter, we proceed to a higher level of analysis, one that examines the relationships between states rather than looking at the attributes of a single state.

Chapter Seven

The Dyadic Level of Analysis, Part I The Nature of Dyads—Really Bad Dyads and Pretty Good Dyads

Before I built a wall I’d ask to know / What I was walling in or walling out, / To whom I was like to give offense. / Something there is that doesn’t love a wall, / That wants it down . . . —Robert Frost Of all the torments, all the cares, / With which our lives are curst, / Of all the plagues a lover bears, / Sure rivals are the worst. —William Walsh Take two democracies and call me in the morning. —T. C. Morgan and V. L. Schwebach

The dyadic level of analysis focuses our attention not on the nature of states individually but on (1) the nature of pairs of states (dyads), that is, their mutual or shared characteristics, and (2) the interaction between these pairs of states. This chapter and the two subsequent chapters will each deal with theories of war that emphasize the importance of factors at the dyadic level of analysis. As a result of empirical research on the causes of war, several things have become obvious. Not all pairs of states have an equal probability of being involved in wars. A small portion of all possible dyads account for a high percent of wars over the past few centuries. In a definitive study Zeev Maoz analyzed empirical data on wars and militarized international disputes (MIDs) for the years 1816–1992. 1 He looked at the behavior of politically relevant dyads— pairs of states that are geographically contiguous or include at least one major power. This is a subset of all potential dyads that constitute the pairs of states we would most expect to be involved in conflict with each other. The data indicate that a majority of all politically relevant dyads (57 percent) are never involved in MIDs and almost 87 percent of them never fight a war. It turns out that a small fraction of politically relevant dyads—the upper 10 percent— account for 34 percent of all MIDs and an astonishing 95 percent of all interstate wars in the period. What accounts for the massive disparity between the majority of relatively peaceful dyads and the minority of extremely conflict-prone dyads? Political researchers have been quite busy working on answers to this question.

237

238

Chapter 7

Political scientists are fond of saying that we know only two things for sure about war: “democracies don’t fight each other and neighbors do.” Thus, political scientists have identified one good/peaceful kind of dyad (jointly democratic states) and one bad/war-prone type of dyad (contiguous states). We should add two other candidates for the bad dyad list, both highly supported by empirical investigation: enduring rival states and states with near parity in their power capabilities. These four propositions are the central focus of this chapter. THE ROLE OF CONTIGUITY AND TERRITORIAL DISPUTES Most wars are fought by neighbors—contiguous states who share land or water borders. 2 An early inspection of the Correlates of War (COW) data indicated that 88 percent of the sixtyseven interstate wars from 1816 to 1980 began as wars between neighbors, and if you delete “imperial wars” involving great powers in their overseas realms, the number is 100 percent. 3 Using Holsti’s qualitative analysis of wars over a much longer period from 1648 to 1989, John Vasquez determined that 91 percent were started by neighbors. 4 Paul Hensel, one of the primary authorities in the field, argues that from 1816 to 1992 about two-thirds of all full-scale interstate wars were begun by contiguous states, and in the later 1945–92 period “almost every full-scale war” began between contiguous states. 5 Wallensteen finds that 93 percent of contiguous pairs of major powers have experienced at least one militarized confrontation, and 64 percent of them have gone to war at least once with each other. 6 Stuart Bremer’s famous study of “dangerous dyads” offers evidence that contiguity is far and away the most important factor associated with war, increasing the probability of war between two states thirty-five times. The other variables in his analysis increase the probability of war only when the dyads are already contiguous. 7 The importance of contiguity is confirmed in a recent study by Bennett and Stam, who conclude that “contiguity increases the risk of war eight times and the risk of reciprocated force nearly twenty times over the baseline level of risk” faced by a pair of states. 8 While the exact magnitude of the danger is open to question, the fundamental proposition is not. Neighbors fight. Contiguous states are not just more prone to wars, they also have a higher likelihood of involvement in conflict in general. Paul Hensel has looked at the links between contiguity and MIDs—dyadic disputes that involved either the threat of military force, the demonstration of military force, or the actual use of military force at some level. More than half of the MIDs in the 1816–2001 period took place between states sharing land or river borders, and another 14 percent were between states contiguous by sea. These numbers were even higher for MIDs with fatalities and those that ended in war. Moreover, Hensel determines that the effect of contiguity has not lessened substantially over time. 9 Paul Diehl’s research on major power rivalries shows that twelve of the thirteen rivalries that ended in war involved confrontation over territory of one of the rivals or over territory contiguous to them. Moreover, of the fifty great-power MIDs involving disputes that were contiguous to one or both of the rivals, twelve escalated to war, while only one of the fifty-four MIDs that were not contiguous to either power escalated to war. 10 Another discovery—one of the more ancient statistical associations in the study of war— provides strong indirect support. Lewis Fry Richardson, one of the first practitioners of empir-

The Dyadic Level of Analysis, Part I

239

ical research in international relations, discovered an extremely strong correlation between the number of borders a country had (both local and colonial) and the number of wars it was involved in. His study of thirty-three states from 1825 to 1946 indicated that the greater the number of borders a state had, the greater the number of wars it experienced. The finding was subsequently confirmed by several other studies. 11 Finally, there is very strong evidence that when armed conflict spreads, it spreads primarily to contiguous states. States located on the borders of a territory with an ongoing war are more likely to join than states lacking geographic proximity. This is true of both interstate wars and intrastate wars. Contiguous states are the states most at risk for war’s diffusion. 12 Of course other factors are also important; alliance ties and the existence of rivalries also help to predict which states will join ongoing conflicts. In one sense the link between contiguity, conflict, and war appears so obvious that it may even be trivial. Most states fight their neighbors because it is not militarily and technologically feasible for them to fight anyone else. Contiguous states are more likely to go to war with each other than with noncontiguous states simply because they have more opportunity to do so. As Kenneth Boulding’s concept of the loss of strength gradient reminds us, the ability of a state to use its military power declines with geographic distance from its home base; thus, relatively fewer wars are fought between noncontiguous countries because few states are able to easily mobilize their militaries for effective action over long distances. 13 We would not expect Peru and Nigeria to fight each other, nor Somalia and Pakistan. Only those major powers with considerable ability to carry out military operations over long distances would experience much involvement in wars with noncontiguous states. Bruce Bueno de Mesquita finds that 89 percent of all “long-distance wars” are indeed initiated by great powers. 14 Sharing a border doesn’t just increase opportunities to fight, it also increases neighbors’ willingness to fight each other. 15 All things being equal, states are more likely to be concerned about their neighbors’ intentions and ambitions, the status of their military forces, and the composition and stability of their regimes compared to less proximate states. Perceived threats from proximate states are seen as more serious, more urgent, and as having a more direct effect on their vital national interests than those from more distant states. Because of these issues, when political scientists attempt to ferret out the causes of war, they normally limit their analysis to what are called politically relevant dyads—dyads that consist of contiguous states or dyads in which at least one state is a great power. Pairs of states that don’t fall into this category of politically relevant dyads have an extremely slim chance of ever ending up in war with each other. However, mere proximity as a primary explanation for war has a number of weaknesses. First, we would also expect that as technology extends the effective reach of militaries, the loss of strength gradient should be reduced over time, and there should be more wars between noncontiguous states in the modern era. There is no evidence of this, and indeed there is some evidence to the contrary. 16 Second, the contiguity argument is that proximity leads to greater contact and interaction between a pair of states and that this leads to a greater mathematical possibility for negative interaction. The more interactions in general, the more likely that states will develop some sort of dispute or disagreement or conflict of interest. However, one might argue just as logically

240

Chapter 7

that greater overall interaction between a pair of states would increase the chances for cooperation. This is essentially the argument made by social communication theorists like Karl Deutsch. These theorists see ease of communication, transportation, and transactions across borders as something that is associated with increased trade and economic interdependence, increased cooperation, and even increased political integration and the creation of security communities. 17 Thus contiguity, and the ease of transaction due to contiguity, is just as likely to have positive outcomes as negative ones. John Vasquez points out a third problem with using contiguity to explain war: Saying that proximity provides an opportunity for war is not the same thing as saying it causes war. 18 Many contiguous states fight often while others not at all. Proximity is a constant. States A & B, C & D, and E & F may share the same borders for two hundred years. During that time, A & B may fight once, while C&D might fight four wars and E & F none. War involvement is not a constant; it clearly varies greatly. While proximity is a necessary condition for war—at least to the extent that states must be physically able to “reach out and touch someone” in order to fight—it is not sufficient by itself to cause war. So why do some pairs of neighbors fight while others do not? One possibility is that the relationship between contiguity and war may derive from that fact that neighbors have real disputes over territory and boundaries. Neighbors fight because they are much more willing to go to war over territorial issues than over other things that might divide them, such as ideology, religion, government policies, regime types, or the balance of power. For Vasquez, contiguity accounts for opportunity, but it is the presence of actual territorial disputes that accounts for a state’s willingness to use force. 19 Contiguity simply masks the fact that the really critical variable is the presence of a dispute over territory. Vasquez places territorial issues as the center of his “steps to war” theory (which we will examine in more detail in a later chapter). He argues that territorial disputes are an underlying cause of war (as opposed to a proximate cause of war) and that real territorial disputes provide a better explanation of the link between contiguity and war than mere proximity. 20 After all, if proximity is the explanation, then it should not matter what the issue was; one issue should be just as likely to touch off a violent exchange as any other. Hensel shows that this is not the case. 21 His test of the relative merits of a contiguity/opportunity explanation versus a territorial issue explanation gives a decided edge to the latter. Looking at conflicts between 1816 and 1992, he finds that the effect of territorial issues is strong for both contiguous states and noncontiguous states. The presence of territorial issues doubles the statistical odds of fatalities in disputes between contiguous adversaries, but it more than triples the odds of fatalities in noncontiguous disputes. A later investigation by Senese and Vasquez, using data from the period 1919–95, comes to a somewhat similar conclusion. They discover that while contiguity provides a very good explanation for why states get involved in MIDs with each other in the first place (the opportunity factor), it does poorly in explaining why these disputes escalate to war (the willingness factor). Territorial issues are much more important at the escalation stage. Most interesting is the discovery that noncontiguous dyads that have territorial disputes are even more prone to war than contiguous dyads that have territorial disputes—a clear refutation of the hypothesis that mere contiguity increases the probability of war. 22 Apparently, the nature

The Dyadic Level of Analysis, Part I

241

of the issue (territory) is so powerful that even distant states will be willing to go to war despite the problems associated with fighting wars with noncontiguous opponents. What, then, might explain why conflicts over territory have such a virulent effect? Three types of answers appear relevant: (1) human territoriality, (2) characteristics inherent in the nature of the territory itself, and (3) domestic and international political considerations. The first possibility is that humans, like other vertebrates, are territorial—in the sense that they “use aggressive displays to keep and gain territory.” 23 Vasquez sees the tendency toward territoriality as the result of neither an innate drive nor an instinct, but as due to the fact that humans are “soft-wired” to learn and retain certain cultural behaviors rather than others. And one such learned behavior is to use a collective form of violence (warfare) as one way to resolve territorial disputes. If humans are indeed territorial, Vasquez suggests that four propositions should logically follow, all of which, he believes, have both intuitive and empirical support: 1. collectivities would divide the Earth, often through the use of force, into territorial units; 2. collectivities would pay a great deal of attention to their boundaries and be vigilant in their defense; 3. on the whole, the boundaries between any two contiguous states that are relatively equal in capability would be expected to be established through a struggle involving the use of force; 4. any new state that emerges could be expected to pose a threat to existing territorial divisions, and if these fears were not allayed, then conflict and violence would become likely. 24 A second explanation for the strong connection between territorial disputes and war is that a particular piece of territory may have special value to the combatants. The land may contain strategic minerals, energy supplies, or important economic resources. It may provide defensive advantages (the Golan Heights for Israel or Syria) serve as a choke point to deny an opponent access to important land or sea routes (islands in the Straits of Hormuz), or simply provide greater defense-in-depth (the western reaches of the old Soviet Union). Some empirical support comes from researchers who have constructed an index of “salience” to measure the importance of a territory to a particular dyad. Hensel demonstrates that rival claims to territories that are deemed to be highly salient lead to nearly double the number of MIDs, compared with the average territorial claim. 25 Territory may hold intangible value for the states. It may be central to a peoples’ national identity (Kosovo to the Serbs or Kiev to Russians) or have special historical importance. Incorporating pieces of territory where co-ethnics live might protect them from harm and also complete a project of national unification (Ethiopia’s Ogaden region for ethnic Somalis). Similarly, land may be endowed with a religious significance (Jerusalem, for instance, or Karbala for Shi’a Muslims in both Iran and Iraq). Thus, while all territory is potentially divisible and contesting states could draw a line through it and “split the difference,” in some geographic conflicts the issues simply becomes “indivisible.” 26 Compromise is not politically possible. Due to the rise of nationalism, states can no longer bargain away parts of “their land”

242

Chapter 7

and some of “their people” without severe domestic repercussions. Oddly enough, territorial conflicts tend to be zero-sum contests. If that is so, the intangible aspects of territory may serve to ramp up political leaders’ determination to use force. This leads us to the next set of arguments. Domestic politics provides a third general reason for the conflict-prone nature of disputes over territory. Territorial issues—especially if they contain intangible and symbolic characteristics—are likely to be manipulated for political purposes. Many international relations theorists assume that one important goal of all political leaders (if not the primary goal) is to maximize their chances of retaining power—by maintaining and increasing their support among various political constituencies. Since territorial issues inherently involve “red meat” themes of nationalism, national security, and territorial integrity, political elites are prone to invoke these issues and use them for political gain. Territorial issues tend to morph into tests of national loyalty. This is particularly true when territorial issues engage ethnic solidarity and national unity. Territorial issues are also likely to develop substantial hawkish domestic constituencies—including military institutions—who pressure government elites to act decisively. As a result, territorial issues may be difficult to control. Given the inherently nationalistic and security-laden natures of these issues, political leaders are unlikely to retreat on them or make concessions. Huth’s study of territorial conflicts generally confirms that political elites have very few domestic political incentives to end territorial disputes; to the contrary, domestic political forces generally reduce political leaders’ incentives to compromise on territorial issues. 27 Political failure on territorial issues is likely to result in more severe consequences than failure on other issues. (After all, the loser in such a contest might suffer military occupation of some of its lands.) Moreover, political elites are always concerned about the state’s reputation, and backing down on territorial issues sends a message of weakness to potential opponents and emboldens others to challenge other territorial assets elsewhere; if a state declines to fight to preserve its own territory, what will it fight for? Political elites are therefore likely to take greater risks and pursue hardline escalatory strategies when territorial issues are concerned compared to nonterritorial issues, and they are less likely to suffer criticism for strong action on these issues, even if they fail. Politically savvy leaders understand that they have a greater capacity to mobilize domestic support for strong action on these issues than on other issues. Indeed, territorial disputes are likely to become a domestic political issue in both countries. 28 As a result of all these things, leaders are probably less willing to make concessions with regard to territorial issues, compared with other issues, more willing to take high risks in pursuit of their goals, and more willing to use military force to achieve their objectives. Territorial Disputes and War: Evidence Two historical surveys on interstate conflict appear to provide somewhat mixed evidence on the pervasiveness of the association between wars and territorial disputes. The first, Evan Luard’s War in International Society, divides modern history into five distinct periods and notes that the issues over which wars have been fought (and the motives of states) change over time. What appears to be vital in one age, and a justifiable cause of war, seems questionable in another. He also notes that difficulty of assigning certain issues as the cause of war. Luard

The Dyadic Level of Analysis, Part I

243

wisely notes that wars are rarely fought just about one specific issue, and that the issues may in fact be seen differently by the combatants. Moreover, developing issue categories is fraught with danger; they can be so broad as to be meaningless, or they can be so narrow as to apply only to one occasion. 29 Luard argues that in the Age of Dynasties (1400–1559), territory was rarely an issue by itself and was likely to be linked to the more important concerns of the age, such as dynastic claims of succession. In the Age of Religion (1559–1648), conflict over religion ignited conflicts over territory, in large part because the norm of the age was that rulers of a particular territory had the right to determine its religion. Nevertheless, territorial issues that lacked any religious dimension were also able independently to touch off wars, especially in areas bordering a particular realm where sovereignty was weak or ambiguous. In the Age of Sovereignty (1648–1789) religion virtually disappeared as an issue leading to war, and territory became the most important single issue associated with war, often associated with questions of dynastic succession and with the increased power that came with control of large or strategic territories. For balance of power reasons, control over strategic territories was especially important. In the Age of Nationalism (1789–1917), wars were most frequently linked to issues of national identity and independence. Territorial issues were rarely a significant issue behind the outbreak of war in Europe, but continued to be important in Latin America and the Far East. And territorial disputes were a major issue in colonial wars. Finally, in the Age of Ideology (1917–86), the issues that gave rise to war were increasingly those linked to conflicting ideologies and what kind of governments should rule—democratic or fascist, communist or noncommunist. There were conflicts over frontiers and the precise location of boundaries, but few over wide-ranging claims to another’s territory, and very few wars of outright conquest, at least after 1945. 30 Other scholars note that while Luard declines to nominate territory as the dominant factor throughout the centuries, nevertheless, territorial issues keep cropping up in his analysis—in connection with issues of succession, dynastic disputes, or state creation. Thus, Luard probably underestimates the importance of territorial issues in fomenting wars. 31 In addition, his analysis is not empirical; he doesn’t present quantitative data but simply offers his judgment on these issues. A second historical survey, K. J. Holsti’s Peace and War: Armed Conflicts and International Order, 1648–1989, provides a somewhat different analysis. 32 Starting in 1648, at the beginning of the modern era of nation-states, Holsti divides the last three and a half centuries into five historical eras (using different periods than Luard). He identifies twenty-four issues, many of which overlap conceptually. While he agrees with Luard that most wars arose over multiple issues, his data indicate that the combination of territory/strategic territory/boundary issues was probably the dominant issue connected to the outbreak of wars over the entire five periods—rather than issues such as balance of power, regional dominance, ideology, commerce, or religion. Although, like Luard, he believes that the importance of territory in producing war has declined in the last two periods (1918–41 and 1945–89), nevertheless, a majority of wars in all five periods are linked to the three territorial issue categories—from a high of 85 percent (1715–1814) to a low of 52 percent (1945–89). 33 He suggests that states are now less

244

Chapter 7

likely to use force to settle territorial issues because of international constraints and because territory is now less important as an index of state power. 34 It is useful to compare Holsti’s data with the standard compilation of war data from the Correlates of War (COW) project. Hensel finds that over half of the wars in both lists can be linked to explicit conflict over territory. But while Holsti’s list indicates a relative decline in the percent of wars over territory issues in the post-1945 era, the COW data actually show an increase in the percent of wars linked to territorial issues in the recent period. 35 Vasquez subjects Holsti’s data to re-analysis. He combines four of Holsti’s original categories (territory, boundary, strategic territory, and irridenta) into a single territory category, and then creates a second combined category of “territory-related issues” (national liberation/state creation, secession/state creation, national unification/consolidation, maintaining the integrity of state/empire, dynasty/succession). This consolidation of issue types leads to the results in table 7.1, which show that a combination of territorial issues are associated with roughly 80 percent to 90 percent of all wars from 1648 to 1989. Vasquez concludes that “territorial issues, defined fairly narrowly, have dominated warfare for almost 350 years. No other single set of issues has persistently been present in wars across such a wide range of historical periods. Indeed, only in the post-1945 period have territorial issues not been in a majority of wars.” 36 Defining territorial issues more broadly just pumps up the percentages even higher. A more recent effort by Vasquez and Valeriano classifies wars according to the issues that were behind the war—territorial, policy, or regime issues. Of the seventy-nine interstate wars between 1816 and 1997, the majority (54.4 percent) arose from territorial disputes; 24.1 percent derived from policy disputes and 11.4 percent from regime disputes. If a broader definition of territorial dispute is used, the proportion rises to 59.5 percent. If one looks only at dyadic wars between neighbors (which constitute 49 percent of all wars), the most common type of war is over territory. Thus, the authors are able to identify dyadic (two-sided) wars between neighbors over territory as the modal war of the nineteenth and twentieth centuries. 37 Several recent studies have added to our understanding of territorial conflicts. Stephen Kocs’s analysis of the 1945–87 period focuses on unresolved territorial disputes. Using a fairly restrictive definition of territorial disputes, he finds that of the twenty-one wars that began in the period, nineteen were initiated by land-contiguous states (the exceptions were Britain’s 1982 war with Argentina over the Falklands and the 1974 Turkish-Cypriot War). Of the twenty-nine dyads involved in these wars, twenty-five were contiguous and eighteen were involved in an unresolved dispute over contiguous territory at the war’s onset. He concludes

The Dyadic Level of Analysis, Part I

245

that war was almost 300 times more likely between contiguous states than noncontiguous states and, most important, “war was about 40 times more likely to break out between contiguous states if they were involved in a never-resolved territorial dispute than if they were not.” 38 On the other hand, he deems war as extremely unlikely if the boundary between contiguous states is “clearly delimited and legally valid.” He also finds that in the post–WWII era, although wars over territories subject to longstanding legal disputes are common, wars that constituted outright grabs for new territory have been rare. And while most territorial conflicts do get resolved at some point, new disputes have been created by wartime conquests (Israel’s occupation of the West Bank, East Jerusalem, Gaza Strip, Sinai Peninsula, and Golan Heights as a result of the 1967 Six Day War) and the breakup and partitioning of states (postwar Germany, Korea, and Vietnam, as well as post–Cold War USSR and Yugoslavia). A large number of studies have now produced solid evidence that of all the types of issues over which dyadic disputes might erupt, territorial disputes are the most dangerous. Vasquez and Henehan directly tested the relative effects of different kinds of issues on the likelihood that militarized international disputes will escalate to war. 39 Using data from 1816 to 1992, they find that disputes over territory have a higher probability of ending in war than one would expect on the basis of pure chance. Most important, they discover strong evidence for the relative importance of territorial issues. Territorial disputes are more likely to escalate to war than dyadic disputes over the policies of opponents or disputes in which one state attempted to overthrow the regime of another. Moreover, 54.6 percent of the ninety-seven wars during this period were associated with territorial disputes, while only 32 percent of wars were associated with policy disputes and 9.3 percent involved regime disputes. When one looks at dyadic disputes, the same pattern appears: two-thirds of the dyadic disputes that escalate to war are over territory, less than one-fourth are dyadic disputes over policy, and a very small fraction involve regime disputes. Switching focus to pairs of states rather than disputes per se, Vasquez and Henehan seek to discover which dyads are most likely to have engaged in at least one war over the period— those whose relations are dominated by territorial disputes, by policy disputes, by disputes over the nature of a regime, or by a combination of these dispute types. Dyads dominated by territorial disputes had a 63 percent chance of at least one war during the period, and dyads with a mixed territory-policy dispute profile had a 70 percent chance of war. Dyads whose relations were characterized by either regime or policy disputes were far less likely to end up in war. When the focus is shifted to the number of dyads having at least one war during the period, 60 percent were associated with territorial disputes (either singly or in combination with policy disputes). Territorial disputes are clearly more dangerous than disputes based on other issues. This is a central finding in international conflict theory, and it has been repeatedly confirmed. 40 Paul Huth has compiled an impressive data bank of territorial conflicts for the period 1950–90. 41 He documented 129 territorial disputes during this period. His data show that approximately one-third of all borders (116 of 356) were disputed at one point during the period. 42 While challengers initiated some sort of militarized action in about 50 percent of these territorial disputes, resort to large-scale force (the use of 1,000 troops or more) was limited to about 20 percent of those militarized confrontations. 43

246

Chapter 7

Huth is interested in what determines whether a territorial dispute, once initiated, will escalate to high levels of diplomatic pressure and militarized confrontations. (He is not trying to explain war, per se, only militarized confrontations, which could involve threats of force, and the use of military force short of war.) As mentioned, Huth’s data show that escalation was not the norm. In well over half of the 129 territorial dispute cases, no threats were issued and no military force of any type was used. Disputes escalated to the use of moderate levels of diplomatic and political conflict in 32 percent of the yearly observations. Militarized confrontations occurred in only 11 percent of the yearly observations, and these were derived from only about twenty dyadic cases. 44 What were the determinants of escalation? Part of the answer goes back to our previous discussion of nation-centered nationalism. Huth finds that ethnic irredentist issues (as indicated by the presence of minority groups along the border who shared ethnicity with the dominant population in the challenging state) and national unification issues (as indicated by shared ethnicity of people in the challenger and target state) both strongly increased the probability the dispute would escalate to high levels of conflict. 45 Cases of ethnic irredentist disputes that escalated to war would include conflicts over Kashmir and the ethnic Somali Ogaden region of Ethiopia claimed by Somalia. Examples of unification disputes that escalated to war would include the Korean War, the Vietnam War, and Iraq’s attack on Kuwait (1990–91), India’s subjugation of Goa (1961), and conflict between North and South Yemen (1972–90). A follow-up study by Huth and Allee for the years 1919–49 confirms that claims that involve ethnic issues are the type of territorial conflict most likely to escalate. The intangible aspects of territorial issues appear to be more dangerous than the tangible dimensions. 46 States also act strategically regarding their decisions to escalate territorial disputes. The strategic value of the territory in question is an important factor, second only to the ethnic dimension. Huth finds that in about two-thirds of the cases in which strategically valuable territory was at stake the conflict escalated to high levels of pressure (e.g., Iran’s 1971 military take-over of islands in the Persian Gulf, Iraqi pressure on Kuwait in 1990 to cede Warba and Bubiyan Islands, war by Egypt and Syria in 1973 to attempt to regain territory previously lost to Israel). On the other hand, in cases in which the economic value of the territory was the main issue, disputes were unlikely to escalate; only when these issues were combined with political and security issues did they become dangerous. 47 In general, the salience or importance of the territory increases prospects that the disputes over territory will escalate. 48 Huth finds that the strategic military balance is also an important factor in territorial disputes. Very weak and very strong challengers were less likely to escalate than were challengers whose capabilities were relatively equal to the target state. Extremely weak states were unable to escalate effectively, while dominant states had little need to control disputed territory. Of the seventeen crises in which challengers actually resorted to large-scale military force, they possessed clear military superiority in only six. 49 Military balance issues showed up in another respect as well. Challengers were less likely to escalate disputes that involved states at whose hands they had suffered military defeats. In these cases challengers exercised caution. Of course defeated neighbors don’t always exercise restraint—witness Egypt and Syria’s initiation of the 1973 Yom Kippur War and Pakistan’s initiation of the 1999 Kargil War. 50

The Dyadic Level of Analysis, Part I

247

Huth concludes that the escalation of territorial disputes conforms to his “modified realist model,” which combines domestic political variables and strategic international variables. Domestic politics often pushed leaders toward diplomatic and military escalation, while calculations of relative military strength would either accelerate or dampen incentives to escalate. 51 Jaroslav Tir has investigated an interesting subset of twentieth-century dyadic territorial conflicts: conflicts between states previously involved in a secession (for example, IrelandGreat Britain, Bosnia-Yugoslavia, or Eritrea-Ethiopia). He finds that although territorial disputes escalate to fatal MIDs in only about 24 percent of the dyads, the existence of an unresolved border issue significantly increases the probability that the two will be involved in MIDs that result in fatalities. He also shows the importance of the intangible factor of ethnic identity (where a minority in the territory shares linguistic and cultural affinity with the dominant ethnic group in the challenger state). Consistent with Huth and Alle, he finds that this ethnic identity factor is far more important in determining which territorial conflicts will result in fatal MIDs than tangible factors such as economic value (natural resources) or strategic value of the territory. Equally important is the process by which secession occurs. Violent secessions are much more likely to lead to subsequent fatal MIDs between dyads than those in which the secession process was peaceful. Peaceful border changes—including borders changed through secession—rarely lead to violent conflicts. Thus a combination of territorial conflict based on ethnicity and a violent original secession process leads to a high hazard of violent conflict over boundaries. 52 Collectively, these empirical findings make a solid argument that the strong statistical association found between contiguity and war is due not to proximity per se (or to borders per se), but to the fact that states fight over disputed territorial claims involving their own land or land contiguous to them. 53 We can now safely interpret the statistical link between contiguity and war as a product of real disputes over adjacent territories by contiguous states. In conclusion, wars over territory appear to be plentiful in the history of states, and states appear much more likely to go to war over territory than over any other issue—economics, religion, ideology, balance of power, or disputes over an opponent’s policies or regime type. It is arguable whether territorial conflicts have declined as a prime motivation for war in the post-WWII period. More recent territorial conflicts tend to be over the precise boundaries between states—the Iran-Iraq conflict over the Shatt-al-Arab waterway, the Soviet-Chinese dispute over the boundaries along the Amur and Ussuri Rivers, and the Ethiopia-Eritrea conflict over border areas, for example—rather than disputes over the possession of entire areas. However, even wars belonging to the latter category may be found in recent times— Iraq’s 1990–91 war to possess Kuwait, for instance. A final caveat is appropriate at this point. Remember that most wars have multiple causes and motivations. The fact that a pair of states has an ongoing territorial dispute and they subsequently engage in war does not necessarily mean that the war was caused by conflict over territory. 54 For instance, the Six Day War (1967) between Israel and her Arab neighbors occurred in the context of unresolved territorial issues dating to the founding of the state of Israel in 1948. However, to conclude that this was a war over territory—even though it ended with the Israelis taking territories from the losers—would be to miss some really important factors that are necessary to explain the outbreak of the war. Ultimately, the war was not about

248

Chapter 7

territory; it was about the lack of security. Even when territorial disputes seem central to the outbreak of war, they may not be the only causes of the war. Territorial conflict was a major feature of the Iran-Iraq relationship, where the disputed Shatt-al-Arab waterway and numerous land border claims festered over several decades, but both sides had other, nonterritorial grievances that were equally compelling. Territorial Issues, Crises, and MIDs If territorial issues are linked to wars, they are also linked to war’s precursors—crises and MIDs. Most wars do not occur out of the blue; they are normally preceded by a series of crises and/or militarized disputes that create a more or less symmetrical conflict spiral in which the two sides respond to each other’s actions with increasingly hostile and threatening behaviors. 55 The International Crisis Behavior (ICB) project headed by Michael Brecher and Jonathan Wilkenfeld has collected data on 412 international crises from 1918 to 1994. 56 Territorial issues are at stake in almost half of these crises. Using updated ICB data for 1918–2001, BenYehuda establishes that crises involving territory (threats to territory, threats to a state’s existence, and general conflict over territory) are the types of crises most likely to trigger violent crisis management techniques and most likely to escalate to war. 57 Brecher and Wilkenfeld observe that protracted conflicts—defined as including at least three crises between a pair of adversaries during a period of at least five years—are much more likely to exhibit extreme violence than nonprotracted conflicts. Their analysis produces a list of thirty-one protracted conflicts (PCs) since 1918, and 60 percent of all crises occur within the context of one of these thirty-one PCs. 58 Huth’s analysis of the ICB data indicates that many PCs involve territorial disputes, just as enduring rivalries are highly correlated to territorial MIDs. Moreover, the ICB data show that crises are more likely to escalate if territorial disputes are involved. 59 International relations scholars have also uncovered a number of interesting links between territorial issues and MIDs. Hensel finds that over one-quarter of the MIDs occurring in the 1816–1992 period involved explicit conflict over territory issues, and two-fifths of the militarized disputes between contiguous states involved territorial issues. Moreover, territory-based MIDs are more likely to produce a militarized response by the target than other types of disputes; they are more likely to escalate to war than MIDs based on other issues; they are more difficult to resolve; and they have a greater probability of recurring than nonterritorial MIDs. 60 Senese and Vasquez report that the presence of a territorial claim creates an eightfold increase in the probability a dyad will engage in a MID on any issue (not just on a territorial issue). Moreover, MIDs based on territorial issues are more highly prone to fatalities than other types of disputes. 61 However, it is necessary to note that most territorial MIDs do not escalate to war, and therefore they do not constitute a sufficient condition for war, though they do increase the probability of war. 62 Senese and Vasquez analyze data from 1919 to 1992 and confirm the hypothesis that a dyad with unresolved territorial claim is more likely to engage in a MID than dyads without such claims and that once a MID has been initiated, MIDs based on territorial disputes are more likely to escalate to war within five years compared with MIDs based on disputes over policy or regime questions. 63 This is a two-stage process; the results indicate that “it is not the

The Dyadic Level of Analysis, Part I

249

mere presence of a territorial claim that leads to war, but the actual threat or use of force on that claim.” 64 What is important about the Senese and Vasquez study is that it shows that it is the use of coercive tactics (as evidenced by a MID) as a way of dealing with the territorial issue that is the problem. The implication here is that the initial MID is not likely to result in war, but is likely to lead to recurring MIDs over the same issue and that such repetitive MIDs are more likely to result in war than those that do not repeat. Additional research has added even more bricks to this solid wall of evidence. For instance, we know that militarized disputes are more dangerous when they recur over and over again for a pair of states. We also know that territorial issues are notoriously hard to put to rest; they are the type of dispute most likely to recur. In fact, Hensel claims the odds of a territorial dispute recurring are twice as high as for nonterritorial disputes. 65 States are more likely to respond to challenges regarding territorial disputes with military threats or actions of their own, and disputes based on territory are more likely to be reciprocated compared with other types of disputes. 66 It also appears that enduring rivals whose relationship is dominated by territorial issues have significantly more MIDs than rivalries preoccupied with other types of issues. 67 To pull things together, we might snatch two conclusions from Paul Huth. First, Huth warns us that while territorial disputes are linked to militarized disputes and war, only a small percent of territorial-based MIDs escalate to war and most territorial disputes end peacefully. Having said that, Huth then concludes that research efforts have resulted in consistent and convergent findings that “the presence of a territorial dispute is correlated with the escalation of MIDs, international crises, and enduring rivalries or protracted conflicts to war.” 68 Thus, there are multiple pathways from territorial disputes to war. Figure 7.1 suggests the linkages between these various factors. Territory and War: Implications for Peace Research makes clear that the absence of territorial disputes substantially reduces the prospect of war, even for contiguous states. If territorial disputes can be settled by mutual accommodation in formal treaty arrangements, those disputes are unlikely to resurface and serve as a pretext for a future war. It is also possible that borders created through the overwhelming dominance of one of the neighbors (assuming this dominance continues over time) will cease to be contested as well—witness the U.S.-Mexican border. As Vasquez notes, “Most borders once satisfactorily settled remain so for long periods of time.” 69 We have a growing body of empirical evidence that settling territorial claims substantially improves a dyadic relationship. While research on this issue is still in its infancy, most signs point in a positive direction. Kocs’s study of post-WWI boundaries shows that when dyads formally accept the legal status of their boundaries, the probability of war between them is about forty times less than for states that have rival claims over the same territory, and this impact is durable rather than temporary. 70 Huth discovers that when previous disputes have been settled by signed territorial agreements, states tend not to raise claims and contest those boundaries. 71 Owsiak finds the same results over a much longer time frame, 1816–2001; moreover, his finding holds for both democratic dyads and nondemocratic pairs. 72 Gibler’s work also shows that dyads who have settled territorial disagreements through alliances tend not to go to war subsequently. More-

250

Chapter 7

Figure 7.1. Relationships between Territorial Disputes and War. Source: Adapted from Paul Huth, “Territory: Why Are Territorial Disputes between States a Central Cause of International Conflict?” in What Do We Know about War?, ed. John Vasquez (Lanham, MD: Rowman & Littlefield, 2000), 88.

over, these alliance settlements generally dampen rivalries and significantly reduce the number of future MIDs between rivals. 73 Territorial settlements may even help to reduce the probability of nonterritorial disputes. 74 We also know that the most successful and long-lasting settlement techniques, and the ones that reduce chances of future armed conflict, involve mutually satisfactory arrangements of some sort. Hensel reports that almost half of all successful attempts to settle dyadic territorial disputes involved bilateral negotiations, but the greatest success rates came from using binding third-party arbitration or adjudication. Better than three-quarters (76.9 percent) of attempts using these methods were successful in ending “most or all of the claim.” Military force had the lowest success rate—only 9.8 percent. 75 Territorial issues imposed by force of arms— especially between countries of relatively equal strength—are not likely to last. Irridentist sentiments among territorial losers have a tendency to recur, especially if ethno-religious diasporas exist. If shifts in relative capability occur, former losers may have a window of opportunity to regain lost territory. The good news is that by the twenty-first century, most borders have now been set for many decades. Some have been undisputed for years; others, once disputed, have been satisfactorily settled and no longer serve as a source of grievance. This represents a kind of ratcheting effect in which the total number of dyadic disputes should decline on a yearly basis. The net effect of this should be to reduce the probability of war in the future. Additionally, as Vasquez notes, the pacification of territorial issues is supported by the continued development and strengthening of peaceful international norms for dealing with territorial claims and transfers of territory—for instance, the norm that territory taken by force cannot confer legal title. Disputes are more likely if norms are ambiguous or absent. 76

The Dyadic Level of Analysis, Part I

251

However, this good news may be tempered by the persistence of ethno-religious nationalism, which Benjamin Barber refers to as the force of “tribalism”—an international force pushing toward the disintegration of states into smaller and more numerous units. 77 When states break apart—as the USSR, Yugoslavia, and Czechoslovakia did in the 1990s—new boundaries are created, some of which are disputed, some not. Much depends on the process of breakup (peaceful or violent), on whether political boundaries are consistent with ethnolinguistic-religious boundaries, and on the treatment of minorities within these new boundaries. THE ROLE OF SHARED ETHNICITY Several of the studies of territory discussed earlier note that territory issues are often intertwined with issues of ethnic nationalism. In this section, we look at the issue of ethnic nationalism from the dyadic perspective rather than from a single-state perspective as we did in the last chapter. Douglas Woodwell constructs three categories of contiguous dyads based on ethnicity: 78 (1) majority-majority dyads made up of two states in which a particular ethnic group exists as a majority in both states—like North and South Korea; (2) minority-majority dyads in which a particular ethnic group comprises the majority in one state and a minority in the adjacent state—like ethnic Russians in Latvia; and (3) minority-minority dyads in which the same group is a minority in both of the neighboring states—like Kurds in Iran, Iraq, Syria, and Turkey. Woodwell proposes that both majority-majority dyads and minority-majority dyads should be significantly more prone to conflict than the average contiguous dyad. The logic is as follows. For majority-majority dyads, disputes are likely to arise for two reasons. First, since both countries are ruled by essentially the same people, there is always the threat that one state will attempt to annex the other to incorporate the entire nation into a single political entity. Second the two states may have conflicting national visions and disputes between them will arise from competing political ideologies, competing claims to group leadership, and competition for prestige and influence. With regard to minority-majority dyads, the basis of potential conflict is different. Here the homeland state (where the group is the majority) is likely to pursue heavy-handed deterrent policies to prevent the adjacent state from repressing its co-ethnic minority. The state that contains the transnational minority is likely to see the homeland state as a threat to intervene and therefore as a fundamental threat to its political sovereignty and territorial integrity. The homeland state may indeed have irredentist goals that can be accomplished only at the expense of its neighbor. As Woodwell says, “Irredentism represents the ultimate dream of the nationalist leader, while involuntary state dismemberment or dissolution represents the ultimate nightmare for any leader.” 79 And while outright land grabs are now relatively rare, as a substitute, the homeland state may foment and support political uprisings by a co-ethnic minority across the border. Woodwell’s analysis dovetails very nicely with Miller’s theory of state-to-nation imbalance discussed in the previous chapter. While Miller looks at the nature of states and regions, Woodwell focuses more directly on dyads. The results of the two studies both lead in the same direction.

252

Chapter 7

Woodwell’s analysis finds that majority-majority dyads are indeed highly prone to conflict. They are roughly 100 percent more likely to be involved in MIDs than the typical contiguous dyad and are over 200 percent more likely to be involved in MIDs with fatalities. The pattern in minority-majority dyads is somewhat less clear and is highly dependent on how the concept of an ethnic group is operationalized. When ethnic groups are defined broadly (rather than simply in terms of language) the effect is strong. Minority-majority dyads are 83 percent more likely to be involved in MIDs and 190 percent more likely to be involved in fatal MIDs. 80 Finally, Woodwell looks at the effect of internal ethnic uprisings on the propensity for conflict between states. For minority-majority dyads the effect, as one would suspect, is extremely destabilizing for the dyad. When a rebellion takes place in the state with a minority diaspora, the adjacent homeland state is likely to become involved in conflict with its neighbor. The probability of MIDs for the pair increases by over 500 percent and the probability of fatal MIDs increases by 1000 percent. Fortunately, such situations are rare; they never make up more than 3 percent of the dyads in Woodwell’s sample. 81 THE ROLE OF RIVALRY Rivals are pairs of states whose leaders see each other as threatening. Relations between these states tend to be fraught with hostility, and their interactions tend to involve serious and repeated conflicts and disputes with each other over time. We can all call to mind compelling examples: India and Pakistan have fought four wars since their joint independence from Britain; Israel fought four wars with Egypt between 1948 and 1973; and the rivalry between France and Germany was at the heart of three wars in Europe between 1870 and 1945. While we might intuitively know what rivalry is, as social scientists we must be more precise about it. Researchers have taken essentially two different approaches to operationally defining rivalry. The first, associated with Paul Diehl, Gary Goertz, and others, conceives of dyadic rivalries as having two essential components: a longstanding dispute and militarized competitiveness. Goertz and Diehl are most concerned with enduring rivalries, in which a pair of states is involved in a minimum number of militarized international disputes (say, six) over a minimum period of time (say, twenty years). 82 Rival dyads that do not meet the criteria for enduring rivals may be isolated rivals (a very short-lived set of rivals having only one or two MIDs) or proto-rivals (an intermediate category, having three to five MIDs). A recently compiled rivalry data set, covering the years 1816–2001, identifies 290 rivals, of which 115 are enduring rivals and 175 are proto-rivals. 83 The second approach to defining rivalries is most associated with William Thompson and his colleagues, who use the term strategic rivalries rather than enduring rivalries. For Thompson, the criteria are that pairs of states must see each other as (1) competitors capable of “playing in the same league”; (2) the source of threats with the potential of becoming militarized, and (3) enemies. 84 There is no requirement for a minimal number of MIDs or a minimal duration for the rivalry to exist; nor is there a requirement that strategic rivals be relatively equal in military capabilities. They categorize strategic rivalries as being either spatial (where the main disputes are over control of territory), positional (over relative influence and status

The Dyadic Level of Analysis, Part I

253

within the international system or a regional subsystem), or ideological. Colaresi, Thompson, and Rasler identify 173 strategic rivalries from 1816 to 1999. 85 While a definitive “rivalry theory of war” does not yet exist, both camps agree on its rudimentary elements. First, rivals share a mutual history that affects their subsequent relations. As Diehl and Goertz say, rivals are “pulled by future expectations and pushed by the not-so-dead hand of the past.” 86 Their past interactions are likely to involve conflicts and disputes—some settled with violence and coercion, some not. The outcome of initial disputes or crises between states is likely to influence the choices leaders make in subsequent encounters. John Vasquez suggests that as conflicts recur, rivals “become more concerned with hurting or denying their competitor than with their own immediate value satisfaction, and with this, hostility deepens and goes beyond that associated with normal conflict.” 87 Moreover, he suggests that there is a tendency for all issues disputed by rivals to become linked into one— “us versus them”—making them less likely to be resolved peacefully. Instead, over the course of their interactions with each other, political elites in rival states come to “learn” that they need to use coercive tools when engaged in conflicts with rivals. 88 In sum, the existence of a rivalry helps to lock in competitive interactions and to accelerate conflict spirals. Conflicts between rival dyads are not independent phenomena, but are connected; previous disputes increase the chances of future disputes. For instance, Peter Wallensteen reports that 75 percent of all states that become engaged in repeated MIDs experience war. 89 When war occurs between rivals it is the outcome of an interactive process involving mutual hostility and distrust, mutual perceptions of threat, and mutual expectations of violence—all ingrained as the result of dyadic interaction over time. As a result, the existence of a rivalry itself can be seen as a cause of war. 90 Second, interactions between rivals have an effect on domestic politics in both states. Rivalries develop political constituencies—groups that benefit politically from the continuation of a hostile relationship. Pro-rivalry constituencies make it difficult for leaders to pursue cooperative or accommodative strategies with rival states, especially during times of crisis. Such strategies are likely to be derided as signs of weakness or appeasement. If cooperative strategies fail, “accommodationists” are likely to be driven from power by “hardliners” who pursue more coercive methods of interaction, as Vasquez points out in his steps-to-war theory. 91 In addition, rivalry relationships are likely to have an impact on national security organizations. The policies and practices associated with rivalry tend to become institutionalized within government bureaucracies and harden over time. Indeed, institutional actors may acquire a role identity as participants in a rivalry. Third, the interaction between rival states probably has individual-level effects on decision makers. Rivalries endow political leaders with a certain amount of psychological baggage. 92 Colaresi, Rasler, and Thompson suggest that the rational abilities of leaders trapped within the confines of a rivalry are likely to be diminished by various decision-making pathologies such as cognitive rigidity, in-group solidarity, out-group hostility, mistrust, misperception, and selffulfilling prophecies. 93 The result of hostile and threatening interactions over time probably means that leaders are more likely to distrust their adversaries, more likely to develop images of the rival as threatening, more likely to expect conflict with the rival in the future, and more likely to conclude that

254

Chapter 7

the only thing the other side understands is brute force. This is likely to affect the way leaders evaluate and process information. 94 Leaders are prone to seeing the acts of a rival as being more hostile than they are, and they are more likely to fear an attack by rivals during periods of crisis. Leaders imbedded in rivalries are prone to use both the fundamental attribution bias and worst case analysis with regard to rival states, especially in situations of uncertainty and ambiguity. These tendencies all increase the problems of the security dilemma—the dilemma of how a state can make itself more secure without doing things that simultaneously decrease the security of others. 95 As a result of these phenomena, we should expect that (compared with nonrival states) rivals should be more likely to (1) fail in efforts to resolve mutual disputes, (2) engage in protracted conflicts and serial crises with each other, (3) initiate and reciprocate the use of force against each other, (4) escalate conflicts, (5) engage in arms races and alliance formation and other coercive acts in response to acts by the other, and (6) go to war with each other—as one theorist says, “They’ll fight about anything and everything.” 96 Finally, (7) we would also expect that certain kinds of potential conflict triggers—internal political unrest, dyadic power transitions, and disputes over territory—would be more dangerous between rivals than between nonrivals. Rivalry and War: Evidence As we know, some pairs of state are much more involved in international conflicts than others. Conflict is not absolutely pervasive throughout the international system as realists insist; instead we can find zones of conflict (inhabited largely by rivals) and zones of relative peace (inhabited by nonrivals and mutually democratic states). 97 Goertz and Diehl’s first major investigation of rivalry uncovered some rather staggering facts. They report that almost half (49 percent) of wars from 1816 through the twentieth century have taken place between enduring rivals. Likewise, enduring rivalries are responsible for roughly half of all militarized disputes and violent territorial changes during the last two centuries. A dispute between enduring rivals is almost twice as likely to result in war as a dispute between isolated rivals who share virtually no extended history of militarized disputes. Moreover, the chance that a pair of enduring rivals will go to war at some point in their relationship is better than 59 percent. 98 An earlier analysis estimated that enduring rival dyads were eight times more likely to wage war than were nonrival dyads. 99 The most recent accounting of rivalry data produces impressive, though slightly different numbers. A relatively small number of enduring rivalries account for the majority (51 percent) of all MIDs, and the most virulent of these rivalries—a subgroup of forty-seven dyads each of which had a minimum of thirteen MIDs—account for fully 30 percent of all militarized disputes. If one broadens the focus to take in both enduring rivalries and proto-rivalries, then these dyads account for over two-thirds of all MIDs. The chance that enduring rivals will go to war with each other at some point in their rivalry is a hefty 36 percent; for the subgroup of “heavy” enduring rivalries, the chances are truly extraordinary—better than 64 percent. 100 Perhaps the most extensive analysis of rivalry and conflict has been done by William Thompson and his associates, and they report equally startling results. Using virtually the same list of wars as Goertz and Diehl, but using their own definition for strategic rivals,

The Dyadic Level of Analysis, Part I

255

Colaresi, Rasler, and Thompson find that strategic rivals have been on opposing sides in 78.4 percent of wars since 1816, 87.2 percent of the wars in the twentieth century, and 91.3 percent of the wars since 1945. 101 (They characterize the sixteen nonrival wars as being brought about by reactions to more or less spontaneous opportunities or problems—such as the 1898 Spanish-American War or the 1956 Soviet intervention in Hungary.) While roughly three-quarters of all wars have involved strategic rivals, less than half (46 percent) of all rivalry pairs have gone to war. Warless rivalries are most prominent among minor-minor pairs (62.2 percent), while only 15 percent of warless rivals were major-major dyads. Major power rivals were clearly more war-prone than minor-minor rivalry dyads or major-minor rivalry dyads. 102 The same study finds that international crises that involve strategic rivals are more dangerous than those that do not. They are more likely to involve the use of medium or high level threats and the use of militarized tactics; they are also more likely to result in war. It takes considerably less to escalate a crisis situation between rivals than between nonrivals; even nonviolent triggers, low-gravity threats, or nonmilitary issues may induce escalation. 103 In addition, participants in serial crises are more likely to be rivals than nonrivals: rivalry dyads made up 80 percent of those cases in which dyads clashed in three sequential crises and 100 percent of the cases of more than three crises. 104 What determines whether a rivalry will end up in war? Several answers have been put forward. Vasquez argues that territorial disputes are the main factor. Rivalries lacking a territorial dimension are far less likely to end in war than those in which contiguous rivals have territorial disputes. 105 For Vasquez, there are two distinct paths to war for rivals: Contiguous rivals fight each other directly in dyadic wars over territory, and noncontiguous rivals tend to be sucked into ongoing multilateral wars initiated by others. Vasquez tests these propositions using three different sets of rivalries in the nineteenth and twentieth centuries, all based on MIDs and all restricted to rivalries between major powers. 106 (Vasquez believes that the concept of rivalry should be applied only to relative equals.) As we would expect, most interstate rivalries are between neighbors; Vasquez puts the estimate at about 85 percent. 107 He finds that regardless of the set of rivalries used, all contiguous major power rivals were involved in wars with each other; there was no case of contiguous major power rivals who did not fight each other. Furthermore, this is not just because contiguity provides rival states with an opportunity to fight. Each set of contiguous rivals had actual territorial disputes centered on claims to each other’s territory or had mutual claims to territory held by a third party. 108 This seems to be confirming evidence that contiguous major power rivals tend to fight dyadic wars over territory. On the other hand, there were numerous cases (depending on the rivalry set used) in which noncontiguous rivals remained at peace with each other for the duration of their rivalry—the primary example being the U.S./ USSR rivalry. It is also true, however, that several noncontiguous rivals fought wars with each other. However, in every case in which noncontiguous rivals went to war, the war was a multilateral war rather than a dyadic war. There were no instances in which noncontiguous rivals fought each other directly and exclusively. 109 They either joined ongoing wars or joined multilateral wars rather early on. As Vasquez notes, his findings are consistent with Thompson’s distinction between spatial rivalries (between states that tend to be contiguous and whose conflict is centered on territorial

256

Chapter 7

issues) and positional rivalries (between mostly noncontiguous states whose conflicts center around issues of status and influence). 110 An analysis of strategic rivalry patterns by Thompson’s team indicates that spatial rivalries are most likely to be found in rivalries involving one or more minor powers. (This is somewhat at odds with the implications of Vasquez’s study, which focuses solely on major powers.) On the other hand, positional rivalries are more prevalent in rivalries involving a major power. They argue that all major-major rivalries were at least partially positional in nature. Thompson maintains that positional rivals are less likely to fight directly and more likely to be involved in multilateral wars, while spatial rivals fight fairly frequently over territorial disputes. 111 For Thompson and his associates, explaining the occurrence of war requires a combination of dyadic factors—territorial conflict, contiguity, and rivalry. The interactive effects are important. In a massive analysis combining data on rivalries, war, territorial conflicts, militarized disputes, crises, and crisis escalation for the years 1919–92, Colaresi, Rasler and Thompson conclude that the triadic combination of contiguity, territorial disputes, and rivalry has a dramatically potent effect on the possibility that a dispute will escalate to war. And when a fourth factor—mixed democratic/nondemocratic regimes—is added, the concatenation of forces impelling dyads toward war is virtually overwhelming. While the three-way combination of territorial disputes/contiguity/rivalry is associated with a 132 percent probability increase in war onset, adding politically mixed dyads increases the probability of war onset by 149 percent. 112 Colaresi, Rasler, and Thompson conclude that “rivalry is critical to conflict escalation. Conflict escalation can occur in its absence but the combination of contested territory, contiguity, and strategic rivalry is an impressive recipe for conflict escalation.” 113 This is an important empirical discovery, and as we will see in the next chapter, it is very much in line with Vasquez’s steps-to-war theory. Another much discussed linkage is that dyadic rivalries undergoing power transitions or capability shifts end up in war. Many of the theories that we will deal with later—such as long cycle theory, power transition theory, hegemonic war theory, and power cycle theory—all propose in some form that war is most likely to happen been rivals during periods in which there is a change in the balance of relative capabilities, either between pairs of states or in the international system as a whole. 114 Indeed, Diehl and Goertz suggest that the logic of power transition theory is that both power transitions and rivalry must be present for war. 115 Theoretically, enduring rivals should be more sensitive to power shifts than other pairs of states because of the psychological baggage accumulated through previous hostile encounters. The normal threat associated with changes in the balance of capabilities is likely to be magnified in the eyes of rivals simply because rivals are seen as most threatening to begin with. Perceived windows of opportunity and vulnerability created by such shifts are likely to be magnified as well. And rivals should be more attuned to taking advantage of capability shifts than nonrivals would be. Some empirical evidence suggests a relationship between unstable balances of power, rivalry, and war. Wayman discovers that the relationship between power shifts and war is stronger for rivals than nonrivals: While power shifts clearly increases the risk of war between nonrivals, power shifts within rival dyads more than doubles the probability of war (from 14

The Dyadic Level of Analysis, Part I

257

percent to 31 percent). 116 Supporting evidence is also provided by Geller. Using a fairly restrictive definition of enduring rivals based on MID data, Geller finds that his group of twenty-nine rivalries in the nineteenth and twentieth centuries accounted for eighty-five militarized disputes, ten of which reached the level of war. While only eight of the ten war dyads had capability balances that could be empirically determined, Geller found that in none of these eight rivalry dyads was the balance of capabilities stable prior to the outbreak of war; each dyadic balance had been in flux prior to the war. On the basis of his statistical findings, he argues that “an unstable military balance among great power rivals approximates a necessary (although not sufficient) condition for the occurrence of war.” 117 Other studies that deal with wars between great powers (though not necessarily with rivalries) also note the significance of changing balances of capability. 118 We should note briefly two additional connections to rivalry. First, it would appear that most arms races take place within the context of rivalries. To the extent that the presence of arms races may contribute to the potential for the outbreak of war, rivalries provide the most fertile conditions for their birth. 119 Second, diversionary use of force is more likely to take place within the context of enduring rivalries. To the extent that leaders operating under conditions of internal unrest and political vulnerability are tempted to engage in diversionary uses of force, the presence of an external rival presents a ready-made scapegoat. Leaders of states lacking such rivals are deprived of these easy targets. 120 In sum, the empirical evidence points to the existence of enduring or strategic rivalry as an important factor in increasing the probability of war. Logically, the elimination of rivalry relationships would be an important factor in establishing a more peaceful global environment. THE ROLE OF REGIME TYPES IN DYADS This section examines whether certain combinations of regime types make a difference in the war-proneness of dyads. The Democratic-Democratic Dyad We have taken a look at some pretty nasty dyads—contiguous states with territorial conflicts and enduring rivals. Are there any good guys and if so who are they? We have seen in chapter 5 that democratic states, as monadic entities, are not necessarily more peaceful than nondemocratic ones. But when we turn to dyads, one piece of evidence seems nearly irrefutable: Democracies hardly ever go to war against each other. This is sometimes referred to as the democratic peace hypothesis. If the dyadic hypothesis is true, it constitutes a serious threat to the view of realist international relations scholars, who assume that internal political arrangements of states matter little and that all states act similarly—on the basis of power realities. The dyadic hypothesis argues instead that regime type makes a difference. The statistical evidence for the dyadic hypothesis first saw the light of day far outside the academic mainstream. Dean Babst published a piece in The Wisconsin Sociologist in 1964 (later republished in an equally obscure journal called Industrial Research in 1972). Based on war data from Quincy Wright’s A Study of War and on Babst’s own coding of regime types,

258

Chapter 7

the article called attention to the fact that between 1789 and 1941 states with elected democratic governments did not seem to have been involved in wars with each other. 121 The article prompted a skeptical rejoinder by Melvin Small and J. David Singer—the founders of the Correlates of War (COW) project. Their 1976 study of the war-proneness of democratic regimes was also published in a relatively obscure Israeli journal. Using early COW data, they examined fifty cases of interstate wars between 1816 and 1965 and found only two “marginal” instances of war between “bourgeois democracies”: (1) Finland’s joining with Germany to fight the USSR and the democratic Allies in World War II, and (2) “an ephemeral republican France attacking an ephemeral republican Rome in 1849.” 122 This fact appeared somewhat anomalous in light of the primary finding of the study, that there was little difference between democracies and nondemocracies at the monadic level with regard to their participation in wars or initiation of wars, and it has piqued the interest of scholars ever since. 123 Around the same time, Rudolph Rummel developed the “joint freedom proposition”: libertarian (democratic) systems mutually preclude violence; violence will occur between states only if one is nonlibertarian. 124 Examining data from 1976 to 1980 he found no instance of violence between politically free states or between states that he termed “free”—that is, both politically free and economically free. 125 While this evidence was quite limited in its time span, more evidence by other researchers—both quantitative and qualitative—soon followed. One result of the early findings of a democratic peace was that international relations scholars “rediscovered” Immanuel Kant’s predictions in the late eighteenth century of a future separate peace between what he called republican states. In recognition that the democracies of the current era are a relatively new phenomenon, and in the desire to determine whether Kant’s claim for a separate peace between republics had finally been achieved two centuries later, several studies have investigated a joint peace between “liberal” regimes or republics. Michael Doyle has emphasized the Kantian basis for the democratic peace. Using COW data and a systematic classification of regime type for all states involved in wars in the last several centuries, Doyle has documented a liberal peace beginning in the eighteenth century. 126 A historical survey by Spencer Weart, aptly titled Never at War, also investigates a separate peace between republican states, a category that can include both democracies and oligarchies (not all republics are democratic, but all democracies are republics). Weart concludes flatly that “well-established democracies have never made war on one another,” but he also finds that “well-established oligarchic republics have scarcely ever made war on one another.” 127 Thus, he claims to have documented a separate peace between the same kinds of republics—a democratic republican peace as well as an oligarchic republican peace. Rather than use the term “republic,” an admittedly slippery concept, most researchers prefer to focus on a peace between modern democratic states. In this form, the proposition— now referred to as “the democratic peace”—seems to have now gained overwhelming support among Western scholars. In the last twenty years, probably no area of inquiry in international relations has received as much examination. The results are heavily supportive of the dyadic hypothesis, even when one simultaneously considers the effects of other potentially relevant variables such as contiguity, joint alliance membership, and level of economic development. 128 Not only is the relationship between joint democracy and peace robust, it appears to

The Dyadic Level of Analysis, Part I

259

be nonspurious. In other words, the strong relationship is not caused by some other unidentified third variable that accounts for variation in both regime type and conflict. 129 There are, inevitably, some disputes about this that we will note shortly. Marginal exceptions aside, there has not been a real war between democracies in well over a century and a half. Indeed, international relations scholar Jack Levy has famously concluded that the lack of war between democracies is “as close as anything we have to an empirical law in the field of international relations.” 130 James Lee Ray has put forward a “friendly amendment” to this law. Since democracy is a continuous concept (democracy is something that exists in degrees—from very little to a lot) the law should really be expressed as, “states that have achieved a certain level of democracy . . . have never fought wars against each other.” 131 The weight of all this evidence leads to the almost inescapable conclusion that joint democracy is a sufficient condition for peace. Regardless of any other conditions, if two states are both democratic, they should be peaceful. On the other hand, it is abundantly clear that joint democracy is not a necessary condition for peace. Most dyads are at peace with each other most of the time, whether or not they are both democratic. Joint democracy is not the only thing that produces peace. 132 Nor is the absence of joint democracy sufficient to cause war. 133 Of course, quite a lot depends upon how one defines a democracy. The more rigid the definition, the less likely one is to discover war between democracies. The broader the definition, the more one is likely to uncover “marginal cases” of interdemocratic warfare. And since war is a relatively rare event, the number of marginal cases and how one interprets them, can make considerable difference on how we assess the validity of the dyadic proposition. 134 Since democracy is something that exists in degrees, it is useful to be able to judge the degree of democracy a state possesses. Social scientists have developed several scales to judge the degree of democracy, based on things such as competitiveness of executive branch selection, competitiveness of legislative selection, level of constraints on executive action, freedom of the press and other media, and so on. The most frequently used measures are the Freedom House scale and the Polity index. While such scales do rate states differently, especially when the analyses extend backwards in time, the scales are largely in agreement overall and produce generally similar results. And support for the democratic peace hypothesis is substantiated regardless of the scale that is used. 135 Let’s take a look at some of the marginal cases of war between democracies. A study by Gleditsch and Hegre actually lists seven “anomalous cases” of democratic-democratic war between 1816 and 1994. 136 Using Polity III data on regime characteristics, their criteria for democracy is to subtract a country’s score on the Polity III autocracy scale from its score on the democracy scale. A sum of +3 is required for a country to be classified as a democracy. Other scholars add further anomalous cases. Russett presents his own list of twelve “candidate” wars between democracies—only two of which overlap with the Gleditsch and Hegre list. 137 The two lists are presented in table 7.2. Some of these cases can be dismissed because they do not meet the test for interstate wars involving sovereign states on both sides: the U.S. Civil War, the Boer War, and the Second Philippine Wars can be rejected on this basis. The first was an internal war of secession and the other two were colonial wars for independence. 138 One case can be rejected on simple pragmatic grounds: In WWII, while Finland was allied with Germany against the Allies, it

260

Chapter 7

fought actively only against the Soviet Union, a nondemocratic state. The level of casualties between Finland the democratic Allies would not even begin to approach the required 1,000. Similarly, Lebanese-Israeli fighting—confined to the air—was so limited in the 1967 war that casualty levels fall far short of the level required for a war. 139 Other cases can be debated on the grounds that one of the countries objectively lacked sufficient democracy. Gleditsch and Hegre conclude that neither Lithuania nor Syria nor Pakistan were stable, established democracies at the time of the outbreak of war. They make the same judgment for Spain in 1898. Russett agrees, emphasizing that in spite of universal male suffrage and a bicameral legislature, the Spanish king had control over the legislative and executive branches—though he says the Spanish-American War is the most difficult case to call. Russett judges that in 1948, Lebanon was not fully democratic and technically Israel was not either—it had not yet held an election. He makes similar assessments of Ecuador and Colombia in 1863 (both governments came to power in revolutions) and France and the Roman Republic (both ephemerally democratic). While both France and Prussia had made steps toward democracy by 1870, neither of them had achieved a sufficient level. Some countries are officially coded as democracies for a particular year, but before the year is out (and before the war begins) they revert to autocracy—Pakistan in 1971 and the Greek Cypriot state in 1974. Gleditsch and Hegre find that raising the definitional bar for democracy from +3 to +8 eliminates all of the anomalous cases except Finland. 140 One classic marginal case is the War of 1812 between the U.S. and Great Britain. The definitional issue revolves around whether Britain was a democracy in 1812. Most political scientists would say no: it was not until the Reform Acts of 1832 or 1867 that enough people were eligible to vote to make elections truly competitive and make Britain truly democratic. Other than the Spanish-American War and the War of 1812, the most contentious case is probably imperial Germany in 1914. Ido Oren reminds us that the concept of democracy is a

The Dyadic Level of Analysis, Part I

261

moving target. 141 Studies of the democratic peace use indicators of democracy drawn from contemporary American thinking about what constitutes a democracy. But political scientists in the United States and elsewhere viewed the concept differently in the nineteenth and early twentieth centuries. Prominent American political scientists such as Woodrow Wilson did not consider Germany to be an autocratic state but a federal constitutional state with the rule of law, efficient administration, a civil service meritocracy, and progressive legislation. And for these earlier political scientists, the presence of a mass electoral process and checks against executive power were not necessarily the hallmarks of a modern democracy. Thus, if imperial Germany is viewed by the standard of the times as a constitutional republic (as “one of us”), then the democratic peace theory is blown out of the water. Doyle’s analysis of imperial Germany straddles the line. He says that Germany was a liberal state under republican law with regard to domestic policy matters but not on foreign policy issues. Here, the emperor’s direct authority over the military and his active role in foreign policy, combined with the independence of the military from civilian control and the lack of effective subordination of the chancellor to the legislature, disqualifies imperial Germany from inclusion, a judgment with which Russett concurs. 142 John Owen quotes other contemporary sources to argue that the prevailing opinion in Britain and the U.S. was that Germany was undemocratic. 143 The most recent marginal case of war between two democracies is the Kargil War between India and Pakistan in 1999. Russett and Oneal count it as a war (there were at least 1,000 battle deaths, though some were clearly guerrilla forces rather than state militaries) and both India and Pakistan rate as democracies on the Polity III index (at the threshold level of +7 on a scale that goes from −10 for extremely authoritarian states to +10 for extremely democratic states). 144 All in all, there is enough suspicion that marginal cases of war between democracies have existed to state the dyadic hypothesis in terms of probabilities rather than certainties: Democratic states are extremely unlikely to fight each other. The Autocratic-Democratic Dyad The most war-prone dyad of regimes seems to be mixed democratic-autocratic dyads. 145 Mansfield and Snyder’s study of dyadic relations shows that in such a dyad, if the least democratic state in the pair becomes more democratic, the probability of war is reduced. However, if the degree of democracy in the most democratic country in the pair is increased by one standard deviation, the predicted probability of war increases from 40 to 65 percent. The most dangerous pairs are those comprised of a strong democracy and a strong autocracy. 146 The degree of political distance between regimes would appear to be important. A sophisticated statistical analysis of dyad conflict from 1816 to 2000 by Scott Bennett confirms this finding, though with a caveat regarding mixed dyads. 147 It also confirms that, with regard to the most serious MIDs, the most dangerous pairs of states are highly autocratic-highly democratic pairs.

262

Chapter 7

The Mixed-Regime Dyad Bennett extends his analysis to mixed-regime dyads. These are dyads in which both states inhabit positions in the middle of the regime classification scale; they are not entirely democratic nor autocratic but have characteristics of both. The evidence indicates that such a combination puts the dyad at risk of war just as much as the autocratic-democratic dyad does. Their regime similarity provides no antidote for war. The Autocratic-Autocratic Dyad Though there is some dispute about this, it appears to be the case that autocratic dyads also have had fewer wars than might be expected by pure chance, creating an “autocratic peace.” Other evidence from the investigation of MIDs suggests that while democratic dyads are by far the most peaceful dyads, there is little difference between autocratic pairs and mixed pairs. For example, Russett and Oneal find little evidence for the “autocratic peace,” at least with regard to MID involvement. They argue that this is to some extent to be expected because there is still significant political distance between different types of autocratic states—a category that includes theocracies, monarchies, fascist states, and communist states. 148 Bennett, on the other hand, does find evidence of an autocratic peace, however its pacifying effect is much weaker than the pacifying effect of joint democracy. He also finds that the effects are most powerful for the most extreme (coherent) autocracies, just as they are for the most extreme (coherent) autocracies—regimes that share all or most of the characteristics of autocracy (or democracy) rather than just a few. These pacifying effects are also most potent with regarding to the most serious MIDs, those in which force is used or those that result in war. And, as noted in the last section, there are no peaceful effects for jointly mixed regimes. Pairs of states that are both partially democratic/partially authoritarian regimes in the middle of the scale are not noticeably peaceful toward each other; therefore it may not be political similarity per se that accounts for reduced conflict. Regime Dyads and Conflict Short of War The dyadic peace enterprise has extended beyond investigations of dyadic war, to investigations of the behavior of democratic dyads in conflict short of war—rivalries, crisis behavior, and the initiation and escalation of militarized disputes. As we know, enduring rivalries are an important cause of international conflict. By all indications when both parties in the rivalry become democratic, the rivalry tends to end. 149 With regard to militarized disputes, Chan’s summary of the evidence is that during the nineteenth century the number of disputes between democracies were actually higher than would be predicted by chance, but lower than expected by chance in the twentieth century. 150 Gleditsch and Hegre look at war involvement and MID involvement from 1816 to 1994. They compare the percent of total dyad years that various pairs of states were involved in conflict (both wars and MIDs). They find that dyads involving two democracies were significantly less involved in wars and MIDs than were mixed dyads (democracy-autocracy) or nondemocratic dyads. 151 Maoz and Abdolali’s dyadic analysis of MIDs from 1816 to 1976 provides evidence that democracies are less likely to engage in lower-level MIDs with each other than other

The Dyadic Level of Analysis, Part I

263

types of dyads, although they do participate in MIDs with limited use of military force and mild levels of violence. Nevertheless, the probability of a democratic-democratic MID at any level of violence appears to be significantly smaller than what would be expected by chance. The frequency of MIDs involving democratic dyads was particularly low in the post-1945 era. Overall, democratic dyads “had significantly less than their fair share of conflict.” 152 A follow-up study by Maoz and Russett looking at politically relevant dyads from 1946 to 1986 found that the most mature democracies—those that had been stable for longer periods of time and that used lower levels of political violence against their own citizens—were the least likely to engage in MIDs with other democracies. 153 Reed’s analysis of MIDs from 1950 to 1985 strongly suggests that while joint democracy inhibits dyads from becoming involved in the onset of a dispute, it has little effect on subsequent escalation of the dispute. Jointly democratic dyads avoid war largely because they rarely become involved in disputes to begin with. 154 This conclusion is seconded by Senese and Vasquez’s study of the outbreak and escalation of territorial disputes. Joint democracy has a dampening effect on the onset of territorial disputes, but a much weaker effect on escalation of disputes to war. 155 Bennett and Stam undertook a massive analysis of all politically relevant dyads (pairs of states that were contiguous or that included at least one major power) over the period 1816–1992. Among other things, they were trying to statistically assess a large number of variables that might account for the initiation of MIDs and the escalation of MIDs to war. 156 Their analysis shows that both democratic dyads and autocratic dyads are less likely than the baseline mixed regime dyads to engage in militarized disputes. But this tendency is even stronger for autocratic pairs than democratic pairs—with the exception that with regard to disputes that escalate to war, the risk of democratic dyads being involved in war is significantly lower than for mixed dyads or purely autocratic dyads. At lower levels of escalation, the pacifying effects of mutual autocracy are somewhat greater than those of mutual democracy. With regard to the initiation of MIDs, they find that democracies are more likely to initiate disputes with autocracies than autocracies are with democracies. However, when democracies initiate disputes, it is likely to be without the use of force. The composition of the dyad has something to do with the chances of going to war. Militarized disputes between democraticautocratic dyads are more dangerous than those between similar types of regimes. The risk of disputes escalating to war lessens as the initiator’s level of democracy rises, but this is only true if the target is another democracy or an anocracy. The risk increases if the target is an autocratic state. 157 As Morgan and Schwebach attest, the vast majority of wars and disputes involve at least one state that is nondemocratic. This suggests to them that the presence of such a dyadic combination is almost a necessary condition for war. 158 These studies make abundantly clear that while dual democracy makes it much less likely that such pairs will initiate militarized disputes, it does not totally eliminate the possibility. Democratic dyads do get involved in disputes with each other in which threats of force or the use of low level force does occur, but the use of force at the highest levels is rare. The fact that disputes between democracies are less violent than those between other pairs of states is buttressed by the fact that disputes between democracies are also shorter—many of them

264

Chapter 7

lasting only one day. More important, territorial disputes between fully institutionalized, mature democracies tend to be increasingly rare in the post-1945 era, virtually eliminating the most dangerous type of conflict issue. 159 The studies cited above are also consistent with the rather robust empirical findings that joint democracy is associated with the use of peaceful methods of resolving international conflict. It is well established that democratic dyads are more likely than other types of dyads to settle conflicts through negotiated settlements, through third-party mediation and binding arbitration, and through international organizations. 160 The Democratic Peace: Theory Despite the observed behavior patterns of democracies, exactly why a pair of democratic states should be free from the scourge of war is not entirely clear. There are several lines of argument, many of which overlap: the normative-cultural argument, the structural-institutional argument, the rationalist approach, and the Kantian approach. We will discuss each of them in turn and also examine some potential flaws in each argument. The normative-cultural argument is that expectations of war and threats of war between democracies are reduced by the presence of common political culture and ideals, such as tolerance for diversity and respect for the freedom of others. Within democracies the expectation is that external conflicts with other democracies will be resolved as they are at home— with resort to mutual accommodation, compromise, the application of legal principles, and peaceful problem-solving. When conflicts arise between democracies, leaders in both states will trust and respect each other. In other words, democratic leaders “externalize” norms of behavior—such as the use of compromise—which they have applied to their internal, domestic conflicts. War is also unlikely due to the creation of mutual identity and sympathy, by stronger people-to-people and elite-to-elite bonds, by the ability of interest groups within these countries to form transnational coalitions, by more frequent communication, and by more positive mutual perceptions. The result is that in times of crisis, other democracies benefit from the assumption of amity and peaceful intent, whereas nondemocracies are likely to suffer from a presumption of ill will and enmity. The normative-cultural argument emphasizes norms, ideas, images, and perceptions held by foreign policy elites (and publics). International relations theorists in the constructivist camp would emphasize that democracies’ views of each other are mutually constructed. The mutual respect and preference for peaceful settlement of disputes that characterizes relations between democratic states is the product of an interactive process that develops over time. Democratic leaders watch other democracies; they compare the political structures and the foreign policy behaviors of other democracies with their own. Over time, they mutually perceive their similarity and develop peaceful attitudes and practices toward the other. 161 According to the normative-cultural argument, democracies will only fight wars that are consistent with liberal democratic principles: wars that are necessary in self-defense or wars to rectify humanitarian abuses and to democratize others. As many critics note, imperial and colonial wars fought by democratic Britain, France, and Belgium in the late nineteenth and early twentieth centuries do not fit this argument very well. 162

The Dyadic Level of Analysis, Part I

265

The normative-cultural argument has drawn heavy fire from critics. For instance, Sebastian Rosato argues that there is little evidence that democracies treat each other with mutual trust and respect. One only has to look at the numerous covert interventions undertaken by the United States to attempt to undermine or overthrow various left-leaning governments during the Cold War: Iran (1953), Guatemala (1954), Indonesia (1957), British Guyana (1961), Brazil (1961, 1964), Chile (1973), and Nicaragua (1984). He argues that in these cases democratic trust and respect was less important than economic and security interests. 163 Democratic peace theorists make two counterarguments. First, the target states were transitional regimes that were not clearly and demonstrably democratic enough to generate mutual trust and respect. Second, however odious these interventions were, democratic norms and constraints were still at work: These interventions had to be covert rather than overt because American officials feared a public backlash if the operations were openly admitted. 164 John Owen amends the normative argument by emphasizing the importance of the perceptions of political elites: The democratic peace only works if liberals in one state actually perceive the other state to be democratic. Slippage between perception and reality may account for marginal cases in which democracies and near-democracies fight each other. 165 The line between democracy and nondemocracy frequently includes substantial gray areas, especially if states are in transition from one type of regime to another. Moreover, what constitutes a liberal democracy changes over time and liberals will disagree over the proper criteria. Generally, however, people tend to identify countries that have the same kind of political institutions as their own state as being democratic. 166 Owen argues that in the nineteenth century, most American believed that only republics were democracies, thus excluding all monarchies. Therefore in 1812, the liberal consensus in the U.S. decreed that Britain was not “one of us.” 167 Owen’s analysis of crises and war emphasizes that there was almost always an internal split over whether the opposing state was a democracy. As Rosato explains, this simply leads to ad hoc explanations of the democratic peace: If countries decline to fight, then they must have perceived each other as democratic; if they do fight, then they must have perceived each other as not democratic. 168 Two of Owen’s cases of transitional democracies—the Mexican-American War and the Spanish-American War—are illustrative. 169 In either case, if the other state had been deemed to be a sister republic, war would have been politically and normatively unthinkable. In the 1840s there was a contentious split in American opinion as to whether postrevolutionary Mexico was still a republic (as Whigs like Henry Clay and John Quincy Adams argued) or whether successive coups had removed it from the list (as Jacksonian Democrats like President Polk and Secretary of State James Buchanan believed). By 1846 Polk and his Democrats had won the argument and war was the outcome. The situation of Spain in 1898 is also problematic; it was at least partly democratic and scholars to this day continue to disagree about how to label it. But Spain had come to be seen by many American leaders as a despotic monarchy or a country with a failing transition rather than a liberal democracy (it had been a constitutional monarchy since 1873); the perception was due in large part to its brutal treatment of its own citizens in Cuba. 170 In both cases liberal elites, using democratic institutions of free speech and threats of electoral punishment, forced reluctant leaders to war with states perceived as not truly democratic.

266

Chapter 7

When elites do perceive an adversary to be democratic, then liberals in the first state can use the democratic institutions to prevent war. Here norms and institutions work together synergistically to create a democratic peace. Thus the U.S. avoided war with Spain in 1873 over the Virginius affair because important political elites in the U.S. argued that Spain had recently become a republic with the overthrow of the monarchy. The structural-institutional argument is that the institutional restraints on executive branch war-making found in democracies are at work independently in both countries in the dyad rather than just one. In each country there are constraints on executive action that make it difficult for a single person or a small group to decide in favor of war: There is civilian control of the military; there is decision making by elected representatives who are accountable to the public through regularly held competitive elections; and war is openly debated in a pluralistic process in which there is a free press and viable opposition groups who can express their opinions and make an impact on government policy. The primary assumption is that if the public (or their representatives) are allowed to make the decision for war or peace, the decision will reflect the desire of the majority to refrain from activity that might bring death, destruction, and severe economic costs to themselves and their homeland. Both the public at large and antiwar interest groups have the ability to constrain and derail a government’s desire for war. If a decision for war by one democratic state is difficult, it should be doubly difficult for two democratic states to obtain the consent of the governed in favor of war. Not only is each individual state constrained, but each also perceives the other to be constrained, mutually reducing perceptions of threat. Moreover, the nature of open, democratic institutions is reputed to have two additional impacts. First, the openness of the system is likely to prevent rapid mobilization of military forces and the use of preemptive, surprise attacks. Other states therefore have less reason to fear democracies and suspect them of launching an attack out of the blue. Second, the openness of democracies presumably means that when democracies issue threats or warnings of their commitment to act in certain ways (or prevent others from acting in certain ways), these statements will be credible. Democracies are less likely to successfully deceive others because information in open societies makes this difficult. Thus, wars based on misinformation are less likely between two democracies. Critics attack the logic of this argument. If democratic institutions effectively constrain violence, logically this should happen regardless of whether the opponent is democratic or autocratic. But this does not appear to be the case. If the logic of institutional constraints is correct, then democracies should be more peaceful in general than other types of regimes, and this is largely untrue. Contrary to the institutional argument, public opinion in democratic states is frequently supportive of war—even against states that might objectively be classified as democratic or against states who realistically pose little threat. Thus pressure from the mass public and from Congress pushed President McKinley to war against Spain in 1898 rather than constraining him. Public opinion was generally, though not totally, supportive of U. S. military action against Iraq in 2003. In part, the failure of democratic institutions to restrain executives may simply be due to the fact that, unless the war takes place on one’s own soil, the costs of war are not shared equally; only a small portion of the population will be directly affected. Moreover, democratic

The Dyadic Level of Analysis, Part I

267

publics are not immune from the forces of nationalism. And certainly executive branch officials play an important role in shaping public opinion and rallying nationalistic sentiment against presumed enemies. There is no guarantee that antiwar groups will have more influence than pro-war groups. Executives may even be able to get around constitutional barriers, knowing that once the war begins, opposition will be overtaken by a rally-round-the-flag effect. The rationalist or strategic choice approach is associated with the work of Bruce Bueno de Mesquita, Robert Fearon, and others who study international relations using rational choice models or bargaining models. 171 It is seen by many as a theoretical buttress to the structuralinstitutional argument. It argues that democracies are more peaceful because the effects of loss are more severe for their leaders. Its interrelated propositions are as follows: 1. It is assumed that all states (and their leaders) are rational and make decisions based on a strategic cost-benefit analysis. 2. A primary motivation for all leaders is the desire to remain in power. 3. To stay in power leaders need to receive the support of a significant segment (a winning coalition) of the society who influence the selection process—“the selectorate.” Leaders in democratic states have to satisfy a much larger selectorate than leaders of autocracies. 4. Leaders who depend on small winning coalitions (typically autocrats) can maintain support through providing private goods to this small group of influential supporters, regardless of the domestic or foreign policies they pursue. On the other hand, leaders who require large coalitions (as in democracies) must stay in office through widespread provision of public goods, which are achieved through successful domestic and foreign policies. Thus, autocrats can afford to lose wars if they still have the resources to provide private goods to their supporters, while democratic regimes must have successful public policies (like winning wars) in order to provide public goods (security) for a large selectorate. 5. The domestic political costs for using force (especially if the use of force fails to achieve the state’s goals) are higher for leaders in a democracy than for autocratic leaders. Losing a war almost always ensures political defeat for democrats, but autocrats are more likely to hold on to office if they lose a war. Therefore, autocrats are likely to take considerably more risks than democratic leaders. Democratic leaders anticipate that they will be held accountable for political-military failures by loss of office; therefore, they will avoid wars, especially if the chances of success are poor and the costs high. This is less true of autocratic leaders. 6. Once involved in a war, democratic regimes will commit a relatively large portion of the state’s resources to winning the war. (This helps to explain why democracies tend to win their wars.) Autocracies are less likely to commit a high level of resources to war since the leader’s survival is less likely to depend on the outcome of the war and since significant resources need to be reserved for the provision of private goods to the regime’s supporters. Thus democracies, with the need to be very sure of victory and knowing that other democracies will invest significant resources to their own war

268

Chapter 7

effort, will be highly unlikely to choose war against a democratic opponent. The democratic peace works in part because deterrence between two democracies is stronger than between other types of dyads. 172 7. The political and economic costs of one democracy’s using force against another democracy will be seen as extremely high, compared with using force against nondemocracies. The mere fact that political elites might find themselves in a situation in which war with another democracy is a policy option would be considered evidence of leadership failure. 8. Democracies might be induced to fight other states (even other democracies) if the likelihood of victory was extremely high. This helps to account for the initiation of wars against weak states and colonial-imperial wars fought by democracies. 9. Democratic states are uniquely able to make credible commitments because their leaders will suffer greater “audience costs” if they back down. Since their opponents are well aware of this, they are inclined to act cautiously. A major criticism of this combined institutional-rational argument is that it may not be true that democratic leaders are more accountable than autocrats and therefore more fearful of the negative political effects of taking their country into a losing or costly war. Rosato’s research suggests that although democratic leaders are more likely to be removed from office for participation in a losing or costly war, the difference is not all that great. A good percent of autocrats lose office as well. And while democratic leaders are highly unlikely to suffer any personal punishment (imprisonment, exile, or death), 29 percent of autocrats who lost wars were punished and 27 percent of autocrats who were involved in costly wars were punished. One can only conclude from this that democratic leaders do not face greater expected costs from bungling a war decision, nor are they more accountable than autocrats. There is no reason to believe that democratic leaders should be more constrained or more cautious than autocrats concerning decisions about war. 173 Likewise, democratic leaders can win wars and still suffer electoral defeat: George H. W. Bush lost the 1992 election to Bill Clinton even though he engineered a very popular win over Iraq in the 1991 Gulf War. The three explanations discussed thus far are probably best seen not as competing explanations for the democratic peace, but as complementary explanations. There is no reason why they cannot all be at work simultaneously. Indeed, the fourth set of arguments, the Kantian approach, suggests that several factors must act simultaneously to produce the democratic peace. The Kantian approach is based on Immanuel Kant’s 1795 essay “Perpetual Peace.” Kant argued that future international peace would be based on three reinforcing achievements, representing three legs of a triangle. 174 First, states should create governments based on “republican” constitutions. Many scholars of the present age have more or less interpreted this to mean the creation of democracies, but Kant had something different in mind. As Michael Doyle explains, Kant’s definition of republican government was confined to regimes that had (1) private property and market-based economies and (2) political systems based on representative government with separation of powers and equal legal or juridical rights for all citizens. 175 In Doyle’s interpretation, “Kant distrusted unfettered, democratic majoritarianism,

The Dyadic Level of Analysis, Part I

269

and his argument offers no support for a claim that all participatory polities—democracies— should be peaceful either in general or between fellow democracies. Many participatory polities have been non-Liberal.” 176 The second leg of the triangle for Kant was that republican governments should enter into a loose “pacific union” based on international law. This union—an international federation of free states based on international organizations—would provide for collective security and generally facilitate cooperative relationships and help to resolve conflicts. As this union or federation expanded to include more republics, it would create a larger and larger zone of peace involving its members in a complex web of integrating institutions. The third side of the triangle was that republican states should adhere to “cosmopolitan law” by which individuals on foreign soil would be treated with hospitality. By extension, cosmopolitan law meant that states would accept a set of rules that supported international commerce and free trade. The resulting increase in trade would make war more difficult because choosing war would mean disrupting beneficial trade. Plus, states will be tied together by a network of commercial interactions that reinforce norms of mutual accommodation and reciprocal adjustment. Thus, republican (democratic) states would join together in loose associations (international organizations) to promote their collective security and other mutual interests, and would engage in global commerce with each other. These three factors would bind democratic states together, and their effects would be interactive and mutually reinforcing. The result would be a “separate peace” among republican states. Thus, peace among democratic states would not be (and cannot be) the result of mere joint democracy, but of the combined effects of joint democracy, mutual trade, and membership in international organizations. Kant realized that republican states would not necessarily be peaceful (the monadic hypothesis). However, republican caution in foreign affairs should replace monarchical caprice, and liberal wars should be fought for popular, liberal purposes only. 177 In Triangulating Peace, their exhaustive study of dyadic militarized disputes from 1885 to 1992, Bruce Russett and John Oneal argue that the democratic peace can best be explained by Kant’s original reasoning. They first argue that democratic governments are more peaceful than other types of governments. Their analysis indicates that increasing the degree of democracy in the least democratic state in any dyad by one standard deviation reduces the risk of a military dispute by 42 percent for that dyad. On the other hand, reducing the degree of democracy in the least democratic state in the dyad by one standard deviation increases the chances of having a MID by 69 percent. 178 Second, they show that trade interdependence between pairs of states also contributes to reduced likelihood of conflict, as does the general openness of a state’s economy to integration in the global trading system. If both dyadic trade interdependence and general openness are increased one standard deviation above the baseline levels, the probability that a dyad will experience a militarized dispute falls by 52 percent. 179 While democracy and trade are highly correlated with each other, they each make a strong independent impact on the likelihood of conflict. Neither factor eliminates the statistical importance of the other. This means that whether states are democratic or not, if they are economically interdependent, this still reduces

270

Chapter 7

their chances of conflict with each other. However, pairs of states that are democratic, interdependent, and open to global trade are most likely to be at peace. 180 Third, Russett and Oneal show that mutual memberships in intergovernmental organizations (IGOs) make an important independent contribution, above and beyond those of trade and democracy, to reducing militarized disputes between pairs of states (though the pacific benefits of IGO membership are smaller than the other two Kantian variables). And this is true in spite of the fact that democracy, trade, and IGO memberships are all interrelated with each other. Moreover, when the values of all three Kantian factors are increased above the baseline values for all states, the combined effect is to reduce the risk of disputes by 71 percent. 181 Democracy, interdependence and international organizations are mutually reinforcing factors. 182 Finally, they show that the three Kantian variables form “virtuous circles” or mutual feedback loops. Not only do these three Kantian variables each directly contribute to reduced likelihood of conflict, but they also have indirect effects by contributing to each other. Joint democracy promotes joint trade and IGO membership. Democracies are more likely to trade with each other and to join IGOs with other democracies. The creation of greater global trade creates the need for international organizations to regulate this commerce. And states who trade with each other are more likely to share memberships in IGOs. 183 Alternative Explanations for the Democratic Peace Some scholars have attacked the dyadic peace as a phenomenon attributable to things other than mutual democracy. The debate is really about whether, in a causal sense, it is mutual democracy that is truly responsible for dyadic peace, or whether some other factor is really doing the heavy-lifting. Several critiques are worth mentioning. First, it is possible that the relationship is simply a statistical oddity. As Ray calculates, “Something on the order of 99 percent of all pairs of states in the world have peaceful relationships in a given year.” 184 Since war is relatively rare and democracies have, until recently, been fairly rare as well, the absence of war between democracies might be the result of pure chance. Moreover, for much of the period prior to 1945, few pairs of democracies were proximate enough to be in a position to fight each other. Thus the absence of war does not necessarily confirm the theory of the democratic peace. 185 The historical universe of possible dyads in which democracies may have had the opportunity to fight each other is small, and this magnifies the importance of exceptions to the rule. 186 And as we have already seen, the number of exceptions (marginal or borderline cases) is actually pretty large. On the other hand, quite a few recent studies have addressed this issue and demonstrate that the absence of war between democracies is statistically significant (and thus unlikely to be due to chance) despite the numerical rarity of both democracies and wars and controlling for contiguity. 187 Second, while it may be true that democracies have rarely fought each other, it is not clear that this is actually due to their being democracies. Realist explanations may be just as effective in explaining why democracies have not fought each other. Christopher Layne’s case studies of four crises—the Trent Affair (1861) between the U.S. and Britain, the AngloAmerican Venezuelan Crisis (1895–96), the Fashoda Crisis between Britain and France (1898), and the Ruhr Crisis (1923) between France and Weimar Germany—indicate that in

The Dyadic Level of Analysis, Part I

271

each case war was quite possible and that at least one party in the crisis was prepared to use force to obtain its goals. Neither democratic institutions nor shared political culture were effective restraints on the use of force. Layne argues that in each case one of the states decided to refrain from force out of considerations of power and geostrategic realities. 188 Third, evidence for the democratic peace suggests that the phenomenon is much stronger in the post-WWII era. An analysis by Farber and Gowa indicates that the difference between democratic dyads and other kinds of dyads is not statistically significant until the post-WWII era, in terms of both war involvement and MID involvement. 189 This suggests to them that the democratic peace is likely to be an artifact of the Cold War bipolar world in which democratic states (led by a democratic hegemonic state, the U.S.) shared common interests in opposing communism and joined in a major alliance system (NATO) in pursuit of this interest. Joint strategic interests and the associated alliance ties might be more important than joint democracy in producing the democratic peace. And since most democracies are in Western Europe or the Western Hemisphere, the democratic peace may only be a regional phenomenon as well. Rosati’s calculations indicate that 90 percent of contiguous democratic dyads from 1950 to 1990 were in the Americas and Western Europe. 190 The counter-argument to this is that certainly democracies in the post-WWII world had common interests, and alliance ties between states are a reasonable indicator of such common interests. However, quite a few studies that support the democratic peace control for the effects of mutual alliance ties and they uniformly show that the relationship between joint democracy and the lack of conflict is statistically significant even when alliance ties are controlled for. 191 Fourth, Douglas Gibler has put forward an alternative known as the “territorial peace.” It argues that removing territorial issues between neighbors and creating stable borders leads both to peace and to an environment that fosters the evolution of democratic governments. 192 In fact, stable borders are seen as a near necessary condition for joint democracy. Looking at all dyads between 1946 and 1999, the presence or absence of joint democracy, the onset of MIDs, and a variety of indirect indicators of stable dyadic borders, Gibler makes an interesting discovery. First, stable borders increase the probability that a dyad will have few MIDs, and it also increases the prospects of joint democracy. More important, democracy has little or no effect on conflict (MID initiation) once controls are added for the presence of stable borders. The indicator of democracy was not statistically significant in any of the models that include border controls. As Vasquez notes, “this is one of the few studies that successfully wipe out the statistical significance between joint democracy and peace.” 193 Gibler concludes that the so-called democratic peace is really a “stable border peace,” and the democratic peace is a spurious result of a larger, territorial peace. The logic behind the theory derives from the effect on states of threats to their homelands. Threats to home territories require vigorous defense, specifically large, standing armies, which, not incidentally, can also be used for political repression at home. Consistent threats to a state’s home territories also create a favorable political environment for placing increased power in the hands of the executive branch and decreasing the checks on executive power. Standing armies also require higher levels of taxation and broad centralization of power, both of which contribute to widening economic gaps between rich and poor. High levels of military

272

Chapter 7

spending and frequent wars and crises dampen domestic consumption and economic growth. Adopting democracy under these conditions is difficult because the state would have to give in to the lower classes’ demands for redistribution of wealth. It costs less to use the military to suppress dissent. All of this works to the detriment of democracy creation. Conversely, states without territorial disputes are much more likely to be at peace, economic growth is more likely, power within the state is more likely to become decentralized, and the state is better able to democratize. Democratic states are at peace, not because they are democratic, but because they lack territorial disputes. (Stable borders between autocratic states should also be peaceful.) An important part of this explanation is that in regions where territorial conflicts are nonexistent, zones of peace are created in which democratic states who have settled their border conflicts with each other tend to cluster in time and space. Gibler maintains that “As neighbors experience fewer and fewer territorial issues, they are more likely to, first, have a peaceful relationship and, second, become more democratic. The democratic peace, as we know it, is simply an outgrowth of a peace between neighbors.” 194 Stabilized borders create a peaceful contagion effect. As Gleditsch’s independent study discovers, democracies rarely ever fight wars on or proximate to their home territories, and democratic states tend to cluster geographically, creating zones of peace. 195 The finding that democracies tend to fight and win shorter wars can be easily explained by the territorial peace. The costliest wars—territorial wars with neighbors—have already been selected out of the pool of potential conflicts facing democracies. What is left are “wars of choice” rather than “wars of necessity,” and democratic states can select wars that can be won at relatively low cost. Democracies will also tend to fight their wars in places far from their own borders, since local borders have already been settled and pacified. A final alternative explanation for peace between democracies offers advanced capitalist economies as the real source of dyadic peace. In the post-WWII world, most democracies are relatively wealthy, economically developed states that are satisfied with the prevailing status quo. The fact that they are rich, satisfied powers may be more important than the fact that they are democratic states. There are several versions of a “capitalist peace,” depending on what one thinks is the best indicator of capitalism—international trade, openness to foreign investment, or low levels of government involvement in the economy. Perhaps the most vigorous recent attempt to prove that peace is due to joint capitalism rather than joint democracy comes from Michael Mousseau. 196 Mousseau refers to his version of capitalism as social-market capitalism. For Mousseau, capitalism means engaging in economic contracts with others and employing free choice in this endeavor. Drawing on Karl Polanyi’s classic work The Great Transformation (1944), Mousseau makes the distinction between countries with contract-intensive (contract-rich) economies and those with contract-poor economies. In contract-rich economies most citizens engage in multiple contracts with strangers (both inside their own countries and outside of it) for wages and for various goods and services. Because of this, they have a common interest in each other’s equal rights and welfare; they also have an interest in having government protect freedom of choice and the legality of contracts and the rule of law in general. Since citizens on a daily basis depend on strangers for their needs, they develop habits of trusting out-groups. As a result of this environment, they learn certain norms or preferences—such as freedom of

The Dyadic Level of Analysis, Part I

273

choice, equal rights, and the rule of law—all fundamental aspects of liberal-democratic forms of government. Indeed, in Mousseau’s calculation, most states with contract-rich economies are democratic, as the theory predicts, but about one-half of all democracies have contractpoor economies. 197 What does this have to do with propensity for war? Mousseau argues that “within and among contract-rich ‘capitalist’ nations, wars cannot happen because they require the harming of others, and citizens in these economies are always better off when others in the market are better off.” 198 On the other hand, citizens in contract-poor economies develop few contracts with strangers; instead, they are generally dependent on reciprocal favors among friends, family, or clan members for obtaining basic goods and services. Since social-economic relations are organized in terms of patron-client relations, they do not develop interests in the welfare of strangers (or society in general). They are not likely to develop the habit of trusting outsiders, either within the state or outside it. They have a mutual interest only in the welfare of their own group. Most less-developed countries (LDCs) are contract-poor economies. In these countries, war may be in the interests of elite in-groups that make up the ruling coalition, especially if foreign conflicts (against strangers) can be used to unify various ethnic, religious, political, social out-groups within the state. What is Mousseau’s operational definition of contract-rich economies? How does he know the extent to which a particular economy is contract-rich or contract-poor? The answer is rather ingenious. He uses World Bank statistics on a country’s level of life insurance contracts. He argues that high levels of life insurance contracts indicate highly institutionalized levels of contracting in general throughout a society, since life insurance contracts require a high degree of trust that the other party (the insurer) will actually honor the contract after one has died. An early study seems to indicate that the democratic peace applies only to jointly socialmarket economies. 199 The logic of Mousseau’s argument is that if social-contract capitalism is to account for the impact of democracy on conflict, then capitalism (contract-rich economies) must correlate with both democracy and conflict. Also, since most contract-rich economies are democratic, but half of all democracies have contract-poor economies, it would be crucial to see if there is a difference between democracies that are contract-rich versus contract-poor. What does he find? Mousseau examines cases of MIDs that end in fatalities from 1961 to 2001. Dividing dyads into three groups—Both Democratic but Both Contract-Poor, Both Contract-Rich, and All Other dyads—he finds that contract-poor democratic dyads are somewhat more likely than other dyads to fight each other, though the difference is not statistically significant and the absolute number is rather small—only nine years with fatal MIDs. The big news is that there is not a single fatal MID for any of the contract-rich dyads, and this is statistically significant. He concludes that while there is no support for the idea that democracy causes peace, a contract-rich economy is a robust force for peace. The presence of a contract-rich economy accounts for both democracy and the absence of MID fatalities. Democratic Dyads: Implications for Peace There is good news in all of this—factually and theoretically. Part of the good news is obvious. To the extent that democratic states do not fight each other, a separate peace exists between and among democracies. And as more countries become democratic—increasing the

274

Chapter 7

number of jointly democratic dyads—this zone of peace is enlarged. A separate issue is the impact of the democratic peace on the international system as a whole. Several considerations are important. First, as the number of democracies increases, not only is the number of joint democratic dyads increased, but so are the number of mixed dyads (democratic-authoritarian). If it is true that the most conflict-prone dyads are mixed dyads rather than joint democratic dyads or joint authoritarian dyads, then this process might initially create more conflict rather than less—at least up to a point. Once the number of democratic states reaches fifty percent of the total states in the system, then every time a democracy replaces an autocracy, the chances of conflict should be reduced. Therefore, in the long run, the creation of new democracies should reduce the total amount of conflict in the international system. According to some estimates, a solid majority of states in the system are now democratic. 200 Sara McLaughlin Mitchell’s estimate is that democracies have increased from 19 percent of the global community in 1900 to about 57 percent in 2009. 201 There is some evidence at the system level that the increase in democratic states and democratic dyads has had a salutary effect. Maoz and Abdolali’s study of conflict from 1817 to 1976 finds that as the number of democratic dyads has increased, the chances of low-level MIDs escalating to war has decreased. Similarly, Crescenzi and Enterline’s study of the post-WWII era reports that as the proportion of democracies in the international system has risen, we have seen a decrease in the proportion of states participating in war. 202 From a normative perspective, the accumulation of democratic states suggests that democratic norms—such as support for nonviolent dispute resolution measures, support for selfdetermination, and support for territorial integrity of states—have an improved chance of becoming adopted by other states and thus determining the nature of a more general international normative consensus. Indeed, Mitchell sees evidence of this in the adoption of democratic norms of international behavior by autocratic states. Moreover, the rise of democratic states in the international system has been accompanied by an increase in the number of international organizations, international courts, and nongovernmental organizations dedicated to peaceful management of conflicts. 203 This systemwide spread of norms is no doubt also enhanced by the fact that the two states purported to have been global hegemonic leaders in the last two centuries, and thus in a position to shape the postwar order, have both been democracies—the U.S. and United Kingdom. Second, democracies tend to win the wars they fight. For instance, Lake finds that of the twenty-six wars fought between democracies and nondemocracies from 1816 to 1988, democracies have won 81 percent. 204 Moreover, wars typically lead to regime changes, with losing states much more likely to undergo violent regime change than winners. 205 (When democracies lose wars, the democratic process provides a mechanism for changing leaders without changing the nature of the regime, thereby preserving the democratic nature of the government.) 206 Once again, the result of this process is that wars are likely to result in greater democracy because democratic states tend to win wars and because autocratic losers are prone to regime change, and some of these regime changes will be in the direction of democracy. There is some evidence in fact that the systemic effect of war has been to increase the proportion of democracies in the international system. And it is also true that interstate wars

The Dyadic Level of Analysis, Part I

275

are declining as the number of democracies has increased, although the reason for this may be something other than the spread of democracy. 207 It is still unclear at this point whether the increase in democratic dyads prior to WWII increased or decreased the potential for war. A potential downside to all of this is that some democratic regimes may in fact be consciously motivated by the concept of the democratic peace and may come to see it as a moral duty of democracies to use war to foster regime change in autocracies. War is used in the short term as an instrument of peace in the long term. This was certainly a major motivation of the neoconservatives (and others) in the Bush administration for the invasion of Iraq in 2003 and of many liberals in the Obama administration for the NATO intervention in the Libyan civil war in 2011. THE ROLE OF DYADIC POWER BALANCES Our last set of potential bad or dangerous dyads has to do with the relative balance of capabilities between states. The fundamental question is whether states are more likely to fight each other if they have relatively equal or unequal power capabilities. Those who believe that equality (parity) promotes war argue that if states are fairly equal in capabilities two things happen. First, both states believe they have a chance of winning and are therefore tempted to initiate a war in times of crisis. Second, since capabilities levels are near equality, it is easy for both states to miscalculate the balance and believe they have superiority, once again tempting them toward war. Those who believe that inequality (preponderance) promotes war argue that states whose power is clearly superior cannot be deterred by lesser states from starting war; they are essentially unchecked and may do as they please. In this case, large, dominant states engage in wars of choice, picking on weaker victims because they are assured of relatively easy victories. The debate has been thrashed out with great vigor over the last several decades. It’s probably fair to say that we finally have a winner in this long-standing dispute, though perhaps it is a TKO rather than a knockout. Most, but not all, studies support the idea that parity leads to war. Three early studies all pointed to the link between dyadic parity and war. A study by Mihalka using COW data for states in the European system from 1816 to 1970 indicated that the chances of a dispute escalating to military violence was significantly lower for dyads with high capability differences. Likewise, studies by Weede on Asian dyads from 1950 to 1969 and by Garnham on thirty dyads from 1816 to 1965 both demonstrated that preponderance of military capabilities is more likely to lead to peace, while parity is more highly associated with war. 208 Another study by Moul investigated dispute escalation among European great powers from 1816 to 1939. He found that over 50 percent of the disputes among relatively equal states escalated to war, while only 3 percent escalated to war for dyads having an unequal capability distribution. 209 Geller’s analysis of twenty-nine enduring rivalries from 1816 to 1986 discovers that parity between rivals is about two times as likely to be associated with war than dyads characterized by disparity preponderance. 210 A fourth study by Huth found that states with military capabilities that diverge greatly from their opponents will be unlikely to escalate

276

Chapter 7

territorial disputes. On the other hand, states with a three-to-one advantage tend not to compromise on territorial disputes. 211 Additional support comes from advocates of power transition theory, which we will examine in chapter 11. Kim, using the COW database 1816–1939, estimated that the probability of war for equal dyads was more than double the probability of war for unequal dyads. A further study extending back to the Peace of Westphalia (1648) further confirms that equality in capabilities (coupled with dissatisfaction of challengers) has a strong impact on the likelihood of war. 212 The most frequently cited evidence comes from Stuart Bremer’s famous study of “dangerous dyads.” Bremer demonstrates that for all dyads in the international system between 1816 and 1965, war is roughly 33 percent more likely for dyads having a small or medium difference in relative capabilities than for dyads with large differences in capabilities. 213 Bennett and Stam’s massive statistical analysis of MID data from 1816 to 1992 shows that as the dyadic balance moves from equality to inequality, the risk of MIDs and wars decreases substantially for “politically relevant dyads.” They find that the risk of war declines 65 percent from dyads that are equal to dyads in which the balance of capabilities is four-to-one. Preponderance clearly reduces the chances that MIDs will be initiated or that they will escalate to war. 214 Most recently, Zeev Maoz’s study of pacifism and “fightaholism” uncovers a large realm of dyadic peacefulness, with about 86.5 percent of all politically relevant dyads never engaging in war. One of the primary correlates of this dyadic peace is a high disparity in capability between dyad members. 215 (The other correlates of dyadic peace are joint democracy, at least one minor power in the dyad, and the evolutionary emergence of the states rather than a revolutionary birth.) The peaceful impact of preponderance is also saluted by historian Geoffrey Blainey, who argues that decisive victories in major power wars have led to longer periods of peace precisely because they illustrated plainly the absolute preponderance of the victor. On the other hand, wars that lacked decisive conclusions led to future wars because they left the system in a situation of relative parity where misperceptions concerning the true balance of capabilities could lead all states to assume the balance was in their favor. In short, there is considerable robust evidence that static balances of equality or nearequality are much more dangerous than balances that tend toward preponderance. There is a general agreement among scholars now on this point, though there is a smattering of evidence on the opposite side. 216 Of course, states on the short end of the military balance do not always give in. As we will see in chapter 9, military preponderance does not guarantee that states with lesser capabilities will be deterred from attacking stronger states; it’s only an empirical tendency. Relative preponderance decreases the probability of war, it doesn’t eliminate it. Strong levels of dissatisfaction and acute rivalry (as well as a variety of factors at the individual level) may get in the way. For instance, Huth finds that for territorial issues that are highly salient, dissatisfied challengers are not deterred from attacking more powerful opponents unless the military advantage is near nine-to-one. 217 And Geller finds that for enduring rivals, the balance may not matter too much. Dissatisfied challengers are equally likely to start wars whether they are

The Dyadic Level of Analysis, Part I

277

superior or inferior in capabilities to their rivals. 218 T. V. Paul’s seminal study of asymmetric conflicts demonstrates that the decision of leaders in weaker states to initiate conflict against stronger opponents is powerfully affected by the advantage of limited fait accompli strategies, the presence of a short-term offensive strike capability, the anticipated support from a great power, and perceptions of domestic political advantage. 219 Finally, we should note that militarily dominant states in unequal dyads do sometimes start wars, though not as often as many would predict. John Vasquez helps us to understand this phenomenon; he distinguishes between two types of wars: (1) wars between relative equals, which he calls “wars of rivalry,” and (2) wars of inequality. 220 He argues that the paths to war for the two types are likely to be significantly different. His “steps to war” theory focuses on wars of rivalry. These wars typically result from conflicts between contiguous rivals having territorial disputes that escalate in a conflict spiral dynamic due to the use of “power politics” tactics—tactics of coercive diplomacy involving demonstrations of force, repeated militarized disputes, the formation of alliances, and arms buildups. Most of these causal factors take place at the dyadic level of analysis. And, of course, it is these wars that are the most typical or representative of interstate wars, and they are essentially wars in which the dyadic balance is close to equality. Wars of inequality exist, but they are less prevalent and less typical of interstate wars in general. Compared with wars of equality, wars of inequality have different dynamics and a different causal logic. Here preponderant states may engage in wars of opportunity or wars of choice: they go to war simply because they can and because they anticipate a positive payoff from the effort. The causal factors of primary importance in these wars are likely to be located at the individual level of analysis, involving cognitive factors, personality variables, perceptual errors, and cost-benefit analyses involving estimates of expected utility. Cashman and Robinson’s examination of six cases of war in the twentieth century includes one war of inequality, the 2003 Iraq War. Their analysis is that, compared with wars of equality, an explanation of the Iraq War requires a fundamentally different set of causal factors—mostly at the individual level. 221 CONCLUSION As we know, war and peace are not distributed randomly. Certain dyadic combinations increase the probability that disputes will arise and that they will escalate to war. Contiguity, territorial disputes, and rivalries are all linked to increased probabilities of war, and they probably act synergistically with each other. A dyad sharing the fivefold combination of contiguous border, a mutual territorial dispute, rivalry, power parity, and the absence of joint democracy is at much greater risk than a dyad without these characteristics. Some of these factors are controllable, at least to some extent. Settling territorial grievances amicably would eliminate a primary source of disputes and war, and it might help to end rivalries. The creation of more democracies might increase the zone of peace, and it also appears to be a well-documented path toward ending enduring rivalries. While contiguity is less amenable to engineering, it might be possible to create “buffer zones” between rival states—though the possibilities for this are geographically slim.

Chapter Eight

The Dyadic Level of Analysis, Part II International Interactions

Eye for eye, tooth for tooth. —Exod. 21:24 All those who take the sword will perish by the sword. —Matt. 26:52 If thine enemy hunger, feed him; if he thirst, give him drink. . . . / Be not overcome of evil; but overcome evil with good. —Rom. 12:20, 21 Absolute security for one power means absolute insecurity of all others. —Henry Kissinger

Imagine a pair of eight-year-old boys making a wreck in the recreation room. Their mothers race in to quell the disturbance. As the mothers enter the room, little Johnny is about to crown little Sammy over the head with his Squid Ray. “What’s going on here? Why are you fighting?” “Johnny won’t let me play with his tank.” “But Sammy won’t let me use his Astro Blaster and it’s my turn. I’m not going to share if he’s not.” “Well, Johnny started it. He refuses to die when I shoot him. He’s not playing fair.” “Yeah, well he started it. He always has to be the good guy.” Sound familiar? From time to time we have all been reminded how the actions of states are not unlike those of small children. The above scenario illustrates a common conflict spiral. One individual responds to actions by another, who in turn responds to those of the first. Each reacts to the prior behavior of the other. The interaction becomes conflictual, the conflict escalates, and fighting breaks out. Many analysts argue that countries go to war because of previous actions directed against them by others. State A is simply responding to the hostile acts of state B. State B feels the same: Aha, you say, then what about state B? Same answer. Its action was prompted by an earlier action by state A. And so it goes. The beginnings of this process of interaction between

279

280

Chapter 8

the two states may be shrouded in history. We may never know who did what to whom first. Ultimately, it may not matter. THE ROLE OF ACTION AND REACTION IN DYADS (I): THE SPIRAL MODEL As an explanation of national behavior, the conflict spiral model of international relations mirrors the stimulus-response theory of individual behavior, which has been a staple of the behaviorist school of psychology. According to these theories, most actions are a response by the actor to a set of stimuli in the environment. Likewise, the behaviors of states are in large part responses to the behaviors of other states (and nonstate actors); the actions of states are interdependent. At the dyadic level of analysis the emphasis is on the interaction of states and on the structure and processes of their interdependent actions. From a strictly dyadic point of view, it is relatively unimportant what individual political leaders think, what their personalities are like, what their perceptions are, what the decision-making process is, and what domestic policy considerations may constrain or compel actions. As J. David Singer once suggested, these variables can all be thought of as being within a black box. 1 We don’t need to know them. We don’t need to understand them. Most states react in the same way to the same stimuli, regardless of their particular national attributes or the personal characteristics of their leaders. Let us now examine a variety of potential interaction patterns. The Reciprocity/Symmetry Pattern A dyadic action-reaction process might proceed in a number of different ways, as illustrated in table 8.1. One possible pattern would be that as conflict develops between a pair of states they begin to develop a tit-for-tat relationship. Each state responds to the other with actions that are similar in type, intensity and magnitude. The response of each to the behavior of the other is proportionate to the preceding act of the other. We might call this a reciprocity/symmetry pattern. This particular pattern indicates a stable relationship in which the level of conflict/ cooperation remains constant, neither escalating nor de-escalating. An example of such an interaction pattern might be twentieth-century Canadian-American relations. As we will see, reciprocity tends to be the normal pattern in dyadic interactions. The Fight/Escalation Pattern The second possible interaction pattern is an escalation process, one that Anatol Rapoport has also labeled as a fight pattern. 2 The escalation/fight pattern describes a situation in which each adversary in a conflict responds to the actions of the other with somewhat bolder, more hostile reactions. The response is not similar to the original action but greater— not an eye for an eye but more like two eyes and a thumb for an eye. Leng and Goodsell have described it as a pattern of “self-aggravating feedback.” 3 The behavior of the states becomes increasingly threatening and aggressive, creating a conflict spiral. Unless this escalatory cycle is broken, the level of violence will eventually reach all-out war. This spiral model is certainly not the most prevalent interaction pattern in international relations, but it is clearly the most danger-

The Dyadic Level of Analysis, Part II

281

ous one and the one that we will focus on. Examples of such a pattern would be the rather rapid escalatory interactions between Israel and her Arab neighbors prior to the 1967 Six Day War and the July 1914 crisis in Europe that followed the assassination of Austrian Archduke Franz Ferdinand. Why the escalation/fight pattern should occur rather than the reciprocity/symmetry pattern is a crucial theoretical question. Several possibilities exist—most of which require skipping to another level of analysis. Leng and Goodsell suggest that escalation occurs because one side in the conflict uses a threat of military violence. In their analysis of five interstate conflicts between 1850 and 1965, they discover that conflicts are more likely to escalate when one adversary uses such a threat. 4 Threats of military violence may represent a kind of qualitative threshold. Once that threshold has been exceeded by one side, the symmetry process is replaced by the escalation process. This may very well be true, but it begs the question of why one side decides to boost the level of hostility by using military threats in the first place. One explanation for escalation, found at the substate level of analysis, is that domestic political considerations compel leaders to escalate their responses in disputes with other states in order to ensure that they cannot be attacked by domestic critics for weakness in foreign affairs. A second set of answers comes from our examination of the individual level of analysis. For instance, it is possible that individual leaders in one or both of the countries simply misperceive the level of hostility and thus respond at a higher level of hostility, even though they believe themselves to be returning it in a proportionate, reciprocal way. Leaders may also react emotionally rather than on the basis of rational calculation. The fundamental attribution error may also predispose leaders to believe that the opponent’s motives are malign. Finally, the operational codes of some leaders may require “one-upmanship” responses to the unfriendly behaviors of rivals. Note that this set of explanations draws on psychological perspectives. As we will see, even though the conflict spiral theory makes its home at the dyadic level, much of the explanatory power of the argument comes from individual-level phenomena. The Asymmetrical Escalation Pattern One variation on this is the asymmetrical escalation pattern, in which one of the states will engage in escalation while the other is content to merely reciprocate. The relationship between Germany and her rivals in the Triple Entente in the weeks prior to the outbreak of World War I is sometimes seen in this light, with Germany in the escalatory role, overreacting to the behavior of Britain, France, and Russia. Prior to the 2003 Iraq War, as U.S. hostility escalated, Iraqi behavior was actually de-escalatory, as it tried to prevent an attack by cooperation with UN inspectors—though such cooperation was clearly sluggish, reluctant, and less than totally forthcoming. The De-escalation Pattern A fourth interaction pattern is a pattern of mutual de-escalation. Each of the states in the deescalation pattern reacts to the behavior of the other by responding with actions that are more conciliatory and less hostile. Thus, the behavior of the states becomes increasingly more

282

Chapter 8

cooperative and harmonious as time goes by. (This process can be either symmetrical or asymmetrical; for simplicity’s sake, it is represented below as symmetrical.) Instead of a conflict spiral, there is a peace spiral. American relations with the Chinese in the early 1970s and with the Soviets in the immediate aftermath of the Cuban Missile Crisis (the period of the “Kennedy experiment”), as well as the detente period of the early 1970s, all illustrate this type of relationship. (Soviet-American and Soviet-Chinese relations in the Gorbachev era are probably best characterized as asymmetrical de-escalation, with the Soviets in the role of the chief de-escalator.) 5 The spiral model is based on the presumption that a state’s behavior depends on the actions directed toward it by other states. Most of the time the interaction pattern tends to be stable: The responses of state A to the actions of state B tend to be similar in type and magnitude. Reciprocity is the general rule. Sometimes, however, the action-reaction pattern escalates. Of the four possible interaction patterns above, the escalation/fight pattern is the one most likely to lead to war. When theorists talk about the spiral model, this is what they are referring to. The spiral model posits that wars do not come “out of the blue” with no hint that dangerous times are upon us. They are preceded by dyadic interactions that become increasingly hostile and threatening over time. The conflict spirals that precede wars may vary considerably— from a few weeks in the 1967 Middle East crisis preceding the Six Day War to the long, slowburning, but gradually accelerating spiral between the U.S. and Japan that began with the “Manchurian Incident” in 1931 and ended a decade later with the attack on Pearl Harbor in December 1941. Let us now examine the spiral model of war more closely.

The Dyadic Level of Analysis, Part II

283

THE ROLE OF CONCILIATORY RESPONSES IN THE SPIRAL MODEL: LIBERAL AND CONSERVATIVE VIEWS 6 Politicians and academics in the United States in the post–World War II era have engaged in a major debate that has an important bearing on this theory. Liberals have argued that tough, militaristic actions by the United States led inevitably to tough, militaristic reactions by the former Soviet Union. Alternatively, conciliatory and cooperative actions by the United States generally led to conciliatory and cooperative reactions by the Soviets. Superpower interaction was believed to be based on the norm of reciprocity. This is the classic formulation of the reciprocity model: Violence begets violence and cooperation begets cooperation. During the Cold War, the implication was that the best way to bring about Soviet cooperation was to treat them with conciliatory and cooperative actions. Conservatives argued just the opposite. They maintained the Soviets acted inversely rather than reciprocally. Cooperative actions by the United States produced only uncompromising, aggressive acts by the Soviet Union, while tough, strong actions by the United States were the most likely to produce accommodation. The implication was that “bullying” the Soviets was the approach most likely to produce the desired results; conciliation would be counter-productive. U.S. relations with communist countries during the Cold War and with so-called rogue states in the post–Cold War era (North Korea, Iran, Iraq under Saddam Hussein) are treated within the same framework. The difference comes down to liberal and conservative ideas about international relations. What accounts for this divergence of opinion between liberals and conservatives? The difference can be traced to their divergent views on the nature of international relations as a whole, but also their differing views of the nature of certain kinds of states. Liberals view most states as essentially cooperative and rational. States (that is, their leaders) are motivated more by fear of others than by any evil intentions to destroy others or to take their property, and cooperation is more likely to be induced by providing benefits and compensation than by threatening to do harm. Conservatives are generally more skeptical of the intentions of states. They saw communist states (especially the Soviet Union) as radically different from democratic states, with a goal of creating a communist world empire. Likewise, current rogue states refuse to abide by international norms, violate the human rights of their citizens, and have leaders who are autocratic and unpredictable, perhaps even irrational. They can neither be trusted nor reliably deterred. Academic supporters of this argument have relied on classical realist thought, especially as it has been extended by deterrence theorists, game theorists, and bargaining theorists. Which theory most correctly explains U.S.-Soviet or U.S.-Iranian interactions or international interactions in general? In order to know, we must observe the behavior of pairs of states on a week-to-week basis and track how one type of action leads to another, a process called the event-interaction approach. A useful approach would be to identify and code daily or weekly actions on, say, a nine-point scale, from those that are the most cooperative (at the low end) to those that are the most hostile (at the high end of the scale). The scale might look something like the one in box 8.1 (this scale was also used in table 8.1 to help depict conflict

Chapter 8

284

interaction patterns). After each event or action has been properly coded, charts and graphs can be constructed to depict the interactions of states, and a number of statistical tests can be performed to help us investigate the relationship between the actions of state A and state B. Box 8.1. Nine-Point Cooperation- Hostility Scale 1. 2. 3. 4. 5. 6. 7. 8. 9.

Most cooperative: signing of cooperative agreements, treaties, and alliances Granting of foreign economic aid or military assistance Reception of political, economic, or military delegations Exchange of cultural delegations Neutral Verbal denunciations or accusations, formal protests, propaganda attacks Economic or diplomatic sanctions, military show of force, troop movements Mobilization or alert of armed forces, issuance of military threats or ultimatums Most hostile: war

Note: Alternatively, we might choose to create a scale in which the cooperative acts are assigned positive values (say, from 1 to 4) and the noncooperative acts are given negative values (say, −1 to −4). For a comparison of some frequently used event-interaction scales, see Joshua Goldstein, “Reciprocity in Superpower Relations: An Empirical Analysis,” International Studies Quarterly 35 (June 1991): 200. Empirical Evidence of Reciprocity Several studies have used variations of this event-interaction approach to examine patterns of conflict interaction between the United States and the Soviet Union. Russell Leng examined data on the bargaining interaction between the two superpowers during three periods of crisis: the Berlin Blockade of 1948–1949, the Berlin Crisis of 1961, and the Cuban Missile Crisis of 1962. Leng reports that the weight of historical evidence overwhelmingly supports the liberal contention. The Soviets were far more likely to be accommodative in response to positive American acts than to American threats. The most effective American influence technique, however, used both carrots and sticks—combining threats and positive inducements. 7 William Gamson and Andre Modigliani studied interaction between the Western and Eastern Bloc nations during the early Cold War period between 1946 and 1963. Their findings also support the notion that cooperation begets cooperation and hostility begets hostility. They conclude that belligerent activity of one side (it doesn’t matter which) was most likely to generate belligerent activity by the other. Likewise, actions of accommodation were most likely to prompt a conciliatory response by the other. 8 Other studies reach similar conclusions concerning Soviet-American relations. Jan Triska and David Finley, two experts on Soviet foreign policy, uncover pronounced patterns of reciprocity in Soviet-American relations throughout the Cold War period. 9 Ole Holsti also finds that during the Cuban Missile Crisis between the United States and the USSR the level of hostility on each side was largely based on the other’s hostility in the immediately preced-

The Dyadic Level of Analysis, Part II

285

ing time period. 10 More recent event-interaction studies consistently confirm the reciprocal nature of U.S.-Soviet interactions. 11 And Soviet-Chinese relations fit the same pattern. 12 Goldstein and Freeman’s analysis of U.S.-Soviet-Chinese relations from three major events–interaction data sets (COPDAB, WEIS-X, and ASHLEY), which collectively cover the period between 1948 and 1989, concludes that bilateral reciprocity was the norm in all three sides of the strategic triangle. Responses of all three states to actions of the others were almost universally reciprocal in nature, though not necessarily completely symmetrical. The United States, the Soviets, and the Chinese each reciprocated each other’s actions on a short-term basis. In none of the interaction patterns was there any evidence of inverse response. None of the superpowers showed any signs of opportunistically exploiting the cooperation of others. Nor did any of the three show signs of backing down when faced with hostility from the others. In sum, Goldstein and Freeman found no evidence of states behaving the way conservative theory hypothesizes that they should. 13 Event data from the interactions among diverse states in several time periods show remarkable similarity. Robert North’s research group at Stanford studied conflict interaction during the 1914 crisis preceding World War I. A major conclusion was that the hostility level of both the Triple Entente and the Dual Alliance powers was based primarily on the hostility targeted at them by the other bloc. 14 Leng and Goodsell investigated interaction patterns in five dyadic conflicts between 1864 and 1962. They found that the conflict behavior of all the states involved was remarkably symmetrical. The negative acts of the pairs of states paralleled each other in typical stimulus-response fashion. Additionally, they discovered that as the amount of interaction increased, the behavior of the two countries became more and more similar. 15 As the quantity of actions the two states directed at each other increased, they seemed to become more aware of their mutual relations and their interactions began to become more directly responsive to each other. The actions of each state became, in a sense, more finely tuned to the actions of the other. Michael Ward used regression techniques to test the relative importance of bureaucratic inertia (which he calls short-term memory) and international interaction on a country’s foreign behavior. He looked at five cases of long-term interaction from 1948 to 1977: U.S.-USSR, U.S.-France, U.S.-Japan, U.S.-Israel, and Israel-United Arab Republic. He concluded that bureaucratic inertia plays only a minor role in determining a state’s behavior, while reaction to the behavior of others seemed to be crucial. He discovered a reciprocal, action-reaction process in each of the five relationships: conflictual behavior by one state leads to conflictual behavior in another, and cooperative behavior by one begets cooperative behavior in others. Ward discovered, however, that the action-reaction process is not always symmetrical. For instance, he discovered that the Soviets reacted to U.S. conflict behavior twice as strongly as the United States did to Soviet conflict behavior. 16 The Middle East has become a much studied area of the world, and analyses of prewar interactions in this region bolster the validity of the stimulus-response process. Regardless of the states involved, the particular time period, or the methods used, the results are pretty similar. The level of threat and hostility expressed between Israel and the Arab states is reciprocal and highly correlated; the best predictor of the level of conflict a state expresses is the level of conflict aimed at it by others. 17

286

Chapter 8

Finally, Leng has conducted a large-scale review of events data from several empirical studies that use three different sources: the Militarized Interstate Dispute (MID) data that code actions as either threats of force, displays of force, and uses of force; the International Crisis Behavior (ICB) data that provide descriptive indicators of international crises; and his own Behavioral Correlates of War (BCOW) data that are a stratified sample of forty militarized interstate crises between 1816 and 1980. 18 His task was to try to assess whether dyadic escalation in MIDs and crises follows the premises of realist-bargaining-deterrence theorists associated with political conservatives or the psychological perspective associated with political liberals. The findings are decidedly mixed, but not necessarily contradictory. The most important finding is that, as the spiral model would predict, reciprocal behavior abounds in the crisis data sets: “Nonviolent triggers are matched by nonviolent responses 69 percent of the time; violent triggers (actual use of force) are met with violent responses 68 percent of the time.” Oddly enough, these reciprocal responses recur regardless of the physical capabilities of the other side and regardless of type of influence attempt tried by the other side—outright coercion, accommodation, or mixed carrot-and-stick approaches. 19 Empirical Evidence: The Spiral Model, Escalation, and War We have seen that the behavior of states is largely reciprocal. Now we need to examine how normal, reciprocal relations might turn into escalating conflict spirals that could lead to war. A series of pioneering studies by Russell Leng have greatly improved our understanding of the crucial dynamics of dyadic interaction. In a study of twenty serious disputes in the twentieth century, Leng and Wheeler discovered that the use of bullying tactics—upping the ante by responding to the opponent with increasingly negative actions—is associated with disputes that escalate to war; it does not lead to an opponent backing down. However, in those rare cases where an opponent does respond to bullying tactics with conciliatory gestures, war can be avoided. On the other hand, a reciprocating (tit-for-tat) strategy is associated with the avoidance of war. In fact, a reciprocating strategy was found to be the only strategy able to successfully counter a bullying strategy. All attempts to counter bullying strategies with bullying led to war. 20 Leng’s analysis of a sample of fourteen serious bilateral disputes between 1850 and 1965 produced similar results. Although Leng found some support for the realistinspired proposition that threats are (modestly) more successful than promises in gaining compliance, it was also true that the use of threats was a highly risky strategy. Weaker states might occasionally be bullied into compliance, but when the states were evenly matched, the situation became much more dangerous. Negative inducements (threats) increased the probability of defiant responses, particularly between states with relatively even capabilities. Threats of military violence produced extreme responses—either abject compliance or counter-threats and punishments, which in turn were associated with war. There was no positive association between the frequency of the use of threats and the outbreak of war, but there was an association between defiant responses to threats and war. 21 Threats lead to defiant responses, which in turn lead to war. Leng and Gochman’s later investigation of thirty militarized dyadic disputes between 1816 and 1975 also produced some interesting results. They examined the effects of three aspects of behavioral interaction in disputes: militarization of the dispute, escalation of the dispute, and

The Dyadic Level of Analysis, Part II

287

the degree of reciprocity by the disputants. As predicted, conflict interactions characterized by a high degree of militarization, escalation, and reciprocity were the most likely to result in war. (Five of six such interactions ended in war.) Interactions characterized by low militarization, non-escalation, and low reciprocity—a pattern called “prudence”—were the least prone to war (zero of four). Generally speaking, when bargaining behavior is highly militarized and highly escalatory, the probability of war is high (eight of twelve), but when behavior is militarized but non-escalatory, war is less probable (three of nine). 22 Once again, the use of bullying strategies to deter or coerce an opponent appears to be a risky proposition. Leng’s more recent analysis of his BCOW data shows that crises are more apt to escalate to higher levels of hostility when they involve physical threats rather than just verbal threats. In general, the higher the level of escalation, the greater the likelihood of war. 23 While Leng finds that only 15 percent of the cases in his BCOW data fit the classic fight pattern, 75 percent of the cases that did exhibit this pattern ended in war. And over half of the cases in which one party escalated and the other responded with high levels of its own ended in war. 24 These findings have rather important implications for crisis behavior. States that respond to minor provocations by attempting to demonstrate toughness and commitment are taking a step down a perilous path. For instance, Israel’s long-standing policy that it will always respond to provocative actions by its opponents with retaliation or reprisals of its own (as a way of indicating its commitment to defending its interests) has dangerous implications for escalation. Sometimes states are aware of the dangerous paths they are on; sometimes they are not. The Case for Dysfunctional Crisis Learning Leng also discusses another interesting pattern that arises out of continued and repeated confrontations between pairs of states. He argues that statesmen are often guided by a conservative realpolitik worldview about bargaining that prescribes the use of tactics that demonstrate power and resolve. Leaders with such a view tend to believe that if they failed to achieve their objectives in a previous crisis with a rival, this was most likely due to insufficient demonstration of toughness. As a result, they adopt a more coercive strategy in the next confrontation with the rival than in the previous crisis. On the other hand, the state that was victorious in the first crisis “learns” that bullying has worked—the rival backs down, the bully gets his way, and war is averted—and continues to rely on coercive techniques. Both states quickly become “locked in” to coercive strategies and the two rivals escalate the level of coercion in each successive encounter. Thus, leaders in both states become victims of what might be called dysfunctional crisis learning. Leng’s discovery that each recurring crisis tends to be more violent is a significant finding. Indeed, Leng discovers that war becomes highly likely in the second crisis. If the states manage to avoid war in their second crisis, then war is almost inevitable in a third crisis. 25 Successive crises are linked. Not only does bullying behavior escalate activity within a crisis, it also creates an escalatory dynamic from one crisis to the next. Serial crises by the same dyad are extremely dangerous. Of course, the large majority of dyadic disputes are ended without war, either because statesmen are incredibly lucky or because they are able to effectively exert control over the escalatory process. But most disputes that become militarized end in stale-

288

Chapter 8

mates rather than in negotiated settlements; 26 so while war is averted, the problem remains, providing kindling for future brush fires. Many bargaining scholars argue that repeated dyadic disputes and crises help states understand each other’s motives, capabilities, and commitments and help them to communicate their own interests, commitments, resolve, and willingness to take risks. But as we saw in chapter 7, the evidence is overwhelming that repeated crises instead lead to violence. Enduring rivals, dyads that by definition engage in a lengthy string of MIDs and crises, are the dyads most likely to see their interactions escalate to war. Repeated interaction does not normally teach rivals how to better achieve accommodation. Paul Diehl reports that about 90 percent of states involved in rivalries go through at least two MIDs before a dispute escalates to war. 27 Brendan Valeriano likewise demonstrates that rivals tend to escalate their interactions with each other from one dispute to the next—and that those who escalate are more likely to go to war than rivals who refrain from escalatory behavior. 28 Brecher and Wilkenfeld confirm that recurring crises (which they refer to as protracted crises) are more likely to be characterized by violent trigger events and a higher level of violence than in nonprotracted crises. 29 What factors might account for the patterns of dysfunctional learning that occurs in recurrent crises and disputes? At least three sets of variables might combine to produce this effect: (1) the internalization of realpolitik strategies and operational codes by both individuals and national security organizations; (2) blowback from domestic political groups, public opinion, and the media who are quick to punish political leaders for any sign of weakness in foreign policy endeavors with rival states; and (3) residual psychological effects of the initial encounter with the rival that sharpen the emotions of hostility, anger, distrust, and perhaps a desire for revenge. The Case for Realism Several aspects of the escalation of crises do fit with realist predictions. First, national interests are related to crisis behavior. Crises are more likely to escalate to war if “basic values” are threatened, while war is less likely if the contestants lack vital security interests. In addition, the higher the stakes, the greater the chance that a state will use more coercive tactics to influence its opponent. 30 Moreover, war is most likely when the stakes involved in the dispute are high for both states, regardless of their relative capabilities and optimism regarding the outcome. Case studies by Lebow and Leng demonstrate that rational considerations of relative power capabilities may also be trumped by fear of domestic criticism for weak responses to a rival’s actions. 31 This is consistent with realist and bargaining perspectives that see “audience costs” and the need to maintain a state’s reputation as crucial constraints on decision-making. There is some evidence consistent with realist-bargaining-deterrence models that escalation of MIDs and crises can be effective bargaining tools that help both sides learn about the other side through actions that signal one’s resolve and interests, leading to a compromise settlement. However, this may involve considerable costs. Leng notes that of the 121 compromise outcomes in the MID data set 67 percent occur after at least one party actually used military force, and all but one of the remaining cases of compromise took place after at least one party demonstrated or displayed its force. 32 Jones, Bremer, and Singer make a similar discovery: High magnitudes of military action are positively associated with the settlement of

The Dyadic Level of Analysis, Part II

289

dispute, but they are also positively associated with the onset of war. In fact, they find that decisive dispute outcomes are most likely to be achieved only after disputes have escalated to war. 33 It appears that states frequently follow the realist prescription of trying to “bargain from strength”; they make compromises only after they have demonstrated resolve by threatening force, displaying force, or actually using force. But this carries with it the heavy risk that war will break out by the time they are ready to negotiate. As Leng points out, bullying is positively associated with high escalation and it is associated with war in 69 percent of the cases in the BCOW sample. 34 In sum, Leng argues that in a world where realism remains the dominant diplomatic culture, leaders involved in crises and MIDs are likely to follow realpolitik prescriptions and engage in coercive actions designed to signal resolve. While they may ultimately achieve some success by doing so, it often comes at the price of triggering a war. As Leng notes, “imposed or negotiated settlements occur more frequently in disputes that have become more militarized, but higher militarization is also associated with a greater likelihood of war.” 35 Thus, the consequences of acting according to realist prescriptions are often consistent with the predictions of the liberal approach. Coercion begets coercion—though not all the time. It would seem that any comprehensive theory of interstate war would have to take these findings into account. While there may be many paths to war, the path of reciprocal escalation of hostilities by states pursuing bullying tactics appears to be an important piece of the theoretical puzzle. IMPLICATIONS FOR THEORY AND POLICY: THE SECURITY DILEMMA In summary, we can conclude that a large array of scientific studies provide evidence to support a conflict spiral theory of international conflict. Whether it is Soviet-American relations or NATO-WTO interaction, whether it is countries in the Middle East or Asia, similar patterns of interaction have been found. States seem to respond to others in the same manner as they are treated. Cooperation begets cooperation; hostility begets hostility. As the level of hostility escalates, conflict spirals develop that may eventuate in war. Arms races may be an important factor in the escalation of this action-reaction pattern toward war, although the evidence is less clear on this point. If the spiral theory is true, it provides a clear, unambiguous challenge to some strongly held tenets of realism and realpolitik: that security lies in strength, that other states will back down from the pursuit of their interests when faced with threats, and that conciliatory acts only lead one’s rivals to believe that you will not defend your interests. Conservatives draw on the Munich analogy here, arguing that the appeasement policies of France and Britain in the 1930s merely whetted Hitler’s appetite for more territory and convinced him the West would not fight. It is argued that a policy of threats backed up by strong actions would have kept Hitler in his place. Conservatives then generalize from this experience and apply the “lesson of Munich”—never appease an opponent—to international politics in general (this is related to deterrence theory, which we examine in the next chapter). But these conservative/realist tenets

290

Chapter 8

appear to be untrue. Evidence seems to prove that threats, bullying behavior, and arms buildups only result in scaring an opponent state, not making it back down. Because demonstrations of strength are intended to decrease violence but actually increase it, Robert Jervis suggests that the central theme of international relations is not evil, but tragedy. 36 The tragedy revolves around what political scientists call the security dilemma: In a situation of international anarchy, where there is no central authority to restrain states from violence, each state is compelled to provide for its own security. However, the security policies of states are interdependent; greater security for one state may mean relatively less security for others. The problem, as Jervis puts it, is that “most means of self-protection simultaneously menace others.” 37 He explains as follows: When states seek the ability to defend themselves, they get too much and too little—too much because they gain the ability to carry out aggression; too little because others, being menaced, will increase their own arms and so reduce the first state’s security. Unless the requirements for offense and defense differ in kind or amount, a status quo power will desire a military posture that resembles that of an aggressor. For this reason others cannot infer from its military forces and preparations whether the state is aggressive. States therefore tend to assume the worst. 38

The result of attempts by each state to create greater security for itself is that no one is more secure. A military buildup by state A could make state B feel less secure in two ways: by reducing B’s physical ability to defend itself and by altering B’s beliefs about A’s motives so that B now believes that A is more motivated by gains than by simply maintaining its own security. 39 If this is the case, B will be more inclined to respond with greater vehemence. The result may be a spiral of ever-increasing levels of belligerence as states try to respond to the prior actions of others. Eventually, this attempt to gain greater security may end in war. The French philosopher Rousseau long ago explained the tragic logic of this unfortunate train of events: It is quite true that it would be much better for all men to remain always at peace. But so long as there is no security for this, everyone, having no guarantee that he can avoid war, is anxious to begin it at the moment which suits his own interest and so forestall a neighbor . . . so that many wars, even offensive wars, are rather in the nature of unjust precautions for the protection of the assailant’s own possessions than a device for seizing those of others. 40

Notice that Rousseau (and the advocates of the spiral model) are in a sense describing a kind of no-fault war. No state desires war, and none seeks territorial gains; the chief motivation is not self-aggrandizement but fear. States initiate wars because they fear the capabilities and intentions of others, and in such circumstances prudence requires that the best defense is a good offense. A sequence of hostile actions and reactions among states convinces each of the hostile intentions of its rivals; the conflict spiral increases in intensity as each responds to the provocations of the other. At some point the threshold of military combat is crossed and war breaks out. No one wanted it. No state can be blamed as having started it. No one is more responsible than any other. It just happened! Many consider the security dilemma to be a fundamental underpinning of the spiral mod41 el. It helps to explain why even defensive-minded states might be trapped in mutually hostile

The Dyadic Level of Analysis, Part II

291

encounters to begin with and why interactions tend to escalate (sometimes) toward war. However, while the security dilemma predicts conflict spirals, not all conflict spirals are due to security dilemmas. The security dilemma takes as its starting point the assumption that states are motivated by the desire to protect themselves rather than the desire to make gains at the expense of others. Conflict spirals may, of course, be the result of interactions between states that really are motivated by the prospect of gains rather than by fear of others. THE SPIRAL MODEL: A DISSENTING OPINION The spiral model is well supported by research, but there are dissenting opinions. One of the most significant is found in Dan Reiter’s influential article, “Exploding the Powder Keg Myth: Preemptive Wars Almost Never Happen.” 42 Reiter’s argument is that the spiral model predicts that runaway escalation leads to a particular kind of war—wars of preemption. Spiral theorists argue that anarchy compels states to protect themselves and to fear the intentions and capabilities of others, which leads to exaggerated fears and worst-case assumptions, worsened by cognitive limitations and misperceptions. As the spiral increases in intensity, fear and anxiety increase, and one side comes to believe that the other intends to attack. The only solution appears to be a preemptive use of force. If the spiral model is valid, we should expect to see quite a few preemptive wars over time. Reiter argues: “The spiral model predicts other outcomes, such as arms races and colonial competitions, but the preemption prediction is especially important, as it taps into a central theoretical dispute in the study of the causes of war.” 43 (He is referring to the dispute between deterrence theory and the spiral model.) As we have already seen, preemptive wars occur because the attacker believes it is about to be the target of an adversary’s immediate attack. (And also it probably believes that there is an advantage in attacking first.) Preemptive wars are defensive in nature; they are motivated by fear rather than desire for gain. So how do we know a preemptive war when we see it? Reiter’s working definition is that war is preemptive if a primary motivation is the belief that the adversary is likely to strike first within sixty days. Using this definition, he claims that between 1816 and the date of his article (1995), only three cases of preemptive war can be found, constituting a mere 4.5 percent (three out of sixty-seven) of the total wars listed in the COW data for the period: Russian-German interactions in World War I, Israel’s attack on Egypt in the 1967 Six Day War, and the Chinese intervention in the Korean War in 1950. He further argues that in each of the three cases leaders in the attacking states had other motives in addition to fear of an imminent attack, indicating that such fear is by itself not a sufficient cause of war. He argues that the two other conditions deemed by spiral theorists to compel preemption—hostile images of the opponent and belief in the advantages of striking first—are present to an extreme degree. The implication is that these causes are so insubstantial that to have an impact, they must exist at very high levels. It would seem to take a lot to provoke states into preemptive war. In sum, the risk of conflict spirals leading to preemption would appear to be exceedingly low, and one would certainly not want to identify conflict spirals as a typical path to war.

292

Chapter 8

Reiter’s limited list of preemptive wars is certainly open to question. Another list, from Stephen Van Evera’s Causes of War, adds several other cases of preemptively motivated wars. 44 While Van Evera agrees with Reiter on the Six Day War and World War I (where there were also preventive motivations), he adds several others to the preemptive list: the Franco-Prussian War and the German attacks on Norway and the Low Countries (1940). At most, then, we have in the neighborhood of six wars in almost two hundred years—still not a very large number. Accepting the paucity of preemptive wars does not, however, sink the spiral model. Reiter is probably incorrectly specifying the connection between conflict spirals and war. The logic of the spiral model leads to wars through several pathways beyond that of a preemptively motivated first strike. Conflict spirals may lead to brinkmanship wars in which leaders perceive their backs are to the wall and they see no viable alternatives to war. After intense crisis interactions, domestic political support for tougher action is ratcheted upward and leaders find it difficult to back down and make concessions—even in the absence of the fear of an impending attack. They may feel that the domestic costs of staying at peace have become intolerable or that the situation is worsening and that they have no steps left to take except for war. The presence of conflict spirals may also provide the catalyst that impels states with preventive motives to act on those motives. Conflict spirals preceded preventive wars such as the Japanese attack on the U.S. in 1941, the Franco-Prussian War of 1871, and the Iran-Iraq War in 1980. They preceded brinkmanship wars such as the Sino-Indian War of 1962, the India-Pakistan War of 1965, and the Russo-Japanese War of 1904. In these instances, while fear of an immediate attack was not at issue, nevertheless, a more general fear of the other state was. And these fears were enhanced by the presence of on-going conflict spirals. Perhaps the chief question for these non-preemptive wars is whether the attacker felt its use of force was a defensive response to previous acts of the rival. As they say in the NFL, that’s a tough call. Conflict spirals are probably endemic. The list of wars that started without some kind of conflict spiral preceding them would undoubtedly be an exceedingly short list. Cashman and Robinson’s case studies of twentieth-century wars indicates that all six of the wars between relative equals were preceded by conflict spirals—some of short duration, some fairly lengthy. 45 Most wars do not come out of the blue but are preceded by rather intense conflict interaction; they are, in some way or other, wars of escalation. One piece of macro evidence that supports this claim comes from Zeev Maoz’s finding that the vast majority (89 percent) of wars have been the culminations of lower-level MIDs that preceded them. 46 At this stage, we should make clear that conflict spirals are not always generated by classic security dilemma situations in which the desire for security by one state led to actions and policies that decreased the perceived security of another. Likewise, spirals do not necessarily take place in situations in which the primary motivation of each state is fear and insecurity. Leaders have multiple motivations, and fear of others is often coupled with desire for gains. What conflict spirals do is to teach each side important things about the other. Frequently, what is learned is actually “mislearned”—the product of misperceptions, dysfunctional worldviews and operational codes, worst-case analyses, improper use of historical analogies, stereotyped thinking, and so on. Of course, it is also logically possible that in a conflict spiral, one

The Dyadic Level of Analysis, Part II

293

side (or both) is actually motivated by gains rather than by fear. In this case the spiral is simply the product of an exercise in which the two attempt to use coercive bargaining to get what they want without war, and when this fails, they go to war. Let us now explore another phenomenon that fits the action-reaction pattern—arms races. THE ROLE OF ACTIONS AND REACTIONS (II): ARMS RACES Arms races have been identified by many as a particular class of international behavior that exhibits an action-reaction pattern. The concept of an arms race is one of those slippery terms like power; everyone seems to have his or her own definition of the term. One of the most quoted definitions is that of Samuel Huntington, who has defined an arms race as “a progressive, competitive peacetime increase in armaments by two states or coalitions of states resulting from conflicting purposes and mutual fears.” 47 Even though the exact wording of definitions varies tremendously when other authors are consulted, several common elements in definitions can be identified: 1. Arms races develop out of a conscious awareness by each state of the dependence of its own arms policies on those of another state. They result from external, competitive impulses. The arms buildups of two countries are simultaneous and interdependent. 2. Arms races consist of rapid increases in arms accumulation that constitute an “abnormal” rate of growth in military procurement. 3. Arms races are best thought of as occurring in times of peace; arms accumulations during war do not count. 4. The weapons involved in the arms race are designed to combat each other in some way. They are either similar (tanks versus tanks) or they are complementary (antitank weapons versus tanks). The Case for Arms Races as an Action-Reaction Pattern Several attempts have been undertaken to determine to what extent arms races are in fact propelled by an interactive, stimulus-response process. The most famous formulation (and one of the earliest) is that of Lewis Fry Richardson, who surmised that the rate of change in any state’s increase in arms was a function of several factors: (1) the opponent’s level of military strength, (2) a state’s own willingness to accumulate arms because of its opponent’s strength and his fear of that strength, (3) the negative factors of fatigue and cost involved in increasing one’s arms levels, and (4) the general level of grievance against the opponent—a kind of hostility or revenge factor. Richardson then constructed a mathematical formula (model) based on these factors. 48 Although the military policies of other nations are not the only determinant of a state’s decision to accumulate arms, Richardson’s formula nevertheless suggests that the rate at which a state’s arms procurement increases is to a large extent dependent on the actions of its rivals. Richardson’s own analysis of nineteenth- and twentieth-century defense spending data concluded that military spending tended to be a reciprocal process. Oddly enough, efforts to reproduce Richardson’s findings have met with only meager success. Indeed, there seems to

294

Chapter 8

be a large amount of evidence that states involved in mutual arms buildups are not reacting primarily to the behavior of others. For instance, a variety of studies suggest that, contrary to popular opinion, neither Soviet nor American arms accumulations in the post–World War II era could be characterized as an action-reaction phenomenon. 49 Indeed, some question whether the term “arms race” can even be applied to the Soviet and American arms accumulation processes. While both superpowers certainly sustained long-term military buildups in the postwar years, neither buildup represented really “abnormal” rates of growth in military spending. 50 Additional research indicates that not only were Soviet and American arms buildups not interactive, but neither was the Chinese arms buildup during the Cold War. 51 Several statistical studies of the military buildups of the two superpowers and their allies during the Cold War era suggest that bureaucratic momentum (as indicated by past levels of military spending) and other domestic processes have a greater impact on a country’s level of weapons spending than does the military expenditure of its rivals. For instance, Rattinger finds this to be true of arms procurements of NATO and Warsaw Pact countries in the 1950s and 1960s. This conclusion is supported by several studies of American and Soviet arms policies. 52 Choucri and North investigated the arms races between various European states as part of their comprehensive analysis of the long-terms causes of World War I. 53 They hypothesized that several factors might plausibly be related to a state’s military expenditure decisions. First, the general socio-economic dynamics of population and technological growth might play a role. Arms expenditures might be a function of the overall growth rate of a state. Second, a country’s own military expenditures in the preceding time period might play a role. This would provide evidence of another domestic factor at work—bureaucratic momentum. Third, a state’s colonial expansion and its intersections with other states’ colonial interests might have effects on military expenditures. As a state expands its activities outward and becomes involved in international trade and in the development of a colonial empire, it would need a larger military presence to protect its new colonial territory and its sea lanes. Finally, military expenditures could be a function of the military expenditures of one’s rivals. Thus, Choucri and North include in their analysis both internal and external factors. Their test produced interesting, but mixed, results. The authors reviewed specifically the famous Anglo-German naval race of 1871–1914, in part, to see whether Richardson’s interactive process was in effect. The judgment of historians seemed to affirm that it was. German colonial expansion had created a desire for a larger merchant marine and navy, and German leaders considered a strong navy to be essential for the defense of its colonies and its international commerce. 54 The German navy was to be constructed in specific relation to that of the British. Admiral Tirpitz’s “Risk Policy” consisted of trying to build a German navy that—although numerically smaller than the British fleet— would be sizable enough to present a risk to the British navy. The British would be deterred from an attack on Germany by the knowledge that an Anglo-German naval confrontation would pose the risk that a severely weakened (though victorious) British fleet might fall prey to a hostile third power. On the other hand, British policy had consistently been one of building and maintaining naval superiority. Her arms accumulation strategy was based on the rule that the British navy ought to be superior to any two powers combined and have a 10

The Dyadic Level of Analysis, Part II

295

percent margin for safety. The policies of both states clearly mandated a close watch on the building programs of rivals. There is much historical evidence that British leaders perceived the German naval buildup as a threat, and we know that both German and British leaders watched each other’s buildup closely. Indeed, in the years prior to the outbreak of the war, Winston Churchill, as First Lord of the Admiralty, repeatedly quoted the most recent German Naval Law to Parliament to justify British naval procurements. Britain’s yearly naval construction was thus in part dictated by German building. 55 France and Russia were also extremely interested in Germany’s military buildup, and the reverse appeared to be equally true. The question was to what extent trends in military growth could be accounted for by domestic factors and to what extent by external, interaction factors. Choucri and North discovered that Germany’s buildup from 1871 to 1914 could be explained much more effectively by internal processes such as population and technological growth and by bureaucratic momentum in arms procurement than by the action-reaction process. This was also true for Britain in the period between 1871 and 1890, but not for the 1890–1914 period. By 1890 Britain seemed to become much more responsive to German moves, though military expansion continued to be influenced by bureaucratic factors. A shift in the dynamics of British arms procurements had apparently taken place as the arms race heated up and as general hostility increased. Additionally, French, Russian, and (especially) Italian military expenditures seemed to be accounted for primarily by domestic factors. The authors report that each state’s military spending at time T-1 (the bureaucratic momentum factor) was a major factor in determining the level of arms expenditure. We should probably not be too surprised at these findings that arms races are more strongly influenced by domestic factors than by interaction with rivals. Arms expenditures would seem to be perfect candidates for government decisions arrived at through bureaucratic and incremental processes. They are long-term, noncrisis, budgetary decisions that ordinarily involve a large number of interested domestic actors—especially in Western democratic countries: legislators, political officials in the executive branch, civilian defense officials, military officers in various rival services, manufacturers of weapons and their subcontractors, citizens groups, political parties and so on. Political leaders may be aware of military procurement increases in a rival country, but this awareness must be filtered through a vast web of domestic constituencies before the final decision is made. Lloyd Jensen has suggested several reasons why action-reaction models of arms races do not adequately explain real-world decisions to accumulate arms. First, arms expenditures are subject to severe domestic pressures emanating from the military-industrial complex. Second, there is probably a tendency for arms races to reflect past spending levels because of the general tendency of military programs to expand in order to absorb whatever monetary funds are available. For instance, the end of the war in Vietnam did not produce a decrease in U.S. military spending, but instead facilitated the continuation of huge military budgets in peacetime. Similarly, the collapse of the Soviet alliance and the disintegration of the Soviet state itself in 1991 did not create a “peace dividend” that led to long-term sizable cuts in the U.S. military. New threats were found to have emerged, some real (Al Qaeda), some not (Iraq). Third, defense spending may be less than responsive to military programs of rivals because of

296

Chapter 8

misperception and miscalculation of the rival’s capabilities and intentions. Fourth, in the real world a state’s arms decisions are made not on the basis of simple bilateral relations, but on the basis of third-party military threats as well. Fifth, it may be that lowered arms levels by rivals actually increase the incentive for a state to add to its military capabilities instead of the reverse. 56 It might be added that a final reason why researchers have failed to validate an actionreaction process for arms races may be methodological. By concentrating on total yearly defense budgets, researchers may not be adequately capturing the arms race phenomenon that is taking place in the development and deployment of particular weapons systems. 57 If, for instance, the American arms race with the USSR focused on the buildup of nuclear warheads and simultaneously reduced its troop levels, the overall effect on the defense budget would likely be a net decrease—even though an intense race was taking place in particular sectors of the total defense program. This actually happened in the 1950s as the United States rearranged its defense priorities in order to get “more bang for the buck.” It is quite possible, then, that the failure to discover evidence of an interactive U.S.-Soviet arms race is due to the use of yearly defense budget figures as indicators of a buildup rather than the actual accumulation of arms themselves. Indeed, Michael Don Ward demonstrates that when data on actual military stockpiles are included in the arms race equation along with figures for defense expenditures, the U.S.-Soviet arms race appears much more reactive. 58 At any rate, due to difficulties in determining whether arms races are a truly interactive phenomenon, many researchers have tended to look instead at mutual, simultaneous military buildups that represent an abnormal rate of expenditure for both—regardless of whether the buildups can be deemed to be the product of a conscious reaction to the arms accumulation of the other. The Case for Arms Races as a Cause of War So far we have been discussing the arms accumulation process, but we have not mentioned war. One of the reasons that there has been so much study of arms races is the belief that arms races somehow play a role in the outbreak of war. If arms races are seen as one manifestation of a generally hostile and reciprocal interaction process between states (or coalitions), then it is reasonable to assume that arms races may accelerate and intensify the conflict spiral and eventually lead to war. The hypothesis that arms races are positively related to war can be derived deductively from the general stimulus-response theory of conflict. If it is true that states act toward others as they are treated, and if it is true that an arms buildup by one side is seen as a sign of hostility by rival states, then we can expect states to reciprocate this hostility with hostile actions of their own—not just with the accumulation of more weapons, but with other kinds of hostile actions as well. Thus, an arms buildup by one side should lead to increased hostility by the other. As we have seen, it is possible for this general conflict interaction to increase in intensity and get out of hand. The conflict spiral might end in war. This, of course, leads to the conclusion that a rapid accumulation of arms might lead not to greater security for either country but to increased hostility and threat. This would seem to flatly contradict the ancient maxim attributed to the Roman military writer, Flavius Vegetius Renatus, “Si vis pacem, para bellum”: if you desire peace, prepare for war. We will have more to say about this later.

The Dyadic Level of Analysis, Part II

297

The theoretical link between arms races and wars is buttressed by the logic of the security dilemma. States in a situation of anarchy must look to their own security, and fearing the intentions and capabilities of others, they will seek to increase their own military capabilities as a form of protection. However, states are likely to interpret the arms buildups of others not as defensive but as potentially offensive. There are several possible reasons for this. One is that it may be difficult to determine whether the particular arms in question are offensive and defensive; they may be dual-purpose weapons. Hence, it is prudent to assume that they may be used for offensive purposes. (This is the logic of worst-case analysis.) Second, the subjective interpretation by state A of state B’s buildup is influenced by the presence of realist norms, which have provided statesmen with foreign policy guidance for the last several centuries. The advice given by these norms suggests that when confronted with potential threats, states should increase their power, thus setting off a conflict spiral. A classic article by Samuel Huntington attempted to discover just how often arms races have led to war and under what circumstances. 59 It seemed clear to Huntington that while some arms races ended in war, not all did. Arms races could also end in an informal mutual agreement to call off the competition or in a victory for one state while the other accepts military inferiority. The question therefore becomes: under what circumstances are these alternatives likely to occur? Huntington suggested that a crucial point may come in the early phases of an arms race. After the challenger makes its initial move to alter the military status quo, the challenged state may choose one of several possible responses. First, it might seek a diplomatic counterbalance through an alliance with a third state or through an arms treaty with the challenger. Second, it might increase its own level of arms—setting the stage perhaps for a series of interdependent arms increases by both sides. Third, it might attempt a preemptive military action while still relatively strong, attacking the challenger while the balance of power is still in its favor. Fourth, the challenged state may take no immediate action. The challenger thereby achieves its goal. If the challenged state then makes a belated, last-ditch effort to catch up and redress the balance of power, war may result from the challenger’s reaction. Thus, Huntington argues that there are two danger points that occur at the beginning of an arms race. The first point is the response of the challenged state to the initial increase in arms of the challenger. The second point is the reaction of the challenger (which has been initially successful in its goal) to the frantic and belated efforts of the challenged state to retrieve its former position. Huntington offers as an example of the first danger point the Israeli decision to attack Egypt in 1956 after that country had been supplied with Soviet bloc weapons. Huntington argues that the French (and British) reaction to German rearmament in the 1930s constitutes a good example of the second pattern. Although the French defense budget remained fairly constant between 1933 and 1936, and then increased somewhat in the next two years, the real reaction didn’t begin until 1939 when France proposed to spend more on arms than in the last five years combined! 60 By the end of the year France and Germany were at war. Huntington concludes that the likelihood of war increases just prior to a change in the military superiority. At this point there would seem to be both great uncertainty and tremendous instability.

298

Chapter 8

The good news is that the probability of war varies inversely with the length of the arms race. Huntington contends that as an arms race continues, the interaction pattern between the states tends to become more predictable and more stable. A situation of “dynamic equilibrium” is produced. Each state continues to increase its arms, but the relative balance remains the same. The two states will probably by this time have come to a tacit, mutual understanding, in which both accept the relative balance. Huntington concludes that a sustained arms race is much more likely to have a peaceful ending than a bloody one. 61 As one critic points out, much of Huntington’s argument about the relative dangers of short- and long-term arms races depends on his initial selection of arms races and the determination of their length. 62 For instance, Huntington decides that the Anglo-German naval race ended in 1912 (with German acceptance of inferiority). If we instead place the end of the arms race two years forward in time, to 1914, it becomes a prime example of a lengthy (sixteenyear) arms race that ends in war instead of peace. Another factor that Huntington identifies as important is whether the arms race is qualitative or quantitative. The latter appears to him to be much more dangerous. A quantitative arms race refers to the addition of larger numbers of men and weapons to one’s military machine. In such an arms race superiority is likely—owing to unequal resources and determination. Thus, quantitative races tend to be settled one way or another. Huntington argues that quantitative races impose greater and greater burdens on those countries involved in them. Meanwhile, governments must mobilize popular support for the sacrifices required by an arms buildup, generating in the process suspicion, fear, and hostility toward the potential opponent. Eventually a time is reached when the increasing costs and tensions of a continued arms race seem worse than the costs and the risks of war. Public opinion once aroused can’t be quieted. . . . Prolonged sufficiently, a quantitative race must necessarily reach a point where opinion in one country or the other will demand that it be ended, if not by negotiation, then by war. 63

This does not exactly square with his earlier proposition that the probability of war varies inversely with the length of the arms race, but it may explain why some long arms races end in war. On the other hand, qualitative arms races—races involving decisions to introduce new weapons systems through technological innovation—involve less risk of war, according to Huntington. Unlike quantitative races, they do not necessarily increase arms budgets. They represent a competition of elite technicians in weapons laboratories, not a competition of masses, and therefore impose no great burden on the general public. Additionally, qualitative races tend toward equality, owing to the tendency for simultaneous invention of parallel military technology. He argues that historically, technological breakthroughs in military weaponry have not resulted in a lasting edge for the innovator. 64 Qualitative arms races are therefore more stable and thus more peaceful than quantitative races. One might wish to amend Huntington’s general argument about the relative stability of qualitative arms races by adding one crucial qualification. It would appear that a danger point is reached when a qualitative jump in weaponry is about to be made by a challenger. Frequently, the dominant state will not wish the challenger to cross a certain qualitative threshold. This threshold may be the acquisition of nuclear weapons, chemical or biological weapons, or some

The Dyadic Level of Analysis, Part II

299

other category of weapons. But at the point at which the challenger is perceived to be about to cross that particular threshold, a resort to arms is sometimes contemplated. Early in the nuclear age, both the Truman and Eisenhower administrations contemplated preventive war against the Soviet Union to forestall its acquisition of nuclear weapons and preserve the American monopoly. China’s acquisition of nuclear weapons was of concern to both the U.S. and the Soviet Union. The Kennedy administration apparently consulted the Soviet government in 1963 about cooperation in a preventive military attack against China; the request was rejected by Khrushchev. Planning for a potential preventive attack against China continued in the Johnson administration, though the use of force was ultimately rejected shortly before China tested its first nuclear weapon in October 1964. 65 In September 1969, after the Chinese had completed several nuclear tests and were preparing to launch their first earth satellite (demonstrating their ability to deliver nuclear weapons with ballistic missiles), the Soviet Union apparently also considered the possibility of a preemptive surgical strike on Chinese nuclear installations. 66 When the Osirak nuclear research reactor neared completion in 1981, providing Iraq with the possibility of making weapons-grade plutonium, the Israeli government undertook an air strike to destroy this facility; and Israel took similar action in September 2007 against a potential nuclear reactor site in Syria. Iraq’s continued search for nuclear weapons technologies was a factor in the decision by the Bush administration to send U.S. forces to liberate Kuwait from Iraqi control in 1991; more important, it was a primary reason for the U.S. invasion in 2003. Proliferation of nuclear weapons to North Korea and Iran have similarly raised the issue of a military option to prevent these two states from crossing the threshold. In early 1994—as North Korea had threatened to withdraw from the Nuclear Non-Proliferation Treaty (NPT), refused inspection of its facilities by the International Atomic Energy Agency (IAEA), and began to remove fuel rods from its nuclear reactor in order to process plutonium so as to construct a weapon—all hell broke loose. Public warnings were issued by the U.S., and the secretary of defense ordered a series of preparations for possible military conflict. Serious consideration was given to an air strike against North Korean nuclear facilities, and political columnists were vocal in their support for the necessity of a second Korean War. War was averted at the last minute by ex-president Jimmy Carter’s diplomatic visit to Kim Il Sung. Similarly, both the Bush and Obama administrations have refused to take the military option off the table regarding Iran’s attempt to gain nuclear weapons status. The Israeli threat of preventive attack against Iran is perhaps even more likely than an American threat. Thus, impressionistic evidence suggests that the breakthrough phase of a qualitative arms race is just as dangerous as the early phase of a quantitative arms race. And one could argue that such a qualitative breakthrough would substantially change the previous balance of capabilities to the benefit of the challenger. The Case for Arms Races: Empirical Evidence Numerous statistical studies have addressed the relationship between arms races and war. The initial studies produced results that were not only mixed but highly contentious. In a muchreported study, Michael Wallace used COW project data to investigate great-power disputes

300

Chapter 8

between 1816 and 1965. His research question was this: “Do serious disputes between nations engaged in an arms race have a significantly greater probability of resulting in all-out war than those between nations exhibiting normal patterns of military competition?” 67 Having identified almost a hundred instances of serious disputes between pairs of major powers, he found only twenty-six disputes that resulted in full-scale war. Of these twenty-six wars, twenty-three were preceded by arms races. Of the seventy-three disputes that did not lead to war, only five were preceded by arms races. Wallace concluded that the presence or absence of an arms race between rival states correctly predicted war (or no war) in over 90 percent of the disputes. 68 Arms races seemed to make a substantial difference in whether a dispute escalated into war or not. Wallace did not address head-on the question of whether arms races cause wars; instead, he was investigating a somewhat narrower proposition, which has been called the tinderbox hypothesis. The tinderbox hypothesis suggests that while arms races don’t necessarily lead to wars directly, they may play an important intervening role in the escalation of disputes to war. That is, arms races do not directly initiate conflagrations, but they do create a flammable condition between racing states in which a small spark that settles down in the context of preexisting tension and hostility might ignite a firestorm. Wallace’s study itself set off a small academic firestorm. Among other things, his methods were attacked as loading the dice in favor of his hypothesis. Several researchers took issue with Wallace’s arms race index, the device he uses to determine whether the level of arms expenditure in a pair of states was great enough to conclude that an arms race was in fact present. The index was obtained by multiplying the arms expenditure scores of two states together. Thus, it is possible for one country to have a very high arms score and the second to have a moderately low score and the product be high enough to determine that the pair was in an arms race. This means that some arms races identified by Wallace were not really mutual races but situations where only one state was rapidly increasing its defense expenditures unilaterally. 69 Altfeld reanalyzed Wallace’s data using a stricter standard for the presence of arms races and found that all arms races thus categorized led to war! On the other hand, many wars broke out without being preceded by arms races. 70 A second problem with Wallace’s original study was that he chose to treat all arms races and all wars as dyadic (bilateral). This meant that instead of having World War I represented by a single case it was represented by nine disputes, and instead of World War II being represented by two cases it was represented by seven disputes. Overall, twenty-six distinct wars were created where only seven or eight integrated wars occurred. 71 The effect of this was to vastly overemphasize the statistical importance of the arms races that preceded the two world wars in the twentieth century. To correct this problem, Erich Weede reconstructed Wallace’s table, but grouped together all dyads that resulted in a single war (while keeping the European and Pacific theaters of World War II separate). This weakened the association between arms races and wars somewhat, but the percent of arms races that escalated to war was still 55 percent, while the percent of disputes not involving arms races that escalated to war was only 3 percent. 72 Further refinements led to a further weakening of Wallace’s original results. Paul Diehl retested Wallace’s work using stricter assumptions and conditions. He devised a new arms

The Dyadic Level of Analysis, Part II

301

race index based on the average rate of change in defense expenditures; he required arms races to consist of mutual buildups in which both sides increase their expenditures at a rate of 8 percent or more for three years; and he treated disputes as multilateral rather than bilateral where this corresponded to reality. His reanalysis of Wallace’s data finds no meaningful covariation existed between arms races and war. The presence or absence of an arms race had no effect on the presence or absence of war. 73 Table 8.2 illustrates the findings of these researchers. More recently, Susan Sample, who has become the premier investigator of the link between arms races and war, seems to have solved many of the nagging methodological problems to the satisfaction of her academic peers. 74 Her bivariate analyses of the relationship between arms races and war controlled for both different samples of dispute dyads and for different measures of mutual military buildups. Her results provide solid evidence for a positive relationship between mutual military buildups and escalation of disputes to war within a five year window. While not as strong as Wallace’s results, Sample’s findings were nevertheless both positive and statistically significant. Virtually all subsequent studies have come to similar results. The presence of mutual military buildups within a dyadic relationship produces a higher empirical probability that disputes between such dyads will escalate to war. Once again, a note of caution is necessary. Most analysts agree that not even a majority of disputes involving arms race lead to war. Using data from six different collections, James Morrow produced a list of thirty-five great-power disputes in the nineteenth and twentieth centuries that were preceded by mutual military buildups. Of the seventeen disputes that were found in all six data sets, only four ended in war (and only five of the total thirty-five). 75 One of the problems of the first wave of research on the link between arms races and wars was that for all practical purposes they include militarized disputes as an intervening variable between arms races and war. They asked whether arms races had an effect on the outcome of militarized disputes. This muddies the question of cause and effect by leaving open the possibility that arms races led to militarized disputes in the first place. Sample addressed this question by looking at major power dyads between 1816 and 1993. Using two different measures of arms races, she finds a positive relationship between mutual military buildups and

302

Chapter 8

the chance that states will be involved in a MID. Thus, an arms race increases both the chances of a MID and the chance that once a MID has occurred that it will escalate to war. 76 Later research by Gibler, Rider, and Hutchison similarly finds that if an arms race is present strategic rivals have a significantly greater chance of being involved in both a militarized dispute and a war. The strong implication is that arms races are linked to wars because an arms race more than doubles the number of MIDs a dyad is likely to experience. Thus, the chances of war are increased because the number of MIDs jumps dramatically. 77 Senese and Vasquez’s work on the steps to war reinforces these arguments. They find that arms races have a separate and independent effect on the probability of war (at least in the prenuclear era) and that “the impact of arms races is so high that they make war almost inevitable once a militarized dispute emerges in their presence.” 78 Arms races exacerbate the general hostility between rivals and this is often expressed through greater propensity to engage in threats of force, demonstrations of force, or uses of force. Sample is also concerned that the statistical relationship between arms races and war may simply be due to the fact that there are other variables that are significantly related to both arms races and the escalation of disputes to war. To test this, she performs a multivariate analysis that includes several variables besides the presence of arms races: the presence of power equality, power transitions, rapid approaches (one state closing the capabilities gap with another without actually achieving parity), defense burdens, the possession of nuclear weapons, and the type of issue involved in the MID. The multivariate test indicated that the relationship between arms races (mutual military buildups) and the escalation of a dispute to war is both positive and significant. Even after considering the effects of other variables, disputes that take place in the presence of an arms race are more than twice as likely to escalate to war than disputes without arms races. (The only other variable that is significant is the possession of nuclear weapons.) When all variables are set at 0, the probability that a major power dispute will escalate to war was a mere 8 percent; but when both states in a dyad were arming, the probability was 21 percent. 79 Sample’s next step was to do something no other researcher had yet done—look at the effect of arms buildups on states other than great powers. Her investigation demonstrates that disputes between two minor powers were significantly more likely to escalate in the context of a mutual military buildup, but this relationship did not seem to hold for mixed dyads (major power–minor power). 80 That arms races have been a preliminary to war is clear. The major wars of our century— World Wars I and II—have each been preceded by arms races. But just as clearly, many wars have not been preceded by such mutual arms buildups, and many arms races never end in war. 81 Just how many wars are preceded by arms races depends on the operational criteria one uses to determine their existence. We must conclude that, while the final evidence is not yet in, it is likely that arms races play only a modest, subsidiary role in the general causation of war. As one analyst put it: “Disputes preceded by arms races do escalate to war more often than other disputes, but they do not overwhelmingly escalate to war.” 82 They are neither a necessary nor a sufficient condition for the outbreak of hostilities. Nevertheless, it is undeniably true that arms races do make it somewhat more likely that a serious dispute will escalate to war.

The Dyadic Level of Analysis, Part II

303

Since some arms races lead to war and some do not, it is appropriate that we ask (as Huntington did) what kind of arms races are most likely to lead to war and under what conditions. As Morrow suggests, some tinderboxes are more flammable than others. 83 There has been some research on these topics, but the answers are far from clear. Several rival theses have been developed. Arms races lead to war (1) if the arms race is unable to attain an equilibrium, (2) if the arms race leads to a change in the dyadic balance of power, (3) if a “revolutionary” power is winning the arms race rather than a “status quo” power, (4) if arms races take place in the presence of war-inducing factors such as territorial disputes, alliances, and rivalries, and (5) if states lack nuclear weapons. One contender for the title of the most war-prone class of arms race comes from the father of arms race theory, Lewis Fry Richardson. In Richardson’s formulation, the races most likely to end violently are “unstable” races—those that fail to reach a point of stability where no further changes take place; instead, they exhibit greater and greater activity and tend toward ever-increasing arms expenditures without meaningful restraint. 84 Theresa Clair Smith uses data from major- and minor-power arms races to investigate the presumed relationship between unstable arms races and war. Her results suggest that unstable races are indeed more prone to war than stable races. Of particular interest is the finding that arms races that begin intensely tend to quickly reach political and economic limits, and thus level off, reaching a stable (and peaceful) equilibrium. On the other hand, races that begin slowly have a tendency to intensify over time and escalate to war. 85 These findings are almost diametrically opposed to Huntington’s original contention that the longer an arms race, the greater the chance that it will end in peace. Another possibility is that the most dangerous arms races are those that alter the dyadic balance of power. Using expected utility theory to analyze a state’s motivation for war, James Morrow argues that arms races that simply lead to the continuous matching of a rival’s capabilities do not alter the relative balance of capabilities and thus do not change either’s calculations of the potential for success in war. Most arms races, however, do not produce a continuous equilibrium. Different rates of arming result in temporary swings in the relative balance. When a state holds a temporary military advantage, its calculations about probable success in war are favorably altered, and it is at this point that disputes are most likely to escalate to overt conflict. The larger the swing in military superiority, the greater the window of opportunity, and the greater the likelihood of war. Fighting a war becomes more attractive, at least in the short run, since this window of opportunity will most likely close later as the rival catches up. Using data from seventeen cases from the nineteenth and twentieth centuries in which major power disputes were preceded by military buildups, Morrow confirms that escalation of disputes to war becomes more likely as the amplitude of the arms race increases. 86 Research by other scholars seems to point in the same direction. Paul Diehl looks at arms races in twenty-two great-power enduring rivalries between 1816 and 1976. He finds that war occurs rather late in these rivalries, usually after two prior militarized disputes. He concludes that military buildups have little direct effect on the escalation of rivalries to war, but that they were dangerous under two conditions: (1) when the buildup is unilateral or asymmetrical and

304

Chapter 8

(2) when the buildup leads to a power shift toward parity (in the nineteenth century) or toward military preponderance (in the twentieth century). 87 Thus, the available (though meager) evidence suggests that long-term arms races that have unstable or asymmetrical characteristics tend to be most likely to end in war. This makes a great deal of intuitive sense. In these races temporary swings in arms accumulation take place that grant to one side or the other an advantage that had not previously existed. The arms race “leader” now has a temporary advantage that permits it to contemplate military actions. For the “loser,” the military gap that has recently opened may engender a new and profound fear of the opponent. Under these conditions, either state might consider taking more extreme actions. Some writers in the realist tradition agree that arms races are dangerous when they lead to a change in the relative balance of military power between rivals, but that this is not the complete story. What is most important is which states are being relatively advantaged or disadvantaged by this shift. The presumption is that the most dangerous situation occurs if the shift in power is in favor of a revisionist state that is dissatisfied with the current international system—an idea similar to the power transition theory, which we will discuss in chapter 11. The argument here is that any apparent causal relationship between arms races and war is really due only to the fact that military buildups have changed the balance of capabilities— diminishing the relative power of status quo states and enhancing the relative power of revisionist states. (Buildup by status quo states should work effectively to signal resolve to the opponent, thus deterring war.) Though the argument appears logical, the empirical evidence is not favorable. Wallace finds, in fact, that neither gains in the relative balance by revisionist states nor revisionist state superiority in the balance are related to the outbreak of war. 88 By the first decade of the twenty-first century, consensus had largely been established that the predictions of deterrence theory have not been borne out by empirical data from the last two centuries. Arms buildups are in the main not particularly useful in deterring war; rather, they are clearly related to the escalation of disputes to war, even when we take into account their impact on the distribution of capabilities. Sample argues that the important question at this point is (1) whether arms buildups really have an independent impact on increasing the likelihood of war (even though they are probably imbedded in a more complex dyadic process like a steps-to-war model) or (2) whether they are just a symptom of a larger phenomenon, like rivalry, without any causal force of their own. 89 For instance, the argument of Paul Diehl, one of the founders of the rivalry concept, is that like war itself, arms races are caused by rivalry. 90 How does one address this question? Sample uses a multivariate analysis that controls for rivalry, but she also looks specifically at pairs of rivals to see whether those with arms races are more prone to war, and she also looks at nonrivals, to see if arms races have an impact on the probability they will go to war. Using her data on mutual military buildups, COW MID data from 1816 to 2001 and three separate measures of rivalry, Sample attempts to get to the bottom of the issue. She finds a strong and significant relationship between mutual arms buildups and the escalation of disputes to war within five years. While controlling for rivalry, “over the whole COW period, states engaging in disputes characterized by mutual military buildups are at least 250 percent more likely to escalate to war within five years than when disputes occur without mutual military buildups.” 91

The Dyadic Level of Analysis, Part II

305

Using a separate check, she looks specifically at states that are not rivals. For this subset of states, she finds that disputes between nonrivals were more than three times as likely to escalate to war within five years if the pair was involved in a mutual buildup compared to pairs that were not involved in mutual buildups. 92 The relationship between arms buildups and escalation to war is not spurious and is not simply an artifact of the existence of dyadic rivalry. These findings are consistent with those of several other researchers who look at military buildups in the context of rivalries and the steps-to-war approach of Vasquez and Senese, and therefore appear to be quite robust. 93 Sample’s research provides strong evidence for the steps-to-war approach that we will take up next. Specifically, the evidence suggests that arms races—like the existence of dyadic rivalries, alliances, and serial crises and disputes—constitute a “step to war,” in which the presence of each step independently increases the likelihood of war for the dyad. The work of Vasquez and Senese makes it clear that the probability of war for a dyad involved in a mutual military buildup increases if other factors—territorial disputes, outside alliances, and rivalries—are also present. Another concern regarding the relationship between arms races and war involves the effects of historical context. Many have suspected that the relationship between arms races and war is primarily an historical artifact of a particular period in time that has now expired. More specifically, the argument is that the presence of nuclear weapons in the post–World War II era has changed the nature of the ballgame. Sample has provided an interesting indirect test of this proposition. In investigating which major power MIDs escalated to war, she broke the data into two periods: 1816–1945 and 1945–93. She found mutual military buildups between 1816 and 1945 to be positively associated with the escalation of MIDs to war, but this result does not hold for the period after 1945. 94 In this later period, which coincides with the age of nuclear weapons, disputes between states that are heavily arming are not highly associated with the escalation to war. This is, of course, not a direct test of whether the possession of nuclear weapons is responsible for the discrepancy. (Sample admits that virtually every major power dispute in the time period occurs when at least one of the disputants possesses a nuclear weapon; thus, the variable—possession of nuclear weapons—does not vary very much.) What we can conclude is that major powers behaved differently after 1945. A second test added data for minor powers as well as major powers, allowing a test for all states. The results once again showed that arms races increased the risk of war in the 1816–1945 period but not afterward. 95 Subsequent examinations by Sample and by Senese and Vasquez all indicate the same thing: the post–World War II era appears to be different. The relationship between mutual arms buildups and escalation to war disappears after World War II. 96 The evidence thus far implies that arms races ought not to be considered as root causes of war, but as part of a dynamic interactive process that also involves other dyadic factors (like militarized disputes, crises, and alliance-making) that serve to ratchet up the level of fear and hostility and escalate interactions between states further and further up the ladder of hostility. 97 Arms races increase the chance of war simply because they increase the amount of tension and hostility and the perceptions of threat that probably already exist between nations. To be fair, of course, spiral theory does not argue that arms races are root causes of war, only that they are part of an overall environment of reciprocated tension and hostility between states and that they play a role in the more general conflict escalation process that leads to war.

306

Chapter 8

THE STEPS-TO-WAR THEORY Since the reader is now familiar with territorial disputes, rivalries, dyadic interaction processes, and arms races, it is appropriate to introduce a theory that makes use of all of them. John Vasquez has put forward a theory of war that, while primarily at the dyadic level, incorporates factors from the substate level and the international system level. It emphasizes the notion that wars do not occur out of the blue but tend to be the product of a cumulative process involving several steps—hence the title steps-to-war theory. 98 It should be mentioned at the outset that Vasquez is not proposing a general theory of war that applies to all cases. He recognizes that there may be several different paths to war; his theory only attempts to ferret out the most common path to war, especially among major states. It is a theory that applies to wars of rivalry and wars of equality—wars between states of relatively equal power. Wars between unequal powers probably have different causes, and the steps to war between such states take a quite different path. For Vasquez, the first step is the existence of a territorial conflict between two states— typically contiguous neighbors. This conflict over territory is seen as an underlying cause of war. Vasquez is on solid empirical ground here; out of all the issues over which states might go to war, territory is the most dangerous. However, the presence of territorial conflicts does not always lead to war. It depends on how the two states handle this conflict. His argument is that the use of realpolitik methods—threats, the demonstration of force, enlarging military capabilities, the creation of alliances, and so forth—vastly increases the probability that territorial conflicts will lead to crises and MIDs, which in turn escalate to war. Territorial disputes, therefore, set off a train of events that lead to war through several stages or steps. The steps-towar model thus incorporates the dyadic conflict spiral model. Drawing on insights from Leng, it also makes use of the fact that spirals take place not just within a particular crisis, but the spiraling takes place across successive crises, as each subsequent crisis between states becomes more intense and antagonistic. 99 This continued hostile interaction also leads to the creation of rivalries—if they do not exist already. The use of realpolitik tactics is the proximate cause of war. As Senese and Vasquez note, they are proximate because they are closer to the outbreak of war, and they are causal since the likelihood of war would be greatly reduced if territorial disputes were handled differently. 100 The core argument here constitutes a frontal assault on realism; the steps-to-war roadmap is essentially a “realist road to war.” 101 A number of key variables are believed to precede the outbreak of war between two states, all of which are essentially dyadic in nature: the existence of territorial disputes (that is, MIDs based on territorial issues), the existence of repeated crises, the existence of rivalry, alliance making, and arms races. These are treated as risk factors. Just as the existence of certain risk factors increases the probability of heart attacks for individuals (obesity, high cholesterol, diabetes, smoking, family history, etc.), the presence of these risk factors increases the likelihood of war. The more risk factors that are present, the greater the probability of war. As the theory is conceptualized, the order in which the steps take place is relatively unimportant. Each step increases the amount of insecurity and the level of threat and hostility perceived by the two states. What matters most is that every step is mutually reinforcing. 102

The Dyadic Level of Analysis, Part II

307

The likelihood of escalation is also affected by factors at the substate level. As we have seen previously, Vasquez makes a distinction between hardline leaders and accommodationists. All things being equal, governments dominated by hardliners will be more likely to use realpolitik practices (what Vasquez refers to as power politics) in disputes with other states. Vasquez argues that the presence of hardliners is an important prerequisite for war. 103 Whether hardliners or accommodationists dominate the domestic political environment hinges largely on the results of the last war the country was involved in: whether the country was victorious and (victorious or not) whether the war was deemed to be worth the costs. If the answer to both is yes, hardline viewpoints are likely to be in the ascendant; if the answer to both is no, accommodationist views are likely to predominate. Regardless of which group is in power, if the opposing state uses power politics in a dispute, the internal political effect in the rival state is to strengthen support for hardline views. Repeated crises will tend to increase the number and the influence of hardliners in power. Once realpolitik policies are adopted by one country, it will be hard for a rival not to respond in kind. As the conflict spiral picks up steam and the policies of the accommodationists are seen as ineffectual, decision makers will be pushed toward more realpolitik tactics, and accommodationists will eventually be driven from power. With hardliners on each side, concrete and tangible stakes (territory) are transformed into intangible stakes that are not easily divisible and are infused with symbolic qualities that make the issues more intractable. In addition, perceptions of threat and hostility increase, the level of anxiety increases, and the prospects for compromise are reduced. Leaders in both states begin to see the other in terms of a rivalry, if such a rivalry does not already exist. Repeated crises and militarized disputes (and failure to successfully resolve issues) exact a psychological, emotional, and political toll on leaders in both countries that lays the groundwork for escalation of future steps. In sum, the dyadic interaction between states fuels a rather nasty mutual feedback loop that affects the domestic political environment in both states in ways that increase the probability of armed conflict. Once this war of rivalry breaks out, whether it will expand to become a world war is largely determined by structural conditions at the international-system level. Vasquez’s original argument is that there are three necessary conditions for the expansion of wars to world wars: (1) multipolar distribution of power, (2) alliance polarization, and (3) the creation of two blocs, neither of which has preponderance of capability. Other factors that have a multiplying effect on the possibility that the war will expand are (4) alliances, (5) contiguity of potential belligerents, (6) rivalry, (7) great-power intervention, (8) economic dependence, and (9) the breakdown of systemic order. 104 Conditions at the international system level also play one other role in the theory. War is more probable to the extent that the system lacks strong norms against the use of force and to the extent that the system lacks effective mechanisms and institutions of nonviolent conflict resolution—especially with regard to the possession and transfer of territory. Vasquez agrees with constructivists who argue that war is a learned activity and is to some extent a product of our culture. He maintains that for several centuries the dominant view of international relations throughout the system has been that of realism. In other words, a major aspect of the political culture of the international system has been a world view that emphasizes the utility

308

Chapter 8

of force and the accumulation of power in order to enhance security. Realism is an ideational system that includes a set of operational codes for statesmen concerning the proper tactics and strategies to pursue in their relations with rival states. The presence of such a culture is a permissive factor that makes war more likely. These systemic variables are treated as “deep background” factors. 105 (An alternative culture that emphasized accommodationist practices would be a dampening factor on the probability of war.) Steps to War: Empirical Findings The primary tests of the theory are found in Senese and Vasquez’s The Steps to War. The study uses MID data for the years 1816–2001. The authors start by confirming the territorial basis of the theory; the type of issue does matter. Pairs of states having territorial claims are more likely to have militarized disputes with each other than if the dispute is based on policy issues or regime types. Once states are involved in territory-based MIDs, this increases the probability they will be involved in a war with each other within five years. Senese and Vasquez estimate that the probability that a territorial MID will escalate to war within five years is about .28—which is five times the probability for MIDs based on policy disputes or regime disputes. 106 The potency of the relationship between territorial MIDs and war holds even when they control for the effects of the original territorial conflict. It isn’t the mere presence of a territorial claim that leads to war, but whether the claim is handled by power politics methods. The next step is to examine the effects (if any) of contiguity on MIDs and war. Using data only from the 1919–95 period, they discover that both contiguity (opportunity) and territorial issues (willingness) are important in the onset of MIDs; the two variables have separate and independent effects. However, territory is a much more important factor in the process by which MIDs escalate to war. The primary role played by contiguity is in increasing the probability that a MID will occur in the first place, not its escalation to war (that depends on the issue). In fact, given the existence of a territorial dispute, contiguous dyads are three times less likely to escalate the dispute to war than are noncontiguous pairs. 107 In the absence of disagreement over territory, MIDs are unlikely to escalate to war. However, when territorial conflict is present MIDs are likely to escalate to war, but this is most likely when the borders of the two states are not contiguous. In sum, Senese and Vasquez argue that it is not mere opportunity (contiguity) that is the key to war, but willingness (as signified by the presence of territorial grievances). 108 Once the importance of territory has been confirmed, the next step is to examine the contention that the various practices involved in power politics (which constitute the steps to war) increase the chances of war. Senese and Vasquez first seek to determine what distinguishes dyads that go to war from dyads that have MIDs but do not end up in war. (They look at all dyads that have at least one MID over their history—a pool of dyads that can reasonably be deemed to face some risk of war.) 109 They hypothesize that the practices associated with power politics should make the difference. Their independent variables are indicative of such practices: dyads dominated by territorial differences, the presence of certain types of alliances, the presence of enduring rivalries, recurring territorial MIDs (four or more), and arms races. They hypothesize that as each of these factors is added, one should see incremental increases

The Dyadic Level of Analysis, Part II

309

in the probability that the dyad will experience war. They break their analysis into three periods: 1816–1945, 1946–89, and 1990–2000. Generally, they find that how states handle their conflicts makes a difference; the tactics associated with power politics are indeed problematic. However, the steps-to-war explanation clearly fits the 1816–1945 period best. For this lengthy era, all five variables have the hypothesized effect, and their effect is cumulative: the probability of war for the dyad increases as each risk factor is added. The variables that have the largest impact are enduring rivalries and the recurring territorial disputes (four or more MIDs). When all five factors are present in a dyad—dominance of territorial disputes, both sides having an outside alliance, recurring territorial MIDs, enduring rivalries, and arms races—the probability of war is an astonishing .935. 110 But even for states that have regime or policy disputes instead of territorial disputes, the presence of power politics factors increases the probability of war. How states handle their disputes matters, even when territory is not the issue. (The fact that power politics techniques substantially boost the probability of war for dyads with nonterritorial disputes suggests a separate path to war.) The Cold War period (1946–89) differs considerably, and in ways that are not entirely consistent with the steps-to-war model. While territorial disputes and rivalries still increase the probability of war, arms races have no effect, and most importantly, alliances may actually make dyads less prone to war. Unlike the earlier period, dyads in which the states were allied with each other perfectly predicted peace. Most importantly, rival dyads in which both states had outside allies had lowered probabilities of war. While the post–Cold War period (1990–2000) is a rather limited slice of history, what can be said about it is that it looks more like the early period than the Cold War. Senese and Vasquez next look at the issue from a slightly different angle. They ask the question of which crises or MIDs—they essentially view the two as synonymous—escalate to war and which do not. Specifically, their dependent variable is whether the current MID or any other MID between the same pair will escalate to war within five years. Their independent variables reflect the steps to war: dispute type, alliances, rivalry (as indicated by the number of prior militarized disputes), and arms races. Overall, they find that the most dangerous condition is when multiple factors are present, as the theory predicts. Once again, the results for the 1816–1945 period fit nicely with the theory. All four risk factors have the expected impact and adding each risk factor on top of another results in a higher probability that the MID will end in war. When all four factors (steps) are present, the probability of war is a whopping .921. 111 And while nonterritorial disputes have a lower probability of escalating to war than territorial disputes, handling them with power politics tactics increases the risk of war in a similar progressive fashion. Once again, the Cold War diverges significantly from this pattern in interesting ways. Alliances may reduce the risk that MIDs escalate to war if both states have outside alliances or if both are allied to each other. Even in a situation where disputes recur and rivalries are created, if both sides have outside alliances, the probability of war is greatly reduced. This is of course contrary to the steps of war logic, but it is consistent with the logic of the Cold War and the deterrent effect of major alliances systems—NATO and the WTO—both of which possessed nuclear weapons. Arms races are, once again, insignificant. On the other hand, the

310

Chapter 8

effects of territory and recurring disputes appear to be similar to the earlier period, and they act in the way predicted—with one interesting exception. Recurring disputes have a decidedly curvilinear pattern; they increase the risks of war, but only up to a point. After a certain point—the number appears to be twenty-nine MIDs—dyads with a very large number of disputes learn to ritualize their conflicts and the likelihood of war actually decreases. While the Cold War is generally an era with less war, the wars that do occur come from the escalation of territorial disputes, which are still the most dangerous type of conflict. As before, the post–Cold War period appears more like the 1816–1945 era than the Cold War. However, one interesting anomaly exists: The probability of war does not increase with repeated disputes. Instead, in this period states have tended to escalate the first or second MIDs to war. This may be a temporary phenomenon, and the authors doubt the pattern will persist. While a fundamental finding is that the highest probability of war occurs when all four steps are present—territorial disputes, rivalries, alliances, and arms races—it is also clear that wars often occur before all four steps take place. Valeriano and Vasquez examine all complex interstate wars in the international system from 1816 to 1997 to determine how many steps are present prior to the outbreak of war. They discover that only a handful of complex wars were preceded by all four steps, and war frequently broke out with just a few steps in place. 112 The general findings of Senese and Vasquez have been supported by a variety of other studies that demonstrate that many of the variables identified by the steps-to-war model interact with each other to increase the probability of war. For instance, the work of Colaresi, Rasler, and Thompson on strategic rivals confirms the significance of territorial disputes, external alliances, rivalries, and mutual military buildups as factors that promote war. An early study using ICB crisis data for the post–World War I period indicated that the combination of rivalries (especially spatial/territorial rivalries), recurrent crises (two or more crises), an arms race, and having external allies increased the probability of war. However, their emphasis in Strategic Rivalries in World Politics is on the combined triple whammy of contiguity, territorial disputes, and rivalry as being the most important factors. 113 In a separate study, Valeriano and Vasquez analyze complex wars to determine whether complex wars are arrived at through a separate path than dyadic wars. 114 (Complex wars are wars involving three or more parties. The presumption is that they represent a different path to war. While contiguous rivals tend to be involved in dyadic wars over territory, noncontiguous rivals fight by joining already-existing wars.) Looking at interstate wars between 1816 and 1997, they categorize wars as dyadic (fifty-one cases for 64.6 percent) or complex (twentyeight cases for 35.4 percent). Complex wars are further delineated by the number of dyads involved: small complex wars (two to five dyads, twenty-two cases), medium complex wars (seven to twenty-eight dyads, four cases), and world wars (thirty-seven to sixty-six dyads, two cases). What differences do they find between the paths to war for complex wars as opposed to dyadic wars? They discover that 70 percent of all complex wars involve rivals with escalatory patterns of interaction, while this pattern was found in only 39 percent of dyadic wars. Second, 78 percent of complex wars were preceded by at least two power politics practices (either arms races, alliance formation, or rivalry), but only 31 percent of dyadic wars had this combi-

The Dyadic Level of Analysis, Part II

311

nation in the prewar period. Moreover, the larger the war, the more likely it was to have at least two of the practices of power politics. The two world wars of the twentieth century were associated with exactly the path one might predict: They were the only complex wars to be preceded by all three practices—arms races, rivalries, and the presence of relevant outside alliances. This finding is especially interesting because in their earlier work, Senese and Vasquez compute that if all three of these factors are present, there is a 90 percent chance that a given dispute will escalate to war. Obviously, the simultaneous existence of these conditions is rare and only a small group of wars, both dyadic and complex, reflect this pattern. Steps-to-War Theory: Summary The steps-to-war theory has been one of the most fruitful and productive research programs in international conflict theory over the past two decades. While the empirical evidence has not always lined up perfectly with the predictions of the theory, it has nevertheless produced solid evidence of a particular path to war pursued by major states throughout most of the last two centuries. The theory provides strong support for the fundamental importance of dyad-level variables in determining whether disputes escalate to war or not. Vasquez and his colleagues provide general confirmation of the pattern that the theory predicts: the use of power politics tools increases the probability of war, especially for dyads having territorial disputes. Territorial disputes, repeated MIDs, and the presence of rivalries are important factors in the escalation to war (and they can probably be seen as implying the presence of a conflict spiral). Alliances and arms races are also important, though this depends on the period of time. The theory provides theoretical buttresses to the spiral theory while simultaneously presenting a major challenge to realist arguments—both theoretical and prescriptive. It especially challenges the realist notion—examined at length in the next chapter—that threats and acts that demonstrate toughness are useful in deterring attacks from others. One important aspect of the steps-to-war model has been temporarily dropped from the research project: the mutual feedback loops linking the dyadic interaction between states and their domestic political environments. One can certainly make an argument for this in the interests of greater simplicity and parsimony. At any rate, the role that domestic political pressures play on ratcheting up the escalation process has not been subjected to analysis. One suspects that these domestic level factors are more amenable to qualitative measures than to quantitative measures. Thus, they have been left behind in the attempt to use large N statistical studies to provide support for the theory. Attempts to examine these important links are perhaps best done through case studies. It’s clear that the steps-to-war theory is to some extent time-bound. While the general notion that hardline policies lead to conflict escalation and to war makes sense, the emphasis on alliance-making and arms races may be misplaced—and their effects are clearly dependent upon the era in question. The Cold War really was different, due to its tight bipolar alliance configuration and the presence of nuclear weapons. The model is clearly better fitted to certain periods of time than others. The steps-to-war theory is also a theory of interstate war intentionally designed to apply within somewhat limited boundaries. It focuses on one path to war (characterized by territorial

312

Chapter 8

disputes between rivals that escalate through the use of power politics) for one type of war (between equals, especially major power equals) during one historical era (1816 to 1945). While the theory implies a dynamic temporal sequence or chain of events, Vasquez and his collaborators have thus far generally refrained from positing a specific sequential pathway as most likely to eventuate in war. Senese and Vasquez do, however, offer their view that the typical sequence is that territorial disputes lead to crises and a perception of rivalry, which then leads to alliance formation and arms races. Vasquez’s case study of the India-Pakistan rivalry and its four wars shows that the steps occur in the predicted order, except for the first Kashmir War (1948). The dyadic rivalry and territorial disputes occur simultaneously: India and Pakistan are “born feuding.” Furthermore, there is no time for either alliances or arms races to occur before the outbreak of war. Only in the last two wars—Bangladesh in 1971 and Kargil in 1999—do all four steps precede the outbreak of war. 115 William Thompson’s team, working with territorial claims data, MID data, and their own data on rivalries for the twentieth century, discover two dominant paths. In one, territorial claims lead to militarized disputes between nonrival pairs. In the second, territorial disputes between rivals lead to militarized disputes. 116 The escalation of disputes in the absence of rivalries is somewhat unexpected, but is not completely inconsistent with the steps-to-war approach. Although the use of the term “steps” may imply a sequential path, the steps are usually presented (and tested) by Vasquez and his associates as a combination of variables that occur simultaneously and that collectively increase the chances of war. As Levy notes, The implicit logic underlying most of these analyses . . . is that the more of these factors that are present and the higher their magnitude, the higher the probability of war. The more intense a territorial conflict, rivalry, or arms race, the greater the probability of war. In addition, a higher value on one variable can compensate for a lower value on another variable. Thus, these factors collectively constitute an additive linear model of the probability of war. 117

Levy suggests that there are several possible sequential paths to war with no single variable common to all of these potential paths, leaving a situation in which no single factor is either a necessary or sufficient condition for war. For instance, one might see the realist road to war as a first-order causal path to war that has nested within it several second-order paths, each with a different combination of variables; in turn each of these second-order paths have nested within them a different sequence of events that constitute third-order paths to war. Of course, the problem is that a theory that delineates so many causal paths to war runs the risk of having too many paths with too few historical cases per path. The cause of any particular war must be explained by a unique set of sequences. This would fundamentally undermine the ability of the theory to make useful generalizations about war. GRIT AS A SOLUTION TO THE CONFLICT SPIRALS The implications of stimulus-response theory, conflict spirals, the steps-to-war theory, and the security dilemma are fairly obvious. A clear-cut strategy for peace can be derived logically from them. If it is true that states are primarily motivated by fear and if war is the result of a reactive process that spirals upward, then it is necessary to intervene in some way in the

The Dyadic Level of Analysis, Part II

313

conflict process to reverse the escalatory spiral. While the bad news is that aggression, violence, and hostility are reciprocated by others, the good news is that so are cooperative and conciliatory actions. Presumably, all that is required is for one country to take the initiative. Instead of returning an eye for an eye, it must turn its cheek and even act charitably toward its antagonist. Many years ago, in the midst of the Cold War, Charles Osgood devised just such a strategy for reducing the hostility that existed between the United States and Soviet Union. He called his strategy GRIT—graduated and reciprocated initiatives in tension reduction. 118 Osgood suggested that unilateral actions, coupled with clear verbal statements, could be used to initiate a reciprocal de-escalatory spiral. By communicating through deeds as well as words, states could learn mutual trust and could reduce the level of tension and hostility that existed between them. Osgood offered several guidelines for the application of this strategy. 119 1. All unilateral actions should be announced publicly prior to their execution and should be specifically identified as part of a deliberate policy of reducing tensions. This enlists world public opinion and augments pressure for reciprocation. 2. The announcement should clearly state an explicit invitation to the rival to reciprocate the initial action with a unilateral response of his own choosing. Reciprocity need not be in the same form as the original act or equal in quantity, but it must be made clear that some form of reciprocity is expected. 3. In order to convey a real commitment to the reduction of tension, unilateral actions that have been announced must be executed on schedule even if the rival does not immediately signal a willingness to reciprocate. 4. Unilateral actions should be in the form of overt deeds rather than positive or negative sanctions. While sanctions are contingent on some prior action of others, Osgood prefers that GRIT actions be deeds that are immediate rather than contingent. The actions should be visible, unambiguous, and clearly verifiable. Overt actions are in this sense also better than non-actions such as moratoriums on nuclear testing. 5. Unilateral actions should be planned and arranged in a graduated series—from least important (and least risky) to most important (and most risky). Osgood suggests that unilateral actions should start out on a small scale and progress from there. As the initiator’s actions are reciprocated, more far-reaching policy actions can be contemplated. 6. It is probably best to begin GRIT with actions that are not in the military/disarmament sphere, but move to these areas later in the process after trust has been built up through reciprocating actions in other areas. Initially, unilateral actions should be designed not to reduce a nation’s ability to deter aggression. GRIT should not involve taking risks that would leave the nation vulnerable to its enemies. The lower the reciprocity, the less risky the actions that should be contemplated. 7. Unilateral actions should be diversified so as not to weaken a nation in any one sphere. Tension reduction can be cumulative over a range of actions in different fields: economic, diplomatic, cultural, and military.

Chapter 8

314

8. Unilateral actions should be continued over a considerable period of time regardless of immediate reciprocity or the lack thereof. In the early phases of the strategy, reciprocation may not be forthcoming, but it is important to continue on one’s schedule of unilateral, tension-reducing actions nevertheless. This is necessary because one’s initial acts may be regarded by the rival as mere tricks or ploys for purposes of propaganda. The genuineness of the acts and the intention of the initiator will become certain to the rival only over a period of time. In essence, GRIT entails a learning process. The initiator must teach his rival that he really wants cooperation and he must provide an environment in which this message can be learned. On the other hand, the rival will eventually have to teach the initiator that he, too, is interested in cooperation rather than taking advantage of the initiator’s acts of generosity. 9. If encroachment occurs (that is, if the rival takes advantage of your unilateral concessions to do you harm), it must be resisted firmly. The resistance should be targeted exclusively in the area of the encroachment; tension-reducing moves should be continued in other areas. In other words, there should be no linkage between an isolated incident of bad behavior in one area and continued progress in another area. 10. GRIT is a flexible, self-regulating policy. The two sides constantly monitor each other’s actions and communicate through deeds. If reciprocity is not attained, the policy can eventually be abandoned without harm to the initiator. GRIT: Evaluation The major problem in implementing GRIT seems to be in initiating the first move. Unilateral concessions are about as popular with world leaders as blind dates are with undergraduates. In both cases a tremendous amount of courage is required to take the first step. Once a country has undertaken a GRIT policy, however, there are some fairly powerful inducements for its rival to play along. GRIT will, of course, be a very controversial policy within the initiating country. The only way to contain the domestic opposition will be to attain success—the quicker the better. Leaders in the rival state will, of course, understand that if they fail to reciprocate, they risk strengthening domestic forces in the initiating state that are more antagonistic toward them. The two sides are therefore mutually dependent on each other for success. A second type of problem with GRIT is that it requires more flexibility and maneuverability in policy making than most governments can muster. Domestic considerations and the vicissitudes of factional, organizational, and bureaucratic politics may place serious limits on the abilities of governments to pursue a GRIT strategy. 120 Your reaction to GRIT may be that it is a rather risky experiment designed to test a harebrained, academic theory of international relations using the real world as a giant laboratory. Clearly, leaders should not subject their citizens to dangerous experiments, but experiments per se cannot be entirely avoided. In a sense, all government policies are experiments based on implicit theories of how the world works. Obviously, a test of GRIT in the real world would be difficult to arrange. Tests of GRIT have occurred primarily through simulation and gaming. For instance, experimental gaming studies carried out by a team of psychologists led by Svenn Lindskold at Ohio University have concluded that GRIT is an effective strategy for fostering trust and cooperation. Furthermore,

The Dyadic Level of Analysis, Part II

315

GRIT is effective in inducing cooperation for both men and women, for both individuals and groups, whether the choices are made sequentially or simultaneously. 121 In their study of U.S.-Soviet-Chinese interaction from 1948 to 1989, Goldstein and Freeman examine six cases of cooperative initiatives, one by each of the superpowers toward each of the others. 122 They find that five of the six attempts were successful in creating cooperative bilateral reciprocity. (The one failure was Gorbachev’s unilateral nuclear testing moratorium aimed at the Reagan administration.) In virtually all cases, the initiator had to overcome significant policy inertia in the target state. As a result, success in gaining reciprocal cooperation depended on the willingness of the initiator to persevere after initially disappointing responses. This suggested to Goldstein and Freeman that an extended GRIT policy (or EGRIT), in which the initiator makes sporadic (rather than continuous) unilateral gestures over an extended period of time might be the best policy to gain mutual cooperation. Goldstein and Freeman performed computer simulations on their Soviet-American interaction data in order to test the effectiveness of several different strategies in inducing cooperation. The strategies were (1) TIT-FOR TAT (TFT), which involves a single initial cooperative move and then simply reciprocates the rival’s previous move; (2) EGRIT; (3) progressive GRIT (PGRIT), which is essentially the GRIT strategy outlined above, which starts with limited initiatives and progressively moves on to actions of greater and greater importance; and (4) a standard GRIT strategy in which the unilateral initiatives do not progress from minor to major, but instead remain constant at an intermediate level. Once again, the results indicate the superior properties of “soft-line” strategies. Although all the strategies were somewhat effective in eliciting some cooperation, EGRIT and PGRIT were the best strategies. Mutual cooperation lasted longest when EGRIT was used, while PGRIT produced higher levels of cooperation. No strategy was successful in the simulation in maintaining cooperation over the long term, however; Soviet-American relations returned eventually to their steady state of hostility—an effect the authors maintain is due to policy inertia in both states. 123 Goldstein and Freeman conclude that the best strategy to induce cooperation might be a “super-GRIT” policy, which would essentially be a policy of permanent EGRIT in which unilateral initiatives would be pursued at a level that could be sustained over a number of years. 124 In fact, the authors argue that Soviet President Gorbachev followed just such a policy in his “new thinking” approach to foreign policy beginning in 1985. Unilateral concessions were made in order to induce cooperation from both the Chinese and the Americans. With regard to the Chinese, Gorbachev’s Vladivostok speech in 1986, announcing Soviet willingness to meet long-standing Chinese conditions for better relations, resulted in normalization of Soviet-Chinese relations. Winning cooperation with the United States was not as easy. A number of Soviet unilateral undertakings were initiated without significant reciprocation: the Soviet nuclear test moratorium of July 1985; the Soviet military withdrawal from Afghanistan in December 1988; Gorbachev’s United Nations speech in December 1988 announcing a unilateral Soviet decision to cut the Soviet military by 500,000 troops over two years and to remove or destroy substantial numbers of tanks, artillery, and aircraft; and the May 1989 promise to unilaterally withdraw 500 nuclear weapons from Europe. Until the end of 1989, according to Goldstein and Free-

316

Chapter 8

man, the Soviets seemed to be doing the initiating and the reciprocating, while the United States rarely initiated and reciprocated only weakly. Improvements in Soviet-American relations took place primarily because of Soviet willingness to persist in making concessions even though the American response had been meager. By the end of 1988 all the Soviets had to show for their efforts was the INF (intermediate range nuclear forces) agreement, a treaty made possible primarily by Soviet willingness to make significant substantive concessions to the United States. The next two years changed all that. A series of spectacular events transformed U.S.-Soviet relations (and international relations in general) in a shockingly brief period of time. The fall of communism in the six former Soviet bloc states in Eastern Europe in the last four months of 1989 was made possible by Gorbachev’s willingness to abandon the Brezhnev Doctrine and allow peaceful transitions. The unification of the two Germanys in October 1990 under Western terms, which maintained German membership in NATO, was also made possible by Soviet concessions. These events were followed logically by the demise of the Warsaw Pact (and COMECON) in April 1991 and the process of removing all Soviet forces from Hungary, Poland, Czechoslovakia, and Germany. In the meantime, the Soviet Union had begun to transform itself from a one-party communist state to a fledgling democracy experimenting with market capitalism. These events led to a virtual landslide of reciprocity by the West. The CFE (conventional forces in Europe) Treaty was signed by the NATO and WTO states in November 1990, and the Soviets and Americans signed the START (strategic arms reduction) Treaty in July 1991. Numerous bilateral and multilateral trade and aid agreements were signed between the Soviets and Western states. Bilateral and multilateral summits proliferated. Then in September 1991, in the aftermath of the abortive coup that had attempted to remove Gorbachev from power, President Bush made a startling television speech. He announced a package of American unilateral actions, which included, among other things: (1) the withdrawal of all U.S. tactical nuclear weapons from Europe, (2) elimination of nuclear cruise missiles from certain types of naval vessels, (3) the removal of U.S. bombers from alert status, and (4) the termination of U.S. mobile ICBM programs. The president asserted that these steps would be implemented with or without reciprocation from the Soviets, but he challenged the Soviets to join the United States in taking equally bold and concrete steps. The American president had adopted the language and the strategy of GRIT. This is the most direct use of the GRIT strategy ever used by a world leader. Gorbachev’s earlier unilateral initiatives were accompanied neither by unambiguous communications about Soviet expectations nor by clear invitations to the United States to reciprocate Soviet actions in an open-ended way. At most, the United States was invited to reciprocate a single specific act, such as the nuclear test moratorium, and this request was not put in the context of further Soviet gestures if the United States reciprocated. The Soviet response to President Bush’s speech was not long in coming. Approximately a week later, Gorbachev announced his own unilateral actions. He directly reciprocated most of the American cuts and then went several steps further by once again suspending Soviet nuclear tests and by pledging that the Soviet strategic arsenal would be kept at a level of 1,000

The Dyadic Level of Analysis, Part II

317

warheads below the American arsenal. He then proposed the two countries agree on a further 50 percent cut in strategic weapons. Within months a second round of GRIT proposals was underway. In his State of the Union address in January 1992, President Bush presented a package that included unilateral defense initiatives (some old, some new) and challenged the new Russian government of Boris Yeltsin to reciprocate. Yeltsin’s reciprocation was virtually simultaneous. (The joint announcements had obviously been worked out in advance—a new twist on GRIT strategy.) It is clear that a number of significant changes had taken place between 1985 and 1991, which had made possible the successful use of GRIT strategies by the two superpowers. In large part, Gorbachev’s continuous unilateral initiatives and his willingness to make concessions were able, over a period of years, to teach the leaders in the United States (and China) that their previous impressions of the Soviet state needed to be changed and that Soviet desire for cooperation was not just a ploy, but was sincere. Trust had begun to replace suspicion in the relationship between the former enemies. Of course, the fact that the democratic experiment in the USSR had led to the demise of communist control not only in that country but in all of eastern Europe also played a role in creating a totally different international environment. Neither should one discount the fact that on both sides of what had once been the Iron Curtain budgetary crises compelled drastic cuts in defense. Regardless of the exact causes, the Cold War climate of mutual hostility, suspicion, and tension had given way to the realization by both sides that cooperation in the pursuit of mutual interests was possible. Under these circumstances, the American president was essentially freed from the domestic political constraints that had bound foreign policy making for decades. He could, for the first time, undertake unilateral military cutbacks without fear of a conservative political backlash leading to defeat at the polls. A “soft line” was at last acceptable. Goldstein and Freeman, whose analysis ended in 1989, argued that decision makers in the three superpowers have operated over the last several decades in an environment in which limited cooperative reciprocity was present but was made difficult by the existence of high levels of policy inertia in all three states. 125 As the Cold War was coming to an end, the environment was one in which policy inertia (at least for the Americans and the Soviets) had given way to a rush toward cooperation. CONCLUSION Examination of the spiral theory and steps-to-war theory at the dyadic level provides some fairly strong insights into the causes of war. In the next chapter, we examine two other theories of international interaction: game theory and deterrence theory. Since we remain at the dyadic level, you will notice a number of distinct similarities in the themes that are examined. Indeed, we will revisit many of the major concepts introduced in this chapter—conflict spirals, the security dilemma, arms races, the strategies of peace through strength, and peace through conciliation and cooperation—providing alternative perspectives on each.

Chapter Nine

The Dyadic Level of Analysis, Part III Game Theory, Bargaining, and Deterrence Theory

Come, Watson, come. The game is afoot. —Sir Arthur Conan Doyle More nations can back themselves into war through retreat and appeasement than by standing up for what they believe. —Ronald Reagan If you strike steel, withdraw; if you strike mush, push on. —V. I. Ulyanov (Lenin)

GAME THEORY Dyadic interaction between states can be seen as a game in the sense that states compete with each other to attain certain rewards. In some types of games only one winner can attain the reward; in other games there can be multiple winners or losers. The essential element of politics as a game, however, is that the strategies followed by each player (country) are interdependent. Each player has to take his opponent’s interests and strategies into account. The best course of action for each player will depend on what he expects the other to do. 1 Game theory has been developed from the disciplines of logic and mathematics to provide a way of understanding certain types of game-like situations and to assist in the development of strategies for making actual decisions. Its purposes are twofold. One is practical and normative—to help decision makers cope with certain situations in the real world by developing strategies for rational behavior. The other is theoretical and empirical—to help explain why certain actions occur in certain situations. Game theory might be useful in helping us understand a variety of international interactions in which strategies are developed by one side to counter the actions and strategies of others: crisis interactions, diplomatic bargaining, arms races, deterrence, prewar mobilizations, and decisions for war. Game theory, like all social science theories, is based on certain simplifying assumptions. The most important assumptions are these:

319

Chapter 9

320

1. People are rational (or at least most people, most of the time), and governments may be thought of as single, rationally calculating entities. Rationality in this sense means that each party seeks to maximize his or her own interests. 2. The utility (worth or payoff) of each outcome can be calculated and quantified on either an interval scale or an ordinal scale of relative desirability. This provides a standard by which strategies can be rationally compared according to their ability to contribute to the maximization of one’s interests. Game theorists frequently use a game matrix to illustrate the choices facing each player and the payoffs that can be expected for each possible outcome. A typical game matrix is displayed in figure 9.1. Each player has two strategies. Player I has strategies A and B; player II has strategies C and D. The strategy options may be the same for each player or they may be different. In this example they are different. Since each player has two strategies, four possible outcomes are depicted by the four cells of the matrix. (If there were more players or more strategy choices, there would be more cells.) The numbers inside each cell indicate the payoffs (utilities) for each player for each of the four possible outcomes. The payoffs for player I (the row or horizontal player) are always designated first and the payoffs for player II (the column or vertical player) are placed second, in parentheses. The choices are usually assumed to be made simultaneously, and the outcome depends on the mix of strategies chosen. This particular game is a game of pure conflict. One player always wins and the other player always loses. There are no chances for mutual gain or mutual loss. Further, what one player wins corresponds exactly to what the other player loses, and the sum of the numbers in each cell is therefore zero. This type of game is called a zero-sum game. The game depicted in figure 9.1 is a two-person, zero-sum game developed by Martin Shubik, a pioneer in game theory. 2 Player I is a police force and player II is a rebel guerrilla unit. The police can choose either (A) to enter the jungle in pursuit of the rebels, a dangerous strategy since this is the rebels’ home territory, or (B) to protect the cities, areas where they presumably have more support. The guerrillas have a choice of (C) full-scale, open battles or (D) small-scale skirmishes fought with unorthodox techniques. From the perspective of the police, city protection is the preferred strategy. If they pursue the rebels into the jungle, they lose—whether in open battle or in skirmishes—it’s just a question of how much. The guerrillas also have a dominant strategy. Whether in the jungle or in the cities, they stand to win more and lose less if they restrict the battle to skirmishes. Thus, both sides have a dominant

Figure 9.1. Game Matrix: Guerrilla Warfare

The Dyadic Level of Analysis, Part III

321

strategy—a strategy that they should try to play consistently given the structure of the situation, the choices available to them, and the payoffs. The interests of each side therefore converge in the cell B,D in the lower right quadrant. The police will stay at home and protect the cities and the rebels will attack with skirmish techniques. This solution represents a saddlepoint, an equilibrium point at which the strategies of each side logically converge. If a game has a saddlepoint (not all do), it usually denotes a minimax (or maximin) solution. A minimax solution represents the best each player can do against a fully rational opponent. It is also a solution in which each player attempts to minimize his losses if the other does his worst. Knowing that the police are not foolish enough to enter the jungle but will protect cities instead, it is rational for the rebels to choose to lose only three by engaging in skirmishes instead of losing nine, which would be the outcome if they chose to attack in open battle. It is a general axiom in two-person zero-sum games that the best strategy for each is based on the minimax principle: each player should try to maximize his minimum gain or minimize his maximum loss. This is a conservative strategy and applies only to zerosum games. Chicken Games Not all situations are zero sum in nature; in fact, most political situations include elements of cooperation as well as conflict. Players may enjoy mutual gains or may suffer mutual losses. Players have common interests as well as an interest in besting the other. Common interests among states might include the avoidance of mutual negative payoffs (such as nuclear annihilation or the cost of an arms race) as well as the achievement of mutual benefits (such as increased trade or mutual access to the wealth of the seabed). Let us look at some non-zerosum games, games in which the sum of the payoffs in each cell do not add up to zero. Figure 9.2 is illustrative of a game commonly called chicken. 3 Each player has a choice of the same two moves: to drive one’s car straight down the center line of the road at high speed against the on-rushing machine of one’s rival or to swerve to avoid hitting the other car (thus “chickening out”). If both refuse to swerve, the result is a cataclysmic clash of chrome and steel with the high probability of injury and death for both drivers. The payoff for this mutual disaster has arbitrarily been set at −20; it could just as easily have been set at −50 or −100. The specific value doesn’t matter as much as the fact that it is a number significantly greater than the next worst payoff. If one driver swerves while the other does not (cells B,C or A,D), then the swerver suffers the indignation of being called a

Figure 9.2. Game Matrix: Chicken

322

Chapter 9

chicken (−5), and his opponent wins the admiration of the crowd (+5). If both swerve, they suffer mild humiliation, but since they share the dubious distinction of both being chicken, the blame is somewhat reduced (−2) for each and, of course, both live to blaze down the highways another day. The minimax strategy here dictates that if you assume the other driver will do his worst, the better part of valor is to swerve, suffer brief humiliation, and wake up alive the next day, thus minimizing your losses. Given the high cost of guessing wrong, you would probably not want to take the chance that the other player will swerve; it is much more prudent to assume she will not. On the other hand, believing that your opponent is a logical and prudent person, one might seek to take advantage of her prudence, but this is risky business. Nevertheless, if your goal is to win rather than to just cope with a bad situation, game theory provides some clues. In order to have a chance of winning a chicken-like confrontation, you must pursue a strategy designed to ensure that your opponent will believe beyond a doubt that you will not swerve and that therefore the only logical action for your adversary is to get out of your way. A winning strategy therefore depends on your establishing your credibility. Winners of the game of chicken are successful because they succeed in conning their opponents into believing they are committed to a strategy of suicide and that they are absolutely inflexible in this strategy. In a sense, success depends on demonstrating that you are not rational, but irrational—that you prefer death (−20) to humiliation (−5). If this belief can be generated, then the choice is purely up to your opponent to choose life or death for both of you. This is sometimes referred to as the “rationality of irrationality” or the “madman” strategy of deterrence in which national leaders actually try to cultivate a reputation for irrational or unpredictable behavior. 4 The success of this strategy depends on two things. First, you must establish your credibility. That is, you must make your opponent believe that you will do precisely what you have threatened to do. This can be done (in terms of the example) by visibly locking the wheel so that it is incapable of movement from side to side. Alternatively, you might stick your hands out the window, indicating that you will not be able to get them back on the wheel in time to turn. Both of these actions indicate to the other driver that you have no alternative but to continue straight ahead. Second, you must establish your credibility before your adversary steals your strategy and does exactly the same thing. The first one to adopt this strategy has the advantage because the burden of choosing life or death is placed on the opponent. Some tantalizing—though indirect—support for chicken tactics can be found in the empirical literature. There is considerable evidence that the first party to initiate a militarized dispute by threatening force or displaying force or using force is the side most likely to prevail in the encounter. 5 The central problem here is that what’s good for the goose is good for the gander. If the best way to win a crisis confrontation is to make an unbending commitment and charge ahead, then it is reasonable to assume that this posture will be adopted simultaneously by both sides. The strategy then becomes an iron-clad rush to destruction. Snyder and Diesing find that coercive (competitive) tactics are frequently effective in real-world chicken situations among countries, especially against weaker opponents. But such tactics run serious risks against rivals with relatively equal capabilities and interests. 6

The Dyadic Level of Analysis, Part III

323

In continuous or indefinitely repeated (iterated) chicken games, each driver may choose to drive straight (defect) rather than swerve (cooperate) on the assumption that this will coerce the adversary into swerving in future encounters. Each may seek to acquire a reputation as a hard-charging, nonswerving player. 7 Chicken games have been seen as applicable to deterrence relationships and to crisis confrontations, particularly brinkmanship crises. We have seen previously that such confrontations frequently lead to war, especially if leaders in one state perceive the rival’s no-swerve strategy to be a bluff rather than a real commitment. In such cases the saddlepoint of swerve/ swerve cannot be reached owing to a misperception of the opponent’s commitment. Of course, the saddlepoint may also be unattainable if one side genuinely perceives the payoffs for swerving to be so severe that humiliation is seen as equal or preferable to death. It might be argued that some of the interactions between the United States and Soviet Union during the Cuban Missile Crisis took on aspects of a chicken game. The United States signaled that it was willing to use force (including the use of nuclear weapons) against the Soviet Union if the latter’s missiles were not removed from Cuba, thus establishing a firm noswerve position. The Soviets could choose to proceed full steam ahead, but the U.S. had already demonstrated the credibility of its willingness to act rashly if pushed. This credibility had been conveyed through words (Kennedy’s public statements) and deeds (the naval blockade of Cuba and the alert of U.S. forces). Although American credibility to fight over the issue had been established, Soviet credibility to fight on behalf of her Cuban ally was lacking, and Soviet leaders probably realized this. The best they could do was to minimize their losses and retreat gracefully. Kennedy’s public promise to refrain from attacking Cuba, along with his secret pledge to remove American missiles from Turkey, made Khrushchev’s payoff for backing down (swerving) more palatable than it would otherwise have been and serves as a good example of how reducing the negative payoffs for the other side can help one side achieve its goals. Prisoners’ Dilemma Games Let us turn now to the most analyzed non-zero-sum game of all, prisoners’ dilemma, as depicted in figure 9.3. Two suspects are apprehended by the police and charged with a crime. A wily district attorney devises an ingenious scheme. He refuses to allow either suspect to communicate with the other, locking them in separate cells, and has them interrogated separately. Each is offered the same deal: 1. If you both refuse to confess, you will get off with ninety days for vagrancy (+10 each). 2. If only one confesses, implicating the other, the informant is freed and his partner gets life in prison (+20, −20). 3. If you both confess to the crime, you will both receive five years in jail (−10 each). What should you do? As always, it depends in part on what you think the other prisoner will do. If you are player 1 and you think 2 will cooperate with you, you can join in an effort to fool the police and get off with a vagrancy charge, or you can “defect” and go free—leaving your buddy to rot. Purely in terms of individual interests, it pays to defect if you think the other

324

Chapter 9

Figure 9.3. Game Matrix: Prisoners’ Dilemma

suspect will cooperate (keep silent). On the other hand, player 2 may decide to confess. If he talks, it would be foolish of you to remain silent and earn a life sentence in jail. If you also defect, you only get five years. So it pays to defect regardless of whether you think the other player will defect or cooperate. No matter what he does, your best response is to defect and assure yourself of the higher payoff. The other player, not being a complete dunce, will undoubtedly come to the same realization as you. He will also defect (confess). Defection is thus the dominant strategy for each player. The resulting solution is a saddlepoint at A,C (−10, −10) in the upper left quadrant, representing mutual defection. The interesting thing about this game is that each player could do better. Through mutual cooperation both could arrive at a satisfactory solution. By refusing to rat on the other, each could achieve +10, representing the relative good fortune of a ninety-day vagrancy sentence. This is both players’ second choice (its payoff is the second highest possible), while the saddlepoint is in fact both players’ third choice! Certainly, both players prefer +10 to −10. The dilemma is that by choosing what appears to be a safer course and one that maximizes individual utility, each player contributes to an outcome relatively less satisfactory for both than is possible to attain. Why are mutual cooperation and mutual gain so difficult to achieve? First, to achieve mutual cooperation each player has to renounce the higher payoff he can attain by unilateral defection. Although there are rewards for mutual cooperation in the game of prisoners’ dilemma, unilateral noncooperation gives the highest reward. Second, the players have no advance knowledge of the other’s choice; they cannot communicate prior to their decision. This makes it extremely difficult to forge a mutual strategy of cooperation. Third, trust, an important ingredient for mutual cooperation, is presumed to be absent. Thus, in the absence of trust and in the absence of communication that might generate such trust, each party is compelled to act in a manner injurious to both. Finally, the ability to cooperate depends on the number of times the game will be played. In the classic example above the two suspects get only one chance to play. However, in an iterated game, in which play is continuous, competitors are able to learn from each other’s moves. If the game is only played once, players don’t have to worry about which option their rivals will choose the next time. But the knowledge that numerous interactions will appear between them in the future will quickly make them realize that neither has anything to gain

The Dyadic Level of Analysis, Part III

325

from continued noncooperation. Realizing that the choices one makes today will influence one’s rival’s moves tomorrow, players learn to signal their willingness to cooperate through their moves. A cooperative choice in the present may induce reciprocal cooperation in the future. 8 Prisoners’ Dilemma Games Applied to International Relations Prisoners’ dilemma games are widely applicable to all sorts of situations in international relations. For instance, the prisoners’ dilemma might be applied to the prewar mobilization of armies, to the formation of alliances, to decisions to intervene in internal conflicts, or the pursuit of free trade versus protectionist policies. The structure of these situations is analogous to prisoners’ dilemma games in that they are all situations where mutual restraint is preferable, but where the best policy is to forgo restraint and avoid being a sucker if others refuse to restrain themselves. Robert Axelrod identifies another interesting prisoners’ dilemma example. He claims that soldiers in World War I developed a tacit live-and-let-live policy in the trenches in Western Europe. Army battalions avoided onslaughts against the enemy under certain circumstances in return for mutual restraint. The important point for Axelrod is that in trench warfare small units faced each other in immobile sectors for extended periods of time. This changed the game from a one move prisoners’ dilemma in which defection is the dominant choice, to an iterated prisoners’ dilemma in which conditional strategies are possible. 9

The result of sustained interaction between units was that mutual cooperation based on reciprocity was possible. Soldiers understood that their actions had indirect consequences: providing discomfort for the enemy was but a roundabout way of providing it for themselves. 10 While artillery barrages against enemy positions would inevitably produce incoming fire, restraint was also likely to be reciprocated. Finally, arms races and arms control would seem to be the classic case of prisoners’ dilemma. Let’s look at figure 9.4. The scenario might go something like this. The leaders of India and Pakistan—two enduring rivals who have fought four wars with each other and who have been involved in many militarized disputes short of war—have gathered their top strategic and political advisers together to discuss recent developments. In each state scientists have determined that a new, highly sophisticated strategic weapon is now possible. The cost, however, will be tremendous. The intelligence community in each country has also discovered that the rival state is also at a similar technological point; they may also decide to build and deploy this new weapon. It is also understood that the decision of each country may be influenced by their evaluation of what their rival may do. Although mutual benefits (slowing the arms race, easing the domestic economic burden associated with that race, and decreasing regional tension) could be derived from mutual restraint, the greatest fear for both would be to scrap the program and find out that the other had begun to deploy it. Thus, any desire to foster mutual cooperation must take into consideration the possibility that one’s opponent might build unilaterally, taking advantage of your naiveté. On the other hand, you could attain military superiority by building while

326

Chapter 9

the other pursued a policy of unilateral restraint. Of course, if both choose to build, no one would be better off: regional tensions would have increased, great expenses would have been incurred, and resources that could have been put to better use would have been squandered in another round of a futile arms race. Unfortunately, the decision to build tends to be impelled by the structure of the situation. Not being able to trust the other side, you build—a choice that minimizes the possibility of great loss and has the added benefit of giving you military superiority if the other side exercises self-restraint. The pursuit of individual interests leads, however, to mutual disaster. Arming is the dominant strategy for both sides. Luckily, arms acquisition processes are not constrained by the formal limits placed on prisoners’ dilemma games. Real-world players may communicate with each other with words and deeds; this greatly improves their understanding of the future moves of their rivals. Second, arms races are not one-shot games. Instead, they are made up of a series of many decisions over many years—repeated decisions about testing, research and development, and deployment of many different weapons systems and their components. In a sense, it is a continuous game with an infinite number of interactions; in other words, it is an iterated game. Knowing that the “game” will continue indefinitely has a significant impact on the way in which players approach the game. 11 In such situations both parties are aware of their adversary’s past behavior and understand that their own present behavior will affect the future behavior of their rival. This makes it possible for players (states) to learn to cooperate. The question is, of course, what strategy is best for attaining mutual cooperation, and how can you prevent yourself from being taken advantage of if your rival is aware of your desire for cooperation? Solving the Prisoners’ Dilemma: TIT-FOR-TAT Robert Axelrod devised a unique way to test various strategies for playing iterated prisoners’ dilemma games. He decided to hold a computer tournament. Game theorists from many countries and from a variety of academic disciplines were invited to submit “decision rules”— computer programs that embody a strategy or rule that determines which move to make in each successive turn of the game. Each entry was paired against every other entry (as well as with itself and with a program representing a random series of choices). The winner would be the decision rule with the most total points accumulated in all the matches. 12

Figure 9.4. Game Matrix: Arms Control

The Dyadic Level of Analysis, Part III

327

Two computer tournaments were held. The same program won both tournaments, TITFOR-TAT, a rule submitted by Anatol Rapoport. TIT-FOR-TAT (TFT) is a very simple rule. Its first move is to cooperate, and thereafter it does exactly what the other player did on the previous round. It rewards cooperative behavior and punishes antagonistic behavior. What makes TFT so successful? Axelrod has suggested that TFT and the other successful tournament rules had several similar properties. 13 First, it paid to be nice. A nice rule is one that will never be the first to defect. Nice rules did very well in both tournaments. Second, the best rules were also provocable. That is, they immediately punished rivals after an uncalled-for defection. Provocable rules immediately defect after the defection of an opponent. If a rule does not retaliate in this manner, it will frequently be taken advantage of by rules that are not nice. It pays to be nice, but it doesn’t pay to be too nice. Third, the best rules were also forgiving—defined as the propensity of the rule to cooperate in the moves after the other defected. If the other program resumes cooperation after its initial defection, TFT is willing to let bygones be bygones and will immediately resume cooperation in response. In other words, it engages in short-term punishment, but is willing to return to cooperation if this is signaled by the opponent. TIT-FOR-TAT represents a perfect illustration of the concepts of niceness, forgiveness, and provocability. “It is never the first to defect, it forgives an isolated defection after a single response, but it is always provoked by a defection no matter how good the interaction has been so far.” 14 Fourth, TFT benefits from its own simplicity and clarity. It is easily recognized by its opponents. It is usually obvious early in the encounter that TFT is offering cooperation in return for reciprocity, but will also reciprocate defection. Since TFT’s strategy is clear, it is evident to the opponent that the best way to respond to TFT is with mutual cooperation. TITFOR-TAT represents a good strategy for dealing with iterated prisoners’ dilemmas because, as Axelrod states, “Its niceness prevents it from getting into unnecessary trouble. Its provocability discourages the other side from persisting whenever defection is tried. Its forgiveness helps restore mutual cooperation. And its clarity makes it intelligible to the other player, thereby eliciting long-term cooperation.” 15 It might be added that not only has TFT produced good results when subjected to computer simulation, but game simulation experiments with human subjects have also indicated that the TFT strategy elicits a good deal of cooperation from rival competitors. 16 Perhaps most important, analyses of adversarial relationships between states in the international arena suggest that reciprocating strategies generally produce better results than alternative strategies—especially if the states are evenly matched. For instance, Leng and Wheeler find that a reciprocating strategy is especially effective against opponents who use a bullying strategy. 17 Although TFT is generally recognized as the single best strategy for playing iterated games of prisoners’ dilemma, 18 it is probably applicable as well to iterated chicken games. For instance, Kenneth Oye notes that a strategy of reciprocity in such situations can offset the perverse effects of no-swerve strategies designed to establish reputations for toughness. 19 Of course, TFT is not a perfect strategy. Perhaps the single most important problem with TFT is that it may prove ineffective in getting reciprocal cooperation started in the midst of a longstanding dispute. In the real world, most relationships are in medias res; they are ongoing.

328

Chapter 9

If the other side has just undertaken a noncooperative action, TFT calls for reciprocal noncooperation. The two sides may then become locked in to repetitive competition. Strategists playing TFT must then wait for the adversary to initiate cooperation. In other words, TFT may not be forgiving enough to induce cooperation. 20 It is possible to modify TFT somewhat to reduce this propensity for locking in to reciprocal hostility. If we assume that players are not faced with the two extreme choices of cooperate or defect, but with gradations of each, players might have a continuum of choices, ranging from total cooperation to total defection. A strategy sometimes called tFT (small tit-for-tat) responds to the opponent’s defection by retaliation one degree less than the opponent’s original move. In fact, this modified strategy won a computer tournament constructed by Theodore To, with TFT coming in second. 21 A TFT strategy permits opponents to get away with a little more defection, but it prevents the escalation of punishment that leads to a locking in of mutually competitive moves. It is therefore better able to restore mutual cooperation when the interactions begin with mutual noncooperation. Other researchers also find that when an adversary attempts exploitation, the most effective method of inducing cooperation is the use of mild retaliation rather than strong retaliation or no retaliation. 22 These experimental findings lend support to Axelrod’s recommendation to reciprocate “only nine-tenths of a tit for a tat.” 23 There are other problems with a TFT strategy. Roy Behr points out that TFT may be an inappropriate strategy for some goals. For instance, the goal in Axelrod’s tournaments was to score as many points as possible; TFT did this admirably. But if the goal is changed to winning against as many opponents as possible, TFT fails miserably. In a tournament of this type staged by Behr, TFT was unable to win over any opposing strategy. 24 The winners were FELD and JOSS, two strategies that were not nice—they attempted to exploit the opposition through unprovoked defections a certain percent of the time. Clearly, in order to score more points than the opponent, one must be willing to defect more than the opponent; TFT can therefore never win in this sense. (And, of course, a strategy that defects in every round can never lose in this sense.) Behr’s point is that there are times when defeating an opponent is more important than maximizing one’s own payoff, and TFT is inappropriate for these situations. The success of a TFT strategy is also affected by the values in the payoff matrix. At a minimum, for TFT to work the payoffs to each side must be greater when they cooperate than when they both choose noncooperation (defect), and they must foresee further interactions in the future. 25 If the penalty (or perceived penalty) for mutual defection is reduced, then we could expect that winners in computer tournaments as well as in real life would be less likely to initiate cooperation, less likely to respond to cooperation with cooperation, and more likely to respond to the opponent’s defections. Stephen Van Evera illustrates this point by showing how misperceptions among European leaders in the 1914 crisis twisted the payoff structures so perversely that a TFT strategy would not have worked to prevent World War I. 26 Generally, beliefs of European leaders—that bullying strategies were effective in coercing others to back down, that rapid mobilizations and offensive military strategies were necessary, that war confers positive benefits, along with the perceptions of the hostility of others—resulted in a payoff matrix unlikely to induce cooperation. Specifically,

The Dyadic Level of Analysis, Part III

329

a. by enlarging the rewards of defect/cooperate outcomes and the penalties of a cooperate/ defect outcome, the attractiveness of opportunistic defection and defensive defection were increased; b. by narrowing the difference between cooperate/cooperate and defect/defect outcomes, the attractiveness of cooperate/cooperate was diminished; c. by raising fear that cooperation would be answered by defection, a greater incentive was given to defection; and d. by raising hopes that defection would be answered by cooperation, they made defection more attractive. 27 Under such conditions, the price of using a nice strategy was too high, and it would probably have been incapable of eliciting cooperative behavior from the opponent as well. TFT and GRIT Compared It might be argued that the attributes that make TFT a success—niceness, provocability, forgiveness, and clarity—also describe the strategy of GRIT, which we discussed in the previous chapter. In fact, TFT and GRIT constitute similar strategies for dealing with prisoners’ dilemma situations. Both are designed to induce cooperative behavior through the communication of conciliatory intentions. Both combine a carrot-and-stick approach to reward cooperation and punish defection. There are, however, several important differences between them. First, while TFT relies on nonverbal communication, GRIT has the advantage of being able to use explicit communication in attempting to induce cooperation. Second, aside from the cooperative cue that it gives on the first move, TFT lets the other party take the initiative and simply follows suit. However, players using the GRIT strategy take the initiative themselves. Leng and Wheeler found that the most successful strategies used by international rivals were not ones of pure reciprocity, such as TFT, but ones that included unilateral initiatives of conciliation. 28 Third, both TFT and GRIT retaliate in response to defection. However, GRIT players always take the lead in returning to cooperation, while the TFT strategy requires defection as long as the opponent defects, leading to the possibility of a prolonged retaliatory spiral. Fourth, in GRIT the initial reciprocation is not expected immediately. The initiator must be prepared to risk some exploitation without retaliating. TFT, on the other hand, lets no defection go without retaliation. Fifth, the GRIT strategy is basically for games involving sequential choice by rivals. TFT is usually seen as applicable to games involving simultaneous choice. GRIT thus probably reflects more precisely the real world of political decision making. Several tests of the comparative effectiveness of TFT and GRIT have been performed. In experimental simulations with human subjects, GRIT has proved to be more effective in eliciting cooperation that TFT. The ability of GRIT players to use explicit verbal communication accounts for much of GRIT’s superiority over the TFT strategy in such experiments. 29 Goldstein and Freeman performed computer simulations on four decades of Soviet-American interaction data to test the relative effectiveness of TFT and three versions of GRIT in induc-

330

Chapter 9

ing reciprocal cooperation. They concluded that while all four strategies were somewhat successful, TFT was the least effective. Progressive GRIT (PGRIT) and extended GRIT (EGRIT) were the most successful. 30 The potency of these strategies was due to their willingness to persist in continued unilateral concessions without reciprocation as a way of inducing mutual cooperation. Since the TFT strategy calls for only one such unilateral initiative, it was less able to overcome the reluctance of the target to cooperate. Critique of Game Theory Game theory does have a few weaknesses. First, one might dispute the assumptions on which the theory is based. As we have seen from the literature on decision making, it is questionable whether policy makers are capable of engaging in rational decision processes all the time. It is also questionable whether meaningful values can be placed on alternative sets of goals and whether hypothetical outcomes can be assigned precise numerical weights according to their desirability. It is especially difficult to try to guess the opponent’s evaluation of the expected outcomes, given the general difficulty of reading others’ minds as well as the common problems of misperception of the opponent’s goals and strategies. Second, and most important, it is debatable whether game theory can be said to actually explain why certain events and behaviors occur—whether it is actually capable of explaining the cause of behavior. For instance, it is probably stretching game theory too far to say that arms races occur because the structure of the situation two states find themselves in is a prisoners’ dilemma structure. Likewise, it would be difficult to say that wars are caused because of the chicken-like situations that may precede them. Perhaps game theory is best seen as a heuristic device rather than an empirical theory. Heuristic devices are merely analytical methods that provide us with insights and analogies that help us to understand why things happen, but do not actually point to causal relationships between variables. 31 Even in this limited role, however, game theory can be a useful guide to policy makers, helping them to clarify their thoughts about possible alternative courses of action and assisting them in developing optimal strategies for attaining their objectives. BARGAINING THEORY Games generally help us to understand how the structure of dyadic interactions might have an effect on state choice; moreover, an analysis of game matrices generally helps us to understand how leaders are confronted with various incentives or disincentives toward the use of force. While the prisoners’ dilemma game gives us some insight into why cooperation is difficult, and the chicken game helps us to understand international crisis confrontations, these heuristic devices still do not move us very far down the road toward a theory of war. James Fearon is responsible for a basic insight that takes us a further step in that direction. His famous article “Rationalist Explanations for War” has provided the groundwork for what has become known as the bargaining theory of war. 32 Drawing on an insight from one of the early giants of game theory, Thomas Schelling, that “most conflict situations are essentially bargaining situations,” Fearon starts from the proposition that decisions for war are interactive choices made in the context of bargaining between states over outcomes desired by each. 33 For

The Dyadic Level of Analysis, Part III

331

instance, one can imagine a two-party game matrix with several options, one of which is war. Fearon argues that, from a rational perspective, war is not only a risky alternative but a costly and comparatively inefficient way to resolve disputes. In principle, there should be a negotiated settlement or a bargain short of war that both sides will prefer to the gamble of war. This will be true for both the stronger state and the weaker state. Both states know that the bargaining leverage they each have and the basic nature of any negotiated settlement will depend on the likely outcome of a potential war and the costs associated with winning it. If the two states could agree on the likely outcome of the impending war, they could avoid it. (This draws on Geoffrey Blainey’s insight that wars are often caused by disagreements over the relative levels of the two states’ strength and resolve.) 34 Logically, both parties should prefer to reach a prewar deal that would reflect the hypothetical postwar settlement—instead of fighting the war, bearing its costs, and then attaining the exact same result. If both states accurately recognize their relative power and resolve, the result of most disputes should be that weaker states make concessions to stronger states and war is avoided. But, of course, this doesn’t always happen. Bargaining theory sees war as a failure of states to reach a negotiated settlement prior to war. The basic question is “what prevents leaders from reaching ex ante (prewar) bargains that would avoid the costs and risks of fighting?” 35 It could be individual-level phenomena having to do with the personality of particular leaders, their cognitive biases, their emotions, their misperceptions, or simply that they operate in an environment of bounded rationality. Alternatively, it could have to do with substate factors such as the power of domestic political interests or the pathologies of the decision-making process. Fearon does not dismiss these lines of explanation out of hand, but neither type of explanation is a rationalist explanation. (Rational choice theorists and game theorists, as well as neorealists, assume that states are unitary rational actors.) Fearon argues that rational leaders may be unable to agree on a mutually negotiated settlement preferable to war if one of three conditions is present. Fearon’s first condition is the existence of private information held by one side but unknown to the other about factors that might affect the outcome of a potential war. This might include information about one’s weapons or capabilities, tactics and strategies, intentions, one’s willingness to use force, or the commitment of allies to lend assistance. If such private information existed, then the two sides might have different estimates about the outcome of a potential military conflict between them, and would thus see their prewar bargaining leverage differently. Just as important, states have logical incentives to misrepresent this kind of information or keep it secret. Fearon takes the position that two rational leaders with the same information about an uncertain event (like the outcome of war) should come to the same conclusion about its likely outcome. “Conflicting estimates should occur only if the agents have different (and so necessarily private) information” about factors that might affect the course of the war or the opponent’s willingness to fight. 36 To avoid the costs and the risks of war, leaders have logical incentives to share such private information by signaling through words and deeds to their opponents, but they usually don’t. This is because although states have an incentive to find a bargain that avoids war, they also have incentives to win the best possible deal. They conceal

332

Chapter 9

or misrepresent their capabilities and their resolve in order to induce better offers and to deter attack, and this leads to miscalculations on both sides. An example used by Fearon is the 1904 Russo-Japanese War that developed over the two countries’ competing interests in Manchuria and the Far East. 37 Russian leaders, believing in Russian military superiority, refused to compromise since they believed Japan would be foolish to attack. The Japanese, on the other hand, due to their superior intelligence service, possessed private information about Russian weaknesses in the Far East, about the positive effects of military reforms and training on the relatively untested Japanese military, about the nature of an effective offensive strategy that could be used against Russian positions in Asia, and therefore on the balance of capabilities between the two. Some (though not all) Japanese military leaders therefore estimated that the chances of success were fifty-fifty. Fearon argues that sharing their private knowledge of the dyadic military situation with the Russians in order to enhance the chances of successful prewar deal would be unlikely to benefit the Japanese: The Russians would probably not have believed them, and revealing their military strategies would reduce the likelihood that they could actually prevail. Fearon’s second condition is the presence of a commitment problem. This problem occurs when there is a perceived shift in the balance of capabilities between the two states: one state is declining relative to another. Declining states will fear that any settlement negotiated with a rising state will not be honored in the future. The rising state has a logical incentive to renege on the terms of an agreement. It will be unlikely to stick to an agreement based on the old balance of capabilities and is apt to make new demands when it has achieved a greater level of relative power. Thus, declining states believe they cannot trust a deal offered by rising states. (It is here that the situation of global anarchy plays a role: In the absence of a global actor capable of enforcing a negotiated settlement, the declining power cannot trust that the rising power will honor its word once the dyadic balance of capabilities has been altered.) Fearon argues that commitment problems might be behind both preemptive wars and preventive wars, though he argues that the latter is most probable. 38 By their very nature preventive war scenarios are dynamic—they occur when dyadic power balances are perceived to be changing. As we will see in chapter 11, many realist theories link dyadic power transitions to war. In these theories declining states are presumed to be motivated by their fear that rising challengers will attack them once they have achieved relative superiority. Declining (but still dominant) states are motivated to attack rising challengers while they still have a chance to defeat them. Fearon argues instead that the declining state attacks because “it fears the peace it will have to accept after the rival has grown stronger.” 39 This is not a situation where leaders disagree about their relative power or each other’s motivations. Moreover, they would each prefer to avoid the costs of war, but the trust necessary for a prewar agreement is lacking because the situation gives one party an incentive to renege on the agreement. Fearon’s third (and least important) condition is the presence of indivisible issues that are not easily amenable to compromise. Typically these are issues that might involve issues of ethnicity, or religion, or issues involving political or moral principles. The most fundamental of these issues might involve existential threats to the sovereignty of a state (where the issue is whether the state will actually continue to exist or not). Territorial issues are usually not regarded as indivisible; historically, geographic boundaries have been subject to a rather

The Dyadic Level of Analysis, Part III

333

infinite progression of divisions, accumulations, partitions, and repartitions. However, it is possible that even territorial issues may take on social, psychological, and political aspects of indivisibility. 40 Robert Powell constructs a formal model that shows that the presence of indivisible issues does not explain the outbreak of war. 41 He argues that even if an issue is indivisible, the fact that war is costly ensures that there will still be a range of agreements that the two players will prefer to fighting. The real problem is that players can’t commit themselves to abide by such agreements. What looks like bargaining indivisibility is really a commitment problem in disguise. In Powell’s view, preventive wars, preemptive wars, and a third type of war, caused by making concessions that improve the opponent’s capabilities, are all closely related. They all involve situations in which the dyadic balance of capabilities changes and thereby brings about commitment problems. David Lake’s bargaining analysis of the Iraq War (2003) illustrates some of the difficulties of using the bargaining model to explain wars. 42 The core question is that since war between the U.S. and Iraq would be costly for both, why was it not possible to reach a negotiated settlement that would leave both sides better off than if war was used. Lake argues that the fundamental issue at stake was not the existence of WMDs in Iraq, but more generally the geostrategic issue of which country would dominate the Persian Gulf region. He establishes that prior to the war, both sides understood that war would create substantial costs, though the estimates on the U.S. side were wildly underestimated. Nevertheless, since both sides would suffer costs, a bargain should have been possible. 43 Lake suggests the U.S. and Iraq could have compromised over the nature of the order imposed on the region, reaching an accommodation that reflected their relative strengths, or they could have divided leadership by issue area . . . At the more proximate level of regime change, Saddam could have fled Iraq and sought exile in some safe haven, perhaps extorting a substantial payment from the United States for doing so. 44

Even if the Bush administration was determined to use force to establish democracy in Iraq or to send a message to potential opponents about U.S. willingness to use force to protect its interests, Saddam’s best response should have been to capitulate and thereby avoid war. Iraqi concessions would have deprived the Bush administration of a casus belli, and would have left both sides better off than if war had taken place. Bargaining theory would suggest that either a commitment problem or the existence of private information might have prevented a deal short of war. Lake argues that a primary U.S. concern was not Iraq’s current WMD program but its future capabilities and intentions, and this created a problem of credible commitment. Without Iraqi WMDs, the U.S. would certainly win a war at relatively low cost. With Iraqi WMDs (in the future), the U.S. would still win, but the costs would be higher and the changed military balance would logically affect any bargain the two might reach in the future: “the United States could not make the same demands on Iraq, and Iraq would not need to concede as much.” 45 Thus, the U.S. had an incentive to use force earlier rather than later. Another aspect of the commitment problem was that if the U.S. could not trust Saddam in the future to carry out any negotiated settlement, then removing him from power had to be part

334

Chapter 9

of the solution. Uncertainty about Hussein’s willingness to keep a bargain meant that regime change was deemed a necessary part of the bargain from the American perspective. However, bargaining theory itself is inadequate in explaining how the Bush administration arrived at this judgment. Presumably the commitment problem was a constant; it applied to the earlier Clinton administration as well as to other rogue states like North Korea. Ultimately, Lake argues that the Bush administration’s extreme perceptions of threat from Iraq and their perceptions of the regime as uniquely evil and untrustworthy depended on prior beliefs that they had brought with them into office—beliefs about Saddam’s personal attributes that are outside of the scope of bargaining theory. Such cognitive attributes and biases do not fit well with the rational assumptions of bargaining theory. What about the view from Saddam Hussein’s perspective? Bargaining theory suggests that Iraq, given its relatively weak position and the high cost of war, should have done everything possible to signal to the U.S. its good intentions—especially by making it unambiguously clear that it had dismantled its WMD program. We now know that it failed to do this because it was actually playing not a two-player game with the U.S., but an n-player strategic game that also involved internal Iraqi opponents and Iran. Based on postwar interviews with Iraqi leaders and on documents captured during the war, it now appears that Saddam Hussein’s top priority was protecting the regime from internal opponents and from external threats from Iran. Since the possession of WMDs might be useful against these foes, Saddam needed to maintain at least the hint that Iraq still possessed these weapons. The downside of this strategy was that the lack of a clear signal to the U.S. of the absence of WMDs merely strengthened the administration’s preconceived notions that such weapons did in fact exist and their numbers were being expanded. Lake suggests that bargaining theory needs to develop n-player bargaining models. 46 Iraq and the U.S. had different views concerning the likely outcome of a military conflict, but—contrary to bargaining theory—this had little to do with incentives to hide or misrepresent private information. Saddam incorrectly perceived that the U.S. lacked resolve to use force in situations that involved taking significant casualties. (The U.S. had withdrawn forces from Lebanon in 1984 and Somalia in 1993 and had ended the Gulf War in 1991 without driving for Baghdad.) Iraqi leaders also believed that while coalition forces might invade Iraq, they would not reach the heavily fortified capital. 47 Thus, Saddam bargained harder and was willing to risk war rather than reveal the absence of WMDs. As Lake explains, this was not the product of any American misrepresentation of its war plans (which were uncommonly transparent) or its failure to signal its resolve; the American threat should have been credible. In Lake’s view, the fault was rather with the fact that Saddam had surrounded himself with toadies who resisted providing him with facts and analysis that he did not want to hear. The “information asymmetry” was due to self-delusion by the Iraqi leader. Of course, the American side suffered from similar misconceptions about the likely costs of war that were also self-delusional. Contrary to the expectations of bargaining theory, the administration never really updated its evaluations of the evidence of threat from Iraq and of the potential costs of war during the long prewar phase. Most important, the high costs of U.S. postwar nation-building activities in Iraq were not factored into the estimates of the war itself—largely because they were believed to be virtually zero. Acceptance of the illusion of

The Dyadic Level of Analysis, Part III

335

extremely low war costs meant that the administration would accept no bargain short of complete Iraqi capitulation. These combined illusory beliefs played a major role in preventing the two sides from achieving an effective bargain short of war. While imperfect and asymmetrical information due to incentives to misrepresent private information is consistent with bargaining theory, information failures due to self-deluding policies stretch the limits of rationality and point to cognitive biases. This is a direct challenge to bargaining theory (and to other rationalist theories such as deterrence theory, as we will soon see). As Lake concludes, “it is hard not to delve into this case without becoming acutely aware of the less than fully rational nature of decisionmaking.” 48 Lake suggests that bargaining theory needs to take account of the literature on cognitive decision making. Cognitive factors can help bargaining theory explain why estimates of costs and resolve and potential bargains become distorted in ways that increase the probability of war. Finally, we should say that interesting developments continue to take place in bargaining theory. While a detailed explanation is beyond the scope of the present analysis, it should be noted that recent bargaining models have attempted to create a single framework that links the outbreak of war, its prosecution on the battlefield, its termination, and its consequences—all as part of a single process. 49 In these models, war (as Carl von Clausewitz has said) is simply the continuation of political bargaining by other means. War is not seen as a failure of diplomacy, but as a continuation of it: it is treated as part of an ongoing bargaining process. The primary role of the actual military combat is to provide further information about the opponent’s capabilities and resolve, thereby reducing uncertainties. Once this occurs, a more accurate picture of the true bargaining range is revealed. As a consequence, as the war continues, the expectations of the two belligerents should converge toward a settlement that will end the war. 50 In sum: Fighting breaks out when two sides cannot reach a bargain that both prefer to war. Each side fights to improve its chances of getting a desirable settlement of the disputed issue. The war ends when the two sides strike a bargain that both prefer to continuing the war, and the outcome is literally the bargain struck. Finally, the duration of peace following the war reflects the willingness of both sides not to break the war-ending bargain. 51

Bargaining Theory: A Short Critique Since bargaining theory is based on rationalist principles, it is fundamentally challenged by theories and empirical evidence that suggest decisions lack rationality. Reiter notes three general challenges. 52 The first comes from cognitive psychology. By now we are familiar with the idea of man as an emotional, intuitive decision maker, relying on heuristic shortcuts, and with images of the world that seem impervious to change in the face of new and contradictory information. We are also well aware of the effects of personality and motivated biases. The result of these factors is twofold: Actors may not be able to use new information to update their beliefs about opponents, and cognitive biases might make war seem more appealing an option than it should be when leaders face unfavorable military conditions. The second challenge comes from the constructivist approach, which treats war as a social construction shaped by norms and culture. From this perspective, leaders may choose war not

336

Chapter 9

because of rational choice based on cost-benefit analysis, but because war plays an important role in the formation of group identities. Identities of states and nations are formed in interaction with (and opposition to) others. The relationship between self and others is frequently one of conflict. The benefits derived from war are social, cultural, and psychological rather than material; they have to do with how war contributes to national identity, and these are factors quite foreign to considerations of bargaining theory. We will address the constructivist approach in chapter 12. The third challenge comes from those who emphasize the central role of domestic politics in international conflict. For instance, it might be argued that according to diversionary war theory leaders might prefer fighting an external war to negotiating a settlement with a rival, regardless of the costs, in order to maintain themselves in power at home. A fourth challenge is related to this. Most bargaining theories of war assume states are unitary rational actors. But this is only a simplifying assumption and reality is, of course, more complex. The influence of multiple interest groups on the decision process may skew governmental choices; the question is simply how much damage does this do to the explanatory value of the theory. One attempt to deal with this added layer of complexity is Robert Putnam’s two-level bargaining theory, to which we now turn. Two-Level Bargaining Fearon’s analysis suggests that wars are not likely to result from situations in which the countries’ preferences are so distinct that there is no mutually acceptable settlement that would prevent war. The real culprits are the existence of private information or commitment problems. Another approach—Robert Putnam’s two-level game theory—offers a different explanation for why such substantive agreements are difficult to find. 53 Virtually all negotiators involved in international negotiation—whether they deal with arms control, postwar peace settlements, international trade, or boundary disputes—tell the same story: They spend as much time and effort negotiating with members of their own team as they do negotiators from the opposing states. Chief negotiators—presidents, prime ministers, secretaries of state, or ambassadors—are involved in bargaining at two levels. In simple dyadic negotiations, level I bargaining takes place between the representatives of the two states, with the presumed goal of an interstate agreement. Level II bargaining takes place in discussions among representatives of relevant domestic political groups within each state. (In arms control negotiations, these may include personnel who represent the chief executive and his staff, the defense ministry, the various armed services, the foreign ministry, arms control specialists, legislators, and various relevant political actors.) Bargaining at level I and level II is generally a simultaneous and an iterative process. The important point for Putnam is that all decisions at level I must be “ratified” by level II participants in both states; without such acceptance, there will be no deal. Whether an agreement is possible depends in part on the nature of the win-sets held by the two sides. A win-set is the set of all possible level I agreements that would be acceptable (gain a majority for ratification) at each country’s level II. In order for an agreement to be possible, the level II win-sets of the two countries must overlap: Any agreement must fall within the level II win-sets of both sides. The important point about the conceptualization of level II is

The Dyadic Level of Analysis, Part III

337

that Putnam, unlike many rationalist theorists, recognizes that states are not really unitary rational actors. A number of important principles can be deduced from this framework. For instance: 1. The larger each country’s win-set, the greater the chance of overlap and thus of a successful agreement. 2. The larger a country’s win-set, the more it can be pushed around by other level I negotiators. 3. The smaller the country’s domestic win-set, the greater its advantages at level I. 4. The larger the number of people that must be satisfied at level II, the smaller the winset, and the greater bargaining power the country has at level I. 5. The smaller the win-sets, the greater the risk of “involuntary defection” due to the inability of negotiators at level I to deliver the agreement at level II. 6. The lower the perceived cost of no agreement, the smaller a side’s win-set and the harder it can bargain. 7. The domestic constituencies whose interests are most affected by an agreement will exert special influence at level II, and the group with the greatest interest is also likely to hold the most extreme position. 8. The politicization of an issue is most likely to activate groups less worried about the costs of no agreement. This reduces the size of the win-set. 9. Level I negotiators are usually badly informed about level II politics in the opposing state. This creates uncertainty about the size of the win-set, which can be a bargaining tool. 10. International pressures may reverberate within domestic politics. Reverberation effects can be positive (facilitating agreement) or negative (inhibiting agreement). At bottom, wars occur when the win-sets of countries do not overlap. Putnam notes that war became inevitable between Britain and Argentina once it became clear that their win-sets regarding the Falklands/Malvinas did not overlap. 54 (As we have seen in chapter 6, the military regime in Argentina also had diversionary motives for war.) Putnam’s analysis is essentially one that is rooted in the interactions between domestic politics and international politics. The crucial factor—the size of each country’s win-set—depends largely on domestic political factors. How might the bargaining approach be applied to situations of war and peace? A two-level analysis of the U.S.-Japanese war in the Pacific would make the following observations about the long negotiations and coercive diplomacy that ended in the Japanese attack on Pearl Harbor. 55 First, the Japanese leadership ultimately realized that U.S. and Japanese win-sets did not overlap. Three sets of issues were involved in the level I negotiations: Japanese membership in the Axis Pact, Japanese expansion into Southeast Asia, and Japanese occupation of China and Manchuria. For Japan, the first two were negotiable, the last was not. Japan had spent too much blood and treasure in the occupation of Manchuria and China. On the other hand, for the U.S. the essential core of any deal with Japan was the inclusion of a Japanese withdrawal from China (after all, Chiang Kai-shek’s Nationalist government was a U.S. ally).

338

Chapter 9

Second, the level II participants for the U.S. involved not only domestic political actors in the United States, but U.S. allies (the British, the Dutch, and the Chinese) who pushed for hawkish American positions. Third, level II agreement in Japan required—at a minimum— acceptance by the imperial army, the navy, and the emperor. Fourth, in Japan, the primary level II constituency was the imperial army, which—having the most at stake—took the most extreme position. The army was unalterably opposed to sacrificing the gains made by Japanese force of arms in China and Manchuria. Fifth, the Japanese were afraid of a no-deal ending, which would clearly mean war. But they were even more afraid of a deal that would lock Japan into a position of inferiority and humiliation; this would have been untenable for domestic political reasons. We can also find evidence of Fearon’s commitment problem here. Japanese leaders were unlikely to believe that the United States would keep an agreement for very long. Since it was clear to all that U.S. military power in the Pacific would only continue to increase rapidly relative to that of Japan, the American side would have substantial incentives to renege on a (modest) deal made in late 1941 or early 1942. As its power increased, it would demand more. Since U.S. leaders could not be trusted to abide by a settlement, the Japanese response was to try to negotiate the strongest possible agreement, thus reducing the potential for a mutual winset. Finally, private information played a role, but not perhaps in the way that Fearon’s theory might predict. One example was Japan’s secretly held ability to effectively use air-launched torpedoes in shallow waters (as in Pearl Harbor). The best example of private information, however, involved Japanese war plans, specifically the decision to initiate the war by attacking Pearl Harbor. This information was withheld from the U.S. for obvious reasons, but the real effect of private information was inside Japan itself. The Pearl Harbor portion of the plan was a closely held secret within the Japanese government. Only top-level planners within the imperial navy knew of the plan until the final days. Within the imperial army, few knew of the secret except the chief of staff and his deputy. Even the army and navy cabinet ministers were probably in the dark until December 1. Civilian cabinet ministers, including the foreign minister, were not told of the plan until the task force had left its rendezvous point. Had the circle of level II leaders aware of the plan been widened, there may have been less support for the decision to end negotiations and initiate war. The attempted destruction of the U.S. fleet at Pearl Harbor would likely ensure a maximum military response from the U.S. and reduce American willingness to negotiate a moderate settlement. In other words, the existence of private information had an effect on the level II bargaining within Japan. It made it more likely that the relevant political elites would approve the plan. Not only did the Japanese military officials have an incentive to withhold information about war plans from the U.S., they also had an incentive to withhold the same information from their domestic rivals within the decision-making unit.

The Dyadic Level of Analysis, Part III

339

DETERRENCE THEORY Deterrence: Theory and Concepts The conflict spiral theory and the GRIT and TFT strategies all suggest that the way to bring about peace is through a willingness to initiate cooperation and to persevere in this. An alternative theory has previously been mentioned. It is encapsulated in the Latin phrase si vis pacem, para bellum: If you seek peace, prepare for war. This is the essence of deterrence theory. The concept of deterrence refers to one’s ability to prevent another person or country from doing something that is undesirable or harmful to one’s own interests. 56 For instance, you don’t want another person to steal your property and you don’t want another country to attack you or your allies. There are typically two ways to prevent this from happening—defense and deterrence. 57 Let’s consider defense first. Suppose you own a collection of rare jewels. To prevent them from being stolen you might consider hiring some muscular Rambo types with heavy artillery to protect your goods twenty-four hours a day. Presumably, this would make it physically impossible for any thief to steal your treasure. Your possessions are defended by making it physically difficult, if not impossible, for the opponent to achieve his goal. This is sometimes referred to as a strategy of deterrence by denial. Deterrence works somewhat differently from defense; it emphasizes retaliatory punishment. Your jewels may not be heavily protected, but you let it be known that you have Rambo and Associates on retainer and if anyone messes with your personal property they will hunt down the perpetrators and generally make life miserable for them. Of course, although deterrence and defense are analytically distinct, they are thoroughly related in practice. A good defense/denial capability may have the side effect of enhancing a state’s deterrence. 58 Deterrence as a theory and as a guide for policy makers was developed in the early Cold War era and owes much to the existence of nuclear weapons. The emphasis on deterrence as retaliation or punishment clearly reflects this. As Patrick Morgan suggests, it would hardly have been a conceptual breakthrough to simply say that deterrence meant: “If you attack us, we will fight back.” 59 Deterrence has, from the beginning, meant imposing unacceptable costs on the would-be attacker. Deterrence is based on a psychological relationship between two people or countries. A attempts to manipulate B’s incentives by impressing B with the idea that if B performs certain actions in the future, he will punish B. The threat of retaliation represents a negative incentive; the hope is that it will raise B’s costs well above any prospective gains from the attack. For the deterrent threat to work, B must be made to believe two things: (1) that A has the capability to retaliate effectively, and (2) that A has the will or intention to retaliate as he has threatened. If both capability and will are present, we say that A’s threat has credibility; it is believable. (It goes without saying that it is also necessary that the threatened punishment is something that B actually fears.) If A’s threat is credible, then B should be prevented (deterred) from taking action. If B actually believes that Rambo and Associates will brutalize him in retaliation for his acts against A, it would not be in his best interest to steal A’s property. The expected losses would

340

Chapter 9

exceed any expected gains. In deterrence, unlike defense, the goal is not to kill the enemy when he invades; the goal is to convince him that he will be killed if he does invade. Like game theory, deterrence theory is based on the assumption that players on both sides are rational—although scholars certainly vary on what the term rationality actually means. This emphasis on rationality has meant that many deterrence theorists have used rational choice frameworks or game-theoretic models to analyze deterrence and to provide guidance for policy makers. In such models, states contemplating an attack compare the expected utility of using force to that of refraining from the use of force, and they select the option with the greater expected utility. The practical aspect of the theory focuses on how military threats can reduce a potential attacker’s expected utility for using force. Put very simplistically, deterrence theory argues that peace is kept through maintaining credible threats against one’s opponents. Ultimately, the opponent must believe that under certain fairly clear circumstances you will fight him. These boundaries are often referred to as red lines. Conversely, according to the theory, war occurs because the threat fails to be received credibly. That is, B either doesn’t believe A has the capacity to retaliate effectively, or B doesn’t believe A has the will to actually do what he has threatened, or both. Opponents are presumed to have aggressive intentions and to be waiting for any window of opportunity that might exist due to deficiencies in the defender’s capability or will. 60 Cut to its essential core, deterrence means threatening military retaliation to prevent a military attack, but in reality deterrence theory is applied to a somewhat wider domain than simply the prevention of war. In Patrick Morgan’s estimation, deterrence applies to preventing four things: an attack, the serious consideration of mounting an attack, nasty confrontations such as a major crisis, and the significant escalation of wars that have already begun. 61 Thus, deterrence is fundamentally about things that don’t happen: it’s about non-occurrences. 62 As we will see, this makes it difficult to empirically assess deterrence success and failure: It’s hard to know why things don’t happen. Deterrence should be distinguished from its near cousin compellence. While deterrence uses threats to persuade others not to do something, compellence uses threats to persuade others to change their current behavior and either stop doing something that you dislike or to do something that they are not currently doing. Deterrence is generally perceived to be the easier task: the target does not have to change his behavior, he just has to continue not doing something. Compellence is harder largely because it requires the target to change his behavior, and people tend to be reluctant to change—especially if it requires giving up something they value, and especially under threats from others. As we know from prospect theory, people feel losses more keenly than gains, they desire to hold on to what they have (the status quo) and are willing to accept risks to hold on to it. 63 Compellence and deterrence are difficult to distinguish and often overlap in practice. They are most properly thought of as twin components of the broader category that we call coercive diplomacy—the use of force or threats to attain a desired outcome. The important point is that it matters whether the target perceives that he is being asked to change his behavior (compellence) or whether he is being asked to continue to refrain from something he is not doing at present (deterrence). 64 For instance, do Iranian leaders see themselves as the targets of American-Israeli attempts to deter them from attaining nuclear weapons—something they

The Dyadic Level of Analysis, Part III

341

proclaim not to be doing at present? Or do they see themselves as the targets of attempts to compel them to reverse and dismantle an ongoing peaceful nuclear program that includes the enrichment of uranium? Another essential distinction is between general deterrence and immediate deterrence. General deterrence is in effect 24/7 during normal times. It consists of broad general threats to respond to an attack by opponents that are backed up by a country’s broad military capabilities. The threats may be somewhat indirect, ambiguous, and amorphous and are not limited to any single opponent or contingency; in fact the threats are simply implicit in the state’s overall military capabilities. The attack to be prevented is at this point merely hypothetical. 65 The purpose of general deterrence is to discourage opponents from making a challenge in the first place and “to ensure that thinking about an attack never goes very far, so crises don’t erupt, and militarized disputes don’t appear and grow.” 66 General deterrence is especially relevant for enduring rivals, since these conflict-fraught pairs tend to see wars between themselves as likely—unless prevented by effective threats. Thus, general deterrence is the norm between these states, but from time to time as militarized disputes break out, rivals transition into the realm of immediate deterrence. 67 Immediate deterrence is the kind of deterrence practiced during a crisis in which there is a distinct possibility of war and the challenger appears to be seriously contemplating and preparing for an attack. Here, a state issues threats related to the immediate situation against a specific opponent (the challenger). Morgan argues that general deterrence is more important than immediate deterrence: “The security achieved is greater. Deterrence fails first as general deterrence.” 68 Once general deterrence fails, we are in the much more dangerous environment of immediate deterrence. We can also make a distinction between direct deterrence of attacks on one’s own homeland, and extended deterrence—preventing attacks on the territory of one’s friends and allies (typically referred to in the literature as protégés). Combining these four categories, we get a four-fold typology of deterrence: (1) general direct deterrence, (2) general extended deterrence, (3) immediate direct deterrence, and (4) immediate extended deterrence. Let’s take a look at what can be thought of as a strong or “muscular” variant of deterrence theory. 69 In this view, a state’s military capabilities are largely taken for granted; what is more problematic for the deterrence theorists and practitioners is the presence (or absence) of a country’s will or intention to use force and how this is communicated to potential aggressors. A state’s most important instrument for inhibiting aggression is the threat of military retaliation, but threats must be communicated in ways that make them credible to the opponent. Threats are typically conveyed through words, but words are cheap. Formal treaty agreements are somewhat better, but the best way to communicate one’s deterrent threat is presumed to be through past actions—especially through the execution of threats. Carrying out threats conveys a willingness to make sacrifices and take risks in order to meet commitments. Recent analyses of deterrence efforts by game theorists take a similar approach by focusing on the concept of costly signals. The argument is that the most credible signals are costly to those who make them. Statements or actions that are clearly irreversible or that would impose high costs on their makers if they were to retreat should convey real intent and therefore credibility. When presidents and prime ministers make highly public statements promising to

342

Chapter 9

come to the aid of an ally, they are essentially mortgaging their honor. They leave themselves with little choice; backing down becomes difficult if not impossible (and this, of course, is the idea). Fearon and other bargaining theorists focus on audience costs—costs incurred when political leaders, most especially in democratic states, suffer political punishment at home for reneging on foreign policy commitments. Fearon argues that deterrence success/failure is better explained by costly signals than by the balance of interests or the balance of capabilities between countries. 70 Of course, while costly signals may give leaders a certain amount of traction in enhancing general deterrence, in crisis situations, they may be seen as provocative and increase the risk of actual military conflicts. (And of course they may back leaders into corners into which they would prefer not to be crammed.) The use of costly signals to convey credible threats is very much at odds with Hans Morgenthau’s pragmatic realist dictum that statesmen should never put themselves in a position from which they cannot retreat without losing face. 71 A strong deterrent policy requires just the opposite; it is necessary precisely to put yourself in a position from which you cannot retreat gracefully. It is only then that the opponent will be absolutely sure of your threat. If you can retreat gracefully, then you might not keep your word, and if you might not, then the opponent will believe you will not. 72 Threats become credible to the extent that you can deprive yourself of freedom of choice. If this can be achieved, your opponents will believe you are automatically bound to carry out your commitment. If this sounds familiar, it should. This was exactly the advice offered as a winning tactic in the game of chicken: convince your opponent that you are incapable of choosing to swerve, that you are committed to driving straight ahead even if it means destruction. This leaves the choice in your opponent’s hands, and the rational thing for him to do is to back down. Thus, deterrence strategy is similar to the chicken strategy in important ways. Both require a clear enunciation of a threat, and both require that the threat be conveyed to the opponent in a highly credible manner. (And both may require a willingness to do what appears to be irrational.) Of course, the problem with treating deterrence as a game of chicken is that both sides may make irrevocable commitments and freeze themselves into incompatible stances, producing precisely the disaster each wanted to avoid. In an interesting insight into the relationship of individual-level factors and dyadic-interaction factors, Charles Lockhart has made the observation that hardline or hawkish statesmen are predisposed (perhaps by their operational codes) to see conflicts with rival states as chicken contests that require strong actions to signal one’s commitment and credibility. On the other hand, “softline” or dovish leaders are more predisposed to view these encounters as prisoners’ dilemma games that call for the development of cooperative and conciliatory strategies. 73 One of the central paradigms of deterrence theory is the Munich Crisis of 1938 and the outbreak of World War II a year later. In 1936 Hitler began a remilitarization of the Rhineland in violation of the Versailles Treaty. Britain and France objected, but took no action beyond this. In March 1938 Hitler bullied the Austrian leadership into accepting German military occupation and incorporation into the Third Reich. Once again France and Britain took no action. In the fall of 1938 Hitler pressured the government of Czechoslovakia for the cession of the Sudetenland to Germany. Although the Czechs had defense pacts with both France and Russia, the latter were unwilling to fight Germany over Czechoslovakia, and at the Munich

The Dyadic Level of Analysis, Part III

343

Conference in October Britain agreed that the Czechs must cede the Sudetenland. In March 1939 German troops marched unopposed into the remainder of Czechoslovakia. The British and French refused to take immediate action, but threatened Hitler that if he attacked Poland, they would come to her defense. As everyone knows, Hitler was not dissuaded by the allied threat and Poland was attacked in the fall of 1939. Britain and France declared war and the second great conflagration of the twentieth century began. What are the lessons of Munich for deterrence theory? The argument made by defenders of a strong deterrence policy is that states that permit their opponents to commit aggression are likely to experience even more aggression. 74 The policy of appeasement lessens the coerciveness of a state’s threat. Tolerance of aggression causes several unfortunate effects to occur simultaneously. The opponent feels a sense of “exhilaration”—his first acts of defiance and aggression have been successful and he is tempted to believe that further moves will meet with similar inaction. The warning signals of other states begin to be discounted by the aggressor. Additionally, the fact that aggression has initially been successful will strengthen the hawkish political faction internally against the accommodationists who have cautioned restraint in the past. For all these reasons there is a high probability that aggression by the opponent will increase rather than abate. The toleration of aggression has effects within the appeasing country as well. Appeasement cannot be sustained indefinitely. The cries of shame and cowardice by internal critics of the regime will eventually force a change in either government policy or personnel or both. Hawks will be brought to power as part of the reaction to the failed policies of the past. Eventually, a new policy of “peace through strength” will be implemented and new threats will be issued. Consequently, war becomes much more likely. On the one hand, the aggressor is predisposed (on the basis of past performances) to discount the appeaser’s threats as empty rhetoric. But internal politics in the appeaser have now forced it to take a hardline approach. The aggressor will make one more move and this time it will be met with war. John Vasquez makes a similar point in his steps-to-war approach. The failure of accommodationist policies in response to the aggressive realpolitik policies of other states is likely to result in an internal political shift in which accommodationists are driven from power and replaced by hardliners. And, of course, the fact that hardliners are then in power in both states increases the likelihood of war. 75 Deterrence theory concludes, therefore, that wars occur because a country’s threat has not been credible enough to deter its opponents. The logical solution is that it may be necessary to confront aggression forcefully in its early stages, even if this means war. It may be better to fight a small war now than a larger war later. Additionally, it is probably safer to confront aggression in its early stages because at this point the aggressor is probably uncertain of the opponent’s reaction and is more likely to disengage if confronted. It is frequently speculated, for instance, that Britain and France would have been successful in halting German aggression if they had taken forceful action from the start. You will have noticed that deterrence theory and conflict spiral theory appear to be the products of alternate realities or strange parallel universes. They start from rather different assumptions and make decidedly different predictions about state behavior. Here are a few major differences: (1) Spiral theory assumes that most states are motivated by their fear of

344

Chapter 9

others and act defensively—except perhaps for the rare predatory state. Deterrence theory assumes states are aggressive and motivated by gains; they constantly seek windows of opportunity through which they make take advantage of others. However, in practice many deterrence supporters depart from this by assuming that “we” are defensive, but our opponents are aggressive. (2) Spiral theory argues that states are most likely to respond in a reciprocal manner to the actions of others; cooperation by A leads to cooperation by B. Deterrence theory expects inverse reactions; toughness by A should lead to acquiescence and submission by B. (3) Spiral theory argues that wars occur because strong actions by A cause B to fear for its security, provoking B to even tougher action, which ignites a spiraling of hostility. Thus, even defensive actions aimed only to deter rivals may have the unintended consequence of provoking the rival into military action. On the other hand, deterrence theory argues that wars are due to signs of weakness. The only thing that prevents war is a strong deterrent threat. If A does not act strongly enough to indicate its military capabilities and its will to carry out its commitments, B will not believe A’s threat is credible and will cross a red line. Cooperation and conciliation are simply fancy words for appeasement. (4) Spiral theory implies that in many cases of war, no one is to blame; neither side consciously desires war or plans for it. Deterrence theorists argue that wars are not likely to occur by mistake; states know where the red lines are. They know that trying to change the status quo will bring a military response. If potential challengers are confused about what they can get away with, it is largely the fault of the defender for not making this clear. Extended Deterrence: Problems of Credibility and Reputation Some forms of deterrence are simpler than others and provide fewer problems. Direct deterrence quickly boils down to a question of capabilities: Does the defender have the ability to respond to an attack and impose costs on the attacker that would overwhelm the benefits of the attack? One’s willingness to respond to a direct attack is largely assumed. States fully intend to respond with force to attacks on their homelands; the question is merely whether they can do this effectively. Extended deterrence is trickier. It has befuddled deterrence theorists engaged in devising rational strategy, national security practitioners who try to implement it, and researchers trying to understand it. One fundamental problem is that extended deterrence by its very nature creates a credibility problem. 76 The problem is best described as a discrepancy between self-interest (of the deterrer) and the interests of others (the ally or protégé). Since deterrence theory was greatly affected by Cold War bipolarity and by the existence of nuclear weapons, the question was frequently posed in this way: Would the United States really be willing to sacrifice Washington, New York, and Peoria to save Bonn, Paris, and the Hague? Would we actually carry out our deterrent threat to destroy the Soviet Union in response to a conventional attack on NATO countries in Europe, knowing that the result might well be the destruction of American cities by Soviet nuclear weapons? How could we convince the Russians of this? How could we convince the French, the Germans, and the Dutch? What could the U.S. do to make such a threat credible? The U.S. typically tried to solve this credibility problem by storming it from several angles at once. We “sunk costs” on our allies’ soil by building military bases and infrastructure to accompany it, by conducting joint training exercises, by selling arms, by creating formal

The Dyadic Level of Analysis, Part III

345

alliances attesting to our commitment, by trading with and investing in these countries, by sinking political capital into our relationships with our allies, and in many cases by basing our soldiers there as a “trip wire” to ensure that if fighting broke out, our soldiers would be killed and we would have to respond to save face. All of this was done to build up our credibility. It was also believed that a crucial tool in attempting to strengthen our credibility was to build up our reputation as a country that did in fact make good on its threats and commitments when they were challenged. This created an important practical question: “When and where must a country like the U.S. respond with force to challenges to our interests?” One widely held answer is that it depends on how vital or how important the interests are: Countries must respond with force only to defend states that possess some kind of intrinsic worth or strategic value. 77 The strategic value of a protégé might be based on the presence of any attribute that would contribute significantly to the military capability of great power opponents if it were to fall into their hands, thus changing the balance of military power. Or its value might be more symbolic. The implication is, of course, that states that do not possess any of these strategic resources may be safely sacrificed; they do not necessarily have to be defended. The strong version of deterrence theory suggests a different policy: Great powers like the United States should oppose any moves by other great power rivals to alter the status quo, not because it would result in a change in the military balance but because failure to act would decrease the credibility of their threats. All commitments—to assorted friends, allies, and clients—should be treated as interdependent. Making good on one commitment, regardless of its intrinsic value, bolsters the credibility of all other commitments. From this perspective the American involvement in Vietnam was laudable. Although South Vietnam possessed little real strategic value, it was necessary for the United States to defend it in order to defend the credibility of our threat to resist aggression against ourselves and our allies. The classic statement about the importance that policy makers attach to reputation comes from an official memo, documented in the famous Pentagon Papers, written during the Vietnam War by Assistant Secretary for Defense for International Security Affairs John McNaughton. McNaughton observed that while there were various reasons the U.S. went to war in Vietnam in the first place, there was really only one reason we had not withdrawn but continued to fight: “to preserve our reputation as a guarantor, and thus preserve our effectiveness in the rest of the world.” At another point he surmised that 70 percent of the U.S. goal in Vietnam was simply to “avoid a humiliating defeat to our reputation as a guarantor.” 78 This creates what might be called a “credibility trap,” and it is inherent in the nature of deterrence theory and strategy. Leaders are likely to believe that potential opponents watch their every move for signs of their willingness to make good on their commitments. (This wariness is reinforced by the egocentric bias—our tendency to believe that the actions of others are aimed at us and in response to our own prior actions.) The suspicion is that an opponent not only carefully scrutinizes our interactions with him but our interactions with third parties as well. Thus, any act or statement that might possibly imply unwillingness or inability to make good on one’s deterrent threat must be treated as crucial. Earl Ravenal notes that taken to its logical conclusions, this strategy means that to burnish its reputation for credibility, states “should intervene in the least significant, the least compelling, and the least rewarding cases, and its reaction should be disproportionate to the immediate provocation or

346

Chapter 9

the particular interest at stake.” 79 The trap is that such over-vigilance (or even paranoia) about one’s own actions locks policy makers into always pursuing the toughest line possible lest opponents detect any sign of weakness. Such acts are not only potentially provocative but probably unnecessary. Critics of this kind of muscular deterrence strategy argue that rational opponents are quite likely to assume that how credible your commitment is depends simply on the intrinsic worth of the target and that it would be irrational for you to spend blood and treasure to create a reputation to defend states with little strategic value. Moreover, as Jervis notes, going to extreme lengths to make a commitment look credible might suggest that one is doing this because the commitment isn’t really that solid and you might not make good on it. 80 Reputation may seem important to some policy makers, but is it really important to the success of deterrence? Jonathan Mercer has studied the issue of reputation’s effects on a country’s credibility using an approach based on social psychology. 81 He defines a reputation as a belief that someone has an enduring characteristic; this characteristic (a disposition) is then used to predict or explain future behavior. We normally explain the behavior of others in two distinct ways: we can deem it to be due to dispositional (character-based) factors or situational (environmental) factors. Mercer argues that a reputation is formed as a result of two things. First, an observer uses a dispositional attribution to explain another’s behavior. (I think President Pugachev broke his promise because he is a pathological liar.) Second, an observer uses the past behavior of another to predict that actor’s future behavior. (President Pugachev backed down when confronted with superior power, so he is likely to act the same way in a similar situation.) This second aspect of reputation formation means that an actor’s commitments are perceived as interdependent or coupled with each other. It is also important to note that in this social psychological approach, situational attributes do not affect one’s reputation. If an actor’s behavior is attributed to a particular situation (rather than to his character or internal disposition), then we recognize that if the situation changes, the actor’s behavior may change as well. He doesn’t always act this way. We also know that attributions depend on the nature of the behavior. We are likely to attribute an opponent’s bad/undesirable behavior to internal dispositions. (President Pugachev is paranoid.) Conversely, we tend to attribute an opponent’s good/desirable behavior to situational factors. (They backed down and made concessions to us because our military was stronger than theirs). You will, of course, recognize this as our old friend the fundamental attribution bias, well known to cognitive psychologists. An important result, Mercer says, is that “while adversaries can get a reputation for having resolve, they rarely get a reputation for lacking resolve.” 82 Moreover, anarchy and the danger of war encourage leaders to be cautious in their assessment of potential opponents and to engage in worst-case analysis; thus, they are more likely to perceive that adversaries are resolute than irresolute. Thus, deterrers should get the benefit of the doubt. Since the central problem in deterrence theory is how to make threats credible, and since a reputation for resolve is deemed by the theory to be essential to credibility, and since reputations are used to predict future behavior, it is essential to determine whether in fact reputations are seen as important in establishing and maintaining a state’s credibility. Mercer notes that the statistical studies by Huth and Russett indicate that a defender’s past behavior and reputa-

The Dyadic Level of Analysis, Part III

347

tion are only important when the two states have a continuing rivalry with prior confrontations. 83 This suggests to him that “A’s past behavior toward B, C, and D is irrelevant to E, which infers A’s future behavior only from A’s past behavior toward E.” 84 Huth and Russett also find that if A retreated from B’s challenge in a previous crisis, then B was likely to challenge A in a subsequent crisis—as deterrence theory would predict. However, they also find that if A defeated B’s challenge in the previous crisis, B was also likely to challenge A again in the next crisis. If either a reputation for standing firm or a reputation for backing down brings about a subsequent challenge, then surely a state’s reputation cannot be the determining factor. Mercer examines the beliefs of decision-makers in three successive crises from 1905 to 1911—the first Moroccan crisis, the Bosnian annexation crisis, and the second Moroccan crisis. All involved the same five European great powers—Britain, France, Russia, Germany, and Austria-Hungary. His finding is that leaders believed that when their adversaries backed down, this was due to external (situational) constraints, not weakness of will or reputation for resolve. One’s reputation for resolve was simply not that critical to deterrence. This being the case, it makes no sense to stand firm—especially to defend rather weak national interests— just to protect your reputation. Other investigations come to similar conclusions. For instance, Daryl Press confirms that a state’s decision to challenge a defender’s commitment is based largely on estimations of current military capabilities and the interests at stake in the situation rather than lessons drawn from previous interactions in similar circumstances. 85 Likewise, Ted Hopf’s study of American extended deterrence commitments in the Cold War concludes that Soviet assessments of American credibility to defend its commitments in Europe were largely unaffected by American behavior in the Third World—despite its losses in Vietnam and Cambodia. 86 On the other hand, opponents are not totally insensitive to one’s credibility-related actions elsewhere; they’re just rationally sensitive to what they look at in drawing their conclusions. For instance, John Orme agrees that while lack of U.S. resolution in the Third World did not encourage a Soviet attack on Europe, it did encourage them to act more forcefully in regions where they perceived weakness. The interdependence of commitments tends to be limited to the interaction of a pair of countries within a particular region. To the extent that reputations are important they “tend to arise from the past record of interactions between the same defender and the same potential attacker within the same geographic region.” 87 Patrick Morgan argues that while reputations are sometimes important, this is not always true, and there is no consistently reliable basis for determining reputation. He presents an interesting thought project to illustrate. If resolve and reputation are most effectively demonstrated by the actual previous use of force against an opponent, then deterrence should be most effective between enduring rivals—because these pairs engage in multiple demonstrations of force against each other over the years. Surely, after not too long a period, both states should realize the clear commitment and resolve of the rival. The perception of credibility should strengthen over time in both states. Thus, deterrence should become stable fairly quickly, crises and militarized disputes should lessen dramatically, and the chance of war should drop dramatically. Unfortunately, the behavior of enduring rivals does not fit this pattern. 88 Rival-

348

Chapter 9

ries tend to escalate from one crisis to the next. Morgan’s analysis of the available information concludes: We lack compelling evidence that commitments are interdependent . . . Case studies and statistical analyses find little evidence that states assess each other’s resolve on the basis of past actions visà-vis third parties, and only modest evidence that past actions vis-à-vis themselves shape images of others’ credibility. 89

The evidence from crisis bargaining studies may add to our understanding here. They strongly suggest that full-bore strategies that attempt to manipulate risk and commitment by demonstrating toughness and resolve are likely to provide poor outcomes. On the other hand, “more moderate bargaining moves seem to serve as effective signal because attackers often distrust defenders’ claims of defensive intentions and they must also face the audience costs associated with backing down.” 90 Russell Leng’s studies of dyadic interactions (see chapter 8) show that combining threats with limited positive inducements is more effective in preventing crises escalation than relying only on deterrent threats. 91 So, instead of hardline attempts to establish credibility of threats, “firm but flexible” strategies do better in convincing potential attackers to retreat by reducing the domestic and international political costs of retreating from military confrontation. The use of the stick side of the strategy continues to send the signal that military conflict will be costly, but the carrot side of the strategy shows the opponent that positive results may accrue from a decision to remain at peace. Theorists have noticed another practical problem with extended deterrence commitments. Iron-clad commitments by great-power defenders may create an incentive for weaker allies to engage in reckless, intransigent, or adventuristic policies at home or abroad—secure (presumably) in the belief that they are protected by the deterrent shield of a great power. Firm commitments may thereby create a situation of moral hazard in which the protégé manipulates the commitment made to it by the great power, ultimately entangling its protector in an unwanted war. 92 Fearon suggests this is why defenders may prefer to keep commitments less than absolute. 93 The U.S. commitment to Taiwan is typically seen in this light. It is variously described as a policy of “strategic ambiguity” or as one of “dual deterrence and dual reassurance.” 94 The U.S. worries that a firm commitment to support Taiwan under any circumstance might embolden it to declare its independence, triggering a military response from the PRC. The U.S. has therefore given Taiwan a less-than-iron-clad commitment, thus attempting to avoid the moral hazard of tempting Taiwan to engage in domestic political behavior that could start a major conventional war in Asia. It tries to deter China from turning to a violent solution to the Taiwan dilemma, while also reassuring them that the U.S. will not back Taiwan if it declares its independence. Thus, it attempts simultaneously to deter both sides from acting recklessly. DETERRENCE: EMPIRICAL RESEARCH Although a tremendous amount has been written about the theory of deterrence, relatively little empirical research on conventional deterrence was undertaken until the 1970s. Most of this research has focused on immediate deterrence rather than general deterrence, and more

The Dyadic Level of Analysis, Part III

349

specifically, it has focused on immediate extended deterrence, reflecting one of the central concerns of the U.S. in the Cold War era. The results have been inconclusive and contentious. Statistical Studies of Deterrence One of the scholars most active in research on deterrence is Bruce Russett. One of the earliest data-based studies on deterrence was reported in his article “The Calculus of Deterrence.” 95 Russett focused on immediate extended deterrence—that is, attempts by one state (the defender) to deter the impending use of military force by another state (the potential attacker) against a friendly third state or ally (the protégé). Russett was particularly interested in discovering what determined whether a deterrent threat failed or succeeded. To investigate this question he examined data from seventeen cases of deterrence between 1935 and 1961 in which a great power tried to prevent an attack on a smaller protégé. The seventeen cases were all situations in which an aggressor was deemed likely to take action against the client state. Russett’s analysis led him to some interesting initial conclusions. First, it did not seem that potential attacks against the more important protégés were more successfully deterred than those against less important clients. In fact, all cases of successful deterrence involved relatively unimportant protégés (in terms of gross national product and population). In other words, successful deterrence did not seem to be linked to the client’s intrinsic strategic value. Second, an explicit commitment by the great-power defender was found to be no guarantee of deterrent success. Third, neither local nor general military superiority in the hands of the greatpower defender seemed to guarantee successful deterrence. These are interesting results because deterrence theory suggests that clear military capability to retaliate coupled with explicit commitments should result in a credible threat, which should result in a successful deterrent effort. If these factors do not lead to successful deterrence, what does? Russett discovered that the best predictor of successful deterrence seemed to be the existence of close military, economic, and political ties between the defender and the client. In all cases of successful deterrence strong military cooperation was present, with the defender supplying the protégé with both arms and advisers. In four of the six cases of successful deterrence there were close political ties, and in five of the six cases there was substantial economic interdependence between the client and defender as indicated by bilateral trade statistics. 96 The general breadth and intensity of relations between the defender and the pawn would seem to signal the credibility of the defender’s commitment to the client state, thus enhancing the probability of successful deterrence. Russett’s early work on deterrence has been significantly expanded. Russett and Paul Huth subsequently identified fifty-four cases of immediate extended deterrence between 1900 and 1980. 97 Using an expected-utility model of deterrence, they assume that deterrence would be achieved if the cost-benefit calculation of the attacker concludes that the expected utility of an attack would be less than the expected utility of forgoing the attack. The potential attacker’s calculations would logically include estimates of the deterrer’s capability to defend the protégé and its commitment and willingness to do so. Huth and Russett devised indicators that indirectly tapped such considerations: the relative military balance between the attacker and the defender and protégé, the economic and military ties that bind the defender to the protégé,

350

Chapter 9

the intrinsic (or strategic) value of the protégé, and the defender’s previous behavior in deterrent situations (his reputation). They found that deterrence was successful in thirty-one of the fifty-four cases. Deterrence attempts were most successful when there were strong economic links (trade) and political links (arms sales) between the defender and the protégé and when the local military balance favored the defender and the protégé. Formal alliance ties did not appear to enhance the chances for deterrent success, nor did the defender’s past behavior in crisis, or even the defender’s possession of nuclear weapons. 98 Huth and Russett subsequently revised and expanded their data set to include the years 1885–1984, resulting in fifty-eight cases of immediate, extended deterrence. Their analysis of the period confirmed that the immediate, short-term balance of conventional military forces had an important impact on deterrence success, though neither the long-term military balance nor the possession of nuclear weapons played a major role. Apparently, challengers do not initiate conflict with the intention of creating long-term wars of attrition, but with the desire to win quickly and establish a fait accompli that cannot be overturned by the defender. What matters most, therefore, is the ability of the defender to prevent such a quick victory. Deterrence through the denial of immediate gains seems more important than a credible threat of punishment in the long term. 99 This means that successful deterrence depends on the ability of the deterrer to rapidly move forces into a position to repulse and attack—an ability that was absent, for instance, in the Iraqi attack on Kuwait in 1990. The British predicament in 1939 may also be instructive here. Some scholars have argued that although Britain tried a strategy of long-run deterrence against Hitler—threatening to win a war of attrition—the only strategy that would have prevented Hitler’s attack on Poland was a credible threat to deny Germany the immediate ability to take Poland, and this could not have been accomplished without British and French cooperation with the Soviet Union. 100 Huth and Russett also examined the credibility of the defender’s threat by looking at his pattern of diplomatic negotiation in previous deterrent situations and at his strategy of military escalation in the crisis. Their findings are illuminating. They discovered that a reciprocal (TFT) strategy of military escalation was more likely to achieve a successful deterrent effect than either a policy of responding more forcefully to the opponent’s actions or a policy of responding at lower levels of military action. Likewise, a reciprocal (“firm but flexible”) strategy of diplomacy was more effective for the defender than either bullying or conciliatory strategies. Deterrence was most effective when combined with conciliation or reassurance. 101 A past record of either bullying or conciliation (backing down) reduced the chances for successful deterrence. Bullying sometimes worked against a much weaker potential attacker, but it was a risky deterrent strategy against a state with similar military capabilities. 102 As one would expect, if a defender was previously forced to concede by the same attacker, the defender’s threat would lack credibility in the next crisis, creating a deterrent failure. However, as we have seen, bullying behavior by the defender in the previous encounter with the same attacker also decreased the chances that deterrence would succeed in the future. Once bullied into concessions, the potential attacker was less likely to retreat again and suffer further damage to its reputation and prestige. 103 Paradoxically, previous deterrent success against the same potential attacker can reduce the chances for success in the future by increas-

The Dyadic Level of Analysis, Part III

351

ing the loser’s grievances. 104 In the long run, a stalemated previous crisis appeared safer than one in which the defender backed down or bullied his way to success. In fact, in Huth and Russett’s study, a stalemate was the only past behavior not associated with the failure of deterrence. 105 Deterrence theorists are becoming increasingly aware that for deterrence to work, the challenger must not only believe the defender’s threats are credible, but the deterrent acts themselves must not increase the potential attacker’s fear of preemptive military action, thus provoking escalation to war. 106 Deterrence can fail because conflict spirals are sparked by states attempting to make credible threats. 107 Decision makers in deterrent situations must walk a fine line. If they demonstrate too much toughness, they risk provoking an overly hostile reaction from a potential challenger who was not previously committed to risking war. If they don’t demonstrate enough firmness, however, they may encourage wishful thinkers in the potential attacker to believe the defender will not make good on its commitment. 108 Leaders need to guard against both appeasement and provocation. Toughness is acceptable, but states should refrain from rigidness and bullying; on the other hand, states need to be conciliatory without being too soft. As a result, Janice Gross Stein has famously argued for the necessity of combining deterrence with “reassurance.” 109 The studies of Huth and Russett show that the nature of deterrent successes and failures falls pretty well within the expectations of deterrence theory. Potential aggressors can be deterred by credible threats of retaliation by a defender, especially if the defender enjoys immediate local military advantages and if the defender’s commitment to the protégé can be clearly discerned from its economic and political-military ties to the protégé. Formal alliance ties seem unimportant as evidence of the defender’s commitment, but this can be explained by the potential attacker’s attention to deeds rather than words and treaties. The defender’s reputation for carrying out commitments seems to be important, but only with regard to the same potential challenger. Strong commitment and a reputation for resolve can be important if absent (though there is disagreement on this); but they are certainly no guarantee if present. Deterrent failures can be laid at the doorstep of less-than-credible military capabilities and less-than-clear-cut commitments to the client state—all predicted by deterrence theory. A final piece of the puzzle deals with alliances. Remember that deterrence theory suggests that one good way to establish a commitment to defend a protégé is to establish an alliance. However, we have already seen that during some periods of time alliances are statistically associated with war rather than the maintenance of peace. We have also noted previously that this may be due to the fact that different alliances may have different purposes. It is important then to devise data sets that are capable of identifying conceptually useful distinctions among alliances. This has now been accomplished with the development of the Alliance Treaty Obligations and Provisions (ATOP) project. Using ATOP data, Brett Benson and others have done research on the ability of alliances to add to deterrence. 110 It is clear that the type of alliance makes a difference. Some offensive alliances are aimed at compellence, others are aimed at deterrence. Another important difference has to do with the conditions that trigger the obligations. In unconditional deterrent alliances, members are obliged to assist their allies militarily under the implied condition that allies not force an offensive change in the status quo. They are obliged to support all defensive objectives,

352

Chapter 9

including preemptive defense. In conditional deterrent alliances, members are only obliged to defend members if they are not attempting to force an offensive change in the status quo and if a target of the alliance attacks an alliance member. Looking at alliance and militarized dispute (MID) data from 1816 to 2000, Benson makes several interesting discoveries. Not all alliances deter. Unconditional compellent (offensive) alliances are actually associated with conflict, as we might imagine. Joining such an alliance actually increases the probability that members will initiate conflict. Unconditional deterrent alliances are also clearly associated with violent conflict. Conditional deterrent alliances do deter MIDs, but the relationship is relatively weak. Nevertheless, members of conditional deterrent alliances are better able to deter challenges than those without alliances. Successful deterrence of violent conflict appears to be limited to states engaged in a conditional deterrent alliance with a great power. The primary beneficiaries of such alliances are minor powers. The finding that different types of alliances produce different effects on the likelihood of conflict helps to explain why some studies suggest alliances lead to war while other studies suggest they help to deter war. Empirical Studies of Deterrence: A Critique One of the major bones of contention among political scientists who study deterrence is how to select actual cases of deterrence. One criticism, originally raised concerning Russett’s initial study but generally applicable to most studies of extended deterrence, is that if one looks only at cases in which the client is threatened, one may in fact miss the most successful cases of deterrence—those where the defender is able to project so great a deterrent threat that no opponents even seriously considered action against the pawn. 111 Many of these are cases of general deterrence rather than immediate deterrence, and detecting success here is difficult because it leaves few fingerprints. 112 Another criticism is equally important. The existence of a case of immediate extended deterrence requires the presence of a potential attacker who is seriously interested in the immediate use of military force and the presence of a defender who actually engages in a serious effort to deter such a use of force. Critics have charged that Russett and Huth (as well as others) have been guilty of using cases in which (1) the potential attacker was not seriously contemplating the use of military force against a client state, thus leading to the erroneous impression of a deterrent success, or (2) the defender did not really engage in a serious effort to deter the attacker, thus giving the erroneous impression of a deterrent failure. 113 For instance, Richard Ned Lebow and Janice Gross Stein, using far stricter definitions for the occurrence of immediate extended deterrence, conclude that of the fifty-four cases in Huth and Russett’s original data set, only nine really fit the definition; and they find only ten real instances of extended deterrence in the later fifty-eight-case set. 114 Of the nine cases of deterrence that they identify in the twentieth century, only three cases were deemed successful and only in the short run—Munich, Egypt’s deterrence of an Israeli attack on Syria in May 1967, and American deterrence of a Turkish attack on Cyprus in 1964. They find immediate deterrence successes to be “uncommon, partial and tenuous.” 115 Looking at deterrent cases identified by Huth and Russett, by George and Smoke, and by Organski and Kugler, they conclude that only twelve cases are coded the same way by all three studies. A more recent assessment of seven studies of deterrence found that the overlap in coding deterrence events

The Dyadic Level of Analysis, Part III

353

was less than 60 percent. 116 Such a wide discrepancy concerning the relevant cases to be examined is astonishing. It points not only to the theoretical difficulty of identifying cases of non-occurrences like deterrence events but also to the wariness with which we should treat the results of any statistical analysis based on a particular set of cases. Determining successes and failures is equally difficult. It has already been suggested that some cases that appear to illustrate successful deterrence may, in fact, be (1) cases in which the adversary never really harbored any intention to act and thus did not need to be deterred, or (2) cases in which the adversary, while indeed wishing to challenge the status quo, was dissuaded from military action not by the expected costs that a defender might impose on him in retaliation, but by other considerations—ideological, political, moral, or legal. 117 Just because an attack does not take place does not mean peace can be attributed to successful deterrence. As Morgan points out, identifying cases of successful deterrence involves proving why something that did not happen did not happen—a task that is a logical nightmare. 118 The old joke about the Ohio farmer may be instructive. It seems that one afternoon an eccentric old farmer rather ostentatiously erected an extremely large pole in the middle of his backyard. His neighbor, spying the novel ornament, asked him why he had placed this oversized bean pole in his backyard. The farmer replied that it helped to keep the elephants away. When the neighbor protested that there were no elephants in the area, the farmer exclaimed triumphantly, “See, it works!” Deterrence may be much the same type of phenomenon. We have already seen that several scholars have argued that the number of deterrence successes found by Huth and Russett is rather inflated and that the number of deterrence failures is much more considerable. In fact, some theorists argue that the number of deterrence failures is so large and the manner in which deterrence fails is so much at odds with rational deterrence theory (RDT) that the theory itself requires major renovations. The greatest challenge comes from those who use case studies rather than statistical analysis, though as usual the results are mixed. Case Studies and Deterrence Scholars have frequently used comparative case studies to tease out information about deterrence. Some of these case studies have looked at situations of direct deterrence as well as indirect deterrence. In their classic work Deterrence in American Foreign Policy, Alexander George and Richard Smoke analyzed the deterrent efforts of the United States in the Cold War period from 1948 to 1962. 119 They rigorously examined and compared eleven cases of U.S. deterrent threats in an effort to explain how deterrence fails and how it succeeds. What leads a potential initiator to mount a challenge to deterrence? The most important factor seems to be the initiator’s view of the risks involved. In almost all the cases the authors find evidence that before taking action, the initiator concluded that the risks of the option he had chosen would be calculable and controllable in a way that would ensure that the risk would be acceptable. Such a feeling on the part of the initiator seems to George and Smoke to be a necessary—but not sufficient—condition for a decision to challenge deterrence. On the other hand, the belief that the range of options available is not calculable or controllable is usually a sufficient

354

Chapter 9

condition for deterrent success. 120 The decision makers in the challenging states seemed to be risk-averse loss minimizers rather than risk-acceptant gain maximizers (just as prospect theory would predict). A second major factor is the initiator’s view of the defending state’s commitment. Contrary to what one might expect, George and Smoke found that an American commitment per se was insufficient to deter a potential initiator. If the opponent possessed an option for challenging the status quo that seemed likely to achieve his objectives at an acceptable cost/benefit ratio, deterrence could fail even in the presence of a seemingly credible U.S. commitment. The authors concluded that the initiator’s perception of the presence of a defender’s commitment may be a necessary condition for deterrent success, but not a sufficient condition. 121 George and Smoke identify three distinct patterns of deterrent failure. Pattern 1 is termed a fait accompli since it involves actions designed to overturn the status quo quickly before the defender has a chance to decide he is committed to reversing the change. Fait accompli patterns occur when the initiator believes that no commitment by the defender actually exists. Of course, if the initiator is correct in his view that no commitment exists, one cannot properly say that there has been a failure of deterrence, since the defender was never committed in the first place. 122 The Soviet intervention in Hungary in 1956 fits this pattern. Soviet intervention was certainly intended to present a fait accompli to the West, but just as certainly the United States was never committed to retaliating against the USSR for any aggression within its own sphere of influence. Sometimes, of course, a commitment does exist in fait accompli gambits, though the opponent believes otherwise. This was probably the situation in the North Korean attack on South Korea in 1950 as well as in the Iraqi attack on Kuwait in 1990. The second pattern of deterrence failure discovered by George and Smoke is called the limited probe. Such actions occur when the initiator believes a defender’s commitment is uncertain or ambiguous. He then initiates a controlled crisis to clarify the defender’s commitment, presumably following Lenin’s famous dictum that “if you strike steel, withdraw; if you strike mush, push on.” The limited probe is frequently followed by a second phase, which George and Smoke identify as controlled pressure. If the initiator regards the defender’s commitment as solid, he may still attempt to find a way around it if he can at the same time keep his options both calculable and controllable. Controlled pressure responses involve lowrisk strategies involving minimum force. 123 George and Smoke conclude that deterrent threats by the United States against opponents that employ low-risk strategies have often been ineffective. Initiators frequently calculate that defenders may not be willing to adequately defend against low-level violations of the status quo. The fact that an initiator possesses multiple options for violating the status quo seems to be a crucial factor in determining whether initiators will risk challenging deterrence. 124 In limited probes and controlled pressure, challenges to the status quo are mounted incrementally in a series of small steps, each of which is reversible if need be. This prompts one commentator to write, “it is not surprising that the effectiveness of deterrence is rather skimpy at the margins, hard to sustain when the challenges come at low levels in places of limited importance over matters not intrinsically of vital national interest. And it is hardly surprising that this is exactly the form in which most challenges appear.” 125 Indeed, many scholars who study Israeli deterrence have found that Israel’s Arab neighbors have frequently resorted to

The Dyadic Level of Analysis, Part III

355

small-scale cross-border raids, skirmishes, and asymmetrical tactics that are not easily deterred by the threat of Israeli reprisal attacks. 126 In this regard, U.S. interactions with Iraq are instructive. Frank Harvey and Patrick James start with the observation that crises and wars are typically not one-and-done encounters but complex and multidimensional events that encompass a sequence of exchanges that frequently mix attempts at both deterrence and compellence. 127 They investigate six protracted crises between the U.S. (and its coalition partners) and Iraq from the immediate aftermath of the Gulf War in 1991 until the outbreak of war in 2003. 128 They found that Iraq continually tested the resolve of the U.S. and its allies in the two no-fly zones, but it almost always backed down when it struck steel instead of mush: “Iraq backed off from extreme confrontation in the face of clear and powerful threats; when the message began to fade, it started to probe again and test for new limits.” 129 Iraqi strategy looked very much like the combination of limited probes and controlled pressure that George and Smoke discovered decades earlier. They conclude that when attempts at deterrence and compellence succeeded and when they failed, it was for “the right reasons”—consistent with expectations derived from the deterrence theory about the need to clearly define behavior deemed unacceptable, to clearly communicate commitment to punish violations, to possess sufficient physical capability, and to demonstrate resolve to carry out retaliation. As we know, the final crisis, from November 2002 to April 2003, ended in a war that probably need not have happened. Harvey and James assign blame to the particular kind of deterrence being practiced. They argue that multilateral approaches to deterrence—those that involved a collective attempt by several parties to deter a challenger—are inherently difficult to sustain over time and played an important role in the failure of compellence and deterrence efforts. Lack of Security Council support for the American-British hard line on Iraq reduced the certainty of deterrence and compellence by 2002–3; and the coalition partners were constrained in their ability to make threats and respond to Iraqi provocations. In this case Saddam Hussein mistook Chinese, Russian, and French opposition to Security Council resolutions permitting force to mean that the Security Council would prevent the use of force. The authors argue that Saddam Hussein perceived that the U.S.-UK threat wasn’t credible because without multilateral approval it lacked legitimacy and therefore would not happen. Ultimately, Hussein underestimated both the probability of an attack and the risks of noncompliance. 130 He had continued to probe the limits of American-British resolve without understanding that he had crossed a red line. Amatzia Baram has sifted through official Iraqi documents captured by American forces (including tape recordings of high-level policy meetings) as well as transcripts of interviews with top Iraqi officials. 131 Like Harvey and James, he finds numerous examples in which Saddam Hussein was deterred: He did not retaliate for Israel’s attack on Iraq’s Osirak nuclear reactor in 1981; he did not use chemical or biological weapons against U.S. and coalition forces in the 1991 Gulf War; and he did not launch a second invasion of Kuwait in 1994. All of these decisions are credited to Hussein’s rational calculation of the balance of capabilities that he faced. Additionally, Saddam also gave in to Iranian compellence efforts to sign the Algiers Treaty in 1975 and surrender the eastern bank of the Shatt al Arab. Once again, he

356

Chapter 9

gave in to superior military capabilities. (But he also got a free hand to deal with the Kurdish revolt in northern Iraq.) However, there were several attempts to deter or compel Iraq that failed in situations where they should have succeeded. In 1980 Iranian deterrence (to the extent that it existed) failed to prevent Saddam’s attack, largely due to his miscalculation of the true balance of capabilities between the two Gulf rivals and strength of Iranian nationalism. In 1990 deterrence efforts by the U.S. (again, to the extent they existed) failed to prevent an Iraqi attack on Kuwait. According to Baram, it is clear that Saddam actually expected an American military retaliation, though he expected it would be limited. He felt confident that the risks were acceptable because the U.S. would not wish to engage in a war that would create large American casualties. In a more limited way, clear Israeli deterrent threats to prevent a Scud missile attack by Iraq failed in 1991. Here Saddam believed that if he attacked with conventional weapons, the Israelis would be confined to a conventional retaliation. These were consequences he did not fear but instead welcomed, since he hoped that an Israeli entry into the war would split the American-led coalition. Finally, American attempts at compellence failed twice. Actions by the multilateral coalition led by President George H. W. Bush failed in 1990–91 to induce Iraq to withdraw from Kuwait, in part because Saddam saw the risks as acceptable. In spite of rather clear evidence of American credibility, Saddam did not believe the U.S. had the will to fight over Kuwait. (He thought the chances fifty-fifty at most.) Moreover, the domestic political risks were high if he were seen to have retreated. In 2003 the George W. Bush administration failed to compel Saddam Hussein to relinquish power or to provide greater transparency in its WMD program. (It’s not entirely clear what the Bush administration’s goal was and whether it would have been willing to take “yes” for an answer if Saddam agreed.) In 2003, in spite of very clear and open evidence of the U.S. intent to invade, Saddam resisted coming to that conclusion. He rationalized dozens of reasons why the American would not invade: the French, German, Russians, and Chinese would prevent a war; global public opinion would be against an attack; the U.S. would not send ground forces but only bomb military targets in the country. Ultimately, according to Baram, “Saddam was able to convince himself that a U.S. assault was unlikely and that if it came, he would still survive it.” 132 He also perceived that the need to keep the threat of WMD for use against Iran and Shi’a opponents in Iraq was essential. A key question is whether Saddam’s unwillingness to give in to superior military capabilities and credible threats was rational. Baram’s argument is that Saddam always “acted according to a coherent set of self-interested preferences,” “tended to ignore inconvenient facts and unpleasant information,” “constructed convoluted scenarios that allowed him to believe events would play out in the way that he wanted,” ignored objections of others, constantly “constructed rationalizations” to explain away things he should have feared, had a sense of infallibility, used a thought process in which “rational elements became fused with entirely fanciful calculations,” and was “delusional and reckless.” 133 Having constructed a list of the extent to which Saddam Hussein distorted rational problem solving, he then says, “This did not make him irrational—but it did make him very dangerous and very, very hard to deter.” 134 He comes to the rather unsatisfactory conclusion that one cannot count on rational regimes to act in predictable ways. Whereas Harvey and James tend to see the success or failure of

The Dyadic Level of Analysis, Part III

357

deterrence and compellence attempts regarding Iraq as consistent with deterrence theory, Baram argues that leaders respond differently and not always as deterrence theory predicts. Individuals are important; they respond differently to attempts by others to coerce their behavior and statesmen need to account for this. Remember that deterrence theory does not say that attempts at deterrence will always be successful; deterrence will fail if the defender’s retaliatory threat is “absent, incredible, or less valuable than the prize.” If deterrent attempts are badly designed or implemented, this does not add up to a falsification of deterrence theory. 135 What would falsify the theory is the prevalence of situations in which the defender meets the requirements of the theory, but the attacker strikes anyway. In other words, when the defender (1) specifically defines its commitment— which may involve the delineation of red lines, (2) communicates this commitment to the potential attacker, (3) possesses sufficient military capabilities to carry out the commitment, and (4) demonstrates its willingness (resolve) to carry them out, and the attacker still initiates military action, then something is wrong. The challenger is not acting as the theory predicts. Some critics contend that far too many such instances can be identified. For instance, Lebow identifies eight brinkmanship crises in which the defender met all four of the standard conditions and failed to deter the potential attacker: Fashoda (1898), Korea (1903–4), Agadir (1911), July Crisis (1914), Korea (1950), Cuban Missile Crisis (1962), Sino-Indian (1962), and Arab-Israeli (1967). 136 He argues that in each of these cases the defender’s commitment and capabilities were solid, but they were not necessarily perceived as such by decision makers in the challenging state. What was important was that the challenger perceived weaknesses in the defender’s capabilities and commitment, not that such weaknesses actually existed. The inability to correctly assess the capabilities and commitment of the protecting state led to deterrence failure under conditions where deterrence success would have been confidently predicted. This is a central finding of many case studies of deterrence: Cognitive factors at the individual level of analysis are crucial. A later case study by Lebow, the Argentine junta’s decision to precipitate the Falklands War in 1982, points out the importance of domestic political motivations in deterrent situations. If a consideration of the military balance in the South Atlantic had been the main consideration, as traditional deterrence theory hypothesizes, then the Argentine generals should simply have waited another year before launching their challenge to Britain. By then the H.M.S. Invincible would have been transferred to the Australian navy, the Hermes gone, and the Intrepid and the Fearless in the scrap heap. In Lebow’s estimation, the generals and admirals in control of Argentina were less concerned with military requirements (which suggested patience) than they were with their domestic political vulnerabilities (which dictated immediate action). 137 Patrick Morgan suggests another obvious case of deterrence failure involving domestic considerations—Kosovo in 1999. Ever since the initial collapse of the Yugoslav state and the ensuing violence—as Serbia sought to prevent the secession of Slovenia, Croatia, and BosniaHerzegovina—the U.S. and its NATO allies had warned Serbian President Slobodan Milosevic not to use force against ethnic Albanians in Kosovo. This general deterrence was successful for several years, though largely because Milosevic was tied down in the fighting against the

358

Chapter 9

Croats and Bosnia Muslims. Once those territories had won their independence in late 1995, Milosevic’s attention turned to Kosovo, moving the situation to one of immediate deterrence. By early 1995 it appeared that Milosevic was pursuing a vigorous campaign of ethnic cleansing to force Albanians out of Kosovo. NATO threats to intervene were clear and American prestige was committed, but there were also simultaneous attempts at conciliation. American Secretary of State Madeleine Albright organized a conference in March 1999 at Rambouillet, a chateau outside Paris, to attempt to devise an internal political settlement between Albanian Kosovars and the Milosevic regime. But the red line appeared clear; if Milosevic did not sign the Rambouillet accord, NATO would use force to prevent further atrocities. Milosevic refused to sign and six days later NATO began a bombing campaign. Why did deterrence fail? Morgan suggests that Milosevic simply had too much at stake domestically; maintaining his shaky control over internal political power was his primary concern. 138 Janice Gross Stein analyzed six cases in which the United States attempted to apply immediate extended deterrence in the Middle East in the 1960s and 1970s. 139 Four of these attempts failed; in the other two the challenger backed down, but it was difficult to determine conclusively whether this was due to the deterrent threat of the United States. In each of the cases the links between the United States and its protégés (Israel, Jordan, and Saudi Arabia) were clear and substantial; thus deterrence failed even though the challengers knew the links between the defender and the target were strong. Reputation was also less important than might have been predicted by deterrence theory; even though the United States stood firm in five of the six cases, deterrence failed in three of these five. 140 Stein maintains that in each case the challenger’s primary motivation was a perception that the costs of inaction were too high. The leaders in the challenging states were vulnerable to domestic and international pressures, increasing their motivation to challenge the status quo. This is consistent with many other studies of deterrence that indicate that a deterioration in either a leader’s domestic political position or in the state’s international security situation may motivate challenges to deterrence. 141 Stein concludes that for challengers who are motivated primarily by their own internal weakness, a strategy of conciliation is perhaps better able to deter unwanted behaviors than a strategy of credible threats. While the earlier statistical studies of deterrence looked to the defender’s weaknesses (in capabilities and credibility) to explain deterrence failure, much of the case study literature has focused on the challenger’s attributes. These studies generally conclude that the strength of a challenger’s motivation is crucial in determining if a challenge emerges and how far it goes. The balance of resolve is important: deterring a highly dissatisfied challenger is difficult. Robert Jervis’s analysis is that it is certainly easier to deter opponents from doing something they don’t care much about than it is to deter them from something they deem vital. Deterrence has always been “easiest where it was least needed” and most difficult “when and where the target was most motivated.” 142 Case studies appear to confirm this. In this respect, as we have seen, issues of territory are usually considered to be particularly important by national leaders, especially if the people who live in the territory are ethnic co-nationals. Thus it is quite likely that policies of deterrence are doomed to fail if the challenger is bent on regaining territory viewed as its own. 143

The Dyadic Level of Analysis, Part III

359

The Egyptian and Syrian attack on Israel in the 1973 October War is a good example of this. The consensus appears to be that Israel had clear military superiority that was recognized by Sadat and other Egyptian leaders, and its resolve to defend itself was clear, with Israeli red lines communicated unambiguously to Arab states. Nevertheless, Israeli deterrence failed due to the combination of strong dissatisfaction with the post-1967 status quo, strong domestic political pressures on Sadat’s government, and therefore a strong motivation to do something. Israeli nuclear weapons appeared to add little to Israeli deterrent ability. 144 Several studies of Israeli deterrence problems conclude that potential attackers, overmatched by Israel’s military capabilities, have “designed around” Israel’s deterrent threats by downsizing their territorial goals and devising less ambitious plans of attack. 145 Case Studies of Deterrence: A Critique The case study approach has come under attack from a number of quarters, but primarily for the common problem of case selection bias. Comparative case studies typically do not use random samples of deterrence interactions, but have generally looked at cases of deterrence failures. This has resulted in a bias in case selection in favor of cases involving particularly resolute challengers who are unlikely to be deterred. Moreover, in focusing on crises, “acute” crises, or cases of immediate extended deterrence, they have selected cases in which general deterrence has already failed. In part this begs the question of whether we can generalize about deterrence from only one type of deterrence—immediate extended deterrence. Critics have argued that selecting cases in this way gives an unwarranted impression that the success/ failure rate of deterrence is one of overwhelming failure. If scholars had looked at the full universe of cases relevant to deterrence, they might actually discover that deterrence works a high percentage of the time. 146 Others respond that in fairness, the case studies advocates are not saying that deterrence fails most of the time, only that it fails frequently, and under conditions contrary to rational deterrence theory. Critics have also disagreed with the interpretations of these cases, arguing that, in fact, the cases are consistent with what we would expect from rational deterrence theory. For instance, John Orme famously reanalyzed Lebow’s eight cases and came to quite different conclusions. He believes that these deterrence failures are consistent with deterrence theory rather than refuting it. He finds that in each case there are serious problems with the defender’s position; either a weak commitment or a weak military capability “tempted an aggressive, perhaps riskprone, but not necessarily irrational opponent.” Badly conducted deterrence efforts account for most of the failures. He also finds that when defenders exercise restraint to prevent provocation, the results are disastrous. 147 Deterrence cases seem especially susceptible to widely differing interpretations. A midpoint argument about the effectiveness of the policy of deterrence is perhaps most useful. Frank Harvey has suggested that the previously mentioned conditions for credible deterrence should only be considered as necessary conditions for deterrent success, not as sufficient conditions. Logically, if states meet these conditions, their chances of success are improved, but since the conditions do not constitute jointly sufficient conditions, rational deterrence theory would predict there will be some deterrence failures even when all the conditions are met. 148 Deterrence should always be seen as a probabilistic phenomenon.

360

Chapter 9

COGNITIVE PSYCHOLOGY AND DETERRENCE Many critics of deterrence who use the case study approach—such as Jervis, Lebow, and Stein—argue that deterrence does not work because of cognitive or perceptual variables at the individual level. These scholars have drawn on recent work in the field of cognitive psychology. Stein argues that a generation of research by cognitive psychologists and neuroscientists has produced a consistent body of knowledge that seriously challenges both a “tight” theory of deterrence and also a “looser” theory of deterrence based on more commonsensical understandings of rationality. She argues that even these relaxed requirements for rationality are seldom met in the real world. 149 Stein points out a number of ways in which nonrational cognitive processes may derail the ability of national leaders to act in accord with deterrence theory. First, as we know from our previous discussions, the default system of human decision making is not deliberative and calculating; it is a “hot” emotional process that precedes reflection and analysis. 150 Decision making in this normal default mode tends to be quick, unreflective, intuitive, and emotional. Deterrence events by their very nature exist within a “hot” context. After all, the essence of deterrence is that it deals with threats, and we know that emotions play an important role in the evaluation of, and response to, threats. Stein argues that “the stability of deterrence is likely much more affected by a ‘hot’ emotional reaction to betrayal, disappointment, shame, or humiliation than it is by a ‘colder’ calculus of capability and resolve.” 151 She suggests that for deterrence to succeed, it should be accompanied with positive inducements early in the process to “reduce the hot emotional sting produced by deterrent threats . . . Deterrence works best when emotions are moderated, in stable environments.” 152 Second, our understanding of cognitive styles shows that hedgehogs and foxes will approach deterrence situations differently. Foxes are inductive, open thinkers who are capable of admitting errors and reevaluating their previous opinions and estimates—including estimates of an opponent’s capabilities and resolve. Hedgehogs are deductive thinkers who have a strong need for coherence and are likely to dismiss inconsistent evidence that contradicts their preexisting images of the world. Since strategic cultures are also important, Stein reasons that leaders who are products of “honor cultures” and who have a strong sense of grievance are especially likely to escalate in response to deterrent threats. When these leaders are also hedgehogs, the risk of escalation increases. 153 Third, human preferences toward risks (and the insights of prospect theory) are clearly applicable to deterrence situations. We know that leaders are likely to be risk averse with regard to prospective gains, but they are more acceptant of risks to protect what they already have and to reverse losses and recapture previous possessions. Stein suggests that Egyptian President Anwar Sadat never accommodated to the Israeli capture of the Sinai in the 1967 Six Day War and was highly motivated to regain a lost possession. Thus, he was undeterred by Israeli threats. Lebow and Stein have found that each party in a deterrence episode tends to see itself as a defender; thus, it is likely to see itself as a victim facing a loss relative to the status quo and therefore more willing to take risks. 154 We also know that how a deterrent situation is framed—as being within the realm of losses or gains—makes a difference. Deterrers pitted against challengers who frame the situation to be within the realm of losses are less likely to

The Dyadic Level of Analysis, Part III

361

be successful. 155 Not only does risk propensity vary with framing, it also varies due to individual personality. Morgan argues that deterrence is most likely to fail if both the challenger and the defender are risk acceptant, or if one is risk acceptant and the other is risk neutral. 156 Finally, we know that humans are subject to serious biases in attribution such as those caused by the fundamental attribution error. Stein suggests that current American and Israeli assessments attribute the Iranian nuclear program to aggressive intentions of Iranian leaders, to their radical fundamentalist ideology, or to their irrationality—all dispositional attitudes. They generally discount situational factors such as Iran’s encirclement by Western forces, the existence of Israeli nuclear weapons, or the American history of the use of force to create regime change against its adversaries in the region. 157 While either attribution may make deterrence or compellence difficult, situational attribution certainly creates the potential for more bargaining room. Just how widespread are these cognitive impairments in the real world of leaders facing deterrent situations? The answer is that they appear to be ubiquitous. Morgan summarizes the results of a wide variety of case studies of the Cold War period that deal in one way or another with situations of deterrence. 158 Many of the studies come to similar conclusions. Instances of nonrational behavior abound. Morgan’s summary paints a disturbing picture: In accounts of how governments fall into sharp confrontations and war there is recurring evidence that governments, elites, and leaders are often barely moved by general deterrence threats that they ought to take into account. Often they are driven by short-term thinking, not attuned to larger implications and potential consequences of what they are considering. They seem caught up in domestic political or ideological preoccupations or moved by wishful thinking or motivated bias. They are affected by serious misperceptions about the deterrer, or discount deterrence because they regard the deterrer and its efforts as illegitimate. They often start by wanting to attack if necessary to get their way, and then keep looking for ways to design around the opponent’s deterrence, which biases their assessments on what might work or when the odds have sufficiently improved. 159

Morgan’s list of findings from these studies indicates that virtually all of the decision-making pathologies identified in chapter 3 by cognitive psychologists crop up when scholars examine decision making by national leaders in deterrence situations. These studies show that national leaders don’t think like deterrence theorists; they make important decisions based on impulse and emotion; they are subject to motivated biases; they don’t update their options in light of new information and become attached to their old views; they are guilty of the fundamental attribution error and other cognitive heuristics and short-cuts; they engage in wishful thinking and defensive avoidance; and they use historical analogies uncritically. 160 The preeminent scholar in this realm, Janice Gross Stein, concludes as follows: The need for simplicity and consistency, impediments to probabilistic thinking, the predisposition to loss aversion, and framing effects are often treated as deviations from the rational model . . . Yet these deviations are so pervasive and so systematic that it is a mistake to consider rational models of deterrence as empirically valid. 161

362

Chapter 9

NUCLEAR WEAPONS AND DETERRENCE A good number of scholars would argue that since the atomic age started in 1945 the nature of deterrence has changed profoundly. At least for those states that possess nuclear weapons, deterrence now means nuclear deterrence. Put simply, the argument is that the devastating effects of the use of nuclear weapons in retaliation should greatly enhance the ability of nuclear weapons states to deter attacks upon their home territories and those of their allies. More specifically, nuclear deterrence against other nuclear-armed states is based on the possession of a secure second-strike force—the ability of one’s nuclear forces to survive a first strike by an adversary and still respond with a second strike that is capable of inflicting unacceptable damage on the initiator. If both states in a dyad possess such capabilities, deterrence is said to be stable: neither state has an incentive to strike the other because the result would be devastation for both. There are a number of empirical questions that flow from this situation. Do nuclear weapons actually prevent the use of force against its possessors (direct deterrence)? Does it deter the use of force against the possessor’s allies (extended deterrence)? Does nuclear deterrence apply to all potential challengers or just to those who also possess nuclear weapons? Does it prevent all types of conflict and violence or just all-out war? While the evidence is somewhat contradictory as always, some agreement seems possible. 162 First, is the possession of nuclear weapons effective in deterring military force initiated by other nuclear weapons states? We have empirical evidence from a number of studies that indicate that, compared with other pairs of states, dyads consisting of two nuclear weapons states (symmetrical dyads) are less likely to be involved in MIDs that escalate to full-scale war and that such dyads are less likely to be involved in crises that end in violence. Conversely, asymmetric dyads (one nuclear/one non-nuclear) and dyads without nuclear weapons are more likely to be involved in MIDs or crises that escalate to war. One study indicates that as the number of states possessing nuclear weapons in a crisis increases, the probability of war decreases. 163 While most would agree that possession of nuclear weapons vastly reduces the chances of war between two possessors, we also know that the probability is not zero. The Kargil War of 1999 was fought between two nuclear weapons states, India and Pakistan. However, that war was limited, localized, and conventional; indeed, it probably just barely met the definitional threshold for an interstate war of 1,000 battle deaths. It did not escalate to the use of widespread force, let alone nuclear weapons. Certainly in the almost seventy years since the dawn of the nuclear age, we can say that war between possessors of nuclear weapons has been rare, if not perhaps nonexistent. (We should note of course that wars are, themselves, rare events and the number of nuclear dyads is also relatively small—only eight current nuclear-armed states. Thus, war between two nuclear states would be a statistically rare event whether deterrence worked or not.) 164 Some scholars believe that nuclear weapons are irrelevant to deterrence. The argument is made most forcefully by John Mueller. 165 His view is that since World War II, the specter of full-scale conventional war among great powers has been enough by itself to deter war among major states. The debate here usually centers on why no World War III took place in Europe,

The Dyadic Level of Analysis, Part III

363

despite antagonism between the NATO and Warsaw Pact countries, or why the Cold War turned out to be the “Long Peace.” Mueller contends that whether or not Stalin (or any Soviet leader) planned to make further territorial grabs in Europe, the presence of nuclear weapons was not necessary to dissuade him from that goal. Having gone through the horrible experience of conventional warfare in the 1940s, no one wanted to risk engaging in this type of conflict again. Morgan observes that, ultimately, we can’t really know how much nuclear deterrence contributed to the absence of war. He thinks that the major contribution of nuclear weapons is that they eliminated the ability of major states to count on winning cheap victories, which after all is what has historically made war seem worthwhile. While conventional war after 1945 probably did this too, nuclear deterrence was pretty much the icing on the cake. 166 Second, does the possession of nuclear weapons deter lesser conflicts between nucleararmed states? Here the empirical evidence is not very encouraging. While the statistical probability of war is comparatively low for nuclear dyads, it would appear that joint possession of nuclear weapons does not prevent nuclear states from becoming involved in crises or militarized disputes with each other. Once again, quite a few studies using different data collections on MIDs and crises agree that while nuclear possessors are unlikely to go to war, joint possession does not prevent nuclear powers from initiating crises or militarized disputes with another nuclear weapons state. 167 Actually, nuclear dyads appear to be engaged in an unexpectedly large number of crises, and once engaged in crises and militarized disputes, they seem unusually likely to escalate these interactions to higher levels of threat and intensity— short of war. 168 Compared with non-nuclear pairs and to asymmetric pairs, joint nuclear dyads actually have a higher probability of escalation short of war. As Daniel Geller notes, these findings are consistent with three propositions regarding nuclear deterrence. 169 The first is Snyder and Diesing’s contention that crises between the great powers have become a surrogate for war. The threat of nuclear devastation means that major powers must avoid the final step in the escalation chain. Participation in international crises becomes the atomic age substitute for war, but deterrence ultimately means that crises will not escalate to war. 170 The second proposition is Glenn Snyder’s famous stability-instability paradox. 171 This is the idea that because deterrent relationships are stable at the level of all-out nuclear war (general deterrence), the possibility of limited wars is actually increased. The presence of offsetting nuclear deterrents liberates states to pursue lesser mischief. Nuclear-armed states may seek to exploit military vulnerabilities at lower levels of violence in order to undermine the opponent’s resolve, to exert pressure, and to win certain concessions. Thus, lower levels of force—including threats, demonstrations of force, limited use of military force or intervention—may be used as a form of coercive bargaining since initiators believe their actions would not be likely to trigger a nuclear response by the opponent. For many, the Kargil War is a perfect illustration of this. 172 Another prominent proposition is often called the nuclear taboo. This is the idea that nuclear states are subject to a kind of limited self-deterrence—a concept rarely discussed in the deterrence literature. The argument is that since 1945 an international norm or tradition has developed that nuclear weapons should exist only as a deterrent and should never be used, especially against non-nuclear states. 173 In spite of the fact that no international treaty to this

364

Chapter 9

effect exists, nuclear-armed states have in fact refrained from using these weapons for over six decades. The argument is that the existence of such a norm acts as an important restraint on decision-makers. While authors have slightly different takes on how this taboo developed and how strong it is, there is also a good deal of agreement. The norm developed gradually in an evolutionary way after 1945. Nina Tannenwald emphasizes that the idea of nuclear non-use has been socially constructed (she uses a constructivist approach). The norm emerged over time and was strengthened as it was pushed forward by domestic and transnational groups and by important individual leaders (norm entrepreneurs), especially in democratic states. It was also strengthened by iterated behavior over time as states consistently refrained from the use of nuclear weapons in situations in which one might have expected them to be employed. Each non-use strengthened the emerging norm and placed more pressure on states that might be the first to break the taboo. 174 T. V. Paul melds this constructivist explanation with one that adds the importance of realpolitik material factors. He argues that the norm has a strong material basis; nuclear weapons are fundamentally different from other types of weapons. Their destructive consequences are unique, and this was widely understood by the 1950s. Thus leaders on both sides of the Cold War divide eventually came to see that there were serious practical limits to what might be accomplished by their use. The nuclear taboo has been the result of states’ selfinterests and the logic of consequences of nuclear use. Most important for Paul, the use of such weapons would create reputational costs. Leaders have rationally calculated that nuclear use may have disastrous effects on their reputation and image, especially with allies and nonaligned states whose support they need. (Reputational costs would include moral and legal considerations, consistent with the constructivist approach.) Paul’s analysis of historical situations in which nuclear use has been considered indicates that restraint is not due simply to tactical and strategic military factors; reputational concerns were important factors in the discussions of political and military leaders. 175 Of course, non-nuclear weapons states understand that such a normative prohibition exists and that nuclear states will be extremely hesitant to use the ultimate weapon. Thus, nonnuclear challengers could calculate that they are unlikely to be the target of a nuclear strike, especially if it is clear that the challenge is not an existential threat to the nuclear state, but is more limited. Simply put: The existence of a nuclear taboo may severely decrease the credibility of deterrent threats between nuclear and non-nuclear states by reducing the challenger’s calculation of the deterrer’s will or intent to carry out a particular retaliatory threat. This was most likely the perception of Egyptian leaders when they initiated the Yom Kippur War in 1973 against nuclear-armed Israel. Israeli nuclear weapons appeared to have little relevance for Egypt. In Paul’s view, as long as Egypt’s strategy was clearly limited to retaking the Sinai and not posing an existential threat to Israel, Egyptian leaders could be confident that Israel would not use nuclear weapons. 176 This is a good example of a strongly motivated challenger who attempts to “design around” the deterrent threats of its opponent by developing what it thinks—correctly or not—are relatively low-risk strategies. 177 Our third question follows from the previous discussion: Does nuclear deterrence help nuclear states prevent attacks from non-nuclear states? These asymmetric dyads appear to pose problems for deterrence theory. As we have already seen, states possessing nuclear

The Dyadic Level of Analysis, Part III

365

weapons do not appear to benefit from any noticeable advantage in their relationships with non-nuclear states. The biggest clue here is that nuclear states have been attacked several times since 1945 by non-nuclear powers: China’s intervention in the Korean War in 1950, the Egyptian-Syrian attack on Israel in 1973, Argentina’s initiation of the Falkland Island War in 1982, and Iraq’s Scud missile attacks on Israel in the Persian Gulf War of 1991. Clearly the possession of nuclear weapons has not deterred the initiation of conventional war by nonnuclear states. One may wish to add to the list the attack by Al Qaeda, a non-state actor, on the United States in 2001. Subnational groups have also attacked Israel, Britain, Russia, Pakistan, and India—all members of the nuclear club. Statistical studies back up impressionistic evidence from case studies. Several empirical analyses look at interactions of nuclear/non-nuclear pairs and come to conclusions that are contrary to predictions of deterrence theory. Rauchhaus finds that compared with disputes between symmetric dyads, disputes involving asymmetric nuclear dyads are more dangerous. Asymmetric pairs are linked statistically to higher chances of crisis involvement, use of force, and war. 178 Likewise, Geller’s analysis of 393 serious militarized disputes from 1946 to 1976 found that possession of nuclear weapons did not at all serve to dampen escalation. His astonishing finding is that the non-nuclear state—whether the initiator or the target in the interaction—was likely to act more aggressively than the nuclear state in the dispute. 179 Studies by Organski and Kugler of fourteen deterrence cases involving nuclear states (seven of which involved asymmetric pairs) and a later study by Kugler of fourteen cases of extreme crises (involving nine asymmetric pairs) both indicate that non-nuclear states are likely to prevail in asymmetric confrontations. 180 Kugler’s analysis is that in all nine asymmetric crises, the outcome favored the non-nuclear challenger. (This is not the same thing as deterrence failure; it doesn’t necessarily mean that nuclear states were unable to deter an actual attack.) The explanation for non-nuclear states’ ability to win such confrontations appears to be connected to their conventional military superiority at the site of the dispute. Geller summarizes: “In asymmetric dyads, the possession of nuclear arms provides no discernible advantage in crises and confrontations. Escalation by the non-nuclear adversary and its use of force against its nuclear-armed opponent—including war—are distinct outcomes with surprisingly high probabilities.” 181 It is also worthwhile mentioning that when nuclear-armed states are the initiators against non-nuclear states, they have, to this point, always refrained from using these weapons even though they faced military defeat of some sort. One thinks here of the Israeli involvement in Lebanon in 1982 and the Soviet intervention in Afghanistan from 1979 to 1989, as well as American involvement in Vietnam. T. V. Paul’s analysis is that the normative taboo against the use of nuclear weapons played some role in decisions for non-use. 182 Finally, does nuclear possession help deter the use of force against the deterrer’s allies in extended deterrence? With some minor exceptions, empirical studies indicate no. The studies cited earlier showing the large number of failures in extended deterrence also indicate that nuclear weapons are almost irrelevant to the question of whether attempts at extended deterrence will be successful or not. What appears to be more important is the combined local conventional military superiority of the defender and its ally/protégé. 183

366

Chapter 9

DETERRENCE: CONCLUSION What can we conclude about deterrence as a way of preventing war? Here are a few modest comments. First, we know that deterrence is hardly foolproof: sometimes it works; sometimes it fails. The difficulty is determining when and how deterrence fails. Second, when it works—the absence, for instance, of a Soviet attack on Central Europe during the Cold War and the absence of a “Second Korean War”—one can always question why. Did a real threat of aggression exist? And if it did, was it the threat of retaliation responsible for the lack of the attack or was it some other factor? 184 Third, wars occur rather frequently even when the target has been diligently applying deterrent strategies to prevent it. Using deterrence theory to devise successful deterrence policies is clearly problematic. When deterrence fails, it frequently fails in ways that are not consistent with deterrence theory. For instance, having sufficient military capability to carry out the threat is important but appears to be salient only if the defender and its protégé have immediate local superiority at the site of the potential attack. Even then, it is no guarantee of deterrent success. Fourth, there are limits to what nuclear weapons can add to the strength of one’s deterrent effort. While they probably play a role in preventing all-out nuclear war between two nucleararmed states, beyond that their utility is questionable. They cannot reliably deter an adversary from using conventional military force against nuclear-armed states or their friends and allies, and they cannot be used to compel a change in the status quo. Fifth, the issues of credibility and reputation are problematic, particularly in extended deterrence. Deterrence theory argues that credibility is important if deterrence is to succeed. Explicit verbal commitment and paper guarantees aren’t enough; the defender must demonstrate his commitment in more overt ways, especially through visible political, military, and economic ties to the pawn. Formal treaty commitments help somewhat, but only if they are conditional deterrent treaties. Signals that have high audience cost would appear somewhat effective in enhancing credibility. But there does not seem to be any magic formula for successfully establishing credibility in the eyes of the potential challenger. Deterrence theory suggests that successful deterrence requires a state to establish a reputation for making good on its commitments. Certainly, the commitment to back up one’s threat must not appear ambiguous or uncertain to others. But even if credibility is apparent, deterrence may still fail. Challengers don’t always pay attention to a deterrer’s reputation, and countries rarely get reputations for lacking resolve. To the extent that reputations for resolve have an impact, they are derived from mutual interactions with the same opponent in the same geographic region. But attempts to gain a reputation for upholding commitments that involve hardline or bullying behavior in deterrent encounters with the same opponent are just as likely to contribute to a future challenge as to a further submission. Sixth, it is possible that many of the actions designed to enhance deterrence create the risk of provocation and escalation to unintended war. Our examination of the spiral theory and Vasquez’s steps to war theory shows that the security dilemma and the problem of conflict spirals apply to deterrence (and compellence) situations. Leaders of deterring states must

The Dyadic Level of Analysis, Part III

367

balance the need for enhanced credibility against the need to avoid inciting fear and provoking an overly hostile reaction from the opponent. (Over-reactions may be due to psychological and emotional pressures on individual leaders or to domestic political pressures.) Our examination of game theory warns us that although tough, noncooperative strategies designed to enhance credibility may be successful, they are also fraught with danger. Iterated chicken games between equal players may turn into conflict spirals as each player tries to demonstrate credibility to the other. Many critics of deterrence theory argue that firm-but-flexible strategies that combine deterrent threats with limited inducements or with reassurance are more helpful in preventing escalation to war than the use of threats alone. Seventh, even when countries establish both sufficient physical capabilities and a solid commitment to make good on the threat to retaliate, deterrence may still fail, because opponents are likely to engage in limited probes and serial attempts to “design around” the deterrer’s commitments and engage in low-risk strategies. Deterrence is skimpy at the margins and difficult to enforce when the challenges come at the low levels. Ultimately deterrence may fail for reasons that deterrence theory does not take into consideration. Leaders may see high costs involved in challenging a deterrer’s commitment, but they may also believe that the domestic costs of inaction are too high. This makes deterrence based on threats both inadequate and unlikely to succeed. 185 Sadat’s decision to initiate war against Israel in 1973 and the Japanese decision to attack Pearl Harbor in 1941 are often cited as examples of this kind of situation. Leaders who decide to challenge deterrent threats are usually highly motivated due to their strong dissatisfaction with the international status quo or due to their internal political vulnerabilities. But they may also be responding based on their own peculiar inner demons. Eighth, while deterrence theory lives at the dyadic level, important factors at other levels of analysis have a profound effect on its application. Deterrence is in part a psychological concept: its success depends substantially on the target’s perception of the threats that have been made against it and the credibility of those threats and the capabilities that back them up. Perceptions often count more than reality. This lands us squarely in the individual level of analysis. Threats may have different effects on different individuals, depending on their personalities. In fact, threats may provoke extreme emotional and nonrational responses in some individuals. If this is so, threats may not be compatible with prompting governments to remain sensible and rational. We also know that crisis situations tend to fray at the edges of human rationality. And, of course, we have frequently referred to the pervasive inability of individuals to correctly perceive the attitudes and actions of others. Unfortunately, the combined effect of these phenomena must surely be that deterrent threats may not be received as intended and may not induce the completely logical responses desired. Robert Jervis makes this summation of attempted deterrence: “in almost no interactions do two adversaries understand each other’s goals, fears, means-ends beliefs and perceptions.” 186 Under such circumstances, the probabilities for the success of deterrence are relatively small. Deterrence theory assumes a rationality that may be lacking in real-world situations. In many cases the defender fulfills all the essential conditions posited by the theory and deterrence still fails. This can be accounted for by individual-level factors such as misperception, the impact of emotions, personality attributes, motivated biases, and cognitive short-cuts asso-

368

Chapter 9

ciated with the operation of system 1. All of these factors may distort crucial information about the cost/benefit ratio, the defender’s commitment, or the likely outcome of a confrontation. A valid theory of deterrence must therefore take irrationality into account. 187 This criticism is also one that might be leveled at game theory and bargaining theory as well, given their attachment to the assumption of rationality. This criticism is not necessarily fatal. There is rationality and then there is rationality. Certainly we find wide variation in what social scientists mean by the term. Morgan argues that deterrence theory does not really need to be based on rationality. He believes deterrence theorists should replace the assumption of rationality with the notion that sometimes (but not always) leaders possess “sufficient reasonableness.” 188 He suggests that any new-and-improved version of deterrence theory should take into account the criticisms of cognitive psychologists. 189 For instance: • Deterrence is more likely to succeed when the challenger is motivated primarily by prospective gains and not fear of losses. (This is consistent with prospect theory.) • Deterrence is more likely to fail if both actors see themselves as aggrieved defenders, and deterrence is more likely to work when the challenger is not seeking to redress losses. • The strength of a challenger’s sense of injury, desperation, or other motivation is very important in determining deterrent success or failure. Ninth, since deterrent threats are intended to affect government decision makers, smallgroup decision processes come into play as well as factors at the individual level of analysis. What might be the effect of a deterrent threat on national leaders under the sway of groupthink? Morgan suggests that the operation of groupthink does not bode well for the success of deterrence. The receiving group would be operating under the illusions of invulnerability and moral superiority, they would be overly optimistic in their assessment of their ability to manage the conflict through brinkmanship, and they might have a tendency to take risks they would ordinarily shun as individuals. 190 One ought to be aware also of the effects of bureaucratic politics on the receptivity of the government to threats. Threats normally strengthen the influence of those who take a hard line against the threateners. And, as we have seen, domestic political considerations play an important role in a potential attacker’s calculations. Tenth, deterrence theory also needs to allow for the insights from the constructivist approach. It is helpful to understand that deterrence and compellence are both social constructions. Deterrence is affected by socially constructed ideas such as rationality, threats, commitments, and resolve; and the success of deterrent strategy depends on the intersubjective understandings that two states develop over time regarding shared norms and practices regarding deterrence and compellence efforts. It also helps to understand that deterrence encompasses a process of learning over time—a process that surely helped to make Soviet-American deterrence more stable and routine over time. It is also useful to understand that obstacles to learning—strategic cultures, bureaucratic inertia, domestic politics, and so on—may interfere with the successful application of a deterrent strategy. 191 American practitioners of coercive diplomacy would probably be wide of the mark if they assumed that Iranian leaders understood our deterrent and compellent actions in the same way that we do.

The Dyadic Level of Analysis, Part III

369

Finally, deterrence is also affected by factors at higher levels of analysis such as the nature of the international system. One analysis of Russett’s study of deterrence points out that most of the cases of successful deterrence in Russett’s study are to be found after World War II. The fact that seven of the eleven cases of attack occurred before 1940 suggests that the nature of the international system may be an important factor. 192 World War II marks a watershed. At the war’s end the international system had made the transition from multipolarity to one characterized by tight bipolarity. As a result, deterrence became bipolar as well: The two superpowers became committed to backing up their deterrent threats against the other, and both possessed nuclear weapons so that deterrence became nuclear deterrence. The status quo had become more clearly delineated compared to the intricacies of the old multipolar past, making red lines more discernible and presumably making deterrence easier. Of course, the bipolar system of the Cold War is now history; deterrence exists in a new environment occupied not only by traditional states, but by rogue states and important nonstate actors like Al Qaeda. This makes deterrence more complex. 193 One of the significant changes in the post–Cold War world is that among the current great powers, immediate deterrence events are exceedingly rare; general deterrence prevails, but it is normally seen by scholars as “recessed” and not very prominent. Deterrence is no longer the central security principle in great-power relations. 194 The lead stories in our twenty-four hour news cycle are no longer related to superpower crises, superpower threats and counter-threats, missile gaps that threaten to undermine our deterrent second strike capability, or news from a superpower proxy war in Asia or Africa or the Middle East. Instead, we find much more complicated and often asymmetric attempts that combine deterrence and compellence attempts, involving states (with or without nuclear weapons), rogue states (like Iran or North Korea), collective actors (NATO and the UN), and nonstate actors (like Hezbollah and Al Qaeda). With this last observation, it is appropriate that we turn our attention to the nature of global politics as a whole. In the next two chapters we devote our attention to the fifth and final level of analysis, the international system.

Chapter Ten

The International System Level of Analysis, Part I Realism, Anarchy, and the Balance of Power

The trouble with the balance of power is not that it has no meaning, but that it has too many meanings. —Inis L. Claude Jr.

We have arrived at the last level of analysis, the international system. Here we shift our focus to the big picture—the way international politics is structured and to the way states interact as part of a global system. Attempts to build systemic-level theories assume that there is in fact some order to the world—that events that appear to be random and unconnected are in reality part of a large-scale system of “organized complexity.” 1 The behaviors of states are seen as containing meaningful similarities and recognizable sets of interaction patterns. In other words, they comprise an international system. Definitions of systems are notoriously vague, but for starters let’s just say that a system is any set of variables in interaction with each other. This interaction is patterned, recurring, and interdependent. Changes in one part of the system affect other parts of the system. In addition to possessing certain units that have a patterned relationship, all systems also have certain rules or norms, definable boundaries, an identifiable organization and structure, and a set of inputs and outputs. The units that make up the international systems are assumed to be primarily sovereign states and organizations of states (such as alliances, trading blocs, and political organizations like the United Nations). But there are nonstate actors as well—multinational corporations, transnational interest groups, even transnational terrorist groups. The interactions between the major actors in the system have developed a regularity over time that conforms to established international legal norms, as well as to certain less formal norms that have developed through custom. Every system has boundaries, which distinguish the system from its external environment. The international system has gradually become a true global system, however—leaving theorists in the uncomfortable position of being unable to find a true external environment to which the system responds. 371

372

Chapter 10

The international system has a variety of structures. One essential aspect of the system’s structure is its lack of an authoritative political organization with the ability to issue and implement commands. In other words, the international system is characterized by anarchy; it is a political system that lacks a government. Paradoxically, we must try to envision a system that has order but has no “orderer.” 2 It does possess a certain degree of stratification and hierarchy, however—such as the international distribution of political and social status. Arguably the most important aspect of the system’s structure is the distribution of military and economic capability among its constituent units. In fact, the variation in the distribution of capabilities is, for many scholars, the defining characteristic of different types of international systems—unipolar, bipolar, tripolar, multipolar. Each system is also characterized by a set of inputs and outputs. Various inputs (or stimuli) from the system’s own internal environment or from its external environment may disrupt the stability or equilibrium of the system. The system then responds with outputs (actions) aimed at restoring a stable balance. Outputs of the system may reenter the system as inputs through the process called feedback, creating a continuous process. One may view the international system as being in the midst of a continuous balancing act; every system presumably responds to disturbances in ways that establish and preserve its essential structure. War may be seen as a way of maintaining the system’s present structure or as the means through which the present structure is destroyed or transformed so drastically that an entirely new system is created. 3 This was much the situation in 1945, when the defeat of the Axis coalition reduced a multipolar system to one dominated by only two superpowers. On the other hand a system may be transformed without war, as was the Cold War bipolar system during the crucial years of 1989–91. Most systems contain within them smaller subsystems. Systems may be thought of as fitting inside each other like a set of Chinese boxes, a concept that Robert North refers to as “nesting.” 4 The global system is comprised of numerous issue-oriented or regional subsystems made up of several states and other actors; each state contains within it a large number of organizational and decision-making subsystems; each decision-making group contains several individuals; each individual is a collection of interacting subsystems—physical, emotional, and so on. At the systemic level of analysis the central assumption is that the structure of the international system plays the most important role in determining the behavior of states. 5 The nature of the state or its leaders is relatively unimportant, and the behavior of states is determined primarily by their positions in the international system or regional subsystem. 6 Different states within the same international system will behave similarly despite differing national attributes. 7 The nature of the international system itself and the state’s position within the system place certain limits or constraints on the behavior of states and compel or dispose states toward certain activities. Evan Luard suggests that the international system molds the character and behavior of its constituent units so that they become similar in their motives, means, roles, norms, institutions, and even internal structures. Perhaps most important, international “society itself teaches conflict to its members: tells them when to conflict, with whom and over what issues.” 8

The International System Level of Analysis, Part I

373

War is typically explained as being linked to a particular kind of international system, brought about by a specific distribution of certain values within the system (military power, economic power, political prestige) or by a state’s position within this overall structure. As you might imagine, there is considerable contention over which kind of international system is the most prone to war. Let’s start our examination of the international system level by exploring the issue of anarchy. INTERNATIONAL ANARCHY The French philosopher Jean-Jacques Rousseau used the now-famous stag hunt analogy to explore the social and political implications of anarchy. The analogy is as follows. Five men all suffer from hunger. Their hunger can be easily satisfied by the fifth of a stag, so they agree to cooperate to trap the unfortunate creature. The fly in the ointment is that the hunger of any single individual can be satisfied by a hare, so as a hare approaches, one of the men chases it down and kills it. He thus satisfies his own hunger, but permits the stag to go free and his erstwhile friends to go hungry. 9 What does the analogy illustrate? First, this is a situation in which the immediate interest of the single individual prevails over the common interests of the group. What the hare snatcher did was good for himself and made a kind of rational sense. Although reason tells him that in the long run developing cooperative working relations with his neighbors is necessary, nevertheless in the short run if someone else were to get the hare, he would have gone hungry instead. 10 States pursue their own national interests and they often do this at the expense of other states and at the expense of the interest of the international community. As you may have guessed from this discussion, the stag hunt analogy is somewhat similar to a prisoner’s dilemma. States have opportunities for mutual gain, but also incentives to defect and pursue individual gains. The difference is that in a prisoner’s dilemma the best outcome for both sides is to defect while the other cooperates, while in the stag hunt the best outcome is derived from mutual cooperation. 11 Second, the international system is made up of sovereign, independent states pursuing their own interests with no world government compelling them to cooperate in the pursuit of common goals. Because any state may refuse to cooperate, states must be constantly ready to engage in measures of self-help. And, to extend the analogy, because any state may use force, all states must be ready to use force in return. As Kenneth Waltz reminds us, in “the state of nature” one man (and, by analogy, one state) cannot begin to behave decently unless he has some assurance that others will not be able to ruin him. 12 The absence of a world government that might enforce rules is a primary structural element of the international system. According to some theorists, the lack of world government means that international society is like a Hobbesian world in which there is a war of “every man against every man.” International anarchy leads to constant suspicion, insecurity, conflict, and violence. Rousseau’s argument is essentially that wars occur because there is nothing to prevent them. 13 The violence within the system doesn’t depend on the nature of states themselves but on the nature of the international system:

374

Chapter 10 It is quite true that it would be much better for all men to remain always at peace. But so long as there is no security for this, everyone, having no guarantee that he can avoid war, is anxious to begin it at the moment which suits his own interest and so forestall a neighbor . . . so that many wars, even offensive wars, are rather in the nature of unjust precautions for the protection of the assailant’s own possessions than a device for seizing those of others. 14

Robert Jervis has wisely noticed that two separate arguments about the nature of humankind are both consistent with the concept of international anarchy, but they lead the theoretician along totally different paths. 15 On the one hand, humans (and therefore states) may be thought of as evil, power-seeking, or motivated by maximizing individual gains. In this case anarchy breeds aggressors and permits them leeway to act in ways injurious to others. Wars result from a failure of a global government to deter or restrain the aggressive actions of states. This first path denotes the familiar deterrence model. Alternatively, humans (and therefore states) may be assumed to be peaceful and cooperative by nature. In this case the lack of world government to enforce cooperation among states leads to aggression out of unwarranted fear that one’s rivals won’t be restrained. International anarchy leads to a tragedy in which war occurs despite the purely defensive intentions of all those involved. This second path is an expanded security dilemma or conflict spiral model. The point is that whichever path one starts down, an anarchic international system still leads to war. In either case the solution would appear to be the creation of some kind of supranational world government with real power to restrain the actions of its constituent members and to compel cooperation. If the problem is the absence of government, then the logical solution would appear to be the creation of government on a global scale. 16 The foregoing discussion has treated international anarchy as a constant condition in international relations. But war itself is not constant, so we need to find variables in the structure of the international system to explain that. As a way out of this dilemma, theorists have suggested that either the international system is not truly anarchic or that the degree of anarchy varies considerably over time. 17 Several aspects of the structure of the intentional system have received close attention. International relations scholars have leveled entire forests turning out academic papers on the distribution of power within the system. Perhaps the best way to start the discussion of those topics is to examine the realist approach toward international relations, since realists are concerned with international system level phenomena and in particular with the distribution (or balance) of power within the system. CLASSICAL REALISM For many decades the dominant approach in international relations has been realism (or realpolitik). Realism is probably best thought of as not a single theory of international relations but as a family of theories. (Some would say a dysfunctional family.) Its adherents have now divided and subdivided themselves into several clans. In the broadest sense, realism is the basis for many of the theories discussed in the last several chapters of this book, including deterrence theory, balance of power theory, theories of system polarity and alliance polarization, offense-defense theory, hegemonic stability theory, long-cycle theory, dynamic differen-

The International System Level of Analysis, Part I

375

tials theory, and power cycle theory. However, realists are interested not only in empirical theories but normative, policy-oriented prescriptions. They write books offering advice to presidents and prime ministers about how to best pursue foreign policy. Some, like Henry Kissinger, have been both academics and policy makers (as the president’s national security adviser and then as secretary of state). Modern realists claim writers such as Thucydides, Machiavelli, Hobbes, and Clausewitz as their intellectual godfathers. In the post–World War II era in the United States, as realism came to dominate the academic discipline of international relations, the dominant scholarly voices were E. H. Carr, Hans Morgenthau, Arnold Wolfers, and Henry Kissinger. Among prominent realpolitik practitioners were Harry Truman, Richard Nixon, John Kennedy, and George F. Kennan (a U.S. diplomat and father of the containment doctrine), as well as Joseph Stalin, Charles DeGaulle, and Winston Churchill. These writers and statesmen are normally categorized as classical realists. 18 Classical realists are bound together (more or less) by several basic assumptions about international politics. First, the international system is anarchic. Second, states are the primary actors in international relations. However, not all states are equal in importance, and classical realists typically concentrate on the so-called great powers. Third, states may be thought of as unitary actors. (I’ve refrained from using the term “unitary rational actors,” since there is some dispute over whether classical realists assume states act rationally.) 19 Regardless of the nature of their particular leaders at a particular time, states pursue policies to promote their national interests. Fourth, the most important factors in the analysis of international politics are power and the distribution of power. Thucydides’ position in the pantheon of realism’s founding fathers is due to the fact that in his History of the Peloponnesian Wars, he identifies in the very first sentence of his gargantuan book the root cause of the war between the Athenian alliance and the Spartan alliance: the growing power of Athens and the fear this caused on the part of Sparta. Fifth, realists assume that power can be measured and calculated. Power includes a state’s military capabilities but also its economic power, geographic position, natural resources, and population. These national capabilities set limits and constraints on a country’s foreign policies. According to realists, states seek to either maximize their power or—at the very least— maintain it (different realist clans emphasize one over the other). Power can be increased through internal means (building up one’s economy and military might) or through external means (creating alliances and undermining potential alliance ties of one’s opponents). For realists, international politics is a struggle for power. 20 Why does this struggle for power exist? Realists give two answers. Classical realists like Morgenthau say that humans are selfish and power-seeking and that, since states are run by humans, states will also seek power. Early realists talked about humans’ “will to power” or urge to dominate. This is a rather pessimistic view of international relations; because one cannot change human nature, there will be no harmony of interests between states. The second answer given by realists is that the struggle for power is due to the anarchic nature of the international system. Because there is no world government that can prevent states from taking advantage of each other, anarchies are “self-help” systems. Sovereign states must look out for their own interests against other unrestrained, sovereign states that are

376

Chapter 10

looking out for their own interests. In this type of international system, it is logical for states to increase their power; this is a defensive response to the condition of anarchy. A final conceptual distinction has also been important to realists—the difference between the goals held by different kinds of states: status quo states and revisionist states. 21 Status quo states are satisfied major powers who wish to maintain the nature of the current global distribution of power, rights, and status. As such, their goal is to maximize security rather than power. Revisionist states are dissatisfied major powers who seek a revision of the rules of the current system as well as a redistribution of power and status in the system. Historically, they have attempted to achieve their revisionist goals through imperial expansion. This in turn requires that they maximize power rather than security; indeed, sometimes power is maximized to the detriment of security. France during the Napoleonic era is the classic example. 22 While a system with strong revisionist states is prone to war, a system without revisionist states is relatively stable and may even create the conditions for a “concert” of great powers who agree to periodic consultations in order to harmonize their policies. 23 Realists of all stripes would agree that the self-help system creates security dilemmas. States see themselves as insecure and they take actions to increase their security; they build up their military machinery and they create alliances. Other states are likely to see this as threatening to their own security; they feel less secure and take actions on their own to increase their security. This creates a spiral of acts intended to increase security, but the actions bring states on all sides closer to war. For classical realists, war is always a possibility. Indeed, it is very much a part of the statesman’s tool chest—one of the many instruments that can be used to balance power and pursue national interests. Given the condition of anarchy, the nature of human beings, the uncertainty about others’ motives, the existence of security dilemmas, the presence of dissatisfied revolutionary states, the difficulty of balancing against complex power configurations, and the incentive that states have to use force on their own behalf, wars— while not inevitable—are seen as pervasive and endemic to the system. Based on these assumptions, classical realists also put forward a number of policy prescriptions. The primary obligation of state leaders is to promote the state’s national interest, primarily the survival of the state itself, along with its territory, its people, and its sovereignty. This requires a realistic assessment of the state’s interests and the interests of others, as well as its physical capabilities and those of others. Goals must fit one’s capabilities, and states must refrain from the pursuit of non-essential goals. A realist policy thus requires prudence, pragmatism, and empathy for the needs and interests of others. It is essentially a conservative approach. Classical realists counsel that the best way to manage international relations and conflict is through a balance-of-power policy. A balance-of-power policy seeks to preserve the stability of the international system by maintaining a more or less equal balance of power between major states and their alliances. Parity (equality) is seen as good; it keeps the peace by deterring would-be aggressors. Preponderance of power is bad; it undermines deterrence and tempts those who have it to go to war with weaker states. Therefore, the essential rule of a balance-of-power strategy is to support the weak against the strong: States should balance against the more powerful actors in the system rather than “bandwagon” with them. When states threaten to become too powerful, other states should take measures to balance against

The International System Level of Analysis, Part I

377

them. Classical realists argue that alliance strategies must be flexible, short term, and nonideological. Finally, classical realists see international politics as something of an ethics-free zone: Statesmen should not let themselves be bound by the same moral and ethical standards that apply to individuals in their relations with others. (This is one reason why Machiavelli is claimed as one of the founding fathers of realism.) In pursuit of the national interest, state leaders may have to do things that are immoral or unethical by individual standards. Statesmen may have to lie, cheat, steal, and murder, but the interests of the state override individual morality. States have no right to pursue ethical policies if the result is the demise of the state. And while treaties, international law, diplomacy and negotiations, and international organizations and international cooperation are all good things, realists have limited faith in their ability to preserve the interests of the state. Classical realist thought in the United States tended to be dominated by historian-statesmen (like Kissinger, Carr, Kennan) or traditionally trained political scientists unfamiliar with quantitative empirical research (like Morgenthau). But in the last several decades academics intent on bringing greater theoretical and scientific rigor to the study of international politics, though still attached to the importance of power and power relationships, have referred to themselves as neorealists. It is to this group—a group that is itself fractured into several sub-groups—that we now turn. NEOREALISM With the publication of Kenneth Waltz’s Theory of International Politics in 1979, the neorealist era officially began. 24 While neorealists hardly make up a united front, they are united by their belief that the anarchic structure of the international structure is the key to understanding international politics; for this reason they are called structural realists. They believe that conflict occurs not because of human nature or the nature of states but because the anarchic system structure compels states to struggle for survival. Anarchy is not just a permissive condition, but a causal force that constrains and shapes the actions of states. The fear of conquest and extinction compel them to undertake self-help measures to increase their security, hence the creation of security dilemmas Although all international systems thus far have been anarchic, the exact distribution of capabilities among players varies over time. Systems may be unipolar, bipolar, tripolar, or multipolar depending on how capabilities are distributed among member states, but all of these systems are balance of power systems. The balance just takes different forms and displays somewhat different dynamics. (Lengthy periods of unipolarity/hegemony are believed to be unlikely due to the tendency to balance against power.)As the system structure changes, so do the interaction patterns among states. The system’s structure determines the behavior of its units. A primary outcome of international anarchy for neorealists is the tendency of units to form balances of power. 25 Given the structure of international anarchy and the relative power of the units, states are virtually compelled to act with balance-of-power strategies in order to prevent a preponderance of power that might lead to aggression. 26 It is essentially a deterrent system through which major wars are prevented through countervailing power. For both neorealists

378

Chapter 10

and classical realists, it is the equality of balance that keeps aggression in check. 27 Preponderance in the hands of a single state or a single bloc is dangerous. As Inis Claude has said, “There is danger when power confronts power, but there is a greater danger when power confronts weakness.” 28 Realists generally expect strong powers to be aggressors, and the only thing that prevents this is the possibility of defeat. The construction of countervailing alliances means that would-be aggressors cannot expect to achieve gains at low cost. 29 Equality of capabilities is presumed to be effective in deterring war at least between the major powers, though it may not produce peace between major powers and lesser states. 30 Note that Waltz’s own balance of power theory is at best a general theory of war that seeks only to explain the recurrence of war over time; it does not attempt to explain specific outbreaks of war. Anarchy means that war is normal. Any given war is explained not by looking at the structure of the international-political system but by looking at the particularities within it: the situations, the characters, and the interactions of states. 31 What remains is how to explain which structural characteristics of the system are most likely to affect the frequency of war. For Waltz, the determining characteristic is the distribution of systemic power, specifically the number of great powers in the system. Finally, Waltz argues that the differing characteristics of specific states and their individual leaders matters little. States are compelled to balance power against power; it is not up to the wisdom of individual statesmen (as Kissinger suggests). Statesmen have to obey the laws of political physics. 32 Nor is the nature of the state important. Waltz does not see states as differentiated in any important way: anarchy produces homogeneous units. 33 All states will behave similarly in similar international systems. This argument has become extremely problematic for neorealists as recent research has supported the existence of a “democratic peace,” which indicates that state characteristics do matter; democratic states act differently than nondemocratic states in the same international system. Neorealists have divided into two main camps—offensive realists and defensive realists. The primary difference between them concerns the reputed effect of anarchy on states (and therefore on the system). DEFENSIVE REALISM Defensive realists—whose ranks would include Robert Jervis, Steven Van Evera, Jack Snyder, Barry Posen, Charles Glaser, and Stephen Walt—largely follow the lead of Kenneth Waltz in arguing that anarchy encourages states to behave defensively. 34 Waltz argues that “the ultimate concern of states is not for power but for security.” 35 States seek primarily to maintain their own power as a way of ensuring their security, not to increase their power through conquest. Put differently, they attempt to maximize relative security, not relative power. 36 Moreover, while it is dangerous for states to have too little power, an excess of power is also dangerous, as it provokes others to increase their own arms. Consistent with the propositions associated with prospect theory, defensive realists lean toward seeing states as risk-averse actors primarily concerned with protecting the status quo and minimizing potential losses rather than maximizing potential gains. 37

The International System Level of Analysis, Part I

379

How and why do wars occur if states seek only their own security and are defensiveminded and if attempts at expansion are generally self-defeating (due to balancing)? Defensive realists offer several answers. One answer has to do with a particular aspect of the structure of the international system—the offense-defense balance. The argument is that the security dilemma is aggravated when offensive forces have an advantage over defensive forces (making defense more difficult and conquest easier) and if it is difficult to distinguish offensive forces from defensive ones. Under these conditions even “status-quo powers must then act like aggressors.” 38 A second answer is that while most states are defensive-minded, defensive realists acknowledge that power-maximizing aggressors who prefer expansion to the territorial status quo do exist. Charles Glaser refers to these states as “greedy states”—they have motives beyond pure security. 39 The extent of a state’s “greed” may be due to individual-level personality variables or it may be created by particular domestic political conditions—variables that have nothing to do with the structure of the international system. One example is Jack Snyder’s theory of cartelized political systems that pursue expansionist policies because they are driven by logrolling political tactics that perpetuate myths of expansion. 40 Nevertheless, states with greedy motivations are deemed to be exceptions. The point is that since states have a certain base level of uncertainty about what motivates an adversary, and if one suspects that the adversary is a greedy state the best strategy may be a competitive one rather than a cooperative one. 41 The last (and primary) answer is that anarchy, combined with misperception and imperfect information about the capabilities and intentions of others, induces fear and insecurity; this drives states to increase their own security, which has the effect of decreasing the security of others. In other words, under anarchy even the defensive-minded orientation of most states leads to the creation of security dilemmas and conflict spirals that may end in unwanted wars. Note that here again the defensive realists admit the influence of individual-level variables— misperception and uncertainty—in addition to systemic variables. In sum, according to Jervis, “Both structural and perceptual reasons conspire to render selfdefeating the actions states take to protect themselves.” 42 The dynamics of international politics resemble a prisoners’ dilemma game. Even defensive-minded states may see expansion or preventive war as the best way to pursue their need for security. Part of the problem is that defensive-minded states do not always recognize each other as such, and given the high costs for mistaking an expansionist state for a status quo oriented state, states often assume the worst. (Are the North Korean and Iranian nuclear programs motivated by need for security or by more aggressive goals?) Uncertainty is endemic to anarchic systems. A state’s attempts to improve its relative security sends ambiguous signals to others. States are not only uncertain about the current capabilities and motives of others, but just as important, they are uncertain about potential shifts in their future motives and relative capabilities. 43 This leads to perceptions of windows of opportunity and windows of vulnerability, which are important to defensive realists. Ultimately, however, defensive realists argue that the system provides states with incentives for war only under certain limited conditions. For this reason, Glaser prefers to call defensive realism “contingent realism.” There is a fairly wide variation in the behavior of

380

Chapter 10

states, and whether war occurs or not is contingent on a number of factors. 44 In the long run, however, security is more plentiful than others (offensive realists) think. One important contribution to modern realist thought comes from Stephen Walt in his Origins of Alliances. 45 Classical realists (and early neorealists) tended to argue that states balanced against power; when a state or coalition of states became too powerful, other states would balance against them. The weak balanced against the strong. However, this general argument failed rather miserably to account for certain historical realities. Why did Britain balance against Germany instead of the United States prior to World War I, and why haven’t Russia, China, and the EU countries balanced against the U.S. in the post–Cold War era? Walt’s answer is that states balance against threats, not against power. Balancing occurs when a state or a coalition becomes especially dangerous and threatening. The level of threat posed by a state depends upon four things: (1) its level of aggregate power; (2) its geographic proximity; (3) its offensive capability; and (4) others’ perception of the degree of its aggressive intentions. Notice that the list includes dyadic level, state-level, and individual-level factors. Finally, in Walt’s view, states do not always balance against the state that represents the greatest overall systemic threat but against the state that poses the most immediate threat. 46 The crucial consideration here is that states evaluate threats based on their own assessment of the danger to themselves. This reformulation or refinement of neorealist balance of power theory helps to explain not only why states don’t always balance against the most powerful states in the system but also why many states actually join the coalition of the strongest state— a practice called bandwagoning. It also dovetails with democratic peace theory. Democratic states (like the members of the EU) do not fear other democracies; thus, even when the Soviet Union collapsed and the United States was left as a dominant state, alliances did not form to counter American power. The United States did not seem a threat to other major powers in the system. In fact, the primary American alliance system—NATO—expanded after the collapse of the USSR. The Soviet Union’s former partners bandwagoned with the winners of the Cold War, allying against the much diminished loser, the Russian Federation. OFFENSE-DEFENSE REALISM Other defensive realists like Stephen Van Evera have refined neorealist theory in their own way. Van Evera acknowledges the importance of system structure, but for him the crucial structural characteristic is not the number of great powers but the offense-defense balance—a concept that originated with Robert Jervis. 47 Jervis argued that the security dilemma is lessened by two variables: (1) whether defensive weapons and doctrines could be distinguished from offensive ones and (2) whether the military advantage was with the defense or the offense. If defensive weapons were clearly distinct from offensive ones, states could—by attaining exclusively defensive weapons—make themselves more secure without threatening others. And when defensive forces held the advantage over offensive ones, the security of status-quo states would not be overly reduced by minor increases in offensive military might by other states. If defenses are superior to offenses and they are clearly distinguishable from each other, “aggression will be next to impossible, thus rendering international anarchy relatively unimportant.” 48 On the other hand, the security dilemma takes its most perilous form

The International System Level of Analysis, Part I

381

when offense forces are dominant and when it is difficult to distinguish offensive forces from defensive ones. These conditions are doubly dangerous: “There is no way to get security without menacing others, and security through defense is terribly difficult to obtain.” 49 Under these conditions, even status-quo powers might have to act like aggressors, initiating war rather than risk being attacked. The offensive-defense balance is part of what Van Evera refers to as the fine-grained structure of power, that is, the distribution of particular types of power—offensive power versus defensive, rising power versus declining power, the power to retaliate after attack, the frequency of power fluctuations, size of first-move military advantages, and so on. He sees this fine-grained structure of power as more important than the gross structure of power (bipolar or multipolar systemic distributions of power, emphasized by neorealists like Waltz) in raising the risk of war. 50 Van Evera puts forward five hypotheses about the relationship between fine-grained structures of power and war. War is believed to be more likely: (1) when states are falsely optimistic about its outcome; (2) when the first side to mobilize or attack has the advantage; (3) when the relative power of states fluctuates sharply, creating windows of opportunity and vulnerability; (4) when resources are cumulative; and, most important, (5) when conquest is easy. 51 He argues that conditions 2 through 5 are actually fairly rare; the structure of international power seldom gives incentives for aggression; indeed, aggression seldom succeeds. However, the above conditions are frequently wrongly perceived to exist. Thus, national decision makers “often exaggerate the size of first-move advantages, the size of windows of opportunity and vulnerability, the degree of resource cumulativity, and the ease of conquest.” 52 Consequently, Van Evera argues that realism has greater explanatory power when its focus is shifted from the gross to the fine-grained structure of power and from power itself to perceptions of power. 53 This is an emphasis on individual-level variables normally absent in classical realist and neorealist analyses. Let’s take each hypothesis in turn. The first hypothesis, that war is more likely if governments fall prey to false optimism about its outcome, is reminiscent of Geoffrey Blainey’s argument, discussed earlier, that wars are caused by disagreements about the relative strengths of countries and that misperceptions about the low costs of war and the high probability of victory are important factors that cause wars. Van Evera provides historical examples to indicate that leaders often held far too optimistic assessments about the real military balance, about the balance of will, about the commitments of allies, and the costs of the war and their ability to win. 54 Second, if the advantage lies with the first side to mobilize or attack, war is hypothesized as more likely. One presumed effect is preemptive war (this hypothesis is the core of what is referred to as “stability theory”). If attacking or mobilizing first is advantageous, one side or both may fear that they will soon be attacked by the other. The incentive to move first is twosided: the best defense for both would seem to lie in using the advantage themselves. They may even be spurred on by a false warning of such an impending attack derived from an accident, by a technical error or the misreading of evidence from the other side. As we have seen, Van Evera agrees with critics like Dan Reiter that there are very few real historical examples of such preemptive wars. His short list of modern preemptive wars includes only World War I (Russia’s mobilization and Germany’s attack on Belgium), the 1967 Six Day

382

Chapter 10

War, Germany’s 1940 attacks on Norway and the Low Countries, and (probably) the 1870 Franco-Prussian War. 55 However, first-move advantages lead to wars through pathways other than the preemptive one. States may launch an opportunistic attack, not because it fears it is about to be attacked, but simply because the first-move advantage creates smaller risks and increases the chances for victory for the initiator. The Japanese attack on Russia in 1904 is probably a good example. Also, first-move advantages lead states to conceal their capabilities and plans, as well as their grievances and perceptions. This has the unfortunate effect of cutting short ongoing diplomatic efforts to solve problems peacefully. Finally, Van Evera argues that real first-move advantages are rare, but unwarranted perceptions of first-move advantages are common and have equally pernicious effects. 56 Third, Van Evera has hypothesized that war is more likely when the relative power of states fluctuates sharply. This more or less incorporates the basic ideas of power transition theory, which Van Evera refers to as window theory. 57 Windows of opportunity or vulnerability can be caused by changes in the relative strength of two or more states—which can be brought about through economic growth, quantitative or qualitative arms buildups, changes in military strategies, diplomatic realignments, and even war itself. But windows are also affected by the offense-defense balance. As Van Evera notes: Windows are larger when conquest is easy. If the offense dominates, a small relative military decline can spell descent from dominance to vulnerability. If the defense dominates, shifts in force ratios produce smaller shifts in the capacity to overrun territory, hence they produce smaller shifts in meaningful political power. 58

The arrival of a power shift (or “window”) creates a one-sided incentive: “the declining state wants an early war, while the rising state wants to avoid war until after the power shift. Specifically, the declining state strikes ‘preventively,’ launching war now to prevent later conflict under worse conditions.” 59 Windows create incentives for preventive war, but they also create two other problems. First, windows lead declining states to risk war more willingly and therefore they adopt more risky external policies that may eventually lead to unwanted wars. Second, they make cooperation between rising and declining states difficult to achieve because they create an atmosphere of haste and impending war and (more important) they undermine the potential for deal-making. Declining states are unlikely to trust offers made by rising states because they suspect that once the adversary has achieved its relative dominance, it will be in a position to repudiate agreements made while relatively less capable, and the declining state will not be able to enforce the deal (you will probably have recognized this as Fearon’s “commitment problem”). Van Evera’s historical cases suggest that preventive motivations were behind Germany’s actions in 1914 and the 1930s, Japanese aggression in 1904 and 1941, and Egyptian aggression against Israel in 1973. He argues that, in fact, preventive motives were evident in most wars he examined in his book, though critics could plausibly argue that in some cases windows explanations may account for when and how military actions took place rather than whether they took place. 60

The International System Level of Analysis, Part I

383

Alternatively, since the logic of windows is clear to all, windows may actually lead a rising state to believe that the declining state is likely to attack, thereby giving the rising state some motivation to take preemptive action against the impending attack. Van Evera’s example here is that Russian mobilization in 1914 against Germany was preemptive in part because the Russians understood that Germany had for years been contemplating a preventive war against rising Russian power. 61 Once again, perceptions are crucial. Van Evera argues that perceptions of windows have probably been far more common than the reality of windows. “Many, perhaps most, preventive wars were disasters for the initiator: the windows were nonexistent or too small to exploit with success.” 62 This was true for Germany in 1914 and in 1939–41; it applies equally to Japan in 1941. This is not to say that all preventive wars are doomed to failure: they did produce a modicum of success for Japan in 1904 and the United States against Iraq in 1991. The fourth hypothesis is that war is more likely when resources are cumulative, that is, when the control of resources enables states to protect or acquire other resources. Resources might include raw materials and food sources as well as industrial capacity, military bases, land for strategic depth, currency, and labor (including military labor). Since cumulativity is affected by the costs of extracting a resource from a particular piece of territory, it is affected by the offense-defense balance. When conquest is easy, resources become more cumulative, and when this happens, international relations becomes more competitive and states are drawn toward expansionist policies. Van Evera argues that both German and Japanese expansion in the 1930s were driven by the notion that “many conquerable resources were highly cumulative.” Each sought economically self-sufficient empires to prevent strangulation by enemy blockades. 63 As Van Evera rightly notes, developments in the modern age—such as high-tech warfare (especially nuclear weapons), the shift to knowledge-based economies, and the rise of nationalism—have combined to reduce the cumulativity of resources in recent times. 64 The fifth (and primary) hypothesis is that war is more likely when conquest is easy. 65 This prediction is deduced directly from offense-defense theory: Conquest is easier when offenses are relatively more potent than defenses. Van Evera refers to offense-defense theory as the “master theory,” in the sense that it helps to explain the other important causes of war. Like the other propositions put forward by Van Evera, it is even more powerful when reformulated in perceptual terms: War is more likely when governments believe conquest is easy. 66 The argument is that when the international system (or regional subsystem) is characterized as “offense dominant,” conquest is fairly easy, and this situation has numerous effects that singly (or in combination) act to cause war. Moreover, false perceptions of offensive dominance create the same dangerous conditions. 67 In sum, if offensive dominance exists or is perceived to exist, the following things are expected. 1. States are more likely to pursue opportunistic expansion because militaristic expansion is easier, its costs are relatively low, its benefits will be high, and it has a greater chance of success. Examples might include Germany’s attack on Poland in 1939 and North Korea’s attack on South Korea in 1950. 2. States are more likely to pursue defensive expansion to attain a better geostrategic position because their borders are less easily defended and more likely to be attacked.

384

Chapter 10

Examples might include Russia’s attack on Finland in 1939 and Israel’s attack on Lebanon in 1982. 3. States tend to resist others’ expansion more fiercely due to increased insecurity. 4. The chances of preemptive war are greater because first-move advantages are larger. Offensive dominance creates a “trigger-happy” environment during crises as no one wants to be caught off guard. Both sides have an incentive for preemption due to mutual fears of an imminent surprise attack by the other. Examples might include World War I and the Six Day War in 1967. 5. Shifts in relative power have greater potential effects in an offense-dominant environment, and therefore windows of opportunity and vulnerability seem larger and more dangerous. The chances of preventive war are therefore greater as states attempt to shut potential windows of vulnerability created by the growing power capabilities of others. Examples might include Japanese aggression against the U.S. in 1941. World War I also had elements of preventive motivation for Central European states fearful of growing Russian power. Finally, the U.S. attack on Iraq in 2003 had preventive motivations of this sort, although there were also elements of opportunistic expansion. 6. States are more likely to use fait accompli diplomatic tactics, which may trigger wars. 7. States tend to negotiate less readily and cooperatively, leaving disputes to fester unresolved. 8. States tend to be more secretive, raising the risk of military and political miscalculations. 9. States tend to react faster and more belligerently to others’ blunders, making blunders more dangerous. 10. Arms racing tends to be faster and harder to control, raising the risk of preventive wars and wars of false optimism. 11. Offense dominance is self-feeding. As conquest grows easier, states adopt policies that make conquest still easier. Van Evera uses case studies—Europe from 1789 to the 1990s, ancient China, and the United States from 1789 to the 1990s—to provide supporting evidence for the core hypothesis. This is not, strictly speaking, a test of the theory, but an attempt to provide enough supporting evidence to indicate the value of the theory. He argues that the case studies corroborate offense-defense theory and indicate that it has great importance: shifts in the offensedefense balance have large effects on the risk of war. . . Offense-dominance is rather rare in history; hence it explains only middling amounts of war. However, perceptions of offense dominance are common, and they explain much warfare. 68

Indeed, he concludes that “Wilhelmine and Nazi Germany, Imperial Japan, Napoleonic France, and Austria-Hungary were all destroyed by dangers that they created through their efforts to escape from exaggerated or imaginary threats to their safety.” 69 The outbreak of World War I is then used to examine a more complete set of hypotheses (twenty-seven predictions in all) on the effects of offense dominance or the perception of it. He concludes that almost all of the predictions of offense-defense theory are fulfilled by the

The International System Level of Analysis, Part I

385

1914 case study. The belief that conquest was easy was a common cause behind many of the phenomena that appeared to be central to the outbreak of the war. In other words, perceptions of offense dominance led to the abundance of intervening factors leading to war. 70 Van Evera argues that the offense-defense balance in the international system is normally strongly in favor of the defense—relatively speaking. 71 Under these conditions, aggression rarely succeeds—because of defensive superiority but also because of balancing. Thus, under normal conditions (and accurate assessment of the security environment) states have little incentive to use force; protection of territory is easy; borders are more secure; states feel less threatened; and war is less likely. On the other hand, at those relatively rare times and places in which the offense is relatively superior, states will be tempted; they will feel more threatened; and war is more likely. (This is the primary prediction of the theory.) Since the perception of offense superiority is as important as its actual existence, war is more likely (a) when conquest is easy and perceived to be easy, and (b) when conquest is perceived to be easy even when defense is dominant and conquest is not easy. This leads to questions about whether the crucial variable is one of systemic structure (offense-defense dominance) or of individual level perception. A final issue of importance is this: How do we know what the offense-defense balance is? How is it defined? This is where the theory has encountered its greatest criticism. We normally think the balance between offense and defense is a function of military technology and doctrine, but this is only part of what determines offense-defense balance according to Van Evera. We have to also consider several other factors: military force posture, deployment and wartime operations; geographic position; the socio-political order (the internal popularity and strength of regimes); and various diplomatic factors (like alliance formation, balancing by neutral states, collective security arrangements, isolationist policies). All of this gets difficult to evaluate for various regions of the world over different periods of time. And those who use this formulation are forced to rely on rather subjective “authors’ estimates” that are bound to involve some analytical bias in order to determine whether a particular era is offense or defense dominant. And it turns a relatively simple military concept into a wide-ranging “megavariable” that leaves virtually nothing out. Finally, as several critics note, the inclusion of so many factors into the concept of offense-defense balance comes close to defining it merely as relative power. 72 Others argue that the elements of offense-defense balance are all derived from the ease of seizing territory; however, if control of another’s territory is not necessary to win a war, then offense-defense theory is founded upon questionable logic. 73 OFFENSIVE REALISM Offensive realists—led by John Mearsheimer—take the position that anarchy compels all states to maximize their relative shares of global power. 74 Mearsheimer argues that since security can only be found in power, states must seek the greatest power to get the greatest security. States act offensively; they seek opportunities to weaken potential adversaries and this sometimes means going to war. In Mearsheimer’s view, there are no status-quo great powers. All great powers have revisionist intentions and seek to change the distribution of power in their favor; indeed, the best way to ensure security is to become the most powerful

386

Chapter 10

state in the system (a hegemon). The tragedy of international politics is that states have little choice in this. All states are compelled by the same logic to act offensively in the face of anarchy. Other offensive realists emphasize the important effect of certain malignant states on the system; offensive-minded states with revisionist motives do exist. Randall Schweller argues that a world of security-seeking states (as described by defensive realists) should be peaceful because such states would surely be able to recognize the defensive motivations of each other and respond accordingly, avoiding the security dilemma and conflict spirals. 75 In Schweller’s view, that world is mythical. The real world is frequently populated by “predatory states” (Glaser’s “greedy states”). Predatory states are motivated by the desire to expand and to maximize their absolute power. Like rotten apples, they spoil the barrel; they cause other states to engage in balancing behavior. Without real external threats from predatory revisionist states, “there is no reason to expect imbalances of threat to emerge nor imbalances of power to act as a catalyst driving balancing behavior.” 76 But for Schweller, the existence of predatory states who are not really interested in mutual security with their neighbors means that “conflict is not only apparent but real; and because it is real, the resulting insecurity cannot be attributed to the security dilemma/spiral model of conflict.” 77 This is markedly different from the security dilemma hypothesis that assumes that most states have defensive motives and there is only a misunderstanding of the other side’s defensively motivated attempts to increase security. In the security-dilemma model, threats are apparent but not real. 78 In offensive realism, threats are real. Schweller argues that history provides numerous examples of wars started by states that pursued expansion, even at the expense of their own security and survival. States that are profoundly dissatisfied with the systemic status quo tend to believe they have little to lose and are willing to take large risks to increase their positions in the system. He cites wars for universal empire begun by Alexander the Great, Charles V, Napoleon, and Hitler as well as wars for more limited gains initiated by Metternich, Napoleon III, Bismarck, Wilhelm II, Mussolini, Stalin, and the Japanese leadership of the 1930s and 1940s. 79 In Schweller’s brand of offensive realism—called balance-of-interest theory—it is not the structure of the international system (anarchy) that creates a motivation for expansion. The cause is to be found at the “unit level” in the nature of certain individuals or certain states. This attention to the unit level also characterizes the final type of realism: neoclassical realism. NEOCLASSICAL REALISM While neorealists generally attempt to explain broad, recurring patterns of international outcomes such as major wars, the rise and fall of hegemonic orders, and patterns of alliance behavior, other types of realists have instead attempted to build theories that explain variation in the foreign policy choices of individual states. These new-style realists have become known as neoclassical realists. 80 To some extent the approach represents an attempt to get out from under the straitjacket of reliance on system-level variables by incorporating variables from other levels of analysis to help explain how states behave. In this, it also represents an attempt to turn back to some of the insights of classical realists (hence the “neoclassical” branding).

The International System Level of Analysis, Part I

387

Neoclassical realists see anarchy as a permissive condition rather than an independent causal force. Taliaferro, Lobell, and Ripsman argue that “In the short run, anarchy gives states considerable latitude in defining security interests, and the relative distribution of power merely sets parameters for grand strategy.” 81 Neorealists (like Waltz) and neoclassical realists both think that systemic variables are of primary importance. However (and this is the important point), neoclassical realists do not see systemic or structural conditions as directly or automatically determining or influencing the behavior of states. Instead, they see “elite calculations and perceptions of relative power and domestic constraints as intervening variables between international pressures and states’ foreign policies.” 82 Neoclassical writers use the term unit-level variables to refer collectively to variables at the state, substate, and individual level of analysis. Unit-level variables “constrain or facilitate the ability of states to respond to systemic imperatives.” 83 The objective reality of any particular systemic distribution of power and what it means for a particular state in its relations with others is mediated through elite perceptions, belief systems, and operational codes; through the various organizational and bureaucratic processes of a state’s national security institutions; and domestic political considerations, including the influence of mass opinion and interest groups. 84 National security executives are frequently compelled to bargain with legislatures, political parties, interest groups, militaries, and economic managers in order to make policies and extract the resources necessary to carry out foreign policies. 85 Taliaferro et al. argue that “leaders define the ‘national interests’ and conduct foreign policy based on their assessment of relative power and other states’ intentions, but always subject to domestic constraints.” 86 Since systemic incentives and threats are probably ambiguous in the first place, this means there is probably not a single, optimal response to any particular systemic condition. 87 “Neoclassical realism posits an imperfect ‘transmission belt’ between systemic incentives and constraints, on the one hand, and the actual diplomatic, military, and foreign economic policies states select, on the other.” 88 The incorporation of elite perceptions (and misperceptions) and domestic political variables as intervening variables means not only that states are not unitary actors, they may not always respond in a perfectly rational manner to their environment. 89 Given a certain set of systemic conditions, different states will respond to them in quite different ways. Or the same state will respond differently, under different leadership—especially if the structure of the system is in flux or is ambiguous. A good example is the U.S. response to the end of the bipolar Cold War system after 1991. 90 Presidents George H. W. Bush (a classical realist with some liberal-idealist notions) and Bill Clinton (a liberal idealist) on the one hand, and George W. Bush (a neoconservative) on the other, perceived the post–Cold War system in quite different ways and formulated quite different policies in response to what they perceived the nature of the system to be. The first two presidents generally perceived a new multipolar world in which the major powers, having no fundamental conflicts between them, would be better able to cooperate through international institutions like the United Nations. Because of great power cooperation, the world would be a safer and more peaceful place. The second President Bush tended to see the international system as unipolar, with the United States in the position of a hegemonic power. The U.S. would be in a position to lead the world and, if necessary, impose its will on

388

Chapter 10

the world. Unfortunately, the new system would not be more peaceful and less threatening than its predecessor. Threats would remain; they would just be different, with the primary threat coming from rogue states that developed weapons of mass destruction and had ties to terrorist groups. To the extent that there would be peace and stability it would be the product of U.S. hegemonic leadership. 91 Here we see that the structure of the international system did not have a direct effect on U.S. actions; instead, it was mediated through images, perceptions, and ideologies of different political leaders. As you can probably tell, neoclassical realism is primarily a theory of foreign policy behavior with implications for war, rather than a theory of war per se. Most neoclassical realists would put themselves in the defensive realist camp. Taliaferro has summarized four auxiliary assumptions of neoclassical defensive realists. 92 These assumptions specify how systemic, structural variables (anarchy and the distribution of power) are translated into the actual foreign policies of states—and whether war is the outcome or not. First, the security dilemma is an intractable feature of the international system. Second, even though the security dilemma is inescapable, it does not always lead to war. Certain “structural modifiers”—such as the offense-defense balance, geographic proximity, dyadic military balances—affect the security dilemma between particular states and act to increase or decrease the tendency toward war. These structural modifiers actually have a greater impact on outcomes of war and peace than do gross distributions of power. Third, material power drives states’ foreign policies, but through the medium of leaders’ calculations and perceptions. Since distributions of power are usually ambiguous, leaders evaluate these things in idiosyncratic ways, relying on their own preexisting images, belief systems, operational codes, and cognitive biases. We have already seen that defensive realists grant an important role to individual perceptions and misperceptions in assessing things such as the intentions of others, the existence of threats, windows of opportunity and vulnerability, and offensive advantages. To this extent there is considerable overlap between defensive realism and neoclassical realism. Fourth, domestic politics can limit the efficiency of a state’s response to the international environment, especially in noncrisis situations and periods lacking immediate external threat. REALISM AND THE ROLES OF POLARITY AND POLARIZATION Perhaps the aspect of the international system most widely discussed by realists of all stripes is the distribution of power among the system’s members. The notion that the balance of power within the system has an important effect on the behavior of states is an idea that can be traced back at least as far as Thucydides’ history of the Peloponnesian Wars. It is a concept dear to the hearts of the realist school of thought. Since the 1960s social scientists have attempted to subject traditional realist hypotheses about the relationship between the balance of power and war to empirical tests. 93 These attempts resulted in a steady stream of studies that fed into a huge swamp of contention over definition of concepts, proper operationalization of variables, and conflicting results and interpretations. If there is any one area of war research that can rightly be called a quagmire, this is it! Now that you have been properly warned, we can plunge ahead.

The International System Level of Analysis, Part I

389

The distribution of military power within the international system has been loosely associated with the concepts of polarization and polarity. Unfortunately, these terms have often been used to mean significantly different things by different people. David Garnham has distinguished four distinct meanings of polarity. 94 1. Polarity is used to refer to the size of the international system. Some scholars use it to distinguish the number of great powers, and others use it to distinguish the number of poles, a term vaguely described as “autonomous centers of power” (which may be individual states or may instead be the number of alliances/blocs/clusters). International systems are then classified as either unipolar, bipolar, or multipolar on the basis of the number of great powers or the number of clusters. (Hardly anyone uses the term tripolar for some reason.) 95 2. Power polarity (also known as power distribution) refers to the concentration or diffusion of national power within the system—whether power is concentrated in a small or large number of states. Although national power is usually thought of as being primarily military power, it also includes economic, technological, and demographic elements. Power polarity may also be unipolar, bipolar, or multipolar. Unipolar systems are characterized by power preponderance: power in the system is concentrated in the hands of a single state. Bipolar and multipolar systems are characterized (presumably) by parity of capabilities. Power is dispersed more or less equally between two superpowers (bipolarity) or among three or more major powers (multipolarity). In some analyses, polarity (size) and power polarity (concentration) seem to be indistinguishable. Although some scholars assume that an increase in the number of poles also increases the dispersion of power throughout the system, others have demonstrated that the number of poles is relatively independent of the distribution of power in the system. 96 3. The term polarization (also known as cluster polarity) describes a specific pattern of alliance bonds. It refers to the degree to which coalitions of states form mutually exclusive blocs of actors with internally friendly and externally hostile relations. Polarization is a function of tightness (the degree of similarity of alliance ties within a cluster) and discreteness (the degree of dissimilarity of alliance ties between clusters.) Polarization occurs when there are many alliance bonds within clusters, but no bonds across clusters. A completely (or tightly) polarized system occurs when all the states in the system are members of one of two alliances and when no member of alliance A is also a member of alliance B. This is sometimes called cluster bipolarity. A system that lacks polarization might have no blocs at all or it might be made up of several blocs with overlapping membership; in this case, the blocs are not mutually exclusive. 97 Nonpolarized systems may also be referred to as being cluster multipolar. (To muddy the waters even further, some scholars think that power distribution ought to be part of the definition of polarization; in other words, the two concepts ought to be combined.) 98 4. Finally, researchers have also sometimes seen power distribution in terms of the magnitude of alliance commitments, that is, the number of alliance commitments in the system or the percent of states in the system with alliance ties. Because the second and third concepts—power polarity and cluster polarity—are the most theoretically pregnant, we shall concentrate on them. We should start by noting that systems that are cluster bipolar are not necessarily power bipolar, nor vice versa. Similarly, systems that are cluster multipolar are not necessarily power multipolar, nor vice versa. 99

390

Chapter 10

The fourfold classification scheme in table 10.1 identifies international systems that occupy all four possible types. The nineteenth-century classical balance of power era in postNapoleonic Europe was both cluster and power multipolar, with many great powers of similar strength arranged into shifting complex alliance arrangements. The World War II era and the years immediately preceding World War I are characterized by numerous relatively equal great powers (power multipolarity) arranged into two tightly polarized alliances. The early Cold War period that immediately followed the end of World War II is power bipolar, with two superpowers militarily dominant, and cluster bipolar, with two very tight alliance systems formed around each of the two superpowers. By 1962 the system was still power bipolar, but no longer cluster bipolar. The Soviet bloc had begun to break down: Yugoslavia had been evicted in 1948, Albania had gradually joined the Chinese camp, Romania had become independent (at least in foreign affairs), and most importantly the two communist giants—China and the Soviet Union—had become wary opponents instead of allies. The Western bloc had suffered deterioration as well, with the defection of the French from the united NATO command structure. The two superpowers were still militarily predominant, but their blocs had been weakened, and strong, nonaligned states such as India had begun to emerge. Empirical research has demonstrated that polarization declined consistently throughout the Cold War, with a rather abrupt decline between 1962 and 1963 and another between 1970 and 1972. 100 BIPOLARITY AND MULTIPOLARITY: THE THEORETICAL DEBATE Political scientists have debated the relative merits of bipolar versus multipolar systems since the 1960s. 101 Though different scholars attached different meanings to the terms, it is probably wisest to look at the bipolar-multipolar debate as if bipolarity meant both cluster and power bipolarity and to think of multipolarity as meaning both cluster and power multipolarity—in other words, in terms of the early Cold War system versus the classic nineteenth-century balance of power system. (We will discuss unipolarity in the next chapter.) The debate over the best system involved two main questions: Which system is most stable? And which system is most peaceful? These questions are not totally distinct. Stability refers to the ability of the system to endure over time. The alternative is for the system to

The International System Level of Analysis, Part I

391

undergo transformation into a different type of international system—for instance, from bipolar to multipolar or unipolar. Such transformations, however, are generally thought to involve a cataclysmic general war that eliminates some of the major actors, creates larger states out of small states through territorial annexation or absorption, and/or splits large states and empires into several smaller, less powerful units. Therefore, what theorists really mean by stability is whether the system can maintain itself over time without experiencing major wars. 102 On the other hand, peacefulness refers to low levels of conflicts of many types, not just major wars. Arguments for Multipolarity Several explanations have been put forward for the reputed stability of multipolar systems. 103 1. As the number of actors in the system increases (especially the number of major powers), the number of possible opportunities for cooperative interactions also increases. The increase in the number of interactions also produces cross-pressures, or cross-cutting loyalties among actors. This decreases the likelihood that any single relationship will become one of implacable opposition; instead, it is more likely that a state’s opponent on one issue may be a friend on another issue. This creates a pluralistic system of complex relationships that prevents rigid cleavages and polarization, reducing the probability that conflicts will lead to war. 2. The larger the number of major actors, the larger the number of potential allies that might be arrayed against an aggressor and the greater the flexibility in alignment. Because states pursue a balance-of-power policy in which they agree to counter any attempt by a state or group of states to become too strong, the greater the number of major powers, the greater the number of potential balancers and the greater the deterrent capability of the system. (Sometimes, of course, deterrence fails and it is necessary to fight limited wars to prevent a state or a coalition of states from achieving dominance.) 3. The greater the number of actors, the larger the number of possible mediators who might be capable of resolving disputes. 4. Multipolar systems help to slow the rate of increase in the arms race, thus reducing tension and hostility. If there are only two major powers in the system and state A increases its armed forces from twenty divisions to twenty-three, state B will be forced to raise its forces by the same number. However, if there are several major powers and state A increases its divisions by three, then states B, C, and F will only have to raise their armies by one division each to offset the increase of state A. 5. In multipolar systems states are unable to become entirely preoccupied by a dispute with any one opponent; consequently, it becomes difficult to allocate sufficient levels of attention to any one state in order to decide for war with that state. It is assumed that a state needs to devote a certain minimum level of attention to another state in order to make a decision for war. Deutsch and Singer once estimated that limit to be 10 percent of the state’s total foreign policy attention. 104 As the number of actors in the system increases (especially the number of major actors), it becomes increasingly difficult to allocate a level of attention to any single actor sufficient to make war possible.

Chapter 10

392

6. In a multipolar system the significance of great-power conflict is reduced. Antagonism is spread out across the entire system, whereas in a bipolar system antagonism is reinforced and echoed throughout the system because of its presumed polarization. 7. Perhaps most important, multipolar systems are inherently full of ambiguity, uncertainty, and unpredictability owing to the large number of major actors and the complex linkages between them. Given the cross-cutting alliance systems, it would be difficult to predict with any precision who would line up on whose side if war were to break out and therefore it would be difficult to predict the outcome of that war. Therefore states are forced to act cautiously. 8. Finally, the historical record indicates that such systems have been characterized by a low intensity of conflict. Minor wars may break out in multipolar systems, but major wars are quite rare—witness the long century of relative peace from the Congress of Vienna in 1815 until World War I. Arguments against Multipolarity Critics of multipolarity dispute the above reasoning with counterarguments of their own: 105 1. A large number of actors increases the number of potential opportunities for conflict as well as the potential opportunities for cooperation. The law of averages argues in favor of greater conflict as a result of greater interaction opportunities. 2. The greater the number of major powers, the more likely the system is to be characterized by diversity of interests and demands rather than by cooperation. 3. The level-of-attention argument assumes a fixed allotment of attention to which states have access and a rather limited ability to reallocate attention among the states in the system. Both these assumptions are too restrictive. States can walk and chew gum at the same time. 4. The increased uncertainty that characterizes multipolar systems leads to a greater probability of misperception and miscalculation and therefore to a greater probability of war. Furthermore, ambiguity tempts aggressors to gamble rather than inducing them to be cautious. 5. If we accept the idea that equality of the distribution of resources among major powers is more conducive to peaceful relations than inequality, mathematical probabilities indicate that the more states in the system, the more likely that resources are to be distributed unequally. This transition from equality to inequality is especially critical as the system moves from two states to three. And an increase in the number of great powers coupled with a decline in valued resources (such as the number of unaligned small states) would lead to a dramatic increase in inequality and instability. 106 6. Finally, the balance-of-power argument requires states to be both risk-averse and riskacceptant at the same time. It assumes that potential aggressors won’t take the risk of launching war when faced with the possibility of a rival coalition but also that some states are expansionistic and are willing to risk overthrowing the status quo.

The International System Level of Analysis, Part I

393

Arguments for Bipolarity The explanations put forward for the reputed stability of bipolar systems are as compelling as those in favor of multipolarity: 107 1. Because the two superpowers have interests everywhere, a solid balance is created, making expansion unlikely and preventing conflict. 2. Since the system is polarized and superpowers have interests all around the globe, a war anywhere could become a general war. This fear of major war and the realization that the stakes are very high in any contest induces caution all around. 3. Certainty and calculation are increased because of the simplicity of alignment in a bipolar system. This lessens the chance of war through misperception or miscalculation. 4. The superpowers are able to pressure their extremist allies or other small states to moderate their behavior. 5. Owing to rigid alliance structures and the salience of superpower issues, bipolar systems have essentially only one dyad across which war might break out—greatly reducing the probability of war—as opposed to multipolar systems, where war might break out across numerous dyads. 6. The balance of power is easier to achieve in a bipolar system. Adjustments and balance are automatic. In fact, balancing mechanisms become regularized, and superpowers develop routines for dealing with crises. A succession of crises even allows the superpowers to fine-tune conflict management techniques. 7. Shifts in alignment among the smaller states will not significantly threaten the balance of power due to the overwhelming military might of the two superpowers. Since the balance of power is secure, changes in alignment do not need to lead to war. 8. Historically, tight bipolar systems—such as the post–World War II international system—have been extremely stable. (While the multipolar camp relied on the lessons of nineteenth century diplomatic history, the bipolarity camp’s arguments are based on the lessons of the post-WWII era.) Arguments against Bipolarity Critics of this view offer several counterarguments: 1. The level of hostility in such a highly polarized system is extremely intense; antagonisms are automatically reciprocated. In effect, a bipolar system is a large zero-sum game in which whatever gains are made by one side must be achieved at the expense of the other. Regardless of whether an alignment shift actually changes the distribution of power capabilities, the losing side will certainly perceive its loss as vital, precipitating a causus belli. 2. A completely polarized system lacks mediators that might be able to moderate or alleviate conflict between the two blocs. 3. Since great-power interests are at stake everywhere, any conflict anywhere can become a general conflagration. The world is continually at the brink of war in such situations,

394

Chapter 10

and the crisis management capabilities of the superpowers are bound to fail at some point. 4. The superpower military stalemate may, paradoxically, lead to situations where conflicts in the less-developed world are tolerated by the bloc leaders because they fear that intervention could lead to direct confrontation. 5. Clarity and certainty of calculation may not be beneficial. It might be that states go to war when they are more certain of the international environment than when they are uncertain. The lack of ambiguity may lead to more war rather than less. The proponents of bipolarity and multipolarity are essentially making the same argument: peace is maintained through the balance of power. They just disagree on which type of international system is most conducive to maintaining a stable balance. They both argue for a policy that maintains an equal balance; some just happen to think this is easier in a system that has only two superpowers and lots of smaller states; others think it is easier when there are many states with equally strong capabilities. What theorists do not advocate is a conscious effort to change one system into the other. Bipolar theorists do not suggest, for example, that during periods of multipolarity we should attempt to reduce the number of great powers to two! Everyone understands that transitions from multipolarity to bipolarity (or vice versa) are frequently accompanied by major wars. Balance-of-power politics are therefore seen as methods of preserving the system as long as possible, whether it is bipolar or multipolar. POLARIZATION: EMPIRICAL RESEARCH How can we discover whether bipolarity or multipolarity is actually better for peace and stability? Researchers have tackled the question in recent decades. Let us first take up the research on polarization or cluster polarity, the degree to which coalitions of states form mutually exclusive blocs of actors with internally friendly and externally hostile relations. Frank Wayman’s study of major power wars finds that for the nineteenth century cluster multipolarity (that is, a lack of polarization) leads to war. Alliance clusters seem to have been disintegrating prior to the outbreak of war. The twentieth century shows just the opposite pattern: cluster multipolarity leads to peace, and alliance patterns become (bi)polarized prior to the outbreak of war. And since the wars of the twentieth century were more severe and were of greater magnitude and duration, high cluster bipolarity (polarization) is therefore associated with severe general wars. 108 Bruce Bueno de Mesquita finds somewhat similar patterns for the entire international system. Like Wayman he finds basic differences between the two centuries. Although he finds no relationship between polarization and the onset of war, he does find that changes in tightness (polarization) precede both great power wars and interstate wars in general, though only in the twentieth century. He discovers that 84 percent of the wars in the twentieth century began in years following a five-year rise in systemic tightness. This is doubtless to be interpreted as the result of alliance making, a process that has a drastic effect on the number of interaction opportunities and that upsets previous patterns of relations. War almost never

The International System Level of Analysis, Part I

395

occurs in periods of declining tightness, according to Bueno de Mesquita, and no multilateral, complex wars occur during times of declining tightness. 109 Using a different indicator of polarization, Michael Wallace finds a curvilinear relationship between polarization and the magnitude and severity of war. 110 Wallace’s independent variable is based on alliance membership, intergovernmental organization (IGO) membership, and diplomatic links. This is combined into a weighted index that takes into account each nation’s military capabilities. Thus, his weighted index combines cluster polarity and power polarity. He finds that systems with either very low or very high levels of polarization have the greatest probability of experiencing severe wars. Moderate levels of polarization minimize the probability of war. This pattern appears to be fairly strong for the twentieth century, but weak for the nineteenth century. Wallace concludes that when there are no alliances, weak and unprotected nations fall victim to the strong; when there is high polarization, rivalries are extremely intense, leading to large, multilateral wars. Tight polarization pulls all members into war. 111 Jack Levy’s historical analysis of great power wars from 1495 to 1975 paints a different picture. He demonstrates that, contrary to the balance-of-power hypothesis, nearly all (five of the six) periods of the most highly flexible (and therefore unpolarized) alliance systems have been followed by relatively high levels of war. 112 The exception is the historically renowned Bismarckian system of 1871–90. Unlike many of the other unpolarized systems, this one was characterized by a static system of complex, cross-cutting bonds rather than by rapidly changing coalitions. Levy concludes that although one can find instances of highly polarized alliance systems that precede war, one can also find examples of those that have resulted in highly stable periods of peace (for instance, the alliances that centered on the League of Augsburg against France’s Louis XIV, compared to the later NATO/WTO polarization). One can find similarly disparate examples for loosely polarized systems. Perhaps the only thing we can conclude with any certainty from these studies is that the relationship between polarization and wars is neither linear nor constant over time. 113 However, there does seem to be some evidence that increased polarization and alliance aggregation leads to war, at least in the twentieth century. And there may be some connection between polarization and the duration, severity (number of battle-deaths), and magnitude of war (total nation-months at war). In various ways, the studies by Wallace, Bueno de Mesquita, and Wayman cited above all appear to link high levels of alliance polarization with bigger wars. ALLIANCES: EMPIRICAL RESEARCH Alliances are inextricably bound up with issues of polarity, polarization, and balances of power. Indeed, early theorizing about alliances was very much tied to balance of power theory. The primary function of alliances was seen as a tool to balance power against power, as a way of deterring war. Alliance construction was also seen as a way to reduce uncertainty, thereby reducing the potential for war by miscalculation. Of course, critics of this approach argued that alliance construction was more likely to increase perceptions of threat, and lead to counter-alliances and counter-threats and an escalatory spiral that ended in war. Let’s take a look at a variety of research that deals specifically with alliances.

396

Chapter 10

The research on alliance construction and aggregation (accumulation) shows the same sort of schizophrenic historical impact as the research on alliance polarization. Alliance aggregation and polarization are not, of course, one and the same. Although it seems logical that high levels of alliance creation, by reducing interaction opportunities and freedom of choice, would increase polarization, alliances can be constructed without greatly increasing the degree of polarization in the system. Alliances may instead be erected in such a way as to increase the complexity and the cross-pressures in the system rather than to reduce these things. Polarization will only be enhanced if the alliances created are mutually exclusive. Singer and Small hypothesize that alliances will most probably reduce cross-pressures, interaction opportunities, and the number of uncommitted actors; it would therefore be reasonable to expect that the greater the number of alliances, the greater the chance of war. Their investigation shows that at least in the twentieth century, the increase in the number of alliances in the system is related to the magnitude and severity of war, although not to the onset of war. In other words, alliance aggregation is related to large wars in the twentieth century. However, this does not apply to the nineteenth century; in that period alliances were related to peace. Over the entire period Singer and Small find a low association between alliance formation or aggregation and the amount of war. 114 An analysis of the 1816–1945 period by Charles Ostrom and Francis Hoole shows that within three years of alliance formation, as the number of alliance dyads in the system increase, so do the number of war dyads. After three years this relationship vanishes, suggesting that the danger of war exists only in the immediate aftermath of alliance formation and then recedes. 115 Jack Levy’s analysis of great power wars from the sixteenth to the twentieth centuries uncovers a similar pattern. Levy discovers that except for the nineteenth century, the majority of alliances have been followed within five years by war involving at least one of the allies. In fact, all great power alliances in the sixteenth, seventeenth, and twentieth centuries have been followed by war within five years and many of these wars were great power wars. In the nineteenth century, however, only a very small percent of the alliances were followed by Great Power wars. Of the fourteen great-power alliances, none was followed within five years by a war involving two of the allies, and only one was followed by a war involving one of the alliance members. 116 These studies provide evidence that alliance formation often leads to war, though the impact is not consistent across all time periods. 117 The major exception appears to be the nineteenth century. What accounts for the apparent different impact of alliances in different periods of time? And what accounts for the more general disagreement on whether (and how) alliances increase the probability of war? There must be something about the nature of certain alliances that make them especially prone to war, while other types of alliances may have precisely the opposite effect. From a political science perspective this requires the creation of a typology of alliances able to ferret out the different characteristics of war-prone alliances and peaceful (deterrent) alliances. (It also requires the construction of data bases that incorporate these distinctions.) Rather early on, various theorists began to think about how different types of alliances might produce different effects. Levy suggested that the purpose of alliances changed after 1815. He asserts that prior to 1815, most peacetime alliances were offensive in nature, with the treaty specifically calling for

The International System Level of Analysis, Part I

397

the initiation of war. After 1815 alliances were mostly defensive, with any military activity dependent on an attack on one of the alliance partners. The early alliances were generally ad hoc and formed as a deliberate prelude to war. However, nineteenth-century alliances were permanent rather than ad hoc and were intended for the purpose of maintaining the status quo and enhancing deterrence. 118 The nature of alliances changed again at the end of the nineteenth century, but in ways that are not totally understood. Michael Wallace agrees that the purpose of the alliance may be an important factor in explaining the differences in the patterns found in various centuries. He asserts that in the nineteenth century alliances were equilibrating mechanisms; they could be formed and broken off without creating a serious threat to the security of any state. On the other hand, alliances in the twentieth century have generally had as their purpose—at least up until World War II—the building of winning coalitions in case of war. This created an arms race reaction in both camps, vastly increasing hostility and the probability of war. 119 It may indeed be true that the nature of the nineteenth-century alliance was different from those of both preceding and following centuries, but Levy’s characterization of them as permanent and Wallace’s description of them as flexible leads one to paraphrase President Kennedy and ask whether these two fellows visited the same century. Finally, the creation of alliances during the Cold War occurred in a vastly different environment that appears to have had important implications for the connection between alliances and war. First, unlike the nineteenth century and the first half of the twentieth century, alliance making occurred in the context of systemic bipolarity as well as alliance polarization. Second, states in both major alliance systems possessed nuclear weapons. Membership in an alliance system generally had a protective deterrent effect, and states from different alliances were unlikely to confront each other in war because of the danger of escalation to thermonuclear warfare. On the other hand, states without membership were unprotected. One way of looking at the question of which alliances are most likely to lead to war is to take the position that alliances that appear most threatening to others should be most likely to be followed by war within the short term—say, three to five years. So which alliances are most threatening? Alliances involving great powers would seem to be one place to start. Singer and Small’s early investigations suggested that the strongest associations between alliances and war occurred when great powers were involved in the alliances (and when the type of pact was a defensive pact). 120 Douglas Gibler uses this as a jumping off place and adds one other important characteristic. His hypothesis is that the alliances that appear most threatening would be those that include great powers, but also those whose members had been successful in their last war. Gibler and Vasquez investigate this hypothesis using an updated version of the COW data from the 1815–1980 period and also Levy’s data on great power wars that covers the period 1495–1815. They find empirical support for the proposition for the entire period of 1495–1980 as well as for both the controversial nineteenth and twentieth centuries. Alliances with these dual characteristics were the most prone to war. (And when alliances that involved territorial settlements were excluded from the sample, the probability of war for alliances with these two properties was extremely high.) Fortunately, the number of alliances sharing these characteristics was fairly low—only 28 of the total number of 223 cases of alliances. 121 While it might

398

Chapter 10

seem that great power alliances are more prone to war simply because great powers themselves are more prone to war than other types of states, this appears not to be the reason. Gibler is able to show that entering an alliance actually increases the probability of war for major powers. 122 In a follow-up analysis of security alliances, Gibler creates a “bellicosity scale” for alliances based on three variables: success in previous war, dissatisfaction with the status quo, and great power status. The higher the bellicosity ranking, the greater the threat the alliance should pose to others. He identifies thirty-five “high bellicosity” alliances and finds that over half of them were followed by war within five years. High bellicosity alliances were almost twice as likely to eventuate in war than the average alliance. Gibler believes that most of these alliances were probably attempts to build war-winning coalitions. As we have seen from our exploration of the dyadic level, alliances are one of the variables that Vasquez and Senese incorporate into the steps-to-war model as contributing to increased probabilities of war for dyads. 123 While Gibler focuses on the types of states that are involved in the alliance, Senese and Vasquez look at who the states are. They suggest that the most problematic alliances are “outside alliances.” In a dyadic relationship between state A and state B, if B has an alliance with a country (state C) that A is not allied with, then the B-C alliance is an “outside” alliance. Senese and Vasquez also recognize that not all alliances are relevant to a particular dispute. They therefore include in their analyses only “politically relevant alliances.” An alliance is politically relevant if it contains a major power, if it contains a minor power contiguous to one of the parties, or if it contains a minor power within the same region as the target state. Their research shows that outside alliances are indeed problematic. Having outside alliances while involved in a territorial militarized dispute increases the probability of war for the entire 1816–2001 period. 124 However, when Senese and Vasquez look at the Cold War period (1946–89), the effect that alliances have on the probability of war changes. For the Cold War era if both states in a dyad had outside alliances the probability that a dispute would escalate to war actually decreased. This was true even in the presence of territorial disputes. 125 A situation of both states having external allies is largely typical of a polarized system in which most states have chosen up sides and belong to one of two alliances. Senese and Vasquez also find that if only one state has an outside alliance, then the chances of escalation to war increase if the dyad is involved in a territorial dispute or if it consists of a pair of rivals. The authors speculate that the Cold War era represented a rather unique set of systemic circumstances— the existence of polarized alliances coupled with the nuclear weapons in the hands of the alliance leaders. With the end of the Cold War and the disappearance of polarized alliances, outside alliances once again increase the probability that dyads will experience war. While this analysis focused on the dyadic level rather than the systems level, it nevertheless suggests a system-wide pattern. While the existence of outside alliances had a bellicose effect until 1945, the Cold War pattern is dramatically different. Based on dyadic-level studies, we might hazard a guess that the connection that Levy and others note concerning the positive relationship between alliances and war in the twentieth century may be due primarily to the events of the first half of the century.

The International System Level of Analysis, Part I

399

Another possible approach to the question of what kind of alliances are the most dangerous is to look at the purpose or function of the alliance. There is general agreement that the nature of the alliance matters; some types of alliances are more likely to increase the chances of war than other. The old COW Formal Alliance data broke alliances into three broad categories: defense pacts, neutrality or non-aggression pacts, and ententes (agreements to consult or cooperate). A new data set—Alliance Treaty Obligations and Provisions (ATOP)—has been constructed by Leeds et al. to provide some clarification. 126 The ATOP categories distinguish somewhat greater variation: offensive pacts, defensive pacts, neutrality pacts, non-aggression pacts, and consultation pacts. Using the COW alliance data, Sabrosky determined that defensive pacts are more prone to war than ententes. (Defensive pacts in the COW data set may have both offensive and defensive purposes.) He also found that neutrality/non-aggression pacts are particularly unreliable promoters of peace and are quite prone to war between the signatories themselves. 127 Nonaggression pacts may be just tactical “time-outs” during which one or both of the sides prepares for a future war. Recent evidence also suggests that predatory great powers sign nonaggression pacts as a tactic to constrain potential adversaries so that they can attack minor powers who are nonsignatories; the classic example is the 1939 Molotov-Ribbentrop Pact that made it possible for Hitler’s Germany to attack Poland without fear of Soviet interference. The making of non-aggression pacts therefore actually increases the probability of war between unequal states (major-minor wars). 128 Analysis of ATOP data and militarized disputes (MID) data by Leeds from 1816 to 1945 supports the proposition that offensive alliances increase the probability an alliance member will be involved in MIDs against a nonsignatory, while defensive alliances reduce the chances that an alliance signatory will be involved in a MID. 129 This follows the traditional view of alliances: Offensive alliances have predatory purposes and generate perceptions of threat and are therefore dangerous; defensive alliances are useful in deterring conflict. However, Leeds’s research only looks at the deterrent effect of defensive alliances on militarized disputes, not on war. A follow-on study that does use war as the dependent variable is unable to conclude that defensive alliances deter war. 130 Benson breaks the ATOP categories down further, creating distinctions between alliances that are for deterrent (defensive) purposes or for compellent purposes. 131 (Compellent alliances establish commitments between partners to assist each other in conflicts, including those waged for offensive purposes.) These alliance types can be further classified as either conditional or unconditional. Benson then seeks to determine the relationship between alliance types and the initiation of MIDs. Two types of alliances appear problematic. Unconditional compellent alliances (in which alliance partners are committed to assisting each other regardless of conditions) are strongly associated with MID initiation; they are associated with a 249 percent increase in the chances an alliance member will initiate conflict. Unconditional deterrent alliances (in which allies agree specifically to defend each other regardless of the circumstances) are also clearly associated with violent conflict—if allies are major powers or revisionist states dissatisfied with the status quo. Now some good news. There is also fairly solid evidence that one kind of alliance actually works to significantly reduce the risk of future war: alliances that are used to cement a

400

Chapter 10

territorial settlement. Gibler’s analysis of data from 1815 to 1980 indicates that only one of the twenty-seven such territorial settlement alliances was followed by war with the ally within five years. 132 Territorial settlement treaties do not increase the perception of threats; on the contrary, they contribute to peace by settling the issues behind the conflict. They lead to peace not in the way balance-of-power theory suggests, by balancing the power of one group with the equal power of another, but by solving the root causes of conflict—territorial disputes. Jack Levy reminds us that we should not just be interested in whether alliances are followed by war; we also need to determine the degree to which wars have been preceded by alliances. Levy does just that. He finds that wars were preceded by alliances less than a fifth of the time in the sixteenth century, less than a third of the time in the seventeenth, and less than one-half of the time in the eighteenth. Only two of the twenty wars (and none of the greatpower wars) in the nineteenth century were preceded by great power alliances and about onehalf of the twentieth century wars were preceded by alliances. Clearly, the large majority of wars have not been preceded by alliances. Levy therefore concludes that alliances are not a necessary cause of war. 133 Overall, the relationship between alliance formation and war (combined into ten-year periods) is negative and relatively low, contradicting the hypothesis that the greater the number of alliances in a given period, the greater the amount of war that is likely. On the other hand, it seems fairly clear that alliances neither prevent war nor foster peace. John Vasquez argues that the association between alliances and war may be traced to the fact that the construction of alliances leads to the creation of counter-alliances, which leads to greater insecurity and greater uncertainty owing to the probability that some alliance partners might welsh on their commitments. 134 He concludes: Since there is often an interval between the alliance and the outbreak of war, it is a legitimate inference that alliances do not directly cause war, but help to aggravate a situation that makes war more likely. They may do this in two ways: by promoting an atmosphere that polarizes the system and by encouraging arms races. 135

Levy warns that even those positive correlations that do exist between the creation of alliances and the onset of war may be spurious. That is, although they occur nearly simultaneously, there may be no direct causal relationship. Instead, the correlation may simply reflect the fact that alliances and wars are both being generated by the same underlying factors. 136 Indeed, even when alliance polarization precedes war, direct causation may not be present. The construction of alliances and the resulting polarization may merely be symptoms that reflect a more basic cause of war. Alliances don’t cause war; states merely form alliances because they believe war is near. The studies on polarization and on alliance formation indicate that alliance making that leads to polarization sometimes (but not always) produces wars of high magnitude, severity, and duration. Polarization increases the perception of threat; it focuses attention on issues that divide states rather than those that bind them together; it reduces the effectiveness of crosspressures; it reduces the number of effective mediators; and it may stimulate arms races. This is not the only pattern, but it is certainly one path by which war may be reached.

The International System Level of Analysis, Part I

401

One of the few patterns that researchers seem to be able to agree on is that, once wars begin, the presence of alliances serves to expand war activity throughout the system to more states. 137 Alliances act as contagion mechanisms for war, but alliances are not the only contagion mechanisms (geographic contiguity is another). 138 A consensus of sorts seems to exist that the construction of alliances is much more likely to be related to war than to peace—at the dyadic level and the international system level—but that this relationship is not perfect and is not consistent across time periods. Alliances don’t always lead to war; wars are not always preceded by alliances; and the relationship depends on the type of alliance and variables pertinent to the particular historic era. It might be useful to mention one side issue regarding alliances: Does the formation of an alliance reduce the chances that the two allies will go to war with each other? There was some initial speculation that just as most disputants in a social context are likely to be close friends or family members, wars should be most likely between closely related states. Familiarity breeds contempt—or something like that. 139 And there appeared to be some evidence to support this. However, this notion was pretty much blown out of the water by Stuart Bremer’s famous study on “dangerous dyads.” 140 Bremer finds that the statistical evidence for allies fighting each other is largely accounted for by the fact that they are geographically contiguous to each other. Once contiguity is controlled for, the relationship between alliance partnership and war disappears. More recently, Senese and Vasquez confirm a pacifying effect of alliances between allies. MIDs between allies are much less likely to escalate to war compared to dyads without a shared alliance. The fly in the ointment is that if one (or both) of the alliance partners have “outside” allies as well, this overwhelms and reverses the pacifying effect of the joint alliance. 141 POLARITY: EMPIRICAL RESEARCH Studies of the polarity—the distribution of power throughout the international system—have largely mirrored those on polarization and alliance formation: they produced widely varied and conflicting results. Some analysts claim to have found a pattern of war proneness in a particular kind of distribution; some have found relationships that differ according to the century investigated; and some have found no relationship at all between the distribution of power and war. The often cited study by Singer, Bremer, and Stuckey of the 1820–1965 period focuses on the magnitude of war in the system (as indicated by the number of nation-months at war), rather than on the number of wars or the onset of war. The independent variable in this study is a statistic labeled CON, which indicates the degree to which military, industrial, and demographic capabilities are concentrated in the hands of a few states in the system or are rather widely dispersed throughout the states in the system. 142 The authors find that in the nineteenth century diffusion of power was associated with low war magnitude, whereas a high concentration of power was associated with a high magnitude of war. The pattern for the twentieth century was just the opposite. In the later period, systems with highly concentrated distributions of power (preponderance) were associated with a low magnitude of war, whereas a diffusion (or balance) of power was associated with a high magnitude of war.

402

Chapter 10

Bruce Bueno de Mesquita reanalyzed the Singer-Bremer-Stuckey study using the presence or absence of war as the dependent variable instead of the magnitude of war. He concludes that the distribution of power is not related to the occurrence of war in either century. 143 Neither the concentration of power in the system nor the change in concentration of power seemed to make a difference. A further replication by Wayman uses CON as well as a more controversial statistic he calls TWO CON, the percent of great power capabilities held by the two most powerful states in the system. Wayman finds that years that were power multipolar (that is, in which power was diffused rather than concentrated) were slightly less war prone, but the wars that broke out in these systems were massive. Three-quarters of the wars in such systems were of high magnitude. On the other hand, three-quarters of the wars in power bipolar systems (that is, in systems where power was highly concentrated) were of low magnitude. The bipolar giants seem to have been able to manage conflict in these systems and prevent them from becoming system engulfing. 144 Ted Hopf examines two phases of the early modern state system in Europe. He finds the 1495–1520 period to be multipolar and the 1521–59 period to be bipolar and concludes that the amount of warfare in the two periods was essentially equal. The polarity of the system made no difference. 145 Similarly, Kegley and Raymond compare six multipolar periods over the last 500 years. They conclude that there is a great variety in conflict behavior across the six systems; multipolar systems are not all the same. 146 Manus Midlarsky’s historical analysis focuses on the transition from bipolar systems to those with a larger number of great powers. 147 His theoretical argument is that an increase in the number of great powers, coupled with a decline in the amount of valued resources in the system, would lead to a dramatic increase in inequality and therefore instability. Midlarsky notices that the Thirty Years’ War (1618–48) began when a transition from bipolarity (Catholic states and Lutheran states in Europe) to tripolarity (Catholic, Lutheran, and Calvinist states) was underway in Europe. World Wars I and II began as the number of resources (small, independent states and colonial areas) decreased while at the same time there was a mild increase in the number of states that joined the club of great powers (for instance, the United States and Japan). Jack Levy investigates the great power system for a larger period of history, 1495–1975, using a subjective classification of particular periods as multipolar, bipolar, or unipolar. 148 His interest is in wars between great powers and wars that involve at least one great power. Like Wayman, he concludes that although war occurs frequently in all systems, multipolar systems are slightly less prone to war than the other systems, though wars in multipolar systems tend to be more serious. Conversely, he discovers that large, general wars have not occurred in bipolar systems (for instance, 1495–1556 and 1945–75). Wars between great powers were more frequent in bipolar systems, though bipolar systems had wars of less severity and magnitude. Bipolar and multipolar systems seem to be equally stable in terms of the relative number of years at peace. Unipolar systems were by far the most war prone, and general wars occurred most frequently in unipolar periods. The results of the studies by both Levy and Wayman turn on its head the traditional realist wisdom that multipolar systems result in more frequent but less serious wars while bipolar

The International System Level of Analysis, Part I

403

systems result in less frequent but more serious wars. Their research suggests that just the opposite may be true. Finally, Bennett and Stam empirically investigate multiple explanations for war using data on MIDs from 1816 to 1992. They find some initial support for the thesis that bipolarity is more peaceful than multipolarity. They divide the data chronologically into three periods: 1816–1917, 1918–45, and 1945–92—the latter period being the only bipolar period. They discover that the risk that disputes will escalate to war were lowest in the post-1945 period; in fact the risk was 95–99 percent lower than for the 1816–1917 period. 149 This sounds good for bipolarity, but as Bennett and Stam note (and as we will see in chapter 11) quite a few theorists claim that the post-1945 period is a unipolar system dominated by American hegemony rather than a bipolar system. Ultimately, the authors argue that “structural arguments that simply break the international system into periods of multipolarity or bipolarity add little to our understanding of the antecedents to disputes or wars.” 150 The relationship between peace and the bipolar 1945 period could be accounted for just as well by a variety of equally plausible explanations, such as the existence of nuclear weapons. COMBINED EFFECTS: POLARITY AND POLARIZATION Since neither the concepts of power polarity nor cluster polarity on their own seem to be overwhelmingly helpful, perhaps a combination of the two concepts would make more sense. Attempts to focus on one factor may be confounded or influenced by the effects of the other. Wayman prefers to combine the concepts of polarity and polarization so that international systems can be classified on the basis of the fourfold schema presented in table 10.1. His research indicates that the systems that have been most prone to serious wars have been those that have been both cluster bipolar and power multipolar—that is, systems composed of a relatively large number of states with equal power that are aligned into two mutually exclusive blocs. The systems that have been most peaceful (at least in the twentieth century) have been both cluster multipolar and power bipolar. 151 Wayman contends that supporters of bipolarity are correct about the stability of bipolarity, as long as their argument is made in terms of power (bipolarity). Likewise, proponents of multipolarity are correct about the stability of multipolarity, as long as they mean cluster/alliance multipolarity. Other variables may interact with polarity and polarization in a way that reduces or increases the probability of war as well. Stoll and Champion suggest that differing results in research on bipolarity and multipolarity may be due to the relative strength of satisfied powers in the system. 152 Different distributions of satisfied/unsatisfied states should produce different behaviors. They hypothesize that when satisfied states hold a relatively low percent of systemic capabilities, there will be a strong positive relationship between concentration of capabilities (i.e., preponderance) and war. A test of this hypothesis gets mild support (though they use nation-months of war as the dependent variable rather than onset of war, and the determination of whether a state is satisfied or not is admittedly “soft”). They argue that if a dissatisfied major power wants to change the status quo, then an equal military distribution can keep the peace. Of course, if the level of satisfaction within the system is high, polarity is less relevant.

404

Chapter 10

The difference between status quo states and revisionist states would seem to be important. Indeed, it is a distinction of great relevance to two theories we will discuss in the next chapter—status discrepancy theory and power transition theory. However, before we move on, we should take one last look at a theory that links systemic polarity to war. POLARITY AND OFFENSIVE REALISM: MEARSHEIMER’S BALANCED AND UNBALANCED SYSTEMS John Mearsheimer’s The Tragedy of Great Power Politics is one of the most well-known efforts to explain great-power war as a result of system polarity. Remember that Mearsheimer argues that anarchy compels states to maximize their relative share of world power in order to attain security and that all great powers therefore seek revisionist goals—changing the distribution of systemic power in their own favor. For Mearsheimer, the only real security lies in hegemony—becoming the most powerful state in the system. However, he argues that global hegemony is virtually impossible due to the difficulty of projecting military power across oceans; he maintains that no state has achieved global hegemony in the modern era. The best that is feasible is the establishment of regional hegemony—as the U.S. has done in the Western hemisphere. Numerous attempts to create a regional hegemony in Europe have all failed. 153 While anarchy is at the root of the problem, it cannot account for war by itself. Anarchy is a constant, but war is not. Therefore, variation in other characteristics of the international system—like the number of great powers (polarity) and the distribution of capabilities among them—must cause variation in war. Variation in the number of great powers determines whether the system is unipolar, bipolar, or multipolar. (Mearsheimer claims unipolarity is only a theoretical category; it has been nonexistent in the real world.) The major consideration concerning the distribution of power is the gap between the number one state in the system and the number two state. If the gap between them is small, the system is balanced; if the gap is large, the system is unbalanced and the most powerful state is an aspiring hegemon. 154 Variation on these two factors gives us three possible real-world international systems: balanced bipolarity, balanced multipolarity, and unbalanced multipolarity. Three other systems are mythical or near mythical: By definition, balanced unipolar systems cannot exist; (unbalanced) unipolarity does not exist in reality because of the difficulty of a single great power establishing global hegemony; and unbalanced bipolar systems would be extremely unstable and not likely to last long. Mearsheimer agrees with Waltz that balanced bipolarity is the most peaceful type of system; unbalanced multipolarity is the most prone to war; and balanced multipolarity (with no regional hegemon) lies in between the two. 155 Why is an unbalanced multipolar system prone to war? In an unbalanced multipolar system there are many great powers, but since the power gap between the number one state and the number two state is significant, the strongest great power is a potential regional hegemon. Such a state is unsatisfied; and it has both the capabilities and the incentives to make gains at the expense of its rivals. Its desire for regional hegemony creates fear among the other great powers, who are enticed by the situation to undertake risky measures to increase their security, thereby creating dangerous conflict spirals. In addition, in multipolar systems, with more great

The International System Level of Analysis, Part I

405

power dyads, there are simply more opportunities for war and a greater chance of miscalculation. In sum, great powers are driven by fear and unbalanced multipolar systems create the most fear. Historical examples of war-prone unbalanced multipolar systems would include the Napoleonic era (1793–1815), which featured France’s bid for regional hegemony in Europe; the Kaiserreich era (1903–18), which was dominated by Germany’s attempt at European hegemony; and the Nazi era (1939–45), which saw Germany’s second grab for hegemony. All systems ended in world wars (Mearsheimer calls them central wars). Overall, for the period 1792 to 1990, Mearsheimer finds only 46 years of bipolarity (the Cold War period from 1945 to 1990), 109 years of balanced multipolarity separated into three periods, and 44 years of unbalanced multipolarity also separated into three periods. His analysis indicates that world wars/central wars only occurred during periods of unbalanced multipolarity, though balanced multipolar systems tended to promote great power-great power wars that did not spread to central wars. His breakdown also shows that almost 80% of the years encompassed by unbalanced multipolar systems were war years—a figure far higher than for bipolarity (2.2 percent) or balanced multipolarity (18.3 percent). 156 Finally, Mearsheimer’s theory includes explanations of the balancing behavior of great powers. Some critics of realism have noted that great powers do not always balance in response to aggression or the build-up of capabilities. Mearsheimer argues that balancing is almost always the response to aggressive behavior; the question is who does it and when. Great powers may “pass the buck” and permit others to balance instead of themselves. In multipolar systems there are plenty of potential “buck-catchers.” What determines which of the great powers balance and which pass the buck? Geography has a lot to do with it: Common borders and contiguity promote balancing while geographic barriers create buck-passing. Mearsheimer argues that all great powers would prefer to buck-pass instead of balance. They will let others balance if at all possible, though when push comes to shove, they will engage in balancing. (Bandwagoning—allying with the aggressor or chief threat—is not an option for great powers.) Of course, in a bipolar system, the other great power must always balance. 157 CONCLUSION What should we conclude from this first round of system-level theories based broadly on the realist approach to international relations? First, with regard to issues of polarity, we can say that the relationship between power distribution and war may vary over time. Second, research efforts produce different results when different indicators of polarity, stability, and war are used. Third, it seems that the distribution of power is only weakly related to the onset of war, but may be associated with certain types of war. If a war occurs, the distribution of power may have a major influence on the kind of war fought. 158 Fourth, war occurs frequently in all types of systems—unipolar, bipolar, multipolar—and none can be credited with making a significant contribution to peace. Given the differing methodologies used and the divergent results achieved by these research efforts, it is necessary to conclude that the relationship between power distribution and war is far from completely understood. In fact, the attempt to use systemic polarity to explain the onset of war may be a theoretical dead end.

406

Chapter 10

With regard to issues of polarization and alliances, conclusions are equally tentative. First, the evidence regarding alliance polarization is just as slippery and unproductive as those on polarity, with results varying considerably from one time period to the next. Second, although the evidence regarding alliance formation is certainly not consistent, it appears that there is an emerging consensus that alliances frequently lead to war rather than peace; they certainly do not have a consistently deterrent effect as one might expect from balance-of-power theory. Whether alliances increase or decrease the probability of war depends on the type of alliances and the nature of the states involved. Third, it is also clear that the effect of alliance formation and alliance polarization on war during the Cold War period is different from previous historical eras. Beyond the exploration of issues of polarity and polarization, attempts to identify significant fine-grained system-level factors such as the offense-defense balance, make intuitive sense, but pose significant problems for testing and are far from achieving universal acceptance. Finally, it is not outside the realm of possibility that political scientists are simply barking up the wrong tree by placing so much importance on system-level factors. But there is a different way of looking at systemic factors that we have not yet addressed. In this chapter we have looked, for the most part, at theories that deal with power in a static form. For example, we have been interested in whether the presence of bipolarity or multipolarity leads to war, whether the presence of a polarized system leads to war, whether the presence of alliances leads to war, and whether the presence of offensive dominance leads to war. In the next chapter, we will take a look at the relationship between power and war in a dynamic way. We will examine several system-level theories of war that emphasize the crucial connection between changes in relative power and the onset of war and the cyclical nature of these changes in power distribution.

Chapter Eleven

The International System Level of Analysis, Part II Power Dynamics, Cyclical Theories, and Historical-Structural Theories of War

The old order changeth, yielding place to new. —Alfred, Lord Tennyson History may not always repeat itself, but it often rhymes. —Mark Twain

Most of the theories in this chapter deal with power dynamics—the effects of change in the relative power capabilities of states in the international system. Many of these theories might be categorized as belonging to a “hegemonic realist” school that attributes international peace and stability to the presence of a dominant hegemonic state that holds a considerable chunk of the system’s economic and military capabilities. Almost all of these theories incorporate insights from two important precursor theories: status discrepancy theory and power transition theory. Status discrepancy theory highlights the idea that states may be motivated for war by their dissatisfaction with the distribution of rewards and status within the system. Power transition theory links the concept of dissatisfaction to changes in the distribution of capabilities in order to explain the outbreak of major power war. PRECURSOR THEORY 1: STATUS DISCREPANCY Status discrepancy theory (also known as rank disequilibrium theory) is based on the sociological concept of stratification, which can be defined as the arrangement of units that make up a social system into “a hierarchy of positions that are unequal with regard to power, property, social evaluation and/or psychic gratification.” 1 The units that make up the social system— individuals, groups, or states—carry out different roles in the social division of labor, and they can be ranked (stratified) on a number of different criteria or dimensions. 2 The chief assumption is that one’s position within this stratified structure plays a role in determining one’s behavior. 407

408

Chapter 11

Since every social system, no matter what size, may be stratified, even the global system can be seen as a stratified social system. States might be ranked according to a wide variety of different criteria: military power, economic stability, internet connectivity, literacy, technological sophistication, diplomatic reputation, productive capacity, possession of nuclear weapons, and so on. The most common approach is to collapse these criteria into three dimensions corresponding to sociologist Max Weber’s classic categories: class (economic might), power (military capability), and status (prestige). 3 Imagine for simplicity’s sake that states might rank in one of two positions—high or low— on these three criteria: economic power, military power, and status. If we refer to high rank as topdog (T) and low rank as underdog (U), then there are a limited number of possible configurations. Total topdog states are ranked TTT on all three dimensions, while total underdog states will be ranked UUU on all three dimensions. Both of these types of states are said to be in rank equilibrium; their ranks on all criteria are the same. On the other hand, states with rankings such as TTU, UTT, TUT, TUU, UTU, or UUT would be in rank disequilibrium; their rankings lack consistency. Just as the profile of individual states will vary, international systems will also have widely different profiles. Some international systems might be made up of states that are totally equilibrated, having no states with rank discrepancies; other systems might include a large proportion of states with rank discrepant profiles. Given these situations, two related sets of hypotheses have been put forward, one at the state level and the other at the systemic level. Johann Galtung has suggested that states that are in rank discrepancy are more likely to participate in war than states that are in rank equilibrium. And Maurice East has hypothesized that the more rank discrepancy that is present in the international system, the more war it should experience. 4 Why should status discrepancy lead to war? A state’s behavior is believed to be causally linked to its position in the international hierarchy. Total topdogs should be relatively peaceful because they have already attained most of the rewards available in the system and are thus satisfied members of the community. On the other hand, total underdogs may be deprived and thus dissatisfied, but they lack the necessary resources to successfully force a change. However, the situation for the rank-discrepant states is different. They are likely to have suffered differential treatment—esteemed because of their objective achievements, say, in the military and economic realms, but lacking in social status ascribed to them by others. This differential treatment produces a destabilizing pressure for upward mobility. If no peaceful channels are available, upward mobility may be sought through violence. Leaders of states, like ordinary individuals, may perceive aggression to be a necessary response to frustration. 5 Rank-discrepant states are relatively deprived (and thus dissatisfied) states, but unlike their cousins, the total underdogs, they possess capabilities that make an attempt to propel themselves upward a realistic possibility. They have motivation and resources, opportunity and willingness. Rank discrepancy would seem to be most dangerous if status lags behind the military and/or economic dimensions. This is the classic situation in which a state has achieved certain heights, but has been denied ascribed status by its peers. These overachievers are much more likely to feel they deserve a fairer deal than rank-discrepant states that are underachievers, with high ascribed status but low actual achievement on military or economic dimensions. Overachievers typically blame the system for constraining them. On the other

The International System Level of Analysis, Part II

409

hand, “underachievers” are usually placed in the position of having to defend their undeserved status against upwardly mobile “overachievers”—a potentially violent situation. 6 With regard to the structure of the international system itself, East posits that systems with a fairly high degree of rank consistency will be more peaceful than those without such consistency. His reasoning is that rank congruence “makes for less ambiguous behavior, more clearly defined roles and role-expectations, and less incentive for social change.” 7 Total topdogs will have little desire to change the system and total underdogs will have neither the material resources nor the inspiration to change. “It is only in the system with status discrepancy where the resources necessary for initiating social change are in the hands of those with the motivation and skill to better themselves.” 8 Rank disequilibrium is not seen as a necessary condition for war: Wars may occur in which neither participant is rank discrepant. Nor is rank discrepancy a sufficient condition: Rank discrepant systems and rank discrepant states don’t always become involved in war. The theory only suggests that given high levels of rank discrepancy for the system or for individual states, aggression becomes more probable. 9 One might well ask, just what level of analysis are we dealing with here? Status discrepancy theorists make reference to individual psychological variables such as the desire of leaders for achievement of their just rewards and their frustrations associated with the inconsistent ranking of their state. And certainly there is a focus on the nature of certain states—whether they are status discrepant or not. However, a central assumption of the theory is that national leaders will react to the situation of rank disequilibrium in similar ways, so that individual personalities may be regarded as unimportant. 10 Likewise, regardless of their other differences, states that share similar ranks within the system will behave similarly. Most important, whether a state is rank discrepant or not can only be determined relative to other states in the context of the international system. And to the extent that dissatisfaction exists, it is dissatisfaction with the distribution of rewards, resources, and privileges in the international system (or in a regional subsystem). Ultimately, therefore, the theory exists at the systemic level. Status Discrepancy: Implications Let us assume for a moment that this theory has some merit. What clues does it present for the possibility of peace? Generally, the theory implies that world leaders should devote their attention to creating and maintaining greater social justice in the system. Peaceful paths toward global social mobility ought to be identified and pursued. At a minimum, states with substantial economic or military capabilities ought to be rewarded with political influence and access to positions and roles that reflect their material attainments. One could argue, for instance, that peace was preserved in the Congress of Vienna era by permitting France to return as a full-fledged member of the great-power system after her defeat in the Napoleonic wars. To have done otherwise would have been to create a permanently dissatisfied state. On the other hand, instability in Asia in the post–World War II era may have been due, at least in part, to the fact that while the People’s Republic of China emerged as a major power after the communist victory in 1949, it had to wait over twenty years to obtain a seat in the United Nations and its rightful place in the international order. In the twenty-first century, great

410

Chapter 11

powers (especially the United States) will have to further accommodate the rise of a resurgent China, as its dynamic (and now semi-capitalist) economy fuels rapid growth. Status Discrepancy: Empirical Evidence Is there any evidence to support a status discrepancy explanation of war? Unfortunately, really strong confirmation is nowhere to be found, though there is some modest support. East investigated the association between status discrepancy (between economic and military ranking, on the one hand, and diplomatic prestige on the other) and the presence of war in the international system between 1948 and 1964, using data for 120 states. He was able to validate the systemic level hypothesis, but found only modest statistical correlations. Lagging the conflict variable two years behind status discrepancy increased the strength of the relationship somewhat. 11 Michael Wallace investigated a slightly longer period, from 1920 to 1964, and once again found a moderately clear association between the level of status discrepancy in the system and battle fatalities (as well as increases in arms levels). Like East, he found that the relationships were stronger when the war data were lagged, this time approximately fifteen years. 12 Investigations by James Lee Ray and Charles Gochman have produced even more modest findings and great ambiguity of interpretation. 13 This may be due in part to the use of different indicators, but it may also be due to the fact that their studies are confined to the European system. Gochman limits his investigation to the major power subsystem from 1820 to 1970. This is a very nice temporal domain, but it is a severely limited political and spatial domain— containing only nine states whose variations in capabilities and status, at least compared with the variation in the rest of the world, are fairly restricted. For example, the United States is not considered to be part of the major power subsystem until 1899. This is the standard historical opinion, based on the American victory over a minor European state, in the Spanish American War of 1898. But if the United States had been included just one year earlier, Gochman would have had another example of a status-discrepant power engaging in a war that results in the improvement of its position in the international system. The point is that status discrepancy theory seems more appropriate for just those sorts of situations—where a state outside the major power subsystem achieves economic and military strength without attaining the political and diplomatic status that usually accompanies it, and then engages in aggression, the result of which is to bring it just that status. Status discrepancy theory would seem to be most applicable, in other words, to those states on the political periphery of the system who desire entrance to the club. Indeed, Gochman finds that his “inequity model” is most applicable to rising powers on the periphery (such as the United States, China, and Japan), but does a rather poor job of explaining war behavior of the traditional European topdogs (such as France, Germany, Austria-Hungary, Italy, and Russia/ USSR). Ultimately, status discrepancy is probably only a subsidiary factor in the cause of war. Not all wars are related to status discrepancy and not all situations of status discrepancy lead to war. Even in those wars in which the participants are status discrepant, one may be hardpressed to conclude that status disequilibrium itself was the primary cause of war. However, as

The International System Level of Analysis, Part II

411

we shall soon see, status discrepancy is frequently linked to other, more important variables at the international system level. PRECURSOR THEORY 2: POWER TRANSITION THEORY We have been looking thus far at systemic variables in a fairly static way—focusing on the structure of the system at a given time, whether it is characterized by bipolarity or multipolarity, by alliance polarization or lack of polarization, by status discrepancy or status equilibrium. But we must also take notice of the importance some scholars have attributed to changes in some of these variables. A. F. K. Organski has put forward a theory of war based primarily on changes in the distribution of power in the international system. 14 The theory—power transition theory—was initially applied only to major wars involving the top great powers in the international system, though it has been extended beyond this limited realm in more recent applications. Organski challenges the classical realist balance-of-power thesis that equality of balance keeps the peace. This idea was based on the argument that equal power should be sufficient to dissuade adventuristic aggression, while superiority tempted its possessor to act like the neighborhood bully. As long as there was an equilibrium in the system and power was balanced against power, war should be deterred. At the very least, realists would agree that states will take action to balance against the threat of hegemony, and the result of this should be that hegemony is rare or non-existent in the international system. Parity and equilibrium should prevail. But as Inis Claude once famously stated, “If an equilibrium means either side may lose, it also means that either side may win”—thus tempting both sides to initiate war. 15 This is Organski’s starting point. Organski and his followers argue that in each historical era a single dominant state usually leads the international order as the head of a coalition of satisfied states. (The international system is seen not as anarchic, but as more or less hierarchically organized.) The dominant state creates rules and norms for the system—regarding trade, diplomacy, the use of force, and so on. Naturally, the leading state creates rules that work to its own benefit and the benefit of its friends and allies. As long as the leader of this status-quo coalition enjoys preponderance of power, peace is maintained. (This makes Organski one of the godfathers of what might be called the hegemonic realist school of thought. However, while other theories refer to this system leader as a hegemon, power transition theorists generally do not use this term.) It is the inequality in the distribution of power between the leading state and its primary challenger— combined with the support for the status quo by the dominant state’s allies—that keeps the peace, not necessarily the overt intervention of the dominant state as a global peacekeeper. Under conditions of power predominance, it would be foolish for a challenger to initiate war, and the dominant state has little to gain or to fear. However, as potential rivals undergo industrialization and modernization, the old leader may be overtaken in terms of physical capabilities, creating a situation that frequently leads to war. The root source of war is the difference in size and rates of growth of the members of the system, which inevitably leads to “overtaking.”

412

Chapter 11

Unlike classical realists who emphasize the static balances of power between states, Organski emphasizes the dynamic character of power and distributions of power. At the core of such shifts in relative power are simultaneous increases in productivity linked to industrialization, increased manpower due to demographic growth, and an increase in the capacity of political elites to mobilize national resources. 16 Power transition theorists focus on economic wealth and power rather than military capabilities; they generally use gross national product (GNP) as an indicator of a state’s capabilities. Sudden changes in national capabilities upset the previous distribution of power, and the resulting unstable balance is dangerous. The more rapid the overtaking, the more likely war is to occur. Specifically, major wars are asserted to be most likely when the challenger catches up to the dominant state, impelling a kind of “rearend collision.” This rear-end collision—the rising challenger catching up and achieving relative parity with the dominant state—is called a power transition. 17 (Actually, the theory suggests the crucial condition is both static parity of power between the dominant state and the contender and dynamic shifts in power—leaving open the question of which one is most important.) The changes wrought by power transitions create a general expectation that someone’s ox is about to be gored. Windows of opportunity open for the challenger and windows of vulnerability appear for the declining leader. Moreover, a reshuffling of the distribution of power changes what each side can expect to gain (or lose) in the case of war. 18 And when the war does occur, it is “fought for nothing less than control of the international status quo.” 19 What about dissatisfaction? While power transitions leading to parity are deemed to be a necessary condition for war, they are clearly not sufficient. Not all power transitions lead to war. More is needed. Partly, what else is needed is for the rising challenger to be dissatisfied with the status quo. Organski believed that dissatisfied powers tended to be late-comers to the ranks of major powers, becoming competitors or contenders some time after the creation of world order by the global leader. The status-quo powers who are the beneficiaries of the global order are understandably reluctant to surrender a portion of their privileges to the newcomers, setting the stage for conflict. If, as international relations scholar Harvey Starr says, war can only occur if there is both an opportunity to go to war and a willingness to go to war, then parity provides the opportunity for war, but dissatisfaction provides the willingness. 20 Ultimately, power transition theorists conclude that “dissatisfaction with the status quo is an essential precondition for conflict.” 21 This should sound familiar. At least some of the explanatory logic for the power transition theory draws on the theory of status discrepancy. 22 There are, of course, some rising states that are satisfied with the status quo and have good relations with the dominant state, and these countries pose no particular threat. The most glaring historical examples of transitions that have not led to war are often explained as cases in which the overtaking state is a supporter of the status quo and is not dissatisfied. The U.S. overtaking of Britain and the Russian overtaking of Britain in the late nineteenth and early twentieth century are both viewed this way. It should be noted that early empirical studies did not deal with this issue directly; rising challengers who engaged in war against declining dominant states—like Germany in 1914 and again in 1939—were simply assumed to be dissatisfied. Why did the empirical research exclude this component of the theory? For the most part the answer lies with the extreme

The International System Level of Analysis, Part II

413

difficulty in devising a valid operational indicator for the concept of satisfaction. It simply defies measurement. More recently, scholars have measured satisfaction through a grab bag of indirect measures, with results that appear promising, but also problematic. 23 In sum, power transition theory proposes that a threefold combination of factors are crucial in bringing about major power wars: There must be a rising state that is also dissatisfied with the status quo and this state must achieve relative parity with the dominant state by overtaking it (achieving a power transition) by means of a faster rate of growth. War is deemed unlikely prior to challengers nearing parity (operationally defined as 80 percent of the leader’s GNP) and also after the challenger has surpassed the previous leader (by a 20 percent margin). It is also proposed that the rising challenger initiates the war prior to the actual transition and that the speed of the transition is important. Most of these propositions are open to question—even by supporters of power transition theory. Let’s look at a few of the arguments. 1. What is the timing of the war? When, exactly, is the war supposed to begin? Power transition theory hypothesizes that major wars are most likely when the power distribution between the dominant state and the challenger is approximately equal. Organski himself initially posited that the challenger would be most likely to initiate war before equality is actually attained, though there is considerable dispute about this point among power transition theorists. 24 Presumably, war initiation represents an attempt by the challenger to hasten the passage, perhaps due to the overconfidence brought on by rapid growth, perhaps due to impatience and frustration with the lack of accommodations for the rising state’s demands for change in the international order, perhaps by the temptation to make use of the opportunity to achieve a complete, outright victory. This seems a strange argument to some. Dale Copeland argues that there is, in fact, no logical reason why a state should attack while it is still rising in power. It should always choose to wait, because as long as its power is rising relative to its primary opponent, it can assure itself of an even better chance of winning a contest that begins later. 25 And as we will see later, in Copeland’s own study of major wars, it is the dominant (and declining) state that is the initiator. It turns out that the idea of a pre-parity attack by a rising challenger is not only logically flawed but also historically inaccurate. In their own investigation of four major power wars in the late nineteenth and early twentieth centuries, Organski and Kugler found that the rising challenger actually initiated war after overtaking the dominant state, but before the combined power of the challenger’s coalition had overtaken the dominant state’s coalition. 26 Tammen et al. later suggest challengers start wars after the point of transition. 27 In one formulation, power transition theorists declare that wars will occur shortly after a transition (after the point of parity is reached) but only if the old dominant state is risk averse and if the challenger is either risk acceptant or risk neutral. 28 2. Who is the war’s initiator? Organski and Kugler assume the weaker challenger initiates the war against the stronger, dominant state. The argument is that powerful, satisfied states do not start wars; they are the primary beneficiaries of the present system and have no interest in changing it. Dissatisfied states with inferior capabilities are unable to successfully challenge the dominant state; at best, they are able to mount sustained “cold wars” against the dominant state. 29 However, powerful dissatisfied states are another breed of cat. They are dissatisfied with the status quo in general and with their position in the system; they therefore desire to

414

Chapter 11

redraft the rules more to their liking. (They are especially dangerous if they have a much larger population than the dominant state.) Kugler and Zagare later add that the logic of the status quo means that the dominant state is risk averse and will not be likely to seek gains by attacking a rising challenger, even though the chances of victory are better early on and they decline the longer the dominant state waits. Logically, the challenger should be both dissatisfied and risk-acceptant. 30 The challenger knows that, by merely waiting, it can attain supremacy and change the status quo without the risk of war. Moreover, the declining state’s ability to resist the challenger’s demands for change will decline as its relative capabilities diminish. At some point, logically, the declining state will simply have to submit to the stronger challenger. Thus, if a rising challenger initiates war, it must perceive the status quo as so onerous that it prefers an early war to the alternative of accepting the status quo for the short term and waiting to challenge it later. 31 A further refinement by Kim and Morrow specify that the probability of war is increased by four conditions: (1) the existence of rough parity, (2) the challenger’s level of dissatisfaction with the status quo, (3) the existence of a risk-acceptant rising challenger and a riskaverse declining state, and (4) low costs of war. Notice that this later version of the theory does not rely on purely structural conditions of power parity and overtaking to explain war, but includes psychological attitudes toward risk. 32 As we have seen, power transition theorists believe war is most likely to be initiated by the rising challenger—at least under certain conditions. 33 One could just as well argue that the logic of the theory suggests that as the gap closes, both states are likely to see the situation as threatening. The two rivals are likely to become increasingly anxious about the situation and sensitive to changes in the distribution of power. The dominant state—fearing that the challenger will surpass it in power and will be unwilling to accept a subordinate position in the international order—might unleash a preventive strike against the challenger, hoping to avoid the inevitable. The lack of clarity in the balance of power would create a situation in which national leaders are likely to see either opportunities or threats in the external environment. But this is not an argument put forward by power transition theorists. They say little to nothing about the dominant state’s preventive motivations. They fall back on the argument that dominant states are satisfied (and risk averse) and will not use war to upset the status quo, even though war may be the only way of preserving the status quo. 34 Daniel Geller contributes some interesting evidence to this discussion. He finds two patterns for wars that occur during power shifts: (1) Prior to the transition, wars are initiated by dominant states (this is a preventive use of force). (2) Once the transition has occurred, the rising challenger is the initiator. 35 Both of these make intuitive sense: Dominant states have their greatest capacity to eliminate threats from rising challengers before the power transition is achieved, and challengers have an overwhelming incentive to delay the use of force until the overtaking has given them a better chance to win. A later study indicates that in both wars and militarized disputes among contenders, the initiators are just as likely to be the superior power as the inferior one. 36 Organski believed that, historically, the challenger’s premature aggression has been a strategic error. Regardless of the dyadic balance between the dominant state and the challenger, the dominant state’s alliance has usually been stronger than that of the challenger. The

The International System Level of Analysis, Part II

415

challenger frequently neglects to entice away the leader’s support and is therefore confronted with a superior coalition, leading to the defeat of the challenger and its bloc. The distribution of power between the rival alliances ultimately determines the outcome of the war. It is important to note here that power transition theory makes predictions exactly counter the balance of power theory. The logic of the hegemonic status quo leads to bandwagoning rather than balancing. States don’t necessarily join alliances against the dominant state; instead, they have powerful incentives to ally with it to defeat challengers. 37 Finally, over the long run, the defeated challenger recovers its power (in fifteen to eighteen years) and even surpasses some members of the winning coalition—a phenomenon Organski calls the “Phoenix Factor.” Wars do not prevent the rise of challengers in the long run, and attempts to arrest the gains of fast-growing sates may be doomed to failure. 3. Does the speed of the transition matter? Organski originally argued that the faster the rate of transition, the greater the probability of war. Presumably, if the rate of growth is relatively slow, the warning time is greater for the dominant state, and both states will have a chance to prepare themselves for the future in a more reasoned and realistic manner. Mutually beneficial arrangements for the transition can be worked out between the challenger and the system leader. Accommodations and compensations can be negotiated, and peaceful resolutions are more likely. On the other hand, if the rate of growth in the challenger is swift, neither state is likely to be adequately prepared for the transition, and miscalculations and precipitous actions are much more likely. 38 This hypothesis has not received much support and it is subject to logical problems as well. As Morrow argues, the growth rate should not matter much. If the growth rate is indeed high, then the declining state will probably be more willing to fight, but the rising challenger will probably be less willing. 39 Perhaps now is the time to mention two things that may already have become apparent to the reader. First, the power transition theory as originally formulated is not a general theory of war; it only attempts to explain certain exceptional cases—major encounters between the most powerful states in the system. 40 On the other hand, it is certainly capable of being generalized to all states within a system, whether a central system or a regional subsystem. There seems to be no compelling theoretical reason why it should apply only to hegemonic contenders. Indeed, Organski and Kugler themselves apply it to the Russo-Japanese War. Second, strictly speaking, the power transition theory is not an international-system-level theory; instead, it operates at the dyadic level. The act of overtaking is a dyadic overtaking, and the parity achieved by the overtaking is a dyadic parity. The theory is concerned with the interrelationship between two states—a dominant power and a rising challenger. It is a version of the systemic distribution of power writ smaller. It is included in this section for several reasons. (1) It essentially deals with balance-of-power issues. (2) It is embedded in a theory about the structure of the international system—as typically led by a dominant power—in which the peace in the system is maintained by the preponderance of the system leader. System stratification is an essential component of the theory. (3) The challenger’s dissatisfaction is with the larger system created by the hegemon, rather than a dissatisfaction with dyadic relations per se. (4) Quite a few international system level theories use power transition theory as a foundation.

416

Chapter 11

Power Transition Theory: Empirical Research In The War Ledger, Organski and Kugler make the first empirical assessment of the power transition theory. Since this study is the object of considerable controversy, it is necessary to discuss it in some detail. Organski and Kugler attempt to test the theory against the major wars involving the great power contenders from 1860 to 1975. They first divide the great powers into central states (European) and peripheral powers (such as Japan, the United States, and China), and then they further divide central powers into contenders and noncontenders. States were deemed to be contenders if they possessed at least 80 percent of the power capabilities of the dominant state. (They use GNP as their measure of national power.) If no such state in a particular twenty year period met this criteria, then the top three states in rank were deemed to be contenders. To determine the relevant wars, they put forward three criteria: (1) wars must include major powers on both sides, (2) the war must have a higher level of battle deaths than previous wars, and (3) the war must be one that would include a loss of territory and population for the loser. Four wars in the time frame appeared to meet these criteria—the FrancoPrussian War, the Russo-Japanese War, World War I, and World War II. However, since in 1904 Japan was a peripheral power and not a contender, the Russo-Japanese War is not considered as a war between contenders. This leaves Organski and Kugler with essentially five wars to consider: the Franco-Prussian War, the eastern and western fronts in WWI, and the eastern and western fronts in WWII. 41 In all, they have a total of twenty contender dyads. What do they find? In the four dyads characterized by preponderance, there were no wars. Likewise, in the six dyads involving parity without a transition, there were no wars. In the ten dyads involving both parity and a transition, half ended in war and half did not. Thus, in each case where there was war between contenders of relatively equal strength, a power transition was under way. Wars between contenders occurred only when there was a transition taking place. All five of the war dyads took place during a power transition or immediately after the transition. In addition, five other dyads experienced both parity and power transitions, but escaped war. Thus, 50 percent of all cases of power transitions led to war. 42 The authors conclude, therefore, that wars among major-power contenders occur only if a power transition is under way. Thus, power transition—the condition of dyadic equality and the dynamic process of overtaking—is a necessary, but not a sufficient, condition for war between contenders. As Lemke and Kugler later put it, “Parity sets the stage where decision makers can, but need not necessarily, choose major war as a viable alternative.” 43 They also find that the speed of transition is important, as hypothesized. Organski and Kugler also analyze the effects of power transitions and parity on noncontender great powers. They find (as expected) that the theory does not seem to fit noncontenders. Unlike the contenders, who only fight when their capabilities are relatively equal, noncontenders fight even in the presence of dyadic preponderance. For peripheral great powers and for central system members as a whole, power transitions appeared to be a poor predictor of war. The analysis is subject to criticism on a number of fronts. First, fairly lengthy (twenty year) time periods are used, thus providing a relatively easy test whether wars were preceded by transitions. Second, the small number of wars makes the analysis crucially dependent on the

The International System Level of Analysis, Part II

417

two world wars. Moreover, the five dyadic cases that supply the evidence for the theory all involve Germany and its rivals in the Franco-Prussian war and the two world wars. Furthermore, as Vasquez notes, the exclusion of the United States from the central system until 1945 (because it generally pursued a policy of isolation and had no alliances with the other great powers) even though it had become the dominant state in the system in terms of capabilities (GNP) has a particularly dramatic impact: If the U.S. had been included, there would have been no power transitions within the twenty years prior to either of the two world wars. Finally, the start of World War I does not fit well with the explanatory logic of the theory. It would be hard to argue that the war started due to a German-British rivalry, and Germany’s overtaking of Britain is largely tangential to the outbreak of the war. Germany did not seek to challenge Britain; indeed, she hoped to keep Britain out of the war. The relevant dispute (involving Austria and Serbia) hardly involved Britain. As Vasquez says, “At best it involved allies of her allies.” 44 Houweling and Siccama note that while Organski and Kugler test to see whether great power wars were preceded by power transitions, they do not really investigate whether power transitions are always followed by wars. Using a broader composite index of national power that incorporated elements of both size and economic development and military capability, an expanded time frame (1816–1975), and a larger list of wars, they find even stronger evidence than Organski and Kugler that power transitions are an important predictor of the outbreak of war. Moreover, their analysis shows that the relationship between power transitions and war applies not just to the small group of contenders at the top of the international pyramid, but to all great powers in the system. Depending on which group of major powers one looks at, power transitions roughly double or triple the prospects of major-power war. However, the relationship is generally pretty weak for all great powers; it is much better for the smaller group of contenders. Once again, power transitions lead to war about half the time or less. 45 Several other studies also lend some support to Organski and Kugler’s findings. Stoll and Champion, using the COW composite indicators of relative capability, agree that all of Germany’s wars with other great powers occurred when predicted by the power transition theory. The Austro-Prussian War, the Franco-Prussian War, and the two world wars all broke out within five years of the intersection of capability scores between Germany and its majorpower rival at the time, thus lending support to the parity-leads-to-war hypothesis. 46 Soysa, Oneal, and Park replicate the Houweling and Siccama study using both GDP and the Correlates of War (COW) index of capabilities. With regard to all great powers, their results were similar. Roughly 42 percent of the dyads that experienced overtakings experienced war within the twenty-year period. The picture was less strong for contenders (the strongest three or four states in the system), though a power transition was still the best predictor of war. One interesting finding was that if Japan and Germany are included as great powers in the post–World War II period, the results were much weaker: only two of thirteen power transitions since 1945 have resulted in war if Japan and Germany are included. This suggests that power transition theory may be less relevant to the post-1945 period. The authors speculate that the less dangerous effects of transitions may be due to the presence of more democratic states, the presence of nuclear weapons, and/or greater economic interdependence. 47

418

Chapter 11

Several studies focus on power transitions between enduring rivals. Charles Gochman looked at the conflict involvement of major power and non-major-power rivals from 1816 to 1980. 48 His analysis generally supports power transition hypotheses. He finds war is more likely among both major power and non-major-power rivals that are relatively equal in total (demographic-industrial-military) capabilities. But parity alone is not sufficient to provoke war; a power transition must be underway. Gochman finds that both major power dyads and non-major-power dyads are more likely to become involved in war when they are rapidly converging toward parity (or diverging away from it, though this finding isn’t as strong). He discovers that great powers are generally more likely to engage in war when their own total capabilities are increasing rapidly, but not when they are decreasing. These findings are consistent with Organski’s rear-end-collision thesis. Gochman concludes that rapid alterations in the relative strength of actors produce uncertainties with respect to intentions, or dissatisfaction with the distribution of benefits, for which the actors have insufficient time to adjust. More gradual change provides greater possibilities for accommodation or adjustment to new realities. 49

A somewhat similar study by Geller performs a Markov chain analysis of rival dyads between 1816 and 1986 using data on militarized disputes. 50 He finds that when one measures power strictly in terms of military capability, then static (military) parity between pairs of states is twice as likely to be associated with war as is static (military) preponderance. With regard to shifts in the distribution of power, he discovers that shifts in power—especially shifts in the direction of increased parity—are associated with a higher probability of war than are actual power transitions in which an overtaking occurs. In fact, power shifts toward parity (convergence) showed the strongest linkage with war. Thus both static parity and dynamic shifts toward parity between rival states are associated with war. One retrospective analysis of power transition theory notes that several researchers have noticed power transitions preceding wars between various small or middle powers: between Austria and Prussia prior to the Seven Weeks’ War, between Iran and Iraq prior to the 1980 war, between China and India prior to the Sino-Indian War of 1962, between North and South Vietnam, and multiple transitions between Israel and her Arab neighbors prior to several of the wars in the Middle East. 51 Finally, other researchers have made interesting arguments about the explanatory range of power transition theory. Lemke and Reed argue that what many see as a “democratic peace” is more likely due to the joint satisfaction of members of the dominant state’s coalition than joint democracy. Since for the last 200 years the dominant state has been a democracy (first Britain, then the United States), and since democratic states aligned with the hegemon are generally satisfied with the international order, the joint peace among these states probably reflects satisfaction with the international system rather than joint political structures. 52 Widening Power Transition Theory to Noncontenders? Since there is some evidence that power transition theory applies to the behavior of noncontender states, the question is raised whether power transition theory can be extended beyond

The International System Level of Analysis, Part II

419

just the top three or four states in the global system. One could certainly argue that the logic of overtaking should apply to all rival dyads, whether they are major powers or minor powers, whether they are contenders or noncontenders. For instance, one argument would simply be that power transitions among any pair of rivals would open windows of opportunity and vulnerability and raise tension, anxiety, and hostility, creating the potential for security dilemmas and conflict spirals that might escalate to war. That, of course, would be a dyadic explanation. The keepers of the power transition tradition have chosen to go in another direction. Douglas Lemke has created what he calls the multiple hierarchy perspective to incorporate minor states into power transition theory. 53 Remember that power transition theory embeds the dynamics of dyadic power shifts into the larger context of the international system led by a dominant state that creates an international order. To preserve this systemic perspective Lemke envisions the global system as containing multiple regional subsystems that operate like miniature global systems inside a particular region. Regional hierarchies are subordinate to the overall global hierarchy. (Lemke pictures this as cones within cones.) Within a regional hierarchy, a locally dominant state may create a regional status quo from which it and its friends benefit; other states in the region may lack such benefits—such as access to valuable territory and resources—and are thus dissatisfied. If one of these regional dissatisfied states increases its power and achieves parity with the old dominant state, then a war can be expected to determine the new regional hierarchy. 54 Wars in regional hierarchies are unlikely to diffuse upward to the global system as they do not threaten the global status quo. On the other hand, major wars in the global hierarchy diffuse downward to engulf regional states. Thus wars within regional hierarchies such as the RussoJapanese War, the Korean War, and the wars in Vietnam and Afghanistan did not spread beyond the region, while conflicts between global contenders in Europe in 1914 and 1939 did. 55 Lemke’s initial study is confined to South America, where he finds four consistent regional hierarchies: Atlantic Coast, Pacific Coast, Northern Rim, and Central States. For the 1865–1965 period he finds two wars relevant to the states in these regional hierarchies: the War of the Pacific (Chile vs. Bolivia and Peru, 1879–84) and the Chaco War (Bolivia vs. Paraguay, 1932–35). His analysis concludes that the relative absence of wars in the South American regional systems was due largely to the stability of local power hierarchies. The two sole instances of war occurred when states were equal in power and the weaker challenger was overtaking the stronger. Of the three power transitions that took place over the entire period, two led to war. (Lemke reckons the third to be a case of satisfied transition.) He concludes that power transition theory applies to both major and minor states and to non-industrial states as well as industrial states. At least with regard to the South American regions, parity approximates a necessary condition for war. Using Lemke’s approach, Tammen et al. expand the geographic scope of the multiple hierarchies perspective beyond South American to examine regional hierarchies in the Middle East, Africa, and the Far East. 56 Oddly enough, this later version of the theory abandons all discussion of power transitions per se; the importance of a rising challenger’s overtaking of the dominant state is simply not included. The analysis is restricted to two variables: the presence/absence of parity and the presence/absence of dissatisfaction. Support for the theory

420

Chapter 11

varies considerably by region. In the South American regions, as we have seen, both wars occur under conditions of parity and dissatisfaction by the challenger with the status quo. In the Middle East three of the four wars occur under the conditions of parity and dissatisfaction; in East Asia only two of the five meet both conditions; and in Africa neither of the two wars occurred under the expected conditions. Their conclusion based on the statistical analysis, is that the presence of parity and dissatisfaction increase the probability of war; war is deemed to be five to ten times more likely in the presence of joint parity and dissatisfaction. However, they note that the degree to which joint parity and dissatisfaction increase the probability of war is generally much higher for great power dyads than for the minor power dyads that they investigate. 57 Frank Wayman also believes that power transition theory can be extended to noncontenders, but for quite different reasons than Lemke. 58 Wayman’s explanation takes a decidedly dyadic route and jettisons the system-level concepts such as satisfaction with a systemic (or subsystem) status quo. His argument is twofold. First, it isn’t just that power transitions (overtakings that establish equality) increase the risk of war; any shift in power might have the same effect, even if it does not lead to an actual overtaking. The more general independent variable, then, is power shifts, which can be either power transitions or rapid approaches— changes in dyadic power that do not reach the point of transition. Second, all such power shifts are dangerous and increase the risk of war, but for reasons that have nothing to do with any systemic status quo. For Wayman, power shifts cause effects that are purely dyadic. First, as a state gains in power its geographic reach extends and the issues over which it is concerned extend as well, impinging on the interests of its neighbors. Second, shifts in power increase “the appetite in the gaining state and apprehension in the declining state.” 59 Third, responding to the growing appetite of rising states is problematic since conciliation and appeasement may increase their appetite for more concessions and reduce the ability to deter them. On the other hand, tough or hawkish responses may provoke them into war. Fourth, any power shift makes it difficult to determine what the actual military balance is and thus may entice both sides into war by permitting them to think that they may prevail. Fifth, power shifts may lead to preventive motivations for war by the declining state. Despite this last reason, Wayman—unlike Organski—makes no particular judgment about which state is likely to initiate war. The logic of power shifts makes both the rising and the declining state equally likely to choose war. 60 Wayman tests several hypotheses using the COW Composite Capabilities index (which combines military, industrial, and demographic capability indicators) for the major powers from the end of the Napoleonic wars to the end of the twentieth century. He compiles dyadic capability data into five year periods. He is then able to distinguish among several dyadic power conditions: power transition, reversed power transition (where power advantage changes and then changes back within the next five years), rapid approach, prolonged parity, and rapid approach with prolonged parity. During the same period he identifies twelve greatpower wars. Finally, he is able to distinguish some pairs of great powers as enduring rivals during certain periods of time. His findings are interesting but not overpowering. They are also remarkably similar to those of Gochman. First, parity in and of itself was not a dangerous condition. Mild prepon-

The International System Level of Analysis, Part II

421

derance and approximate equality were equally (and only mildly) war-prone. Second, all types of power shifts are linked to increased probabilities of war in the subsequent decade, although rapid approaches were actually more dangerous than real power transitions. Power transitions led to war 19 percent of the time, but rapid approaches led to war 30 percent of the time. (Rapid approaches with prolonged parity led to war 25 percent of the time; reversed transitions led to war 14 percent of the time. Prolonged parity led to war only 10 percent of the time.) 61 It would appear that the crucial condition is that any kind of power shift produces an increased probability of war. The dynamic aspect of the shift is the important part, not that actual power parity exists—a conclusion quite opposite of that taken by many new wave power transition theorists. Wayman recognizes that power shifts are neither a necessary or sufficient condition for war. Power shifts only lead to war a fraction of the time. And wars can occur in the absence of power shifts. Given that the strength of the relationship is rather weak, what are we to make of the situation? John Vasquez suggests that we should see power shifts simply as “one factor that increases the probability of certain types of wars, especially when combined with other factors.” 62 If power shifts don’t lead to war on a regular basis, then what kind of power shifts are most likely to lead to war? Wayman’s answer is that power shifts among enduring rivals are much more likely to lead to war than power shifts between nonrivals. (The existence of rivalries implicitly suggests the presence of mutual dissatisfaction between states—an important connection with power transition theory.) According to Wayman, any type of power shift among rivals approximately doubles the risk of war from 14 percent to 31 percent. For nonrivals it increases the risk of war from 8 percent to a meager 14 percent. 63 Wayman’s findings are mirrored by those of Huth and Russett, who discover that enduring rivals are especially sensitive to power transitions. The occurrence of power transitions within the context of an enduring rivalry increases the probability of war for that dyad by more than 21 percent. 64 Power Transition: Second Thoughts Before we get overexcited that one of the theories in this chapter seems to have received consistent validation, we ought to mention several studies that provide contrary evidence. An early effort by Peter Wallensteen looks at the dyadic relationships of all great powers from 1816 to 1976 and finds no relationship between transitions in economic strength and war within a ten-year period. Of the fifty-three dyads examined, eleven had power transitions but only three resulted in war. The first two were Prussia overtaking Austria prior to the Seven Weeks War and Prussia overtaking France prior to the Franco-Prussian War. The third was China overtaking France prior to the Korean War—an anomaly created by France’s marginal participation in the UN effort. 65 Several studies by Woosang Kim provide negative evidence as well. 66 Kim’s initial investigation of Organski’s theory uses an expanded list of wars and the same composite capability indicator as Stoll and Champion, but uses a different set of statistical tests. He finds most of Organski’s hypotheses to be too weakly supported to grant confirmation. The distribution of power between two states seems to have little to do with the incidence of war. In fact, neither the equality-of-power-leads-to-war hypothesis nor the preponderance-leads-to-war hypothesis

422

Chapter 11

is supported. In addition, Kim discovers a negative relationship between power transitions and war; war, in fact, seems less likely when one state is overtaking another. Subsequent analyses which investigate power relationships among the great powers going back to 1648 appear to confirm that neither power shifts (toward parity) or power transitions are associated with major wars, although rough parity was deemed to increase the probability of major power war if alliance considerations are included—specifically when one adds the expected contributions of third party joiners to the dyadic power relationship. 67 Finally, Bennett and Stam’s gigantic empirical analysis designed to judge the comparative power of various hypotheses about war is a decided downer with regards to power transition theory. The authors use data for MIDs, MID escalation, and war for all dyads and for politically relevant dyads between 1816 and 1992. Among the variables included in their study are static parity of power, the relative satisfaction of states (determined by the degree to which their alliance portfolios of the two states in the dyad match each other), and their own indicator of power transition (a measure of the difference in growth rates for states having equal power). They find that most variables associated with the power transition theory have weak and inconsistent effects on the probability of war. While static preponderance clearly makes a dyad more likely to be peaceful than equality, the transition and satisfaction variables don’t add much to the finding. Even when the analysis is restricted to dyads involving the system leader, the effect of power transition on the probability of disputes and their escalation to war is virtually negligible. 68 A small digression is in order at this point. We have noted in chapter 7 the rather robust evidence from researchers at the dyadic level that static balances of equality (or near equality) are much more dangerous than balances characterized by preponderance. In response to these findings, many scholars in the power transition community have now essentially jettisoned the transition part of the theory and simply argue that what matters is the fact of power parity rather than the actual process of overtaking and transition toward parity. Even some of the most vigorous proponents of the power transition approach now suggest that power transition theory should be renamed parity theory. There is no consistent theoretical justification to expect war immediately prior to or after transitions. Theoretically, it is parity that is important to war initiation. The closer to parity a dyad is, the greater the threat of war. Parity, not the actual transition, is of theoretical importance. For this reason it would have been better if Power Transition Theory had been named Power Parity Theory. 69

Woosang Kim’s formulation of an “alliance transition model” can be seen as buttressing these arguments. 70 Kim seriously modifies Organski’s original theory. He believes that the original limitation of the power transition model to the industrial age is misguided. He also disagrees with Organski’s emphasis on internal growth as the only way for states to add to their power; alliances have always played an essential role here. Thus Kim’s measure of power consists of a state’s internal capabilities adjusted for the support that the country can expect from other great powers. In this way he can create an “alliance equality” variable, an “alliance transition” variable, and an “alliance growth rate” variable. (He also incorporates a measure of dissatisfaction that is based on the degree of congruence between the alliance portfolios of each state

The International System Level of Analysis, Part II

423

and the dominant state.) His collective findings indicate that for the entire period after 1648, neither dyadic power transitions nor alliance transitions have a significant impact on war; the speed of the transitions and the actual overtaking also appear unimportant. On the other hand, alliance equality and dissatisfaction both increase the probability of war. Two points are in order here. First, we should also note that wars do, of course, occur in situations of dyadic inequality (preponderance); large countries fight small countries—think U.S. versus Iraq, for instance, or the Russo-Finnish Winter War. But as John Vasquez notes, preponderance is probably associated with different kinds of war—imperial wars and wars of choice—and have their own dynamics and a different set of causes. 71 Second, while the evidence seems clear that parity of power is conducive to war at the dyadic level, this does not necessarily mean that power equality at the international system level is associated with war. Power Transitions: Conclusions While the validity of power transition theory has by no means been determined, it appears as one of the more interesting attempts to explain great-power war. It is clear that dyadic equality (or near equality) of power doesn’t always lead to war between opponents; it is also clear that power transitions don’t always eventuate in war. However, the coincidence of power transitions and dyadic parity does seem important. It appears possible that rapid changes in power capabilities, especially those that lead toward dyadic parity, make war between major power rivals and (perhaps) between non-major-power rivals more likely. It is also likely that power transitions toward equality are associated with other factors—such as systemic power deconcentration, arms races, the perceptions of threat—which further contribute to the likelihood of war. However, the evidence is not nearly as strong as one would hope for. Vasquez suggests that given the relatively weak direct connection between power transitions and war, their primary effect may be to produce militarized disputes, which may then escalate to war if states resort to hawkish policies. There is in fact some evidence that dyadic power shifts (power transitions and rapid approaches) are followed by MIDs. 72 The power transition theory makes intuitive sense; it is consistent with our sense of the probable; it has internal logic; it is fairly parsimonious; and it is linked to important factors at other levels of analysis that also contribute to the initiation of war. Unfortunately, the pile of empirical evidence put forward on its behalf is contradictory. That being said, major shifts in the relative balance between rival states seem to be an important part of the war puzzle. And despite a smattering of findings to the contrary, a significant amount of evidence suggests that we should also have serious doubts about the traditional balance of power hypothesis that equality of power leads to peace. 73 However, the power transition theory itself is probably not the all-purpose, industrial-strength theory that many had hoped for. Since power transitions seem to be neither necessary nor sufficient conditions for war, perhaps the essential question is when or under what conditions do power transitions lead to war? Power transitions theory itself argues that dissatisfaction is one such condition. A second condition, previously noted as well, is the existence of rivalries; rival states are particularly sensitive to power transitions. Two other conditions worth noting occur at the systemic level and these are incorporated in theories that we will view shortly: systemic deconcentration of

424

Chapter 11

power and systemic shifts in power such as critical points in a country’s power cycle (which we will explore later in this chapter). Geller’s studies of contenders and other great powers from 1816 to 1976 demonstrate a synergistic interaction between the distribution of power at dyadic and systemic levels. Dyadic power parity (and power transitions) are associated with war, but only if there is a systemwide deconcentration of power underway. (Systemic power deconcentration means that the percent of capabilities controlled by the system leader is declining and the distribution of power in the system is moving away from unipolarity and toward multipolarity.) 74 Houweling and Siccama also demonstrate a link between dyadic power transitions and more general systemic transitions. A dyadic power transition, combined with at least one state in the dyad going through a critical point in its relative power cycle, combines synergistically to increase the probability of war. 75 Let us continue our examination of the relationship between power preponderance and peace (and between power equality and war) by turning our attention to theories of the international system that focus on the cyclical rise and fall of global hegemonic leaders and the connections of these cycles to war. COPELAND’S DYNAMIC DIFFERENTIALS THEORY Having explored balance of power issues and polarity in the previous chapter, and having looked at power transitions, we can now profitably look at a theory that combines insights from all three—Dale Copeland’s dynamic differentials theory. This is a theory that, like most of the theories in this chapter, has been devised to explain what have variously been called world wars, global wars, hegemonic wars, or major wars. As Copeland defines them, major wars are those that involve all the major powers in the system, are fought at the highest level of intensity by all parties, and include the possibility that at least one great power might be eliminated as a sovereign state as a result of the conflict. 76 Copeland starts with the assumption that states are unitary rational actors whose primary goal is to maximize security. In the fashion of realist thinkers, he also assumes that states are largely uncertain about the current intentions of other states in the system and are completely uncertain about their future intentions. They also understand that in attempting to ensure their own security, they may run the risk of major war—either directly or indirectly, through the possibility that crisis interactions may escalate to inadvertent war. To intentionally risk either of these requires substantial motivation and this is likely only when a dominant military great power is faced with a deep and inevitable decline relative to other states in the system. In both bipolar and multipolar systems the declining power is most likely to start a war. Logically, rising states have no need to initiate war; it makes more sense simply to wait until they have achieved superiority. Copeland’s theory is essentially a theory of preventive war—war undertaken by a dominant state in the realization that its power position relative to others is deteriorating, that there is little short of war that can reverse the decline, that when the rival state achieves superiority it is likely to act aggressively, and that since war in the future is likely, it would be preferable to fight now when the chances of victory are still reasonable. Preventive war is seen as an

The International System Level of Analysis, Part II

425

option of last resort. 77 The theory takes as its starting point Organski’s pivotal setting: a looming power transition between a dominant (but declining) state and a rising challenger. However, contrary to Organski, he argues that the war’s initiator will be the declining state rather than the rising challenger. When the declining power initiates the war depends on its evaluation of two things: the extent of its decline and the inevitability of the decline. The deeper the perceived decline and more inevitable it appears, the more likely a state is to turn to preventive war. This is especially true if the dominant state has an advantage in military power, but is inferior in economic power and potential power (land, natural resources, large population, etc.). 78 In Copeland’s theory the current leader is typically a state that is militarily dominant but is modest in its land area, resources, and population. It is challenged by a state whose potential power gives it the economic and military ability to overwhelm the current leader once it is able to mobilize its vast potential. All of the foregoing is set off by dynamic changes or shifts in the distribution of relative power. The situation that Copeland describes is one in which a true power transition has not necessarily taken place: The rising challenger has not yet achieved parity. And although the theory is primarily at the international system level, the power shifts are for the most part shifts in the dyadic balance between the top state in the system and a rising challenger. Copeland’s concept of dynamic differentials refers to the “simultaneous interaction of the differentials of relative power between the great powers and the expected trends of those differentials.” 79 Notice that the theory does not require that states have innate aggressive motives (found at the individual, substate, or state level of analysis). Aggressive motivation is not a sufficient condition because even leaders with aggressive individual level motivations will be deterred by the lack of the proper power capabilities. Neither is aggressive motivation a necessary condition, because even if the declining state’s primary adversary is defensively motivated and currently nonaggressive, the declining state has reason to fear future aggression once the adversary has become militarily dominant. It is the uncertainty of future motives and the fear of future aggression that drives the logic of preventive war. 80 These features place dynamic differentials theory solidly in the camp of defensive realism. Incorporating ideas of polarity, Copeland argues that the constraints on the declining state’s ability to initiate major wars (or crises that might escalate to wars) differ depending on whether the system is bipolar or multipolar. His argument is that in a multipolar system, the declining state must have a significant level of military superiority (perhaps just less that 50 percent of the system’s total military capability) to rationally entertain thoughts of war. This is so because in a system of many great powers, the declining state may have to defeat more than one enemy. In both 1914 and 1939 Germany (a militarily dominant, but declining state) had to fight against Britain and France in order to destroy Russia (the rising challenger), and the fact that it nearly did so, despite little real help from its allies, proved that it began the war from a position of military superiority. On the other hand, in a bipolar system the declining state needs only rough parity with the other great power, and may even get away with slight inferiority. Success is easier since there is only one great power adversary to defeat. 81 In

426

Chapter 11

neither system is decline by itself enough to bring about major war; declining states must have sufficient power to make the possibility of victory in such a war feasible. Copeland suggests that there are five pathways to major war, four of which are consistent with his theory. First, given its declining relative power, a state may simply initiate a preventive war, or a crisis that will become the pretext for a major war. The declining state clearly prefers major war to the status quo (this is called a “deadlock preference”). Second, as in pathway number one, the declining state prefers major war to the status quo. But instead of directly resorting to preventive war, it provokes a crisis to coerce concessions that might moderate or reverse its relative decline. When those concessions are rejected by the opponent, preventive war is initiated. Third, the declining state (and its opponent) prefer peace to war—a “staghunt” preference. Crisis initiation is the first option for the declining state, but given an environment in which there are advantages in striking first, interaction between the declining state and its rival are characterized by a hair trigger. Inadvertent preemptive wars are likely, as both sides will see striking first as preferable to waiting. Thus declining power trends lead to preventive motives, which lead to the initiation of crises, which lead to inadvertent war through preemption. Fourth, once a crisis is initiated by a declining state, leaders on both sides find it difficult to back down. Their reputations are overcommitted, and they tend to prefer war than see their reputations diminished. 82 To test his theory, Copeland uses primary documents that reveal the motivations of the leaders of major powers in three periods: the pre–World War I and pre–World War II periods (both multipolar) and the early Cold War period up to the 1962 Cuban Missile Crisis (a bipolar era). He argues that the documentary evidence indicates that in both 1914 and 1939 Germany was the dominant military power, but German leaders saw the country sliding into an inevitable relative decline due to concerted attempts by Russia to modernize its industry and military and make use of its tremendous potential power based on its land area, natural resources, and population—all vastly larger than that of Germany. In both cases, German leaders consciously launched preventive wars while they were still militarily superior in order to forestall Russia from overtaking Germany as the dominant state in the system. Copeland also examines the seven major wars prior to 1900. The first three occur in bipolar systems: the Peloponnesian Wars between Sparta and Athens (431–404 BCE), the Second Punic War between Carthage and Rome (218–202 BCE), and France-Hapsburgs wars (1521–56). As predicted, the initiators—Sparta, Carthage, and France—had all previously been the dominant military power, were all experiencing a deep and terminal decline, and all initiated preventive war at a point at which they were relatively equal to their rising rival. The next four cases occur in multipolar systems and three of them fit the predicted pattern: the Thirty Years’ War (1618–48), the wars of Louis XIV (1688–1713), and the Napoleonic wars (1803–15). The initiators—Spain, France, and France—were dominant states militarily with preventive motives due to impending long-term decline, but they still possessing a measure of superiority at the war’s outbreak. The Seven Years’ War (1756–63), was another war initiated by a declining formerly dominant state: For preventive reasons, Austria (in alliance with Russia) provoked Prussia into a preemptive war, but Austria was no longer superior (as the theory predicts), and the military balance was one of parity.

The International System Level of Analysis, Part II

427

Copeland claims his theory to be superior to three other theories at the systems level: (1) classical realism—which expects war when a state possesses preponderance but not when there is a dyadic balance of power or a balance between alliance blocs; (2) structural realism— which believes war is likely in multipolar systems but unlikely in bipolar ones; and (3) hegemonic stability theory—which believes that shifts in power bring about wars, but expects rising states rather than declining states to be the initiators and expects war when the balance approaches parity. HISTORICAL-CYCLICAL THEORIES OF WAR Several theorists have turned their attention to the historical evolution of the international system and in particular to fluctuations in the concentration of power as reflected in the rise and decline of the system’s leader and the wars that attend these phenomena. Proponents of these historical-structural theories believe that to understand the current international system, one needs to know how this system has evolved historically. In many ways the international system of today constitutes merely a modification of the structures and processes of the international system that existed, say, in the sixteenth century. Many of the patterns and cyclical processes can be understood by examining their origins and development in previous international systems. Let us turn our attention to several of these theories that deal with cycles of war between dominant states and challengers. The main contenders are Robert Gilpin’s theory of hegemonic war, George Modelski’s long cycle theory, and Emmanuel Wallerstein’s theory of the capitalist world system. We will then look at Charles Doran’s cycle of relative power. Though this last theory departs significantly from the first three, it also looks at the rise and fall of relative power in the international system—though it declines to put these cycles in the context of cycles of global hegemonic leadership. GILPIN’S THEORY OF HEGEMONIC WAR Robert Gilpin’s theory is outlined in his War and Change in World Politics. 83 Like Organski’s original power transition theory (and Copeland’s dynamic differentials theory), Gilpin’s theory is not a general theory of war but a medium-range theory that addresses wars fought between all major powers for leadership in the international system—hegemonic wars. Other types of wars are believed to be due to a different set of dynamics. The theory is also limited to the period following the Thirty Years’ War and the creation of modern nation-states after 1648. Like Organski, his attention is focused on the dominant state in the system, which Gilpin calls a hegemon. These states’ leadership of the system has, historically, been set in motion by military victory in a systemwide hegemonic war. The hegemon’s leadership is based on its simultaneous military and economic dominance and on its ability to provide certain public goods to system members: military security, investment capital, an international currency, a secure environment for trade and investment, a set of rules for economic transactions and the protection of property rights, and the general maintenance of the status quo. In return for the provision of these collective goods, the hegemon receives revenues and other benefits. 84

428

Chapter 11

Gilpin is interested primarily in wars fought for dominance in the international system. Hegemonic wars are direct contests between the dominant power(s) and a rising challenger over the governance and leadership of the international system. In his view, war arises because of an increasing disequilibrium between the political organization of the system, on the one hand, and the actual distribution of power capabilities, on the other. As the reigning hegemonic state gradually loses the dominant economic and military position it once held, the hierarchy of prestige in the system and the hierarchy of power are no longer compatible. (If this sounds a lot like status discrepancy theory, then you are paying attention.) The state of affairs just described is largely due to the law of uneven growth, which virtually assures that the distribution of power in the system will be unstable. Uneven rates in the growth of national power result in a cycle of growth and decline for all states and the rise and fall of hegemonic powers. Gilpin’s conception of the law of uneven growth is somewhat similar to Lenin’s use of the law of uneven development, which he asserted would lead to war between capitalist states. Gilpin, however, is a neorealist; he believes that the clash between major powers is not primarily economic but a more fundamental clash of strategic and national interests: it is a power struggle, not an economic struggle. What is important here is the uneven growth of a state’s total power in a broader sense, not the uneven development of national economies. For Gilpin, this uneven growth of power is created not only by changes in transportation, communication, industrial technology, population, prices, and the accumulation of capital but also by changes in military technology and strategy. Gilpin places equal emphasis on the decline of the dominant power and the rise of the challenger. The decline in the relative position of the hegemon is virtually inevitable and is due to several factors: (1) the costs of maintaining dominance in the system—which include military expenditures, aid to allies, and provision of collective economic goods necessary to maintain the global economy; (2) the loss of economic and technological leadership to other states owing to uneven rates of growth, decreasing innovation and risk taking in the hegemon, and the tendency for the hegemonic state to consume more and invest less; (3) the tendency for military and economic technology to diffuse to other states; (4) the erosion of the hegemon’s resource base; and (5) the tendency for power to shift from the center to the periphery as fighting among states in the central system weakens them all. Incorporating a rational choice model in his theory, Gilpin argues that states will attempt to change the international system “in response to developments that increase its relative power or decrease the costs of modifying political arrangements.” 85 The propensity to initiate change in the international system continues until an equilibrium is reached between costs and benefits. The international system is stable as long as no state perceives that it is profitable to alter it. Essentially, this means stability is ensured as long as the strongest states in the system are satisfied with the prevailing distribution of rights and benefits. A change in relative power alters the cost of changing the international system, granting some states a powerful incentive to seek revision. As their relative power increases, rising states attempt to change the rules of the system, the division of spheres of influence, and the distribution of benefits and territories, but only when the expected benefits of altering the system are perceived to exceed the expected costs. Due to persistent disequilibrium “the international system is beset by tensions, uncertainties, and crises.” 86 The stage is now set for hegemonic war, though Gilpin argues that

The International System Level of Analysis, Part II

429

several preconditions are associated with the outbreak of the war: (1) intensification of conflicts; (2) perception that a fundamental historic change is approaching and the growing fear by at least one state that time is working against it, creating preventive motives for war; and (3) the course of events begins to escape human control. Moreover, the war itself is frequently preceded by a systemic polarization as the great powers join mutually exclusive alliances systems. 87 Gilpin argues that war has historically been the primary method of resolving the disequilibrium between the structure of the international system and changing distribution of power. Hegemonic war itself then transforms the system. The result is that a new equilibrium will reflect a new distribution of power, and another cycle of growth-expansion-decline will begin. Gilpin sees two examples of hegemonic war in the premodern world—the Peloponnesian War and the Second Punic War. In the modern world he identifies four: the Thirty Years’ War (1618–48), the wars of Louis XIV (1667–1713), the wars of the French Revolution and Napoleon (1792–1814), and the combined world wars. 88 Not all hegemonic wars produce dominant states who provide systemic leadership. Only the hegemonic wars of the nineteenth and twentieth centuries have produced a global hegemonic power—Britain in the nineteenth century and the United States in the twentieth. Who initiates this war? Gilpin is relatively agnostic about this and even suggests that it may ultimately be impossible to determine. He suggests that the rising challenger is expected to be the most likely culprit, as it attempts to expand its influence to the limits of its new capabilities. However, he also recognizes the possibility that the hegemon itself may attempt to weaken or destroy the challenger by initiating a preventive war to forestall its loss of position. Gilpin argues that there aren’t too many examples historically of hegemonic powers willing to concede their dominance over the system to a rising challenger in order to avoid war. Neither are there many examples of rising states that prefer not to press their advantage. 89 Neither bipolarity nor multipolarity guarantees peace, according to Gilpin. The most important factor is not the distribution of power, but the dynamics of power relations over time. In both bipolar and multipolar systems, changes in the relative power among the principal actors lead to war and change. 90 (Without have said it in so many words, Gilpin has explained hegemonic wars in terms of power transitions.) 91 Central to Gilpin’s theory of war and change is the idea that there is an inverse relationship between the power of the hegemon and the likelihood of war. Unipolar systems are seen as the most stable, while instability accompanies the decline of the hegemon’s military preponderance. A weakening of the hierarchy of prestige and increased ambiguity places the system on the road to war. 92 For this reason his theory is often referred to as hegemonic stability theory. Unlike many of the theorists we will discuss shortly, Gilpin does not see the rise and fall of global hegemony as being associated with economic cycles or phases of the global capitalist system. The strong implication of Gilpin’s theory is that the near seven decades of peace following the end of World War II have been due to the continued hegemony of the United States but that a decline of American preponderance could usher in a new era of global warfare. 93 But Gilpin also argues that it is possible that systemic change can be accomplished without hegemonic warfare. That this has not happened historically does not mean we are doomed to the pattern. Alternative methods for systemic transformation can be found.

430

Chapter 11

Hegemonic War: Empirical Research Gilpin’s work only illustrates his theory with historical examples; he does not subject it to an empirical test. He is able neither to demonstrate that all hegemonic wars have resulted from the type of systemic disequilibrium he discusses, nor that all such systemic disequilibria have led to hegemonic wars. One full-fledged test has been made of Gilpin’s hypothesized relationship between hegemonic decline and war by Edward Spiezio. 94 Since the creation of the modern international system in 1648 with the Peace of Westphalia, hegemonic governance has been accomplished only twice according to Gilpin, by the British (1815–1939) and the United States (1939 to the present). 95 Because the American cycle is incomplete, the period of British hegemonic rule is taken as a test case. Spiezio attempts to discern whether war was in fact inversely related to Britain’s relative military and economic position, that is, whether war was more frequent during the period of hegemonic decline than the period of Britain’s ascendancy. He concludes that Gilpin’s theory is generally supported, but the relationship is not particularly strong, and there are some serious anomalies. As hypothesized, the frequency of international conflict is inversely related to Britain’s relative power during her entire cycle of leadership. However, although wars occurred more frequently during Britain’s decline than during her ascendancy, the difference was not overwhelming (54 percent to 45 percent). Wars occurred frequently in both phases. Interestingly, ten of the twelve wars involving a great power on one or both sides that occurred during Britain’s rising phase were actually clustered around the years 1845–60, the high point of Britain’s relative power in the system! On the other hand, the majority (79 percent) of wars involving great powers on both sides—the type of wars most relevant to the theory—did begin during Britain’s declining phase. However, Spiezio concludes that ultimately, the degree of hegemonic power cannot be seen as a primary determinant of the occurrence of war. Boswell and Sweat’s time series analysis of conflict from 1496 to 1967 finds stronger support for Gilpin’s assertion that periods of hegemony help deter wars of high intensity, especially after 1790. 96 Brian Pollins has also tested the fit between Gilpin’s periods of hegemony and the number of militarized international disputes in the international system from 1816–1976. He found that the periods of both British and American hegemony were associated with significantly fewer MIDs. 97 Bennett and Stam perform a relatively simple test of Gilpin’s theory. Using data on MIDs and wars from 1816 to 1992, they compare those periods identified as hegemonic by Gilpin with nonhegemonic periods. They find little support for the thesis that interstate conflict is reduced during periods in which there is a clear hegemonic leader. 98 A more sophisticated test produces results that are even more challenging for a theory of hegemonic stability. Using a measure of the degree to which economic and military capabilities in the international system are concentrated in the hands of a single state (versus being dispersed among several great powers), they discover that systemic power concentration turns out to be one of the most potent predictors of militarized disputes and war. They estimate that war is between three and six times more likely when the concentration of military and economic capabilities is one standard deviation above the average for the system as compared to one standard deviation below the average. Stable periods of high concentration appear to be the most dangerous— exactly the opposite of what Gilpin predicts. 99 (This is also contrary to the predictions of long

The International System Level of Analysis, Part II

431

cycle theory—the next theory we will look at.) In spite of some evidence in support of the theory, we have to conclude—as Robert Jervis has—that hegemonic stability theory has been rather “battered by encounters with evidence.” 100 MODELSKI’S LEADERSHIP LONG CYCLE THEORY A long cycle theory of international politics and war has been put forward by George Modelski and William Thompson. 101 According to this view, there are three principal structures in the world system: the global political system, the world economy, and the world cultural subsystem. Their focus is on the global political system, which they see as only partially anarchic: Global order is intermittent—there is more of it during some periods than in others. 102 Management of the system is sometimes absent altogether or is often shared among several states, but from time to time the management of this interdependent global system is in the hands of a single unit. This state, which Modelski calls a world power, dominates the keeping of order in the system through its monopoly of military resources. The position of the world power prior to 1945, according to Modelski and Thompson, was based primarily on naval capabilities—the factor that contributes most to a state’s command of global reach; 103 in the post–WWII era, airpower augments sea power as a facilitator of global reach. The military preponderance of the world power makes it possible for this state to provide public goods such as military security, world organization, and a set of rules for international economic relations. The world power is the system’s leader politically, militarily, economically, financially, and culturally. World powers have a temporary monopoly over either critical global markets (think West African gold and Asian spice for Portugal and the Baltic and Asian trade routes for the Netherlands) or the production of globally significant commodities (think wool or cotton textiles, iron, steam engines and railroads for Britain and steel, chemicals, petroleum, automobiles, electronics, and aerospace industry for the United States). Fundamentally, though, its position is based on its edge in the type of innovation that is able to radically transform global commercial and industrial activities and its consequent domination of lead industries. (And naval supremacy is typically based on technological innovation.) Historically, of course, each innovator is succeeded by another state that outperforms the initial innovator in its particular specialty and/or becomes the primary innovator of a new lead industry or complex of industries. 104 Since 1494 there has been a succession of world powers that have shaped the global system. 105 The rise and fall of these world leaders has been cyclical in nature. At the macro level, the power capabilities in the system go through a process of concentration, deconcentration, and reconcentration. Each cycle begins with a global war, which determines how the system is to be constituted and which world power will be able to organize the system. The war leaves military capabilities (particularly sea power) highly concentrated, at least temporarily, in the hands of a single state. This state also possesses, again temporarily, the world system’s leading economy. Indeed, victory in the global war is essential to a second burst of innovation and long-term growth for the leading state. (Its first innovative spurt accelerates the relative decline of the old global leader and immediately precedes the onset of global war.) Eventually, the power and the political legitimacy of this state decline and it attracts competi-

432

Chapter 11

tors. Order gives way to disorder, concentration of power to deconcentration. Long cycle theorists identify four stages of the leadership cycle, each associated with different levels of conflict: (1) global war, associated with the highest level of warfare; (2) world power, the lowest level of conflict; and (3) delegitimization; and (4) deconcentration, with intermediate levels of conflict. It is the world power stage that is associated with peace and stability. Like Gilpin and Organski, the long cyclists believe that concentration of power in the hands of a world leader is associated with systemic stability, though this does not last. The long cycle of world leadership endures approximately one hundred years. The world power manages the system alone for the initial part of this cycle, but the system does not remain unipolar; as the legitimacy and power of the global leader erode, the system drifts from unipolarity to bipolarity and multipolarity. Long cycle theorists claim the international system has completed four full cycles and part of a fifth, as illustrated by table 11.1. To what does Modelski attribute the cyclical nature of world leadership? First, each cycle is born in warfare—a situation that is not the most conducive to the creation of a stable world order. Second, monopolization of global power is a double-edged sword: Monopolies create benefits, and benefits attract competition. Monopolization is also expensive; the burden of managing the system requires the expenditure of tremendous resources. Debt levels and the inability to secure uninterrupted access to credit are particularly important in the decline of world powers, especially the early global leaders. 106 Third, there is a tendency for the world power to respond to challenges by defending fixed positions and distant frontiers. Fourth, there are the interrelated problems of complacency, institutional rigidities, and increased costs of living that retard future growth. Fifth, the long cycle is related to shifts in the distribution of economic resources among the states in the system, though this has more relevance for the post-1815 world system than for the pre–nineteenth century. Sixth, and perhaps most important, there is a tendency for the leader to lose its advantage in technological innovation. Finally, the coalition constructed in order to win the previous global war has a tendency toward fracturing. Thompson notes that in each of the long cycles one member of the winning coalition changed sides after the war and became the primary challenger in the next succession struggle. 107 The long cycle theory is concerned primarily with global wars—wars that result in the selection of a new world power (or the confirmation of the previous world power). In that sense, global wars are succession struggles. Each global war has been relatively long-lasting

The International System Level of Analysis, Part II

433

and has taken place at intervals of approximately one hundred years. And each, according to Modelski, has had a distinctive oceanic-maritime component. 108 Although the primary focus of the theory is on global wars, the gradual deconcentration of power in the system is also related to smaller wars that serve as a prelude to the more cataclysmic global wars. 109 Long cyclists maintain that the cause of war is to be found primarily in the processes and dynamics of the system, specifically the changing distribution of power that derives from uneven rates of development among members of the international system. Research by long cycle theorists generally supports the power transition contention by Organski that, immediately prior to global war, the challenger’s capabilities are on the rise while those of the world power’s are declining, though Modelski and Thompson argue that this does not necessarily mean the dissatisfied challenger will attack the leading power. The conflict is usually initiated not by a direct confrontation between challenger and hegemon, but by the challenger’s attempt at expansion on the European continent in what is initially a localized conflict, an attempt that frequently occurs before the challenger has actually surpassed the hegemon in power capabilities. With one exception (prior to the wars of the French Revolution), the global situation has been triangular: the rising continental power is confronted by two maritime powers, one of which is the world power. When the dust settles at the end of the global war, the challenger is defeated, and the other maritime power emerges as the successor to its ally as the new world power. 110 Thompson explains: Invariably, the challenger appears to act on the hope, belief, or mistaken assumption that one or more of the globally oriented powers will not oppose its continental expansion. These misperceptions may be mixed with impatience and overconfidence stemming from the challenger’s rapid capability improvements and encouraged by a system characterized by declining order and increasing strife. But in any event, global wars tend to begin as relatively localized affairs, becoming global in scope only after the globally oriented power(s) decides to participate. 111

The fact that global wars tend to start as local affairs is theoretically puzzling. It suggests that challengers may not have actually intended an all-out challenge to the status quo, raising the question of whether global wars for systemic leadership are (or need be) deliberate acts. 112 Whether global wars are deliberate or not, their results are similar. In fact, long cyclists uncover a pattern that virtually all cyclical theorists affirm: The challenger invariably loses the military contest with the hegemon. The postwar transfer of global leadership is often accomplished in a fairly cooperative manner, with the new leader frequently arising from the coalition of allies surrounding the old hegemon. 113 The events of the first half of the present century are illustrative. German challenges were mounted—prior to World War I and World War II—before it overtook either Britain or the United States in economic terms, but especially important was its lack of naval power relative to either the declining hegemon (Britain) or the eventual new world power (the United States), thus limiting its global reach capabilities and leaving it vulnerable to naval blockades. 114 After World War II hegemony was smoothly transferred from the reigning world leader, Britain, to its coalition partner, the United States. Historically, each successive global leader has been stronger than its predecessors, though its opponents have been stronger as well. 115

434

Chapter 11

Long Cycles: Empirical Research Modelski’s long cycle theory hypothesizes that unipolar systems will be the most stable and peaceful, followed by bipolarity, while multipolar systems are the most unstable. Thus, the level of stability (and order) in the system is directly related to the power structure of the international system. Thompson tests this set of propositions using a special indicator of polarity based on the distribution of sea power. Thompson discovers that, with two exceptions, a unipolar or near-unipolar distribution of power does in fact emerge from periods of global warfare. 116 These periods are characterized by less warfare than would be expected by chance, and destabilizing warfare is least common in unipolar systems. As the distribution of power in the system becomes increasingly less concentrated, the amount of war increases, with multipolarity being the least stable system. 117 The 1816–1945 period deviates somewhat from this pattern in that bipolarity in this period resulted in a high proportion of weighted warfare. The British-German bipolar pair was especially war prone, as were Britain-Netherlands and Britain-France in the seventeenth and eighteenth centuries, respectively. On the other hand, the British-French rivalry and the British-American rivalry in the late nineteenth and early twentieth centuries were not especially dangerous. Thus, the relative stability of bipolarity seems to be somewhat dependent on nonsystemic factors. 118 The important point remains, however, that unipolarity—at least as defined by long cyclists—is the most peaceful. Boswell and Sweat’s time series analysis of data from 1496 to 1967 largely agree with this finding. Using Modelski and Thompson’s measure of sea power concentration and their own measure of war intensity, they confirm that high concentrations of global power inhibit severe warfare, especially for the period of the industrial revolution (after 1790). 119 This is exactly the opposite of the study by Bennett and Stam (using a different measure of power concentration) reported earlier, which finds a link between power concentration in the system and increased military conflict. Thompson also investigates the possible connection between Organski’s power transition and the hegemonic war cycle, examining the five long cycle wars for evidence of power transitions. Using his own data on national power (based on naval capabilities), he finds each global war was preceded by a decline in the capability of the dominant state relative to the primary contender at the dyadic level, in part validating Organski’s theory for the five global wars. He concludes that there may be some linkage between systemic and dyadic processes; if a world system cycle and a power transition occur simultaneously, even small local crises could trigger a global war. 120 We know from the theories of Organski, Gilpin, and Modelski that the declining power of states, especially of hegemons, is a dangerous situation; however, declining power doesn’t always lead to war. It would be helpful to know when it does and when it doesn’t. Theorists have put forward various conditions. 121 It has been suggested that declining power leads to war when (1) the transition is rapid and moves toward rough equality, (2) the magnitude of the shift is high, (3) the challenger and hegemon are rivals who lack a history of friendship, (4) the challenger is dissatisfied with the status quo, (5) offensive forces have a relative advantage over defensive forces and there is an expected probability of the adversary initiating war first, (6) there is a probability of victory within tolerable costs, and (7) leaders in the initiating state

The International System Level of Analysis, Part II

435

are risk-acceptant. Whether all these conditions need to be present or just some, and in which combinations, is not clear. Finally, other theorists have attempted to extend the explanatory range of long cycle theory beyond the club of major powers and their wars. (You knew this was coming.) Brian Pollins believes that the historical-cyclical theories can be applied to all states in the system and to conflicts below the level of war. He finds long cycle theory helpful in explaining patterns of militarized disputes in the post-1816 period, but finds two distinct phases of conflict rather than the four identified by Modelski and Thompson. In short, he finds the combined world power/delegitimization phases to have a relatively low amount of militarized conflict and the combined deconcentration/global war phases to have a high amount of militarized conflict. 122 Long Cycles: A Critique Critics, of course, have challenged some of the analyses of long cycle theorists. For instance, Jack Levy argues that the classification of global wars is dubious. Long cycle theorists omit several major power wars with important consequences for the European system that other scholars classify as major wars. The second phase of Charles V’s Italian wars with France (1672–78) began early in the first long cycle’s world-power phase; the Thirty Years’ War (1618–48) broke out during the world-power phase of the second cycle; the war of Jenkin’s Ear/Austrian succession (1739–48) began barely a year after the end of the world-power phase of the third long cycle. That such important major power wars occurred during the very periods deemed to be most peaceful by the long cycle theory ought to undermine our confidence in it more than just a little. 123 A second concern is the emphasis on naval power at the expense of land-based military power and the de-emphasis of the importance of a dominant military position of the European continent. Levy summarizes: The primary cause of the great wars of the past . . . has been the perception by most of the great powers that one state was threatening to gain a dominant position in Europe. The great powers have always perceived the most serious threats . . . as coming from the great land powers of Europe—which could threaten their territorial integrity—rather than from the more wealthy naval and commercial powers. This is why the great European military coalitions have always been formed against the most threatening continental power rather than against the leading naval power. 124

This criticism should be coupled with our earlier observation that global wars do not break out as direct challenges in which large maritime powers confront each other. Instead, they tend to break out when one land power in Europe attacks another and the global leader then intervenes—making global war look like an accident rather than a concerted attempt by the rising challenger to take on and bring down the prevailing system leader. Long Cycles Reformulated: Fusion of Global and Regional Cycles In an effort to address some of the criticisms of long cycle theory Karen Rasler and William Thompson have reformulated the original theory somewhat to make it more congruent with

436

Chapter 11

the historical realities of the last several centuries. 125 They believe that historically, there have been two types of great powers: (1) global powers who are essentially maritime powers oriented toward global trade and (2) regional powers who are primarily land powers oriented toward territorial expansion. For both global powers and regional powers there are historical cycles of concentration and deconcentration of capabilities, and these cycles are fused together. However, the patterns are not parallel or synchronous but dissynchronous: When one declines the other peaks. At those times when global capabilities are concentrated in the hands of a single hegemonic sea power, regional capabilities tend to be dispersed among several states; and when global capabilities are dispersed, regional capabilities tend to be concentrated in the hands of a single land power. From 1494 until the end of World War II the most important global region has been Europe—the home of several global powers and the site of global wars to determine systemic leadership. Rasler and Thompson note that no single state has been able to establish hegemony over Europe for long. This has been due to the presence of military counterweights from adjacent areas: primarily Russia in the east and Britain and the United States in the west. As Thompson says, “The east supplies the armies, the west supplies the sea power.” 126 The state with the largest share of army capabilities in Europe—referred to as the leading power—has usually also controlled the largest population and the largest economy in Europe. These European leading powers have been Spain (1494–1644), France (1645–1870), and Germany (1871–1945). The concentration of power at the regional level is almost entirely a function of territorial expansion coupled with the existence of an ambitious ruler. 127 The most important threat to the international system—and the one that leads to global wars—occurs when regional capabilities (in Europe) become concentrated in the hands of a land power that aspires to global hegemony and that also endeavors to acquire naval power, posing a threat to the global leader. The fluctuation in the relative strengths of European regional powers—like Spain, France, and Germany—create the possibility for new regional leadership. And the concentration of capabilities at the regional level tends to happen—given the nature of the two cycles—at a time when the relative capabilities of the global hegemon are in decline. This creates a window of opportunity for land-based powers on the rise. (Global deconcentration eases the way for rising regional contenders.) It is possible then for a European land power to dominate the continent and to simultaneously challenge the old/declining world power. Any state that becomes the leading power in Europe is poised to take on global supremacy, and therefore the world power organizes a coalition to counter the looming threat. Another way of saying all of this is that power transitions are essential to the dynamic. Power transitions between regional land powers or between global sea powers may lead to war, but the most dangerous and destabilizing transitions occur when a rising regional land power with global (and maritime) aspirations overtakes a declining maritime power (the old system leader). Global wars are not the result of a direct military attack on a declining global leader by a challenger. Instead, the challenger attacks lesser continental powers, generally hoping to avoid intervention by the global leader and its allies. The possibility of a regional hegemony in Europe however threatens the global structure of power, increasing substantially the probability of intervention. What begins as a local war in Europe, therefore, escalates to a global

The International System Level of Analysis, Part II

437

war. 128 Essentially, the global leader acts as an “external balancer”—a state outside the regional European system that intervenes when the balance between European land powers moves away from equality, threatening the stability of the regional system. 129 Global war thus results from a global deconcentration of power that coincides with a regional concentration of power. Historically, the result has been that the declining global power is able to organize a coalition strong enough to defeat the regional challenger, and a new global power emerges from the winning coalition. Additionally, the regional challenger has tended to expand territorially to both the west and the east, generating opposition on both flanks, virtually assuring its own defeat. 130 With the defeat of the regional challenger, capabilities in the global system once again become concentrated in the hands of the (new) world power and become dispersed in the European regional system. Rasler and Thompson are able to chart the fluctuating concentration of capabilities for both global powers and regional powers, using measures of sea power and land armies, respectively. 131 The paths are largely as hypothesized. The two cycles seem largely dissynchronous: The concentration of sea power alternates with the concentration of regional land power. Moreover, global wars appear to take place as predicted—during periods of global deconcentration of power and regional concentration. Rasler and Thompson argue that the combination of these two structural conditions significantly increases the probability of global war during the last five centuries. World War I is somewhat of a deviant case, in the sense that there is not much concentration of regional power before the war’s onset. They suggest that in some cases the “perception of regional hegemony may work as an effective substitute for objective regional capability concentration.” 132 In the aftermath of global war, power is once again reconcentrated in the hands of a single global power. (Nonglobal wars are followed by a continued deconcentration of power.) Thus the Napoleonic wars are presaged by the concentration of power in Europe by France, while that of Britain (the global power) is in decline. Likewise, the global wars of the twentieth century occur because of the twin spikes during which regional capabilities in Europe become more concentrated in the hands of Germany. Their investigation of power transitions indicates that transitions are indeed an important ingredient in the structural transformations leading to global war. 133 Most importantly, the hypothesis that war is more likely when there is a power transition between a rising regional leader and a declining global leader receives considerable support after the sixteenth century. Wars, especially global wars, are more likely when a regional leader catches up to and passes the relative capability position of a declining global leader. If the Spanish challenge to Portugal in the sixteenth century is excluded, then four out of the six five-year periods involving overtaking transitions (66.7 percent) led to war, while only four out of forty-six periods in which the dyads were either equal or unequal with no overtaking (8.7 percent) led to war. (Once again, WWI does not entirely fit the model.) Additionally, transitions between old regional leaders and rising regional challengers are prone to war. 134 The authors also find some evidence that power transitions between global leaders (the declining global leader and a maritime challenger) are connected to wars between major powers, but this hypothesis has the least predictive success. Moreover, the most recent historical cases provide interesting counter examples. With regard to the transition between Netherlands and Britain, the two fought rather inconclusive wars, but at the end British leadership

438

Chapter 11

was not militarily resisted by the exhausted Dutch. Britain and its successor, the United States, fought each other (1812–14), but well before the actual transition. In both cases, both parties—the challenger and the declining global leader—determined that a greater threat to their security existed, and instead of the titanic clash of maritime titans one might expect, they formed a coalition against the greater menace. 135 On the other hand declining and aspiring global leaders did fight each other in the extended Dutch-Portuguese wars of the seventeenth century and the first and second Anglo-Dutch Wars (1652–54 and 1665–67). One note of comparison should be added here. Rasler and Thompson’s analysis of European regional powers excludes both Ottoman Turkey and Russia, treating them instead as outside counterweights that occasionally intervene in the European subsystem. The relative rise and decline of Russian and Turkish military power play no role in Rasler and Thompson’s calculations. This fits with the theoretical framework they have developed, and it makes some sense in terms of data analysis. For example, figures for the size of Russia’s army tend to overwhelm the smaller armies of western Europe; poorly armed and organized, they never had anywhere near the impact that their raw numbers might suggest. On the other hand, remember that Copeland sees Russia as the primary threat to Germany’s military dominance in the first half of the twentieth century—a threat based on her large population and large army. For Copeland, Russia is not an external balancer, but a central player in the balance of power in Europe. It is the looming threat that Russian will overtake German military power that is the motivation for major war. A further theoretical refinement flows from the distinction between land powers and sea powers. Thompson argues that the creation of balancing coalitions against the global maritime system leader is not particularly common. Balancing is more likely to take place against the leading regional land power. In part this is due to the differing strategic orientation of the two. Global maritime powers do not seek to expand territorially, at least not at the expense of neighboring land powers, and none has sought to control European affairs. Instead, their maritime-commercial-industrial outlook means that they are more concerned with access to overseas markets. As Thompson says, “their economic success may well be resented, but they are not seen as representing overt threats to the sovereignty of other major powers.” 136 Moreover, the economic and commercial nature of the leading maritime state means other great powers may reap advantages from allying with it rather than against it. On the other hand, expansionary land powers are seen as threats to the political sovereignty and territorial integrity of neighboring states and are thus much more likely to be the targets of balancing coalitions—helping to explain the lack of success by rising regional challengers. All of this is, of course, very much in line with Stephen Walt’s “balance of threats” argument that states tend to balance against threats rather than against power per se. Levy and Thompson would add an amendment to Walt’s idea that threat depends on a state’s aggregate power, its geographic proximity, its offensive power, and its perceived aggressive intentions by arguing that land and sea powers affect these categories differently. They would argue that “a proximate threat on land is far more threatening than a proximate threat at sea,” and “the extent of an adversary’s ‘offensive power’ is significantly influenced by whether the adversary is a land power or a sea power. Sea powers are significantly less capable than land powers of invading or occupying land powers.” 137

The International System Level of Analysis, Part II

439

Levy and Thompson compare balancing behavior by great powers against leading European land powers and against leading global maritime powers from 1495 to the present. 138 In their initial inquiry they find—as balance-of-power theory proposes—that major states formed balancing coalitions against the leading European land powers if these leaders were disproportionately strong (they controlled a third or more of the total military capabilities in the system). However, the tendency to balance was much weaker against leading states that controlled fewer resources. The greater the concentration of power in the hands of the leading land power, the greater the size of the balancing coalition tended to be. Over half of the coalitions aimed at balancing against the leading land power (with a minimum of 33 percent of the system’s capability) were broad coalitions of three or more states. When Levy and Thompson turn their attention to sea powers, they discover an entirely different pattern. They discover that great powers form alliances against the leading sea power only 16 percent of the time. Moreover, the stronger the leading sea power’s margin of superiority, the lower the probability that a balancing coalition will be constructed. Not only that, the larger the margin of superiority, the fewer the number of great powers that will generally join a coalition against a global leader. No broad coalitions (with three or more members) formed against leading sea powers that possessed 50 percent or more of the system’s naval capabilities. Coalitions may form against the leading sea powers, but they’re generally small. Consistent with their expectations, they find that great powers are actually more likely to ally with the dominant sea power than ally against it. Not only that, but the tendency to ally with the maritime leader increases as the maritime leader’s margin of superiority increases. With regard to maritime leaders with a 50 percent margin of superiority, “great powers are more than four times likelier to ally with the lead sea power than to balance against it.” 139 Levy and Thompson report that historically, great powers in the European system have bandwagoned by allying with the dominant land power only 13 percent of the time, but have balanced against it with alliances 55 percent of the time. By contrast, in the global maritime system, great powers have bandwagoned with the leading sea power 34 percent of the time, but have only balanced against it an astonishingly small 8 percent of the time. They conclude: “great powers ally against the dominant land great powers more than four times more frequently than they ally with them, while they ally with dominant sea powers more than four times more frequently than they ally against them.” 140 The traditional balance of power theory, originally constructed to explain balancing among land power on the European continent in order to prevent hegemony, needs to come to grips with the fact that states have balanced quite differently against global sea powers. Long Cycles: Implications As for the present situation, the long peace since 1945 can be explained by the structural condition of the international system: for most of the last forty-five years (1946–68 according to Thompson), it has been in the world power phase with clear American hegemony. In 1969 it entered a bipolar delegitimation phase, but the United States was still able to provide political, military, and economic order to the system. Indeed, Thompson demonstrates that the United States has been in a stronger relative position than other world powers at a comparable point in the long cycle, along both naval concentration indicators and leading economic sector indica-

440

Chapter 11

tors. 141 The American position has been buttressed by the anomalous situation in which the leading military contender and the leading economic contender in the 1970s and 1980s were different states—the USSR and Japan, respectively. The Japanese economic miracle faded. At the same time, the collapse of the Soviet Union and its fragmentation into fifteen independent states in 1991 has left the U.S.—at least temporarily—without a real military competitor. Moreover, the status of Europe as the primary regional location for great power competition is no longer valid, with no immediately apparent successor—though Asia might assume this role in the future. The current focus appears to be on the rise of China as a potential challenger. With China now ranked second in the world behind the U.S. in total economic power, and with a growing military, it is becoming a dominant regional power, yet it currently lacks the military global reach of the United States. Thus, the structural conditions of the international system continue thus far to discourage global war. However, if long cycle projections remain on course the probability of global war might increase substantially by the fourth decade of the twenty-first century. 142 Long cycle theorists argue that there is nothing inherent in the logic of long cycles that requires them to begin with global wars. Disequilibria in the international system could be resolved through peaceful change. Alternative mechanisms for the transfer of global leadership may be found. The cause of war in the contemporary world has simply been the lack of a substitute mechanism of the making of global decisions about political leadership. 143 WALLERSTEIN’S WORLD SYSTEMS APPROACH Immanuel Wallerstein is the leader of a rather diverse school of thought variously called the world systems or world economy approach. 144 The world systems perspective is essentially a political economy approach to international relations that focuses on international inequality and dependence. Even though its primary focus is not on war, its influence on the study of war is widespread. It shares much in common with the approaches of Gilpin and Modelski. As with the other historical-structural approaches, the world systems perspective comes complete with its own terminology. Wallerstein classifies world systems into two types: world empires, in which a single political unit controls the world system’s economy (e.g., the Roman empire), and world economies, which are essentially multicentric, with no single state in control. Wallerstein argues that although limited world empires existed in the past, the modern age (beginning about 1450) has been characterized by the emergence of a European-based capitalist world economy that has gradually become a truly global economy. Although attempts have been made by certain states to construct a world empire, in the modern age none has succeeded. Like the realists, Wallerstein emphasizes the anarchic nature of the international system. Essentially, the competitive nature of the system prevents monopolization, while the balance of power in the interstate system prevents any one state from controlling the world economy. 145 Political anarchy leads to a particular form of global economic system—a capitalist world economy with an international division of labor. This world economy is divided by world systems analysts into three segments: the core, the periphery, and the semiperiphery. Core states are the “haves”; they have the most efficient and

The International System Level of Analysis, Part II

441

productive economies and are the most technologically advanced. Production in core states is capital-intensive and uses skilled, high-wage labor. Core states also have the strongest military establishments. Needless to say, they also receive a disproportionate share of the world economy’s rewards. The periphery is made up of economically weak states whose production is primarily in low-wage, labor-intensive goods. Their economies are highly dependent on those of the core state to which they are most closely associated. The semiperiphery constitutes an intermediate category of states, with some production that is similar to that of core areas and some similar to that found in the periphery. Thus, the semiperiphery acts in some respects as an exploiter (of periphery states) and as an exploited area (by the core states). There is constant conflict over membership in these groups; all states hope to be upwardly mobile. Mobility within the core area is also conflictual. The core itself is divided between hegemonic powers and regular core states. A hegemonic power is a core state that develops a position of dominance throughout the entire world economy. Wallerstein sees this dominance primarily in terms of comparative advantage: the concentration of certain kinds of enterprises (called lead industries) within the core state. The hegemonic state holds a decisive superiority in agricultural-industrial productivity, in finance and investment, and it dominates world trade, amassing the largest single share of the world market and therefore the largest economic rewards. As a result, the hegemonic state is able to impose a set of rules on the system. True hegemonic status has been achieved by only three states and for only brief periods, according to Wallerstein: the United Provinces, 1620–72; Great Britain, 1815–73; and the United States, 1945–67. World wars have played a key role in consolidating the hegemonic status of each. 146 The United Provinces’ position was developed by the Thirty Years’ War; British hegemony resulted from its victory over France in the Napoleonic wars; and American hegemony resulted from World Wars I and II. Each attempt to impose world empire—by Louis XIV, by Napoleon, and by the Germans in the twentieth century—came during periods of relative hegemonic (that is, Dutch or British) weakness. Like Modelski, Wallerstein argues that the hegemonic states have been primarily maritime powers, though they become land powers as well to counter land-based challengers. World systems theorists place war in the framework of the development and expansion of the capitalist world economy. At bottom, “the capitalist system is a system that has pitted all accumulators of capital against one another.” 147 Wallerstein argues that wars can be seen as “struggles to shape the institutional structures of the capitalist world-economy so as to construct the kind of world market whose operation would automatically favour particular economic actors.” 148 Although world wars accommodate expanded levels of economic development, a point is reached when the political framework under the domination of the old hegemon is inadequate to facilitate world commodity production and distribution on a larger scale. Thus, says Christopher Chase-Dunn, “world wars and the rise and fall of hegemonic core powers . . . can be understood as the violent reorganization of production relations on a world scale,” so as to increase the internationalization of capitalist production. 149 World wars are essentially attempts to restructure the interstate political structure to reflect changing economic realities and

442

Chapter 11

to convert the political-military strength of rising challengers into a greater share of the world surplus. 150 World wars result in the crowning of a new hegemonic power; however, the victor’s hegemony does not last. According to world systems theorists, its demise has much more to do with economic factors than military ones. Maintaining hegemony is expensive in terms of military and bureaucratic overhead, and tax burdens rise within the hegemon. Uneven capitalist development leads to a change in the distribution of productive capabilities among the core states, and the hegemonic power loses its competitive edge in the production of leading industries; agricultural and industrial production declines. Wages in the hegemonic state tend to rise, reducing competitiveness, and profit rate differentials change, leading to the export of capital from the hegemon. The free market results in the flow of capital and technology to other states, and innovations that have given the hegemonic power a competitive edge can always be copied by others. Hegemonic core states simply can’t control this process. It is important to understand that for Wallerstein these changing distributions of power are rooted in the underlying economic order and in uneven rates of capitalist development. Shifts in the balance of political-military power may be important, but they are brought about by economic processes. Almost as soon as power is concentrated in the hands of the hegemon, it begins to wane, creating a cycle of concentration and diffusion of power among hegemonic powers and other core states. World systems theorists describe a four-phase cycle: (1) ascending hegemony, in which there is acute warfare among rival states to succeed the older hegemon; (2) hegemonic victory, a period of scattered conflict in which the contender passes the declining, older hegemon; (3) hegemonic maturity, or true hegemony, a period of low conflict; and (4) declining hegemony, a period of competition between the hegemon and potential successors. (For some reason, the phases of ascending hegemony and declining hegemony do not overlap.) 151 Hegemonic decline initiates a period of increasing military and economic competition. The hegemonic power’s attempt to impose direct and exclusive control over the periphery leads to conflict between core states as well as resistance from the periphery. These conflicts are compounded by the increasingly stagnant world economy and by protectionist trade policies in the core. Over time, the hegemon’s alliance system begins to disintegrate, and two states usually emerge as contenders—the winner usually being the one that remains allied with the declining hegemon. The United Provinces, Great Britain, and the United States all achieved their hegemonic positions after other competing core powers weakened themselves in intracore warfare, according to Chase-Dunn, but their success depended more on their economic competitive advantage than on their military superiority, though both were important. 152 The creation of a new hegemonic core state ushers in a long period of economic growth in which the hegemonic core state provides economic and military order and support. This is also a period of relative peace. 153 Thus, unicentricity is associated with relative peace and multicentricity with relatively higher levels of war. World wars are not initiated by hegemonic core states but by a rising challenger within the core. World systems theorists agree with other theorists of war cycles that the challengers in these world wars have always failed. The reason seems to be twofold. The first reason is strategic. The challenger attempts to put areas under its control that are simply too large to be

The International System Level of Analysis, Part II

443

conquered and subjugated, and this error is combined with a failure to develop a strategy of competitive production. The second reason derives from the logic of the capitalist world economy itself. The challenger has a hard time lining up allies, probably because the potential allies fear that their interests will not be well served by the change. 154 In other words, the challenger finds it difficult to find other dissatisfied states; most core states are status-quo powers. Thus, for world systems theorists, the balance of power plays an important role in determining the outcome of world wars, but the objective basis of the balance of power is the nature of the capitalist world system itself. 155 If the cycle of war is inherent in the basic logic of uneven development, exploitation, competition, and conflict found in the capitalist world economy, how is the war cycle to be broken? Presumably, the chances of peace will be enhanced with the dissolution of the present system and its replacement by a socialist world system. In one version (that of Chase-Dunn), a socialist world system would essentially mean that capitalism would be replaced with a collectivist economic rationality, and systemic leadership would be provided by a democratic and federal system of world government. 156 World Systems Theory: Evidence and Criticism Matching Wallerstein’s four hegemonic phases with great power wars creates certain problems. As others note, Wallerstein’s three world wars—the Thirty Years’ War, the wars of the French Revolution and Napoleon, and World Wars I and II—each fall in different phases of the hegemonic cycle. Indeed, the Thirty Years’ War (1618–48) occurs during the hegemonic maturity phase of the United Provinces—presumably the phase that should be most peaceful. Moreover, while world systems theorists agree that the hegemonic maturity phase should be most peaceful, they disagree somewhat on the degree of conflict to be found in the other phases. To make things worse, various world systems theorists date the different phases differently. All of this makes it difficult to test propositions and to uncover the relationship—if any—between hegemonic cycles and war. 157 An attempt by Pollins to match Wallerstein’s phases with militarized disputes for all states in the system finds that, as predicted, the ascent phase was the most conflict-ridden. However, the level of conflict in the other phases did not fit his predictions. 158 Boswell and Sweat’s time series analysis, which attempts to fit Wallerstein’s phases to periods of war intensity, finds rather poor results, especially for the preindustrial period (1496–1790). 159 Finally, we should take note of a problem that bedevils all the historical-structural theorists to some degree. It is universally agreed that prior to World War I (and World War II, for that matter) Britain was the hegemonic power and Germany was the primary challenger. What, then, are we to do with the fact that by 1914 the United States had already passed both Britain and Germany in terms of industrial production, leading sector position, and gross national product? The statistics for the pre–World War II period are even more lopsided in favor for the United States. 160 If the two world wars were truly wars of hegemonic succession, one would expect that the United States would be involved in the initial fighting either as the rising (and, indeed, overtaking) challenger or as the new defending hegemon. In neither case was the German challenge directed at the United States. Modelski is able to save the argument by pointing to British naval superiority even in the 1930s (though this still begs the question of

444

Chapter 11

the importance of sea power), but for those who emphasize economic and industrial criteria as the essence of hegemony, belated American participation in the two world wars presents a problem. If the United States is neither the initiator nor the chief defender, how can these wars be about global leadership? The fact that the wars do indeed result in a change of global leadership is not enough. Results and causes should not be confused. K-Waves Both long cyclists and world systems theorists have incorporated certain economic cycles called Kondratieff waves—also known as K-waves or long waves—into their theories of the cycle of world leadership. The Russian economist Nikolai Kondratieff claimed in the 1920s to have discovered fifty-year waves (called long waves) in prices, production, and consumption in the economies of the major capitalist nations. He argued that these cycles were indicative of rhythms—upswings and downswings—within the international economic system as a whole. His research also suggested that upswings in economic long waves were related to the occurrence of major war. While Kondratieff did not devise a systematic theory to explain how economic cycles might cause wars, he tentatively speculated that wars were due to the increased economic struggle for markets and raw materials that accompanied the accelerated pace of economic activity, rising prices, and growth in production that took place in upswing phases. 161 While many economists today do not believe that K-waves even exist, long cyclists not only claim to have found them but insist that they operate parallel to world leadership cycles: one-hundred-year world leadership cycles are linked to pairs of K-waves. And since K-waves are linked to leadership cycles, they are also associated with cycles of global war. 162 Wallerstein also identifies A-phases (global growth) and B-phases (stagnation) in his cycles. A number of international relations specialists have investigated the links between Kwaves and war. Thompson and Zuk, using primarily price data for Britain and the United States, demonstrate that the majority of wars from 1780 to 1914, and most of the severe wars, were initiated in the upswing phase of a K-wave, as predicted by Kondratieff (and Wallerstein). The price upswings tended to precede major wars, and K-wave upper turning points also coincided with the termination of major wars. 163 Thompson and Zuk also argue that major wars have been largely responsible for the upswing curve of K-waves. They conclude that while it is difficult to say that K-waves cause major wars or that major wars cause K-waves, a relationship does exist. Global wars and K-waves appear to reflect a common underlying process, and instabilities in one structure will be transmitted to the other. 164 Edward Mansfield uses data from a longer period, 1495–1980, and finds that the average number of wars begun during the upswing phase of K-waves is greater than the average number of wars begun in the downswing phase. However, this only holds true when the analysis was for all wars; when the analysis is confined only to great powers, there was no significant difference. 165 Joshua Goldstein has compiled a marvelous set of economic indicators covering the years 1495–1975. 166 His effort required the stitching together of a variety of economic indicators from different geographic areas (a good number of them associated with Britain). The result is an economic data set encompassing 481 years that maps the trough and peak dates for fiftyyear-long K-waves. In fact, he finds long waves in production, investment, innovation, prices, and wages. (These waves do not occur simultaneously, but sequentially, beginning with pro-

The International System Level of Analysis, Part II

445

duction.) Goldstein is unable to find any relationship between K-waves and the frequency of great power wars; the number of wars is roughly equal in upswing and downswing periods. However, Goldstein uncovers a clear association between K-waves and the cycle of war severity (average battle deaths per year). Like Thompson and Zuk, he finds that severe wars are more likely in the upswing phase of K-waves. For the nine waves that have occurred from 1495 to 1918, he finds that each peak in warfare occurs near the end of an upswing phase. He maintains that nine of the ten war severity peaks since 1500 have occurred near the end of the upward phase of the K-wave. The exception was World War II. 167 Typically, peaks in war severity are sandwiched between production and price upswings—with production upswings preceding war by ten years or so and war preceding price upswings by one to five years. Of course, the interpretation of the results is highly dependent upon the correct assignment of years to economic upswing or downswing phases. Goldstein’s explanation for the link between K-waves and war is twofold. First, increasing production produces a greater demand for resources, which in turn leads to international competition for these resources and to greater conflict over markets. Second, this increased competition occurs during a period when production increases have made increased supplies of war materiel available to the military sector, drastically increasing the probability of war. The first argument is essentially a lateral pressure argument without the emphasis on population growth. The second argument is essentially a resource theory of war—wars are most likely when states have the resources to pursue them. 168 War is most likely to occur, then, near the end of the long-wave upswing, when resources are most abundant. 169 Note, however, that Goldstein falls back on state-level argument (the presence of economic wherewithal) to help explain a theory based on systemic-level factors (K-waves). Goldstein initially argued that the hegemonic cycle and the economic long-wave cycle, though they are not in phase with each other, operate in conjunction. Thus, hegemonic decline does not by itself lead to war; it is only dangerous when it coincides with an expansionary phase of the economic cycle. Economic expansion by itself is not dangerous either; it must be accompanied by hegemonic stagnation. For example, the economic expansion of the 1960s was not associated with major wars because of the strong hegemonic position of the United States. Goldstein (writing in 1988) predicted new economic upswings to coincide with the continuation of American hegemonic decline between 2000 and 2030. 170 In subsequent work, Goldstein has clarified his thinking somewhat. Emulating other historical-cyclical theorists he argues that each long wave has four phases: expansion, war, stagnation, and rebirth. He also argues that long waves are independent from hegemonic cycles. There is a direct connection between economic phases and war, but these are not linked to the hegemonic political cycle. Pollins, however, devises a “coevolution model” that combines long wave phases and leadership long cycle stages and finds them superior at explaining the timing of militarized disputes to the alternatives—Goldstein’s original model, the ModelskiThompson long cycle model, Wallerstein’s model, or Gilpin’s model. 171 Independent analyses provide some support for a link between K-waves and war. Bennett and Stam, using Goldstein’s coding rules to determine upswings and downswings, find that periods of system-wide economic growth are associated with increased risks that militarized disputes will escalate to higher levels of conflict, including war. Their assessment is that the

446

Chapter 11

risk that a dispute will escalate to war is 80 percent to 100 percent higher during periods of sustained economic growth. Thus K-wave upswings appear to have a rather dramatic effect on the potential for conflict. 172 Others are not so sure. Jack Levy reexamines the issue, matching Goldstein’s data on economic production cycles against the ten general wars of the last five centuries. 173 He is interested not in peaks of war severity but in war initiation. When the production cycle alone is considered (after all, Goldstein’s theory is based on the rise and fall of production, rather than on prices or other variables), Levy discovers a picture at odds with Goldstein’s theory. Four of the ten wars were begun during the middle or end of a production downswing phase, and two occurred at the beginning of an upswing—rather than near the end of the upswing, as Goldstein’s theory suggests. Many of the wars broke out near the transition from downswing to upswing, so that the casualties associated with them belonged in the upswing phase even though the wars might have begun in the downswing—explaining why Goldstein found an association between K-waves and severity of war, but not between K-waves and war initiation. Not only do long cyclists see K-waves as important aspects of the structure of the world economy, so do world systems theorists. 174 The rise and decline of hegemonic powers is believed to be synchronized with pairs of K-waves, though there appears to be no exact number of K-waves that make up a hegemonic cycle. 175 While Goldstein’s research indicated that war severity peaks during the upswing phase of K-waves, specifically between the production cycle peak and the price cycle peak, Chase-Dunn turns this argument on its head. If war severity is at its apex after the production peak, this means war severity peaks after the beginning of the downswings in both the investment and production cycles. He is thereby able to make a more directly Marxist explanation. Downswings are periods in which states have a lot of resources available for war (as Goldstein suggests); simultaneously, overproduction during this phase leads to increased competition for foreign markets and investment opportunities. Pressure to use state power to protect or expand market shares and investment opportunities during an economic downturn becomes an important factor in the slide toward war. 176 DORAN’S RELATIVE POWER CYCLE THEORY Charles F. Doran’s relative power cycle theory places a theory of decision-making about war in the context of the rise and decline of the relative power of major states. 177 Doran argues that the power capabilities of states, relative to other members of the great-power central system, follow a cyclical path of growth, maturation, and decline. Though the highs and lows for each state will be different, as well as the length of time required to attain these peaks and troughs, nevertheless each great power will go through this cycle. The general pattern is that of an “accelerating rise in relative power that ultimately slows down until relative power peaks, and the decline in relative power that likewise ultimately proceeds at a slower pace.” 178 The cycle is largely due to uneven rates of internal economic development. Since major powers tend to get involved in large wars, these power dynamics would seem to be helpful in explaining the timing of the larger wars in history, or what Doran calls extensive wars. 179

The International System Level of Analysis, Part II

447

Although this theory may seem to operate at the nation-state level, dealing with cycles of national power, it is really a systemic theory. Doran’s power cycle is a cycle of relative power. A state’s relative position at any given point in its ride through the cycle depends as much on the power of others as on its own internal growth or decline. As with most theories that are systemic in nature, relative power cycle theory maintains that the policies and behaviors of states depend largely on their position in the system—in this case on their position in the cycle of relative power. War is most likely to occur as a state reaches four critical points along the cycle. At each of these points an abrupt inversion occurs in the path of relative capabilities. A complete cycle contains two inflection points and two turning points, as indicated in figure 11.1. The lower turning point (where most states enter the great-power system) is the point at which the state’s relative position changes from a declining power to a rising power, as its capabilities, relative to others, begin to increase. The first inflection point represents the point at which the state’s relative capabilities, while still rising, begin to rise at a much slower rate, signifying that the initial rapid relative power accumulation cannot go on indefinitely. The upper turning point is the point at which a state’s capabilities, relative to those of other system members, begins to decline. The state changes from a rising power to a declining power. The second inflection point is the point at which the state’s relative decline, which has been rapid at first, begins to drop more slowly, indicating that the future may be brighter. The crucial theoretical question is why these specific circumstances should be particularly dangerous. The answer is complicated, but plausible. It incorporates theories about government planning and decision-making and national role conceptions as well as a theory of uneven development. Doran argues that national leaders make their long-range predictions by simply making a linear extrapolation (i.e., a straight-line projection) from past experience. If the past decade or so has seen a rapid rise in the state’s power relative to others in the system, planning will be based on a continuation of this trend. However, each of the four critical points in the cycle represents a change in the previous trend. Each is essentially unpredictable, and each suddenly indicates that the previous thinking was radically mistaken, just at the point where being mistaken is most threatening to the state’s position. 180 National behavior is based in part on the leader’s conceptions of the state’s role within the system; furthermore, this national role conception is based primarily on the state’s position within the system in terms of its relative power capabilities. A change in relative capabilities will mean a change in roles (for instance, from leader to follower, from aid giver to aid recipient, from a state that gives counsel to one that is ignored). These transitions in role, triggered by passing through the critical points in the power cycle, are difficult for states to make. Inertia characterizes the system; it does not adjust very quickly to changes in relative power. The inevitable result is that power and role get out of sync. On the upside of the power cycle, states with increased relative power tend to lag behind in the acquisition of new and enlarged roles; on the downside, the tendency is for role to exceed the actual power of states, as elites in declining states are not predisposed to yield roles to up-and-coming rivals. This gap between role and power in the system is an important source of conflict; if the gap is not reduced by successful accommodation, states may use force to rectify the situation. 181 You will have noticed that all of this flows quite directly from status discrepancy theory.

448

Chapter 11

Figure 11.1. The Cycle of Relative Power. Source: Adapted from C. F. Doran, “War and Power Dynamics: Economic Underpinnings,” International Studies Quarterly 27 (1983): 420.

While critical points call for major role transformations, their sudden appearance eludes advanced detection; uncertainty is therefore high and national leaders are vulnerable to overreaction, miscalculation, and the misperceptions that might lead them to choose war. 182 At critical points, leaders are more likely to act impulsively, emotionally, recklessly, antagonistically, and opportunistically. System-level phenomena (power shifts) trigger individual-level phenomena (perceptions of threat or opportunity) among the political elite in all major states. The specific reasons for war differ, however, depending on which critical point the state has arrived at. Each point has its own logic and dynamic. For instance, as the state’s relative power decreases, its role and interests should decrease accordingly, but states are usually reluctant to accept this reduction in status and influence. On the other hand, when a state’s relative power increases, its interests and roles increase as well, but the other members of the system are frequently unwilling to permit the expansion of the rising state’s activities. Thus, in both cases, disparities exist between a state’s capabilities and its role, but in different ways. 183 It is not just the reaction of the state that is undergoing the critical change that is important, but the reaction of other members of the system to the change as well. 184 Misperceptions and anxieties also afflict other system members. Under normal circumstances, all states in the system are able to plan for transitions in relative power, but these planning exercises break down for all states when a great power undergoes a critical change. Power cycle theory has implications for the behavior of states, but it also has systemwide implications. When several

The International System Level of Analysis, Part II

449

states in the system simultaneously go through critical points in the relative power cycle, this has the unfortunate effect of generating pressures for role transformations while at the same time initiating changes in expectations about the relative balance of power. This leads to massive structural uncertainty in the system and detracts from the ability of system members to manage change. This disequilibrium makes the system extremely vulnerable to extensive wars, which may lead ultimately to a transformation of the international system itself. 185 This seems to have been what happened in the pre–World War I period. According to Doran, between 1885 and 1914 every member of the central great-power system passed through at least one critical point in its power cycle. Altogether, the major powers passed through nine critical points, creating a period of abrupt and wrenching change in the international system. 186 Systemic stability prevails when the movement of system members though the power cycle is routine and anticipated. At these times the gap between interests and capabilities is small for major actors, and the system is in equilibrium. According to Doran, bipolar and multipolar systems are equally stable (or unstable). The decisive factor is not the difference in structure between the two systems, but the transition between multipolar and bipolar systems that occurs when a number of the major actors in the system pass through critical points in the cycle of relative power. 187 Doran has essentially put forward a mediated stimulus-response model in which systemlevel changes are correctly perceived by leaders of major powers. 188 The model does not posit that system-level changes will produce extensive war in a deterministic fashion; instead, it is necessary for the systemic turbulence to be actually recognized by decision makers. Individual perceptions are an important intervening variable between critical points and war. The proximate causes of major warfare are the actions taken by states in response to the perceived systemic changes and to their evaluation of these changes, but the systemic changes themselves are the root or ultimate causes of war. 189 Indeed, the theory is somewhat similar in structure to Vasquez’s steps-to-war theory. The root cause (systemic structural change) does not directly initiate war, but instead triggers a set of risky, reckless, and threatening behaviors. These behaviors are then perceived as threatening by others, and are then reciprocated, usually by a set of great power rivals. Critical points don’t trigger war directly, but they trigger the first in a series of steps to war. A fundamental question is why should statesmen be so overcome with shock and anxiety at critical points in the power cycle? Doran provides several interrelated reasons. The arrival at a critical point in the power cycle is generally unpredictable in advance. Thus, there is always an element of surprise when leaders discover what has taken place. Moreover, the nonlinearity of the event adds to this surprise. This is especially true at the upper turning point of the curve, when states realize that although their absolute level of capability continues to increase, they are nevertheless in a decline relative to others, which creates a feeling of helplessness and fatalism. Additionally, at critical points leaders sense that they are in a new ballgame—in a place where they have not been before. They are confronted with complexity and uncertainty and feel largely unable to anticipate the consequences of their actions. In many ways, the arrival at critical points creates a crisis-like situation. Leaders perceive a threat to fundamental values and interests; they are confronted with increased complexity and a lack of information;

450

Chapter 11

communication is subject to more frequent misinterpretation; the effects of stress are more pronounced and rationality is degraded, while thinking becomes more simplistic; and judgment is increasingly affected by emotion. All of this is compounded by the possibility that there will be multiple simultaneous critical points. Finally, in this atmosphere of crisis, uncertainty, anxiety, and threat, other sources of systemic instability—hegemonic decline, systemic deconcentration of power, status discrepancies, and dyadic power transitions—become much more salient and threatening. 190 The Cycle of Relative Power: Empirical Research How does the theory fare in the face of empirical testing? Pretty well according to its proponents. Doran and Parsons investigated the cycles of the nine states that made up the major power system from 1816 to 1965. They hypothesized that the propensity to initiate extensive war was greatest during the four critical points and should diminish as the distance in time from such points increased. There were twenty-three critical periods during the years investigated and twenty-six wars were initiated during these critical periods (actually in “critical intervals” just before, during, and after the critical points). The more numerous noncritical periods contained fifty-one initiations of war. However, the average magnitude, severity, and duration of wars were much higher for the critical periods, allowing Doran and Parsons to conclude that major powers are more likely to initiate wars that become extensive during one of the critical periods. The frequency of war is not related to relative power cycles. Only 34 percent of the wars fell during the critical periods, but wars that were initiated by the great powers during a critical period escalated far beyond wars initiated at other times. 191 That is, relative power cycles are not associated with all wars, just major wars among great powers. The inflection points seemed to be more prone to war initiation than the turning points. Great powers appeared most likely to start wars when their rate of growth in relative capabilities reaches a maximum (first inflection) or a minimum (second inflection). This runs somewhat contrary to popular thought, which presumes war to be most likely when states are at either the apex or nadir of their relative power curves, and further research by Doran in fact seems to put more emphasis on the upper and lower turning points. Generally speaking, the nineteenth century after 1815 was free of major war because there was little movement of actors into or out of the system, and the leading powers of the system were traversing through the portions of their power trajectories that were smooth and predictable. Marginal adjustments based on early detection were routinely handled. Conversely, major power war engulfed the first half of the twentieth century because a large number of states entered critical points in their cycles and because new states entered the system (Germany in the late nineteenth century, followed by the United States and Japan) while some old members (Austria-Hungary) dropped out, radically transforming relative power positions and roles of other states in the system. Together, these abrupt changes indicated a critical transformation in the system. 192 Doran reexamined the impact of power cycles using recalculated critical points (treated as five-year periods) and a more restrictive definition of the critical interval in which a state’s behavior should be affected (ten years: three years prior to the midpoint year and six years

The International System Level of Analysis, Part II

451

after). Once again the critical interval was found to be strongly associated with major war initiation. 193 Doran also claims that there is a pattern in major wars regarding the identity of both attackers and targets. The attacker is typically a state that has recently passed through its upper turning point and the target is a relative “pygmy” at the lower turning point whose rapid growth is most responsible for the attacker’s relative reduction in power. 194 The archetype of this pattern is Germany’s attack on Russia in 1914. (Russia reaches its lower turning point between 1894 and 1899, and Germany traverses through its upper turning point from 1902 to 1907.) The crux of the issue was that although it was increasingly dissatisfied with its role and status in the system, Germany could wait for the achievement of these goals as long as it could anticipate that it would continue to grow in relative power. This rosy situation came to a screeching halt a decade prior to 1914 when Germany reached its peak of relative power. 195 Another test by Doran pitted Organski’s power transition theory against his own power cycle theory. The test investigated the five wars from 1816 to 1975 that involved major power contenders (the top three or four states in the system) on both sides—the Crimean War, the Franco-Prussian War, World Wars I and II, and the Korean War. The results indicate that critical changes in relative power are much better predictors of extensive war among contenders than either equality of power or the presence of power transitions. Many of the warring dyads were characterized by power symmetry (relative equality) and many were also characterized by a power transition; nevertheless, both symmetrical dyads and transitions led to war less than half the time. On the other hand, 90 percent of the critical points (100 percent if one counts the USSR as a participant in the Korean War) led to major war. And while only 36.5 percent of the states experienced critical points in their power cycles during the period, all of the sixteen states at war either were undergoing a critical change or were fighting against a state that was. All the war participation of the noncritical-point states was against a criticalpoint state. Perhaps most interesting of all, Doran found that in every case in which a power transition led to war, a critical point was also present. While it was not necessary for power transitions to take place in order for critical points to lead to war, power transitions led to war only when a critical point was also present. 196 Given the existence of a critical point, the presence of a power transition would appear to add nothing to increase the statistical probability of war. Using Doran’s data set on critical points, Houweling and Siccama re-analyzed the relationship between critical points and power transitions among major power contenders. 197 They concluded that the presence of a critical point by itself is a necessary condition for involvement in war by contending great power dyads. Each of the eleven wars in the period was preceded by a critical point, and no great power went to war without having passed through a critical point. But since nine critical-point states did not go to war, critical points are not a sufficient condition for war—something else is needed. Houweling and Siccama find that the presence of a critical point in a dyad that is experiencing a power transition is much more dangerous than a critical point in a nontransition dyad. They argue that a power transition in dyads of major powers, together with the passing (by at least one of the nations concerned) through a critical point on its capability trajectory, is a complex sufficient condition of warfare in these dyads: No pair of great powers that meets this complex condition

452

Chapter 11 remained at peace in the period of observation. However, this complex sufficient condition is far from also being necessary for war: Seven dyads in which the complex sufficient condition was not met did, in fact, fight. The implication is that the complex sufficient condition transition/critical point is but one of many conditions that allow for warfare between major powers. 198

While Doran developed power cycle theory to explain extensive wars among the great powers in the central system, his followers have now decided that the theory can be extended to explain wars between middle powers within regional subsystems and even militarized disputes short of war. Andrew Parasaliti has analyzed data for seven Middle Eastern states—Iran, Iraq, Saudi Arabia, Egypt, Syria, Israel, and Jordan—from 1951 to 1991. In accounting for Iraq’s war with Iran (1980–88) and with Kuwait (1990–91) he notes an astounding number of critical points for relevant states between 1975 and 1990. Iran goes from a high (1975) to second inflection point (1978) to a low (1981). Iraq goes from a low (1975) to its first inflection point (1980) to its high point (1987). 199 Sushil Kumar has analyzed the three-state South Asia regional system of India, China, and Pakistan for the years 1950–95. He also finds a tremendous amount of structural volatility. China progresses through an entire cycle and emerges at a second lower turning point—five critical points in all. Pakistan hits its second inflection point and a lower turning point, and India marks an upper turning point. Against this background of eight critical points, the region experienced four wars between India and Pakistan, two situations of near-war brinkmanship between the same pair, and the 1962 war between India and China. 200 The two studies indicate that regional power cycles are shorter and more volatile, with more critical points in a shorter period of time, than in the central system. This seems to explain the greater amount of conflict within regional systems. Research on power cycle theory has continued to assess connections between critical points and other phenomena associated with war. Hebron and James investigate critical points in the context of international crises. 201 They use prospect theory to generate hypotheses about the relative risk propensity of states during each of the four critical points in the state’s cycle of relative power. State leaders are likely to frame their situation differently—in terms of whether they believe they are in the realm of gains or losses—at different critical points. They hypothesize that crisis activity will be greater for major powers when they are near critical points. Specifically, states are more likely to be challengers in a crisis during the two critical points on its upswing (at the lower turning point and the first inflection point), and they will tend to be on the defensive during their relative declines (at the upper turning point and the second inflection point). Using data on critical points from Doran and foreign policy crises from the ICB Project from 1918 to 1985, their analysis produces mixed results. It appears generally true that the maximum crisis activity of the great powers occurred in the same decade as their critical points, but this was not true for all great powers, especially for the USSR. It also appeared generally true that great powers were more involved as challengers in crises during their upward phases in the cycle of relative power. Nevertheless, the statistical results were far short of overwhelming. There is also evidence that critical points are associated with other phenomena that are involved in conflict spirals that precede war. In a separate analysis, Hebron and James also put forward modest evidence that the timing of critical points has an impact on a state’s participation in militarized disputes. 202 They also find some evidence that states tend to be the aggres-

The International System Level of Analysis, Part II

453

sor in MIDs during the upslope of their power cycles and defenders on the downslope. Once again, however, the results were far from perfect. Similarly, Chiu finds that critical points are statistically associated with alliance formation. As critical points lead to high levels of insecurity, state leaders create alliances as one way to protect themselves in the face of the massive uncertainty and role transformations that are associated with critical points. 203 Tessman and Chan expand the inquiry into the realm of deterrence confrontations. 204 Like Hebron and James they use prospect theory to suggest that leaders whose states are at certain critical points will find themselves facing potential losses and will therefore be more prone to risk-taking. They will therefore be more likely to initiate deterrence challenges and less likely to use accommodation to resolve such situations by measures short of war. Leaders whose states are at the upper turning point, will clearly face the domain of losses, and should thus be most risk-acceptant. Leaders of such states are both overconfident (because they have risen to the top) and insecure (because their power has now peaked and will henceforth decline relative to others) at the same time—a dangerous combination. Leaders whose states are at the lower turning point will face a domain of future gains and should therefore be least risk-acceptant. Leaders whose states are at either of the two infliction points fall in the middle; they will simultaneously face potential losses and gains. Of the two, the authors believe leaders at the first inflection point will be more risk-acceptant than those at the second infliction point. Using eighty-eight cases of immediate (direct or extended) deterrence involving major powers as both defenders and challengers, and extending Doran and Parsons data on critical points to 1995, their findings strongly confirm the importance of critical points. They find that great powers are significantly more likely to mount deterrent challenges during critical points than during noncritical periods, especially around the first inflection points and the peak of their power curve. Moreover, challenges initiated during critical periods were more likely to escalate to war. This reinforces the argument that at critical points states are fearful of their security. Deterrence encounters among major powers rarely result in war; but decisions for both challengers and defenders to pull back from armed confrontation are much more likely to occur during noncritical periods than critical periods. Doran’s original research linking critical points and extensive wars has now been repeatedly replicated over the years by various scholars, and at present no empirical research has directly challenged the power cycle theory. Power Transitions and Historical-Structural Theories: A Comparison Despite significant differences, the power transition theory, Copeland’s dynamic differentials theory, Doran’s power cycle theory, and the three historical-structural theories discussed here share several similarities. (1) They were initially formulated to address major wars rather than all interstate wars. (2) They all link changes in the distribution of relative power capabilities to war, though some emphasize military capabilities and others economic capabilities. (3) Except for Doran’s cycle of relative power theory, they focus on the struggle between the dominant state in the system and a rising challenger. (4) They all emphasize long-term causes of war. (5) In all the theories, uneven rates of growth among the system’s members lead to a changing distribution of capabilities (economic, military, or both) and play an important role in creating a kind of systemic disequilibrium that may lead to war. (6) Most would accept the insights of status discrepancy theory and agree that a state’s relative satisfaction or dissatisfaction with its

454

Chapter 11

position in the international system is an important factor in its war behavior, though this is more important to some than to others. (7) With the exception of Copeland and Doran, they all tend to see a unipolar concentration of power in the hands of a system leader as representing a relatively stable and peaceful situation, though they also agree that this situation cannot last indefinitely. Some of these hegemonic realists identify phases of hegemonic rise and fall and some do not; other than the common identification of Britain and the U.S. as hegemonic powers, they disagree on the existence of other potential hegemons and the specific time periods during which the system was dominated by a hegemonic state. A theoretically interesting point that frequently emerges from research at the systemic level is that many of the factors proposed as having a causal effect on war may operate in conjunction with certain other systemic or dyadic variables. For example, drawing on this chapter and the previous chapter: (1) Long cyclists, world systems theorists, and others agree that global leadership cycles somehow interact with global economic phases to produce major wars, though they don’t agree exactly why. (2) Some, like Stoll and Champion, suggest that a concentration of power within the system may be dangerous or peaceful, depending on the distribution of satisfaction in the system. (3) Rasler and Thompson posit that world wars are most likely when power at the regional level is concentrated but is deconcentrated at the systemic level. (4) Doran has argued that Organskian power transitions only lead to war when combined with a nation’s passing through one of its critical points in the power cycle, and Houweling and Siccama find concurrent power transitions and critical points to be a sufficient condition for great-power war. (5) Research by Daniel Geller also indicates that dyad-level factors and systemic-level factors are interrelated. Focusing on the links between hegemonic stability theory and power transition theory, he concludes that power shifts among contender states had to be accompanied by conditions of systemic deconcentration in order for war to result. 205 Thompson also found that the conjunction of deconcentration in the world-system cycle with a dyadic power transition was especially dangerous. SYSTEMIC THEORIES: IMPLICATIONS Is it possible to apply anything that we have learned from the theories in this chapter to the present age or the immediate future? Many social scientists would raise the question whether theories that use changing balances of power capabilities to predict major war between the leading state and a challenger “remain relevant in a world in which declining benefits of conquest, nuclear deterrence among most major powers, the spread of democracy, and changing collective norms and ideas reduce the probability of major war among great powers to a historical low level.” 206 We should at least raise the possibility that theories of hegemonic transitions—such as Gilpin’s hegemonic stability theory and Modelski and Thompson’s long cycle/world leadership theory—are no longer relevant. It is entirely possible that they are applicable only to a particular set of international systems in a relatively narrow window of time in which several European great powers lived in close proximity to each other without the benefit of mutual democracy or nuclear deterrence. That system—thankfully—no longer exists. To the extent

The International System Level of Analysis, Part II

455

that these theories provide a useful explanation of the behavior of states, that explanatory power quite probably has historical limits. Having said this, let us try to kick the can a little ways down the street. To the extent that changing distributions of power and status in general and the decline of systemic leadership in particular precipitate war, what are the implications? We know that maintaining the status quo is impossible; if we have learned anything from systemic (and dyadic) studies it is that differential rates of growth make a stable balance of power impossible in the long run. Whatever stability exists, hegemonic or otherwise, must be merely temporary owing to the more general disequilibrium of uneven development. The systemic changes that have taken place over the last quarter century have been massive. Political turmoil in Eastern Europe led to the peaceful replacement of communist governments with pro-Western regimes who then embarked on transitions to democracy. Equally dramatic political changes in the Soviet Union led to the dissolution of the union itself, the rapid decline of Soviet/Russian military capabilities, and the dismantling of the Soviet military bloc. NATO, now enlarged by many post-Soviet states, appears an alliance without a mission. The creation of a unified Germany has brought a theoretically powerful (yet peaceful) old actor back into the heart of Europe. Meanwhile in 2010, China—with an annual growth rate of 10 percent over ten years—surpassed Japan as the world’s second largest economy. The United States, while still the most powerful state in the system by almost any measure, continues to decline relative to its chief economic competitors. A number of things can be said about the transformation that began in the crucial years, 1989–91. The changes were by all measures truly revolutionary. One analyst even refers to the transformation as the “functional equivalent of a hegemonic war” without the violence. 207 Virtually all the systemic changes that structural-historical theorists talk about when they mention hegemonic transitions have taken place. The late 1980s and early 1990s witnessed power shifts of the first magnitude; several major states passed through critical points in the power cycle; and alliance polarization decreased. As polarization decreases and the distribution of power changes, we are witnessing a fundamental transformation of the international system. A new distribution of power has developed; territorial adjustments have been undertaken in Europe and in the former USSR; a new political alignment has emerged; a new hierarchy of prestige is evident; and a new set of norms may (or may not) be taking effect. There is little doubt we are witnessing a fundamental transformation of the international system. Perhaps most remarkable is the fact these changes have been largely peaceful; hegemonic or global wars have not been part of the package. Based on our previous discussion of systemic-level theories, this is an outcome one would not necessarily have expected. Explaining our good fortune will take some doing. The following analysis should be considered both rudimentary and preliminary. First, a power shift has taken place, but it is not the war-inducing power transition that theorists usually talk about. This transition was one in which the declining state was the challenger (the former USSR) rather than the hegemonic power (the United States); and in this case the challenger never caught up. The power shift that took place was not toward parity, but away from parity, widening the gap between the United States and its primary adversary.

456

Chapter 11

Although there is some reason to believe that power transitions away from parity are dangerous, transitions in which potential challengers move toward parity with the dominant state have historically been more likely to lead to war. 208 Additionally, the potentially dangerous decline in Soviet power was offset by the fact that the Gorbachev administration had begun to seek accommodation with the U.S. as its primary strategy for dealing with that decline. The United States and Russia were at the time actually engaged in a mutual disarmament race, at least with regard to strategic weapons. Russian conventional military power continued to decline precipitously, and the Bush-Putin and Obama-Medvedev presidencies have been successful in further reducing strategic nuclear weapons. However, while the U.S. has reduced its nuclear arsenal, the effects of the September 11 attacks by Al Qaeda, the subsequent war on terrorism, and two asymmetric wars in Iraq and Afghanistan (along with bureaucratic inertia) have meant that the conventional U.S. military has actually expanded in size in the decade after 2001—increasing the gap with its competitors. Second, systemic polarization—what we have called cluster polarity—has decreased drastically. The Cold War bipolar alliance system no longer exists. NATO is no longer an antiRussian organization: The Yeltsin-Bush summit in early 1992 produced a statement that the two countries were friends rather than enemies. The Warsaw Pact is defunct and many of its former members have asked for, and attained, membership in NATO; even Russia has a partnership agreement with NATO. Since military alliances have little meaning unless they are organized against another state or group of states, one can only conclude that the system is unpolarized. On balance, this lack of alliance polarization would seem to be a good thing. It has clearly been associated with reduced international tension. It has provided the background against which the Russian-American arms race has been reversed. And to the extent that Russia will now be more likely to cooperate with the other major powers to reduce international conflict through the United Nations (as it did in the Persian Gulf crisis) or through regional organizations (such as the Conference on Security and Cooperation in Europe), the loss of polarization should prove to be beneficial to world peace. However, while conflict between core members of the system may become less likely, the former superpowers may be less willing or able to constrain themselves with regard to lesser states in the periphery or semi-periphery. The core members of the system are still tempted to intervene militarily in areas where they perceive their security and interests to be threatened: witness the Russian intervention in Georgia to help “liberate” South Ossetia and Abkhazia in 2008, the U.S. and British invasion of Iraq in 2003, and NATO intervention in Libya in 2011 to bring down the government of Muammar Gaddafi. What about power distribution (polarity)? The fashionable argument is that the demise of the Soviet Union has created a unipolar international system in which the United States is the only true superpower—the only state with superior capabilities on all relevant domains of power: economic, military, and technological. The logic of this assessment is simple: Most theorists agreed that the years 1945 to 1989 were characterized by a bipolar system; one of those poles (the USSR) ceased to exist in 1991 and no state has taken its place; ergo, the system must be unipolar. 209 The most forceful argument for unipolarity has been put forward by William Wohlforth and his associates. The statistical facts here are almost staggering. In 2006 the U.S. accounted

The International System Level of Analysis, Part II

457

for about a quarter of the world’s total gross domestic product (GDP) and almost half of the GDP belonging to the great powers. The American relative economic advantage over its competitors surpasses that of any major power in history, except for the U.S. itself as it emerged from World War II in 1945. With regard to military power, the United States probably spends more on its military than all of the other major states in the world combined—a staggering fact to contemplate. Its military research and development (R&D) is six times greater than Germany, Japan, France, and Britain combined. It may in fact account for half of all military R&D in the world. 210 Few would disagree with the statement that no other country has the military wherewithal to mount a major military expedition outside its own region except with the assistance of the United States. Thompson adds that one of the unusual things about the current system is that the U.S. military is both the leading sea power and the leading land power. While it does not possess the largest army in sheer size, it is the leader in lethality, technological sophistication, and the ability to project force over long distances. As Thompson notes, “Normally the leading whale is not also the leading elephant.” 211 As the twenty-first century dawned, the U.S. confronted an environment in which its major economic competitors (Japan and the European Union) were not necessarily military competitors and its major military competitors (Russia and China) were not sufficiently advanced economically. That assessment may now be changing. China is clearly an economic competitor (and one with whom the U.S. has a huge trade deficit). However, even though China’s economy is now the second largest in the world, it is still roughly one tenth the size of the American economy and its per capita GDP lags far behind. Nevertheless, its economic growth rates are astounding. Furthermore, it has begun to modernize its military. While it is true that the United States is the only state with the combined military-economic-technological capacity to be called a true superpower, there is nevertheless an element of multipolarity in the system. Russia remains a state with a large nuclear arsenal, though some suggest that American advances combined with Russia’s state of military and technological disrepair may have reduced considerably Russia’s second-strike retaliatory capability. 212 China continues to grow both economically and militarily. Germany and Japan have surpassed the United States on some indicators of economic strength. The European Union has become stronger politically as well as economically—the economic difficulties of 2008–13 notwithstanding. It is doubtful that the United States, though a unipole, could be classified as a hegemonic power. It plays a major role (with others) in developing norms for the system and enforcing these norms, and it is itself probably above those rules whose strictures it wishes to avoid—though not without consequence. 213 However, it is not always able to maintain order; it can’t reliably exert influence over friends or opponents; and it is unable to prevent a contrary international consensus on issues that it deems important. A few examples should suffice. 214 1. While the United States was clearly the only military power with the ability to transport and deploy sufficient force to counter Iraq’s conquest of Kuwait, the insolvency of the American government meant that the United States had to go begging for contributions from its friends to defray the cost of the expedition. 2. The United States and its allies were unable to prevent the collapse of the Doha round of the World Trade Organization negotiations in 2008—a serious blow for the

Chapter 11

458

3. 4. 5. 6. 7.

8. 9.

American position of further arranging the world trading system around the principle of free trade. The U.S. was unable to prevent international consensus on the Kyoto Treaty and on the creation of the new International Criminal Court at the Hague. The U.S. failed to restore order in Somalia and other failed states. The U.S. has failed to prevent nuclear proliferation in India, Pakistan, North Korea, and possibly Iran. The U.S. failed to attain a UN Security Council consensus on the need to use military force against Iraq in 2003. The U.S has been unable to get major European actors like Germany and Britain to pursue American stimulus policy toward recession rather than the road toward fiscal austerity and budget balancing. The U.S. has failed (though some would say it did not try very hard) to deter or punish genocidal-type massacres in Congo, Rwanda, and Sudan. The U.S. was unable to prevent Russian intervention in regional conflict in Georgia and the subsequent dismemberment of that state.

Thus, we have to conclude that the current international system moved from bipolarity to unipolarity, but the unipole (the United States) is probably not a hegemon. The presence of a unipolar system has not, of course, meant the end of interstate warfare. It may portend a situation in which major power war is extremely unlikely, but this was also the case for most of the previous Cold War era. It may very well be that major power war is unlikely not because of unipolarity but because of several other conditions. 215 First, most major states have nuclear weapons. Second, even a conventional war between major states would be extremely costly. Third, the spread of democracy and liberal values (and European integration) means that the developed states of North America and Europe can hardly conceive of the possibility that there might be war between them. Finally, in a unipolar system more than any other type of system, the regime and leadership characteristics of the leading state also matter. It can be argued that American unipolarity is less threatening to other major states than if the system were led by, for instance, an autocratic state in the heart of Eurasia with plentiful neighbors. 216 At any rate, the crucial point is probably not whether the system is unipolar, bipolar, or multipolar, but what happens in the transition from one distribution to another. We seem to have weathered the latest transition in good shape. Systemic-level theories tell us that war is the result of broad socio-political-economic forces beyond the day-to-day control of individuals or governments. Individual decisions to start wars merely represent the culmination of these large-scale impersonal forces. Consequently, attempts to prevent the outbreak of wars and to manage international structural change will be precarious at best. National leaders have had, and will continue to have, little control over the long-term cycle of national growth and decline and the changes in political relations that flow from them. Systemic theorists offer somewhat different analyses of the paths to peace. Polarity theorists rely on the ability of major actors to construct balances of power in either bipolar or

The International System Level of Analysis, Part II

459

multipolar environments. A few theorists (such as those in the world-economy school) pin their hopes on a peaceful transition to a new international system based on different principles. However, the message of most of the theories discussed at this level of analysis—status discrepancy, Gilpin’s theory of hegemonic war, Organski’s theory of power transitions, Modelski’s theory of world leadership, and Doran’s cycle of relative power—is that the major powers will need to peacefully and cooperatively manage the inevitable transitions in power, role, and status that occur in the international system. While in the past these changes have been accompanied by war, we must develop new, more peaceful mechanisms of change. Some of the theorists in this group, notably Gilpin, Doran, Modelski, and Thompson, believe this is possible. While systemic forces are crucial causal factors, they are not entirely deterministic; political responses to these phenomena are to some extent subject to freedom of choice. The world has recently witnessed what may be the first major systemic transformation to have taken place without large-scale interstate warfare. During this potentially explosive period, international leaders have taken great care to manage these momentous systemwide changes by building new international institutions or enhancing old institutions and by consciously seeking ways to reduce the grievances of the major loser (the former Soviet Union) in the transformation. Perhaps there is reason for optimism. The critical issue in the long run in the eyes of most historical-structural theorists is whether and how the declining global leader (whether it is actually in relative decline now or whether its tenure at the top of the international system will be a prolonged one is open for discussion) and the other major powers will accommodate what is seen as the inevitable rise of China. This must be done in a way that accords the status and roles that China will deserve as a result of its economic and military capabilities, thereby preventing status discrepancy and dissatisfaction. This currently seems to be happening. A level-headed analysis sees China not as a dissatisfied, revolutionary state bent on challenging the global leader but as a relatively satisfied member of the great-power system. China has a permanent seat on the UN Security Council; it sought and obtained membership in the World Trading Organization; it is presently at peace with all of its neighbors; it participates in regional multilateral organizations such as the Asia-Pacific Economic Cooperation Forum (APEC), the Shanghai Cooperation Organization, and various groups associated with the Association of Southeast Asian Nations (ASEAN); it is a participant with the U.S., Russia, South Korea, and Japan in the Six Party talks with North Korea; its territorial claims are local rather than regional (Taiwan, the Spratley and Paracel Islands); and its military rebuilding effort has been modest and aimed primarily at the regional rather than global level. China is satisfied, in David Wilkinson’s words, “to be treated as a formal equal—for instance, to be consulted, to criticize, to delay, to limit and then to abstain from U.S.-led joint undertakings.” 217

Chapter Twelve

Constructivism A Digression

Anarchy is what states make of it. —Alexander Wendt I’m not crazy about reality, but it’s still the only place to get a decent meal. —Groucho Marx

In this chapter we take a quick look at one of international relations theory’s younger children—constructivism. In a book organized by levels of analysis, constructivism poses a considerable challenge because it transcends and transects our five levels. The following will be a brief tour through the land of constructivism and its potential contributions to the question of what causes war. Constructivism is a relatively new approach to international relations that emphasizes the importance of ideational and cultural factors in explaining the behavior of states and their leaders. The term was first introduced in 1989 by Nicholas Onuf in his book World of Our Making. 1 Constructivism emerged to help scholars deal with questions for which realism and liberalism struggled to find answers: Why did the Cold War end? Why did NATO expand after the collapse of the Soviet Union and the Warsaw Pact alliance? Why don’t states in Western Europe fight each other anymore? Why does an Iran with nuclear weapons scare the United States but not an Israel with nuclear weapons? Why have most states refrained from using chemical weapons since WWI even though they have significant military utility? What accounts for the increased importance of human rights in the calculations of states and their leaders? Like the realist and liberal schools of theory, constructivism is fairly diverse. However, one can identify a central core of themes common to most constructivist theorists. THE CONSTRUCTIVIST APPROACH: BASIC PRINCIPLES 1. Constructivists argue that international politics are socially constructed. Reality is what people “construct” as reality through their interactions with each other. Here is a short list of some things we can consider to be the products of social construction: war, threats, enemies, 461

462

Chapter 12

national identities, national interests, operational codes and worldviews, rivalries, the standard operating procedures of governmental institutions, anarchy, sovereignty, appeasement, commitments, the democratic peace, and international law. Constructivists generally focus on three types of ideational phenomena: identities, interests, and norms. 2. An important part of international politics is the creation of identities, defined as “mutually constructed and evolving images of self and other.” 2 By their very nature, identities are relational and “mutually constitutive”: Individuals, groups, and states have identities only in relation to others. The concept of “we” is unthinkable outside of a relationship to “they.” As one scholar puts it, “identity doesn’t belong to one person alone, and no one belongs to a single identity.” 3 Identities of states, nations, and individuals are malleable. In this sense, there may be no “original” or “natural” identities since all identities are subject to a continual process of construction and reconstruction. Identities also depend on the social context; as individuals, we act differently when interacting with different “others” in different situations or contexts. Government leaders have multiple identities, many of which are linked to institutional roles—as party chairs, as prime ministers, as “honest brokers,” as “defender of the faith,” or as ultimate “deciders” at whose desk the buck stops. States also have multiple identities: faithful allies, leaders of the free world, offshore balancers, or victims. Each of these identities “is an inherently social definition of the actor grounded in the theories which actors collectively hold about themselves and one another and which constitute the structure of the social world.” 4 States not only have a sense of their own identities but assign identities to others—allies, neutrals, rogue states, competitors, and enemies. States (and their leaders) understand others according to the identities they attribute to them. From the constructivist perspective, war is very much intertwined with the establishment of the identity of others as rivals, threats, opponents, and enemies. 3. There is an important connection between a state’s identity and its interests. National interests are intersubjective understandings about how a state advances its power and wealth. 5 They are inextricably bound up with the process of identity construction. As Ted Hopf notes, identities perform three important functions: they tell you who you are; they tell you who others are; and, in telling you who you are, they imply a particular set of interests or preferences with respect to policy choices in a particular context with respect to particular others. 6 The identity-interest relationship is not a “which came first—the chicken or the egg” issue. Interests, identities, and norms are all the result of an ongoing, iterative, interactive social process, and they are all mutually constitutive. 7 4. Constructivists emphasize the importance of norms and social roles that develop out of social interactions. Norms are formal or informal standards or rules for behavior. They create expectations about how we should behave and how others should behave in certain circumstances. Rules also endow individuals with different capabilities (agency) based on shared understandings. Norms constrain and regularize behavior, and they create order. As Onuf says, “rules create rule.” Norms always exist within specific contexts; different contexts contain different norms and identities. The context may be within the larger international system, within a specific regional subsystem, within an alliance, within in a dyadic relationship between states, or within the context of a particular organization, like the Department of De-

Constructivism

463

fense. Norms shape the interests of individuals and states, and interests shape actions. Thus, norms guide actions, but they do not determine them. Constructivists typically distinguish between regulative norms (which regulate behavior) and constitutive norms (which tell us who we are and which may create new actors and new interests). Norms, like all social phenomena, change. This is important because over time a change in norms may lead to a change in state interests, and therefore state behavior. 5. Constructivists argue that the realist and liberal view of states as rational actors and as utility maximizers is too confining. Constructivists would say that actors are guided by the logic of appropriateness, as compared with the logic of consequences used by rationalist theories. 8 As individuals, we act differently in different contexts. For instance, Ba and Hoffmann ask students “Would you ever wear a blue leisure suit to class?” 9 In the context of the 1970s this would probably have been considered cool. But in 2013? Probably not. Likewise, would modern industrial states ever use chemical weapons against each other? They did in the context of World War I, but today it would be virtually unthinkable. In the real world, “what is rational is a function of legitimacy, defined by shared values and norms within institutions or other social structures rather than purely individual interests.” 10 Social structure matters as well as self-interest. For instance, human rights norms constrain the actions of states not because of considerations of power capabilities (and of consequent sanctions), but because human rights in part constitute the identity of states, especially states that are democratic. It is part of who they think they are. 6. Constructivists focus on change. Since the identities and interests of states are based on continual interactions, they can change over time; they’re not fixed. Reality is, in a sense, always in the process of “becoming.” There is a continual interaction between agents and their structural and contextual environments. As identities change, interests and norms will change and so will state behavior. And changing behaviors and practices will then change the intersubjective knowledge that constitutes systems and subsystems. In sum, all the important elements of constructivist theory are moving targets that are involved in continuous interaction with each other in a giant loop of mutual feedback. 7. To a large extent constructivism was born as a critique of neorealism’s view that the anarchic structure of the international system determined state behavior. Constructivists like Alexander Wendt rejected the notion of systemic determination. They believed that states had much more freedom or agency. But agents (actors) don’t exist independent from the particular social environments in which they are embedded. By their actions, the agents in international politics (that is, individuals, groups, and states) help to create structures—such as treaties, rules, norms, domestic and international institutions, and the global distributions of power— that in turn affect and constrain them; the agents and structures are “mutually constituted” or “co-constituted.” 11 From this point of view, the structure of the international system by itself cannot cause states to act in any particular way. In Wendt’s famous phrase: “Anarchy is what states make of it.” 12 Its meaning depends upon the context. Thus anarchy (or hegemony) can have different meanings for different actors depending on the intersubjective understandings and practices that they have developed among themselves. 13 Anarchy has no inherent logic of mutual suspicion and competition. For example, there is no particular reason that when two states

464

Chapter 12

“meet” in an anarchic system, they should immediately fear each other. After all, prior to any social interaction between two actors, there is really no such thing as a threat. 14 If anarchy leads to a security dilemma and to a competitive and conflictual self-help system, it is because of a history of interactions through which actors defined (constructed) their identities and interests in a certain way and which created certain fearful expectations for all parties. Self-help is neither an automatic nor an essential feature of anarchy. Anarchies can be conflictual or cooperative. In Wendt’s often quoted words: “Anarchy as such is not a structural cause of anything. What matters is its social structure, which varies across anarchies. An anarchy of friends differs from one of enemies, one of self-help from one of collective security, and these are all constituted by structures of shared knowledge.” 15 An anarchy of friends can even create “security communities” among themselves. 16 States create their own “hell”; it doesn’t necessarily come with the territory of anarchy. On the other hand, actors can’t just will their way out of that hell; they’re already embedded in social structures not necessarily of their own making. They can, however, work to change these structures. 17 8. For constructivists, the structure of the international system—its polarity, for instance— is not strictly material (that is, based on the distribution of power capabilities). There is also an ideational component. Of course, material power is not irrelevant for constructivists; after all, it makes a difference whether, for instance, the right to use preventive force is advocated by a hegemonic power like the United States. Strong states are better positioned to set precedents and to reinforce (or de-legitimize) international norms. States are not equal in terms of normative weight. But the international system is also made up of social and ideational relationships—including mutually constituted ideas, norms, and practices—that have been constructed by certain states at a particular time and in a particular setting. International structure is essentially determined by the distribution of ideas. As Finnemore and Sikkink put it, “Shared ideas, expectations, and beliefs about appropriate behaviors are what give the world structure, order and stability.” 18 The effects of material factors (like military capabilities) aren’t predetermined; their impact depends on how we think about them. The international system is anarchic (or bipolar or hegemonic) in large part because international elites perceive, interpret, and evaluate it to be anarchic (or bipolar or hegemonic). This social construction of the international system then shapes actors’ identities and interests as well as their behavior. 19 As a result, many constructivists would argue that the distinction between material and ideational phenomena is a false one, since all material things are endowed with social meaning. Constructivists, like realists, believe that the structure of the international systems helps determine state behavior, but the structure does not necessarily compel certain behavior by states and their leaders; states have more agency, although this ability to act and to choose is rigorously constrained by the social structure. 20 Since change is important, “idea shifts and norms shifts are the main vehicle for systemic transformation. Norm shifts are to the ideational theorist what changes in the balance of power are to the realist.” 21

Constructivism

465

CONSTRUCTIVIST THEORY AND WAR There is considerable debate on the status of constructivism. While it is sometimes referred to as a theory, many others refer to it as simply an “approach” to international relations—a lens or framework through which to make sense of international phenomena. Many constructivists assert that their task is not to provide causal explanations for social phenomena but to simply understand them better. So most constructivists attempt to answer the question ‘How is war possible?” rather than “What causes war?” They look at the meaning of war or how actors understand war, the imagery of war, the narratives of war and discourses of war, without putting forward causal explanations per se. 22 They focus not on war as a dependent variable that is caused by some set of independent variables but on understanding why peace has prevailed in situations in which other theories would predict conflict. Constructivism has also been criticized because it makes few unambiguous predictions (hypotheses) about what we can expect to happen in international politics. As D’Anieri says, while constructivism tells us that ideas matter, its adherents are not very good at providing general rules about how they matter, why they matter, or which ideas will matter more than others. 23 How would constructivists examine the problem of war (or its absence)? A constructivist explanation of war would try to show how the social and ideational structure of the system— the international system, a regional subsystem, or a dyadic system—would make violence possible by endowing (constituting) actors “with certain identities and interests and by endowing material capabilities with certain meanings.” 24 A constructivist explanation of war would be highly contextual, individualistic, complex, and highly contingent. Each war would have a unique explanation. Constructivists tend to emphasize the differences between states rather than their similarities. As Hopf puts it: “The same state is, in effect, many different actors in world politics, and different states behave differently toward other states, based on the identities of each.” 25 At what level of analysis does constructivism operate? Wendt says that constructivists (like realists) see states as the primary actors and the primary unit of analysis, and he also argues that system-level theorizing is important. He categorizes constructivists as “structuralists”; they believe that the structure of the international system “constitutes” state identities and state interests. In other words, a state’s identities and interests are “made possible by, and would not exist in the absence of” the type of system in which they are embedded. 26 Wendt also assumes that states can be treated as unitary actors, thus eliminating the need for analysis of individuals or substate institutions as primary agents or actors. 27 On the other hand, most international relations scholar view constructivism as an approach that is applicable to multiple levels of analysis. Let us think about the ways that the constructivist approach might help us explain interstate war (or its absence) from each of the five levels of analysis. The Individual Level of Analysis Jeffrey Checkel notes that since constructivists focus on norms that are intersubjective and therefore mutually or collectively held, they tend not to focus on individuals as agents. 28 Nevertheless, constructivists have much to say about individual-level phenomena. Constructi-

466

Chapter 12

vists, like cognitive theorists, are interested in ideas and beliefs. But while the cognitive approach focuses on the ideas, beliefs, and perceptions of individuals, constructivists emphasize how the interaction of individuals (and the interaction between states and their leaders) co-create these things. As Theo Farrell says, “constructivists are not interested in the beliefs actors hold so much as the beliefs actors share.” 29 Ultimately, they don’t really care about individuals per se, but rather the discourse, the identities, and norms that individuals draw on when making sense of the world and when making policies. Constructivists see individuals as embedded in social institutions and are interested in how institutions shape them and vice versa. If individual perceptions, beliefs, and worldviews are created through social interactions, then several types of social interactions might be important for constructivists. For instance, the views held by the leader of a state might result from interaction with his peers within a particular domestic institutional environment. President Putin’s worldview may be largely a product of his interactions with his peers in the KGB in his formative adult years, and these views may represent an intersubjective consensus of shared beliefs, norms, and expectations held by the national security elites of the old Soviet Union. One could also argue that the relevant set of interactions that form Mr. Putin’s beliefs are his interactions with other states in his capacity as a world leader. The bottom line is that war is more likely if political leaders have developed—due to their interactions with their peers and with other states—certain identities and views of other states (as rivals, enemies, threats) and certain techniques for dealing with such states (brinkmanship strategies, escalation strategies, preemptive strategies). Remember, if you can, the dysfunctional operational code that was outlined in chapter 3; constructivists would see this worldview as the product of a leader’s social interactions with others rather than simply a product of the individual’s own cognitive appraisal of the environment. The Substate Level of Analysis At the substate level, constructivism sees the same processes of socially constructed meaning taking place, but with a focus on particular political institutional subcultures within a state: A constructivist analysis might focus on how military elites in France came to emphasize offensive doctrines before World War II or how political-military cultures in Germany and Japan might help explain aggression in the 1930s and 1940s. 30 Here the focus would be on how specific influential security organizations developed a constellation of views—identities of themselves and of opponents, values, assumptions, views of state interests and of institutional interests, a set of norms and standard operating procedures for dealing with others—that might be conducive to antagonistic, aggressive, or reckless behavior. New and emerging norms—whether they are norms of human rights or norms legitimating the preemptive use of force—lead to changes in state behavior. How do ideas like human rights or preemption get translated into policy? At the substate level there exist certain idea men or norms entrepreneurs—intellectual peddlers who attempt to promote and gain support for their preferred views and policies, some of which are clearly novel and contrary to the received wisdom of the times. Norms entrepreneurs may operate within the political and bureaucratic institutions in the bowels of government or they may work outside government in

Constructivism

467

think tanks, NGOs, or INGOs. (Neoconservatives centered on Paul Wolfowitz in the Defense Department played this role in U.S. policy toward Iraq.) Norms entrepreneurs may be attached to epistemic communities—networks of knowledgebased experts—either within a single country or as part of a transnational network. Peter Haas defines an epistemic community as a network of professionals with recognized expertise and competence in a particular policy area who have (1) a shared set of normative and principled beliefs that provide a value-based rationale for actions, (2) a shared set of causal or analytical beliefs, (3) shared norms of validity or criteria for weighting and validating knowledge, and (4) a common policy enterprise. 31 Essentially, members of epistemic communities share the same worldview (episteme). Epistemic communities may frame issues, provide information and advice, and advocate on behalf of preferred policy options. They may even capture and consolidate power within governmental institutions, thus institutionalizing their influence over policy. Policy makers are especially likely to rely on epistemic communities during conditions of major systemic change that brings about uncertainty, when top decision makers lack the knowledge to assess the potential outcomes of different policy options and where standard operating procedures seem inadequate to the tasks at hand. We can look briefly at two incidents of this type, both dealing with the collapse of the USSR and the end of the Cold War bipolar system. In the first incident, we have Gorbachev’s “New Thinking,” which Robert Herman describes as a major reason for the end of the Cold War. 32 In this instance, ideas and new intellectual frameworks were an important determinant of state behavior. Gorbachev’s foreign policies represented an improbable U-turn in Soviet foreign policy from the aggressive “old thinking” based primarily on Marxist-Leninist dogma. How was this possible? Where did these ideas come from? Even within the relatively closed Soviet society, there were different currents of thought, and there were norms entrepreneurs in charge of think tanks, some of whom were part of transnational epistemic communities of liberal international relations experts. In this case Soviet idealists adopted ideas associated with liberal Western scholars— such as the security dilemma—and produced a Soviet version of policy consistent with its prescriptions. (Herman uses the term specialist networks rather than epistemic communities.) Thus, Gorbachev and his foreign minister, Eduard Shevardnadze, championed policies of “mutual security,” “reasonable sufficiency,” “defensive (non-provocative) defense,” interdependence, and a “nonclass” approach to international affairs. They generally pursued an accommodative policy (including some unilateral arms reductions reminiscent of a GRIT strategy) based on a new understanding of capitalism—as both internally vital and externally nonthreatening—that eventually drew reciprocal behavior by the Reagan and Bush administrations in the United States. Without the change in ideas, concepts, intellectual frameworks (and operational codes, if you will), there would have been no changes in policy and perhaps the Cold War would have continued. Herman’s focus is at the individual and substate (institutional) level of analysis. He identifies the Soviet think tank IMEMO (Institute of the World Economy and International Relations) as a prime actor in developing and advocating foreign policy ideas. More specifically, its directors in the Gorbachev era, first Alexander Yakovlev and then Yevgeny Primakov were

468

Chapter 12

clearly norms entrepreneurs who were instrumental in pushing new ideas. They were aided by their personal closeness to Gorbachev and by the changing nature of the international system. 33 The political and cultural roots of New Thinking had their origins in the Khrushchev era, the de-Stalinist campaign, and the era of peaceful coexistence and détente in the 1960s. But, of course, these idealist perspectives did not last. What permitted them to resurface in a more radical “idealism on steroids” version was the rise of a reformist leader in Gorbachev who was receptive to these views and the change in the international system and the Soviet Union’s place in it (especially after the debacle in Afghanistan). As Jeffrey Checkel has noted, dramatic system-level changes often create opportunities for policy change, providing “policy windows through which policy entrepreneurs could jump.” 34 In 1985 the policy window was open, norms entrepreneurs were available with policy ideas that could address the problems presented by the new international system, and there was a leader who was receptive to these ideas. Central to Herman’s analysis is the proposition that radical idealists within Soviet think tanks (and their elite supporters in the party and government hierarchy) underwent a fundamental change in identity. The idealists’ new vision of the USSR was as a “normal country” that would take its place among the world’s advanced industrial democracies, within a “common European home.” As Herman says, “When Gorbachev, Shevardnadze, and others spoke of depriving the West of an enemy . . . they were expressing aspirations of joining the community of democratic states. They now understood that . . . only a common identity would permit the two sides to transform their relationship.” 35 The West would no longer be seen as “the other.” The construction of this new Soviet identity by foreign affairs specialists was, of course, part of a larger, and longstanding struggle going back to the days of Peter the Great between “Westernizers” and “Slavophiles” over Russian and Soviet identity in the modern world. The new identity logically meant moving in the direction of greater democracy at home and abandoning Marxism-Leninism as a guide to defining state interests. The construction of a different Soviet identity was the key to reconceptualizing Soviet security interests. Once the policy elite’s vision of national identity changed, policy changes consistent with that identity would flow freely: unilateral and multilateral cuts in conventional and nuclear forces, restructuring of Soviet forces in Eastern Europe to emphasize defense rather than offense, and eventually permitting Eastern European bloc states to hold democratic elections and leave the Warsaw Pact. The outcome was not determined by the changing structure of the international system; much was contingent and affected by human agency. The second incident is the situation facing the George W. Bush administration in the new post–Cold War system and in the immediate aftermath of the Al Qaeda attacks of September 11, along with his administration’s concern regarding Iraq’s possession of WMDs. With regard to the decision to mount the 2003 Iraq War, the relevant epistemic community would be made up of former national security officials in the first Bush presidency (1989–92) who during the Clinton years made up a community of neoconservative policy intellectuals located in think tanks like the Prospect for the New American Century (PNAC), wrote policy briefs for various conservative journals, and then returned to governmental power in the second Bush presidency. These policy analysts had developed shared understandings about the nature

Constructivism

469

of the new international system (unipolar), the nature of U.S. power (largely unconstrained), the nature of threats to the United States (the combination of transnational terrorist groups, rogue states, and WMDs), and the most appropriate solutions (preventive use of military force to bring about regime change). Now occupying important posts in executive branch national security institutions, they were thus in a position to interpret and frame the situation and to provide solutions to the perceived problem; they also found a president receptive to their ideas. The bottom line at the substate level of analysis is that war is more likely to the extent that critical national security institutions have developed dysfunctional subcultures or institutional worldviews and aggressive standard operating procedures. Conversely, if social-political-economic interactions create a context in which national security institutions can develop nonthreatening worldviews linked to accommodationist policy preferences, the chances of war diminish. State Level of Analysis For many constructivists, states remain the primary actors in international relations. The crucial question for all state-centric theories is “What is it about the nature of some states that make them more prone to war than other states?” Constructivists would most likely address this by focusing on the nature of state identities—both identities of self and of others. A state’s own identity plays a role in determining whether states will be inclined to employ war as a foreign policy tool in general. Its identification of others as enemies or friends also helps to determine which states it will go to war with. Unlike neorealists who see identities as fixed and immutable, for constructivists a state’s identities (and therefore its interests) change over time. The history of great powers like the United States suggests that there is a tendency to engage in a serial search for potential enemies. After Germany and Japan were defeated and no longer to be feared, American elites and the mass public swiftly found new enemies in the Soviet Union and China; with the collapse of the Soviet Union, Washington rather quickly focused on rogue states with nuclear aspirations as well as transnational terrorist groups. The identification of certain significant others as threats—communist states, rogue states, nondemocratic states—not only leads to the creation of new interests but helps redefine the state’s own beliefs about who it is. The notion that states actually have identities presents confounding philosophical dilemmas about whether large, aggregate, corporate entities can have a meaningful collective identity. Certainly, not all theorists would agree with the idea of a unitary state with its own collective ideational attributes. Individuals have beliefs, worldviews, identities, and perceptions, but not states. Moreover, since identities and interests are both intersubjective and mutually constructed, the real focus slides upward toward the systemic or dyadic interstate interactions that helped produce specific identities and interests, or downward to interactions of leaders with their peers embedded in various substate institutions.

470

Chapter 12

The Dyad Level of Analysis One could make the case that the single most relevant political culture might be the structure of the dyadic relationship between a pair of countries, like France and Germany or India and Pakistan. The intersubjective realities that are created through dyadic interaction may make war more likely or they may make peace and cooperation more likely. It all depends. How states treat each other affects the social structure of their relationship through reciprocity. As Wendt says, “History matters.” 36 Future interactions between states are based on past interactions; learning occurs. This is the logic underlying the existence of both enduring rivalries and, on the other hand, security communities. It is also central to the ability of players to cooperate in iterative games. This constructivist argument is accepted, at least in part, by some realists. Stephen Walt has attempted to explain why states do not necessarily balance against the accumulation of power by other states, as realists predict. His answer (as you will remember from chapter 10) is that states do not balance against the accumulation of material power by a state alone; they also consider geographic proximity, offensive military capabilities, and whether the state appears to be a threat. In other words, they balance against the most potent threats rather than against power per se. 37 Threats are, of course, socially constructed (by both material factors and ideational factors). They are mutually constituted understandings that arise out of interactions between states. Britain, France, Germany, and Japan did not balance against American power in the Cold War era but against the Soviet Union. Western states have not balanced against Israeli nuclear power because it is not perceived as a threat to them in the way that nuclear weapons in the hands of Iran and North Korea (and previously Iraq) are. Enduring rivalries is another dyadic-level concept that fits comfortably in the constructivists’ playbook. Enduring rivalries have a structure all of their own that is clearly intersubjective. States generally come to identify each other as rivals, as enemies, and as threats through a lengthy interactive process in which their mutual reactions to each other’s words and deeds create an intersubjective meaning. Soviet and American Cold War identities as rivals developed out of a series of diplomatic games begun as allies in World War II; mutually antagonistic interactions were bound up with each state’s emerging postwar identities, their interests and policies, their rhetoric and discourse. Out of this interaction emerged a set of relatively firm expectations and norms of proper behavior. On the other hand, some states seem to be “born” rivals, with their identities formed even prior to their emergence as sovereign states— for instance, the India-Pakistan rivalry and that of the two Koreas. Another important dyadic concept is the security dilemma. For constructivists, a security dilemma “is a social structure composed of inter-subjective understandings in which states are so distrustful that they make worst-case assumptions about each other’s intentions, and as a result define their interests in self-help terms.” 38 In their view, security dilemmas are the result of uncertainty, since states are unable to understand the intentions of other states with any confidence. However, identities, norms, and social practices reduce uncertainty by providing meaning. 39 Constructivists generally accept the logic of reciprocity; cooperation begets cooperation, and hostility begets hostility. Through negative social interactions, states may become convinced (rightly or wrongly) of another’s malign intentions, thus exacerbating the security

Constructivism

471

dilemma and bring on a conflict spiral. Through positive social interaction, states may become more certain of another state’s benign intentions, thus avoiding the security dilemma. Social interaction also has a role in the democratic peace because the uncertainty about the intentions of some states can be reduced if they are identified as democratic. For realists, the nature of the state’s political organization should not matter; all states should respond in similar ways to the dangers of anarchy. But constructivists argue that the democratic peace occurs because interactions between these states have created norms of compromise and cooperation and the expectation that disputes will be solved peacefully rather than with coercion. They have constructed similar identities and interests. Between states who perceive each other to be liberal democracies a different structure of relationship exists. 40 For the dyadic level of analysis, the bottom line is that war is more likely if interaction between members of a dyad produces antagonistic and exclusive identities, distrust, shared expectations of hostility, and norms of interaction that emphasize individual gains instead of mutual interests. The International System Level of Analysis At the international systems level, constructivists typically focus on global norms, the changes in these norms over time, and the effect of norm change on state behavior. For instance, John Vasquez has seen the existence within the international community of realist norms that emphasize power politics as a permissive factor that makes it easier for states and their leaders to use military force and war as an acceptable tool in their foreign policy toolbox. As Ba and Hoffmann note, throughout much of modern history, the great powers of Europe have interacted in pretty much exactly the way that realists would predict that states would behave under anarchy—as insecure, suspicious, and competitive entities. 41 Gregory Raymond and Charles Kegley have studied the relationship between global norms and international behavior. 42 As Raymond says, “international norms are more than modal regularities; they are intersubjectively shared understandings that oblige national leaders to act in a certain way.” 43 International norms create a web of mutually understood expectations that help to give meaning to the behavior of states. Indeed, Raymond argues that “all independent political entities with regular intercourse have developed conventions defining appropriate behavior for certain situations, including the resort to war.” 44 (The earliest evidence of this dates to the fourteenth century BCE in the Amarna tablets that document treaties binding the kings of Egypt, Hatti, Mittani, Babylonia, and Assyria.) Raymond and Kegley make a macro distinction between international systems (or regional subsystems) having permissive normative orders and those having restrictive normative orders. Permissive normative orders give leaders considerable leeway to take actions to protect their states in spite of the existence of general norms prohibiting violence. Typically, states in these systems have mutually granted each other considerable moral flexibility based on the realpolitik notion of “reason of state”: States (and their leaders) may do things that might be seen as immoral on a personal level in order to promote the higher interests of the state. Leaders may properly argue that given the nature of systemic anarchy, necessity requires them to use force even if this means a breach of international obligations.

472

Chapter 12

On the other hand, restrictive normative orders consist of norms that apply narrower criteria to the circumstances under which one might resort to force (and how to wage war once it has begun). Leaders have less discretion to argue they are compelled by necessity or reason of state to use force in contravention of international norms. In a restrictive normative order states typically agree to place limits on the use of force, to limit their unilateral ability to ignore international agreements, to geographically restrict their competition with each other, and to create mechanisms for the amicable resolution of disputes. Raymond surveys various studies of international norms and concludes that they suggest that the escalation of disputes is less likely when the ability of states to act in a unilateral, opportunistic, and self-interested way is “encapsulated within a cocoon of norms that establish mechanisms for refereeing controversies and upholding agreements.” 45 Historical evidence supports the conclusion that normative support for the use of good offices, mediation, conciliation, and arbitration lower the odds of war for major states. A consensus among great powers concerning the criteria for the use of force—both in war and the use of force short of war— also appears to lower the chances that MIDs will escalate to war. More generally, when states accept the basic legal principle of pacta sunt servanda (treaties must be observed) as a universal norm, serious disputes between major powers decline and fewer disputes escalate to war. On the other hand, when states are given greater latitude to exempt themselves from community norms and where there is a greater willingness of major states to accept force short of war—reprisals, retorsions, and other tit-for-tat measures—the incidence of war is greater. In sum, permissive international orders are generally associated with war. 46 We can speculate that what is true at the international system level almost certainly holds true for regional normative orders and for dyadic, partner-specific normative orders. Martha Finnemore and others have noted the emergence of international norms over the last 150 years that include the abolition of slavery, the end of the slave trade, the end of colonization, the renunciation of apartheid, and the promotion of a broad set of human rights. This has also come to include a new norm supporting external intervention for humanitarian purposes, especially if the intervention is multilateral in form and sanctioned by international authority. This new norm of humanitarian intervention clearly erodes the previous norm of state sovereignty. 47 She argues that these changes have occurred as the result of continual and highly visible (and sometimes violent) contestation over norms—a contest involving individuals, states, and governmental organizations and nongovernmental organizations. Over time some new norms won out and others lost. Changes in norms then create permissive conditions for changes in state behavior. Constructivists focus on the social and cultural meanings that actors attach to certain things and practices—many of which are specifically related to war. Several scholars have noticed significant change in realpolitik norms throughout the twentieth century. For instance, Richard Price has looked at the emergence of the norm of non-use of chemical weapons in the postWWI era, especially against civilians. He argues that the taboo was essential to how states viewed themselves (their identities) as members of a civilized society and what it meant to be a civilized people. 48 Of course, the taboo was most effective with regard to those who were also identified as being members of the same civilized global community. Thus, American use of napalm against the Vietnamese was guided by a different set of normative considerations.

Constructivism

473

The norms for nuclear weapons use have also been significant. As we saw in chapter 9, nuclear weapons have not been used since 1945, despite opportunities in which realpolitikoriented leaders actually considered their use and even when not using them meant losing wars to minor powers. 49 They attribute this to a growing norm that possessors of nuclear weapons not use these devices in actual combat, especially not against non-nuclear opponents. From a constructivist perspective, Tannenwald sees this newly emerged norm as socially constructed and intersubjective, based on the iterated practice of non-use over time, as well as by conscious efforts of norms entrepreneurs, including both domestic and transnational interest groups, to foster an international stigma associated with nuclear weapons use. Paul accepts the constructivist argument that evolving norms have made the use of nuclear weapons almost impossible, but adds that there are other, nonconstructivist reasons as well. Because of their unique nature as absolute weapons with tremendous destructive power, there are severe limits to what they can accomplish. Most importantly for Paul, the use of nuclear weapons comes with costs concerning “reputation and image, maintenance of alliance relationships, allegiance of nonallied state, and non-proliferation.” 50 As Paul explains: First, the material character of the weapon gives rise to reputational concerns through the logic of consequences, that is, in the calculations of decision makers. Reputation concerns are highlighted due to the efforts of norm entrepreneurs or reputational intermediaries, who make arguments based on the logic of consequences and the logic of appropriateness. These reputational concerns lead to self-deterrence on the part of nuclear states. . . Second, when self-deterrence is practiced over a period of time, it becomes a tradition which in turn helps to self-deter in contexts of potential nuclear uses vis-à-vis nonnuclear states. This feedback loop becomes stronger over the passage of time; as more years go by without use, the more times an opportunity for use is not taken, the harder it becomes to cross the threshold. 51

Perhaps even more controversial is John Mueller’s argument, put forward in the late 1980s, that international norms have changed fundamentally regarding the use of conventional war between major powers. 52 While Mueller is not, strictly speaking, a constructivist, his arguments are consistent with the ideational approach. His argument is that war is becoming obsolete among the states of the developed world because in those states attitudes about war are changing. War is no longer seen as a normal part of international relations; it is no longer viewed as required by human nature; it is no longer believed to be necessary for human progress; and it is no longer seen as serving a necessary social-political-economic function. Instead, it is now widely viewed as irrational, immoral, and unacceptable as part of the relations between civilized states. Offensive war is not acceptable. (This is the logic of appropriateness.) This ideational change was propelled, in part, by the enormous destruction of World Wars I and II and the realization that future wars would be so costly as to be irrational. (This is the logic of consequences.) Mueller argues that while most political elites had come to this conclusion shortly after WWI, one important leader, Adolf Hitler, had not—thus largely explaining the anomaly of WWII. Mueller’s primary argument is that war, like any other cultural institution, is subject to change over time. Changes in moral values led eventually to changes in social institutions. While a good many societies at one time supported institutions such as slavery, dueling, and

474

Chapter 12

even human sacrifice, these institutions have been gradually discredited and then completely abandoned (not just replaced by other social institutions). The same is beginning to happen to attitudes about war in the developed world. One can point to the fact that several European states have forgone war for centuries now—the Swiss, the Scandinavians, and the Dutch, for instance. And Europe itself, an area that was once the most war-prone in the entire world, has not experienced interstate war since 1945—unless one counts the multiple secessionist wars within the former Yugoslavia. Nevertheless, it should be emphasized that the trend Mueller discusses is not yet global in scope. (Indeed, the trend may not encompass all of Europe, as the conflict between Slovenia, Croatia, and Bosnia-Herzegovina against the remnants of the Yugoslav state led by Serbia illustrates.) The majority of the world’s states are not members of the developed area (and neither are they overwhelmingly democratic). War is not yet obsolete; it is only in the process of becoming obsolete—a process whose end is uncertain and whose time frame is unknown. Mueller suggests that social scientists, by ignoring the autonomous power of ideas, have left out an important independent variable in their search for the origins of war. He suggests that war is ultimately an idea that has been adopted by large portions of international society, perhaps temporarily, as a method for dealing with conflict. Since it is required neither by human nature nor by the nature of the international system, it may be abolished like any other cultural creation. According to Mueller, war can disappear without requiring that there be any notable change or improvement on any of the level of analysis categories. Specifically, war can die out without changing human nature, without modifying the nature of the state or the nation-state, without changing the international system, without creating an effective world government or system of international law, and without improving the competence or moral capacity of political leaders. It can also go away without . . . enveloping the earth in democracy or prosperity; without devising ingenious agreements to restrict arms or the arms industry; . . . and without doing anything whatsoever about nuclear weapons. 53

The conclusion at this level of analysis is simply that war is more likely to the extent that interactions between states produce systemic norms, expectations, intersubjective understandings, and policy preferences that tend to be based on realpolitik assumptions and prescriptions and which create a permissive normative order regarding the use of force. THE IRAQ WAR: A CONSTRUCTIVIST ANALYSIS Finally, let’s take a quick look at one attempt from the constructivist perspective to explain the decision for one specific interstate war, the U.S. invasion of Iraq in 2003. Peter Howard is interested in answering the following question: Why did the U.S. invade Iraq instead of North Korea? 54 (Notice that he is not asking what caused the U.S. to invade Iraq in 2003.) President George W. Bush’s 2002 State of the Union message identified a set of rogue states belonging to an “axis of evil”—Iran, Iraq, and North Korea. Classical realists would predict that given certain potential threats to the U.S. from rogue states within the axis of evil, the gravest danger would be presented by the country with the largest material capabilities. The Bush administration argued, however, that Iraq represented the largest and most signifi-

Constructivism

475

cant threat to the U.S. Howard does the math and it doesn’t work. Whatever one counts—size of armed forces, naval power, air power, missiles, possession of chemical, biological, and nuclear weapons (WMDs)—the numbers clearly show that North Korea had greater material capabilities and should therefore have been seen as a greater threat. Why didn’t the U.S. therefore choose North Korea as the target of its military invasion? Howard argues that the U.S. decision centered on who was more likely to actually use its WMDs. As Howard says, “Threat perception is necessarily based on ideational and identity issues, not material capabilities, and necessitates a turn away from a materialist explanation to an ideational one.” 55 Moreover, perceived intentions are intersubjective, the result of rules and identities developed through interactions. And the meaning ascribed to material capabilities is more important than the capabilities themselves. This is essentially Walt’s argument that states balance against perceived threats rather than against power per se. Constructivists argue that norms and identities both create and constrain policy options available to states. They help determine what is acceptable (or appropriate). But in this case, focusing on norms doesn’t help understand why the U.S. was willing to talk to North Korea about its development and possession of WMDs while it chose to invade Iraq. Both Iraq and North Korea were members of Bush’s axis of evil, but while the rules for the two rogue states had a “family resemblance,” there was no single doctrine applied to both. Basic rules were set out in the National Security Strategy document and by the identification of the three states as rogue states within the “axis of evil” that defied international norms and threatened U.S. interests. However, these basic rules intersected with other rules in ongoing “security games” being played out through separate dyadic interactions of the U.S. with North Korea, Iran, and Iraq. All games (like poker) have rules; and subgames (five card stud, Texas Hold’em, seven card draw with low hole card wild) have more precise rules within the overall context of the category. As we have seen, prewar, war, and postwar situations can also be seen as bargaining games. The day-to-day games played between the U.S. and the different rogue states intersected with and modified the way the basic axis of evil rules were applied in each state. Howard focuses on the language used by the participants in these games, although by language he means both words and deeds. Essentially, interaction through words and deeds is a way of conveying shared meaning between the actors. With regard to North Korea, the U.S. was entangled in a game of nuclear negotiation that dated back to the 1994 Agreed Framework, and this negotiation involved multiple partners— South Korea, China, Russia, and Japan, in addition to the U.S. and North Korea—who were parties to the Agreed Framework. The U.S. relationship with Iraq was different because the rules of this subgame stemmed from the 1991 Gulf War, the subsequent UN sanctions regime, daily tit-for-tat aerial games centered on the no-fly zones, and “cheat and retreat” verification games regarding Iraq’s possession of WMDs. It was essentially a game of unilateral enforcement rather than one of multilateral negotiation. Engagement through diplomacy was never an option. As constructivists would say, “context is everything and history matters.” At last we have reached the end of our discussion of theories of international conflict. We can now proceed to a final summary.

Chapter Thirteen

Conclusion

Keep fear alive. —Stephen Colbert

THE CAUSES OF WAR: SOME BASICS I hope that in the preceding pages the reader has become acquainted with some of the complexities involved in discovering the causes of war. It is easy to be overwhelmed by the sheer multitude of hypotheses and empirical findings concerning the causes of war. By now you are probably hoping that the author will sum up by sorting among the various theories presented in the preceding pages and give the REAL ANSWER to the question of what causes war. However, such simple and easy answers don’t exist. We have been unable to identify a single theory that alone can account for war as a general phenomenon. Instead, we have discovered several islands of theory that have achieved partial validation: they seem to apply to a good many wars, but not all wars; or they are useful in explaining wars between great powers, but not between states of lesser power; or they pertain to certain periods of time, but not all; or they apply under certain conditions, but not under all. Although no single theory seems to have anywhere near universal validity, social scientific research has not been completely fruitless. Some theories have been identified as lacking a factual basis and have been relegated to the status of mythology. They have expired due to lack of confirmation and have been abandoned by scholars. In this category we may place explanations built around humans’ instinctual aggressiveness, war weariness, business cycles, the nature of governments, population growth, lateral pressure, and the polarity of the international system. However, several theories, having been validated often enough by real-world events and systematic research, have emerged as more useful and warrant closer attention. In this category we place theories that emphasize changes in dyadic or systemic balances of capabilities, theories that focus on the effects of various kinds of political unrest and the political vulnerability of political elites, theories that address state-to-nation imbalances, and the steps-to-war theory that incorporates the critical factors of territorial disputes, dyadic rivalries, and conflict spirals. Although there is as yet no psychological theory of war, we find cognitive, personal477

478

Chapter 13

ity, and emotional factors important to the understanding of war, even though their exact effect will differ greatly from one situation to another. While game theory, bargaining theory, and constructivism may not help us to directly pinpoint a specific cause of war, they may on occasion help us (in a heuristic way) to understand things where other theories fail to enlighten us. Certain patterns and tendencies have been identified as likely precursors to war; other reputed patterns have turned out to be dead ends. Some processes thought to be helpful in promoting peace—like balancing through arms accumulations, the creation of alliances, and making deterrent threats—have been discovered to be two-edged swords. They are just as likely to increase the probability of conflict as they are to deter war. Thus, a winnowing-out process is occurring in the scholarship of war studies. Even though the investigations undertaken by social scientists during the last five decades have not culminated in the creation of a single, unified theory of war, they have certainly added greatly to our understanding of the causes of war and have filled in some dark holes in the cosmic puzzle of war. And by so doing they have added as well to our understanding of how peace may be maintained. Here are a few things we think that we know: (1) A single “master theory” capable of explaining interstate war does not at this point exist. (2) Scholarly efforts have shown that there is no single causal factor that is overwhelmingly responsible for interstate war. As Bennett and Stam have shown, the vast majority of statistical correlations between causal variables and interstate war are fairly weak. 1 (3) Most wars are probably due to multiple factors that interact with each other. Interstate war is a complex phenomenon and is the result of a multicausal process. (This is probably one of the biggest messages of this book.) (4) Moreover, important causal factors are found at different levels of analysis. Not only is there a winnowing-out process going on, but there is also a process of cross-level integration going on as researchers find important connections between levels of analysis. (5) The factors that cause war will vary from case to case, and no two cases of war are likely to result from exactly the same combination of forces. (6) It must be admitted that whether wars actually occur is subject to a certain amount of randomness, chance, and serendipity. There will always be some unpredictability involved in the outbreak of war (and much of this uncertainty may have to do with idiosyncratic individual level variables). (7) It may be possible, nevertheless, to identify certain repeated dangerous patterns of interaction between causal variables; these patterns may take the shape of causal chains that sequentially link causal factors together, or they may simply consist of a particular constellation of factors that occur together and reinforce each other. However, the construction of a single, integrated, cross-level theory is still in its rudimentary stages. Interstate war is a relatively rare event, and many states never experience it. So it probably requires the presence of a rather special combination of conditions. There is, in fact, a consensus on the existence of certain empirical regularities that can act as risk factors. Their presence increases the risk of war (and the more risk factors, the greater the likelihood of war), and their absence helps diminish that risk. Observing these risk factors can act as a kind of early warning system to detect the likelihood of war. Here are a few empirical regularities or patterns social scientists agree on: (1) When interstate war happens, it almost always involves

Conclusion

479

contiguous neighbors. If neighbors are not involved, then one of the belligerents is very likely to be a major power. (2) The most likely underlying cause of fighting between contiguous states is a dispute over territory, though not all territorial disputes lead to war (and states also fight over other issues). (3) Large power disparities between states seem to promote peace rather than war. Therefore, it is quite likely that most interstate wars (though not all) involve states who are relatively equal in capabilities. (4) A disproportionately large percentage of wars involve enduring rivals or strategic rivals. These states share an extended mutual history of hostile interactions that probably include participation in serial crises and/or militarized disputes with each other, and perhaps even a history of previous wars. (5) Historically, large, powerful states have been more likely to be involved in war than small, less powerful states. (6) Mature democracies are highly unlikely to ever fight each other; therefore, dyadic pairs of mature democratic states are excluded from the pool of states likely to be involved in war. It logically follows therefore that virtually all wars will involve dyads in which at least one state is nondemocratic. (7) Most wars are preceded by militarized disputes or crises that involve escalatory behavior preceding the outbreak of war that looks like a conflict spiral—though the temporal length of the spiral may vary considerably. While the existence of these patterns themselves are beyond doubt, several commentators have noted that the theories that attempt to explain how these patterns lead to war are often inadequately specified. The facts are robust, but the theoretical explanations are still weak. John Vasquez argues that instead of “islands of theories” it would be more accurate to call them “islands of findings.” 2 The combined effect of the previous regularities is that any theory of war will, in reality, apply almost entirely to a relatively small group of dyads consisting of “politically relevant actors” who are probably rivals with relatively equal capabilities who are contiguous, and at least one state in the dyad is a nondemocratic state. Perhaps the best that can be done at the moment is to take these recurrent patterns as a baseline and illustrate how factors at the various levels of analysis might operate together in an interactive and synergistic way to increase the probability of war. One might, in fact, build a hypothetical case that represents the coming together of so many factors that war would be virtually inevitable—a perfect storm, so to speak. An alternative approach would be to recognize that different types of wars have different causes; distinct patterns made up of different combinations of factors may combine to produce different kinds of wars. We will try our hand at both approaches. THE PERFECT STORM Let us place our initial focus at the dyadic level of analysis because this is perhaps where the most fruitful results have been produced. We’ll start by loading up our initial dyadic scenario with a number of high-risk factors. We will begin with the occurrence of an international crisis between two contiguous states; the two states are relatively equal in power; one of the states is nondemocratic; and the dyad is classified as an enduring rivalry. Contiguity creates a situation of opportunity and willingness for the use of force for both states; it also creates the potential for territorial disputes. Relative equality creates a military context in which each side can believe it can prevail in a potential war, and attempts at deterrence will therefore be less

480

Chapter 13

effective. The lack of joint democracy deprives the pair of the institutional and cultural similarities that reduce conflict and create the democratic peace. Rivalry creates historical memories based on past interactions that predispose leaders to distrust the rival, to perceive threats, to engage in worst-case analysis, to believe that the rival will respond inversely to cooperation and conciliation, and to prefer the use of coercive tactics to attain its goals. In the hypothesized scenario, the leaders of both states perceive that the situation represents a serious threat to their states’ vital interests and that the use of force by their opponent is within the realm of possibility in the immediate future. Wars normally develop as a climax to a set of hostile and threatening interactions that escalate as the dispute becomes militarized. But what lights the spark in the first place? There are multiple contenders, but perhaps the most likely is a flare-up of an unresolved dyadic territorial dispute. We know that this type of issue is the most dangerous problem for contiguous states. These issues are important both for the tangible and strategic attributes of the territory but, even more importantly for their intangible and symbolic aspects. The latter relates especially to issues of national identity, and territorial issues that are related to state-tonation imbalances are particularly dangerous. Territorial issues inherently involve “red meat” political themes of nationalism, national identity, national security, and territorial integrity. Given the nature of these issues, political leaders are unlikely to retreat on them or make concessions. Reputation is also at stake: if one doesn’t fight for its own territory, what will it fight for? Thus, territorial issues have the potential to unleash a runaway conflict spiral that escalates to war. Other potential sparks can also be identified. In general, attempts by states to enhance their security may create security dilemmas and conflict spirals, whether they affect the real balance of capabilities or not. Some sparks may be related to the perception of impending dyadic or systemic shifts in capabilities—power shifts, critical points in a state’s relative power cycle, systemic deconcentration of power. All of these—especially in combination—are important for great power rivals at the top of the global system. A related phenomenon having to do with a potential change in the military balance may also create a crisis trigger—the perceived attainment of a qualitative military threshold by a rival. Finally, at the state level, political instability, democratic transitions, internal discord, civil wars, revolutions, and the change of political regimes may trigger international crises. In all of these situations, the spark may be lit by defensive-minded states simply seeking security for themselves or by predatory states seeking to maximize their gains. Each is possible, and there is no empirical data available for use to establish an educated guess about which is more likely. Once the spark is lit, a great deal depends upon the interaction between states as they attempt to deal with the situation provided by the trigger. Most wars do not appear out of the blue without any warning signals; they are preceded by militarized disputes and crises that tend to spiral upward with actions and counter-actions marked by increasing levels of threat and hostility. The escalation can be symmetric or asymmetric. An important issue for conflict theory is which disputes are likely to escalate to war and how (and why) this happens. Considerable evidence indicates that the use of realpolitik tools—such as bullying and the use of threats, ultimatums, brinkmanship tactics, coercive actions, and demonstrations of

Conclusion

481

force—which are meant to demonstrate toughness and to deter opponents are likely to lead to hostile reciprocal actions rather than to the desired result of backing down. As the steps-to-war theory indicates, various phenomena affect the probability of escalation to war, both independently and cumulatively—repeated crises, militarized disputes (especially over territory), the creation of rivalries, arms buildups, and alliance construction. Each action that is taken increases the probability the conflict will escalate. The more of these phenomena that occur, the greater is the likelihood of war. Dyadic interaction processes are certainly abetted by factors at the individual, small-group, nation-state, and international system levels. At each level of analysis, several factors may contribute to the likelihood of war or may retard the march toward violence. Phenomena at the individual level of analysis have been generally believed to be of limited importance in explaining war. In part, this is because it is difficult to incorporate individual level variables into general theories of war in a parsimonious way. They are also not easily amenable to large-N data studies. On the other hand, case studies often cite their importance, and most scholars accept the proposition that at least in some wars, the crucial determinant of whether a crisis escalates to war or not will be idiosyncratic attributes of particular leaders. Thus, individual level factors may be especially important short-term or immediate causes of war, at least in some instances. They probably deserve a larger place in our theoretical thinking about the causes of wars. Individual-level factors are a kind of wild card in the escalation of conflict to war. They may be important, but in ways that are not too easy to predict in advance. Nevertheless, certain tendencies are problematic. For reasons having to do with the emotional and psychological makeup of leaders, with their operational codes, and their reliance on “lessons of history,” realpolitik tactics are not likely to be effective—especially between equals. Threats are likely to be met with counter-threats, defiance with intransigence, and demonstrations of resolve with counterdemonstrations. The degree of hostility escalates and energizes the conflict spiral. Security crises may begin as prisoners’ dilemma games, but they have a tendency to become transformed into games of chicken in which leaders on both sides come to believe two things: that backing down is (for themselves) unacceptable but that their opponents will relent when confronted with clear evidence of commitment. Within this context, attempts at deterrence have a limited possibility of success. In fact, even superior military capabilities and formal commitments may not suffice to deter violent conflict. Threats and coercive actions intended to prevent the violent actions of others in fact serve to confirm the worst-case suspicions of one’s opponents. The perceived threat to their own interests often impels them to initiate or continue belligerent activities of their own. Whether or not an original threat in fact existed, the result is the same: deterrence fails. As the field of international relations has expanded over the years to take in larger chunks of intellectual territory belonging to more and more branches of social science, we have become increasingly aware of the effects of bounded rationality. We now understand much more about the pervasive impact of cognitive biases in information processing, heuristic short cuts, and the influence of the emotional side of our reasoning on decision-making. Explanations of war that omit these things may miss something important.

482

Chapter 13

Actions that take place at every level of analysis are filtered through individual leaders’ perceptual screens and are shaped and interpreted by their images and worldviews. The individual’s response to these actions is given direction through his or her current images and operational code. The dynamics of conflict spirals will be affected by leaders’ perceptions—of their own internal political vulnerability, of the hostility of interactions with others, of the intentions of others, and the meaning of their actions. Other perceptions are also important. Perceptions of impending power transitions, the lack of justice in the systemic distribution of status, the construction of alliances, the relative balance of offensive and defensive forces, and the general balance of power within the international system are all important to the extent that they are perceived as important by political elites with the authority to make decisions about war and peace. While phenomena at the dyadic, state, and international system level may have independent effects on the behavior of states, perhaps the best way to conceptualize the role of phenomena at the systemic, dyadic, and nation-state levels of analysis is as stimuli that trigger individual perceptions (and misperceptions), which then guide policy decisions about war and peace. Misperceptions about the opponent’s actions, his intentions and capabilities—and hence about the degree of threat to one’s own security—may create the conditions necessary for the onset of a crisis. Once the crisis begins, these misperceptions may accelerate and exacerbate the level of tension. Particularly important are the combination of an over-perception of the rival’s hostility, treachery, and threat, coupled with an under-perception of both the capabilities of the rival and the amount of risk involved. Unwarranted confidence in the ability to compel one’s opponents to back down short of war, or of one’s ability to defeat the adversary with little cost if war does begin, seem to be an important part of the picture. Another significant individual level factor is the framing of the problem. To the extent that one side frames the problem as being within the realm of potential losses (or, more dangerously, within the realm of certain losses), the propensity to engage in risky escalatory behavior increases. The situation is even more dangerous if both sides frame the problem within the realm of losses. While images and perceptions play a decisive role in determining the extent to which leaders believe the interests of their states are being threatened, operational codes play a crucial role in determining how these leaders will respond to the perceived threats. Leaders with realpolitik operational codes, which are based on the belief that bullying and threats will work and which emphasize the use of aggressive tactics, are probably most likely to find themselves trapped in conflict spirals from which they cannot extricate themselves short of war. Individual personality factors may also play a role here, affecting the ability of national leaders to realistically appraise and react to the international situation. Since many national leaders probably possess psychological characteristics such as power orientation and high dominance, and employ various ego defense mechanisms to protect against feelings of low self-esteem, the likelihood that they will be psychologically predisposed to back down in the face of threats from external opponents is probably fairly low. If you add a high level of risk acceptance to the aforementioned traits, the personal fuel mixture becomes highly volatile.

Conclusion

483

Finally, psychological stress brought on by the crisis may have a deleterious effect on the ability of decision makers to make rational calculations. Ultimately, decisions for war are likely to be predicated on the assumption that war is inevitable, or that it can be waged successfully, or, at the very least, that it can be waged at an acceptable level of cost. These individual perceptions (or misperceptions) can be reinforced by factors at the small-group level. Like individual-level factors, substate phenomena have been given little prominence as factors leading to war. And like individual-level factors, they are difficult to incorporate in large-N studies of international conflict. Moreover, substate factors interact with individuallevel factors in complex ways that are not well understood. While case studies point to their importance in some circumstances, they are not likely to hold an important place in any general theory of interstate war. Nevertheless, they may enhance our understandings of what causes war at the margins. At the small-group level, decision makers may ignore warnings of impending policy failures and military disasters through the process of groupthink. Small-group dynamics may be coupled with individual cognitive processes that prevent the decision makers from re-examining incorrect assumptions and from seriously considering points of view at variance with their own. National security organizations suffer collectively from many of the same information processing errors as individuals; for instance, they use procedural shortcuts or heuristics. They also tend to rely on relatively rigid standard operating procedures, which may play a role in determining which options are presented to policy makers, as well as determining how policies, once decided upon, are actually carried out in the field. On the other hand, the dynamic of bureaucratic politics may be predominant. Crisis escalation is more likely if the decision process is dominated by political elites whose politicalbureaucratic-economic interests will be maintained or enhanced by a decision to go to war. One should be especially mindful of the effects of domestic political pressures on national leaders and of the tendency of hardline factions to increase their power in times of crisis and confrontation. The fact that hardliners dominate the decision-making machinery of the government means, almost by definition, that realpolitik tactics are more likely to be employed in the crisis. 3 This is perhaps one of the most important effects of domestic politics on the problem of war—though its effects are not well documented at this point. Efforts to defuse conflict spirals are difficult because of one’s own previous acts and pronouncements and because of the domestic implications of compromising or reversing previous hardline policies. Elite fears that domestic opponents will be able to brand compromise and conciliation as appeasement may stifle creative conflict resolution. That these dynamics occur in one country is bad enough, but the internal politics of both rivals are likely to be similarly affected, thus creating reciprocal, reinforcing effects. In this environment, threats are likely to lead to counter-threats and demonstrations of force to counter-demonstrations. Rivalry interacts with domestic politics in malignant ways. The likelihood of a crisis escalating to war is probably enhanced in cases of rivalries—especially if the present crisis has been preceded by other crises with the same opponent. If a regime in one (or both) of the states has recently lost a previous crisis encounter with its rival, the need to prevent a second (or third) loss of face makes necessary a demonstration of resolve and makes a policy of compro-

484

Chapter 13

mise and conciliation difficult to achieve. Internal political considerations may outweigh international considerations. Domestic politics tend to reinforce rivalry dynamics. It impedes the ability of rival states to cooperate and to find overlapping win-sets. This is especially true with regard to territorial issues. The domestic insecurity or vulnerability of elites—especially if combined with economic problems—is probably a factor that promotes militant postures and the taking of risks and leads to the creation of crises and their escalation to war. The presence of a rivalry automatically creates potential demons or scapegoats who can be blamed for internal woes. Scapegoat wars, while not overly plentiful, stem directly from the stuff of domestic politics. At the nation-state level, the propensity for the more powerful states to become disproportionately involved in war is a worrisome factor. The connection between large states and war is related to factors at the individual and substate levels. At the individual level, leaders of major powers are most likely to have national role conceptions that define their state’s roles as protectors, defenders, interveners, and activists in the global system who carry the burden of providing world order. (They have identities that predispose them toward the use of force.) At the substate level, the more powerful states also have well-developed national security institutions (military-industrial complexes) whose leaders or supporters are likely to be well represented in the governing coalition. Some characteristics of states, such as their internal political instability and their ethnoreligious composition (a nation-to-state imbalance), appear important in some instances. Internal instability is likely to create perceptions of elite vulnerability, perceptions of windows of vulnerability and opportunity, and serve as a motivation for diversionary wars or kick-’emwhile-they’re-down wars. While these patterns are not prevalent enough to be considered a primary cause of war, they are present frequently enough to be identified as a crucial factor in some wars. Such factors might usefully be treated as risk factors that might increase a state’s chances of being involved in war. The nature of a state’s political system (or its economic system) tends not to be a fundamental cause of war. Whether any particular crises will escalate to war probably depends not so much on the nature of a particular state’s political or economic system as on the degree of political-economic difference between a pair of states. At least in modern times, democratic systems have refrained from mutual war, while wars between states with different systems have been commonplace. While sharing similar political systems tends to discourage war, sharing borders with states governed by different principles tends to make it easier to attribute negative traits to the other. These state-level factors—size and power, political instability, ethno-religious composition, and the nature of the political system—are probably more important in the initial creation of conflict than in determining whether conflicts escalate to war. We will come back to many of these variables when we deal with paths to war. Although a great deal of the early work on causes of war dealt with the nature of the international system, very little has come out of it. The particular configuration of power within the system—unipolar, bipolar, tripolar, or multipolar—may not matter. Interstate wars occur in all types of systems. Alliance creation and aggregation have been related to war in

Conclusion

485

many time periods but not all. (Alliances are generally a step to war, but not during the Cold War era.) Changes in the structure of the international (or regional) system are probably much more dangerous than any static feature of the system. Whether the structural transition is brought about by industrial-technological-demographic growth, by the conscious accumulation of military capabilities by a challenging state, or by the rearrangement of military alliances, the fortunes of some states are being diminished while the fortunes of others are being enhanced. The results are many: role transformations, status disequilibria, changes in the perceived degree of security, and increased systemic uncertainty. It is at these times that the degree of insecurity and threat in the international system is highest. Responses to these structural threats are likely to take on the characteristics of conflict spirals. Serious great-power crises have in the past been most likely to occur during periods of transition in the international system (or in regional subsystems) where there are significant shifts in the balance of capabilities, especially between the dominant power in the system and its major rival(s)—but also between any set of rivals. (Dyadic and systemic power shifts are probably related.) But even here there is a great deal of indeterminacy. Systemic and dyadic power shifts don’t always lead to war. Other factors—rivalries, contiguity, arms races—play a role in determining which transitions lead to war and which do not. And theorists are unsure about who the likely initiator will be. Power transition issues ultimately involve issues of prestige and status. Thus, power transitions can set off conflict over the proper distribution of prestige and status within the system as well as conflict over the proper distribution of political, military, and economic power. It would seem especially dangerous if several system-level and dyad-level phenomena were occurring simultaneously—as is likely. For instance, the occurrence of power shifts, critical points in the cycle of relative power, hegemonic decline, systemic deconcentration of power, regional concentration of power, and alliance polarization—or just a few of these— would be much more dangerous than the existence of any single factor. Any of these structural changes at the international systems level may trigger perceptions of threat by national leaders at the individual level. Systemic and dyadic factors are closely linked to individual and perceptual factors. Perceptions of coming changes in dyadic or systemic balances create the perception of windows of vulnerability and opportunity as well as preventive and/or preemptive motivations for war. Rivals are especially sensitive to power shifts. We should also add that systemic variables appear to be more helpful in explaining world wars or general wars rather than simple dyadic confrontations. They may also be less relevant in the post-1945 era. One of the themes that runs through several theories at many levels of analysis is the connection between fear and war. While some individual leaders may be motivated by gains, fear is a primary consideration. At the individuals level humans are soft-wired to fear deeply those who are different. We also fear losses more than we desire gains, as prospect theory tells us. At the state level, fear of political vulnerability creates incentives for violence. At the dyadic level, the rivalry approach, the security dilemma, spiral theory, and power transition theory all deal with fear of the other state’s capabilities and intentions. At the systemic level, changes in the distribution of capabilities, status, and roles create fears of the future. Factors

486

Chapter 13

that promote fear promote war. (While cognitive theorists would emphasize the possibility that fear may be the product of misperception, constructivists would rightly remind us that fear is an individual or group response to the perception of threat, and threats are socially constructed.) DIFFERENT PATHS TO WAR? Another way to look at all this is to propose that there are different types of wars that have different causes. 4 This helps to explain our inability to discover a single set of causes for interstate wars or a single general theory that explains war. As Levy and Thompson point out, the possibility that the same outcome (interstate war) can be the result of different causal paths is known as equifinality. There is no agreement among analysts about how many paths to war there are and what defines them. This actually gets tricky theoretically, because the more narrowly we define what a pathway is, the more pathways we will find; and the more pathways we find, the less general our explanations of war become, and the more they are tailored to individual wars. 5 At any rate, let’s try to designate some fairly obvious discrete pathways to war. Our Path 1 will be Vasquez and Senese’s steps-to-war path. This pathway explains wars between equals who typically engage in disputes about territory, which leads to rivalry relationships, repeated MIDs or crises, and perhaps (but not necessarily) to arms races and alliance formation. This theoretical pathway is essentially a dyadic path that is consistent with conflict spiral theory, the security dilemma, dysfunctional crisis learning, and the dangerous impact of rivalries. Dyadic escalation is central to this pathway, and its virulence may depend on several factors, such as individual-level misperceptions and cognitive dysfunctions. But the steps-towar theory specifically identifies a malicious feedback loop in which state A’s actions have a malignant impact on state B’s domestic political interest groups, on public opinion, and on its political leaders. The escalation exacerbates the mutual understanding of threat and produces greater levels of fear, anxiety, and hostility. Escalation also makes it more difficult for political leaders to seek accommodative solutions and it strengthens hardliners in the internal balance of political power within both states. Path 1 is deemed by many to be the primary path to war among major powers; other paths are seen as less frequent. Path 2 is a domestic instability path. It originates with the nature of the state, specifically with states undergoing domestic unrest—from mere riots and turmoil, to all-out civil war or revolution. It actually consists of two subpathways. Path 2(a) is a diversionary path to war. It starts with a political leadership that is vulnerable to domestic opponents and that seeks to initiate a militarized dispute or a war with a contiguous rival in hopes of rallying support for his regime and preserving his hold on power at home. Leaders tend to frame the issue in terms of a tradeoff between accepting near certain losses at home versus highly risky external adventures that might reverse these losses. Path 2(b) is the kick-’em-while-they’re-down road to war. In this path, a state’s vulnerability—due to political unrest or revolution—creates a window of opportunity that entices a rival state to attack in order to settle old scores. This subpathway may attain part of its logic through the fact that internal unrest in the first state

Conclusion

487

changes the balance of power between the rivals such that the once-dominant state in the dyad is reduced to parity or less. Path 3 is a hegemonic struggle path. This path is described by historical-cyclical theorists—Organski, Kugler, Gilpin, Copeland, Modelski, Thompson, Rasler—and suggests an entirely different road to war, though it is generally a highway confined to great powers and it leads to world war. Here patterns of unequal growth throughout the international system lead to power transitions (or rapid approaches) that lead to wars of prevention by the declining hegemonic state or to wars of opportunity by the rising challenger. These theories draw on both the logic of dyadic power transitions and changes in the overall balance of systemic capabilities, the willingness of states to take large risks to prevent perceived losses, the inability of major states to adjust to changes in power, role, and status in the system, and the logic of status discrepancy theory. Misperceptions of dyadic and system balances, of windows of opportunity and vulnerability, and of the offense-defense balance may be involved. The conflict may begin as a regional contest between land powers and then escalate as global sea powers become engaged. Rising challengers are motivated by potential gains, while declining dominant states are motivated by the desire to avert potential losses by engaging in preventive war. While this path may explain great power wars of the nineteenth and twentieth centuries, it may describe a category of wars that may no longer be relevant. Path 4 applies to wars of inequality. This probably comes in two versions. Path 4(a) results in wars initiated by the strong against the weak. These are most likely wars of opportunity or choice rather than wars of necessity. The logic of conflict spirals, runaway escalation, security dilemmas, and dyadic balances probably do not apply. Explaining these wars most likely requires resort to individual-level analyses that draw on expected utility theory and on misinterpretations of the environment caused by personality variables, cognitive filters, and emotional processes. Operational codes and worldviews are probably important as guides to how leaders interpret the situation and also as guides to the proper responses to the actions of others. Path 4(a) probably also subsumes the “great man” theory of war, in which the primary factor in war is the personal ambition of leaders, driven in large part by subconscious personality variables. In the other variant, Path 4(b) wars are initiated by the weak against the strong. Once again, war depends heavily on individual computations of expected utility, misperceptions of power balances, and windows of opportunities. Domestic political considerations are also likely to be important for this path, especially if the domestic political costs of inaction are greater than the costs of going to war. To the extent that leaders frame the problem as being within the realm of losses, they are more likely to opt for risky military ventures against stronger states. Path 5 is connected to nation-state imbalances, the desire of some states to carry out expansionist policies toward their ethno-linguistic-religious diasporas, the treatment of minorities and the desire of co-ethnics in neighboring states to rescue them, or the desire of ethnic communities to break away from their former states and the desires of others to assist them in this endeavor. The root causes of these wars stem from issues of identity, the fear of others, and the social construction of “us” and “them.” In this pathway, internal wars (especially wars of secession) are linked to external wars and interventions.

488

Chapter 13

THE CAUSES OF PEACE: IS WAR BECOMING OBSOLETE? It’s always good to try to end such a lengthy endeavor with some good news. At the very least, we should entertain a brief discussion of the thesis that interstate war is now becoming obsolete—at least that major wars among countries of the developed world are becoming obsolete. Virtually all commentators who look at the phenomenon conclude that the empirical evidence is fairly clear: Interstate wars are declining. Not only has the proportion of states involved in interstate wars declined, but wars between major powers has decreased substantially, with a “long peace” prevailing since the end of the Korean War, in which no major states have directly fought each other. The majority of armed conflicts today are intrastate rather than interstate. 6 How might we explain this? Our discussion of the causes of wars should give us a few clues. Several interrelated factors seem relevant here: resolution of outstanding territorial claims, the spread of democracy, the spread of peaceful norms, and the spread of global interdependence. 7 First, one of the most widely reported global trends of the last few decades has been the replacement of autocratic governments with democratic governments. If we believe, as most theorists do, that the likelihood that democratic states will fight each other is virtually nil, and if we couple that law with the recognition that the number of democratic states in the world is continually increasing (while the number of autocratic states is decreasing), then we come to the conclusion that war is slowly and surely becoming obsolete. Once we reach the point at which true democracies make up 50 percent of the global population of states, from that point forward, every time a democracy replaces an autocracy, not only is the zone of joint democratic peace enlarged, but the number of conflict-prone mixed dyads is reduced, which should reduce the potential for conflict in the system as a whole. It should be noted that joint democracy is probably a sufficient condition for peace. If any two states are mature democracies, nothing else is required to create the condition of dyadic peace between them. In this sense, explaining dyadic peace does not require a multicausal explanation. Second, it is certainly beyond dispute that territorial issues are the single most dangerous issues over which states might face off against each other. Over the past several centuries, a sizable proportion of cases of interstate war involved disputes over territorial claims. It is also true that once territorial issues are settled to the satisfaction of the states involved, these issues tend to remain settled. Moreover, we have seen evidence that the settling of territorial issues helps to keep states from becoming involved in militarized disputes over nonterritorial issues. This suggests that over time, as boundaries of state become settled and gain legitimacy, interstate wars should decline. Statistically speaking, in the absence of serious territorial claims, states (whether contiguous or not) have a very low probability of being involved in a dispute that will escalate to war. Stable borders may even help promote joint democracy among contiguous states. The good news is that since the days of decolonization in the late 1950s and early 1960s, a good many territorial disputes have been ended. While new ones occasionally erupt, the trend is in a positive direction. 8 Third, John Mueller and others—some constructivists, some not—focus on the spread of international norms, at least among advanced industrialized states. As we have seen, Mueller believes that the connection between the spread of democracy and peace is spurious. He

Conclusion

489

contends the primary causal factor is simply the geographic spread of the idea that offensive interstate war is unacceptable. That idea has diffused throughout the globe in a manner similar to the diffusion of democracy: both started in the developed world. Thus, the countries in which “moral progress” has occurred with regard to war have also experienced democratic political progress. The two are associated geographically but not causally. The spread of global norms undermining the legitimacy of the use of interstate war is part of a more general process in normative change at the global level. The change arguably includes taboos on the use of nuclear weapons and chemical weapons (and probably biological weapons), the proposition that states may legitimately intervene inside other states to prevent humanitarian atrocities, and the notion that state officials may be subject to prosecution in international courts (and the courts of other states) for gross violations of international norms, like genocide and crimes against humanity. Nevertheless, it should be emphasized that the trend Mueller discusses is not yet global in scope. Indeed, the trend may not encompass all of Europe, as the recent conflict between Slovenia, Croatia, Bosnia-Herzegovina, and Kosovo against the Serbian remnants of the Yugoslav state illustrates. War is not yet obsolete, it is only in the process of becoming obsolete—a process whose end is uncertain and whose time frame is unknown. Fourth, following the lead of Emmanuel Kant, modern liberal theorists like Bruce Russett and John Oneal find that international organization links and trade links have an impact on reducing the chances of war that is independent from the positive effects of democracy. 9 In Europe, of course, the creation of the European Union and its common market has enshrined these principles. Membership in the EU requires new applicants to be politically democratic and to have resolved all border disputes with their neighbors. Many commentators view the European region as an important test case. It was once an extremely war-prone region, and the home region of the two global wars of the twentieth century. What has made it possible to change its history so dramatically in the period since 1945? The combined effects of legitimately accepted territorial boundaries, the spread of democracy, international organization membership and trading links, along with the spread of peaceful norms goes a long way toward explaining the “long peace”—at least among the states of Europe. Optimists see this as a trend that will spread globally.

Notes

1. EMPIRICAL THEORY AND THE CAUSES OF WAR 1. St. Louis Post Dispatch, April 12, 1985, p. 2; quoted in Ronald J. Glossop, Confronting War: An Examination of Humanity’s Most Pressing Problem, 2nd ed. (Jefferson, NC: McFarlane, 1987), 3. 2. J. David Singer and Melvin Small, The Wages of War, 1816–1965: A Statistical Handbook (New York: Wiley, 1972). 3. J. David Singer, “Introduction,” in Explaining War: Selected Papers from the Correlates of War Project, J. D. Singer and Associates (Beverly Hills, CA: Sage, 1979), 11–20. See also David Dessler, “Beyond Correlations: Toward a Causal Theory of War,” International Studies Quarterly 35, no. 3 (September 1991): 337–55. 4. Glenn H. Snyder and Paul Diesing, Conflict among Nations: Bargaining, Decision-Making, and System Structure in International Crises (Princeton, NJ: Princeton University Press, 1977), 21–22.

2. THE INDIVIDUAL LEVEL OF ANALYSIS, PART I 1. My thanks to Professor James Rosenau for this observation. 2. William James, “The Moral Equivalent of War,” in War: Studies from Psychology, Sociology, Anthropology, rev. ed., ed. Leon Bramson and George W. Goethals (New York: Basic Books, 1968). 3. Sigmund Freud, “Why War?” in International War: An Anthology, ed. Melvin Small and J. David Singer (Homewood, IL: Dorsey Press, 1985), 158–63. Although Freud himself did not use the term “thanatos,” others have used it as an appellation for the death instinct. 4. Konrad Lorenz, On Aggression (New York: Bantam Books, 1966). 5. Robert Ardrey, African Genesis (New York: Atheneum, 1961); The Territorial Imperative (New York: Atheneum, 1966); and The Social Contract (New York: Atheneum, 1970). 6. Lionel Tiger and Robin Fox, The Imperial Animal (New York: Holt, Rinehart & Winston, 1971), 21, 22. 7. See, for instance, Raymond Dart, “The Predatory Transition from Ape to Man,” International Anthropological and Linguistic Review 1 (1953): 207–8. 8. C. K. Brain, The Hunter of the Hunted? An Introduction to African Cave Taphonomy (Oxford: British Archaeological Reports, 2005). See also the earlier commentary by Richard Leakey, The Making of Mankind (New York: Dutton, 1981), 221–25, who argues that the damage was due to the compression of bones and other debris over a lengthy period of time. 9. Instincts and drives are somewhat different phenomena. A drive (hunger, sex) generates a chemical or hormonal state within an organism that motivates the subject to engage in certain types of action. Once the action is taken, the internal state is satisfied, until the chemical or hormonal conditions recur. An instinct requires an external

491

492

Notes

stimulus to give rise to a behavior. See John Vasquez, “Why Do Neighbors Fight? Proximity, Interaction, or Territoriality,” Journal of Peace Research 32, no. 3 (1995): 277–93, 282. 10. Edward O. Wilson, Sociobiology: The Abridged Edition (Cambridge, MA: The Belknap Press of Harvard University Press, 1980), 120–21. 11. Lorenz, On Aggression, 42. 12. See Samuel S. Kim, “The Lorenzian Theory of Aggression and Peace Research: A Critique.” in The War System, ed. Richard Falk and S. S. Kim (Boulder, CO: Westview, 1980), 84. 13. Lorenz, On Aggression, 52–53; and see Kim, “The Lorenzian Theory,” 85. 14. Anthony Storr, “Aggression Is an Instinct,” in Are Humans Aggressive by Nature?, ed. D. Bender and B. Leone (St. Paul, MN: Greenhaven Press, 1983), 16–21. 15. J. P. Scott, “That Old-Time Aggression,” in Man and Aggression, ed. Ashley Montague (New York: Oxford University Press, 1968). 16. Luigi Valzelli, Psychobiology of Aggression and Violence (New York: Raven Press, 1981), 81, cited in John A. Vasquez, The War Puzzle (Cambridge: Cambridge University Press, 1983), 139. 17. Kim, “The Lorenzian Theory,” 97. 18. James A. Schellenberg, The Science of Conflict (New York: Oxford University Press, 1982), 30, and Scott, “That Old-Time Aggression,” 53. 19. Kim, “The Lorenzian Theory,” 88. 20. Ashley Montagu, “The New Litany of ‘Innate Depravity’ or Original Sin Revisited,” in Man and Aggression, ed. Ashley Montagu (New York: Oxford University Press, 1968), 9–11. 21. See, for instance, David Pilbean, “The Fashionable View of Man as a Naked Ape is: 1. An Insult to Apes, 2. Simplistic, 3. Male-Oriented, 4. Rubbish,” New York Times Magazine, September 3, 1972, 10, cited in Ronald Glossop, Confronting War (Jefferson, NC: McFarlane, 1987), 45. Most of these arguments were put forward prior to Goodall’s observations about chimp behavior. Chimps are now seen as quite territorial. On the territoriality of primates, see Jane Goodall, “Life and Death at Gombe,” National Geographic 155, no. 5 (May 1979): 499. See the discussion below in this chapter. 22. Scott, “That Old-Time Aggression,” 56. 23. See Robert Carneiro, “War and Peace: Alternating Realities in Human History,” in Studying War: Anthropological Perspectives, ed. S. P. Reyna and R. E. Downs (Langhorne, PA: Gordon and Breach, 1994), 3–27; and Keith Otterbein, The Anthropology of War (Long Grove, IL: Waveland Press, 2009), 62–64. 24. Robert Carneiro, “War and Peace: Alternating Realities in Human History,” 3–27. 25. Barbara Ehrenreich, “Men Hate War Too,” Foreign Affairs 78, no. 1 (1999): 118–22, 118. 26. Kim, “The Lorenzian Theory,” 95. 27. Jane Goodall, Through a Window: My Thirty Years with the Chimpanzees of Gombe (Boston: Houghton Mifflin, 1990), 210. 28. See, for instance, Diane Fossey, Gorillas in the Mist (Boston: Houghton Mifflin, 1983), and Jane Goodall, Through a Window. 29. Of the 3 billion letters that make up the human genome, only 15 million (less than 1 percent) have changed since human and chimp lineages diverged 6 million years ago. Katherine S. Pollard, “What Makes Us Human?” Scientific American (May 2009), 44–49. Many of the changes are clustered around parts of the genetic code dealing with the brain, permitting the enlargement of the cerebral cortex. (The human brain’s volume has more than tripled since the time of the common human-chimp ancestor.) 30. Jane Goodall, The Chimpanzees of the Gombe: Patterns of Behavior (Cambridge, MA: Harvard University Press, 1982); and Goodall, Through a Window. On collecting rocks, see Seth Borenstein, “The Thinker, Nonhuman Variety,” Washington Post, July 17, 2012, E2. 31. Michael Wilson and Richard Wrangham, “Intergroup Relations in Chimpanzees,” Annual Review of Anthropology 32 (2003): 363–92, especially 369. 32. Goodall, Through a Window, especially 75–84, 98–111, and 206–16; and Goodall, The Chimpanzees of Gombe, 526–28. 33. Goodall, The Chimpanzees of Gombe, 530. 34. M. Wilson and R. Wrangham, “Intergroup Relations in Chimpanzees,” 370–76. The five sites are Gombe and Mahale in Tanzania, Tai in Tanzania, and Kibale and Budongo in Uganda. See also John Mitani, David Watts, and Sylvia Amsler, “Lethal Intergroup Aggression Leads to Territorial Expansion in Wild Chimpanzees,” Current Biology 20, no. 12 (June 2010): 507–9. 35. R. Carneiro, “War and Peace: Alternating Realities in Human History,” 9.

Notes

493

36. M. Wilson and R. Wrangham, “Intergroup Relations in Chimpanzees,” 370. 37. M. Wilson and R. Wrangham, “Intergroup Relations in Chimpanzees,” 372. There is, of course, plentiful evidence of intragroup violence unconnected to territorial behavior—both in the wild and in sanctuaries. For instance, Frans de Waal and his colleagues at the Arnhem Zoo in the Netherlands detected the violent death of a chimpanzee that resolved a particularly nasty three-way dominance struggle among the senior males in the community. Frans de Waal, Peacemaking among Primates (Cambridge, MA: Harvard University Press, 1989), 61–66. 38. M. Wilson and R. Wrangham, “Intergroup Relations in Chimpanzees,” 384–85. 39. Robert Sapolsky, “A Natural History of Peace,” Foreign Affairs 85, no. 1 (January/February, 2006): 106–7; Goldstein, War and Gender, 187–89, 220. Some scholars suggest that the difference between bonobo and chimpanzee behavior may be due to environmental factors: Bonobos inhabit large, undisturbed rainforests. See I. J. N. Thorpe, “Anthropology, Archaeology, and the Origins of Warfare,” World Archaeology 35, no. 1 (2003): 145–65. 40. Frans de Waal, Peacemaking among Primates (Cambridge, MA: Harvard University Press, 1989), and Good Natured: The Origins of Right and Wrong in Humans and Other Animals (Cambridge, MA: Harvard University Press, 1995). 41. Sapolsky, “A Natural History of Peace,” 109; de Waal, Good Natured, 177. De Waal’s working hypothesis about variation in peacemaking is that those species that show the most reconciliation behavior are those in which group cohesiveness is most crucial. 42. De Waal, Good Natured, 176. Attempts at peacemaking are also made by third parties (usually kin) in some societies (De Waal, Good Natured, 204). 43. De Waal, Peacemaking among Primates, 95–96. 44. De Waal, Peacemaking among Primates, 53. 45. De Waal, Peacemaking among Primates, 38–39, 261. As de Waal notes, “Reconciliation relates to both past and future; it serves to ‘undo’ previous events with an eye to future relationships” (Peacemaking among Primates, 39). 46. De Waal, Peacemaking among Primates, 243. 47. De Waal, Peacemaking among Primates, 229. 48. De Waal, Peacemaking among Primates, 270. 49. Albert Somit, “Humans, Chimps, and Bonobos,” Journal of Conflict Resolution 34, no. 3 (September 1990): 578–79. 50. Edward O. Wilson, Sociobiology: The New Synthesis (Cambridge, MA: Harvard University Press, 1975). See also Wilson’s On Human Nature (Cambridge, MA: Harvard University Press, 1978). 51. Wilson, On Human Nature, 20–22. 52. Bradley Thayer, “Bringing in Darwin: Evolutionary Theory, Realism, and International Politics,” International Security 25, no. 2 (Fall 2000): 124–51, 130. 53. Steven Mithen, “How Fit Is E. O. Wilson’s Evolution?” New York Review of Books (June 21, 2012), 26–28. 54. The theory of inclusive fitness has become the dominant approach to modern evolutionary thinking, but Wilson, in The Social Conquest of the Earth (New York: Liveright, 2012), now supports a different and much more controversial variant of evolutionary theory called group selection. Group selection means that genes or combinations of genes become fixed in a population due to the benefits they bestow on the whole group, regardless of the effects on the inclusive fitness of individuals and their close kin. He now sees warfare as completely universal and as the epitome of group selection. See Mithen, “How Fit Is E. O. Wilson’s Evolution?” 55. Wilson, On Human Nature, 99–100 and throughout chapter 5, 99–120. 56. Thayer, “Bringing in Darwin,” 141–42. 57. See Thayer, “Bringing in Darwin.” 58. Azar Gat, War in Human Civilization (Oxford: Oxford University Press, 2006), 140. 59. Gat, War in Human Civilization, 44–46. 60. Gat, War in Human Civilization, 47–52; 135–36. 61. Frans de Waal, Good Natured, 220, note 11. 62. Wilson, Sociobiology, 1980 edition, 122. 63. Goldstein, War and Gender, 139. 64. Wilson, Sociobiology, 1980 edition, 122. 65. Wilson, On Human Nature, 106–7 and 114, as well as 101–2. 66. Gat, War in Human Civilization, 40–41. 67. Wilson, On Human Nature, 116–17. 68. Wilson, On Human Nature, 106.

494

Notes

69. Wilson, On Human Nature, 119. 70. See, for instance, Stephen Jay Gould, “Darwinian Fundamentalism,” New York Review of Books, June 12, 1997; Richard Lewontin, Steven Rose, and Leon Kamin, Biology, Ideology and Human Nature: Not in Our Genes (Pantheon, 1984); and Frans de Waal, Peacemaking among Primates. 71. Wilson, Sociobiology, 1975 edition, 345. 72. Montague, Sociobiology Examined, 7–8 73. Montague, Sociobiology Examined, 8–10. See also Marshall Sahlins, The Use and Abuse of Biology: An Anthropological Critique of Sociobiology (Ann Arbor: University of Michigan Press, 1976). 74. Duncan Bell and Paul MacDonald, “Correspondence: Start the Evolution without Us,” International Security 26, no. 1 (Summer 2001): 187–94. 75. Bell and MacDonald, “Correspondence,” 192. 76. Keith Otterbein, The Anthropology of War (Long Grove, IL: Waveland Press, 2009), 53–54. 77. Otterbein, The Anthropology of War, 53. 78. See Richard A. Gabriel, The Culture of War: Invention and Early Development (New York: Greenwood Press, 1990), 16–17. 79. See Goldstein, War and Gender, 157–58. 80. Otterbein, The Anthropology of War, 55–56. 81. See Virginia Morell, “Evidence Found for a Possible ‘Aggression Gene,’” Science, 260 (June 18, 1993), 1722–23; Boyce Rensberger, “Science and Sensitivity,” Washington Post, March 1, 1992; and Associated Press, “Aggressive Behavior, Genetic Flaw Linked,” Washington Post, October 22, 1993, A3. 82. Genes are not the only reason for low levels of serotonin; chronic alcoholism decreases serotonin production, as does the taking of amphetamines, like the drug Ecstasy, which destroys brain cells that make serotonin. 83. Goldstein, War and Gender, 136. 84. Martin Daly and Margo Wilson, Sex, Evolution and Behavior, 2nd ed. (Boston: Willard Grant Press, 1983), 146, quoted in David Barash, “Evolution, Males, and Violence,” The Chronicle of Higher Education, May 24, 2002, B7–B9, B9. 85. On the other hand, at least one review of psychological research indicates that, “In laboratory studies of human aggression, where the use of physical aggression is controlled and the possibility of escalation of violence is eliminated, there is little difference in the frequency of aggression in males and females.” Quoted in R. Brian Ferguson, “Perilous Positions,” Foreign Affairs 78, no. 1 (1999): 125–27, 125. 86. See Francis Fukuyama, “Women and the Evolution of World Politics,” Foreign Affairs 77, no. 5 (1998): 24–40, especially 30. 87. Robert Sapolsky, The Trouble with Testosterone: And Other Essays on the Biology of the Human Predicament (New York: Scribner, 1997), 149–59, cited in Goldstein, War and Gender, 150. 88. Goldstein, War and Gender, 131. 89. Goldstein, War and Gender, 139. 90. On these points, see Evan Luard, War in International Society (New Haven, CT: Yale University Press, 1986). 91. Bronislaw Malinowski, “War—Past, Present, and Future,” in War as a Social Institution, ed. J. D. Clarkson and T. C. Cochran (New York: Columbia University, 1941), 21–31, 23, quoted in Robert Carneiro, “War and Peace: Alternating Realities in Human History,” 9. 92. Montagu, “The New Litany,” 6, 16. 93. Leakey, The Making of Mankind, 229. 94. Leakey, The Making of Mankind, 229–30, 237. 95. Andrew Bard Schmookler, The Parable of the Tribes: The Problem of Power in Social Evolution (Boston: Houghton Mifflin, 1984). 96. The strategic choices facing Neolithic societies provide a good example of the Prisoners’ Dilemma game, which we will explore in chapter 8. 97. Robert Carneiro, “War and Peace: Alternating Realities in Human History,” 14. See also R. Carneiro, “A Theory of the Origin of the State,” Science 169 (1970): 733–38. 98. Keith F. Otterbein, The Anthropology of War, especially chapter 6. 99. Otterbein, The Anthropology of War, 75–77. 100. Otterbein, The Anthropology of War, 24–26. The development of fraternal interest groups is aided by patrilocal residence arrangements and also by polygyny. For Otterbein, social-political organization is the primary underlying cause of warfare.

Notes

495

101. Otterbein, The Anthropology of War, 67. 102. Gat, War in Human Civilization, 160–72. 103. Gat, War in Human Civilization, 166. 104. David Fabbro, “Peaceful Societies,” in The War System, ed. R. Falk and S. S. Kim (Boulder, CO: Westview, 1980), 180–203. 105. Fabbro, “Peaceful Societies,” 181. 106. Fabbro, “Peaceful Societies,” 199. 107. Fabbro, “Peaceful Societies,” 181. 108. Elizabeth Thomas, The Harmless People (New York: Knopf, 1959), quoted in Seyom Brown, The Causes and Prevention of War (New York: St. Martin’s, 1987), 14. 109. Gwynne Dyer, War (New York: Dorsey, 1985), 6. 110. Dyer, War, 9. 111. Dyer, War, 9. 112. Dyer, War, 10. 113. Quincy Wright, A Study of War (Chicago: University of Chicago Press, 1965), 63. 114. Werner Levi, “The Causes of War and the Conditions of Peace,” in Toward a Theory of War Prevention, ed. Richard Falk and S. H. Mendlovitz (New York: World Law Fund, 1966). 115. Lawrence Keeley, War before Civilization (New York: Oxford University Press, 1996), 8–15. 116. Keeley, War before Civilization, 27–29. 117. Carol Ember and Melvin Ember, “Violence in the Ethnographic Record: Results of Cross-Cultural Research on War and Aggression,” in Troubled Times: Violence and Warfare in the Past, ed. Debra Martin and David Frayer (Langhorne, PA: Gordon & Breach, 1997), 1–20; and Melvin Ember and Carol Ember, “Cross-Cultural Studies of War and Peace: Recent Achievements and Future Possibilities,” in Studying War: Anthropological Perspectives, ed. S. P. Reyna and R. E. Downs (Langhorne, PA: Gordon & Breach, 1994), 185–208. Both cited in Goldstein, War and Gender, 26. 118. Goldstein, War and Gender, 32–33. 119. Gat, War in Human Civilization, 17–35. 120. Gat, War in Human Civilization, 35. 121. See Keeley, War before Civilization, 30–32, and Goldstein, War and Gender, 32–34. 122. Ember and Ember, “Cross-Cultural Studies of War and Peace,” 189. 123. Keeley, War before Civilization, 29–36. 124. Mervyn Meggitt, Blood Is Their Argument: Warfare among the Mae Enga Tribesmen of the New Guinea Highlands (Palo Alto, CA: Mayfield Publishing, 1977), 201, reported in Ember and Ember, “Cross-Cultural Studies of War and Peace,” 189–90. 125. Keeley, War before Civilization, 29. 126. For killing rates see Azar Gat, “The Pattern of Fighting in Simple, Small-Scale, Prestate Societies,” Journal of Anthropological Research 55 (1999): 563–83. 127. Keeley, War before Civilization, 32. 128. Azar Gat, “The Pattern of Fighting in Simple, Small-Scale, Prestate Societies,” and War in Human Civilization, 114–32. The differences between human and animal patterns are these: (1) it is quite difficult for animal attackers to surprise their rivals due to their possession of rather acute senses; (2) it is more difficult for animals to “finish off” their rivals due to the lack of external weapons; and (3) the young are frequently the primary victims in the animal world, whereas male adults are the primary fatalities in human raiding activity. 129. Gat, War in Human Civilization, 131. 130. Ember and Ember, “Cross-Cultural Studies of War and Peace,” 190–96. 131. Ember and Ember, “Cross-Cultural Studies of War and Peace,” 194. 132. Keeley, War before Civilization, 37. One burial site has been dated with radiocarbon testing at circa 13,500 BCE. See I. J. N. Thorpe, “Ancient Origins of Warfare and Violence,” in Warfare, Violence, and Slavery in Prehistory, ed. I. J. N. Thorpe (Oxford: British Archaeological Reports, 2005), 1–18. 133. Keeley, War before Civilization, 38–39. 134. Raymond Kelly, Warless Societies and the Origin of War (Ann Arbor: University of Michigan Press, 2000). 135. Kelly, Warless Societies, 37. 136. Kelly, Warless Societies, 18. 137. Kelly, Warless Societies, 7.

496

Notes

138. For instance, Ember’s data, which Keeley cites, lump these different types of violence together, inflating the frequency of war. Kelly, Warless Societies, 123. 139. Otterbein, The Anthropology of War, 72–74. 140. Social substitutability assumes both group injury and group liability. An injury to one member of the group is seen as a collective injury demanding a collective response, and those who are liable for the injury include members—though perhaps not all members—of the opposing group. Keith Otterbein (The Anthropology of War, 43–45) notes that many scholars find it virtually impossible to distinguish feuding from war and consequently lump them together in their analyses (like Ember and Ember). Otterbein simply says that warfare is armed combat between political communities and feuding is within political communities. He uses the term “internal war” for armed combat between local groups that are culturally the same and are also separate political communities. 141. Kelly, Warless Societies, 42. 142. Kelly, Warless Societies, 60, 62, Indeed, the conceptualization of marriage as a transaction between social groups marks an important step in the transition from unsegmented societies to segmented ones. 143. Kelly, Warless Societies, 54–55. 144. Kelly, Warless Societies, 56. 145. Kelly, Warless Societies, 51–52. There are only two cases of unsegmented foraging societies with frequent warfare, and one of these, the Slave tribe of Canada, were largely victims of attacks by the Cree (a segmented neighbor) and tended to respond by flight rather than defensive violence. The only other example was the Andaman Islanders. 146. Kelly, Warless Societies, 148. Note that, for Kelly, the decisive difference between warless and war-fighting societies is not the transition from a mobile existence to a sedentary one, as the traditional explanation predicts. Many mobile foraging societies are warlike. The difference is how these groups are organized (Kelly, 66–67). While unsegmented organization is the dominant factor, Kelly (73) is able to draw a more composite picture of warless societies that includes these attributes: the absence of marriage payments and concepts of kin group member liability, frequent marriage with members of outside communities, migratory or seminomadic settlement, absence of food storage, and a population density typically below one person per square mile. 147. Kelly, Warless Societies, 147. 148. Kelly, Warless Societies, 75, also 130–31. 149. Kelly, Warless Societies, 147–61. 150. Kelly, Warless Societies, 141–43, 147–49. 151. Thorpe, “The Ancient Origins of Warfare and Violence,” 11. 152. Jonathan Haas, “Warfare and the Evolution of Culture,” in Archaeology at the Millennium, ed. G. M. Feinman and T. D. Price (New York: Kluwer/Plenum, 2001), 329–50. 153. Scott, “That Old-Time Aggression,” 54. 154. See Margaret Mead, “Warfare Is Only an Invention—Not Biological Necessity,” in Peace and War, ed. Charles Beitz and Theodore Herman (San Francisco: WH Freeman, 1973), 112–18. 155. See Albert Bandura, “The Social Learning Theory of Aggression,” in The War System, ed. R. Falk and S. S. Kim (Boulder, CO: Westview, 1980), 141–56. 156. Edwin I. Megargee and Jack E. Hokanson, The Dynamics of Aggression (New York: Harper & Row, 1970), 43, cited in Lloyd Jensen, Explaining Foreign Policy (Englewood Cliffs, NJ: Prentice-Hall, 1982), 20. 157. De Waal, Good Natured, 178–80. Also cited in Sapolsky, “A Natural History of Peace,” 113–14. 158. Sapolsky, “A Natural History of Peace,” 6–8. 159. Sapolsky, “A Natural History of Peace,” 114–17. 160. Sapolsky, “A Natural History of Peace,” 106. 161. Jared Diamond, Guns, Germs and Steel (New York: W. W. Norton, 1997). 162. Diamond, Guns, Germs and Steel, 54. 163. Diamond, Guns, Germs and Steel, 54. 164. Diamond, Guns, Germs and Steel, 55. 165. Diamond, Guns, Germs and Steel, 56. 166. John Vasquez, The War Puzzle (Cambridge: Cambridge University Press, 1993), 140. 167. Sapolsky, “A Natural History of Peace,” 111–12.

Notes

497

3. THE INDIVIDUAL LEVEL OF ANALYSIS, PART II 1. For an article that makes this point, see Daniel Byman and Kenneth Pollack, “Let Us Now Praise Great Men,” International Security 25, no. 4 (Spring 2001): 107–46. The authors note (112) that “Because personalities differ, it is entirely possible that variance in the traits of individuals explains the differences in international relations. For instance, although not all wars have been caused by aggressive, risk-tolerant, greedy, or vainglorious leaders, those leaders who did manifest these traits regularly went to war.” 2. John Stoessinger, Why Nations Go to War, 3rd ed. (New York: St. Martin’s, 1982), 135. Emphasis added. 3. The following draws from Ole Holsti, “Foreign Policy Decision-Makers Viewed Psychologically: ‘Cognitive Process’ Approaches,” In Search of Global Patterns, ed. James Rosenau (New York: Free Press, 1972), 127; Margaret Hermann, “Effects of Personal Characteristics of Political Leaders on Foreign Policy,” in Why Nations Act: Theoretical Perspectives for Comparative Foreign Policy Studies , ed. M. East, S. Salmore, and C. F. Hermann (Beverly Hills, CA: Sage, 1978), 51–52; and Byman and Pollack, “Let Us Now Praise Great Men,” 141. 4. Abraham Maslow, “A Theory of Human Motivation,” Psychological Review 50 (1943): 394. 5. Although empirical evidence exists for the presence of physical and security needs, and indeed for their hierarchical nature, empirical validation for the existence of Maslow’s three higher needs is lacking. See Ross Fitzgerald, “Abraham Maslow’s Hierarchy of Needs—An Exposition and Evaluation,” in Human Needs and Politics, ed. R. Fitzgerald (Rushcutters Bay, Australia: Pergamon Press, 1977), 36–51. 6. Roy Baumeister, Laura Smart, and Joseph Boden, “Relation of Threatened Egotism to Violence and Aggression: The Dark Side of High Self-Esteem,” Psychological Review 103, no. 1 (1996): 5–33. 7. Lloyd S. Etheredge, A World of Men: The Private Sources of American Foreign Policy (Cambridge, MA: MIT Press, 1978); see also L. S. Etheredge, “Personality and Foreign Policy,” Psychology Today, March 1975, 37–42. 8. David Winter, “Measuring the Motives of Political Actors at a Distance,” in The Psychological Assessment of Political Leaders, ed. Jerrold Post (Ann Arbor: University of Michigan Press, 2005), 153–77. 9. Winter, “Measuring the Motives of Political Actors at a Distance,” 158. 10. See discussion of findings in Martha Cottam, Beth Dietz-Uhler, Elena Mastors, and Thomas Preston, Introduction to Political Psychology, 2nd ed. (New York: Psychology Press/Taylor & Francis, 2010), 28–29. 11. D. Winter, “Measuring the Motives of Political Actors at a Distance,” 158; and D. Winter, “Assessing Leaders’ Personalities,” in The Psychological Assessment of Political Leaders, ed. J. Post (Ann Arbor: University of Michigan Press, 2005), 32–33. See also Kenneth Terhune, “Studies in Motives, Cooperation, and Conflict within Laboratory Microcosms,” Buffalo Studies 4 (1968): 29–58, cited in Lloyd Jensen, Explaining Foreign Policy (Englewood Cliffs, NJ: Prentice Hall, 1982), 22–23. 12. David G. Winter and Abigail J. Stewart, “Content Analysis as a Technique for Assessing Political Leaders,” in A Psychological Examination of Political Leaders, ed. M. G. Hermann (New York: Free Press, 1977), 60. 13. Harold Lasswell, Psychopathology and Politics (Chicago: University of Chicago Press, 1930), and Power and Personality (New York: Norton, 1948). 14. K. W. Terhune and J. M. Firestone, “Psychological Studies in Social Interaction and Motives (SIAM), Phase 2: Group Motives in International Relations Games,” CAL Report VX-2018-6-2, Cornell Aeronautical Laboratory, 1967, and D. G. Winter, The Power Motive (New York: Free Press, 1973), both cited in M. Hermann, “Effects of Personal Characteristics of Political Leaders on Foreign Policy,” 66. 15. Winter, “Measuring the Motives of Political Actors at a Distance,” 159. 16. Margaret G. Hermann, “Explaining Foreign Policy Behavior Using the Personal Characteristics of Political Leaders,” International Studies Quarterly 24 (1980), 7–46, cited in Winter, “Measuring the Motives of Political Actors at a Distance,” 163. 17. Winter and Stewart, “Content Analysis as a Technique for Assessing Political Leaders,” 60. 18. K. W. Terhune, “Motives, Situation, and Interpersonal Conflict within Prisoners’ Dilemma,” Journal of Personality and Social Psychology, Monograph Supplement 8, No. 3 (1968) Part 2: 1–23, cited in Hermann, “Effects of Personal Characteristics of Political Leaders on Foreign Policy,” 66. 19. David G. Winter, “Saddam Hussein: Motivations and Mediations of Self-Other Relationships,” in The Psychological Assessment of Political Leaders, ed. J. Post (Ann Arbor: University of Michigan, 2005), 372. Hussein’s above-average ranking on affiliation should be seen in context: he had close political relationships only with those who were seen as similar to himself and in whose company he felt safe. He tended to rely on an inner circle of

498

Notes

trusted advisers who were either members of his family or friends from his home town of Tikrit. His need for affiliation did not extend beyond this rather closed circle. 20. Milton Rokeach, The Open and Closed Mind (New York: Basic Books, 1960). 21. Milton Rokeach, “The Nature and Meaning of Dogmatism,” Psychological Review 61 (May 1954): 200, cited in Jensen, Explaining Foreign Policy, 26. 22. T. W. Adorno, The Authoritarian Personality (New York: Harper & Row, 1950). 23. Jensen, Explaining Foreign Policy, 25. 24. Winter, “Assessing Leaders’ Personalities,” 32–33. 25. Lloyd Etheredge, “Personality Effects on American Foreign Policy, 1898–1968,” American Political Science Review 72 (June 1978): 434–51; and Graham H. Shepard, “Personality Effects on American Foreign Policy, 1969–1984: A Second Test of Interpersonal Generalization Theory,” International Studies Quarterly 32, no. 1 (March 1988): 91–123. 26. Etheredge, “Personality Effects,” 449. Shepard is unable to confirm any difference between the effects of introverted and extroverted personality on U.S. policy toward the Soviet Union in the 1969–1984 period. 27. See Bruce Bueno de Mesquita, “Risk, Power Distribution and the Likelihood of War,” International Studies Quarterly 25, no. 4 (December 1981): 542–46; and Bueno de Mesquita, The War Trap (New Haven, CT: Yale University Press, 1981). 28. Byman and Pollack, “Let Us Now Praise Great Men,” 137. 29. Margaret Hermann, “Assessing the Foreign Policy Role Orientations of Sub-Saharan African Leaders,” in Role Theory and Foreign Policy Analysis, ed. S. G. Walker (Durham, NC: Duke University Press, 1987), 161–98. 30. Margaret G. Hermann, “Saddam Hussein’s Leadership Style,” in The Psychological Assessment of Political Leaders, ed. J. Post (Ann Arbor: University of Michigan Press, 2005), 375–86. 31. Stephen Benedict Dyson, “Personality and Foreign Policy: Tony Blair’s Iraq Decisions,” Foreign Policy Analysis 2, no. 3 (July 2006): 289–306. 32. Jerrold Post, “Assessing Leaders at a Distance: The Political Personality Profile,” in Post, The Psychological Assessment of Political Leaders, 69–104; see 69 and 77–78 especially. 33. Post, “Assessing Leaders at a Distance,” 80. 34. Post, “Assessing Leaders at a Distance,” 80. 35. Post, “Assessing Leaders at a Distance,” 79–80. 36. Post, “Assessing Leaders at a Distance,” 100. 37. The following relies on Post, “Assessing Leaders at a Distance,” 93–100; and Raymond Birt, “Personality and Foreign Policy: The Case of Stalin,” Political Psychology 14, no. 4 (1993): 607–25. 38. Birt, “Personality and Foreign Policy: The Case of Stalin.” 39. For other analyses of Stalin’s failure in June 1941, see Simon S. Montefiore, Stalin: The Court of the Red Tsar (New York: Vintage, 2003); Adam Ulam, Stalin: The Man and His Era (New York: Viking, 1973); and William Taubman, Khrushchev: The Man and His Era (New York: W. W. Norton, 2003). 40. See J. Post, “Assessing Leaders at a Distance”; R. Raskin, J. Novacek, and R. Hogan, “Narcissism SelfEsteem Management,” Journal of Personality and Social Psychology 60, no. 6 (June 1991): 911–18; R. Raskin and H. Terry, “A Principle Components Analysis of the Narcissistic Personality Inventory and Further Evidence of its Construct Validity,” Journal of Personality and Social Psychology 54, no. 5 (May 1988): 890–902; and L. Carroll, “A Comparative Study of Narcissism, Gender, and Sex Role Orientation among Body Builders, Athletes and Psychology Students,” Psychological Reports 64, no. 3 (June 1989): 999–1006. 41. U.S. House of Representatives, Committee of Foreign Affairs, “The Persian Gulf Crisis,” Statement by Jerrold Post, December 11, 1990 (pp. 381–401), USGPO, Washington, D.C., 1991; McNeill-Lehrer News Hour, February 2, 1991, interview with Betty Glad and Jerrold Post. J. Post; “Saddam Hussein’s Leadership during the Gulf Crisis,” in The Political Psychology of the Gulf War, ed. Stanley Renshon (Pittsburgh: University of Pittsburgh Press, 1993), 49–66. See also the analyses of Saddam Hussein all found in Jerrold Post (ed.) The Psychological Assessment of Political Leaders: J. Post, “Saddam Hussein of Iraq: A Political Psychological Profile,” 335–66; Walter Weintraub, “General Personality Traits and Ego Defenses,” 367–70; David G. Winter, “Motivations and Mediation of Self-Other Relationships,” 370–74; Margaret G. Hermann, “Saddam Hussein’s Leadership Style,” 375–86; Stephen G. Walker, Mark Schafer, and Michael D. Young, “Saddam Hussein: Operational Code Beliefs and Object Appraisal,” 387–91; Peter Suedfeld, “Saddam Hussein’s Integrative Complexity under Stress,” 391–96. 42. Post, “Saddam Hussein: A Political Psychological Profile,” 342. 43. Post, “Saddam Hussein: A Political Psychological Profile,” 349. Stanley Renshon generally agrees with this analysis, but adds that narcissistic rage—an extreme and emotionally violent response to narcissistic injury to one’s

Notes

499

grandiose self-image—needs to be taken into account in Hussein’s case. “Both grandiosity and rage interfere with good judgment . . . Grandiosity interferes with judgment because it leads to an overestimation of one’s skills and ability to affect circumstances. Rage interferes with judgment because the emphasis is on getting even and proving one’s opponents wrong (and oneself right) rather than finding an effective strategy to pursue realistic goals.” While the quote refers to Hussein’s judgment during the 1990–91 Gulf War, it is just as relevant to the events leading up to the 2003 war. Stanley Renshon, “The Gulf War Revisited,” in The Political Psychology of the Gulf War, ed. Stanley Renshon (Pittsburgh: University of Pittsburgh Press, 1993), 329–57; quote is from 344–45. 44. For case studies of the 1980–88 Iran-Iraq War and the 2003 Iraq War, see Greg Cashman and Leonard Robinson, An Introduction to the Causes of War (Lanham, MD: Rowman & Littlefield, 2007). 45. Jerome Frank, Sanity and Survival: Psychological Aspects of War and Peace (New York: Random HouseVintage, 1967), 59. 46. John C. G. Rohl, The Kaiser and His Court (Cambridge: Cambridge University Press, 1987); Robert Waite, “Leadership Pathologies: The Kaiser and the Fuhrer and the Decisions for War in 1914 and 1939,” in Psychological Dimensions of War, ed. Betty Glad (Newbury Park, CA: Sage, 1990), 143–68; and Richard Ned Lebow, Between Peace and War: The Nature of International Crises (Baltimore: Johns Hopkins University, 1981). This summary draws on a previous summary in Cashman and Robinson, An Introduction to the Causes of War, 76–79. 47. Rohl, The Kaiser and His Court, 21. 48. See Alexander George and Juliette George, Woodrow Wilson and Colonel House: A Personality Study (New York: Dover Publications, 1964). 49. Rohl, The Kaiser and His Court, 11–27; Waite, “Leadership Pathologies,” 153; Lebow, Between Peace and War, 143. 50. Waite, “Leadership Pathologies,” 154. 51. Lebow, Between Peace and War, 142–43. 52. See especially Rose McDermott, “The Feeling of Rationality: The Meaning of Neuroscientific Advances for Political Science,” Perspectives on Politics 2, no. 4 (December 2004): 691–706; Rose McDermott, “Mutual Interests: The Case for Increasing Dialogue between Political Science and Neuroscience,” Political Research Quarterly 62, no. 3 (September 2009): 571–83; Neta C. Crawford, “The Passion of World Politics: Propositions on Emotion and Emotional Relationships,” International Security 24, no. 4 (Spring 2000): 116–56; Janice Gross Stein, “Foreign Policy Decision-Making: Rational, Psychological, and Neurological Models,” in Foreign Policy: Theories, Actors, Cases, ed. Steve Smith, Amelia Hadfield, and Tim Dunne (Oxford: Oxford University Press, 2008), 102–14; and Jonathan Mercer, “Rationality and Psychology in International Politics,” International Organization 59 (Winter 2005): 77–106. 53. Crawford, “The Passion of World Politics,” 124–25. A basic list of emotional states would include love, fear, anger, joy, sadness, and shame. 54. McDermott, “The Feeling of Rationality,” 692. 55. McDermott, “The Feeling of Rationality,” 693. 56. Antonio Damasio, Descartes’ Error: Emotion, Reason, and the Human Brain (New York: Putnam & Sons, 1996); Hanna Damasio, Thomas Grabowski, Randall Frank, Albert Galaburda, and Antonio Damasio, “The Return of Phineas Gage: The Skull of a Famous Patient Yields Clues about the Brain,” Science 264, no. 5162 (1994): 1102–5; Antoine Bechara, Hanna Demasio, Daniel Tranel, and Antonio Damasio, “Deciding Advantageously before Knowing the Advantageous Strategy,” Science 275, no. 5304 (1997): 1293–95; and Antoine Bechara, Daniel Tranel, and Hanna Damasio, “Characterization of the Decision-Making Deficit of Patients with Ventromedial Prefrontal Cortex Lesions,” Brain 123, no. 11 (2000): 2189–2201. Summarized in McDermott, “The Feeling of Rationality”; J. G. Stein, “Foreign Policy Decision-Making”; and Mercer, “Rationality and Psychology in International Politics.” 57. Mercer, “Rationality and Psychology in International Politics,” 93–94. It’s also true that people with brain damage similar to Elliot’s are unable to accurately judge the trustworthiness of people based on their faces. 58. Damasio, Descartes’ Error, summarized in McDermott, “The Feeling of Rationality,” 694. 59. Daniel Kahneman, Thinking, Fast and Slow (New York; Farrar, Straus and Giroux, 2011), 139. 60. Mercer, “Rationality and Psychology in International Politics,” 94 and 95. Mercer also suggests (97) that emotion is important to identity: It is the emotional content of one’s identity that provides the basis for trust. Identity produces emotion and emotion creates trust, which then leads to cooperation. “The emotion in identity explains why group members trust each other and why they may distrust out-group members.” 61. Summarized in Stein, “Foreign Policy Decision-Making,” 111. 62. Kahneman, Thinking Fast and Slow, 29. 63. Stein, “Foreign Policy Decision-Making,” 110.

500

Notes

64. D. Overbye, “Free Will: Now You Have It, Now You Don’t,” New York Times, January 2, 2007, D1. 65. Robert Jervis, “Understanding Beliefs,” Political Psychology 27, no. 5 (2006): 641–63; quote is from 645. 66. See, for instance, Stein, “Foreign Policy Decision-Making,” 112–13. 67. Paul Slovic, Melissa Finucane, Ellen Peters, and Donald MacGregor, “The Affect Heuristic in Judgments of Risks and Benefits,” in Heuristics and Biases, ed. Thomas Gilovich, Dale Griffin, and Daniel Kahneman (New York; Cambridge University Press, 2002), 397–420, referred to in Kahneman, Thinking, Fast and Slow, 103. 68. Crawford, “The Passion of World Politics,” 136. 69. A mood is “an amorphous state—like emotions, but without specific objects or referents.” McDermott, “The Feeling of Rationality,” 692. 70. According to mood theory, we are more likely to remember things that are consistent with our present mood. Stein also suggests that analogies are triggered by emotional cues: a leader’s mood may “affect which lessons from history are drawn on—events that generated similar emotions of anger and fear in the past become more accessible. Learning from history can thus be driven by affect as well as by rational thought.” Stein, “Foreign Policy DecisionMaking,” 696; see also McDermott, “Mutual Interests,” 576. On mood theory see Gordon Bower, “Mood and Memory,” American Psychologist 36, no. 2 (1981): 129–48. 71. Kaheman, Thinking, Fast and Slow, 69. 72. McDermott, “Mutual Interests,” 576. 73. Crawford, “The Passion of World Politics,” 146–47. 74. Crawford, “The Passion of World Politics,” 141 75. Crawford, “The Passion of World Politics,” 138. 76. Crawford, “The Passion of World Politics,” 143. 77. McDermott, “Mutual Interests,” 576. People with positive moods are actually more acceptant of risk if they perceive the risk is low. Crawford, “The Passion of World Politics,” 144. 78. Y. Vertzberger, Risk Taking and Decisionmaking (Stanford, CA: Stanford University Press, 1998), 76. Some researchers note that fear promotes relatively pessimistic assessments of risk, but anger leads to relatively optimistic assessment. Thus, those who respond with anger to an opponent’s behavior would be generally more likely to respond with a risky military action than one whose primary reaction was fear. In the latter case, the more likely response would be caution. See McDermott, “The Feeling of Rationality,” 696–97. This draws on research from Jennifer Lerner and Dacher Keltner, “Beyond Valence: Toward a Model of Emotion-Specific Influences on Judgment and Choice,” Cognition and Emotion 14, no. 4 (200): 473–93. Others suggest that risk calculations are affected by emotional assessment of the target. As Paul Slovik notes, those who hated Saddam Hussein were likely to feel his possession of WMD created a large risk for the U.S., while those who saw him as sane probably saw the threat as presenting the U.S. with less risk. Paul Slovic, “Trust, Emotion, Sex, Politics and Science: Surveying the Risk Assessment Battlefield,” Risk Analysis 19, no. 4 (1999): 689–701, quoted in McDermott, “The Feeling of Rationality,” 697. 79. Crawford, “The Passion of World Politics,” 134. 80. For two excellent reviews of this material, see Janice Gross Stein, “Foreign Policy Decision-Making: Rational, Psychological, and Neurological Models,” in Foreign Policy: Theories, Actions, Cases, ed. Steve Smith, Amelia Hadfield, and Tim Dunne (Oxford: Oxford University Press, 2008), 102–14; and Jerel A. Rosati, “The Power of Human Cognition in the Study of World Politics,” International Studies Review 2, no. 3 (Autumn 2000): 45–75. See also Kahneman, Thinking, Fast and Slow. 81. Stein, “Foreign Policy Decision-Making,” 109. 82. John Steinbruner, The Cybernetic Theory of Decision (Princeton, NJ: Princeton University Press, 1974), 112. 83. Rosati, The Power of Human Cognition,” 51. 84. Stein, “Foreign Policy Decision-Making,” 107–8. Stein’s discussion draws on A. Tversky and D. Kahneman, “Availability: A Heuristic for Judging Frequency and Probability,” Cognitive Psychology 5 (1973): 207–32; D. Kahneman, P. Slovic, and A. Tversky, Judgment under Uncertainty: Heuristics and Biases (New York: Cambridge University Press, 1982); and R. Jervis, “Representativeness in Foreign Policy Judgments,” Political Psychology 7, no. 3 (1986): 483–505. 85. Kahneman, Thinking, Fast and Slow, 144. 86. A. Tversky and D. Kahneman, “Judgment under Uncertainty: Heuristics and Biases,” in Kahneman et al., Judgment under Uncertainty, 3–20. Cited in Cottam et al., Introduction to Political Psychology, 39–40. 87. Stein, “Foreign Policy Decision Making,” 108. See also Cottam et al., Introduction to Political Psychology, 39–40. It’s also true that desirable outcomes are seen as more likely to occur than undesirable ones, but this is a motivated bias rather than an unmotivated bias. It’s related to the general tendency toward wishful thinking.

Notes

501

88. R. Jervis, “Perception and Misperception in International Politics,” 319–23. 89. See Martha Cottam, Beth Dietz-Uhler, Elena Mastors, and Thomas Preston, Introduction to Political Psychology, 2nd ed. (New York: Psychology Press, 2010), 40–41. The concept of the fundamental attribution error dates to observations of Fritz Heider, The Psychology of Interpersonal Relations (London: Lawrence Erlbaum, 1958). An early, classic study of the fundamental attribution bias is E. E. Jones and V. A. Harris, “The Attribution of Attitudes,” Journal of Experimental Psychology 3 (1967): 1–24. 90. Paul D’Anieri, International Politics, 2nd ed. (Boston: Wadsworth/Cengage, 2012), 168. 91. Kahneman, Thinking, Fast and Slow, 199–200. Since the sequence in which we notice things matters, the halo effect increases the weight or importance of our first impressions of other people, groups, or countries. Kahneman, Thinking, Fast and Slow, 83. 92. Cottam et al., Introduction to Political Psychology, 41. For the halo effect, also known as the positivity effect, see S. Plous, The Psychology of Judgment and Decision Making (New York: McGraw-Hill, 1993). 93. Kahneman, Thinking, Fast and Slow, 75. 94. Kahneman, Thinking, Fast and Slow, 79. This search for causality begins at a very early age in humans. Apparently, even six-month-old infants will interpret a sequence of events as a cause and effect relationship in which the first act in the sequence causes the second. Kahneman, Thinking Fast and Slow, 76. 95. Kahneman, Thinking, Fast and Slow, 62. 96. Kahneman, Thinking, Fast and Slow, 87. 97. Kahneman, Thinking, Fast and Slow, 114. 98. Kahneman, Thinking, Fast and Slow, 81. 99. Kahneman, Thinking, Fast and Slow, 122. 100. Kahneman, Thinking, Fast and Slow, 153. 101. Jervis, Perception and Misperception, 217. 102. Steinbruner, The Cybernetic Theory of Decision, 116. To some degree, reliance on historical analogies is underpinned by the principle of representativeness—the well-known cognitive tendency to “exaggerate similarities between one event and a prior class of events.” Stein, “Foreign Policy Decision Making,” 107–8; see also Rosati, “The Power of Human Cognition,” 63. 103. Stein, “Foreign Policy Decision-Making,” 104. 104. Ernest May, “Lessons” of the Past: The Use and Misuse of History in American Foreign Policy (New York: Oxford University Press, 1973); Paul Diesing and Glen Snyder, Systems, Bargains, Decisions (Princeton, NJ: Princeton University Press, 1977), 321. 105. Jervis, Perception and Misperception, 252. 106. Jack S. Levy and William R. Thompson, Causes of War (Chichester, UK: Wiley-Blackwell, 2010), 147. 107. Garry Wills, Reagan’s America (New York: Penguin, 1985), 286–307. 108. See Simon Sebag Montefiore, Young Stalin (New York: Knopf, 2007), and Stalin: The Court of the Red Tsar (New York: Vintage/Random House, 2003). 109. Jervis, Perception and Misperception, 238. 110. Harry S. Truman, Memoirs, II (Garden City, NY: Doubleday, 1956), quoted in Ernest May, “Lessons” of the Past, 81–82. 111. The following draws on May, “Lessons” of the Past, 91–101. 112. Kennedy and his advisers also saw the situation in Vietnam as symbolically similar to that faced by the Truman administration with regard to China in 1949. Here the lesson was that if Vietnam were to go communist, the administration on whose watch this occurred would be blamed for the “loss” and would suffer politically in a country where being “soft on communism” was the kiss of death. 113. Pentagon Papers, II, ed. Sen. Gravel (Boston: Beacon Press, 1971), 650, quoted in May, “Lessons” of the Past, 98–99. 114. Jervis, Perception and Misperception, 244. 115. See Greg Cashman and Leonard C. Robinson, An Introduction to the Causes of War: Patterns of Interstate Conflict from World War I to Iraq (Lanham, MD: Rowman & Littlefield, 2007), chapter 7, “Iraq”, esp. 333–50. 116. Yuen Foong Khong, Analogies at War: Korea, Munich, Dien Bien Phu, and the Vietnam Decision of 1965 (Princeton: Princeton University Press, 1992), 10, 21. 117. Khong, Analogies at War, 253. 118. Khong, Analogies at War, 24–28. See also Deborah Welch Larson, “The Role of Belief Systems and Schemas in Foreign Policy Decision-Making,” Political Psychology 15, no. 1 (1994): 17–33.

502

Notes

119. Khong, Analogies at War, 25–26. The primary difference between the two seems to be that the term schema is used to refer to a more abstract and general theory about how the world works, while analogies are more concrete and specific. Nevertheless, many scholars use the terms more or less interchangeably. 120. Khong, Analogies at War, 245. 121. Khong, Analogies at War, 28–29. 122. As Khong notes, the Munich/1930s analogy and the Korean analogy were somewhat fused together and were mutually reinforcing, but this sixth point is where the Koran analogy departs from the Munich analogy. 123. Khong, Analogies at War, 170. 124. The perseverance effect is largely due to the fact that the use of schemas is a top-down process that is easily able to withstand information deemed to be discrepant or irrelevant, not out of any need for cognitive balance per se. 125. Khong, Analogies at War, 35–36. 126. D. Kahneman and A. Tversky, “Prospect Theory: An Analysis of Decision Making Under Risk,” Econometrica 47 (1979): 263–91; Kahneman and Tversky, eds., Choices, Values, and Frames (Cambridge: Cambridge University Press, 2000); and A. Tversky and D. Kahneman, “Advances in Prospect Theory: Cumulative Representation of Uncertainty,” Journal of Risk and Uncertainty 5 (1992): 297–323. 127. Jack Levy, “Prospect Theory, Rational Choice, and International Relations,” International Studies Quarterly 41 (1997): 87–112; Levy, “An Introduction to Prospect Theory,” in Avoiding Losses/Taking Risks: Prospect Theory and International Conflict, ed. Barbara Farnham (Ann Arbor: University of Michigan, 1994), 7–22; Levy, “Prospect Theory and International Relations: Theoretical Applications and Analytical Problems,” Political Psychology 13, no. 2 (1992): 283–310; Levy, “Prospect Theory and International Relations: Theoretical Applications and Analytical Problems,” in Farnham, Avoiding Losses/Taking Risks, 119–46; Levy, “Loss Aversion, Framing Effects and International Conflict,” in Handbook of War Studies II, ed. Manus Midlarsky (Ann Arbor: University of Michigan Press, 2000), 193–221. See also Jeffrey Taliaferro, Balancing Risks: Great Power Intervention in the Periphery (Ithaca, NY: Cornell University Press, 2004), 29–30, for a good summary. 128. Kahneman, Thinking, Fast and Slow, 282. 129. Levy, “Prospect Theory, Rational Choice, and International Relations,” 89. 130. McDermott, Risk-Taking in International Politics, 29. Kahneman suggests that the loss aversion ratio—the gain needed to balance a loss—is typically in the range of 1.5 to 2.5 to 1. Kahneman, Thinking Fast and Slow, 284. 131. R. Thaler, “Towards a Positive Theory of Consumer Choice,” Journal of Economic Behavior and Organization 1 (1980): 39–60. 132. Actually, this status quo bias is more accurately a reference point bias. Levy, “Prospect Theory, Rational Choice, and International Relations,” 90–91. 133. Kahneman, Thinking, Fast and Slow, 282; see also 287, 305. He argues that the fact that among territorial animals, the defenders almost always win, suggests that animals (including humans) probably fight harder to prevent losses than to accumulate gains. 134. Levy, “Prospect Theory and International Relations,” 284. The magnitude of the loss does not need to be particularly high; the fact that a loss (any loss) exists is more important than the magnitude. Levy, “Prospect Theory and International Relations,” 284, drawing on Robert Jervis, The Meaning of Nuclear Revolution (Ithaca, NY: Cornell University Press, 1989), 170. 135. Levy, “An Introduction to Prospect Theory,” 13. As he notes elsewhere, “framing is a variable.” Levy, “Loss Aversion, Framing Effects, and International Conflict,” 197. 136. McDermott, Risk-Taking in International Politics, 24–25. 137. Robert Jervis, “Political Implications of Loss Aversion,” in Avoiding Losses/Taking Risks, ed. Barbara Farnham (Ann Arbor: University of Michigan, 1994), 23–40. Quote is from 24. One of the problems regarding framing is that prospect theory does not provide as yet any neat theoretical explanation for how individuals might choose a reference point in order to frame a problem. 138. Levy, “Prospect Theory, Rational Choice, and International Relations,” 91. 139. Levy and Thompson, Causes of War, 152. 140. Levy, “Loss Aversion, Framing Effects, and International Conflict,” 202. 141. Levy, “Loss Aversion, Framing Effects, and International Conflict,” 198. 142. Levy and Thompson, Causes of War, 150. 143. Jervis, “Political Implications of Loss Aversion,” 25. 144. Levy, “Prospect Theory and International Relations,” 287. 145. Rose McDermott, Risk-Taking in International Politics (Ann Arbor: University of Michigan Press, 1998). 146. Kahneman, Thinking, Fast and Slow, 143, 315–16.

Notes

503

147. McDermott argues this effect is worsened if leaders use analogies linking their opponents to historical bogeymen like Hitler. Rose McDermott, Risk-Taking in International Politics, 30–33. 148. This is sometimes referred to as the “1% Doctrine,” a term put forward in Ron Suskind, The One Percent Doctrine (New York: Simon & Schuster, 2006). 149. Jervis, “Political Implications of Loss Aversion,” 31. 150. Audrey McInerney, “Prospect Theory and Soviet Policy Towards Syria, 1966–1967,” in Avoiding Losses/ Taking Risks, ed. Barbara Farnham (Ann Arbor: University of Michigan Press, 1994), 101–18. 151. McInerney, “Prospect Theory and Soviet Policy Towards Syria,” 116. 152. Charles J. Sullivan, “The Kremlin and Kabul: The 1979 Soviet Invasion of Afghanistan in Retrospect,” The Washington Review of Turkish and Eurasian Affairs, September, 2011. http://www.thewashingtonreview.org/ articles/the-kremlin-and-kabul-the-1979-soviet-invasion-of-afghanistan-in-retrospect.html. 153. J. G. Stein, “Calculation, Miscalculation, and Conventional Deterrence I: The View from Cairo,” in Psychology and Deterrence, ed. R. Jervis, R. N. Lebow, and J. G. Stein (Baltimore: Johns Hopkins University Press, 1985), 34–59. 154. Taliaferro, Balancing Risks, 37. 155. Taliaferro, Balancing Risks, 51–52. 156. It is probably true that while the decision to proceed north of the parallel was riskier than the alternative of simply stopping the military effort at the 38th parallel, many civilian and military leaders believed the risks were fairly low: the North Korean forces were clearly routed and on the run, the Soviet Union had shown no signs of being willing to intervene, the Chinese had threatened to intervene but appeared weak, making their threat less than fully credible. A military victory was seen as relatively easy in this context. 157. Taliaferro, Balancing Risks, 21–22. 158. Paul Kowert and Margaret Hermann, “Who Takes Risks? Daring and Caution in Foreign Policy Making,” Journal of Conflict Resolution 41, no. 5 (October 1997): 611–37. 159. Levy, “Prospect Theory and International Relations,” 305. 160. Yaacov Vertzberger, “Rethinking and Reconceptualizing Risk in Foreign Policy Decision Making.” 161. Yaacov Vertzberger, Risk Taking and Decisionmaking (Stanford: Stanford University Press, 1998), 77. 162. Kowert and Hermann, “Who Takes Risks?” 625. 163. Kowert and Hermann, “Who Takes Risks?” 631–32. 164. Mercer, “Prospect Theory and Political Science,” 10. 165. Jervis suggests that prospect theory is most useful in explaining those situations in which leaders persevere in deteriorating situations in which they face major losses if they choose to do nothing. LBJ’s escalation of the war in Vietnam and the Japanese attack on Pearl Harbor fit fairly well here. So does Sadat’s decision for war in 1973. Note that in each case, domestic political losses were as important as international losses. Robert Jervis, “Political Implications of Loss Aversion.” On the Japanese decision, see Cashman and Robinson, An Introduction to the Causes of War, chapter 3, and also Jeffrey Taliaferro, Balancing Risks: Great Power Intervention in the Periphery (Ithaca, NY: Cornell University Press, 2004), chapter 4. On the 1973 war, see Stein, “Calculations, Miscalculations, and Conventional Deterrence.” 166. Levy and Thompson, Causes of War, 138; Vertzberger, Risk Taking and Decisionmaking, 77. 167. Definitions of concepts such as images, perceptions, and schemas vary from author to author. This definition of image is a matter of personal taste and should not be considered as definitive. 168. Ole Holsti, “Foreign Policy Decision Makers Viewed Psychologically,” 122. 169. Charles Duelfer and Stephen Benedict Dyson, “Chronic Misperception and International Conflict: The U.S.Iraq Experience,” International Security 36, no. 1 (Summer 2011), 77–100; quote from 76. 170. Harold Sprout and Margaret Sprout, The Ecological Perspective on Human Affairs (Princeton, NJ: Princeton University Press, 1965). 171. Alexander George, “The Operational Code: A Neglected Approach to the Study of Political Leaders and Decision-Making,” in The Conduct of Soviet Foreign Policy, ed. Erik Hoffman and Frederick Fleron (New York: Aldine, 1980), 170. 172. Nathan Leites, A Study of Bolshevism (Glencoe, IL: Free Press, 1953). 173. George, “The Operational Code,” 172–73, and Ole Holsti, “Foreign Policy Decision Makers Viewed Psychologically,” 138. 174. See Starr, Henry Kissinger, 45–47. 175. On Dulles, see Ole Holsti, “The ‘Operational Code’ Approach to the Study of Political Leaders: John Foster Dulles’ Philosophical and Instrumental Beliefs,” Canadian Journal of Political Science 3 (1970): 123–57; for Soviet

504

Notes

leaders, see Douglas Blum, “The Soviet Foreign Policy Belief System: Beliefs, Politics, and Foreign Policy Outcomes,” International Studies Quarterly 37 (1993): 373–94; for Mao, see Huiyun Feng, “The Operational Code of Mao Zedong: Defensive or Offensive Realist?” Security Studies 14, no. 4 (2005): 637–62; for Rabin and Peres, see Scott Crichlow, “Idealism or Pragmatism? An Operational Code Analysis of Yitzhak Rabin and Shimon Peres,” Political Psychology 19, no. 4 (1998): 683–706; for LBJ, see Stephen Walker and Mark Schafer, “The Political Universe of Lyndon B. Johnson and His Advisers: Diagnostic and Strategic Propensities in Their Operational Codes,” Political Psychology 21, no. 3 (2000): 529–43; for Jimmy Carter, see Stephen Walker, Mark Schafer, and Michael Young, “Systematic Procedures for Operational Code Analysis: Measuring and Modeling Jimmy Carter’s Operational Code,” International Studies Quarterly 42, no. 1 (1998): 175–89; for G. H. W. Bush and Clinton, see Stephen Walker, Mark Schafer, and Michael Young, “Presidential Operational Codes and Foreign Policy Conflicts in the Post-Cold War World,” Journal of Conflict Resolution 43, no. 5 (October 1999): 610–25; for G. W. Bush, see Jonathan Renshon, “Stability and Change in Belief Systems: The Operational Code of George W. Bush,” Journal of Conflict Resolution 52, no. 6 (December 2008): 820–49. 176. Stephen G. Walker, “The Interface between Beliefs and Behavior: Henry Kissinger’s Operational Code and the Vietnam War,” Journal of Conflict Resolution 21, no. 1 (March 1977): 151–52. 177. Starr, Henry Kissinger, 159. 178. Robert Jervis, “Hypotheses on Misperception,” in International Politics and Foreign Policy, rev. ed., ed. James Rosenau (New York: Free Press, 1969), 240. 179. Margaret Herman, “Assessing Foreign Policy Behavior Using the Personal Characteristics of Political Leaders,” International Studies Quarterly 24 (1980): 7–46; P. Suedfeld and P. Tetlock, “Integrative Complexity of Communications in International Crises,” Journal of Conflict Resolution 21 (1977): 169–84; P. Suedfeld, P. Tetlock, and C. Ramirez, “War, Peace and Integrative Complexity,” Journal of Conflict Resolution 21 (1977): 427–42; P. Tetlock, “Identifying Victims of Groupthink from Public Statements of Decision-Makers,” Journal of Personality and Social Psychology 37 (1979): 1314–24; and the summary in D. G. Winter, “Assessing Leaders’ Personalities,” 27–29. 180. The classic work is Leon Festinger, A Theory of Cognitive Dissonance (Evanston, IL: Row, Patterson, 1957). 181. Ole Holsti, “Cognitive Dynamics and Images of the Enemy,” in Image and Reality in World Politics, ed. John Farrell and Asa Smith (New York: Columbia University Press, 1967). The tendency toward cognitive dissonance affects not only how new information is processed, but also how the images in our belief system are organized. For instance, some beliefs are created and maintained not on the basis of direct evidence but because they are logically connected to other established beliefs: Consistency of belief requires they be present. On this point see John Steinbruner, The Cybernetic Theory of Decision (Princeton, NJ: Princeton University Press, 1974), 97 and 114. 182. Charles Hanley, “Half of U.S. Still Believes Iraq had WMD,” Washington Post, August 7, 2006, http://www. washingtonpost.com/wp-dyn/content/article/2006/08/07/AR2006080700189.html. 183. Shankar Vendantam, “Persistence of Myths Could Alter Public Policy Approach,” Washington Post, September 4, 2007, A3. The number declined by 2010, but was still significant. See, for instance, “September 11 Attacks Opinion Polls,” http://en.wikipedia.org/wiki/september_11_attacks_opinion_polls. 184. Kahneman, Thinking, Fast and Slow, 302. 185. Vedantam, “Persistence of Myths.” 186. Steinbruner, The Cybernetic Theory of Decision, 100–101. See also Dina Zinnes, “Some Evidence Relevant to the Man-Milieu Hypothesis,” in The Analysis of International Politics, ed. J. Rosenau, V. Davis, and M. East (New York: Free Press, 1972), 245. 187. Steinbruner, The Cybernetic Theory of Decision, 102. 188. See Rosati, “The Power of Human Cognition,” 56–59. Jonathan Renshon’s analysis of operational codes suggests that in fact beliefs are not internally consistent; however, this may be due in part to the way that operational code indices are constructed using the Verbs in Context System. J. Renshon, “Stability and Change in Belief Systems,” 840. 189. Larson, “Belief Systems and Schemas,” 25–27. Larson draws on S. T. Fiske and S. E. Taylor, Social Cognition, 2nd ed. (New York: McGraw-Hill, 1991), and H. Markus and R. B. Zajonc, “The Cognitive Perspective in Social Psychology,” in The Handbook of Social Psychology, 3rd ed., ed. G. Lindzey and E. Aronson (New York: Random House 1985), 137–230. 190. Khong, Analogies at War, 246, notes that it took two years of contrary evidence from the battlefield in Vietnam for McNamara to abandon his belief of the efficacy of aerial bombing (based on the 1944 bombing of Germany analogy), while nothing was able to change Walt Rostow’s support for the WWII bombing analogy. 191. Rosati, “The Power of Human Cognition,” 57–58.

Notes

505

192. Steinbruner, The Cybernetic Theory of Decision, 124–36. 193. Steinbruner, The Cybernetic Theory of Decision, 129. 194. Larson, “Belief Systems and Schemas,” 28. Larson is drawing on evidence in Fiske and Taylor, Social Cognition; and J. Crocker, S. T. Fiske, and S. E. Taylor, “Schematic Bases of Belief Change,” in Attitudinal Judgment, ed. J. R. Eiser (New York: Springer-Verlag, 1984), 197–226. 195. Philip Tetlock, Expert Political Judgment: How Good Is It? How Can We Know? (Princeton: Princeton University Press, 2005). 196. See, for instance, Charles Krauthammer, “No, The Cold War Isn’t Really Over,” Time, September 5, 1988. 197. He then became director of the CIA from 1991 to 1993 and was appointed secretary of defense for both George W. Bush and Barack Obama from 2006 to 2011, after a hiatus from government as the president of the University of Texas. 198. Robert Kaiser, “Documents Detail U.S. Intelligence Insights, Mistakes,” Washington Post, March 11, 2001, A15. 199. Kaiser, “Documents Detail U.S. Intelligence Insights, Mistakes.” 200. See Ole Holsti, “Cognitive Dynamics and Images of the Enemy,” and Holsti, “The Belief System and National Images: A Case Study,” in International Politics and Foreign Policy, 543–50. 201. Henry Kissinger, The Necessity for Choice (New York: Doubleday, 1962), 210. 202. Jervis, Perception and Misperception, 308. 203. Karl Deutsch and Richard Merritt, “Effects of Events on National and International Images,” in International Behavior, ed. H. Kelman, 132–87. Learning may occur because of major successes as well as failures. See Dan Reiter, Crucible of Beliefs: Learning, Alliances, and World Wars (Ithaca, NY: Cornell University Press, 1996). On foreign policy learning in general, see Jack Levy, “Learning and Foreign Policy: Sweeping a Conceptual Minefield,” International Organization 48 (1994): 279–312; George Breslauer and Philip Tetlock, eds. Learning in U.S. and Soviet Foreign Policy (Boulder, CO: Westview Press, 1971); and Jerel Rosati, Joe Hagan, and Martin Sampson, eds. Foreign Policy Restructuring: How Governments Respond to Global Change (Columbia, SC: University of South Carolina Press, 1994). 204. Stein, “Foreign Policy Decision-Making,” 106. 205. Renshon, “Stability and Change in Belief Systems.” 206. See the discussion in Stein, “Foreign Policy Decision-Making,” 106–7. 207. Cashman and Robinson, An Introduction to the Causes of War, 335–36. 208. Stephen Walker, Mark Schafer, and Michael Young, “Systematic Procedures for Operational Code Analysis: Measuring and Modeling Jimmy Carter’s Operational Code,” International Studies Quarterly 42, no. 1 (1998): 175–89. On the other hand, traumatic events don’t always lead to a reversal of beliefs. Feng found that the Korean War simply reinforced and strengthened all five of Mao Zedong’s philosophical indexes, making them even more hostile. Huiyun Feng, “The Operational Code of Mao Zedong.” Tetlock argues that external shocks and dramatic events may trigger a kind of defense mechanism that consolidates preexisting beliefs rather than changing them. Phillip Tetlock, Expert Political Judgment: How Good Is It? How Can We Know? (Princeton: Princeton University Press, 2005), 131–32. 209. Jervis, Perception and Misperception, 191. 210. Jervis, Perception and Misperception, 145. 211. Jervis, Perception and Misperception, 208. 212. Nancy Cooper and John Barry, “Seven Minutes to Death,” Newsweek, July 18, 1988, 18–24. A similar situation occurred in April 1989 when an Iraqi aircraft erroneously shot down a friendly Egyptian training jet armed with missiles that was flying to an international air show in Baghdad. The Iraqi air defense command had been at a heightened state of alert for several weeks after Israeli officials publically announced their concern over Iraq’s nuclear power program. Additionally, Iraqi air defenses had become accustomed to a shoot-first strategy during the Iran-Iraq War. Patrick Tyler, “Iraqis May Have Thought Downed Jet Was Israeli,” Washington Post, April 29, 1989, A16 and A26. 213. Dominic D. P. Johnson and Dominic Tierney, “The Rubicon Theory of War: How the Path to Conflict Reaches the Point of No Return,” International Security 36, no. 1 (Summer 2011): 7–40. The theory is based on Heinz Heckhausen and Peter Gollwitzer, “Thought Contents and Cognitive Functioning in Motivational versus Volitional States of Mind,” Motivation and Emotion 11, no. 2 (June 1987): 101–20. 214. Jack S. Levy, “Misperception and the Causes of War: Theoretical Linkages and Analytical Problems,” World Politics 36, no. 1 (October 1983): 76–99. 215. Cashman and Robinson, An Introduction to the Causes of War, 376.

506

Notes

216. Jack Levy, “Misperceptions and the Causes of War: Theoretical Linkages and Analytical Problems,” World Politics 36, no. 1 (1983): 76–99; Levy and Thompson, Causes of War, 135. 217. Quoted in Jervis, Perception and Misperception, 74. 218. Stoessinger, Why Nations Go to War, 211. 219. Robert North, “Perception and Action in the 1914 Crisis,” in Image and Reality in World Politics, 122. 220. Levy, “Misperception and the Causes of War,” 88–89. 221. Lebow, Between Peace and War, 200. 222. Charles Duelfer and Stephen Benedict Dyson, “Chronic Misperception and International Conflict: The U.S.Iraq Experience,” International Security 36, no. 1 (Summer 2011): 73–100. 223. Geoffrey Blainey, The Causes of War (New York: Free Press, 1973), 246. 224. Lebow, Between Peace and War, 242–43. 225. William J. Barnds, India, Pakistan and the Great Powers (New York: Praeger, 1972), 200, cited in Stoessinger, Why Nations Go to War, 125–26. 226. Lebow, Between Peace and War, 245–46. 227. Levy, “Misperception and the Causes of War,” 84. See also Dominic Johnson, Overconfidence and War: The Havoc and Glory of Positive Illusions (Cambridge, MA: Harvard University Press, 2004). 228. See Thomas Ricks, Fiasco: The American Military Adventure in Iraq (New York: Penguin Press, 2006). 229. Lebow, Between Peace and War, 62 and 83. 230. Lebow, Between Peace and War, 238–41. 231. Lebow, Between Peace and War, 97. 232. Lebow, Between Peace and War, 164–66. 233. Jervis, Perception and Misperception, 52. 234. Lebow, Between Peace and War, 158–61. 235. Joseph de Rivera, The Psychological Dimension of Foreign Policy (Columbus, OH: Charles Merrill, 1968), 247–57. 236. Lebow, Between Peace and War, 157–78. 237. Evan Luard, War in International Society (New Haven, CT: Yale University Press, 1986), 329. 238. Luard, War in International Society, chapter 8, 329–78. 239. See John Mueller, Retreat from Doomsday: The Obsolescence of Major War (New York: Basic Books, 1989); and also Steven Pinker, The Better Angels of Our Nature: Why Violence Has Declined (New York: Viking, 2011). 240. Levy, “Misperceptions and the Causes of War,” 81. 241. For instance, Barbara Tuchman, The Guns of August (New York: Dell, 1962). 242. Lebow, Between Peace and War, 254; and Fritz Fischer, War of Illusions: German Policies from 1911 to 1914 (New York: Norton, 1975), 45, 55, and 398–402. 243. Luard, War in International Society, 360–61. 244. Lebow, Between Peace and War, 264–65. 245. On these points see Blainey, The Causes of War, 208–9, and Tuchman, The Guns of August, 45–62 and 142. 246. Stephen Van Evera, The Causes of War: Power and the Roots of Conflict (Ithaca, NY: Cornell University Press, 1999). 247. Blainey, The Causes of War, 57–67. 248. Blainey, The Causes of War, 57–67, and Levy, “Misperception and the Causes of War,” 91–93. For the Soviet, Chinese, and North Korean perceptions of the situation, see Sergei N. Goncharov, John W. Lewis, and Xue Litai, Uncertain Partners: Stalin, Mao, and the Korean War (Stanford, CA: Stanford University Press, 1993). 249. Washington Post, August 4, 1990, and October 21, 1990. 250. Blainey, The Causes of War, 53. 251. Lebow, Between Peace and War, 102. 252. Irving Janis and Leon Mann, Decision-Making: A Psychological Analysis of Conflict, Choice and Commitment (New York: Free Press, 1977). 253. This discussion of Janis and Mann is based on Lebow, Between Peace and War, 107–10. 254. Lebow, Between Peace and War, 111. 255. Lebow, Between Peace and War, 275–76. 256. Jonathan Renshon, Why Leaders Choose War: The Psychology of Prevention (Westport, CT: Praeger, 2006). 257. Renshon, Why Leaders Choose War, 143. 258. Renshon, Why Leaders Choose War, 151–53. 259. Renshon, Why Leaders Choose War, 145.

Notes

507

260. Renshon, Why Leaders Choose War, 159. 261. Renshon, Why Leaders Choose War, 146. 262. Renshon, Why Leaders Choose War, 147. Several factors were deemed to be less useful in explaining preventive war. Declining power was seen as less important than the perception of declining relative power. Indeed, a real decline in material power was not present between Britain and Egypt in 1956, Israel and Iraq in 1981, or the U.S. and Iraq in 2003. Perceptions of closing windows of opportunity appeared important only to the extent that they interact with perceptions of declining power and the belief that conflict is inevitable. In the Suez Crisis there was no evidence that British leaders thought in terms of windows at all. (And of course in the U.S.-Soviet early Cold War situation and the India-Pakistan relationship, windows were clearly perceived, but leaders still opted against war. Finally, perceptions of offensive advantage did not seem important. In two of the three cases of preventive war, such perceptions were absent. 263. Jonathan Renshon and Stanley Renshon, “The Theory and Practice of Foreign Policy Decision Making,” Political Psychology 29, no. 4 (2008): 512; Yaacov Vertzberger, Risk Taking and Decisionmaking (Stanford, CT: Stanford University Press, 1998), 75. 264. Post, “The Impact of Crisis-Induced Stress on Policy Makers,” 486. 265. McDermott, “The Feeling of Rationality,” 697. 266. Thomas Wiegele, “Decision-Making in an International Crisis: Some Biological Factors,” International Studies Quarterly 17, no. 3 (September 1973): 302–4. 267. Sally Squires, “Stress, Poor Communications Cited in Vincennes’ Downing of Jet,” Washington Post, October 17, 1988. 268. From Jerrold Post, “The Impact of Crisis-Induced Stress on Policy Makers,” in Avoiding War: Problems of Crisis Management, ed. Alexander George (Boulder, CO: Westview Press, 1991), 471–94; Yaacov Vertzberger, Risk Taking and Decisionmaking (Stanford, CA: Stanford University Press, 1998); Jonathan Renshon and Stanley Renshon, “The Theory and Practice of Foreign Policy Decision Making,” Political Psychology 29, no. 4 (2008): 509–36; Barry Schneider, Danger and Opportunity: Decision Making, Bargaining, and Management in Three United States and Six Simulated Crises, Ph.D. dissertation, Columbia University, New York, 1974, 67–68, cited in Patrick Morgan, Deterrence (Beverly Hills, CA: Sage, 1977), 186–87; Donald Kinder and Janet Weiss, “In Lieu of Rationality,” Journal of Conflict Resolution 22, no. 4 (December 1978): 722; Ole Holsti, “Theories of Crisis Decision Making,” in International Relations Theory, ed. Paul Viotti and Mark Kauppi (New York: Macmillan, 1987), 244–81; Ole Holsti, “Crisis Management,” in Psychological Dimensions of War, ed. Betty Glad (Newbury Park, CA: Sage, 1990), 116–42; Ole Holsti and Alexander George, “The Effects of Stress on the Performance of Foreign Policy-Makers,” in Political Science Annual, vol. 6, Individual Decision Making, ed. C. P. Cotter (Indianapolis, IN: Bobbs-Merrill, 1975), 255–319; G. Keinan, N. Friedland, and Y. Ben-Porath, “Decision-Making Under Stress: Scanning of Alternatives Under Physical Threat,” Acta Psychologica 64 (1987): 219–28. 269. Ole Holsti, Crisis, Escalation, War (Montreal: McGill-Queens University Press, 1972), 199–200. 270. Crisis situations create other problems that impair decision-making that are not directly related to stress. Crises frequently reduce the size of the policy-making group, further decreasing the likelihood that leaders will get wide-ranging advice and analysis. Leaders also tend to rely on ad hoc communication channels and reduce their communication with adversaries during a crisis. Rosati, “The Power of Human Cognition,” 69–70 271. Post, “The Impact of Stress on Policy Makers,” 473. 272. Uri Bar-Joseph and Rose McDermott, “Personal Functioning under Stress: Accountability and Social Support of Israeli Leaders in the Yom Kippur War,” Journal of Conflict Resolution 52, no. 1 (February 2008): 144–70. 273. Post, “The Impact of Stress on Policy Makers,” 490. 274. Post, “The Impact of Stress on Policy Makers,” 476–83. 275. Vertzberger, Risk Taking and Decisionmaking, 75; Holsti, “Crisis Management,” 127. 276. Brecher and Wilkenfeld, A Study of Crisis, 171–84. 277. Vertzberger, Risk Taking and Decisionmaking, 395. 278. See Mercer, “Rationality and Psychology in International Politics,” 87–88. 279. Herbert C. Kelman, “Social-Psychological Approaches to the Study of International Relations,” in International Behavior: A Social-Psychological Analysis, ed. H. Kelman (New York: Holt, Rinehart & Winston, 1965).

508

Notes

4. THE SUBSTATE LEVEL OF ANALYSIS 1. It is also assumed that one’s preferences are consistent, in the sense that they are ranked in a transitive order. That is, if A is preferred to B, and B is preferred to C, then A will be preferred to C. 2. See Jack S. Levy and William R. Thompson, Causes of War (Malden, MA: Wiley-Blackwell, 2010), 131–32. 3. Paul D’Anieri, International Politics, 2nd ed. (Boston: Wadsworth, Cengage, 2012), 156. Technically, according to rational choice theorists, choice involves rather difficult computations: (1) Identify all the possible consequences or outcomes for each policy option. (2) Determine the probability that each outcome will occur. (3) Calculate the utility (net benefits minus costs) of each potential outcome. (4) Weigh (multiply) the utility of each outcome by its probability. And (5) determine the expected utility of each option by adding together all the weighted sums of each possible outcome for the option. The option with the highest expected utility should be your choice. Levy and Thompson, Causes of War, 132. 4. Bruce Bueno de Mesquita, “Foreign Policy Analysis and Rational Choice Models,” http://www.isacompass. com/info/samples/foreignpolicyanalysisandrationalchoicemodels_sample.pdf, 4, 11. 5. James Fearon, “Domestic Political Audiences and the Escalation of International Disputes”, American Political Science Review 88 (1994): 577–92, and Fearon, “Rationalist Explanations for War,” International Organization 49 (1995): 379–414. 6. See Bruce Bueno de Mesquita, The War Trap (New Haven, CT: Yale University Press, 1981). Bueno de Mesquita sees governments as unitary decision makers who approach the problems of war and peace as expected utility maximizers. 7. Bueno de Mesquita, The War Trap, 29. 8. Bueno de Mesquita, The War Trap , 46. The expected utility of a particular war strategy is a function of the sum of the utilities of the possible outcomes times their probabilities. 9. Bueno de Mesquita, The War Trap, 182–83. 10. Bruce Bueno de Mesquita and David Lalman, War and Reason (New Haven, CT: Yale University Press, 1992). 11. D. Scott Bennett and Allan C. Stam, “A Universal Test of an Expected Utility Theory of War,” International Studies Quarterly 44 (2000): 451–80. See also Bennett and Stam, “A Cross-Validation of Bueno de Mesquita and Lalman’s International Interaction Game,” British Journal of Political Science 30 (200): 541–60. 12. Bruce Bueno de Mesquita, Alastair Smith, Randolph Sivereson, and James Morrow, The Logic of Political Survival (Cambridge, MA: MIT Press, 2003). 13. See Paul D’Anieri, International Politics: Power and Purpose in Global Affairs, 2nd ed. (Boston: Wadsworth, 2012), 186. 14. Philip Tetlock, Expert Political Judgment: How Good Is It? How Can We Know? (Princeton, NJ: Princeton University Press, 2005). 15. Charles W. Kegley Jr., and Eugene R. Wittkopf, American Foreign Policy: Patterns and Process, 3rd ed. (New York: St. Martin’s, 1987), 472. 16. McGeorge Bundy, Danger and Survival: Choices about the Bomb in the First Fifty Years (New York: Random House, 1988), 451. That Khrushchev’s decision was initially made with a very small circle of advisers and that he failed to consult diplomats and those who were knowledgeable about the United States is confirmed by William Taubman, Khrushchev: The Man and His Era (New York: Norton, 2003), 541–45. 17. Glenn H. Snyder and Paul Diesing, Conflict among Nations: Bargaining, Decision Making, and System Structure in International Crises (Princeton, NJ: Princeton University Press, 1977), 369. 18. Paul A. Anderson, “What Do Decision Makers Do When They Make Foreign Policy? The Implications for the Comparative Study of Foreign Policy,” in New Directions in the Study of Foreign Policy, ed. Charles F. Hermann, Charles W. Kegley Jr., and James N. Rosenau (Boston: Allen & Unwin, 1987), 285–308. 19. Herbert Simon, “Human Nature in Politics: The Dialogue of Psychology with Political Science,” American Political Science Review 79, no. 2 (June 1985): 293–304. 20. Simon, “Human Nature in Politics,” 295. 21. The foundational works are Herbert Simon, Administrative Behavior (New York: Macmillan, 1959), and Simon, Models of Man (New York: Wiley, 1957); James March and Herbert Simon, Organizations (New York: Wiley, 1958); and Richard Cyert and James March, A Behavioral Theory of the Firm (Englewood Cliffs, NJ: Prentice-Hall, 1963).

Notes

509

22. If no appropriate alternative appears, decision makers then lower the level of aspiration and review the options again. See Snyder and Diesing, Conflict among Nations, 344. 23. See Snyder and Diesing, Conflict among Nations, 342. 24. Anderson, “What Do Decision Makers Do?” 296–97. 25. David Baybrooke and Charles Lindblom, “Types of Decision-Making,” in International Politics and Foreign Policy, rev. ed., ed. James Rosenau (New York: Free Press, 1969), 207–16. 26. Leslie Gelb and Richard Betts, The Irony of Vietnam: The System Worked (Washington: Brookings Institution, 1979). 27. Gelb and Betts, The Irony of Vietnam, 278. 28. Gelb and Betts, The Irony of Vietnam, 295. 29. See Arthur Schlesinger Jr., The Bitter Heritage: Vietnam and American Democracy, 1941–1966 (Boston: Houghton Mifflin, 1967). 30. Alex Mintz, “The Decision to Attack Iraq: A Noncompensatory Theory of Decisions Making,” Journal of Conflict Resolution 37, no. 4 (1993): 595–618; A. Mintz, “How Do Leaders Make Decisions? A Poliheuristic Perspective,” Journal of Conflict Resolution 48, no. 1 (February 2004): 3–13; Alex Mintz and Nehemia Geva, “The Poliheuristic Theory of Decision,” in Decision Making on War and Peace: The Cognitive-Rational Debate, ed. N. Geva and A. Mintz (Boulder, CO: Lynne Rienner, 1997): 81–101; see also the chapters in Alex Mintz, ed. Integrating Cognitive and Rational Theories of Foreign Policy Decision Making (New York: Palgrave Macmillan, 2003). 31. Mintz, “How Do Leaders Make Decisions?” 8. 32. Alex Mintz, “Integrating Cognitive and Rational Theories of Foreign Policy Decision Making: A Poliheuristic Perspective,” in Integrating Cognitive and Rational Theories of Foreign Policy Decision Making, 1–9. See also P. James and E. Zhang, “Chinese Choices: A Poliheuristic Analysis of Foreign Policy Crises, 1950–1996,” Foreign Policy Analysis 1 (2005): 31–54. 33. Karl DeRouen Jr., “The Decision Not to Use Force at Dien Bien Phu: A Poliheuristic Perspective,” in Integrating Cognitive and Rational Theories of Foreign Policy Decision Making, 11–28. 34. Steven B. Redd, “The Influence of Advisers and Decisions Strategies on Foreign Policy Choices: President Clinton’s Decision to Use Force in Kosovo,” International Studies Perspectives 6 (2005): 129–50. 35. Alex Mintz and Karl DeRouen Jr., Understanding Foreign Policy Decision Making (New York: Cambridge University Press, 2010), 79. 36. As Mintz points out, “Policymakers typically expect the use of force to help them politically (at least in the short run) through the ‘rally ‘round the flag’ effect, and thus the two dimensions (political and military) are not always orthogonal (independent) of each other. Mintz, “The Decision to Attack Iraq,” 602. 37. Mintz, “The Decision to Attack Iraq.” 38. Mintz and DeRouen, Understanding Foreign Policy Decision Making, 81–86. 39. DeRouen, “The Decision Not to Use Force at Dien Bien Phu.” 40. Karl DeRouen, Jr., Politics, Economics, and Presidential Use of Force Decision Making (Lewiston, NY: Edwin Mellen Press, 2001). 41. Kanishkan Sathasivam, “‘No Other Choice’: Pakistan’s Decision to Test the Bomb,” in Mintz, Integrating Cognitive and Rational Theories of Foreign Policy Decision Making. 42. Patrick James and Enyu Zhang, “Chinese Choices: A Poliheuristic Analysis of Foreign Policy Crises, 1950–1996,” Foreign Policy Analysis 1, no. 1 (March 2005): 31–54. 43. Redd, “The Influence of Advisers and Decision Strategies on Foreign Policy Choices.” 44. Jonathan W. Keller and Yi Edward Yang, “Empathy and Strategic Interaction in Crisis: A Poliheuristic Perspective,” Foreign Policy Analysis 5, no. 2 (April 2009): 169–89. 45. Keller and Yang, “Empathy and Strategic Interaction in Crises,” 173. 46. Graham Allison, Essence of Decision: Explaining the Cuban Missile Crisis (Boston: Little, Brown, 1971). The revised edition is Graham Allison and Philip Zelikow, Essence of Decision: Explaining the Cuban Missile Crisis, 2nd ed. (New York: Longman, 1999). See also Allison’s “Conceptual Models and the Cuban Missile Crisis,” American Political Science Review 63 (September 1969), 689–718; Allison and Morton Halperin, “Bureaucratic Politics: A Paradigm and Some Policy Implications,” in Theory and Policy in International Relations, ed. Raymond Tanter and Richard Ullman (Princeton, NJ: Princeton University Press, 1972), 40–79; and Morton Halperin, Bureaucratic Politics and Foreign Policy (Washington, DC: Brookings Institution, 1974). 47. Allison and Halperin, “Bureaucratic Politics: A Paradigm.” They also label the BPM a paradigm rather than a theory in this revised effort.

510

Notes

48. Allison and Zelikow refer to model II as the organizational behavior model instead of the organizational process model in their revised edition, but most scholars continue to use the original terminology. 49. See Bob Woodward, Bush at War (New York: Simon and Schuster, 2002). 50. Snyder and Diesing, Conflict among Nations, 373. 51. Allison, Essence of Decision, 79. 52. Jonathan Bendor and Thomas Hammond, “Rethinking Allison’s Models,” American Political Science Review 86, no. 2 (June 1992): 301–22; see especially 309. 53. This is largely contrary to the original tenets of organizational theory as put forward by Simon and March, who make the argument that organizations enhance decision making under bounded rationality and can, for instance, process information more effectively than individuals. The second edition of Essence of Decision takes a less pessimistic approach to organizational routines. See the argument of Bendor and Hammond, “Rethinking Allison’s Models.” 54. See David Welch, “Organizational Process and Bureaucratic Politics Paradigms: Retrospect and Prospect,” International Security 17, no. 2 (Fall 1992): 112–46; see especially 117. 55. John Steinbruner, The Cybernetic Theory of Decision (Princeton, NJ: Princeton University Press, 1974). This cybernetic tinkering is, of course, a method by which large organizations can reduce the overwhelming complexity of the policy environment. The focus on a few variables helps to reduce complexity by eliminating variety. Keeping critical variables within a tolerable range reduces both complexity and uncertainty. Complexity is further reduced through the breaking down of problems into smaller “miniproblems”—each dealt with by separate organizational subunits. 56. Allison, Essence of Decision (1971), 130. 57. Welch, “Organizational Process and Bureaucratic Politics Paradigms,” 126. 58. Scott Sagan, “The Origins of the Pacific War,” in The Origin and Prevention of Major Wars, ed. Robert Rotberg and Theodore Raab (Cambridge: Cambridge University Press, 1988), 323–52, especially 346–47; John Dower, War without Mercy: Race and Power in the Pacific War (New York: Pantheon, 1986), 260; and Greg Cashman and Leonard C. Robinson, An Introduction to the Causes of War (Lanham, MD: Rowman & Littlefield, 2007), 143. 59. Jack S. Levy and William R Thompson, Causes of War (Malden, MA: Wiley-Blackwell, 2010), 168–80. 60. Jack S. Levy, “Organizational Routines and the Causes of War,” International Studies Quarterly 30, no. 2 (June 1986): 193–222. See also Levy and Thompson, Causes of War. 61. Levy, “Organizational Routines and the Causes of War;” see also Stephen Van Evera, “The Cult of the Offensive and the Origins of the First World War,” International Security 9, no. 1 (1984): 58–107; see especially 49, 63, 86–89. 62. Levy, “Organizational Routines,” 211. 63. See Jerel Rosati, “Developing a Systematic Decision-Making Framework: Bureaucratic Politics in Perspective,” World Politics 33 (January 1981): 234–52, for a somewhat more restricted view of the central propositions of the BPM. 64. Thomas Preston and Paul ‘t Hart, “Understanding and Evaluating Bureaucratic Politics: The Nexus Between Political Leaders and Advisory Systems,” Political Psychology 20, no. 1 (March 1999): 49–98. 65. Since according to the BPM each player attempts to maximize his or her own interests—primarily organizational interests but also personal interest—the model assumes rational actors. The model is rational, but it is nonunitary. See Bendor and Hammond, “Rethinking Allison’s Models.” 66. Juliet Kaarbo and Deborah Gruenfeld, “The Social Psychology of Inter- and Intragroup Conflict in Governmental Politics,” Mershon International Studies Review 42, no. 2 (November 1998): 226–33. 67. Empirical evidence does indicate a causal relationship between attitudes and roles. See Kegley and Wittkopf, American Foreign Policy, 3rd ed., 464, n2. Conformity is especially evident in formal roles. 68. Roger Hilsman developed a variant of the BPM that he calls the political process model. While Allison assumes that the organization is the single most important determinant of the policies that players espouse, and that larger, powerful bureaucracies are the most important determinants of policy outcomes, Hilsman regards government organizations as only one factor—and not necessarily the most important. He notes the importance of factionalism within government organizations so that, for instance, certain offices within the State Department align with certain offices within the Defense Department in opposition to rival cross-institutional factions. Most important, Hilsman’s model gives a larger role to the effects of domestic politics on foreign policy and includes congressional politics as well as executive branch politics. It also gives more attention to interest groups and to the public. Roger

Notes

511

Hilsman, The Politics of Policy Making in Defense and Foreign Affairs (Englewood Cliffs, NJ: Prentice-Hall, 1987), 77–78. 69. In their study of international crises, Snyder and Diesing found that all strategies put forward for consideration could be accounted for by bureaucratic politics; Conflict among Nations, 407. 70. Charles F. Hermann, “The Impact of Single Group Decision Units on Foreign Policy,” paper presented at the International Studies Association conference, St. Louis, March 1988. See also Margaret Hermann and Charles F. Hermann, “A Look Inside the ‘Black Box’: Building on a Decade of Research,” in Biopolitics, Political Psychology and International Politics, ed. Gerald Hopple (New York: St. Martin’s, 1982). 71. Kenneth Arrow, Social Choice and Individual Values (New York: Wiley, 1951); see also Bruce Bueno de Mesquita, The War Trap, 12–18; and C. F. Hermann, “The Impact of Single Decisions Units.” 72. See Miriam Steiner’s critique of the BPM, “The Elusive Essence of Decision,” International Studies Quarterly 21, no. 2 (June 1977): 389–422. 73. Snyder and Diesing (Conflict among Nations, 350) define a majority coalition as “a portion of the decisionmaking group that can carry out a strategy without the help of the remaining group members, and if necessary against their active opposition.” Snyder and Diesing see the building of majority coalitions as the essence of the BPM, while Charles Hermann (“The Impact of Single Group Decision Units”) seems to disagree, emphasizing compromise. 74. Snyder and Diesing, Conflict among Nations, 353. 75. Snyder and Diesing, Conflict among Nations, 519. 76. Snyder and Diesing, Conflict among Nations, 519. 77. Bundy, Danger and Survival, 446. 78. Kegley and Wittkopf, American Foreign Policy, 3rd ed., 498. See also Philip G. Roeder, “Soviet Policies and Kremlin Politics,” International Studies Quarterly 28, no. 2 (June 1984): 171–93, on the lack of consistency and coherence in Soviet foreign policy. 79. Philip Roeder has argued that “pluralistic” and “oligarchic” decision-making regimes in the Soviet Union are usually characterized by compromise, incrementalism, and aversion to risk. It appears that political competition leads to risk taking in the USSR, primarily in situations where a single leader who has consolidated his authority over foreign policy is being challenged by rivals. See Roeder, “Soviet Policies and Kremlin Politics.” 80. Roger Hilsman, The Politics of Policy Making in Defense and Foreign Affairs, 61. 81. Barbara Hill, “A General Model of International Conflict: Dynamics, Problems and Prospects,” paper presented at the International Studies Association conference, St. Louis, March 1988, 25. 82. A third theoretical issue has also received a lot of attention: Is government action a result unintended by any single player? This argument (relating to Arrow’s paradox) appears faulty on several grounds. First, it assumes that the result of “pulling and hauling” is always a compromise of some sort, when it is clear that another possible result may be that one faction gets its way and another faction loses. In this latter case, one faction clearly gets its intended result. Second, it assumes the chief executive is one of the bargainers. As Bendor and Hammond (“Rethinking Allison’s Models,” 315) point out, if outcomes are really unintended, one has to conclude that the choice of the blockade was unintended by Kennedy; but there is no evidence that he wanted a different solution. 83. Allison and Zelikow, Essence of Decision, 2nd ed., 307. 84. Martin Hollis and Steve Smith, “Roles and Reasons in Foreign Policy Decision Making,” British Journal of Political Science 16, no. 3 (July 1986): 269–86. Quote is from 276. 85. Stephen Krasner, “Are Bureaucracies Important? (Or Allison Wonderland),” Foreign Policy 7 (Summer 1972): 159–79. Quote is from 165. 86. Robert Axelrod, “Bureaucratic Decisionmaking in the Military Assistance Program: Some Empirical Findings,” in Readings in American Foreign Policy: A Bureaucratic Perspective, ed. Morton Halperin and Arnold Kanter (Boston: Little, Brown, 1973), 154–72. 87. Andrew K. Semmel, “Some Correlates of Attitudes to Multilateral Diplomacy in the United States Department of State,” International Studies Quarterly 20, no. 2 (June 1976): 301–24. 88. See, for instance, Cashman and Robinson, An Introduction to the Causes of War, chapter 3. 89. See Kaarbo and Gruenfeld, “The Social Psychology of Inter- and Intragroup Conflict in Governmental Politics.” 90. Edward Rhodes, “Do Bureaucratic Politics Matter? Some Disconfirming Findings from the Case of the U. S. Navy,” World Politics 47 (1994): 1–41. 91. Anderson, “What Do Decision Makers Do,” 299.

512

Notes

92. Graham H. Shepard, “Personality Effects on American Foreign Policy, 1969–1984: A Second Test of Interpersonal Generalization Theory,” International Studies Quarterly 32, no. 1 (March 1988): 121. 93. Hilsman, The Politics of Policy Making, 87. 94. Krasner, “Are Bureaucracies Important?” 95. Harriet Fast Scott, “Soviet Military Doctrine: From the Great Patriotic War to the Gorbachev Era,” address to the 2nd International Security Conference on Soviet Military Doctrine in an Era of Change, Norfolk, May 26, 1989. Also Michael Dobbs, “Soviet Cities Dissent on Afghan War,” Washington Post, March 20, 1989. 96. Madeleine Albright, Madame Secretary: A Memoir (New York: Miramax Books, 2003), 182. 97. Samuel Huntington, The Soldier and the State: The Theory and Politics of Civil-Military Relations (Cambridge, MA: Harvard University Press, 1957). 98. Richard Betts, Soldiers, Statesmen, and Cold War Crises (Cambridge MA: Harvard University Press, 1977), 4, 210, and 216. 99. Snyder and Diesing, Conflict among Nations, 512 and 359. 100. Todd S. Sechser, “Are Soldiers Less War-Prone than Statesmen?” Journal of Conflict Resolution 48, no. 5 (October 2004): 746–74. 101. Halperin and Kanter, “Introduction,” in Readings in American Foreign Policy: A Bureaucratic Perspective, 9–10. 102. Krasner, “Are Bureaucracies Important?”; Jerel A. Rosati, “Developing a Systematic Decision-Making Framework: Bureaucratic Politics in Perspective,” World Politics 33, no. 2 (January 1981): 234–52; Amos Perlmutter, “The Presidential Political Center and Foreign Policy: A Critique of the Revisionist and Bureaucratic-Political Orientations,” World Politics 27, no. 1 (October 1974): 87–106; Robert Art, “Bureaucratic Politics and American Foreign Policy: A Critique,” Policy Sciences (December 1973): 467–90. 103. Perlmutter, “The Presidential Political Center,” 92. 104. Dan Caldwell, “Bureaucratic Foreign Policy Making,” American Behavioral Scientist 21, no. 1 (October 1977): 97. 105. Rosati, “Developing a Systematic Decision-Making Framework.” 106. On this point see Fen Osler Hampson, “The Divided Decision-Maker: American Domestic Politics and the Cuban Crisis,” International Security 9, no. 3 (1984/85): 130–65. 107. Bundy, Danger and Survival, 400–401. 108. Bendor and Hammond, “Rethinking Allison’s Models,” 316. 109. Welch, “Organizational Process and Bureaucratic Politics,” 132–33. 110. See Rosati, “Developing a Systematic Decision-Making Framework,” 246; and Krasner, “Are Bureaucracies Important?” 111. Thomas Preston and Paul ‘t Hart, “Understanding and Evaluating Bureaucratic Politics: The Nexus between Political Leaders and Advisory Systems,” Political Psychology 20, no. 1 (March 1999), 49–98. They argue that leaders shape the decision-making process in different ways depending on their need for power (which affects their desire to control the process and its decisions), their degree of integrative complexity (which affects their need for information), and their prior experience and expertise (which affects their need for advisers). 112. See Paul ‘t Hart and Uriel Rosenthal, “Reappraising Bureaucratic Politics,” Mershon International Studies Review 42, no. 2 (November 1998): 233–40. 113. Snyder and Diesing, Conflict among Nations, 355–56. 114. On intermestic politics see Bayless Manning, “The Congress, the Executive and Intermestic Affairs: Three Proposals,” Foreign Affairs 55, no. 2 (January 1977): 306–24, and John Spanier and Eric Uslaner, How American Foreign Policy Is Made, 2nd ed. (New York: Holt, Rinehart & Winston, 1978). 115. C. F. Hermann, “The Impact of Single Decision Units,” and Bueno de Mesquita, The War Trap. 116. C. F. Hermann, “The Impact of Single Decision Units.” 117. Snyder and Diesing, Conflict among Nations, 512. Snyder and Diesing also see the BPM as more relevant to modern times, “when agencies other than the foreign offices have become involved in foreign policy making.” 118. Kim Richard Nossal, “Bureaucratic Politics and the Westminster Model,” in International Conflict and Conflict Management, ed. Robert O. Matthews, Arthur Rubinoff, and Janice Gross Stein (Scarborough, ONT: Prentice-Hall, 1984), 120–27. 119. For example, H. Gordon Skilling and Franklin Griffiths focus on “interest groups” in the Soviet system in their Interest Groups in Soviet Politics (Princeton, NJ: Princeton University Press, 1971); Carl Linden uses a “conflict model” to examine the Khrushchev administration in Khrushchev and the Soviet Leadership (Baltimore: Johns Hopkins University Press, 1966); and Dennis Ross emphasizes “coalition maintenance” in his “Coalition

Notes

513

Maintenance in the Soviet Union,” World Politics 32, no. 2 (January 1980): 258–80. For a dissenting view on these pluralistic perspectives, which advocates a return to a totalitarian model of Soviet policy making, see William Odom, “A Dissenting View on the Group Approach to Soviet Politics,” World Politics 28, no. 4 (July 1976): 542–67. 120. Jiri Valenta, Soviet Intervention in Czechoslovakia, 1968: Anatomy of a Decision (Baltimore: Johns Hopkins University Press, 1979), 4. 121. Ross, “Coalition Maintenance in the Soviet Union”; see also Dennis Ross, “Risk Aversion in Soviet Decisionmaking,” in Soviet Decisionmaking for National Security, ed. Jiri Valenta and William Potter (Boston: Allen & Unwin, 1984), 237–51. 122. Levy and Thompson, Causes of War, 169–74. 123. Trumbull Higgins, The Perfect Failure: Kennedy, Eisenhower, and the CIA at the Bay of Pigs (New York: Norton, 1987), quoted in Levy and Thompson, Causes of War, 174. 124. See for instance, James Bamford, A Pretext for War: 9/11, Iraq, and the Abuse of America’s Intelligence Agencies (New York: Doubleday, 2004). 125. See Artemy Kalinovsky, “Decision-Making and the Soviet War in Afghanistan,” Journal of Cold War Studies 11, no. 4 (2009): 46–73; and Charles J. Sullivan, “The Kremlin and Kabul: The 1979 Soviet Invasion of Afghanistan in Retrospect,” Washington Review of Turkish and Eurasian Affairs, http://www.thewashingtonreview.org/articles/ the-kremlin-and-kabul-the-1979-soviet-invasion-in-retrospect.html. 126. John Vasquez, The War Puzzle (Cambridge: Cambridge University Press, 1993), chapter 6, 198–224. 127. See for instance the chapter on World War II in the Pacific in Greg Cashman and Leonard C. Robinson, An Introduction to the Causes of War (Lanham, MD: Rowman & Littlefield, 2007). 128. See, for instance, Donald Kagan, On the Origin of Wars (New York: Doubleday, 1995), 187. 129. Levy and Thompson, Causes of War, 95. 130. Jiri Valenta, Soviet Intervention in Czechoslovakia, 1968 (Baltimore: Johns Hopkins University Press, 1979). For an excellent use of the BPM to explain the U.S. involvement in the Vietnam War, see Robert Gallucci, Neither Peace Nor Honor (Baltimore: Johns Hopkins University Press, 1975). David Halberstam’s brilliant Vietnam study, The Best and the Brightest (Greenwich, CT: Fawcett, 1972) is consistent with the BPB approach, although Halberstam never explicitly identifies his analytical framework as such. 131. Karen Dawisha, in “The Limits of the Bureaucratic Politics Model: Observations on the Soviet Case,” Studies in Comparative Communism 13, no. 4 (Winter 1980): 300–346, points out that many of the institutional arrangements of the old Soviet system—especially the dominance of the CPSU—increased centralization, decreased factionalism, and decreased organizational autonomy, thereby making bureaucratic politics unlikely in the Soviet Union. These points are well taken, but they do not preclude factional politics. She also argues that Valenta’s analysis of the Soviet decision in 1968 fails to prove that bureaucratic politics were very important in the decision on the grounds that “a single and unified military view could not be discerned. Differences of opinion within the military were judged to be at least as great as those within the Party” (306–7). But this misses the point. Bureaucratic politics exists within large organizations as well as between them. They are not monoliths. Transorganizational alliances existed in both the pro-intervention and the anti-intervention camp. This would seem to be the essence of the BPM, not its absence. 132. Steve Smith, “Policy Preferences and Bureaucratic Position: The Case of the American Hostage Rescue Mission,” International Affairs 61, no. 1 (Winter 1984–85): 9–25; and Martin Hollis and Steve Smith, “Roles and Reasons in Foreign Policy Decision Making,” British Journal of Political Science 16 (July 1986): 269–86. Other analysts have provided BPM explanations for other Soviet decision. Spechler traces the change in Soviet policy toward the Middle East in 1973—in particular the decisions to permit Egyptian purchases of Soviet military equipment and to give a green light for war with Israel—to a change in the relative balance of power among elite factions with the Politburo. Those who held a “competitive” image of Soviet-American relations (led by Prime Minister Kosygin) shifted their support from those who held a “cooperative” image of Soviet-US relations (led by General Secretary Brezhnev) to those who held an “antagonistic” image (including Defense Minister Grechko, party ideologist Suslov, KGB Chair Andropov, and other). In this case, pressure for the shift came from events in the external environment, such as Egypt’s expulsion of Soviet advisers in 1972. Dina Rome Spechler, “The USSR and the Third World Conflicts: Domestic Debate and Soviet Policy in the Middle East, 1967–73,” World Politics 38, no. 3 (April 1986): 436–61. 133. See the chapter on the Iraq War in Cashman and Robinson, An Introduction to the Causes of War. 134. Martin Smith, “US Bureaucratic Politics and the Decision to Invade Iraq,” Contemporary Politics 14, no. 1 (March 2008): 91–105. Quote is from 103.

514

Notes

135. David Mitchell and Tansa Massoud create an integrated model to explain the Iraq War decision that combines insights from the literature on individual leadership, management style, bureaucratic politics, and small group dynamics. They argue that Bush’s personal style led him to create a management structure that was formal, but also had a low degree of centralization. This decision structure, in turn, created incentives for both anticipatory compliance and also the formation of bureaucratic factions and infighting and ultimately to defective decision-making process. D. Mitchell and T. G. Massoud, “Anatomy of Failure: Bush’s Decision-Making Process and the Iraq War,” Foreign Policy Analysis 5, no. 3 (July 2009): 265–86. 136. Allison and Halperin, “Bureaucratic Politics: A Paradigm.” Allison’s second edition of Essence of Decision maintains the use of the term paradigm. 137. Welch, “Organizational Process and Bureaucratic Politics Paradigms,” 114–15. 138. Bendor and Hammond, “Rethinking Allison’s Models,” 318. 139. Caldwell, “Bureaucratic Foreign Policy Making,” 100. 140. David Welch, “A Positive Science of Bureaucratic Politics?” Mershon International Studies Review 42, no. 2 (November 1998): 210–16. Quote is from 214. 141. ‘t Hart and Rosenthal, “Reappraising Bureaucratic Politics.” 142. On this point see David Dessler, “Beyond Correlations: Toward a Causal Theory of War,” International Studies Quarterly 35, no. 3 (September 1991): 337–55. 143. Irving Janis, Groupthink, 2nd ed. (Boston: Houghton Mifflin, 1982), 9. The first edition was called Victims of Groupthink (Boston: Houghton Mifflin, 1972). 144. Janis, Groupthink, 242–43. 145. Janis, Groupthink, 258. 146. Sally Riggs Fuller and Ramon J. Aldag, “Challenging the Mindguards: Moving Small Group Analysis beyond Groupthink,” in Beyond Groupthink: Political Group Dynamics and Foreign Policy-Making, ed. P. ‘t Hart, E. Stern, and B. Sundelius (Ann Arbor: University of Michigan Press, 1997), 55–93. 147. For reviews of groupthink research, see James K. Esser, “Alive and Well after 25 Years: A Review of Groupthink Research,” Organizational Behavior and Human Decision Processes, 73, no. 2/3 (February/March 1998): 116–41; Paul ‘t Hart, “Irving L. Janis’ Victims of Groupthink,” Political Psychology 12, no. 2 (1991): 247–78; and W. W. Park, “A Review of Research on Groupthink,” Journal of Behavioral Decision-Making 3 (1990): 229–45. 148. C. McCauley, “The Nature of Social Influence in Groupthink: Compliance and Internalization,” Journal of Personality and Social Psychology 57 (1989): 250–60. 149. P. E. Tetlock, R. Peterson, C. McGuire, S. Chang, and P. Feld, “Assessing Political Group Dynamics: A Test of the Groupthink Model,” Journal of Personality and Social Psychology, 63 (1992): 403–25. 150. On the Iranian hostage crisis, see S. Smith, “Groupthink and the Hostage Rescue Mission,” British Journal of Political Science 15 (1984): 453–58; on the Falklands crisis, see J. Heller, “The Dangers of Groupthink,” The Guardian, January 31, 1983, 13; on Iran-Contra, see P ‘t Hart, Groupthink in Government: A Study of Small Groups and Policy Failure (Baltimore: Johns Hopkins University Press, 1990); on the Gulf crisis, see Steve. A. Yetiv, “Groupthink and the Gulf Crisis,” British Journal of Political Science 33, no. 3 (2003): 419–42. 151. Yetiv argues that what accounts for the relative policy success of Bush’s handling of the 1990–91 Gulf crisis is that, in this case, directive leadership was carried out by a leader with a high level of foreign policy skill and experience, reducing the negative effects of the lack of methodical norms of decision making. S. Yetiv, “Groupthink and the Gulf Crisis.” 152. C. R. Leana, “A Partial Test of Janis’ Groupthink Model: Effects of Group Cohesiveness and Leader Behavior on Defective Decision Making,” Journal of Management 1 (1985): 5–17. 153. For a review, see Esser, “Alive and Well after 25 Years,” 127–28, 134; and ‘t Hart, “Irving L. Janis’ Victims of Groupthink,” 253–54, 263–64. 154. Yetiv, “Groupthink and the Gulf Crisis,” 437. 155. ‘t Hart, “Irving L. Janis’ Victims of Groupthink,” 253. 156. Esser, “Alive and Well after 25 Years,” 131–32. See also Leana, “A Partial Test of Janis’ Groupthink Model.” 157. Fuller and Aldag, “Challenging the Mindguards,” 70. This is also the conclusion of Park, “A Review of Research on Groupthink.” 158. Janis, Groupthink, 12 and 127. 159. Richard Barnet, Roots of War: The Men and Institutions behind U.S. Foreign Policy (New York: Penguin, 1981), 109.

Notes

515

160. See, for instance, Dean Pruitt, “Choice Shifts in Group Discussion: An Introductory Review,” Journal of Personality and Social Psychology 20 (1971): 339–60. 161. See D. Cartwright, “Risk-Taking by Individuals and Groups: An Assessment of Research Employing Choice Dilemmas,” Journal of Personality and Social Psychology 20 (1971): 261–78; and B. Wilpert, P. Burger, J. Doktor, and R. Doktor, “The Risky Shift in Policy Decision Making: A Comparative Analysis,” Policy Sciences 7 (1976): 365–70. 162. D. G. Myers and H. Lamm, “The Polarizing Effect of Group Discussion,” in Current Trends in Psychology: Readings from the American Scientist, ed. Irving Janis (Los Altos, CA: Kaufman, 1977). 163. See Andrew K. Semmel, “Small Group Dynamics in Foreign Policymaking,” in Biopolitics, Political Psychology, and International Politics, ed. Gerald Hopple (New York: St. Martin’s, 1982), 94–113; A. K. Semmel and Dean Minix, “Small Group Dynamics in Foreign Policy Decision Making: An Experimental Approach,” in Psychological Models in International Politics, ed. L. Falkowski (Boulder, CO: Westview, 1979); and Dean Minix, Small Groups and Foreign Policy Decision Making (Lanham, MD: University Press of America, 1982). 164. Semmel, “Small Group Dynamics in Foreign Policymaking.” 165. Minix, Small Groups and Foreign Policy, 127–44. 166. Semmel also finds that the psychological rigidity of group members is an important factor. Groups composed of more flexible members tended to be less risk-prone than those composed of members whose psychological profiles were more rigid. Semmel, “Small Group Dynamics,” 108. 167. Glen Whyte, “Groupthink Reconsidered,” Academy of Management Review 14, no. 1 (1989): 40–56. 168. Whyte, “Groupthink Reconsidered,” 43. 169. Whyte, “Groupthink Reconsidered,” 44–45. 170. Whyte, “Groupthink Reconsidered,” 52. 171. McCauley, “The Nature of Social Influence in Groupthink.” 172. Paul Hoyt and Jean A. Garrison, “Political Manipulation within the Small Group: Foreign Policy Advisers in the Carter Administration,” in Beyond Groupthink, 249–74. 173. ‘t Hart, Groupthink in Government. 174. ‘t Hart, Groupthink in Government, 123. 175. ‘t Hart, Groupthink in Government, 49. 176. ‘t Hart, Groupthink in Government, 123–24. 177. ‘t Hart, Groupthink in Government, 139. 178. ‘t Hart identifies a third, relatively rare, pathway as well. In this pathway, group cohesion is fostered by group prestige and group rituals as well as social amicability. Group members possess external anonymity and lack accountability and are therefore likely to become subject to deindividuation—“an intrapsychic state characterized by an absence of self-awareness, and disinhibited behavioral properties.” Deindividuation means that group members accept risky ventures because they are negligent of risks and are even insensitive to the possibility that they are taking risks. ‘t Hart, Groupthink in Government, 157. 179. ‘t Hart, Groupthink in Government, 124. 180. ‘t Hart, “Irving L. Janis’ Victims of Groupthink,” 262. 181. ‘t Hart, Groupthink in Government, chapter 14. 182. In addition to ‘t Hart, Groupthink in Government, see Stern and Sundelius, “Understanding Small Group Decisions in Foreign Policy,” 128, on this point. 183. Stern and Sundelius, “Understanding Small Group Decisions in Foreign Policy,” 129. 184. Cashman and Robinson, An Introduction to the Causes of War. 185. See Cashman and Robinson, An Introduction to the Causes of War, chapter 7, “The Iraq War,” and chapter 8, “Conclusion,” 373–74. 186. Dina Badie, “Groupthink, Iraq, and the War on Terror: Explaining US Policy Shift toward Iraq,” Foreign Policy Analysis 6, no. 4 (October 2010): 277–96. 187. E. Stern and B. Sundelius, “The Essence of Groupthink,” Mershon International Studies Review 1 (1994): 101–8; and E. Stern “Probing the Possibility of Newgroup Syndrome: Kennedy and the Bay of Pigs,” in Beyond Groupthink, 153–89. 188. Stern, “Probing the Possibility of Newgroup Syndrome,” 164–65. The concept of “idiosyncrasy credit” was coined by Hollander. E. Hollander, “Conformity, Status, and Idiosyncrasy Credit,” Psychological Review 65 (1958): 117–27. 189. Stern, “Probing the Possibility of Newgroup Syndrome,” 165. 190. Stern and Sundelius, “Understanding Small Group Decisions in Foreign Policy,” 136.

516

Notes

191. Stern and Sundelius, “Understanding Small Group Decisions in Foreign Policy,” 138. 192. Stern, “Probing the Possibilities of Newgroup Syndrome,” 166. 193. Stern, “Probing the Possibilities of Newgroup Syndrome,” 185. 194. Stern, “Probing the Possibilities of Newgroup Syndrome,” 182. Another facet of this is that new groups are also rather susceptible to manipulation by some group members. This appears to have happened as well. 195. Gregory M. Herek, Irving L. Janis, and Paul Huth, “Decision Making during International Crisis: Is Quality of Process Related to Outcome?” Journal of Conflict Resolution 31, no. 2 (June 1987): 203–26. 196. Mark Schafer and Scott Crichlow, “Antecedents of Groupthink: A Quantitative Study,” Journal of Conflict Resolution 40, no. 3 (September 1996): 415–35. 197. Janis, Groupthink, 196. 198. Janis, Groupthink, 197. 199. Yetiv, “Groupthink and the Gulf Crisis.” 200. ‘t Hart, Groupthink in Government, 129. 201. Both group cohesion and group conflict have beneficial effects—in moderation. Bureaucratic conflict helps to offset the potential for rushed decisions and false consensus created by groupthink; and group cohesion helps to offset the dysfunctional consequences of extreme conflict between top-level players. 202. C. F. Hermann, “The Impact of Single Group Decision Units.” 203. C. F. Hermann, “The Impact of Single Group Decision Units.” 204. Janis, Groupthink, 262–71. 205. Cited in Daniel Kahneman, Thinking Fast and Slow (New York; Farrar, Straus and Giroux, 2011), 264–65. 206. Alexander George, “The Case for Multiple Advocacy in Making Foreign Policy,” American Political Science Review 66 (September 1972): 751–85. 207. Richard K. Betts, “Analysis, War, and Decision: Why Intelligence Failures Are Inevitable,” World Politics 31, no. 1 (October 1978), 76. 208. Richard Ned Lebow, Between Peace and War: The Nature of International Crises (Baltimore: Johns Hopkins University Press, 1981), 296–305.

5. THE STATE LEVEL OF ANALYSIS, PART I 1. See Michael P. Sullivan, International Relations: Theories and Evidence (Englewood Cliffs, NJ: PrenticeHall, 1976), 102–3. 2. J. David Singer and Melvin Small, Wages of War, 1816–1965: A Statistical Handbook (New York: Wiley, 1972), 287. The nations involved in the most wars in this period were Britain and France, nineteen; Turkey, seventeen; Russia, fifteen; Sardinia, twelve; and Spain, nine. 3. Zeev Maoz, “Pacifism and Fightaholism in International Politics: A Structural History of National and Dyadic Conflict, 1816–1992,” International Studies Review 6 (2004): 107–33. 4. See Kenneth Waltz, Man, the State and War (New York: Columbia University Press, 1959), chapter IV, from which this discussion draws heavily. 5. Edmund Burke, Letter to William Smith, January 9, 1795, quoted in John Bartlett, Famous Quotations, 14th ed., ed. Emily Morison Beck (Boston: Little, Brown, 1968), 454. 6. A. J. P. Taylor, Rumors of War (London: Hamish Hamilton, 1952), 44, quoted in Waltz, Man, the State and War, 114. It should be mentioned that liberals suggest several other solutions. The Manchester liberals of the eighteenth and nineteenth centuries suggested that a policy of free trade would bind nations together economically in such a way as to make war unthinkable. War would make all states subject to international economic chaos since it would imperil the international trading system in which all nations were mutually dependent on each other for important goods and services. Some modern liberals argue that a world government is the only long-term solution. Just as modern liberals argue that the central government should involve itself to a much greater degree in resolving economic conflicts within the state (abandoning laissez faire policies), they also argue that a world government is necessary to resolve political and economic conflicts between states. 7. Quincy Wright, A Study of War, vol. II (Chicago: University of Chicago Press, 1942), 833–42. 8. Melvin Small and J. David Singer, “The War Proneness of Democratic Regimes,” Jerusalem Journal of International Relations 1 (1976): 49–69.

Notes

517

9. Bruce Russett and R. J. Monsen, “Bureaucracy and Polyarchy as Predictors of Performance: A CrossNational Examination,” Comparative Political Studies 8 (April 1975), 5–31. 10. Michael Doyle, Ways of War and Peace (New York: Norton, 1997), 268. 11. Michael Haas, “Societal Approaches to the Study of War,” in The War System: An Interdisciplinary Approach, ed. Richard A. Falk and Samuel S. Kim (Boulder: Westview Press, 1980), 354–55. 12. Dina Zinnes and Jonathan Wilkenfeld, “An Analysis of Foreign Conflict Behavior of Nations,” in Comparative Foreign Policy: Theoretical Essays, ed. Wolfram F. Hanrider (New York: David McKay, 1971), 167–213; Steven A. Salmore and Charles F. Hermann, “The Effects of Size, Development and Accountability on Foreign Policy,” Peace Research Society Papers 14 (1970): 15–30. 13. R. J. Rummel, “Libertarianism and International Violence,” Journal of Conflict Resolution 27, no. 1 (March 1983): 27–71. See also his “Libertarian Propositions on Violence Within and Between Nations: A Test against Published Research Results,” Journal of Conflict Resolution 29, no. 1 (September 1985): 419–55. 14. Steve Chan, “Mirror, Mirror on the Wall . . . Are the Freer Countries More Pacific?” Journal of Conflict Resolution 28, no. 4 (December 1984): 617–48. 15. Erich Weede, “Democracy and War Involvement,” Journal of Conflict Resolution 28, no. 4 (December 1984): 649–66. 16. T. Clifton Morgan and Sally H. Campbell, “Domestic Structures, Decisional Constraints, and War: So Why Kant Democracies Fight?” Journal of Conflict Resolution 35 (1991): 187–211. Morgan and Campbell also find that decisional constraints reduce the probability of war for major powers, but increase the probability of war for minor powers. 17. In addition to the previously mentioned studies by Wright, Small and Singer, Salmore and Hermann, Russett and Monsen, Chan, and Weede, see also Zeev Maoz and Nasrin Abdolali, “Regime Types and International Conflict, 1816–1976,” Journal of Conflict Resolution 33 (1989): 3–35; Erich Weede, “Some Simple Calculations on Democracy and War Involvement,” Journal of Peace Research 29 (1992): 377–83; Zeev Maoz and Bruce Russett, “Normative and Structural Causes of Democratic Peace, 1946–1986,” American Political Science Review 87 (1993): 624–38; Henry Farber and Joanne Gowa, “Polities and Peace,” International Security 20, no. 2 (1995): 123–46; Charles Kegley and Margaret Hermann, “Military Intervention and the Democratic Peace,” International Interactions 21 (1995): 1–21; Steve Chan, “In Search of the Democratic Peace: Problems and Promise,” Mershon International Studies Review 41, no. 1 (1997): 59–91; James Lee Ray, “Democracy: On the Level(s), Does Democracy Correlate with Peace?” in What Do We Know About War?, ed. John Vasquez (Lanham, MD: Rowman & Littlefield, 2000). 18. Maoz and Abdolali, “Regime Types and International Conflict,” 20. 19. Maoz and Abdolali, “Regime Types and International Conflict,” 23–26. 20. T. Clifton Morgan and Valerie L. Schwebach, “Take Two Democracies and Call Me in the Morning: A Prescription for Peace?” International Interactions 17, no. 4 (1992): 305–20. 21. Kenneth Benoit, “Democracies Really Are More Pacific (In General),” Journal of Conflict Resolution 40, no. 4 (1996): 636–57. 22. Nils Petter Gleditsch and Havard Hegre, “Peace and Democracy: Three Levels of Analysis,” Journal of Conflict Resolution 41, no. 2 (1997): 283–310; see especially 295–96. 23. Gleditsch and Hegree, “Peace and Democracy,” 293–94, 307. 24. Sven Chojnacki, “Democratic Wars and Military Interventions, 1946–2002: The Monadic Level Reconsidered,” in Democratic Wars, ed. A. Geis, L. Brock, and H. Muller (New York: Palgrave Macmillan, 2006), 13–37. 25. Chojnacki, “Democratic Wars and Military Interventions,” 19–21. 26. Jack Levy, “Preventive War and Democratic Politics,” International Studies Quarterly 52, no. 1 (March 2008): 1–24. 27. See Randall Schweller, “Domestic Structure and Preventive War: Are Democracies More Pacific?” World Politics 44, no. 2 (1992): 235–69. 28. Rousseau, Gelpi, Reiter, and Huth, “Assessing the Dyadic Nature of the Democratic Peace, 1918–88.” 29. J. Joseph Hewitt and Jonathan Wilkenfeld, “Democracies in International Crises. International Interactions 22 (1996): 123–34. 30. Paul Huth, Standing Your Ground (Ann Arbor: University of Michigan, 1998), 187. 31. Huth, Standing Your Ground, 136–37. 32. Bruce Russett and John Oneal, Triangulating Peace: Democracy, Interdependence and International Organizations (New York: Norton, 2001). Politically relevant dyads are pairs of states that are contiguous or pairs that contain at least one great power.

518

Notes

33. Russett and Oneal, Triangulating Peace, 115. Their study finds that pairs of democracies are much more peaceful than autocratic pairs or mixed pairs. The risk of conflict for mixed pairs is 73 percent above the baseline, while the risk for autocratic pairs is 67 percent above it. The difference is not statistically significant. The risk of two democracies (both at +7 on the democracy scale) is 41 percent below the baseline risk. 34. Russett and Oneal, Triangulating Peace, 116. 35. Russett and Oneal, Triangulating Peace, 50. 36. On this point see also Errol Henderson, Democracy and War: The End of an Illusion? (Boulder, CO: Lynne Rienner, 2002), 55–58. 37. David Lake, “Powerful Pacifists: Democratic States and War,” American Political Science Review 86, no. 1 (1992): 24–37. 38. Dan Reiter and Allan Stam, “Democracy, War Initiation and Victory,” American Political Science Review 92, no. 2 (1998): 377–90; see table 1, p. 384. They also find that democracies won 76 percent of the wars they were involved in, regardless of whether they were initiators or targets. See also Reiter and Stam, “Democracy and Battlefield Effectiveness,” Journal of Conflict Resolution 42, no. 3 (1998): 259–77; and Reiter and Stam, Democracies at War (Princeton, NJ: Princeton University Press). For a critique, see Michael Desch, “Democracy and Victory: Why Regime Type Hardly Matters,” International Security 27, no. 2 (2002): 5–47, and responses by Ajin Choi, David Lake, and Reiter and Stam in International Security 28, no. 1 (Summer 2003). 39. Randolph Siverson, “Democracies and War Participation: In Defense of the Institutional Constraints Argument,” European Journal of International Relations 1, no. 4 (1995): 481–89. 40. See, for instance, Michael Desch, Power and Military Effectiveness: The Fallacy of Democratic Triumphalism (Baltimore: Johns Hopkins University Press, 2008); Desch, “Democracy and Victory: Why Regime Type Hardly Matters,” International Security 27, no. 2 (Fall 2002): 5–47; Desch, “Democracy and Victory: Fair Fights or Food Fights?” International Security 28, no. 1 (Summer 2003): 180–94; Stephen Biddle and Stephen Long, “Democracy and Military Effectiveness: A Deeper Look,” Journal of Conflict Resolution 48, no. 4 (August 2004): 525–46; Jane Kellett Cramer, “Militarized Patriotism: Why the U. S. Marketplace of Ideas Failed before the Iraq War,” Security Studies 16, no. 3 (July 2007): 489–524; Chaim Kaufmann, “Threat Inflation and the Failure of the Marketplace of Ideas: The Selling of the Iraq War,” International Security 29, no. 1 (Summer 2004): 5–48. 41. Alexander Downes, “How Smart and Tough Are Democracies? Reassessing Theories of Democratic Victory in War,” International Security 33, no. 4 (Spring 2009): 9–51. 42. John Owen, “How Liberalism Produces Democratic Peace,” International Security 19, no. 2 (1994): 87–125, 91. 43. The argument was originally put forward in Edward Mansfield and Jack Snyder, “Democratization and the Danger of War,” International Security 20, no. 5 (1995): 5–38. A fuller discussion is found in their later Electing to Fight: Why Emerging Democracies Go to War (Cambridge, MA: MIT Press, 2005). 44. Mansfield and Snyder, Electing to Fight, 57. 45. Mansfield and Snyder, Electing to Fight, 54. 46. See Mansfield and Snyder, “Pathways to War in Democratic Transitions,” International Organization 63 (Spring 2009): 381–90. 47. Mansfield and Snyder, Electing to Fight, 36. The evidence cited here is from Mansfield and Snyder’s updated and revised analysis in Electing to Fight, rather than the earlier analysis found in their article in International Studies. 48. Edward Mansfield and Jack Snyder, “Incomplete Democratization and the Outbreak of Military Disputes,” International Studies Quarterly 46, no. 4 (2002): 529–49; Patricia Weitsman and George Shambaugh, “International Systems, Domestic Structures, and Risk,” Journal of Peace Research 39, no. 3 (2002): 289–312; and Christopher Gelpi and Joseph Grieco, “Attracting Trouble: Democracy, Leadership Tenure, and the Targeting of Militarized Challenges, 1918–1992,” Journal of Conflict Resolution 45, no. 6 (2001): 749–815. 49. Mansfield and Snyder, Electing to Fight, 6–7. 50. Mansfield and Snyder, Electing to Fight, 101–24. 51. Mansfield and Snyder, Electing to Fight, 129–35. 52. Mansfield and Snyder, Electing to Fight, 4. 53. William R. Thompson and Richard M. Tucker, “A Tale of Two Democratic Peace Critiques,” Journal of Conflict Resolution 41 (June, 1997): 428–51; John Oneal and Bruce Russett, “The Classical Liberals Were Right: Democracy, Interdependence, and Conflict, 1950–1985,” International Studies Quarterly 41(1997): 267–94; Russett and Oneal, Triangulating Peace; Andrew Enterline, “Regime Changes, Neighborhoods, and Interstate Conflict, 1816–1992,” Journal of Conflict Resolution 42, no. 6 (1998): 804–29; Kurt Dassel and Eric Reinhardt, “Domestic

Notes

519

Strife and the Initiation of Violence at Home and Abroad,” American Journal of Political Science 43, no. 1 (1999): 56–85; and Graeme Davies, “Domestic Strife and the Initiation of International Conflicts,” Journal of Conflict Resolution 46, no. 5 (2002): 672–92. 54. Andrew J. Enterline, “Driving While Democratizing,” International Security 20, no. 4 (Spring 1996): 183–96. 55. Michael D. Ward and Kristian S. Gleditsch, “Democratizing for Peace,” American Political Science Review 92, no. 2 (March 1998): 51–61. 56. D. Scott Bennett and Allan C. Stam, The Behavioral Origins of War (Ann Arbor: University of Michigan, 2004), 117–20. 57. Russett and Oneal, Triangulating Peace, 116–22. 58. Vipin Narang and Rebecca M. Nelson, “Who Are These Belligerent Democratizers? Reassessing the Impact of Democratization on War,” International Organization 63 (Spring 2009): 357–79. 59. See Enterline, “Driving While Democratizing,” and Thompson and Tucker, “A Tale of Two Democratic Peace Critiques.” 60. This is also the judgment found in James Lee Ray, “Democracy: On the Level(s), Does Democracy Correlate with Peace?” in What Do We Know about War? ed. John Vasquez (Lanham, MD: Rowman & Littlefield, 2000), 299–316. 61. Mary Caprioli and Peter F. Trumbore, “Rhetoric versus Reality: Rogue States in Interstate Conflict,” Journal of Conflict Resolution 49, no. 5 (October 2005): 770–91. 62. John A. Hobson, Imperialism: A Study (Ann Arbor: University of Michigan Press, 1965), 71–82. 63. Hobson, Imperialism, 71–93. 64. Hobson, Imperialism, 46–63. 65. V. I. Lenin, Imperialism: The Highest Stage of Capitalism (New York: International Publishers, 1939). 66. See Waltz, Man, the State and War, 154–58. 67. Michael Haas, “Societal Approaches to the Study of War,” 349. 68. Kenneth N. Waltz, Theory of International Politics (Reading, MA: Addison-Wesley, 1979), 20. The discussion here draws on pp. 18–30. 69. D. K. Fieldhouse, “Imperialism: An Historiographical Revision,” in Economic Imperialism, ed. Kenneth Boulding and Tapan Mukerjee (Ann Arbor: University of Michigan Press, 1972), 110. 70. Herbert Feis, Europe, the World’s Banker, 1870–1914 (New York: August M. Kelley, 1930), 23, cited in Waltz, Theory of International Politics, 24. 71. Robert Gilpin, US Power and the Multinational Corporation (New York: Basic Books, 1975), 74. 72. Waltz, Theory of International Politics, 24. 73. Waltz, Theory of International Politics, 25. 74. William Appleman Williams, Tragedy of American Diplomacy, rev. ed. (New York: Dell, 1962); see especially 18–50. 75. See Reuven Brenner, Betting on Ideas: Wars, Inventions, Inflation (Chicago: University of Chicago Press, 1985), chapter 1, cited in Bruce Russett, “Economic Decline, Electoral Pressure and the Initiation of Interstate Conflict,” 124, in Prisoners of War? Nation-States in the Modern Era, ed. Charles Gochman and A. N. Sabrosky (Lexington, MA: Lexington Books, 1990). 76. C. W. Ostrom and B. L. Job, “The President and the Political Use of Force,” American Political Science Review 80 (1986): 554–66. 77. Russett, “Economic Decline, Electoral Pressures, and the Initiation of Interstate Conflict,” especially 124–34. 78. Geoffrey Blainey, The Causes of War (New York: Free Press, 1973), 89–90. See also Bruce Russett, “Prosperity and Peace,” International Studies Quarterly 27 (1983): 381–87. 79. Alec Laurence Macfie, “The Outbreak of War and the Trade Cycle,” Economic History 3 (February 1938): 89–97. 80. Joshua Goldstein, Long Cycles: Prosperity and War in the Modern Age (New Haven, CT: Yale University Press, 1988), 239–48. Goldstein’s work will be discussed more fully in chapter 7, since he sees economic cycles not as national cycles, but as part of the structure of the international system. 81. Goldstein, Long Cycles, 260–62. He adds a lateral pressure argument as well—that economic growth stimulates the need for greater resources, and this leads to lateral pressure, intersections, and war. 82. On this point see Russett, “Prosperity and Peace,” 386. 83. Blainey, 93. 84. For Macfie it was not only the present mood of optimism that was important, but also a nervousness that the prosperity might not last. Blainey does not mention this and prefers to emphasize only the optimistic part.

520

Notes

85. Blainey, 94. 86. Dexter Perkins, The American Approach to Foreign Policy, rev. ed. (New York: Atheneum, 1968), 136–55. 87. William R. Thompson, “Phases of the Business Cycle and the Outbreak of War,” International Studies Quarterly 26 (June 1982): 301–11. While Russett claims that American involvement in conflict since 1898 has been associated with weak economic conditions, his use of two- and three-year lags throws this conclusion into doubt. He also looks primarily at militarized disputes rather than war involvement. B. Russett, “Economic Decline, Electoral Pressures, and the Initiation of Interstate Conflict.” 88. There are many different indicators of the power of states; among the most widely used are geographic size, population, GNP or GNP per capita, iron and steel production, energy production or consumption, defense budget, and size of armed forces. 89. These same countries are in Diehl and Goertz’s list of ten countries most involved in territorial changes in the last century and a half (minus Austria-Hungary) and in Gochman and Maoz’s list of the ten states most involved in militarized disputes (minus Spain). Paul Diehl and Gary Goertz, “Territorial Changes and Militarized Conflict,” Journal of Conflict Resolution 32, no. 1 (March 1988): 103–22; Charles Gochman and Zeev Maoz, “Militarized Interstate Disputes, 1816–1976: Procedures, Patterns and Insights,” Journal of Conflict Resolution 28 (1984): 585–616. Of the sixteen wars (both interstate and extrasystemic) in which the most war-prone great powers were not involved, ten were in Latin America. See Singer and Small, Wages of War, table 11.2, 275–80, and tables 4.2 and 4.4, 60–69 and 72–75; see also the summary on 286–87. 90. Melvin Small and J. David Singer, “Patterns in International Warfare, 1816–1965,” Annals of the American Academy of Political and Social Sciences 391 (September 1970): 151–52, quoted in Lloyd Jensen, Explaining Foreign Policy (Englewood Cliffs, NJ: Prentice-Hall, 1982), 222–23. 91. Stuart Bremer, “National Capabilities and War Proneness,” in The Correlates of War II: Testing Some Realpolitik Models, ed. J. David Singer (New York: Free Press, 1980), 57–82. Bremer also finds that while the topranked states in the system had the highest frequency of war involvement, the number 2 ranked states had the highest frequency of war initiation. Indeed, of the states in ranks 1–5, the states ranked as number 1 had the lowest rate of initiation. This gives indirect support for those who emphasize the centrality of conflict between the system leader and its challenger(s). These theories will be addressed in chapter 9 and 10. 92. Paul Senese and John Vasquez, The Steps to War (Princeton: Princeton University Press, 2008), 270. The top ten states involved in war were the United States (Persian Gulf War, 1991); Soviet Union (Afghanistan, 1979); China (Sino-Viet War 1979); United Kingdom (Falklands, 1982 and Persian Gulf 1991), and France and Italy (Persian Gulf, 1991). 93. Michael Haas, “Societal Approaches to the Study of War,” in The War System, ed. Richard Falk and Samuel Kim (Boulder, CO: Westview, 1980), 355–56 and 365. Other analysts also find a positive association between foreign conflict involvement and national power indicators. See Patrick McGowan and Howard Shapiro, The Comparative Study of Foreign Policy (Beverly Hills, CA: Sage, 1973). 94. Gochman and Maoz, “Militarized Interstate Disputes.” 95. Maoz, “Pacifism and Fightaholism in International Politics.” 96. See, for instance, Maurice A. East and Phillip Gregg, “Factors Influencing Cooperation and Conflict in the International System,” International Studies Quarterly 11 (September 1967): 266; Lewis Richardson, Statistics of Deadly Quarrels (New York: Quadrangle/New York Times, 1960), xi; R. J. Rummel, “Some Attributes and Behavior Patterns of Nations,” Journal of Peace Research IV, no. 2 (1967): 197; and R. J. Rummel, “The Relationship between National Attributes and Foreign Conflict Behavior,” in Quantitative International Politics, ed. J. David Singer (New York: Free Press, 1968), 204. See also R. Rosecrance, A. Alexandroff, B. Healy, and A. Stein, “Power, Balance of Power, and Status in Nineteenth Century International Relations,” Sage Professional Papers in International Studies, Vol. 3 (Beverly Hills, CA: Sage, 1974). The authors find little association between capability indicators and conflict of the five great powers in Europe between 1870 and 1881. 97. See R. J. Rummel, “Testing Some Possible Predictors of Conflict Behavior Within and Between Nations,” Peace Research Society (International) Papers 1 (1964): 79–111; and R. J. Rummel, “The Relationship Between National Attributes and Foreign Conflict Behavior,” 187–214. 98. See especially Salmore and Hermann, “The Effects of Size, Development and Accountability on Foreign Policy,” 16–30; Maurice East, “Size and Foreign Conflict Behavior,” World Politics 25 (July 1973): 556–76; and Michael Sullivan’s excellent discussion of the literature in International Relations: Theories and Evidence, 109–14. 99. The most recent COW list of interstate wars is found in Meredith Reid Sarkees and Frank Wayman, Resort to War: 1816–2007 (Washington, DC: CQ Press, 2010).

Notes

521

100. See James Dougherty and Robert L. Pfaltzgraff Jr., Contending Theories of International Relations, 2nd ed. (New York: Harper & Row, 1981), 66–68. 101. William L. Shirer, The Rise and Fall of the Third Reich (New York: Fawcett Crest, 1960), 77. 102. Dougherty and Pfaltzgraff, Contending Theories of International Relations, 2nd ed., 67. 103. Adolf Hitler, Mein Kampf, American ed. (Boston, 1943), quoted in Shirer, The Rise and Fall of the Third Reich, 123. 104. Marcel Leroy, for instance, suggests a causal path from an environment-induced internal conflict to diversionary external conflict. Marcel Leroy, “Human Population as a Factor in Strategic Policy and Action,” in Global Resources and International Conflict, ed. Arthur Westing (Oxford: Oxford University Press, 1986), 159–82. 105. Stuart Bremer, J. David Singer, and Urs Luterbacher, “The Population Density and War Proneness of European Nations, 1816–1964,” Comparative Political Studies 6 (1973): 329–48. See also, J. David Singer, “The Correlates of War Project: An Interim Report and Rationale,” World Politics 24 (January 1972): 267. Singer’s findings refer to members of the European system. 106. Quincy Wright, A Study of War, vol. II (Chicago: University of Chicago Press, 1942), 1132. 107. Nazli Choucri, Population Dynamics and International Violence (Lexington, MA: Lexington Books, 1974). 108. Nils Petter Gleditsch, “Armed Conflict and the Environment: A Critique of the Literature,” Journal of Peace Research 35, no. 3 (1998): 381–400, 389. 109. Jaroslav Tir and Paul F. Diehl, “Demographic Pressure and Interstate Conflict: Linking Population Growth and Density to Militarized Disputes and Wars, 1930–1989,” Journal of Peace Research 35, no. 3 (1998): 319–39. 110. Tir and Diehl, “Demographic Pressure and Interstate Conflict,” 336. 111. Nazli Choucri and Robert North, Nations in Conflict: National Growth and International Violence (San Francisco: W. H. Freemen, 1975), 14–24. 112. Nazli Choucri and Robert North, “Lateral Pressure in International Relations: Concept and Theory,” in Handbook of War Studies, ed. Manus Midlarsky (Boston: Unwin Hyman, 1989), 296. They conclude that the more immediate cause of war is “human and subjective.” 113. Choucri and North, Nations in Conflict, 234–54. See the discussion in the next chapter. 114. Nazli Choucri, Robert North, and Susumu Yamakage, The Challenge of Japan before World War II and After (London: Routledge, 1992). 115. Richard Ashley, The Political Economy of War and Peace (New York: Nichols, 1980). 116. Choucri and North, “Lateral Pressure: Concept and Theory,” 310. 117. Choucri and North, “Lateral Pressure: Concept and Theory,” 311. 118. See Thomas Homer-Dixon, “On the Threshold: Environmental Changes as Causes of Acute Conflict,” International Security 16, no. 2 (Fall 1991): 76–116; Homer-Dixon, “Environmental Scarcities and Violent Conflict: Evidence from Cases,” International Security 19, no. 1 (Summer 1994): 5–40; and Homer-Dixon, Environment, Scarcity, and Violence (Princeton, NJ: Princeton University Press, 1999).

6. THE STATE LEVEL OF ANALYSIS, PART II 1. Geoffrey Blainey, The Causes of War (New York: Free Press, 1973), 71. 2. R. J. Rummel, “Dimensions of Conflict Behavior within and between Nations,” General Systems: Yearbook of the Society for General Systems Research 8 (1963): 1–50. Also, Rummel, “The Relationship between National Attributes and Foreign Conflict Behavior,” in Quantitative International Politics, ed. J. David Singer (New York: Free Press, 1968), 202 and 208. In a later study Rummel found foreign conflict to be inversely related to one of the three domestic conflict dimensions—subversion. This would seem to buttress at least part of Blainey’s argument that severe internal conflict is likely to inhibit a country’s desire to engage in external adventurism. R. J. Rummel, “Testing Some Possible Predictors of Conflict Behavior within and between Nations,” Peace Research Society (International) Papers 1 (1964): 79–111. 3. Raymond Tanter, “Dimensions of Conflict Behavior within and between Nations, 1958–1960,” Journal of Conflict Resolution 10 (March 1966): 41–64; Phillip Gregg and Arthur Banks, “Dimensions of Political Systems: Factor Analysis of ‘A Cross-Polity Survey,’” American Political Science Review 59 (September 1965): 602–14; and Zinnes and Wilkenfeld, “An Analysis of Foreign Conflict Behavior of Nations,” 167–213. See also Michael Haas, “Social Change and National Aggressiveness, 1900–1960,” in Singer, Quantitative International Politics, 215–45. Jonathan Wilkenfeld hypothesized that the lack of relationship between internal and external conflict in these studies

522

Notes

was because Rummel and his followers had grouped together different kinds of political systems in their analyses. He reanalyzed the data, but divided the countries into three groups—personalist (primarily Latin American states), centrist (primarily communist states and some Middle Eastern states), and polyarchies (democracies). He discovered different internal-external conflict patterns for different types of states. The results represented a fairly mixed bag. Jonathan Wilkenfeld, “Domestic and Foreign Conflict Behavior of Nations,” Journal of Peace Research 5, no. 1 (1968): 59–69. 4. The best review of this is Jack Levy, “The Diversionary Theory of War: A Critique,” in Handbook of War Studies, ed. Manus Midlarsky (Boston: Unwin and Hyman, 1989), 259–88. 5. Daniel Geller, Domestic Factors in Foreign Policy: A Cross-National Analysis (Cambridge: Schenkman, 1985). 6. Edward Mansfield and Jack Snyder, Electing to Fight (Cambridge, MA: MIT Press, 2005), 161. While the relationship holds for interstate wars in general, it does not exist for world wars or for extra-systemic or colonial wars. 7. Mansfield and Snyder, Electing to Fight, 101, 161–62. 8. Henk W. Houweling and Jan G. Siccama, Studies of War (Dordrecht: Martinus Nijhoff, 1988), chapter 7. 9. Greg Cashman and Leonard C. Robinson, An Introduction to the Causes of War: Patterns of Interstate Conflict from World War I to Iraq (Lanham, MD: Rowman & Littlefield, 2007), 371–72. The exception was the Six Day War in 1967. The other case in the study is the U.S.-Iraq War of 2003, which is a war between unequal states. 10. The situation is sometimes described as a principal-agent problem. In all states, democratic or autocratic, political leaders (agents) work on behalf of others and are subject to various constituencies (principals) who are able to punish them by removal from office if they are displeased. And leaders wish to remain in positions of power and are willing to take actions on their own behalf that are contrary to the interests of the principals and community as a whole—going to war, for instance—in order to achieve this. See Christopher Gelpi, “Democratic Diversions: Governmental Structure and the Externalization of Domestic Conflict,” Journal of Conflict Resolution 41, no. 2 (1997): 255–82, 259. 11. Georg Simmel, Conflict and the Web of Group Affiliations (Glencoe, IL: Free Press, 1955); Lewis Coser, The Functions of Social Conflict (New York: Free Press, 1956). See also the review in Jack Levy, “The Diversionary Theory of War: A Critique.” 12. The classic study here is John Mueller, War, Presidents and Public Opinion (New York: Wiley, 1973). For evidence of the brief but positive impact of the use of external force on presidential approval ratings, see P. James and J. Rioux, “International Crises and Linkage Politics: The Experiences of the United States, 1953–1994,” Political Research Quarterly 51 (1998): 781–812, and K. De Rouen, “Presidents and the Diversionary Use of Force: A Research Note,” International Studies Quarterly 44 (2000): 317–28. 13. See the review by Birger Heldt, “The Dependent Variable of the Domestic-External Conflict Relationship: Anecdotes, Theories and Systematic Studies,” Journal of Peace Research 34, no. 1 (1997): 101–6. 14. Levy, “The Diversionary Theory of War,” 282. 15. Levy, “The Diversionary Theory of War,” 271. 16. Quincy Wright, A Study of War, 2nd ed. (Chicago: University of Chicago Press, 1965), 140. 17. Richard Rosecrance, Action and Reaction in World Politics (Boston: Little, Brown, 1963), 306. 18. Richard Ned Lebow, Between Peace and War: The Nature of International Crises (Baltimore: Johns Hopkins University Press, 1981), 57–70. 19. Jack Levy, “The Diversionary Theory of War,” 259–88. 20. In this case the Argentine government engaged in a diversionary reaction that they probably did not expect would lead to full-scale war with Britain. See M. Hastings and S. Jenkins, The Battle for the Falklands (New York: Norton, 1983). 21. Hastings and Jenkins, The Battle for the Falklands; William K. Domke, War and the Changing Global System (New Haven: Yale University Press, 1988), 104–6, also argues that the war was aimed at alleviating general public discontent. 22. Hastings and Jenkins, Battle for the Falklands, 48. 23. Jack Levy and Lily Vakili, “Diversionary Action by Authoritarian Regimes: Argentina in the Falklands/ Malvinas Case,” in The Internationalization of Communal Strife, ed. M. Midlarsky (New York: Routledge, 1992), 118–46. 24. Lawrence Freedman and Virginia Gamba-Stonehouse, Signals of War: The Falklands Conflict of 1982 (Princeton, NJ: Princeton University Press, 1991), 49–83. 25. Freedman and Gamba-Stonehouse, Signals of War, 67–68.

Notes

523

26. Freedman and Gamba-Stonehouse, Signals of War, 62. 27. Richard Ned Lebow, “Miscalculation in the South Atlantic: The Origins of the Falklands War,” in Psychology and Deterrence, ed. Robert Jervis, Richard Ned Lebow, and Janice Gross Stein (Baltimore: Johns Hopkins University Press, 1985), 89–124. 28. James Meernik, “Presidential Decision Making and the Political Use of Military Force,” International Studies Quarterly 38, no. 1 (1994): 121–38; and James Meernik and Peter Waterman, “The Myth of Diversionary Use of Force by American Presidents,” Political Research Quarterly 49 (1996): 573–90. See also Leeds and Davis, “Domestic Political Vulnerability and International Disputes.” 29. Bruce Russett, “Economic Decline, Electoral Pressure, and the Initiation of Interstate Conflict,” in Prisoners of War? Nation-States in the Modern Era, ed. Charles Gochman and Alan Ned Sabrosky (Lexington, MA: Lexington Books, 1990), 123–140. 30. Charles Ostrom and Brian Job, “The President and the Political Use of Force,” American Political Science Review 80 (1986): 541–66; Patrick James and John Oneal, “The Influence of Domestic and International Politics on the President’s Use of Force,” Journal of Conflict Resolution 35 (1991): 307–34. 31. Kevin Wang, “Presidential Responses to Foreign Policy Crises: Rational Choice and Domestic Politics,” Journal of Conflict Resolution 40, no. 1 (1996): 68–97. 32. See B. O. Fordham, “Another Look at ‘Parties, Voters and the Use of Force Abroad’” Journal of Conflict Resolution 46 (1998): 572–96; K. DeRouen, “Presidents and the Diversionary Use of Force: A Research Note,” International Studies Quarterly 44 (2000): 317–28; and D. Clark, “Can Strategic Interaction Divert Diversionary Behavior? A Model of U.S. Conflict Propensity,” Journal of Politics 65, no. 4 (2003): 1013–39. 33. Bruce Russett shows that the U.S. has been more likely to be involved in MIDs in presidential election years than in non-election years. James and Oneal see the use of force as more likely in the third quarter of presidential election years, and Wang finds that forceful U.S. responses in crises are more likely as elections draw near. There is, of course, contrary evidence. Studies by Gaubatz and by Huth and Allee both find democratic states are more likely to get involved in international wars in the early stages of electoral cycles, but rarely around election times. And Leeds and Davis’s examination of eighteen advanced industrial states indicates no significant relationship at all between election cycles and diversionary behavior. Bruce Russett, “Economic Decline, Electoral Pressure, and the Initiation of Interstate Conflict”; James and Oneal, “The Influence of Domestic and International Politics on the President’s Use of Force”; Kevin Wang, Presidential Responses to Foreign Policy Crises”; Kurt Taylor Gaubatz, “Election Cycles and War,” Journal of Conflict Resolution 35 (1991): 212–44; Paul Huth and Todd Allee, The Democratic Peace and Territorial Conflict in the Twentieth Century (Cambridge: Cambridge University Press, 2002); B. A. Leeds and D. R. Davis, “Domestic Political Vulnerability and International Disputes,” Journal of Conflict Resolution 41 (1997): 814–34. 34. T. Clifton Morgan and Kenneth N. Bickers, “Domestic Discontent and the External Use of Force,” Journal of Conflict Resolution 36, no. 1 (1992): 25–52. 35. Levy and Vakili, “Diversionary Action by Authoritarian Regimes,” 120. 36. See Jack Levy, “The Diversionary Theory of War,” 272–74. 37. Alastair Smith, “Diversionary Foreign Policy in Democratic Systems,” International Studies Quarterly 40 (1996): 133–53. 38. Blainey, The Causes of War, 81. Hazelwood refers to this preference for a direct focus on the domestic situation as encapsulation. Leo Hazelwood, “Diversion Mechanisms and Encapsulation Processes: The Domestic Conflict-Foreign Conflict Hypothesis Reconsidered,” Sage International Yearbook of International Studies 3 (1975): 213–43. 39. On these issues, see Arno J. Mayer, “Internal Causes and Purposes of War in Europe, 1870–1956,” Journal of Modern History 41 (1969): 291–303; Hazelwood, “Diversionary Mechanisms and Encapsulation Processes”; Geoffrey Blainey, The Causes of War, 68–86; Jack Levy, “The Diversionary War Hypothesis”; and Birger Heldt, “The Dependent Variable of the Domestic-External Conflict Relationship.” 40. Bruce Russett, “Economic Decline, Electoral Pressure, and the Initiation of Interstate Conflict.” 41. Graeme Davies, “Domestic Strife and the Initiation of International Conflicts,” Journal of Conflict Resolution 46, no. 5 (2002): 672–92. 42. Davies, “Domestic Strife and the Initiation of International Conflicts,” 684–89. 43. Gelpi, “Democratic Diversions.” 44. One issue here is whether the willingness of the state to use diversionary force is linked to the presence or absence of strategic opportunity. One popular argument is that democratic states use diversionary force less because they have fewer strategic opportunities for it. The argument is that autocracies understand that democracies are

524

Notes

dependent on popular support and are sensitive to anything that reduces that support and are therefore likely to escalate disputes in a diversionary way if there is domestic unrest due to economic downturns, unpopularity of political leaders, and/or upcoming elections. Autocracies act in a more conciliatory way and do not seek to create crises that democracies can exploit. The result is that there are fewer crises for democracies to intervene in. The proposition is difficult to assess empirically and has only been validated by formal modeling. Given what we know about misperceptions in international relations, it perhaps begs credulity to suggest that the presumed lack of democratic scapegoating is due to clever strategic anticipation by autocratic elites. The above idea is most associated with Alastair Smith, “Diversionary Foreign Policy in Democratic Systems.” See also Brett Ashley Leeds and David R. Davis, “Domestic Political Vulnerability and International Disputes,” Journal of Conflict Resolution 41, no. 6 (1997): 814–34. 45. Sara McLaughlin Mitchell and Brandon C. Prins, “Rivalry and Diversionary Uses of Force,” Journal of Conflict Resolution 48, no. 6 (2004): 937–61. 46. Mitchell and Prins, “Rivalry and Diversionary Uses of Force,” 945. 47. Their control variables also have a clear impact. If both states in the dyad are democracies, this reduces MID initiation. So does distance. States are also more likely to initiate MIDs with states over whom they have a distinct military advantage. Mitchell and Prins, “Rivalry and Diversionary Uses of Force,” 951–56. 48. Jaroslav Tir and Michael Jasinski, “Domestic-Level Diversionary Theory of War: Targeting Ethnic Minorities,” Journal of Conflict Resolution 52, no. 5 (October 2008): 641–64. 49. Ted Robert Gurr, Peoples Versus States: Minorities at Risk in the New Century (Washington, DC: United States Institute of Peace, 2000). 50. Tir and Jasinski, “Domestic-Level Diversionary Theory of War,” 655. 51. This is a point made by Levy and Vakili, “Diversionary Action by Authoritarian Regimes.” 52. Ostrom and Job, “The President and the Political Use of Force”; Russett, “Economic Decline, Electoral Pressures, and the Initiation of Interstate Conflict”; Gelpi, “Democratic Diversions.” 53. Smith, “Diversionary Foreign Policy in Democratic Systems,” 145–47. 54. Leeds and Davis, “Domestic Political Vulnerability and International Disputes.” 55. Ross Miller, “Regime Type, Strategic Interaction, and the Diversionary Use of Force,” Journal of Conflict Resolution 43, no. 3 (1999): 388–402. 56. Miller, “Regime Type, Strategic Interaction, and the Diversionary Use of Force,” 398–99. 57. Miller, “Regime Type, Strategic Interaction, and the Diversionary Use of Force,” 394–96. 58. Mitchell and Prins, “Rivalry and Diversionary Uses of Force,” 951–56. 59. Davies, “Domestic Strife and the Initiation of International Conflicts.” 60. Gelpi, “Democratic Diversions.” 61. Jeffrey Pickering and Emizet Kisangani, “Democracy and Diversionary Military Intervention: Reassessing Regime Type and the Diversionary Hypothesis,” International Studies Quarterly 49 (2005): 23–43. 62. Kurt Dassel, “Civilians, Soldiers, and Strife: Domestic Sources of International Aggression,” International Security 23, no. 1 (1998): 107–40. 63. In a later work, Dassel and Reinhardt argue that domestic strife will lead to foreign aggression only if it threatens the interests of the military. Kurt Dassell and Eric Reinhardt, “Domestic Strife and the Initiation of Violence at Home and Abroad,” American Journal of Political Science 43, no. 1 (1999): 56–85, especially 57. Dassel and Reinhardt argue that domestic strife is also most likely in the presence of contested institutions. 64. Dassel, “Citizens, Soldiers, and Strife,” 116. 65. Samuel Huntington, Political Order in Changing Societies (New Haven, CT: Yale, 1968), 196, quoted in Dassel, “Citizens, Soldiers, and Strife,” 117. 66. On the role of the Japanese military and the Japanese political system in World War II in the Pacific, see Greg Cashman and Leonard C. Robinson, An Introduction to the Causes of War: Patterns of Interstate Conflict from World War I to Iraq (Lanham, MD: Rowman & Littlefield, 2007), chapter 3. The authors find substantial merit in applying Dassel’s line of argument to the Japanese case. 67. Dassel and Reinhardt, “Domestic Strife and the Initiation of Violence at Home and Abroad.” 68. Davies, “Domestic Strife and the Initiation of International Conflicts.” 69. Blainey, The Causes of War, 72–81. 70. Kristian Gleditsch, Idean Salehyan, and Kenneth Schultz, “Fighting at Home, Fighting Abroad: How Civil Wars Lead to International Disputes,” Journal of Conflict Resolution 52, no. 4 (2008): 479–506, especially 483. 71. Blainey, The Causes of War, 82. Michael Haas seems to imply just the opposite. He cites the example that when internal disputes broke out in Switzerland in 1801 Napoleon dispatched 20,000 men to obtain an armistice that

Notes

525

brought Switzerland under French domination. On the other hand, he mentions that large nations with vexing problems were seen by others as too difficult to control, thus deterring attempts at conquest. Michael Haas, “Societal Approaches to the Study of War,” in The War System, ed. Richard Falk and Samuel Kim (Boulder, CO: Westview, 1980), 352. 72. See the case study in Cashman and Robinson, An Introduction to the Causes of War, chapter 6. 73. Blainey, The Causes of War, 81. 74. Zeev Maoz, “Joining the Club of Nations: Political Development and International Conflict, 1816–1976,” International Studies Quarterly 32, no. 2 (June 1989): 199–231. 75. The presence of revolutionary change within states has an international system level effect as well. Maoz found that the level of stability in the international system is sensitive to the way in which new states join the system and to the way in which states transform themselves politically. The more revolutionary change in the system, the greater the number of militarized disputes in the system. K. J. Holsti’s study of war and issues is indirectly supportive of Maoz’s results. Holsti finds that the creation of nation-states has been a major source of war since the late eighteenth century and has been the most prevalent war-generating issue in the post-1945 era—a period in which over 50 percent of the wars were related to state creation. (See Holsti, Peace and War, 311–12.) Not all the wars involved with this issue, however, could be categorized as interstate wars. Many would be colonial wars or wars of national liberation. 76. Stephen M. Walt, Revolution and War (Ithaca, NY: Cornell University Press, 1996). See also an earlier version in Stephen M. Walt, “Revolution and War,” World Politics 44, no. 3 (1992): 321–68. 77. Some revolutions, like the Mexican Revolution, did not produce this effect. The Mexican Revolution was not viewed as a threat largely because it was not deemed likely to spread. (This was also true of the elite revolution in Turkey and perhaps the American Revolution.) Walt, Revolution and War, 287. 78. This is an argument in Walt’s earlier article that is not present in the later book. 79. Walt, “Revolution and War,” 325, lists this as an example of war between the U.S. and Nicaragua even though the U.S. did not participate directly in the conflict. 80. See Walt, Revolution and War, chapter 5 and Cashman and Robinson, An Introduction to the Causes of War, chapter 6. 81. Dilip Hiro, The Longest War: The Iran-Iraq Military Conflict (New York: Routledge, Chapman and Hall, 1991), 32; quoted in Cashman and Robinson, An Introduction to the Causes of War, 277. 82. Blainey, The Causes of War, 83. 83. For a review of the literature, see John Vasquez, “Factors Relating to the Contagion and Diffusion of International Violence,” in The Internationalization of Communal Strife, ed. Manus Midlarsky (New York: Routledge, 1993), 149–72; and also Patrick M. Regan, “Choosing to Intervene: Outside Interventions in Internal Conflicts,” Journal of Politics 60, no. 3 (1998): 754–79. 84. Vasquez, “Factors Related to the Contagion and Diffusion of International Violence,” 160–61. 85. Gleditsch et al., “Fighting at Home, Fighting Abroad.” 86. Gleditsch et al., “Fighting at Home, Fighting Abroad,” 487–88. The three data sets they use are the COW Intra-State Wars, the Uppsala Armed Conflict Data, and a list generated by Fearon and Laitin. 87. Gleditsch, Salehyan, and Schultz, 488 and Table 1 on 489. 88. Gleditsch et al., “Fighting at Home, Fighting Abroad,” 488. 89. Gleditsch et al., “Fighting at Home, Fighting Abroad,” 500. 90. And, of course, all of this assumes that internal conflict precedes external conflict. The relationship may proceed from external conflict to internal conflict as well. In fact, there may be a reciprocal, interactive effect. See Levy, “The Diversionary Theory of War,” 259–88. 91. See Cashman and Robinson, An Introduction to the Causes of War, chapter 3, especially 100–102. 92. Paul Huth, Standing Your Ground (Ann Arbor: University of Michigan, 1996), 78–83, 103–12. Huth cautions that while the presence of bordering minorities is an excellent predictor of whether disputes will escalate, they are not a good predictor of why countries get involve in disputes in the first place. There are just as many situations in which one state’s people exist as minorities in a neighboring state and there is no territorial dispute. 93. Stephen Van Evera, “Hypotheses on Nationalism and War,” International Security 18, no. 4 (1994): 5–39. 94. Van Evera, “Hypotheses on Nationalism and War,” 14. 95. Van Evera, “Hypotheses on Nationalism and War,” 15. 96. Van Evera, “Hypotheses on Nationalism and War,” 8–9. 97. Benjamin Miller, States, Nations, and the Great Powers: The Sources of Regional War and Peace (Cambridge: Cambridge University Press, 2007).

526

Notes

98. Miller, States, Nations, and the Great Powers, 2. 99. Miller, States, Nations, and the Great Powers, 56. 100. Miller, States, Nations, and the Great Powers; see table 2.2 on 58 and 3.2 on 101. 101. Miller, States, Nations, and the Great Powers, 97–99. 102. Miller, States, Nations, and the Great Powers, 156. 103. Miller, States, Nations, and the Great Powers, 20. 104. Miller, States, Nations, and the Great Powers, 82–83. 105. Miller, States, Nations and the Great Powers, 121–27. 106. Miller, States, Nations, and the Great Powers, 108–10. 107. Miller, States, Nations, and the Great Powers, 64–65. 108. Miller, States, Nations, and the Great Powers, 370. 109. Miller, States, Nations, and the Great Powers, 111–12. 110. Miller, States, Nations, and the Great Powers, 111; see Kalevi J. Holsti, Peace and War: Armed Conflicts and International Order 1648–1989 (Cambridge: Cambridge University Press, 1991), 311–12. 111. Miller, States, Nations, and the Great Powers, 112–15. 112. Miller, States, Nations, and the Great Powers, 115–20. 113. Arnold Toynbee, A Study of History, Vol. IX (London: Oxford University Press, 1954), 322–23. 114. Blainey, The Causes of War, 7. 115. Levy and Morgan suggest, however, that regime type may make a difference. It is possible that war may be positively addictive for some regime types and negatively addictive for others. Jack Levy and T. Clifton Morgan, “The War Weariness Hypothesis: An Empirical Test,” American Journal of Political Science 30 (1986): 31. 116. Lewis Richardson, Arms and Insecurity (Chicago: Quadrangle, 1960), 232. 117. Richardson makes a feeble attempt to salvage what he can of the theory by suggesting that at least it provides a good explanation for the rapid demise of the French army and that government’s subsequent surrender to the Nazis. He says, “The French had entered an inter-war breathing space in a mood of disillusionment and discouragement that had been registered in action eventually in France’s collapse and capitulation in June 1940.” Richardson, Statistics of Deadly Quarrels, 495. 118. On positive reinforcement, see John Nevin, “War Initiation and Selection by Consequences,” Journal of Peace Research 33, no. 1 (1996): 99–108. For other potential reasons for a link between war winning and repeated wars, see Levy and Morgan, “The War Weariness Hypotheses,” 28; and Vasquez, The War Puzzle, chapter 6, especially 207–10. 119. Vasquez, The War Puzzle, 207–10. 120. P. Karsten, Soldiers and Society: The Effects of Military Service and War in American Life (Westport, CT: Greenwood, 1978), cited in Francis A. Beer, Peace and War (San Francisco: W. H. Freeman, 1981), 293. 121. Robert J. Lifton, Home From the War: Vietnam Veterans—Neither Victims nor Executioners (New York: Simon & Schuster, 1973), and M. Janowitz, The Professional Soldier (Glencoe, IL: Free Press, 1960), cited in Beer, 298. 122. See, for instance, Richard J. Barnet, The Roots of War: The Men and Institutions Behind U.S. Foreign Policy (New York: Penguin, 1973). 123. See Blainey, The Causes of War, 17 and 108–24; and Levy and Morgan, “The War Weariness Hypothesis,” 28–29. 124. David Garnham, “War-Proneness, War-Weariness, and Regime Type: 1816–1980,” Journal of Peace Research 23, no. 3 (1986): 279–89, and his “Explaining Major Power Bellicosity and Pacifism,” paper presented at International Studies Association conference (April 1982). 125. Levy and Morgan, “The War Weariness Hypothesis,” 35–39. Levy and Morgan conclude that while their findings on the distribution of war unambiguously contradict the war-weariness hypothesis, they are not strong enough to support the contrary hypothesis of positive contagion. 126. Levy and Morgan, “The War-Weariness Hypothesis,” 46–47. 127. Singer and Small, Wages of War, 283–84. 128. J. David Singer and Melvin Small, “Foreign Policy Indicators: Predictors of War in History and the State of the World Message,” Policy Sciences 5, no. 3 (September 1974): 271–96. 129. J. David Singer and Thomas Cusack, “Periodicity, Inexorability and Steermanship in International War,” in From National Development to Global Community, ed. Richard Merritt and Bruce Russett (London: Allen and Unwin, 1981), 413–15.

Notes

527

130. Singer and Cusack, “Periodicity, Inexorability and Steermanship in International War,” 415–17. This relationship is statistically significant for international wars, but not for interstate wars. Singer and Cusack find no relationship between the costliness of war and the length of the interwar interval. 131. Nevin, “War Initiation and Selection by Consequences.” 132. Kevin Wang and James Lee Ray, “Beginners and Winners: The Fate of Initiators of Interstate Wars Involving Great Powers Since 1945,” International Studies Quarterly 38, no. 1 (March 1994): 139–54. 133. Geller and Singer’s review of the literature comes to a similar conclusion. Geller and Singer, Nations at War, 49–50.

7. THE DYADIC LEVEL OF ANALYSIS, PART I 1. Zeev Maoz, “Pacifism and Fightaholism in International Politics: A Structural History of National and Dyadic Conflict, 1816–1992,” International Studies Review 6, no. 4 (2004): 107–33. 2. For a review of the evidence see John A Vasquez, The War Puzzle (Cambridge: Cambridge University Press, 1993), 123–39; John A. Vasquez, “Why Do Neighbors Fight? Proximity, Interaction or Territoriality,” Journal of Peace Research 32, no. 3 (1995): 277–93; Paul R. Hensel, “Theory and Evidence on Geography and Conflict,” in What Do We Know about War? ed. John A. Vasquez (Lanham, MD: Rowman & Littlefield, 2000), 57–84; Paul R. Hensel, “Territory: Geography, Contentious Issues, and World Politics,” in What Do We Know About War? 2nd ed., ed. John Vasquez (Lanham, MD: Rowman & Littlefield, 2012), 3–26; Paul K. Huth, “Why Are Territorial Disputes between States a Central Cause of International Conflict?” in Vasquez, What Do We Know About War? 85–110. 3. Melvin Small and J. David Singer, Resort to Arms: International and Civil Wars, 1816–1980 (Thousand Oaks, CA: Sage, 1982). 4. Kalevi J. Holsti, Peace and War: Armed Conflicts and International Order, 1648–1989 (Cambridge: Cambridge University Press, 1991). 5. Paul Hensel, “Theory and Evidence on Geography and Conflict,” 64–65. 6. Peter Wallensteen, “Incompatibility, Confrontation, and War: Four Models and Three Historical Systems, 1816–1976,” Journal of Peace Research 18, no. 1 (1981): 57–90. 7. Stuart Bremer, “Dangerous Dyads: Conditions Affecting the Likelihood of Interstate War, 1816–1965,” Journal of Conflict Resolution 36, no. 2 (June, 1992): 309–41. According to Bremer, the conditions that characterize the most war prone dyads are, from most important to least important: (1) contiguity, (2) absence of mutual alliance, (3) at least one of the states lacks an advanced economy, (4) at least of the states is nondemocratic, (5) absence of overwhelming preponderance by one state over the other, and (6) at least one state is a great power. Bremer does not include rivalry as a variable in his study. 8. D. Scott Bennett and Allan C. Stam, The Behavioral Origins of War (Ann Arbor: University of Michigan Press, 2004), 115. 9. Paul Hensel, “Territory: Geography, Contentious Issues, and World Politics,” 6, table 1.1. 10. Paul F. Diehl, “Contiguity and Military Escalation in Major Power Rivalries, 1816–1980,” Journal of Politics 47, no. 4 (November, 1985): 1203–11. 11. Lewis Fry Richardson, Statistics of Deadly Quarrels (New York: Quadrangle/New York Times, 1960),176. See also Harvey Starr and Benjamin Most, “The Substance and Study of Borders in International Relations Research,” International Studies Quarterly 20 (1976): 581–620, and H. Starr and B. Most, “A Return Journey: Richardson, ‘Frontiers,’ and Wars in the 1946–1965 Era,” Journal of Conflict Resolution 22 (1978): 441–67. Starr and Most find the general relationship to be true, but demonstrate that it varies according to the time period, data set, states investigated, types of borders, and statistics used. 12. Harvey Starr and Benjamin Most, “A Return Journey: Richardson, ‘Frontiers,’ and Wars in the 1945–1965 Era,” Journal of Conflict Resolution 22, no. 3 (1978): 441–62; Randolph Siverson and Harvey Starr, “Opportunity, Willingness and the Diffusion of War,” American Political Science Review 84 (1990): 47–67; Randolph Siverson and Harvey Starr, The Diffusion of War: A Study of Opportunity and Willingness (Ann Arbor: University of Michigan Press, 1991); Kristian Gleditsch, Idean Salehyan, and Kenneth Schultz, “Fighting at Home, Fighting Abroad: How Civil Wars Lead to International Disputes,” Journal of Conflict Resolution 52, no. 4 (2008): 479–506. 13. Kenneth Boulding, Conflict and Defense (New York: Harper & Row, 1962), 227–47. See also Douglas Lemke, “The Tyranny of Distance: Redefining Relevant Dyads,” International Interactions 21 (1995): 23–38. 14. Bruce Bueno de Mesquita, The War Trap (New Haven, CT: Yale University Press, 1981), 166.

528

Notes

15. For a discussion of the concepts of opportunity and willingness see Benjamin Most and Harvey Starr, Inquiry, Logic, and International Politics (Columbia: University of South Carolina Press, 1989). Most and Starr argue that opportunity and willingness are jointly necessary conditions for some types of international behavior. For a general review of the theoretical explanations for the link between territory and conflict, see Paul Hensel, “Theory and Evidence on Geography and Conflict,” 58–63. See also Paul Diehl, “Geography and War: A Review and Assessment of the Empirical Literature,” International Interactions 17, no. 1 (1991): 11–27. 16. See John Vasquez, “Why Do Neighbors Fight?” 280, citing Charles Gochman, “The Geography of Conflict: Militarized Interstate Disputes since 1816,” paper presented to the annual meeting of the International Studies Association, Washington, DC, 1990, 10–14. And see Paul Hensel, “Theory and Evidence on Geography and Conflict.” 17. See for instance, Karl Deutsch et al., Political Community and the North Atlantic Area (Princeton, NJ: Princeton University Press, 1957), and R. W. Cobb and C. Elder, International Community (New York: Holt, Rinehart and Winston, 1970). 18. John A. Vasquez, “Why Do Neighbors Fight?” 19. Paul Senese and John Vasquez, The Steps to War: An Empirical Study (Princeton, NJ: Princeton University Press, 2008), 104–11. 20. John A. Vasquez, The War Puzzle; John A. Vasquez, “Why Do Neighbors Fight?”; John Vasquez and Marie T. Henehan, “Territorial Disputes and the Probability of War, 1816–1992,” Journal of Peace Research 38, no. 2 (2001): 123–38; Paul D. Senese and John A Vasquez, “A Unified Explanation of Territorial Conflict: Testing the Impact of Sampling Bias, 1919–1992,” International Studies Quarterly 47 (2003): 275–98. Territorial disputes are an underlying cause of war because while they do not lead to war immediately, they touch off a sequence of events that end in war. Territorial disputes are a causal factor in the sense that the presence of territorial disputes makes war possible and the absence of such disputes makes war highly unlikely. (Vasquez, The War Puzzle, 124.) 21. Hensel, “Theory and Evidence on Geography and Conflict.” 22. Senese and Vasquez, The Steps to War, 104–25. The finding suggests that one reason contiguous states are involved in a large number of wars is because they are involved in a large number of militarized disputes to begin with, not because contiguity causes MIDs to escalate. This builds on the initial finding in Paul Senese, “Territory, Contiguity, and Interstate Conflict: Assessing a New Joint Explanation,” American Political Science Review 4 (2005): 769–79. 23. Vasquez, “Why Do Neighbors Fight?” 282. 24. Vasquez, “Why Do Neighbors Fight?” 283, drawing on Vasquez, The War Puzzle, 140–41. In Vasquez’s view, however, neither territoriality nor territorial issues lead directly to war. Much depends on how states handle their territorial conflicts. Given the fact that humans have territorial tendencies, under certain conditions, this provides a motivation or willingness for war. A territorial dispute is simply the most likely first step of a path toward war. (This is Vasquez’s steps-to-war theory, which we will discuss later.) The interactions of the states as they handle their disputes will determine whether war will occur or can be avoided. If boundaries are mutually accepted, then the probability of war between neighbors will be reduced dramatically. On the other hand, if states resort to realpolitik policies and coercive tactics to press their territorial claims, then militarized disputes and crises are likely to recur and escalation to war becomes more likely as leaders tend to adopt more coercive strategies. 25. Hensel, “Territory: Geography, Contentious Issues, and World Politics,” 18–19. 26. James Fearon, “Rationalist Explanations for War,” International Organization 49 (1995): 379–414; see especially 386–90. 27. Paul Huth, Standing Your Ground: Territorial Disputes and International Conflict (Ann Arbor: University of Michigan Press, 1996), 171–79. 28. This argument is central to Paul Huth’s “modified realist model” for explaining territorial conflicts. Huth, Standing Your Ground. See also Huth, “Why Are Territorial Disputes between States a Central Cause of International Conflict?”, in Vasquez, What Do We Know About War? 85–110, especially 100–101; and Paul D. Senese and John A. Vasquez, “A Unified Explanation of Territorial Conflict: Testing the Impact of Sampling Bias, 1919–1992,” International Studies Quarterly 47 (2003): 275–98; see 278. For a general theory of the domestic political considerations of decision makers in a rational choice model, see Bruce Bueno de Mesquita and David Lalman, War and Reason (New Haven, CT: Yale University Press, 1992), and B. Bueno de Mesquita, A. Smith, R. Siverson, and J. Morrow, The Logic of Political Survival (Cambridge: MIT Press, 2003). 29. Evan Luard, War in International Society (New Haven, CT: Yale University Press, 1986), 83–84. He also devotes a separate chapter to the motives of states, compared with the issues over which wars are fought. The difference is not entirely distinct in practice. The following discussion is based on chapter 3, pp. 83–131.

Notes

529

30. Although Luard’s book was published in 1986, we can assume that he would want to extend the Age of Ideology to 1989 or 1991, that is, until the end of the Cold War. 31. John Vasquez, The War Puzzle, 131–32; Paul Huth, “Why Are Territorial Disputes between States a Central Cause of International Conflict?” 89. 32. Kalevi. J. Holsti, Peace and War: Armed Conflicts and International Order 1648–1989 (Cambridge: Cambridge University Press, 1991). 33. K. J. Holsti, Peace and War, 308, table 12.2. 34. K. J. Holsti, Peace and War, 307–10, tables 12.1 and 12.2. 35. Hensel, “Territory: Theory and Evidence on Geography and Conflict,” 65–66. 36. Vasquez, “Why Do Neighbors Fight?” 284–85. 37. John A. Vasquez and Brandon Valeriano, “Classification of Interstate Wars,” Journal of Politics 72 (2010): 292–309. 38. Stephen A. Kocs, “Territorial Disputes and Interstate War,” Journal of Politics 57, no. 1 (February 1995): 159–75, 172. 39. Vasquez and Henehan, “Territorial Disputes and the Probability of War, 1816–1992,” Journal of Peace Research 38 (March 2001): 123–38. 40. The findings are confirmed for the years 1816–2001 by Senese and Vasquez, in their The Steps to War, as well as Paul Hensel, “Charting a Course to Conflict: Territorial Issues and Interstate Conflict, 1816–1992,” Conflict Management and Peace Science 15 (Spring 1996): 43–73. See the review of this literature in John A. Vasquez, The War Puzzle Revisited (Cambridge: Cambridge University Press, 2009), chapter 10. 41. Paul Huth, Standing Your Ground. 42. Huth, Standing Your Ground, 69. 43. Huth, Standing Your Ground, 34. 44. Huth, Standing Your Ground, 103. 45. Huth, Standing Your Ground, 103–13. Since Huth had earlier found that irredentist claims were not a very strong predictor for dispute initiation, due largely to strong international norms against such claims, he reasons that only those states most committed to pursuing the claims would make them and that these states were also the ones most likely to engage in the use of strong actions to overturn the status quo. 46. Paul Huth and Todd Allee, The Democratic Peace and Territorial Conflict in the Twentieth Century (Cambridge: Cambridge University Press, 2002). 47. Huth, Standing Your Ground, 108–113. 48. Hensel, “Territory: Geography, Contentious Issues and World Politics,” 18–19; Paul Huth and Todd Allee, The Democratic Peace and Territorial Conflict in the Twentieth Century (Cambridge: Cambridge University Press, 2002). Several of Huth and Allee’s indicators of territorial salience are related to ethno-nationalism: ties to a bordering minority and the presence of shared ethnic or linguistic groups along the border. 49. In many of these instances the challenger used force merely to probe the target’s response or create a fait accompli that would buttress its bargaining position, rather than achieve a territorial victory. 50. Huth, Standing Your Ground, 116–128. Three other hypotheses produced strong results. Disputes were likely to escalate if (1) negotiations in the previous year had ended in stalemate, (2) the target state had taken political, economic, or military actions to alter the territorial status quo, and (3) there was a history of militarized conflict over territorial issues. In these situations, domestic pressures push leaders toward the use of confrontational policies, though the preference for escalation can be reduced by the hard realities of military weakness (128–36). 51. Huth, Standing Your Ground, 183. 52. Jaroslav Tir, “Keeping the Peace after Secession: Territorial Conflict between Rump and Secessionist States,” Journal of Conflict Resolution 49, no. 5 (October 2005): 713–41. 53. Vasquez, The War Puzzle, 132–33. 54. See Greg Cashman and Leonard C. Robinson, An Introduction to the Causes of War: Patterns of Interstate Conflict from WWI to Iraq (Lanham, MD: Rowman & Littlefield, 2007), 370–71. 55. Cashman and Robinson, An Introduction to the Causes of War, 367–69. 56. Michael Brecher and Jonathan Wilkenfeld, A Study of Crisis (Ann Arbor: University of Michigan Press, 2000). 57. Hemda Ben-Yehuda, “Territoriality and War in International Crises: Theory and Findings, 1918–2001,” International Studies Review 5 (2004): 85–105. 58. Brecher and Wilkenfeld, A Study of Crisis, 156, 163.

530

Notes

59. More appropriately, the ICB data show that crises are more likely to escalate if military-security issues are involved, and the ICB project codes territorial disputes as a military-security issue. Huth, “Why are Territorial Disputes between States A Central Cause of International Conflict?” 91. 60. Paul Hensel, “Territory: Theory and Evidence on Geography and Conflict,” 65, 72; Hensel, “One Thing Leads to Another: Recurrent Militarized Disputes in Latin America, 1816–1986,” Journal of Peace Research 31 (August 1994): 281–98; and Hensel, “Interstate Rivalry and the Study of Militarized Conflict,” in Conflict in World Politics: Advances in the Study of Crises, War and Peace, ed. F. P. Harvey and B. D. Mo (London: Macmillan 1998), 162–204; P. Huth, Standing Your Ground. See also the discussion in Huth, “Why Are Territorial Disputes between States a Central Cause of International Conflict?” 90. 61. P. D. Senese and J. A. Vasquez, The Steps to War, 92, 97; P. D. Senese, “Geographic Proximity and Issue Salience: Their Effects on the Escalation of Militarized Interstate Conflict,” Conflict Management and Peace Science 15 (Fall 1996): 133–61; and P. Hensel, “Territory: Theory and Evidence on Geography and Conflict.” 62. Senese and Vasquez, “A Unified Theory of Territorial Conflict,” 280. 63. Senese and Vasquez, “A Unified Theory of Territorial Conflict,” 292 and table 3. 64. Senese and Vasquez, “A Unified Theory of Territorial Conflict,” 290, 295. Senese and Vasquez use an extension of Huth’s original 1950–1992 data on territorial conflict, which also covers the period between 1919 and 1950. 65. Paul Hensel, “Charting a Course to Conflict;” 1996; see also Hensel, “One Thing Leads to Another”; Hensel, “Interstate Rivalry and the Study of Militarized Conflict”; and Huth, Standing Your Ground, 200. 66. P. Hensel and P. Diehl, “It Takes Two to Tango: Non-Militarized Response in Interstate Disputes,” Journal of Conflict Resolution 38 (September 1994): 479–506; Sara McLaughlin Mitchell and Brandon Prins, “Beyond Territorial Contiguity: Issues at Stake in Democratic Militarized Interstate Disputes,” International Studies Quarterly 43 (March 1999): 169–83, 174. 67. Jaroslav Tir and Paul Diehl, “Geographic Dimensions of Enduring Rivalries,” Political Geography 21 (February 2002): 263–86. 68. Huth, “Why Are Territorial Disputes between States a Central Cause of International Conflict?” 92–93. 69. Vasquez, The War Puzzle, 146. 70. Stephen Kocs, “Territorial Disputes and Interstate War, 1945–1987,” Journal of Politics 57 (February 1995): 159–75. 71. Huth, Standing Your Ground, 86, 90–91. 72. Andrew Owsiak, “Signing Up for Peace: International Boundary Agreements, Democracy, and Militarized Interstate Conflict,” International Studies Quarterly 56, no. 1 (March 2012): 51–66. 73. Douglas Gibler, “Alliances That Never Balance: The Territorial Settlement Treaty,” Conflict Management and Peace Science 15 (Spring 1996): 75–97; and Gibler, “Control the Issues, Control the Conflict: The Effects of Alliances that Settle Territorial Issues on Interstate Rivalries,” International Interactions 22 (April 1997): 341–68. 74. Hensel, “Territory: Geography, Contentious Issues, and World Politics,” 13. 75. Hensel, “Territory: Geography, Contentious Issues, and World Politics,” 20. 76. Vasquez, The War Puzzle, 146–51. 77. Benjamin Barber, Jihad vs. McWorld: How Globalism and Tribalism Are Reshaping the World (New York: Ballantine Books, 1996). 78. Douglas Woodwell, “Unwelcome Neighbors: Shared Ethnicity and International Conflict during the Cold War,” International Studies Quarterly 48 (2004): 197–223. 79. Woodwell, “Unwelcome Neighbors,” 201. 80. Minority-minority dyads show no clear pattern. Woodwell concludes that there is no theoretical or empirical reason to believe that such dyads foster interstate conflict. 81. Woodwell, “Unwelcome Neighbors,” 217. 82. See, for instance, Paul F. Diehl and Gary Goertz, War and Peace in International Rivalry (Ann Arbor: University of Michigan Press, 2001), 17–48. See also Gary Goertz and Paul F. Diehl, “(Enduring) Rivalries,” in Handbook of War Studies II, ed. Manus Midlarsky (Ann Arbor: University of Michigan Press, 2000), 222–267; and Gary Goertz and Paul F. Diehl, “Rivalries: the Conflict Process,” in Vasquez, What Do We Know About War? 197–217. A militarized international dispute (MID) involves an interaction between a pair of states that involves either a threat to use military force, a display of military force, or an actual use of military force. The Correlates of War (COW) project collects and updates a MID data set. 83. James Klein, Gary Goertz, and Paul F. Diehl, “The New Rivalry Data Set: Procedures and Patterns,” Journal of Peace Research 43, no. 3 (2006): 34. A more up-to-date report, using slightly different coding rules, identifies 330

Notes

531

rivals, of which 152 are enduring rivalries and 178 are proto-rivalries. Paul F. Diehl and Gary Goertz, “The Rivalry Process: How Rivalries Are Sustained and Terminated,” in Vasquez, What Do We Know About War? 2nd ed., 83–109. 84. William R. Thompson, “Identifying Rivals and Rivalries in World Politics,” International Studies Quarterly 45 (2001): 557–86; Michael P. Colaresi, Karen Rasler, and William R. Thompson, Strategic Rivalries in World Politics (Cambridge: Cambridge University Press, 2007). 85. Colaresi, Rasler, and Thompson, Strategic Rivalries in World Politics, chapter 3. 86. Diehl and Goertz, War and Peace in International Rivalry, 77. This is similar to the effect that the “shadow of the future” has on sequential or iterative games, which we will discuss in chapter 9. 87. John A. Vasquez, “Distinguishing Rivals That Go to War from Those That Do Not: A Quantitative Case Study of the Two Paths to War,” International Studies Quarterly (1996) 40: 531–58, 532. 88. Diehl and Goertz, War and Peace in International Rivalry,78, treat learning as “an intervening variable mediating the impact of the past on the present.” 89. Wallensteen, “Incompatibility, Confrontation, and War.” 90. Diehl and Goertz, War and Peace in International Rivalry, 7. War can also contribute to the initiation of rivalries. 91. Vasquez, The War Puzzle, 210–16. 92. Colaresi et al., Strategic Rivalries in World Politics, 101. 93. Colaresi et al., Strategic Rivalries in World Politics, 28. 94. Colaresi et al., Strategic Rivalries in World Politics, 101–9. 95. Colaresi et al., Strategic Rivalries in World Politics, 111. 96. Brandon Valeriano, “Becoming Rivals: The Process of Rivalry Development,” in Vasquez, What Do We Know About War? 2nd ed., 63–82; see 63, 74. 97. Frank Whelon Wayman, “Rivalries: Recurrent Disputes and Explaining War,” in Vasquez, What Do We Know About War? 219–34; see especially 227. 98. Goertz and Diehl, “Rivalries: The Conflict Process,” 200; Diehl and Goertz, War and Peace in International Rivalry, 60–64. 99. Gary Goertz, Contexts of International Politics (Cambridge: Cambridge University Press, 1994), 208–12. 100. Klein, Goertz, and Diehl, “The New Rivalry Data Set.” Diehl and Goertz, “The Rivalry Process,” 87–88, use somewhat different coding, and report slightly different numbers. 101. Colaresi et al., Strategic Rivalries in World Politics, 88–89. 102. Colaresi et al., Strategic Rivalries in World Politics, 90. 103. Colaresi et al., Strategic Rivalries in World Politics, 119–27. On the other hand, democracy within a rivalry actually seems to have more of a pacifying effect on crises than it does between nonrivals, and rivals do not exploit domestic instability of others in a crisis significantly more than do nonrivals. 104. Colaresi et al., Strategic Rivalries in World Politics, 155–56. 105. Vasquez, The War Puzzle, 123–52; Vasquez, “Distinguishing Rivals That Go to War from Those That Do Not.” 106. Vasquez, “Distinguishing Rivals That Go to War from Those That Do Not.” 107. Vasquez, “Why Do Neighbors Fight?” 285. 108. Vasquez, “Distinguishing Rivals,” 542–47. 109. Vasquez, “Distinguishing Rivals,” 449–50. 110. Vasquez, “Distinguishing Rivals,” 553. 111. Colaresi et al., Strategic Rivalries in World Politics, 80; William R. Thompson, “Principal Rivalries,” Journal of Conflict Resolution 39 (June 1995): 195–223. 112. Colaresi et al., Strategic Rivalries in World Politics, 256, report that the three-way combination of territorial disputes/contiguity/rivalry is associated with a 132 percent probability increase in war onset, compared with a 24 percent probability increase for the combination of territorial dispute/contiguity only. Adding mixed dyads increases the probability of war onset to 149 percent, compared with a 32 percent increased probability for mixed dyads involved in territorial disputes outside of a rivalry. See also K. Rasler and W. Thompson, “Contested Territory, Strategic Rivalries, and Conflict Escalation,” International Studies Quarterly 50, no. 1 (March 2006): 145–67. 113. Colaresi et al., Strategic Rivalries in World Politics, 257. 114. Since most of these theories predate the use of the term “enduring rivals,” these theories have instead used the terms “contenders,” “world powers” and “regional powers,” “hegemonic powers,” and “great power challengers.” However, the states involved would surely make virtually any current list of enduring rivals or strategic rivals.

532

Notes

115. Diehl and Goertz, War and Peace in International Rivalry, 7. 116. Frank W. Wayman, “Power Shifts and the Onset of War,” in Parity and War, ed. Jacek Kugler and Douglas Lemke (Ann Arbor: University of Michigan Press, 1996), 145–62. 117. Daniel Geller, “The Stability of the Military Balance and War among Great Power Rivals,” 165–190, 184, in The Dynamics of Enduring Rivalries, ed. P. F. Diehl (Urbana: University of Illinois Press, 1998). 118. Henk Houweling and Jan Siccama, “Power Transitions and Critical Points as Predictors of Great Power War: Toward a Synthesis,” Journal of Conflict Resolution 35 (1991): 642–58; Daniel Geller, “Power Transition and Conflict Initiation,” Conflict Management and Peace Science 12 (1992): 1–16; Paul Huth, D. Scott Bennett, and Christopher Gelpi, “System Uncertainty, Risk Propensity, and International Conflict among the Great Powers,” Journal of Conflict Resolution 36 (1992): 478–517. 119. Toby Rider, Michael Findley, and Paul F. Diehl, “Just Part of the Game? Arms Races, Rivalry, and War,” Journal of Peace Research 48 (2011): 85–100. 120. Sara McLaughlin Mitchell and Brandon Prins, “Rivalry and the Diversionary Uses of Force,” Journal of Conflict Resolution 48 (2004): 937–81. 121. Dean Babst, “Elective Governments—A Force for Peace,” The Wisconsin Sociologist 3, no. 1 (1964): 9–14; Babst, “A Force for Peace,” Industrial Research (April 1972): 55–58. 122. Melvin Small and J. David Singer, “The War Proneness of Democratic Regimes,” Jerusalem Journal of International Relations 1 (1976): 49–69. 123. Small and Singer suggested the difference might simply be due to lack of contiguous democracies during a period of time when democracies were relatively rare on the world scene. 124. Found originally in Rudolph J. Rummel, Understanding Conflict and War, Vol.4: War, Power and Peace (Beverly Hills, CA: Sage, 1979), 277–79. 125. Rudolph J. Rummel, “Libertarianism and International Violence,” Journal of Conflict Resolution 27, no. 1 (March 1983): 27–71. 126. Michael Doyle, Ways of War and Peace (New York: Norton, 1997), chapter 8. Doyle’s definition of a liberal regime, based on Kant’s criteria, includes the existence of private property and market economy, external sovereignty, citizens who possess juridical rights, republican representative government (including parliamentary monarchy), and a legislature with an effective role in policy making and being formally and competitively elected. 127. Spencer Weart, Never at War: Why Democracies Will Not Fight One Another (New Haven, CT: Yale University Press, 1998), 13–14. 128. In addition to Babst, Rummel, and Small and Singer, cited above, for evidence of the democratic peace see Zeev Maoz and Nasrin Abdolali, “Regime Types and International Conflict, 1816–1976,” Journal of Conflict Resolution 33(1989): 3–35; Zeev Maoz and Bruce Russett, “Normative and Structural Causes of Democratic Peace, 1946–1986,” American Political Science Review 87 (1993): 624–38; James Lee Ray, “Wars Between Democracies: Rare or Nonexistent?” International Interactions 18 (1993): 251–76; James Lee Ray, Democracy and International Conflict (Columbia: University of South Carolina Press, 1995); Bruce Russett, Grasping the Democratic Peace (Princeton, NJ: Princeton University Press, 1993); Steve Chan, “In Search of Democratic Peace: Problems and Promise,” Mershon International Studies Review 41 (1997): 59–91; Nils Petter Gleditsch and Havard Hegre, “Peace and Democracy: Three Levels of Analysis,” Journal of Conflict Resolution 41 (April 1997): 283–310; David Rousseau, C. Gelpi, D. Reiter, and P. Huth, “Assessing the Dyadic Nature of the Democratic Peace, 1918–88,” American Political Science Review 90, no. 3 (1996): 512–33; Spencer Weart, Never At War: Why Democracies Will Not Fight One Another (New Haven, CT: Yale University Press, 1998); and Bruce Russett and John Oneal, Triangulating Peace: Democracy, Interdependence, and International Organizations (New York: Norton, 2001). 129. It has been suggested, for instance, that political stability really accounts for peace between two democracies. The presence of joint democracy simply masks the underlying fact that these states don’t fight each other because they are both politically stable. In fact, Maoz and Russett show that controlling for the effects of political stability pretty much wipes out the strong relationship between democracy and peace. See Zeev Maoz and Bruce Russett, “Alliance, Contiguity, Wealth, and Political Stability: Is the Lack of Conflict among Democracies a Statistical Artifact?” International Interactions 17 (1992): 245–67. However, Ray notes that although it is quite likely that democracy leads to stability, it is not very likely that stability leads to democracy and peace. Therefore stability is best seen as a variable that intervenes between democracy and peace, rather than a confounding variable that leads to both democracy and peace. This would make the relationship between democracy and peace indirect, but not spurious. See James Lee Ray, “On the Level(s), Does Democracy Correlate with Peace?” in Vasquez, What Do We Know About War? 299–327.

Notes

533

130. Jack Levy, “Domestic Politics and War,” in R. Rotberg and T. Rabb, The Origin and Prevention of Major Wars (Cambridge: Cambridge University Press, 1988), 88. 131. James Lee Ray, “Does Democracy Cause Peace?” Annual Review of Political Science 1 (1998): 27–46. 132. Ray, “On the Level(s), Does Democracy Correlate with Peace?” 133. Chan, “In Search of Democratic Peace,” 71. 134. Chan, “In Search of Democratic Peace,” 72. 135. Russett and Oneal, Triangulating Peace, 46. 136. Gleditsch and Hegre, “Peace and Democracy,” table 2, 288. 137. Bruce Russett, “The Fact of Democratic Peace,” in Debating the Democratic Peace, ed. M. Brown. S. LynnJones, and S. Miller (Cambridge, MA: MIT Press, 1996), 58–81; see table 1.1, 75. 138. Russett, “The Fact of Democratic Peace,” 75–76. 139. Russett, “The Fact of Democratic Peace,” 76. 140. Gleditsch and Hegre, “Peace and Democracy,” 286–89; Russett, “The Fact of Democratic Peace,” 76–78. 141. Ido Oren, “The Subjectivity of the ‘Democratic’ Peace,” International Security 20, no. 2 (1995): 147–84. 142. Doyle, Ways of War and Peace, 288; Russett, “The Fact of Democratic Peace,” 77. 143. John Owen, “How Liberalism Produces Democratic Peace,” International Security 19, no. 2 (1994): 87–125. 144. Russett and Oneal, Triangulating Peace, 48. 145. Maoz and Abdolali, “Regime Types and International Conflict, 1816–1976”; Chan, “In Search of Democratic Peace,” 82–83; John Oneal and Bruce Russett, “The Classical Liberals Were Right: Democracy, Interdependence, and Conflict, 1950–1985,” International Studies Quarterly 41, no. 2 (1997): 267–94. 146. Edward Mansfield and Jack Snyder, Electing to Fight (Cambridge, MA: MIT Press, 2005), 162–63. 147. D. Scott Bennett, “Toward a Continuous Specification of the Democracy-Autocracy Connection,” International Studies Quarterly 50 (2006): 313–38. See also Suzanne Werner, “The Effects of Political Similarity on the Onset of Militarized Disputes, 1816–1985,” Political Research Quarterly 55, no. 2 (2000): 343–74. 148. Russett and Oneal, Triangulating Peace, 114–16. 149. D. Scott Bennett, “Democracy, Regime Change, and Rivalry Termination,” International Interactions 22, no. 4 (1997): 367–97. 150. Chan, “In Search of Democratic Peace,” 84. 151. Gleditsch and Hegre, “Peace and Democracy: Three Levels of Analysis,” 289–91. 152. Maoz and Abdolali, “Regime Types and Conflict,” 20–26. 153. Zeez Maoz and Bruce Russett, “Normative and Structural Causes of Democratic Peace,” American Political Science Review 87, no. 3 (1993): 624–38. 154. William Reed, “A Statistical Model of Conflict Onset and Escalation,” American Journal of Political Science 44, no. 1 (2000): 84–93. 155. Paul Senese and John Vasquez, “A Unified Explanation of Territorial Conflict: Testing the Impact of Sampling Bias, 1919–1992,” International Studies Quarterly 47 (2003): 275–298. 156. D. Scott Bennett and Allan C. Stam, The Behavioral Origins of War (Ann Arbor: University of Michigan, 2004). 157. Bennett and Stam, The Behavioral Origins of War, 128–131. 158. T. C. Morgan and V. L. Schwebach, “Take Two Democracies and Call Me in the Morning: A Prescription for Peace?” International Interactions 17, no. 4 (1992): 305–20. 159. Sara Mitchell and Brandon Prins, “Beyond Territorial Contiguity: Issues at Stake in Democratic Militarized Disputes” International Studies Quarterly 43, no. 1 (1999): 169–83. 160. William J. Dixon, “Democracy and the Peaceful Settlement of International Conflict,” American Political Science Review 88, no. 1 (1994): 14–32; William J. Dixon, “Democracy and the Management of International Conflict,” Journal of Conflict Resolution 37, no. 1 (1993): 42–68; William Dixon and Paul Senese, “Democracy, Disputes, and Negotiated Settlements,” Journal of Conflict Resolution 46, no. 4 (2002); 547–71; Gregory Raymond, “Democracies, Disputes, and Third-Party Intermediaries,” Journal of Conflict Resolution 38, no. 1 (1994): 24–42; Holley Hansen, Sara McLaughlin Mitchell, and Stephen Nemeth, “IO Mediation of Interstate Conflicts: Moving Beyond the Global vs. Regional Dichotomy,” Journal of Conflict Resolution 52, no. 2 (2998): 295–325. See also Huth and Allee, The Democratic Peace and Territorial Conflict, and also Todd Allee and Paul Huth, “Legitimizing Dispute Settlement: International Legal Rulings as Domestic Political Cover,” American Political Science Review 100, no. 2 (2006): 219–34. 161. See Harald Muller and Jonas Wolff, “Democratic Peace: Many Data, Little Explanation?” in Democratic Wars, ed. A. Geis, L. Brock, and H. Muller (New York: Palgrave Macmillan, 2006), 41–73.

534

Notes

162. See Sebastian Rosato, “The Flawed Logic of Democratic Peace Theory,” American Political Science Review 97, no. 4 (2003): 585–601. 163. Rosato, “The Flawed Logic of Democratic Peace Theory,” 590–91. 164. Bruce Russett, Grasping the Democratic Peace (Princeton, NJ: Princeton University Press, 1993), 120–24. 165. John Owen, “How Liberalism Produces Democratic Peace”; and John Owen, “Perceptions and the Limits of Liberal Peace: The Mexican-American and Spanish-American Wars,” in Paths to Peace, ed. Miriam Fendius Elman (Cambridge, MA: MIT Press, 1997), 153–89. 166. Owen (“How Liberalism Produces Democratic Peace”) distinguishes between liberal ideals and democratic institutions. It is possible for states with democratic institutions to be illiberal due to the prevailing norms and political culture. Thus, contemporary Iran may have democratic institutions, but its political elites “do not view the world through a liberal lens.” Elites in Western democracies are unlikely to see Iran as truly democratic therefore. On the other hand, states with some liberal elements may be undemocratic, like Britain prior to the Great Reform Act of 1832. Finally, liberal democracies may, from time to time, be governed by illiberal leaders who may attempt to implement foreign policies autonomously, but these leaders will be constrained from fighting other democracies by democratic institutions. 167. As Owen admits, there was not unanimity on this point. Republicans in the U.S. considered only republics (like France) to be democracies, while Federalists turned against France after the execution of Louis XVI in 1793. While Federalists under President Adams wanted to go to war with France in 1798, they were constrained by Republicans in Congress and the public. Owen, “How Liberalism Produces Democratic Peace,” 136. 168. Rosato, “The Flawed Logic of Democratic Peace Theory,” 592. 169. These cases are from Owen, “Perceptions and Limits of Democratic Peace.” 170. Mark Peceny, using a constructivist approach, also concludes that although Spain possessed democratic institutions, most Americans did not perceive Spain to be part of the global community of liberal-democratic states. Mark Peceny, “A Constructivist Interpretation of the Liberal Peace: The Ambiguous Case of the Spanish American War,” Journal of Peace Research 34, no. 4 (1997): 415–30. 171. For instance, see Bruce Bueno de Mesquita and David Lalman, War and Reason (New Haven, CT: Yale University Press, 1992); B. Bueno de Mesquita, Alastair Smith, Randolph Siverson, and James Morrow, The Logic of Political Survival (Cambridge: MIT Press, 2003); and J. D. Fearon, “Domestic Political Audiences and the Escalation of International Disputes,” American Political Science Review 88 (1994): 577–92. 172. See Muller and Wolff, “Democratic Peace: Many Data, Little Explanation?” 56. 173. Rosato, “The Flawed Logic of Democratic Peace Theory,” 593–94. See also Michael C. Desch, “Democracy and Victory: Why Regime Type Hardly Matters,” International Security 27, no. 2 (2002): 5–47. 174. Russett and Oneal use the triangle imagery in their Triangulating Peace, and I adopt their usage in this section. 175. Michael Doyle, “Liberalism and World Politics,” American Political Science Review 80, no. 4 (1986): 1151–63. See also Michael Doyle, Ways of War and Peace (New York: Norton, 1997), especially chapter 8. 176. Doyle, Ways of War and Peace, 287. 177. Doyle, “Liberalism and World Politics,” 1160. 178. Russett and Oneal, Triangulating Peace, 108–110. 179. Russett and Oneal, Triangulating Peace, 145–49. Their measure of trade interdependence or “economically important trade” is a state’s total trade with the other state divided by its GDP. Their measure of openness is the ratio of a state’s total global trade to its GDP. 180. Russett and Oneal, Triangulating Peace, 154–55. 181. Russett and Oneal, Triangulating Peace, 171–73. 182. Even controlling for the effects of trade and joint IGO membership, dual democracies are still 33 percent less likely than the average dyad to become involved in a MID (p. 275). Russett and Oneal’s analysis also shows that realist factors are also important. Great distances, preponderance of power, minor power status, and (after 1945) mutual alliances all reduce the probability of conflict for a pair of states. They note that “Kantian influences have not abolished power politics” (p. 195). 183. Russett and Oneal, Triangulating Peace, 197–228. 184. Ray, “Does Democracy Cause Peace?” 33. 185. David Shapiro, “The Insignificance of the Liberal Peace,” International Security 19, no. 2 (1994): 50–86. 186. Christopher Layne, “Kant or Cant: The Myth of the Democratic Peace,” International Security 19, no. 2 (1994): 5–49.

Notes

535

187. Bremer, “Dangerous Dyads”; Maoz and Abdolali, “Regime Types and International Conflict, 1816–1976”; Maoz and Russett, “Alliance, Contiguity, Wealth and Political Stability: Is the Lack of Conflict among Democracies a Statistical Artifact?”; Nils Petter Gleditsch, “Geography, Democracy and Peace,” International Interactions 20 (1995): 297–323. 188. Layne, “Kant or Cant.” See also Stephen R. Rock, “Anglo-US Relations, 1845–1930: Did Shared Liberal Values and Democratic Institutions Keep the Peace?” in Paths to Peace, ed. Miriam Fendius Elman (Cambridge, MA: MIT Press, 1997), 101–49, which comes to the same conclusion that liberal values are not an adequate explanation for the lack of war between the U.S. and Britain from 1845 to 1930. 189. Henry Farber and Joanne Gowa, “Polities and Peace,” International Security 20, no. 2 (1995): 123–46. 190. Rosato, “The Flawed Logic of Democratic Peace Theory,” 599–600. 191. Ray, “Does Democracy Cause Peace?” 38. 192. Douglas Gibler, “Bordering on Peace: Democracy, Territorial Issues, and Conflict,” International Studies Quarterly 51 (2007): 509–32; and Gibler, “The Implications of the Territorial Peace,” in Vasquez, What Do We Know about War? 2nd ed., 211–34. 193. John Vasquez, The War Puzzle Revisited (Cambridge: Cambridge University Press, 2009), 370. 194. Gibler, “Bordering on Peace,” 516. 195. Kristian S. Gleditsch, All Politics Is Local (Ann Arbor: University of Michigan Press, 2002). 196. Michael Mousseau, “Market Prosperity, Democratic Consolidation, and Democratic Peace,” Journal of Conflict Resolution 44, no. 4 (2000): 472–502; Mousseau, “An Economic Limitation to the Zone of Democratic Peace and Cooperation,” International Interactions 28 (2002): 137–64; Mousseau, “The Social Market Roots of Democratic Peace,” International Security 33 (2009): 52–86; and Mousseau, “A Market-Capitalist or a Democratic Peace?” in Vasquez, What Do We Know About War? 2nd ed., 189–209. 197. Mousseau, “A Market-Capitalist or a Democratic Peace?” 205. 198. Mousseau, “A Market-Capitalist or a Democratic Peace?” 194. 199. Mousseau, “Market Prosperity, Democratic Consolidation, and Democratic Peace.” 200. Russett and Oneal, Triangulating Peace, 123. 201. Sara McLaughlin Mitchell, “Norms and the Democratic Peace,” in Vasquez, What Do We Know about War? 2nd ed., 167–88. 202. Maoz and Abdolali, “Regime Types and International Conflict, 1816–1976”; and Mark Crescenzi and Andrew Enterline, “Ripples from the Waves? A Systemic, Time-Series Analysis of Democracy, Democratization, and Interstate War,” Journal of Peace Research 36, no. 1 (1999): 75–94. See the review in Mitchell, “Norms and the Democratic Peace.” 203. Mitchell, “Norms and the Democratic Peace,” 180. 204. David A. Lake, “Powerful Pacifists: Democratic States and War,” American Political Science Review 86, no. 1 (1992): 24–37. As always, coding of the regime type of states and the coding of winners and losers is subject to some disagreement. See also Dan Reiter and Allan C. Stam, “Democracy and Battlefield Military Effectiveness,” Journal of Conflict Resolution 42 (1998): 259–77; and Allan C. Stam, Win, Lose, or Draw: Domestic Politics and the Crucible of War (Ann Arbor: University of Michigan, 1996). There are several potential reasons why democracies tend to win their wars. The most widely accepted seems to be that since democratic governments know they will pay a high political cost for losing wars, they select the wars they participate in (and initiate) very carefully. See Dan Reiter and Allan Stam, “Democracy, War Initiation and Victory,” American Political Science Review 92, no. 2 (1998): 377–90. 205. Bruce Bueno de Mesquita, Randolph Siverson, and Gary Woller, “War and the Fate of Regimes: A Comparative Analysis,” American Political Science Review 86 (1992): 638–46. 206. Sara McLaughlin Mitchell, Scott Gates, and Havard Hegre, “Evolution in Democracy-War Dynamics,” Journal of Conflict Resolution 43, no. 6 (1999): 771–792, 777. 207. Mitchell, Gates, and Hegre, “Evolution in Democracy-War Dynamics,” 788–90. 208. Michael Mihalka, “Hostilities in the European State System, 1816–1970,” Peace Science Society Papers 26 (1976): 100–116; Erich Weede, “Overwhelming Preponderance as a Pacifying Condition among Contiguous Asian Dyads, 1950–1969,” Journal of Conflict Resolution 20 (September 1976): 395–411. In Weede’s study preponderance is defined as an overwhelming 10–1 disparity, and lesser levels are interpreted as equality. David Garnham, “Dyadic International War, 1816–1965: The Role of Power Parity and Geographic Proximity,” Western Political Quarterly 29 (June 1976): 231–42. See also the review in Siverson and Sullivan, “The Distribution of Power and the Onset of War.”

536

Notes

209. William Moul, “Balance of Power and the Escalation of Serious Disputes among European Great Powers, 1815–1939: Some Evidence,” American Journal of Political Science 32 (1988): 241–75. 210. Daniel Geller, “Power Differentials and War in Rival Dyads,” International Studies Quarterly 37 (1993): 173–93. 211. Huth, Standing Your Ground, 115–156. 212. Woosang Kim, “Alliance Transitions and Great Power War.” American Journal of Political Science 35 (1991): 833–50; and Kim, “Power Parity, Alliance, and War from 1648 to 1975,” in Parity and War: Evaluations and Extensions of the War Ledger, ed. Jacek Kugler and Douglas Lemke (Ann Arbor: University of Michigan Press, 1996), 93–105. 213. Stuart Bremer, “Dangerous Dyads: Conditions Affecting the Likelihood of Interstate War, 1816–1965,” Journal of Conflict Resolution 36 (1992): 309–41. 214. Bennett and Stam, The Behavioral Origins of War, 124–25. 215. Maoz, “Pacifism and Fightaholism in International Politics.” 216. Some support for the “equality leads to peace” hypothesis comes from a study by Wayne Ferris. His analysis of the forty-two wars from 1850 to 1965 found that thirty-four (81 percent) exhibited a power disparity of 1.45 to 1.0 or greater between the contesting dyads. He concludes that “few wars are seen to occur when two sides to the conflict approach equality in power capabilities below the ratio level of 1.45. Once that threshold is exceeded, however, the number of war events increase markedly.” Wayne Ferris, The Power Capabilities of Nation-States (Lexington, MA: Lexington Books, 1973), 76. Seyom Brown also maintains that while challenges to the prevailing balance (or imbalance) may trigger war, preponderance in a dyadic relationship may lead to war as well. He argues that if there is an intensely antagonistic relationship between two states and a large military imbalance, this tempts a dissatisfied superior state to escalate conflict to war—for instance, Japan’s attack on China, Hitler’s aggression in Europe, North Korea’s attack on South Korea, and India’s attack on Pakistan in 1971. Seyom Brown, The Causes and Prevention of War (New York: St. Martin’s, 1987), 103–5. Perhaps the key here is that the militarily superior power is deemed dissatisfied, while Organski and Kugler generally assume the opposite. 217. Huth, Standing Your Ground, 87. 218. Daniel Geller, “Status Quo Orientation, Capabilities, and Patterns of War Initiation in Dyadic Rivalries,” Conflict Management and Peace Science 18, no. 1 (2000): 73–96. 219. T. V. Paul, Asymmetric Conflicts: War Initiation by Weaker Powers (Cambridge: Cambridge University Press, 1994). 220. Vasquez, The War Puzzle, 64–65. 221. Cashman and Robinson, An Introduction to the Causes of War.

8. THE DYADIC LEVEL OF ANALYSIS, PART II 1. J. David Singer, “The Level of Analysis Problem in International Relations,” in International Politics and Foreign Policy, rev. ed., ed. James Rosenau (New York: Free Press, 1969), 23. 2. Anatol Rapoport, Fights, Games and Debates (Ann Arbor: University of Michigan Press, 1960), discussed in Russell Leng and Robert Goodsell, “Behavioral Indicators of War Proneness in Bilateral Conflicts,” in Sage International Yearbook of Foreign Policy Studies, Vol. II, ed. Patrick J. McGowan (Beverly Hills, CA: Sage, 1974), 193–94. 3. Leng and Goodsell, “Behavior Indicators of War Proneness,” 194. 4. Leng and Goodsell, “Behavior Indicators of War Proneness,” 207–17. 5. Joshua S. Goldstein and John R. Freeman, Three-Way Street: Strategic Reciprocity and World Politics (Chicago: University of Chicago Press, 1990). 6. The terms “liberal” and “conservative” are used rather loosely here. They are intended to serve as an umbrella term that covers a wide range of labels. The liberal label as used here encompasses that who might be called “doves,” “accommodationists,” or Wilsonians. The Conservative label encompasses those who might be called “hawks,” “hardliners,” and “realists.” 7. Russell Leng, “Reagan and the Russians: Crisis Bargaining Beliefs and the Historical Record,” American Political Science Review 78 (June 1984): 338–55.

Notes

537

8. William A. Gamson and Andre Modigliani, Untangling the Cold War: A Strategy for Testing Rival Theories (Boston: Little, Brown, 1971), reported in Michael P. Sullivan, International Relations: Theories and Evidence (Englewood Cliffs, NJ: Prentice Hall, 1976), 286–87. 9. Jan F. Triska and David D. Finley, “Soviet-American Relations: A Multiple Symmetry Model,” in International Political Analysis: Readings, ed. David V. Edwards (New York: Holt, Rinehart & Winston, 1969). 10. Ole Holsti, Crisis, Escalation and War (Montreal: McGill-Queens University Press, 1972). 11. Richard K. Ashley, The Political Economy of War and Peace (New York: Nichols, 1980); Michael Don Ward, “Cooperation and Conflict in Foreign Policy Behavior,” International Studies Quarterly 26 (March 1982): 87–126; William J. Dixon, “Measuring Interstate Affect,” American Journal of Political Science 27 (November 1983): 828–51. 12. See Frank Mogdis, “The Verbal Dimension in Sino-Soviet Relations: A Time Series Analysis,” paper presented at American Political Science Association convention, September 1970, cited in Sullivan, International Relations, 289–90. 13. Goldstein and Freeman, Three-Way Street, especially chapter 3. See also Joshua Goldstein, “Reciprocity in Superpower Relations: An Empirical Analysis,” International Studies Quarterly 35, no. 2 (June 1991): 195–209, which deals solely with U.S.-Soviet relations. 14. Ole Holsti, Robert North, and Richard Brody, “Perception and Action in the 1914 Crisis,” in Quantitative International Politics, ed. J. David Singer (New York: Free Press, 1968); and North, Brody, and Holsti, “Some Empirical Data on the Conflict Spiral,” Peace Research Society (International) Papers 1 (1964): 1–14. 15. Leng and Goodsell, “Behavior Indicators of War Proneness.” The five conflicts are the Schleswig-Holstein conflict between Germany and Denmark (1864–66), the Moroccan Crisis between France and Germany (1904–6), the Austro-Serbian conflict on the eve of WWI (1914), the Suez Crisis involving Britain and Egypt (1956–57), and the Cuban Missile Crisis between the U.S. and USSR (1962). 16. Michael Don Ward, “Cooperation and Conflict in Foreign Policy Behavior: Reaction and Memory,” International Studies Quarterly 26 (March 1982): 87–126. Israel’s response to UAR behavior was also escalatory, returning both conflict and cooperation at a ratio of 1.7 to 1.0. 17. See, for instance, Jeffrey S. Milstein, “American and Soviet Influence, Balance of Power, and Arab-Israeli Violence,” in Peace, War and Numbers, ed. Bruce Russett (Beverly Hills, CA: Sage, 1972), 139–62; R. Burrowes and J. Garriga-Pico, “The Road to the Six Day War; Relational Analysis of Conflict and Cooperation,” Peace Science Society (International) Papers 22 (1974): 47–74; J. M. McCormick, “Evaluating Models of Crisis Behavior: Some Evidence from the Middle East,” International Studies Quarterly 19 (1975): 17–45; J. Wilkenfeld, V. Lussier, and D. Tahtinen, “Conflict Interactions in the Middle East, 1949–1967,” Journal of Conflict Resolution 16 (1972): 135–54; J. Wilkenfeld, “A Time Series Perspective on Conflict in Foreign Policy Behavior in the Middle East,” in Sage International Yearbook of Foreign Policy Studies III, ed. P. McGowan (Beverly Hills, CA: 1975), 177–212. 18. Russell Leng, “Escalation: Competing Perspectives and Empirical Evidence,” International Studies Review 6 (2004): 51–64. 19. Leng, “Escalation: Competing Perspectives and Empirical Evidence,” 56–58; Leng is drawing on Michael Brecher and Jonathan Wilkenfeld, A Study of Crises (Ann Arbor: University of Michigan Press, 1997); and Leng, Interstate Crisis Behavior, 1816–1980: Realism vs. Reciprocity (Cambridge: Cambridge University Press. 1993), 121–33. 20. Russell J. Leng and Hugh B. Wheeler, “Influence Strategies, Success and War,” Journal of Conflict Resolution 23 (December 1979): 655–84. 21. Russell J. Leng, “Influence Strategies and Interstate Conflict,” in Correlates of War II: Testing Some Realpolitik Models, ed. J. David Singer (New York: Free Press, 1980), 124–57, especially 154. 22. Russell J. Leng and Charles S. Gochman, “Dangerous Disputes: A Study of Conflict Behavior and War,” American Journal of Political Science 26 (November 1982): 664–87. 23. Leng, ““Escalation: Competing Perspectives and Empirical Evidence,” 56–57. Leng draws on Gochman and Leng, “Dangerous Disputes”; Daniel Jones, Stuart Bremer, and J. David Singer, “Militarized Interstate Disputes, 1816–1992”; and Zeev Maoz, “Resolve, Capabilities, and the Outcomes of Interstate Disputes, 1816–1976,” Journal of Conflict Resolution 27 (1983): 195–229. 24. Leng, “Escalation: Competing Perspectives and Empirical Evidence,” 57. Leng is drawing on Leng, Interstate Crisis Behavior. 25. Russell J. Leng, “When Will They Ever Learn? Coercive Bargaining in Recurrent Crises,” Journal of Conflict Resolution 27 (September 1983): 379–419. See also Russell J. Leng, “Reagan and the Russians: Crisis Bargaining

538

Notes

Beliefs and the Historical Record,” American Political Science Review 78 (September 1984): 338–55; and Russell J. Leng, “Crisis Learning Games,” American Political Science Review 82 (March 1988): 179–94. 26. Leng, “Escalation: Competing Perspectives and Empirical Evidence,” 62. 27. Paul Diehl, “Arms Races to War: Testing Some Empirical Linkages,” Sociological Quarterly 26 (1985): 331–49. 28. Brandon Valeriano, “The Steps to Rivalry: Power Politics and Rivalry Formation,” Ph.D. dissertation, Vanderbilt University, 2003, cited in John Vasquez, The War Puzzle Revisited (Cambridge: Cambridge University Press, 2009), 401. 29. Brecher and Wilkenfeld, A Study of Crises, 826–34, cited in Leng, “Escalation: Competing Perspectives and Empirical Evidence,” 61. 30. Leng, “Escalation: Competing Perspectives and Empirical Evidence,” 55, 58; Leng is citing Russell Leng, “Will They Ever Learn?,” 57–58; Michael Brecher and Jonathan Wilkenfeld, Crisis, Conflict and Instability (Oxford: Pergamon Press, 1989), 216; and Leng, Interstate Crisis Behavior, 1816–1980, 121. 31. Richard Ned Lebow, Between Peace and War: The Nature of International Crisis (Baltimore: Johns Hopkins University Press, 1981); and Russell Leng, Bargaining and Learning in Recurring Crises: The Soviet-American, Egyptian-Israeli, and Indo-Pakistani Rivalries (Ann Arbor: University of Michigan Press, 2000). 32. Leng, “Escalation: Competing Perspectives and Empirical Evidence,” 56. 33. Daniel Jones, Stuart Bremer, and J. David Singer, “Militarized Interstate Disputes, 1816–1992: Rationale, Coding Rules, and Empirical Patterns,” Conflict Management and Peace Science 15 (1996): 163–213. 34. Leng, “Escalation: Competing Perspectives and Empirical Evidence,” 58–59. 35. Leng, “Escalation: Competing Perspectives and Empirical Evidence,” 62. 36. Robert Jervis, “Cooperation Under the Security Dilemma,” World Politics 30, no. 2 (January 1978): 167–214; and Jervis, “Perception and Misperception: The Spiral of International Insecurity,” in Theory and Practice of International Relations, 6th ed., ed. William Olson, David McClellan, and Fred Sondermann (Englewood Cliffs, NJ: Prentice-Hall, 1983), 201. 37. Jervis, “Perception and Misperception,” 200. 38. Jervis, “Perception and Misperception,” 201. 39. Charles Glaser, “The Security Dilemma Revisited,” World Politics 50 (October, 1997): 171–201. 40. Jean-Jacques Rousseau, A Lasting Peace through the Federation of Europe, trans. C. E. Vaughan (London: Constable, 1917), 78–79, quoted in Jervis, “Perception and Misperception,” 200. 41. Glaser, “The Security Dilemma Revisited,” 171. 42. Dan Reiter, “Exploding the Powder Keg Myth: Preemptive Wars Almost Never Happen,” International Security 20, no. 2 (1995): 5–34. 43. Reiter, “Exploding the Powder Keg Myth,” 8. 44. Stephen Van Evera, Causes of War (Ithaca, NY: Cornell University Press, 1999), especially notes on 40 and 41. Van Evera places the Chinese intervention in Korea in the preventive war category rather than preemption. 45. Greg Cashman and Leonard C. Robinson, An Introduction to the Causes of War: Patterns of Interstate Conflict from WWI to Iraq (Lanham, MD: Rowman & Littlefield, 2007), 367–69. The six wars were WWI, WWI in the Pacific, the Six Day War, the 1971 India-Pakistan War, Iran-Iraq, and Ethiopia-Eritrea. (The last war is reported on the book’s website, rather than the book itself.) The seventh war covered was a war of inequality—the United States versus Iraq in 2003. In this case there was a decided attempt by Iraq to de-escalate conflictual interactions prior to the outbreak of war. 46. Zeev Maoz, Paths to Conflict: International Dispute Initiation, 1816–1976 (Boulder, CO: Westview, 1982). 47. Samuel P. Huntington, “Arms Races: Prerequisites and Results,” in Public Policy, vol. 8, ed. C. J. Friedrich and S. E. Harris (Cambridge, MA: Graduate School of Public Administration, Harvard University, 1958), 41. 48. Lewis F. Richardson, Arms and Insecurity (Pittsburgh: Boxwood Press, 1960). For those interested, the precise equation is as follows: dx / dt = ky – ax + g, in which dx / dt = the rate at which one increases arms levels; x = one’s own strength; y = the opponent’s strength; k = one’s readiness to accumulate arms; a = fatigue and cost; and g = the general level of grievance. 49. W. Ladd Hollist, “An Analysis of Arms Processes in the United States and Soviet Union,” International Studies Quarterly 21 (September 1977): 503–28; Benjamin Lambeth, “The Sources of Soviet Military Doctrine,” in Comparative Defense Policy, ed. B. Horton et al. (Baltimore: Johns Hopkins University Press, 1974); A. F. K. Organski and Jacek Kugler, The War Ledger (Chicago: University of Chicago Press, 1980); Stephen J. Majeski and David L. Jones, “Arms Race Modelling: Causality Analysis and Model Specification,” Journal of Conflict Resolution 25 (1981): 259–88; Charles Ostrom Jr., “Evaluating Alternative Foreign Policy Decision Making Models,”

Notes

539

Journal of Conflict Resolution 21 (June 1977): 235–66; and Charles Ostrom Jr. and R. F. Marra, “U.S. Defense Spending and the Soviet Estimate,” American Political Science Review 80, no. 3 (September 1986): 819–42. 50. Mike Horn, “Arms Races and the Likelihood of War,” paper presented to International Studies Association convention, Atlanta, 1984. 51. Thomas R. Cusack and Michael Don Ward, “Military Spending in the United States, Soviet Union and the Peoples’ Republic of China,” Journal of Conflict Resolution 25 (September 1981): 429–67. Studies of arms races in the Middle East have produced mixed results. For some periods and for some states there has been a Richardsontype reaction process, but this has not been found to hold true for all states and all time periods. For instance, see Hans Rattinger, “From War to War: Arms Races in the Middle East,” International Studies Quarterly 20 (December 1976): 501–31. 52. Hans Rattinger, “Armaments, Détente, and Bureaucracy: The Case of the Arms Race in Europe,” Journal of Conflict Resolution 19 (December 1975): 571–95. See also W. Ladd Holllist, “Alternative Explanations of Competitive Arms Processes: Tests on Four Pairs of Nations,” American Journal of Political Science 21 (May 1977): 315–40. 53. Choucri and North, Nations in Conflict. 54. Arthur J. Marder, From the Dreadnought to the Scapa Flow (New York: Oxford University Press, 1961), 121 quoted in Choucri and North, 206. 55. See Choucri and North, Nations in Conflict, 207. 56. Lloyd Jensen, Explaining Foreign Policy (Englewood Cliffs, NJ: Prentice-Hall, 1982), 239–40. 57. Goldstein and Freeman, Three-Way Street (26–27), note that, generally, stimulus-response studies that use data aggregated on a yearly basis fail to find reciprocity, while those using data aggregated on a sub-annual basis generally do find reciprocity. They suggest that this is because using annual aggregations lumps together reciprocal interactions and exchanges, thus masking them and washing out lagged correlations and making them appear simultaneous. 58. Michael Don Ward, “Differential Paths to Parity: A Study of Contemporary Arms Races,” American Political Science Review 78 (1984): 297–317. Joshua Goldstein argues that event and spending data are both prone to biases that cause the significance of reciprocity to be understated in statistical analyses. The aggregation of such data over long periods of time tends to have the same effect. Studies based on annually aggregated military expenditure data rarely find evidence of reciprocal arms racing. J. S. Goldstein, “Reciprocity in Superpower Relations,” 198. 59. Huntington, “Arms Races: Prerequisites and Results.” 60. Huntington, “Arms Races: Prerequisites and Results,” 61. 61. Huntington, “Arms Races: Prerequisites and Results,” 65. Richardson’s model points in a somewhat different direction. In his formulation, if armaments increase indefinitely with little restraint, war will eventually break out. These “runaway” arms races, in which no equilibrium point is reached, are the most dangerous. 62. Patrick Morgan, Theories and Approaches to International Politics, 3rd ed. (New Brunswick, NJ: Transaction, 1981), 268. 63. Huntington, “Arms Races: Prerequisites and Results,” 75–76. 64. Huntington, “Arms Races: Prerequisites and Results,” 71–72. 65. See Marc Trachtenberg, History and Strategy (Princeton, NJ: Princeton University Press, 1991), chapter 3; William Burr and Jeffrey Richelson, “Whether to ‘Strangle the Baby in the Cradle’: The United States and the Chinese Nuclear Program, 1960–64,” International Security 25 (Winter 2000/01): 54–99; and T. V. Paul, The Tradition of Non-Use of Nuclear Weapons (Stanford, CA: Stanford University Press, 2009), 64–72. 66. Jonathan Adelman and Deborah Palmieri, The Dynamics of Soviet Foreign Policy (New York: Harper & Row, 1989), 276–80. 67. Michael D. Wallace, “Arms Races and Escalation: Some New Evidence,” Journal of Conflict Resolution 23 (March 1979): 7. 68. Wallace, “Arms Races and Escalation,” 14. 69. On these points, see Erich Weede, “Arms Races and Escalation: Some Persisting Doubts,” Journal of Conflict Resolution 24 (June 1980): 285–87; and Randolph Siverson and Paul Diehl, “Arms Races, the Conflict Spiral, and the Onset of War,” in Handbook of War Studies, ed. Manus Midlarsky (Boston: Unwin Hyman, 1989), 203. 70. M. Altfeld, “Arms Races?—and Escalation? A Comment on Wallace,” International Studies Quarterly 27, no. 2 (June 1983): 225–31. 71. Paul F. Diehl, “Arms Races and Escalation: A Closer Look,” Journal of Peace Research 20, no. 3 (1983): 206–7. This is a problem that Wallace himself addresses and partly resolves in a later study: Michael D. Wallace, “Armaments and Escalation: Two Competing Hypotheses,” International Studies Quarterly 26 (March 1982):

540

Notes

37–56. Using a stricter definition of arms races and a combination of dyads in which two or more allies entered simultaneously into war with a common enemy, his results are still quite supportive of the connection between war and arms races. 72. Erich Weede, “Arms Races and Escalation: Some Persisting Doubts.” And see Wallace’s reply, M. D. Wallace, “Some Persistent Findings: A Reply to Professor Weede,” Journal of Conflict Resolution 24 (June 1980): 289–92. 73. See Diehl, “Arms Races and Escalation: A Closer Look,” 207–11. Also, research by Lambelet, unconnected to the Wallace study, finds arms races and wars to be independent of each other. John Lambelet, “Do Arms Races Lead to War?” Journal of Peace Research 12, no. 2 (1975): 123–28. 74. Susan Sample, “Arms Races and Dispute Escalation: Resolving the Debate,” Journal of Peace Research 34 (1997): 7–22. 75. James D. Morrow, “A Twist of Truth: A Reexamination of the Effects of Arms Races on the Occurrence of War,” Journal of Conflict Resolution 33 (September 1989): 518–19. 76. Susan G. Sample, “Military Buildups: Arming and War,” in What Do We Know About War?, ed. John Vasquez (Lanham, MD: Rowman and Littlefield, 2000), 165–95. 77. Douglas Gibler, Tony Rider, and March Hutchison, “Taking Arms against a Sea of Troubles: Conventional Arms Races during Periods of Rivalry,” Journal of Peace Research 42 (2005): 131–47. 78. John Vasquez, “What Do We Know about War?” in What Do We Know about War? ed. John Vasquez (Lanham, MD: Rowman & Littlefield, 2012), 301–30; quote is from 316. 79. Sample, “Military Buildups: Arming and War,” 176–77, tables 8.7 and 8.8. The probability goes to .69 if, in addition to arms races, the states have a high defense burden, territorial disputes, parity, power transitions, rapid approaches, and no nuclear weapons. 80. Susan Sample, “The Outcomes of Military Buildups: Minor States vs. Major Powers,” Journal of Peace Research 39 (2002): 669–91. 81. Perhaps one reason that not all wars are preceded by arms races is that not all wars are between relative equals. Although arms races tend to escalate disputes between relative equals, they rarely play a role in disputes between nonequals. On this point see John Vasquez, “The Steps to War,” World Politics XL, no. 1 (October 1987): 136, n85. 82. Morrow, “A Twist of Truth,” 502. 83. Morrow, “A Twist of Truth,” 502. 84. L. F. Richardson, Arms and Insecurity. 85. Theresa Clair Smith, “Arms Race Instability and War,” Journal of Conflict Resolution 24 (June 1980): 253–84; and T. C. Smith, “Curvature Change and War Risk in Arming Patterns,” International Interactions 14 (1988): 201–28. Smith includes only those races that last for four years or more. 86. Morrow, “A Twist of Truth.” 87. Another necessary condition seems to be that the dispute in question arises in an area contiguous to one of the rivals. Paul Diehl, “Arms Races to War: An Analysis of Some Underlying Effects,” Sociological Quarterly 26 (1985): 331–49. 88. Wallace finds no support either for the hypothesis that relative rates of military growth are related to the outbreak of war. M. D. Wallace, “Armaments and Escalation,” International Studies Quarterly 26 (1982): 37–56. See also M. D. Wallace, “Armaments and Escalation: A Reply to Altfeld,” International Studies Quarterly 27 (1983): 233–35. 89. Susan Sample, “Arms Races: A Cause or a Symptom?” in What Do We Know about War? 2nd ed., ed. John Vasquez (Lanham, MD: Rowman & Littlefield, 2012), 111–38. 90. Paul Diehl, “Contiguity and Escalation in Major Power Rivalries, 1816–1980,” Journal of Politics 47, no. 4 (1985): 1203–11. 91. Sample, “Arms Races: A Cause or a Symptom?” 125, table 6.1. 92. Sample, “Arms Races: A Cause or a Symptom?” 131, table 6.3. 93. Other studies that support the independent effect of mutual arms buildups would include Gibler, Rider, and Hutchison, “Taking Arms against a Sea of Troubles”; Paul Senese and John Vasquez, The Steps to War: An Empirical Study (Princeton, NJ: Princeton University Press, 2008); and Michael Colaresi, Karen Rasler, and William Thompson, Strategic Rivalries in World Politics: Position, Space and Conflict Escalation (Cambridge: Cambridge University Press, 2007). See Sample’s discussion of these studies, “Arms Races: A Cause or Symptom?” 119–20. 94. Sample, “Military Buildups: Arming and War,” 185.

Notes

541

95. Susan Sample, “The Outcomes of Military Buildups: Minor States vs. Major Powers,” Journal of Peace Research 39 (November 2002): 669–92. 96. Sample, “Arms Races: A Cause or a Symptom?” and Senese and Vasquez, The Steps to War. 97. Research by Diehl and Kingston suggests that large arms increases don’t make states or a set of rivals more likely to be involved in disputes. That is, they find no evidence for a causal chain that starts with arms races and proceeds to disputes and then to war. Instead, it is more likely that disputes arise because of other reasons and then arms races develop due to the tensions involved in the preexisting disputes. Paul Diehl and J. Kingston, “Messenger or Message? Military Buildups and the Initiation of Conflict,” Journal of Politics 49 (1987): 789–99. 98. The steps-to-war theory was first put forward in John A. Vasquez, “The Steps to War: Toward a Scientific Explanation of Correlates of War Findings,” World Politics 40 (October 1987): 108–45. A longer explanation followed in John A. Vasquez, The War Puzzle (Cambridge: Cambridge University Press, 1993). Other major works associated with the theory are John A. Vasquez, “Reexamining the Steps to War: New Evidence and Theoretical Insights, in Handbook of War Studies II, ed. Manus Midlarsky (Ann Arbor: University of Michigan Press, 2000), 371–406; John A. Vasquez, “The Probability of War, 1816–1992,” International Studies Quarterly 48 (March 2004): 1–28; and John A. Vasquez and Marie T. Henehan, “Territorial Disputes and the Probability of War, 1816–1992,” Journal of Peace Research 38 (March 2001): 123–38. At this date, the final summation of the theory and the statistical analysis that supports it can be found in Paul D. Senese and John A. Vasquez, The Steps to War: An Empirical Study (Princeton, NJ: Princeton University Press, 2008). 99. Senese and Vasquez, The Steps to War, 34. 100. Senese and Vasquez, The Steps to War, 14. 101. The steps-to-war model, like the spiral model, assumes that at least one state is defensively motivated. 102. Senese and Vasquez, The Steps to War, 22–23. 103. Vasquez, The War Puzzle, 199. 104. Vasquez, The War Puzzle, chapter 7. 105. Senese and Vasquez, The Steps to War, 33. 106. Senese and Vasquez, The Steps to War, 97. The section is essentially a reprinting of Senese and Vasquez, “A Unified Explanation of Territorial Conflict,” International Studies Quarterly 47 (June 2003): 275–98. 107. Senese and Vasquez, The Steps to War, 121. 108. Senese and Vasquez, The Steps to War, 123. 109. Their unit of analysis is the dyad itself. As they readily admit, one of the problems with this design is that one cannot tell if the various steps to war actually precede a war (if the dyad has a war) or occur after it. Nevertheless, they argue that what we can infer from these results is that the absence of (or the presence of only a few) steps to war correctly predicts the absence of war. 110. Senese and Vasquez, The Steps to War, 145–52, and table 5.2. 111. Senese and Vasquez, The Steps to War, 191–99, and table 6.3. 112. Brandon Valeriano and John Vasquez, “Identifying and Classifying Complex Interstate Wars,” International Studies Quarterly 54 (2010): 561–82. 113. Michael Colaresi and William Thompson, “Alliances, Arms Buildups and Recurrent Conflict: Testing a Steps-to-War Model,” Journal of Politics 67 (May 2005): 345–64; Colaresi et al., Strategic Rivalries in World Politics. 114. Valeriano and Vasquez, “Identifying and Classifying Complex Interstate Wars.” 115. John Vasquez, “The India-Pakistan Conflict in Light of General Theories of War, Rivalry, and Deterrence,” in The India-Pakistan Conflict: An Enduring Rivalry, ed. T. V. Paul (Cambridge: Cambridge University Press, 2005), 54–79. 116. Karen Rasler and William Thompson, “Contested Territory, Strategic Rivalries, and Conflict Escalation,” International Studies Quarterly 50 (March 2006): 145–67; Colaresi et al., Strategic Rivalries in World Politics, chapter 9. See the discussion in Vasquez, The War Puzzle Revisited, 396–98. 117. Jack Levy, “The ‘Paths-to-War’ Concept,” in What Do We Know about War? ed. John Vasquez (Lanham, MD: Rowman & Littlefield, 2012), 281–90; quote is from 281. 118. Charles E. Osgood, “Graduated Unilateral Initiatives for Peace,” in Conflict Resolution: Contributions from the Behavioral Sciences, ed. Clagett G. Smith (Notre Dame, IN: University of Notre Dame, 1971), 515–25. The following section draws on this work. See also Osgood, An Alternative to War of Surrender (Urbana, IL: University of Illinois Press, 1962). 119. Osgood, “Graduated Unilateral Initiatives for Peace.”

542

Notes

120. This applies as well to TIT-FOR-TAT, a strategy to be discussed later. On this point see Kenneth Oye, “Explaining Cooperation Under Anarchy: Hypotheses and Strategies,” World Politics 38 (October 1985): 16. 121. GRIT appears to be more effective than TIT-FOR-TAT in inducing early cooperation. See Svenn Linskold and Michael Collins, “Inducing Cooperation by Groups and Individuals,” Journal of Conflict Resolution 22 (December 1978): 679–90; S. Lindskold, P. S. Walters, and H. Koutsourais, “Cooperation, Competitors, and Responses to GRIT,” Journal of Personality and Social Psychology 8 (1968): 122–33. For an excellent overall review of the experimental literature on GRIT, see Svenn Lindskold, “Trust Development, the GRIT Proposal, and the Effects of Conciliatory Acts on Conflict and Cooperation,” Psychological Bulletin 85, no. 4 (1978): 772–93. See also S. Lindskold, “Conciliation with Simultaneous or Sequential Interactions,” Journal of Conflict Resolution 23 (December 1979): 704–14. 122. Goldstein and Freeman, Three-Way Street, chapter 4. 123. Goldstein and Freeman, Three-Way Street, chapter 5, especially 134–36. 124. Goldstein and Freeman, Three-Way Street, 153. 125. Goldstein and Freemen, Three-Way Street, 152.

9. THE DYADIC LEVEL OF ANALYSIS, PART III 1. Thomas Schelling, The Strategy of Conflict (New York: Oxford University Press/Galaxy Books, 1963), 9–10. 2. Adapted from Martin Shubik, “Game Theory and the Study of Social Behavior: An Introductory Exposition,” in Game Theory and Related Approaches to Social Behavior, ed. Martin Shubik (New York: Wiley, 1964), 15–17. See also the discussion in James Dougherty and Robert Pfaltzgraff Jr., Contending Theories of International Relations: A Comprehensive Survey, 2nd ed. (New York: Harper & Row, 1981), 515–16. 3. See Dougherty and Pfaltzgraff, Contending Theories of International Relations, 514–17. 4. The “rationality of irrationality” idea is generally traced to Schelling, The Strategy of Conflict, and also to Herman Kahn, Thinking about the Unthinkable (New York: Avon Books, 1962). Richard Nixon, especially in his dealings with the Chinese, is reputed to have been a practitioner of the madman strategy of trying to persuade foreign leaders that he was capable of acting unpredictably and a little irrationally. 5. Melvin Small and J. David Singer, Resort to Arms: International and Civil Wars, 1816–1980 (Beverly Hills, CA: Sage, 1982), see 196–97; and Zeev Maoz, “Resolve, Capabilities, and the Outcomes of Interstate Disputes, 1816–1976,” Journal of Conflict Resolution 27 (1983): 199–200. 6. Glenn Snyder and Paul Diesing, Conflict among Nations: Bargaining, Decision Making, and System Structure in International Crises (Princeton, NJ: Princeton University Press, 1977), 118–22. 7. Kenneth Oye, “Explaining Cooperation Under Anarchy: Hypotheses and Strategies,” World Politics 38 (October 1985): 14. 8. On the importance of repeated prisoners’ dilemma interactions, see Robert Axelrod, The Evolution of Cooperation (New York: Basic Books, 1984), 11–12. 9. Axelrod, The Evolution of Cooperation, 77. 10. Axelrod, The Evolution of Cooperation, 84. 11. Murnighan and Roth find that as the probability of continued play is increased, for particular games, cooperation is also increased. J. Keith Murnighan and Alvin E. Roth, “Expected Continued Play in Prisoner’s Dilemma Games,” Journal of Conflict Resolution 27 (June 1983): 279–300. 12. Axelrod’s matrix awards 1 point for each player for mutual defection and 3 points for each player for mutual cooperation; the defect/cooperate cell yields either 0 points or 5 points. Axelrod, The Evolution of Cooperation, 8. 13. See Axelrod, The Evolution of Cooperation; and Axelrod, “Effective Choice in the Prisoner’s Dilemma,” Journal of Conflict Resolution 24 (March 1980): 3–25; and Axelrod, “More Effective Choice in the Prisoner’s Dilemma,” Journal of Conflict Resolution 24 (September 1980): 379–403. 14. Axelrod, “More Effective Choice in the Prisoner’s Dilemma,” 394. 15. Axelrod, The Evolution of Cooperation, 54. 16. S. Oskamp, “Effects of Programmed Strategies on Cooperation in Prisoner’s Dilemma and Other Mixed Motive Games,” Journal of Conflict Resolution 15 (1971): 225–59. 17. Russell J. Leng and Hugh B. Wheeler, “Influence Strategies, Success and War,” Journal of Conflict Resolution 23 (December 1979): 655–84; Russell J. Leng, “Reagan and the Russians: Crisis Bargaining, Beliefs and the

Notes

543

Historical Record,” American Political Science Review 78 (June 1984): 338–55; also Paul Huth and Bruce Russett, “Deterrence Failure and Crisis Escalation,” International Studies Quarterly 32 (March 1988): 29–45. 18. See, for instance, Raymond Dacey and Norman Pendergraft, “The Optimality of TIT-FOR-TAT,” International Interactions 15, no. 1 (1988): 52. 19. Oye, “Explaining Cooperation Under Anarchy,” 15. Martin Patchen believes TFT should be even more effective in chicken situations than in prisoners’ dilemmas because the rival suffers maximum damages when his noncooperative move is reciprocated. Martin Patchen, “Strategies for Eliciting Cooperation from an Adversary: Laboratory and International Findings,” Journal of Conflict Resolution 31 (March 1987): 171. 20. Patchen, “Strategies for Eliciting Cooperation from an Adversary,” 176–77. Leng also demonstrates the absolutely critical importance of initiating the interaction with a cooperative move. Without such initial strategies, it is almost inevitable that the two sides will become locked into an escalation that could lead to war. Leng attributes this escalation to “realpolitik” assumptions about bargaining strategies. Russell J. Leng, “Crisis Learning Games,” American Political Science Review 82 (March 1988): 179–94. 21. Theodore To, “More Realism in Prisoner’s Dilemma,” Journal of Conflict Resolution 32 (June 1988): 402–8. 22. C. L. Gruder and R. J. Dulak, “Elicitation of Cooperation by Retaliatory and Nonretaliatory Strategies in a Mixed-Motive Game,” Journal of Conflict Resolution 17 (1973): 162–74. 23. Axelrod, The Evolution of Cooperation, 138. 24. Roy Behr, “Nice Guys Finish Last—Sometimes,” Journal of Conflict Resolution 25 (June 1981): 289–300. TIT-FOR-TAT was also the only strategy to lose to RANDOM. The scoring in Behr’s tournament was: win = 2, tie = 1, and loss = 0. TIT-FOR-TAT did equally poorly when point differentials were used as the standard of success instead of win-loss records. 25. Dacey and Pendergraft, “The Optimality of TIT-FOR-TAT,” 45–52; Oye, “Explaining Cooperation Under Anarchy,” 1–24; Patchen, “Strategies for Eliciting Cooperation from an Adversary,” 176. (The success of TIT-FORTAT is also affected by the number of players.) 26. Van Evera, “Why Cooperation Failed in 1914,” World Politics 38 (October 1985): 80–117. 27. Van Evera, “Why Cooperation Failed in 1914,” 99. 28. Leng and Wheeler, “Influence Strategies, Success and War.” 29. Svenn Lindskold and Michael Collins, “Inducing Cooperation by Groups and Individuals,” Journal of Conflict Resolution 22 (December 1978): 679–90. 30. Joshua S. Goldstein and John R. Freeman, Three-Way Street: Strategic Reciprocation in World Politics (Chicago: University of Chicago Press, 1990), chapter 5. 31. For instance, Jack Levy’s attempt to reconstruct the preference orderings of state leaders in Austria, Germany, France, Britain, Russia, and Serbia in the 1914 crisis helps to explain why it was virtually impossible to construct a cooperative solution in prewar Europe. Jack S. Levy, “Preferences, Constraints and Choices in July 1914,” International Security 15 (Winter 1990–91): 151–86. 32. James Fearon, “Rationalist Explanations for War,” International Organization 49, no. 3 (Summer, 1995): 379–414. Fundamental works on bargaining theory would also include Robert Powell, In the Shadow of Power: States and Strategies in International Politics (Princeton, NJ: Princeton University Press, 1999); Harrison Wagner, “Bargaining and War,” American Journal of Political Science 44, no. 3 (July 2000): 469–84; and Dan Reiter, “Exploring the Bargaining Model of War,” Perspectives on Politics 1, no. 1 (March 2003): 27–43. 33. Thomas Schelling, The Strategy of Conflict (Cambridge: Harvard University Press, 1960), 5. 34. Geoffrey Blainey, The Causes of War (New York: Free Press,1973). 35. Fearon, “Rationalist Explanations for War,” 380. 36. Fearon, “Rationalist Explanations for War,” 392. 37. Fearon, “Rationalist Explanations for War,” 398–400. Two excellent sources on the origins of the RussoJapanese War are Ian Nish, The Origins of the Russo-Japanese War (London: Longman, 1985), and William Langer, “The Origins of the Russo-Japanese War,” in Explorations in Crisis, ed. Carl Schorske and Elizabeth Schorske (Cambridge: Harvard University Press, 1969), 3–45. 38. In a situation in which offensive advantages tempt states toward preemptive war, large first-strike incentives might make any potential prewar agreement unenforceable and thus all bargains would lack credibility. Leaders would probably think as follows: “The first-strike advantage is so great that regardless of how we resolve any diplomatic issues between us, one side will always want to attack the other in an effort to gain the (huge) advantage of going first.” Ultimately, Fearon thinks that the effect of an offensive advantage is that it magnifies other causes of war by narrowing the bargaining range. For instance, if leaders in WWI had thought defenses were superior to

544

Notes

offenses, the set of potentially enforceable agreements available would have been much larger. Fearon, “Rationalist Explanations for War,” 403. 39. Fearon, “Rationalist Explanations for War,” 406. 40. Stacie Goddard argues that indivisibility is a matter of social construction. Territorial issues may be constructed in such a way as to make them indivisible during the bargaining process, due to attempts by statesmen to provide legitimacy for their positions and deny legitimacy to the positions of their opponents. Stacie Goddard, “Uncommon Ground: Indivisible Territory and the Politics of Legitimacy,” International Organization 60 (Winter 2006): 35–68. See also Monica Toft, “Indivisible Territory, Geographic Concentration, and Ethnic War,” Security Studies 12, no. 2 (2002): 82–119. 41. Robert Powell, “War as a Commitment Problem,” International Organization 60 (Winter 2006): 169–203. 42. David Lake, “Two Cheers for Bargaining Theory: Assessing Rationalist Explanations of the Iraq War,” International Security 35, no. 3 (Winter 2010/2011): 7–52. 43. Lake rejects the idea that the Bush administration simply wanted war and that the decision can be explained by U.S. preferences alone. He argues that while the administration certainly expected large benefits from a war, these could not have been enough to offset the actual costs of fighting, though they certainly reduced what the administration would find as an acceptable bargain. 44. Lake, “Two Cheers for Bargaining Theory,” 17. 45. Lake, “Two Cheers for Bargaining Theory,” 25. 46. Kevin Woods, James Lacey, and Williamson Murray, “Saddam’s Delusions: The View from the Inside,” Foreign Affairs 85, no. 3 (May/June 2006): 2–26; and Amatzia Baram, “Deterrence Lessons from Iraq: Rationality Is Not the Only Key to Containment,” Foreign Affairs, 91, no. 4 (July/August 2012): 76–90. 47. Lake, “Two Cheers for Bargaining Theory,” 31, quoting Michael Gordon and Bernard Trainor, Cobra II: The Inside Story of the Invasion and Occupation of Iraq (New York: Pantheon, 2006), 121. 48. Lake, “Two Cheers for Bargaining Theory,” 48. 49. For instance, Darren Filson and Suzanne Werner, “A Bargaining Model of War and Peace: Anticipating the Onset, Duration, and Outcome of War,” American Journal of Political Science 46, no. 4 (October 2002): 819–38; Harrison Wagner, “Bargaining and War,” American Journal of Political Science 44, no. 3 (2000), 469–84; Robert Powell, “Bargaining Theory and International Conflict,” Annual Review of Political Science 5 (2002): 1–30. 50. Dan Reiter, “Exploring the Bargaining Model of War,” Perspectives on Politics 1, no. 1 (March 2003): 27–43; see 31. 51. Dan Reiter, “Exploring the Bargaining Model of War,” 29. 52. Dan Reiter, “Exploring the Bargaining Model of War,” 34–36. 53. Robert D. Putnam, “Diplomacy and Domestic Politics: The Logic of Two-Level Games,” International Organization 42, no. 3 (Summer 1988), 427–60. 54. Putnam, “Diplomacy and Domestic Politics: The Logic of Two-Level Games,” 438. 55. This discussion draws on Greg Cashman and Leonard C. Robinson, An Introduction to the Causes of War: Patterns of Interstate Conflict from World War I to Iraq (Lanham, MD: Rowman & Littlefield, 2007), 89–154, especially 129–48. The essential sources for the Pacific War, which this section draws on, are Herbert Bix, Hirohito and the Making of Modern Japan (New York: Harper Collins, 2000); Robert Butow, Tojo and the Coming of the War (Princeton, NJ: Princeton University Press, 1961); Herbert Feis, The Road to Pearl Harbor (Princeton, NJ: Princeton University Press, 1950); Akira Iriye, The Origins of the Second World War in Asia and the Pacific (London: Longman, 1987); Scott Sagan, “The Origins of the Pacific War,” in The Origin and Prevention of Major Wars, ed. Robert Rotberg and Theodore Rabb (Cambridge: Cambridge University Press, 1988), 323–52; and Roberta Wohlstetter, Pearl Harbor: Warning and Decision (Stanford, CA: Stanford University Press, 1962). 56. George and Smoke define deterrence as “the persuasion of one’s opponents that the costs and/or risks of a given action he might take outweigh the benefits.” Alexander L. George and Richard Smoke, Deterrence in American Foreign Policy: Theory and Practice (New York: Columbia University Press, 1974), 11. 57. Glenn Snyder originally made the distinction between deterrence through denial (or defense) and deterrence through punishment. Glenn Snyder, Deterrence and Defense: Toward a Theory of National Security (Princeton, NJ: Princeton University Press, 1961). 58. Patrick Morgan, Deterrence: A Conceptual Analysis (Beverly Hills, CA: Sage, 1977), 30. 59. Patrick Morgan, Deterrence Now (Cambridge: Cambridge University Press, 2003), 13. 60. Richard Ned Lebow long ago debunked the idea that states jump through windows of opportunity. There are plentiful examples of states that decline to attack their rivals even though they have an advantage in military

Notes

545

capabilities and that advantage will not last long into the future. Richard Ned Lebow, “Windows of Opportunity: Do States Jump through Them?” International Security 9, no. 1 (Summer 1984): 147–86. 61. Morgan, Deterrence Now, 117. Although some authors might disagree, deterrence is generally not about using military threats to achieve gains. Morgan argues this excessively broadens the definition of deterrence. 62. Morgan, Deterrence Now, 118. 63. See Robert Jervis, “Deterrence, Rogue States, and the U. S. Policy,” in Complex Deterrence, ed. T. V. Paul, P. Morgan, and J. Wirtz (Chicago: University of Chicago Press), 133–57. 64. Morgan, Deterrence Now, 2–3. 65. Morgan, Deterrence Now, 9, and more generally 80–115. Morgan’s idea of general deterrence is similar to Lawrence Freedman’s concept of “internalized deterrence.” Lawrence Freedman, Deterrence (Cambridge, UK: Polity, 2004). 66. Morgan, Deterrence Now, 80. 67. Patrick Morgan and T. V. Paul, “Deterrence among Great Powers in an Era of Globalization,” in Paul et al., Complex Deterrence, 259–76; see 265–71. Morgan and Paul argue that today no great power dyads are enduring rivals, and great powers spend most of their attention on deterring rogue states than other great powers. Only the U.S.-China-Taiwan relationship is influenced directly by deterrence. 68. Morgan, Deterrence Now, 84. 69. See, for instance, James L. Payne, The American Threat: The Fear of War as an Instrument of Foreign Policy (Chicago: Markham, 1970), and the revised edition, Payne, The American Threat: National Security and Foreign Policy (College Station, TX: Lytton, 1981). The discussion of deterrence in this chapter follows the discussion in Payne. 70. James Fearon, “Signaling Versus the Balance of Power and Interests,” Journal of Conflict Resolution 38, no. 2 (1994): 236–69; and Fearon, “Domestic Political Audiences and the Escalation of International Disputes,” American Political Science Review 88 (September 1994): 577–92. The basic conceptual foundations for costly signals can be found in Thomas Schelling, Arms and Influence (New Haven, CT: Yale University Press, 1966). 71. Payne, The American Threat (1970), 172. 72. Payne, The American Threat (1970), 139. 73. Charles Lockhart, “Problems in the Management and Resolution of International Conflicts,” World Politics 29 (April 1977): 378–403. 74. The following section is based on Payne, The American Threat (1970), 1–22. 75. John Vasquez, The War Puzzle (Cambridge: Cambridge University Press, 1993), 198–224. 76. See for instance, Timothy W. Crawford, “The Endurance of Extended Deterrence: Continuity, Change, and Complexity in Theory and Practice,” in Paul et al., Complex Deterrence, 277–303. 77. Payne, The American Threat (1970), 7 and 113–26. 78. Neil Sheehan, Hedrick Smith, E. W. Kenworthy, Fox Butterfield, and Daniel Ellsberg, Pentagon Papers (New York: Bantam Books, 1971), 491–92, quoted in Jonathan Mercer, Reputation and International Politics (Ithaca, NY: Cornell University Press), 12. 79. Earl Ravenal, “Counterforce and Alliance: The Ultimate Connection,” International Security 6, no. 4 (1982): 28, quoted in Morgan, “Saving Face for the Sake of Deterrence,” 135. 80. Robert Jervis, System Effects: Complexity in Political and Social Life (Princeton, NJ: Princeton University Press, 1997), 146, 255–58, cited in Morgan, Deterrence Now, 51 81. Jonathan Mercer, Reputation and International Politics (Ithaca, NY: Cornell University Press, 1996). 82. Mercer, Reputation and International Politics, 10. 83. Paul Huth and Bruce Russett, “What Makes Deterrence Work? Cases from 1900–1980,” World Politics 36 (July 1984): 496–526; Huth and Russett, “Deterrence Failure and Crisis Escalation,” International Studies Quarterly 32 (1988): 29–45; Paul Huth, “Extended Deterrence and the Outbreak of War,” American Political Science Review 82 (1988): 423–43; and Huth, Extended Deterrence and the Prevention of War (New Haven, CT: Yale University Press, 1988). 84. Mercer, Reputation and International Politics, 22. 85. Daryl Press, Calculating Credibility: How Leaders Assess Military Threats (Ithaca, NY: Cornell University Press, 2005). 86. Ted Hopf, Peripheral Visions: Deterrence Theory and American Foreign Policy in the Third World (Ann Arbor: University of Michigan Press, 1995). On the other hand, see Gregory Miller, The Shadow of the Past: Reputation and Military Alliances before the First World War (Ithaca, NY: Cornell University Press, 2012) who

546

Notes

argues that having a poor reputation has costs. States with bad reputations find it difficult to find allies and must pay a higher price to entice potential allies. 87. John Orme, Deterrence, Reputation and Cold-War Cycles (New York: Macmillan, 1992), cited in Huth, “Deterrence and International Conflict,” 42. 88. Morgan, Deterrence Now, 102–3. 89. Morgan, Deterrence Now, 102. See also Timothy Crawford’s summary of the evidence in Timothy Crawford, “The Endurance of Extended Deterrence.” 90. Huth, “Deterrence and International Conflict,” 38, and generally 37–41. 91. Leng, “When Will They Ever Learn: Coercive Bargaining in Recurrent Crises,” Journal of Conflict Resolution 27 (1983): 379–419; Leng, “Reagan and the Russians,” American Political Science Review 78 (1984): 338–55; Leng, Interstate Crisis Bargaining Behavior, 1816–1980 (Cambridge: Cambridge University Press, 1993). 92. These ideas go back to the modern classics of deterrence literature: George and Smoke, Deterrence in American Foreign Policy, and to Snyder and Diesing, Conflict among Nations. See also Timothy Crawford, “The Endurance of Extended Deterrence.” 93. James Fearon, “Signaling Foreign Policy Interests,” Journal of Conflict Resolution 41 (February 1997): 84. 94. See, for instance, Kenneth Lieberthal, “Preventing War over Taiwan,” Foreign Affairs 84, no. 2 (2005): 53–63. 95. Bruce Russett, “The Calculus of Deterrence,” in International Politics and Foreign Policy, rev. ed., ed. James Rosenau (New York: Free Press, 1969), 359–69. 96. Russett, “The Calculus of Deterrence,” 364–68. 97. Paul Huth and Bruce Russett, “What Makes Deterrence Work? Cases from 1900–1980,” World Politics 36 (July 1984): 496–526. 98. Huth and Russett, “What Makes Deterrence Work?” 516–18. 99. Paul Huth, “Extended Deterrence and the Outbreak of War,” American Political Science Review 82 (June 1988): 423–43; Huth, Extended Deterrence and the Prevention of War (New Haven: Yale University Press, 1988); and Huth and Russett, “Deterrence Failure and Crisis Escalation,” International Studies Quarterly 32 (1988): 29–45. Their analysis of deterrence in this extended period was unable to support the earlier finding that strong economic and political ties between defender and protégé increase the chance of deterrence success. They attribute this to the inclusion of six new cases from the pre-1900 period. In this earlier era arms transfers were a less frequent and probably less relevant aspect of the political-military relationship between two states. On this point see Huth, “Extended Deterrence and the Outbreak of War,” 436. 100. Alan Alexandroff and Richard Rosecrance, “Deterrence in 1939,” World Politics 29 (April 1977): 404–24. 101. Huth, “Extended Deterrence and the Outbreak of War,” 436. While the behavior of the defender in most recent confrontations with the attacker had a significant impact on deterrent success, the past behavior of the defender in deterrent situations in general did not. 102. Huth and Russett, “Deterrence Failure and Crisis Escalation,” 39. 103. Huth, “Extended Deterrence and the Outbreak of War,” 438. 104. Robert Jervis, “Rational Deterrence: Theory and Evidence,” World Politics 41 (January 1989): 198–99. 105. Huth and Russett, “Deterrence Failure and Crisis Escalation,” 39. 106. Huth, “Extended Deterrence and the Outbreak of War,” 426. 107. Jervis, “Rational Deterrence: Theory and Evidence,” 192. 108. John Orme, “Deterrence Failures: A Second Look,” International Security 11 (Winter 1986–87): 97. The idea that the attacker may be pushed into an attack on the client state is based, in part, on the idea that attackers have both defensive and offensive motivations for their actions and that this should be incorporated into deterrence theory. See Huth, “Extended Deterrence and the Outbreak of War,” 424. 109. Janice Gross Stein, “Deterrence and Reassurance,” in Behavior, Society, and Nuclear War, vol. 2, ed. Philip Tetlock et al. (New York: Oxford Press, 1991), 9–72. 110. Brett V. Benson, “Alliances: ATOP Data and Deterrence,” in What Do We Know About War? 2nd ed., ed. John Vasquez (Lanham, MD: Rowman & Littlefield, 2012), 45–62. See also Benson, “Unpacking Alliances: Deterrent and Compellent Alliances and Their Relationship with Conflict, 1816–2000,” Journal of Politics 73, no. 4 (2011): 1111–27, and Brett Leeds, “Do Alliances Deter Aggression? The Influence of Military Alliances on the Initiation of Militarized Interstate Disputes,” American Journal of Political Science 47, no. 3 (2003): 427–39. 111. Clinton Fink, “More Calculations About Deterrence,” Journal of Conflict Resolution 9 (1965): 54–66. 112. Morgan, Deterrence Now, 128.

Notes

547

113. A further problem is, according to George and Smoke, that Russett defines deterrence success as a situation that occurs when an attack on a pawn is (1) prevented or (2) repulsed without conflict between the attacking force and the regular combat units of the great power defender. The inclusion of point 2 leads Russett to include as examples of success some cases that others would identify as failures—for example, the unsuccessful French and British attack on Egypt in 1956 and the Soviet blockade of Berlin in 1948. George and Smoke conclude that Russett seems to blur the distinction between successful deterrence and successful defense. George and Smoke, Deterrence in American Foreign Policy, 516–17. 114. Richard Ned Lebow and Janice Gross Stein, “Deterrence: The Elusive Dependent Variable,” World Politics 42 (April 1990): 336–69. Critics of Lebow and Stein retort that by using too stringent criteria for the selection of deterrent cases they miss far too many real cases. See Paul Huth and Bruce Russett, “Testing Deterrence Theories: Rigor Makes a Difference,” World Politics 42 (July 1990): 466–501. 115. Lebow and Stein, “Deterrence,” 348. 116. Frank Harvey, “Deterrence and Compellence in Protracted Crises: Methodology and Preliminary Findings,” International Studies Notes 22, no. 1 (1997): 12–23. 117. George and Smoke, Deterrence in American Foreign Policy, 516–17. A third possibility also exists—one of mixed outcomes—cases in which some actions are deterred, but some other, lesser acts are not. 118. Morgan, Deterrence: A Conceptual Analysis, 23. 119. George and Smoke, Deterrence in American Foreign Policy. 120. George and Smoke, Deterrence in American Foreign Policy, 527, 529. 121. George and Smoke, Deterrence in American Foreign Policy, 526. Other factors that may play a secondary role in determining the success or failure of deterrence are (1) the initiator’s perception of the adequacy and appropriateness of the defender’s military capability, (2) his perception of a defender’s motivation, (3) his perception that only force can bring about the desired change, and (4) his willingness to accept compensation elsewhere as an alternative to challenging the defender (530–31). 122. On this point, see Morgan, Deterrence: A Conceptual Analysis, 141–42. 123. George and Smoke, Deterrence in American Foreign Policy, 540–44. 124. George and Smoke, Deterrence in American Foreign Policy, 508, 532. 125. Morgan, Deterrence: A Conceptual Analysis, 143. 126. Studies cited in Paul Huth, “Deterrence and International Conflict: Empirical Findings and Theoretical Debates,” American Review of Political Science 2 (1999): 25–48. Huth’s summary draws on Uri Bar-Joseph, “The Conceptualization of Deterrence in Israeli Strategic Thinking.” Security Studies 7 (1998): 145–81; E. Lieberman, “The Rational Deterrence Theory Debate,” Security Studies 3 (1994): 384–427; Lieberman, “What Makes Deterrence Work?” Security Studies 4 (1995): 833–92; B. Morris, Israel’s Border Wars (Oxford: Clarendon Press, 1993); and J. Shimshoni, Israel and Conventional Deterrence (Ithaca, NY: Cornell University Press, 1988). 127. Coalition goals included attempting to compel Iraq to accept no-fly zones in the north and south, to permit a UN humanitarian presence in Iraq, to end its various attacks on Iraqi citizens, to withdraw its military from the Kuwait border and from areas near the two NFZs, and to permit unimpeded inspections by UNSOM and the IAEA. The coalition also attempted to deter Iraq from further actions against its civilians, from violating UN agreements, and from further aggression against Kuwait. Simultaneously, Iraq was involved in attempts to compel UN members to loosen or end the sanctions regime and to deter further encroachments on Iraqi sovereignty as well as a military invasion by the U.S. coalition. 128. Frank P. Harvey and Patrick James, “Deterrence and Compellence in Iraq, 1991–2003: Lessons for a Complex Paradigm,” in Paul et al., Complex Deterrence, 222–56. 129. Harvey and James, “Deterrence and Compellence in Iraq,” 230. 130. Harvey and James, “Deterrence and Compellence in Iraq,” 244–45. 131. Amatzia Baram, “Deterrence Lessons from Iraq: Rationality Is Not the Only Key to Containment,” Foreign Affairs 91, no. 4 (July/August 2012): 76–90. 132. Baram, “Deterrence Lessons from Iraq,” 89. 133. Baram, “Deterrence Lessons from Iraq,” passim. 134. Baram, “Deterrence Lessons from Iraq,” 90. 135. Christopher H. Achen and Duncan Snidal, “Rational Deterrence Theory and Comparative Case Studies,” World Politics 41 (January 1989): 152. Additionally, a state can rationally choose to fight a war it thinks it will probably lose if the gains from winning and/or costs of alternative policies are great enough. See Jervis, “Rational Deterrence,” 187.

548

Notes

136. Richard Ned Lebow, Between Peace and War: The Nature of International Crisis (Baltimore: Johns Hopkins University Press, 1981), 82–87. 137. See R. N. Lebow, “Miscalculations in the South Atlantic: The Origins of the Falklands War,” in Psychology and Deterrence, ed. R. Jervis, R. N. Lebow, and J. G. Stein (Baltimore: Johns Hopkins University Press, 1985), 122. 138. Morgan, Deterrence Now, 110. 139. J. G. Stein, “Extended Deterrence in the Middle East.” None of these deterrence attempts were undertaken to deter an impending military attack per se against an American protégé; most were attempts to deter escalations of ongoing wars in the region. 140. Consistent with deterrence theory, Stein finds that in three of the four cases of failure, the deterrence strategy was weakly articulated, but she concludes that faulty articulation was, by itself, not a sufficient explanation of failure. 141. Huth, “Deterrence and International Conflict,” 40. 142. Robert Jervis, “Deterrence, Rogue States, and the U. S. Policy,” in Paul et al., Complex Deterrence, 137. 143. For these conclusions, see Morgan, Deterrence Now, 113, 162–63, and 149–50. 144. See T. V. Paul, “Nuclear Taboo and War Initiation in Regional Conflicts”; Janice Gross Stein, “Calculation, Miscalculation, and Conventional Deterrence 1: The View from Cairo,” in Jervis et al., Psychology and Deterrence, 34–59; Yair Evron, “Deterrence Experience in the Arab-Israeli Conflict,” in Deterrence in the Middle East: Where Theory and Practice Converge, ed. A. Klieman and A. Levite (Boulder, CO: Westview Press, 1994), 98–121; Uri Bar-Joseph, “Variations on a Theme: The Conceptualization of Deterrence in Israeli Strategic Thinking,” Security Studies 7, no. 3 (1998): 145–81. 145. Huth, “Deterrence and International Conflict,” 35. 146. Christopher Achen and Duncan Snidal, “Rational Deterrence Theory and Comparative Case Studies,” World Politics 41 (January 1989): 143–69; James Fearon, “Selection Effects and Deterrence,” International Interactions 28, no. 5 (January 2002): 5–29; and see Knopf, “Three Items in One.” 147. John Orme, “Deterrence Failures: A Second Look,” 121–22. See also Achen and Snidal, “Rational Deterrence Theory and Comparative Case Studies,” and Fearon, “Selection Effects and Deterrence.” 148. Frank Harvey, “Rigor Mortis, or Rigor, More Tests: Necessity, Sufficiency, and Deterrence Logic,” International Studies Quarterly 42 (December 1998): 675–707. 149. J. G. Stein, “Rational Deterrence against ‘Irrational’ Adversaries? No Common Knowledge,” in Paul et al., Complex Deterrence, 58–82. In the first version a state will only undertake a challenge to a deterrent threat if the estimated benefits exceed the costs. (Challengers who choose not to maximize expected utilities are classified as nonrational and are thus outside the scope of the theory.) In the looser version, in order to be logical, people must at a minimum be able to rank their preferences in a transitive way, update their estimates in response to new evidence that is valid and reliable, and learn from history and apply propositions from the past in an appropriate way. 150. Stein, “Rational Deterrence against ‘Irrational’ Adversaries?” 78. 151. Stein, “Rational Deterrence against ‘Irrational’ Adversaries?” 70. 152. Stein, “Rational Deterrence against ‘Irrational’ Adversaries?” 79. 153. Stein, “Rational Deterrence against ‘Irrational’ Adversaries?” 64–65, 78. 154. Lebow and Stein, “Deterrence: The Elusive Dependent Variable.” 155. Stein, “Rational Deterrence against ‘Irrational’ Adversaries?” 67. 156. Morgan, Deterrence Now, 171. 157. Stein, “Rational Deterrence against ‘Irrational’ Adversaries?” 66. 158. Morgan, Deterrence Now, 140–48. 159. Morgan, Deterrence Now, 115. 160. Morgan, Deterrence Now, 134–148. 161. Stein, “Rational Deterrence against ‘Irrational’ Adversaries?” 162. The following discussion is based on the excellent review by Daniel Geller, “Nuclear Weapons and War,” in What Do We Know About War ?, 2nd ed., ed. John Vasquez (Lanham, MD: Rowman & Littlefield, 2010), 139–63. 163. Bruce Bueno de Mesquita and William Riker, “An Assessment of the Merits of Selective Nuclear Proliferation,” Journal of Conflict Resolution 26 (1982): 283–306; Robert Rauchhaus, “Evaluating the Nuclear Peace Hypothesis: A Quantitative Approach,” Journal of Conflict Resolution 53, no. 2 (2009): 258–77; Victor Asal and Kyle Beardsley, “Proliferation and International Crisis Behavior,” Journal of Peace Research 44 (2007): 139–55. 164. We might add one borderline case here as well—the Sino-Soviet border clashes in 1969. China possessed nuclear weapons at the time but her capability was minimal; moreover, the number of battle deaths did not reach the threshold of 1,000.

Notes

549

165. John Mueller, “The Essential Irrelevance of Nuclear Weapons,” International Security 13, no. 2 (Fall 1988): 55–79; and Mueller, Retreat from Doomsday: The Obsolescence of Major War (New York: Basic Books, 1989). 166. Morgan, Deterrence Now, 38–39, 132–33. 167. Paul Huth, D. Scott Bennett, and Christopher Gelpi, “System Uncertainty, Risk Propensity, and International Conflict among the Great Powers,” Journal of Conflict Resolution 36 (1992): 478–517. 168. Kyle Beardsley and Victor Asal, “Nuclear Weapons as Shields,” Conflict Management and Peace Science 26 (2009): 235–55; Daniel Geller, “Nuclear Weapons, Deterrence, and Crisis Escalation,” Journal of Conflict Resolution 34 (1990): 291–310; Rauchhaus, “Evaluating the Nuclear Hypothesis.” 169. Geller, “Nuclear Weapons and War.” 170. Glenn Snyder and Paul Diesing, Conflict among Nations: Bargaining, Decision Making, and System Structure in International Crises (Princeton, NJ: Princeton University Press, 1977), 455–56. 171. Glenn Snyder, “The Balance of Power and the Balance of Terror,” in The Balance of Power, ed. Paul Seabury (San Francisco: Chandler, 1965), 185–201. 172. See, for instance, T. V. Paul, “Causes of the India-Pakistan Enduring Rivalry,” in The India-Pakistan Conflict: An Enduring Rivalry, ed. T. V. Paul (Cambridge: Cambridge University Press, 2005), 3–26. 173. T. V. Paul and Nina Tannenwald are the two scholars most associated with the concept of the nuclear taboo. See T. V. Paul, The Tradition of Non-Use of Nuclear Weapons (Stanford, CA: Stanford University Press, 2009); and Paul, “Nuclear Taboo and War Initiation in Regional Conflicts,” Journal of Conflict Resolution 39, no. 4 (1995): 696–717; Nina Tannenwald, The Nuclear Taboo (Cambridge: Cambridge University Press, 2007); and Tannenwald, “Stigmatizing the Bomb: Origins of the Nuclear Taboo,” International Security 29, no. 4 (2005): 5–49. 174. Tannenwald, The Nuclear Taboo, and “Stigmatizing the Bomb.” 175. Paul, The Tradition of Non-Use of Nuclear Weapons. 176. Paul, “Nuclear Taboo and War Initiation in Regional Conflicts.” 177. Stein, “Calculation, Miscalculation, and Conventional Deterrence: The View from Cairo.” 178. Rauchhaus, “Evaluating the Nuclear Hypothesis.” 179. Geller, “Nuclear Weapons, Deterrence, and Crisis Escalation.” 180. A. F. K. Organski and Jacek Kugler, The War Ledger (Chicago: University of Chicago Press, 1980); Jacek Kugler, “Terror without Deterrence: Reassessing the Role of Nuclear Weapons,” Journal of Conflict Resolution 28 (1984): 470–506. 181. Geller, “Nuclear Weapons and War,” 161. 182. See T. V. Paul’s discussion of the failure to use nuclear weapons in ten significant wars from 1950 to 1995: “Nuclear Taboo and War Initiation in Regional Conflicts.” 183. Paul Huth, Extended Deterrence and the Prevention of War (New Haven, CT: Yale University Press, 1988); Huth, “Extended Deterrence and the Outbreak of War”; Huth and Russett, “What Makes Deterrence Work?”; Huth and Russett, “Deterrence Failure and Crisis Escalation.” The exceptions are Erich Weede, “Preventing War by Nuclear Deterrence or by Détente,” Conflict Management and Peace Science 6 (1981): 1–8; and Weede, “Extended Deterrence by Superpower Alliance,” Journal of Conflict Resolution 27 (1983): 231–54. 184. For an interesting argument about deterrence being responsible for the absence of a Second Korean War, see David C. Kang, “International Relations Theory and the Second Korean War,” International Studies Quarterly 47 (2003): 301–24. 185. Leaders may be motivated by both opportunity and vulnerability under different circumstances or simultaneously; the two are not mutually exclusive. See Achen and Snidal, “Rational Deterrence Theory and Comparative Case Studies,” 148–49, and Stein, “Extended Deterrence in the Middle East,” 329–34. 186. Jervis, “Rational Deterrence: Theory and Evidence,” 198. 187. This is the message of Robert Jervis, R. N. Lebow, and J. G. Stein’s Psychology and Deterrence. 188. Morgan, Deterrence Now, 53–55. 189. Morgan, Deterrence Now, 110–13, 149–51. 190. Morgan, Deterrence: A Conceptual Analysis, 61–62. 191. For a review of constructivist insights applicable to deterrence theory, see Amir Lupovici, “The Emerging Fourth Wave of Deterrence Theory—Toward a New Research Agenda,” International Studies Quarterly 54 (2010): 705–32. 192. Fink, “More Calculations About Deterrence.” 193. For an analysis of deterrence in this complex environment, see the articles in Paul et al., Complex Deterrence. 194. Morgan and Paul, “Deterrence among Great Powers in an Era of Globalization.”

550

Notes

10. THE INTERNATIONAL SYSTEM LEVEL OF ANALYSIS, PART I 1. Michael Sullivan, International Relations: Theories and Evidence (Englewood Cliffs, NJ: Prentice-Hall, 1976), 144. 2. The presence of either anarchy or hierarchy comprises the basic principle by which international systems are ordered according to Kenneth Waltz, Theory of International Politics (Reading, MA: Addison-Wesley, 1979), 89. 3. James E. Daugherty and Robert L. Pfaltzgraff Jr., Contending Theories of International Relations, 3rd ed. (New York: Harper & Row, 1990), 137. 4. Robert North, War, Peace, Survival: Global Politics and Conceptual Synthesis (Boulder, CO: Westview, 1990), 10. 5. For a contrary view—that the international system plays no determining role—see North, War, Peace, Survival. 6. International systems theory assumes that variables at other levels of analysis have a random effect on the behavior of states. Sullivan, International Relations, 153. 7. Waltz, Theory of International Politics, 72. 8. Evan Luard, War in International Society (New Haven, CT: Yale University Press, 1986), 385. See also his Types of International Society (New York: Free Press, 1976). 9. Found in Rousseau’s “A Discourse on the Origin of Inequality,” in The Social Contract and Discourses, trans. G. D. H. Cole (New York: Dutton, 1950), 238. 10. Kenneth Waltz, Man, the State and War (New York: Columbia University Press, 1959), 168–69. 11. On comparisons between the prisoners’ dilemma and stag hunt games, see Robert Jervis, “Cooperation under the Security Dilemma,” World Politics 30, no. 2 (1978): 167–214. 12. Waltz, Man, the State and War, 6–7. 13. Waltz, Man, the State and War, 188. 14. Jean-Jacques Rousseau, A Lasting Peace Through the Federation of Europe, trans. C. E. Vaughan (London: Constable, 1917), 78–79, quoted in Waltz, Man, the State and War, 180. 15. Robert Jervis, Perceptions and Misperceptions in International Politics (Princeton, NJ: Princeton University Press, 1976), 58–113. 16. For the classic critique of world government, see Inis Claude, Jr., Power and International Relations (New York: Random House, 1962). 17. See, for instance, William Thompson, “Systemic Leadership, Evolutionary Processes, and International Relations Theory: The Unipolarity Question,” International Studies Review 8 (2006): 1–22. 18. E. H. Carr, The Twenty Years Crisis: 1919–1939 (London: Macmillan, 1951); Hans Morgenthau, Politics among Nations (New York: Alfred A. Knopf, 1948); Arnold Wolfers, Discord and Collaboration (Baltimore: Johns Hopkins, 1962); Henry Kissinger, A World Restored (Boston: Houghton Mifflin, 1964) and Diplomacy (New York: Simon and Schuster, 1994); George F. Kennan, American Diplomacy, 1900–1950 (New York: Mentor Books, 1952). 19. See Jeffrey W. Legro and Andrew Moravcsik, “Is Anybody Still a Realist?” International Security 24, no. 2 (Fall 1999): 5–55; and Jeffrey Taliaferro, “Security Seeking under Anarchy,” International Security 25, no. 3 (Winter 2000/01): 128–61, especially 154–57. 20. Hans Morgenthau, Politics among Nations, 3rd ed. (New York: Knopf, 1961), 38–39. 21. Classical realists use somewhat different terminology: Kissinger uses the terms “revolutionary states” and “status-quo states”; Morgenthau refers to “imperialistic” and “status-quo powers”; Carr distinguishes between satisfied and dissatisfied powers. See the discussion in Randall Schweller, “Neorealism’s Status-Quo Bias: What Security Dilemma?” Security Studies 5, no. 3 (Spring 1996): 90–121, especially 98. 22. As we will see later, John Mearsheimer, an offensive realist, argues that there are no status quo great powers; all great powers seek revision of the international system in their favor. More specifically, they all seek to be a global or regional hegemon. John Mearsheimer, The Tragedy of Great Power Politics (New York: Norton, 2001), 2–3. 23. See Schweller, “Neorealism’s Status-Quo Bias,” 100. 24. Kenneth Waltz, Theory of International Politics (Reading, MA: Addison-Wesley, 1979). 25. “Neorealism” is an attempt to refine classical realist thought by injecting greater theoretical rigor, especially by defining concepts more clearly and by developing testable hypotheses. Power is still the key variable for neorealists, but the focus has been shifted from the nation-state level to the international system level. Waltz is the father of this approach, but neorealism has spawned many children who go by different names.

Notes

551

26. Waltz, Theory of International Politics. This is not to say that the system’s structure completely determines state behavior; it simply exerts a strong constraining influence. In return, the units of the system shape the structure of the system. Neorealists emphasize that states will behave differently in different international structures. Waltz identifies three aspects of international structure that are important in this respect: (1) the principle by which the system is organized—hierarchy or anarchy, (2) the specification of functions of the units (there is none in anarchy), and (3) the distribution of capabilities among the units. Waltz, Theory of International Politics, 82. 27. Waltz argues that to some extent, the behavior of states is subject to socialization: States tend to emulate the successful policies of others, and balancing is one of those successful policies. Waltz, Theory of International Politics, 74–77, 127–28. 28. Claude, Power and International Relations. 29. Jacek Kugler and A. F. K. Organski, “The Power Transition: A Retrospective and Prospective Evaluation,” in Handbook of War Studies, ed. Manus Midlarsky (Boston: Unwin Hyman, 1989), 176. 30. Randolph Siverson and Michael Sullivan, “The Distribution of Power and the Onset of War,” Journal of Conflict Resolution 27, no. 3 (1983): 477. See also Richard J. Stoll and Michael Champion, “Capability Concentration, Alliance Bonding, and Conflict among the Major Powers,” in Polarity and War: The Changing Structure of International Conflict, ed. Alan Ned Sabrosky (Boulder, CO: Westview, 1985), 67–94. Stoll and Champion remind us that the realists associated an equal balance with stability, not with peace (73–75). 31. Waltz, “The Origins of War in Neorealist Theory,” 44. 32. Classical realists like Morgenthau, Kissinger, and Wolfers did not dwell on international system level constraints. 33. Waltz, Theory of International Politics, 93–97. 34. The theoretical foundation for defensive realism can be found in Robert Jervis, “Cooperation under the Security Dilemma,” World Politics 30, no. 2 (January 1978): 167–214. The primary core of defensive realist thought includes Jervis, Perception and Misperception in International Politics (Princeton, NJ: Princeton University Press, 1976); Stephen Van Evera, “The Cult of the Offensive and the Origins of the First World War,” International Security 9, no. 1 (Summer 1984): 58–107; Van Evera, Causes of War: Power and the Roots of Conflict (Ithaca, NY: Cornell University Press, 1999); Charles Glaser, “The Security Dilemma Revisited,” World Politics 50, no. 1 (October 1997): 171–201; Stephen Walt, The Origins of Alliances (Ithaca, NY: Cornell University Press, 1995); and Jack Snyder, “Civil-Military Relations and the Cult of the Offensive, 1914 and 1984,” International Security 9, no. 1 (Summer 1984): 108–46. 35. Waltz, “The Origins of War in Neorealist Theory,” 40. 36. Jeffrey W. Taliaferro, “Security Seeking under Anarchy: Defensive Realism Revisited,” International Security 25, no. 3 (Winter 2000/01): 128–61; see 159. 37. Randall Schweller interprets Waltz’s dictum that states seek to maximize security rather than power to mean that they seek the minimum power needed to assure the state’s survival. And they seek relative advantage rather than absolute gains. Schweller, “Neorealism’s Status-Quo Bias,” 101–2. 38. Robert Jervis, “Cooperation under the Security Dilemma,” 169. 39. Glaser, “The Security Dilemma Revisited.” 40. Jack Snyder, Myths of Empire: Domestic Politics and International Ambition (Ithaca, NY: Cornell University Press, 1991). 41. Glaser, “The Security Dilemma Revisited,” also argues that states may have “unit level knowledge” of the adversary’s motives. For instance, leaders of democratic states may believe that nondemocratic states are not pure security-seekers; thus, they would be more likely to feel threatened by military buildups by those states, while if the same acts were performed by democratic states, they would be less likely to feel threatened since they recognize that democratic states are pure security-seekers. 42. Robert Jervis, “Realism, Neoliberalism, and Cooperation: Understanding the Debate,” International Security 24, no. 1 (Summer 1999): 42–63, 49. 43. Taliaferro, “Security Seeking under Anarchy,” 144–45. Taliaferro (147–49) cites Japanese expansion in the 1930s and 1940s and China’s intervention in the Korean War as examples of how fear of adversaries’ future intentions can provoke aggressive military actions. 44. Glaser, “The Security Dilemma Revisited,” 189; Taliaferro, “Security Seeking under Anarchy,” 128–29. 45. Stephen Walt, The Origins of Alliances (Ithaca, NY: Cornell University Press, 1987). 46. Walt, The Origins of Alliances, 21–34, 262–85. 47. Robert Jervis, “Cooperation Under the Security Dilemma,” World Politics 30, no. 2 (January 1978): 167–214. 48. Jervis, “Cooperation under the Security Dilemma,” 187.

552

Notes

49. Jervis, “Cooperation under the Security Dilemma,” 211. 50. Stephen Van Evera, Causes of War: Power and the Roots of Conflict (Ithaca, NY: Cornell University Press, 1999), 7–8. 51. Van Evera, Causes of War, 4. These conditions seem to be neither necessary nor sufficient conditions for war. 52. Van Evera, Causes of War, 6, 9. 53. Van Evera, Causes of War, 9. 54. However, as Richard Betts notes, it is pretty easy to cite historical quotes from leaders who were overoptimistic about the coming wars. Reality is somewhat more complex. Domestic policy makers in all states are likely to be split, with some being optimistic and others pessimistic, and analysts can cite ample evidence of both points of view for any country. The real state of mind of policy makers is more likely to be ambivalence. Betts, “Must War Find a Way?” 172–73. 55. Van Evera, Causes of War, 40–45, and n19. 56. Van Evera, Causes of War, 71. On the other hand, Richard Betts argues that real first-move advantages may not be all that rare; in three of Van Evera’s examples—WWI, Chinese intervention in 1950, and the 1967 ArabIsraeli War—real first move advantages existed. Richard K. Betts, “Must War Find a Way?” International Security 24, no. 2 (Fall 1999): 166–98; see 173. 57. Van Evera distinguishes between windows of opportunity (which are fading offensive opportunities for a state) and windows of vulnerability (which are growing defensive vulnerability). Van Evera, Causes of War, 74. 58. Van Evera, Causes of War, 104. 59. Van Evera, Causes of War, 73. 60. See Betts, “Must War Find a Way?” 174–75. 61. Van Evera, Causes of War, 81–82. 62. Van Evera, Causes of War, 104. 63. Van Evera, Causes of War, 108–9. 64. Van Evera, Causes of War, 115–16. 65. The logic of the argument seems to be that when conquest is easy, war is more likely for the international system as a whole, for dyads, and for individual states. 66. Van Evera, Causes of War, 117. Emphasis added. 67. Van Evera, Causes of War, 119–22, especially diagram 3. Under certain, relatively rare, circumstances offense dominance can decrease the risk of war. For instance, purely one-sided possession of offensive capabilities by a status quo power can increase deterrent capabilities against potential aggressors under certain conditions, 121–22, 152–60. 68. Van Evera, Causes of War, 122. 69. Van Evera, Causes of War, 192. 70. Kier Lieber has put forward a comprehensive critique of the defensive realist analysis of World War I. He argues that the most recent historiography of that war runs counter to many of the positions taken by Van Evera and other defensive realists regarding fundamental issues such as the existence of a security dilemma, preemptive motives, inadvertent war, conflict spiral dynamics, and the importance of offense-defense theory. This appears crucial because much of the concepts and theoretical arguments central to defensive realism appear to have been built upon the history of the July crisis. Keir Lieber, “The New History of World War I and What It Means for International Relations Theory,” International Security 32, no. 2 (Fall 2007): 155–91. 71. An easy way—but certainly not the only way—to think about the offense-defense balance is simply as the numerical advantage needed for offensive forces to overcome defensive forces on the battlefield. Thus the O-D balance may be seen as a ratio (5–1, 3–1, 2–1, etc.) of offensive forces to defensive forces. 72. Betts, “Must War Find a Way?” see especially 185–90. See also the comments by James W. Davis, Jr., Bernard I. Finel, and Stacie Goddard in “Correspondence: Taking Offense at Offense-Defense Theory,” International Security 23, no. 3 (Winter 1998/1999): 179–206. 73. See letters by James W. Davis Jr. and Bernard I. Finel in “Correspondence: Taking Offense at OffenseDefense Theory.” 74. John Mearsheimer, The Tragedy of Great Power Politics (New York: Norton, 2001), 1–3, 32–36. 75. Randall L. Schweller, “Bandwagoning for Profit: Bringing the Revisionist State Back In,” International Security 19, no. 1 (Summer 1994): 72–107; and Schweller, “Neorealism’s Status-Quo Bias.” 76. Schweller, “Neorealism’s Status-Quo Bias,” 115–16. 77. Schweller, “Neorealism’s Status-Quo Bias,” 104. 78. Schweller, “Neorealism’s Status-Quo Bias,” 117.

Notes

553

79. Schweller, “Neorealism’s Status-Quo Bias,” 106–7. Revisionist states tend to bandwagon with each other, rather than engage in balancing behavior. See also Randall Schweller, Deadly Imbalances: Tripolarity and Hitler’s Strategy of World Conquest (New York: Columbia University Press, 1997). 80. For a summarization of neoclassical realism, see Steven E. Lobell, Norrin M. Ripsman, and Jeffrey W. Taliaferro (eds.), Neoclassical Realism, the State, and Foreign Policy (Cambridge: Cambridge University Press, 2009), especially chapter 1, “Introduction: Neoclassical Realism, the State and Foreign Policy,” by Taliaferro, Lobell, and Ripsman, 1–41. There is no single neoclassical theory of foreign policy per se; rather there are several theories of foreign policy behavior associated with this approach. 81. Taliaferro et al., “Introduction,” 7. 82. Taliaferro et al., “Introduction,” 28. 83. Taliaferro et al., “Introduction,” 4. 84. Taliaferro, “Security Seeking under Anarchy,” 157, explains that “For any foreign policy theory to explain state behavior, it must specify the mechanism through which the independent variable translates into policies.” 85. Taliaferro et al., “Introduction,” 25. 86. Taliaferro et al., “Introduction,” 26. 87. Taliaferro et al., “Introduction.” 29. 88. Taliaferro et al., “Introduction,” 4. 89. Taliaferro et al., “Introduction,” 22. Neoclassical realists argue that this is in fact consistent with the beliefs of older classical realists like Morgenthau. 90. See, for instance, Greg Cashman and Leonard C. Robinson, chapter 7 “The Iraq War,” in An Introduction to the Causes of War (Lanham, MD: Rowman & Littlefield, 2007), 297–363. See also Robert Jervis, “Understanding the Bush Doctrine,” Political Science Quarterly 118, no. 3 (Fall 2003): 365–88; Chaim Kaufmann, “Threat Inflation and the Failure of the Marketplace for Ideas: The Selling of the Iraq War,” International Security 29, no. 4 (Summer 2004): 5–48; and Jonathan Monten, “The Roots of the Bush Doctrine: Power, Nationalism, and Democracy Promotion in Grand Strategy,” International Security 29, no. 4 (Spring 2005): 112–56. 91. See Cashman and Robinson, An Introduction to the Causes of War, 338–39. 92. Taliaferro, “Security Seeking under Anarchy.” Although he argues that these four assumptions characterize both defensive neorealism and defensive neoclassical realism, they are clearly more reflective of the latter. 93. For a review of empirical tests of realist assumptions, see John Vasquez, The Power of Power Politics: A Critique (New Brunswick, NJ: Rutgers University Press, 1983). 94. This classification is drawn from David Garnham, “The Causes of War: Systemic Findings,” in Polarity and War: The Changing Structure of International Conflict, ed. Alan Ned Sabrosky (Boulder, CO: Westview, 1985), 7–23. 95. For an approach that calls for a twofold classification based on the number of military centers of power and the number of political centers of decision, see Patrick James and Michael Brecher, “Stability and Polarity: New Paths for Inquiry,” Journal of Peace Research 25, no. 1 (1988): 31–42. 96. Jack S. Levy, “The Polarity of the System and International Stability: An Empirical Analysis,” in Sabrosky, Polarity and War, 47. 97. Levy defines nonpolarized systems as characterized by the presence of a number of alliances and the absence of two well-defined clusters and either (1) a series of cross-cutting alliance bonds or (2) rapidly shifting alliance partners. Jack S. Levy, “Alliance Formation and War Behavior: An Analysis of the Great Powers, 1495–1975,” Journal of Conflict Resolution 25 (1981): 607. 98. Michael Wallace, “Polarization: Toward a Scientific Conception,” in Sabrosky, Polarity and War, 95–113. 99. Frank Wayman demonstrates that power polarity and cluster polarity are two uncorrelated and empirically separate dimensions. Frank Wayman, “Bipolarity, Multipolarity, and the Threat of War,” in Sabrosky, Polarity and War, 115–44. See also Jeffrey Hart, “Symmetry and Polarization in the European International System, 1870–1879: A Methodological Study,” Journal of Peace Research 11 (1974): 229–44, and Jeffrey Hart, “Power and Polarity in the International System,” in Sabrosky, Polarity and War, 25–40. See also Levy, “The Polarity of the System and International Stability,” 47–48. 100. David Rapkin, William R. Thompson, and Jon Christopherson, “Bipolarity and Bipolarization in the Cold War Era,” Journal of Conflict Resolution 23, no. 2 (1979): 261–95. They use a “bloc flow” approach to the measurement of (bi)polarization, which determines the extent to which conflict behavior (of a variety of types, not necessarily military) is concentrated between blocs and cooperative behavior is concentrated within blocs. 101. Several analysts suggest that alternative systems containing elements of both bipolarity and multipolarity may prove the most stable. For instance, Richard Rosecrance examines the system of “bimultipolarity” and Manus

554

Notes

Midlarsky discusses a system of “hierarchical equilibrium.” See R. Rosecrance, “Bipolarity, Multipolarity, and the Future,” 325–35 in International Politics and Foreign Policy, rev. ed., ed. J. Rosenau (New York: Free Press, 1969); and M. Midlarsky, “Hierarchical Equilibria and the Long-Run Instability of Multipolar Systems,” 64–74 in Midlarsky, Handbook of War Studies. Both Rosecrance’s “bimultipolarity” and Midlarsky’s “hierarchical equilibrium” model share some features with Morton Kaplan’s “loose bipolar” model (and his “very loose bipolar” model). See M. Kaplan, “Variants on Six Models of the International System,” in Rosenau, International Politics and Foreign Policy, 291–303. 102. See Henry Kissinger, A World Restored: The Politics of Conservatism in a Revolutionary Age (New York: Grossett and Dunlap, 1964). 103. See Karl Deutsch and J. David Singer, “Multipolar Power Systems and International Stability,” World Politics 16, no. 3 (1964): 390–406; and Lloyd Jensen, Explaining Foreign Policy (Englewood Cliffs, NJ: PrenticeHall, 1982), 255. 104. Deutsch and Singer, “Multipolar Power Systems and International Stability.” 105. One of the classic critiques of the Deutsch and Singer argument for the stability of multipolarity is found in Rosecrance, “Bipolarity, Multipolarity, and the Future.” 106. Midlarsky, “Hierarchical Equilibria and the Long-Run Instability of Multipolar Systems.” 107. The classic formulation in support of bipolarity is found in Kenneth Waltz, “International Structure, National Force, and the Balance of World Power,” in Rosenau, International Politics and Foreign Policy, 304–14. See also his Theory of International Politics. 108. Wayman, “Bipolarity, Multipolarity, and the Threat of War.” Wayman’s measure of cluster polarity is the ratio of the actual poles (alliance blocs and nonaligned great powers) to the number of potential poles. 109. Bruce Bueno de Mesquita, “Systemic Polarization and the Occurrence and Duration of War,” Journal of Conflict Resolution 22 (June 1978): 241–67. 110. Wallace, War and Rank Among Nations, and “Alliance Polarization, Cross-Cutting, and International War, 1815–1964,” Journal of Conflict Resolution 17 (December 1973): 576–604. 111. Similar results were found by two other scholars, Charles Kegley and Gregory Raymond, “Alliance Norms and War: A New Piece in the Old Puzzle,” International Studies Quarterly 26 (December 1982): 572–95. They find that extremely flexible and rigid alliances are associated with a high magnitude and severity of war. 112. Levy, “Alliance Formation and War Behavior,” 608. 113. This is a point made by Wallace, “Polarization: Toward a Scientific Conception,” 110. 114. J. David Singer and Melvin Small. “Alliance Aggregation and the Onset of War, 1815–1945,” in Quantitative International Politics, ed. J. David Singer (New York: Free Press, 1967), 247–86; Singer and Small, “Foreign Policy Indicators: Predictors of War in History and in the State of the World Message,” Policy Sciences 5: 271–296. 115. Charles W. Ostrom and Francis W. Hoole, “Alliances and War Revisited: A Research Note,” International Studies Quarterly 22 (June 1978): 215–36. 116. Levy, “Alliance Formation and War Behavior.” 117. One dissenting study is Frank Wayman, “Alliances and War: A Time-Series Analysis, in Prisoners of War? ed. C. Gochman and A. Sabrosky (Lexington, MA: Lexington Books, 1990), 93–113. Wayman’s is one of the few studies finding no evidence that alliances increase the probability of war within three to five years. He finds that at the dyadic level, defensive pacts produce no significant increase in war for their members. 118. Levy, “Alliance Formation and War Behavior,” 590, 605. 119. Wallace, “Polarization: Towards a Scientific Conception.” 120. J. David Singer and Melvin Small, “National Alliance Commitments and War Involvement, 1815–1945,” Peace Research Society (International) Papers 5 (1966): 109–40; and Singer and Small, “Formal Alliances: 1815–1939: A Quantitative Description.” Journal of Peace Research 3 (1966): 1–32. 121. Douglas Gibler and John Vasquez, “Uncovering Dangerous Alliances, 1495–1980,” International Studies Quarterly 42 (1998): 785–807. 122. Douglas Gibler, “Reconceptualizing the Alliance Variable: An Empirical Typology of Alliances,” PhD dissertation, Vanderbilt University, 1997, cited in Gibler and Vasquez, “Uncovering Dangerous Alliances.” 123. Paul Senese and John Vasquez, The Steps to War (Princeton, NJ: Princeton University Press, 2008). 124. Senese and Vasquez, The Steps to War, chapters 5 and 6. 125. Senese and Vasquez, The Steps to War. 126. Brett A. Leeds, Jeffrey M. Ritter, Sara McLaughlin Mitchell, and Andrew W. Long, “Alliance Treaty Obligations and Provisions, 1815–1944,” International Interactions 28, no. 3 (2002): 237–60.

Notes

555

127. Alan Ned Sabrosky, “Interstate Alliances: Their Reliability and the Expansion of War,” in The Correlates of War: Testing Some Realpolitik Models, vol. 2, ed. J. David Singer (New York: Free Press, 1980). 128. Paul W. Schroeder. “Alliances, 1815–1945: Weapons of Power and Tools of Management,” in Historical Dimensions of National Security Problems, ed. K. Knorr (Lawrence: University of Kansas Press, 1976), 227–62; and William B. Moul, “Great Power Nondefense Alliances and the Escalation to War of Conflicts between Unequals, 1815–1939,” International Interactions 15, no. 1 (1988): 25–43. 129. Brett Ashley Leeds, “Do Alliances Deter Aggression? The Influence of Military Alliances on the Initiation of Militarized Interstate Disputes,” American Journal of Political Science 47, no. 3 (2003): 427–39. 130. Brett Ashley Leeds, “Alliances and the Expansion and Escalation of Militarized Interstate Disputes,” in New Directions in International Relations, ed. Alex Mintz and Bruce Russett (Lanham, MD: Lexington Books, 2005), 117–34. See the commentary in John Vasquez, The War Puzzle Revisited (Cambridge: Cambridge University Press, 2009), 382. 131. Brett Benson, “Alliances: ATOP Data and Deterrence,” in What Do We Know about War? 2nd ed., ed. John Vasquez (Lanham, MD: Rowman & Littlefield, 2012), 45–62. 132. Douglas Gibler, “Alliances That Never Balance: The Territorial Settlement Treaty,” Conflict Management and Peace Science 15 (1996): 75–97; Gibler, “Control the Issues, Control the Conflict: The Effects of Alliances That Settle Territorial Issues on Interstate Rivalries,” International Interactions 22 (1997): 341–68; and Gibler, “Alliances: Why Some Cause War and Why Others Cause Peace.” 133. Levy, “Alliance Formation and War Behavior,” 600. 134. John Vasquez, “The Steps to War: Toward a Scientific Explanation of Correlates of War Findings,” World Politics 40, no. 1 (October 1987): 121. 135. Vasquez, “The Steps to War,” 123. 136. Levy, “Alliance Formation and War Behavior,” 604. 137. See, for instance, Randolph Siverson and Joel King, “Alliances and the Expansion of War,” in To Auger Well: Early Warning Indicators in World Politics, ed. J. David Singer and Michael Wallace (Beverly Hills, CA: Sage, 1982), 37–49. Also, Siverson and Harvey Starr, “Opportunity, Willingness, and the Diffusion of War, 1816–1965,” American Political Science Review 84 (March 1990): 47–67. 138. Benjamin Most and Harvey Starr, “Diffusion, Reinforcement, Geopolitics, and the Spread of War,” American Political Science Review 74 (December 1980): 932–46. Both geographic proximity and alliance ties are correlated with war involvement, although alliance membership has the greater effect of the two. However, the combined effects of both factors have an even greater effect than either variable by itself. See B. Most, P. Schrodt, R. Siverson, and H. Starr, “Border and Alliance Effects in the Diffusion of Major Power Conflict, 1816–1965,” in Prisoners of War? Nation-States in the Modern Era, ed. C. Gochman and A. N. Sabrosky (Lexington, MA: Lexington Books, 1990), 209–29. 139. Bruce Bueno de Mesquita, The War Trap (New Haven, CT: Yale University Press, 1981). 140. Stuart Bremer, “Dangerous Dyads: Conditions Affecting the Likelihood of Interstate War, 1816–1965,” Journal of Conflict Resolution 36 (1992): 309–41. James Lee Ray also shows that some of the problems were due to some of the original coding rules in the COW data set. J. L. Ray, “Friends as Foes: International Conflict and Wars between Formal Allies,” in Gochman and Sabrosky, Prisoners of War? 73–91. 141. Senese and Vasquez, The Steps to War. 142. J. David Singer, Stuart Bremer, and John Stuckey, “Capability Distribution, Uncertainty, and Major Power War, 1820–1965,” in Peace, War and Numbers, ed. Bruce Russett (Beverly Hills, CA: Sage, 1972), 19–48. CON ranges from 0 to 1.0; it is 0 when the distribution of capabilities is perfectly equal across the system, and it is 1.0 when one state holds 100 percent of the capabilities in the system. 143. Bruce Bueno de Mesquita, “Measuring Systemic Polarity,” Journal of Conflict Resolution 19 (June 1975): 187–216. 144. Wayman, “Bipolarity, Multipolarity, and the Threat of War.” 145. Ted Hopf, “Polarity, the Offense-Defense Balance, and War,” American Political Science Review 85 (1991): 475–94. 146. Charles Kegley and Gregory Raymond, A Multipolar Peace? Great Power Politics in the Twenty-First Century (New York: St. Martins, 1994). 147. Midlarsky, “Hierarchical Equilibria and the Long-Run Instability of the Multipolar System,” 70. 148. Levy, “The Polarity of the System and International Stability.” 149. D. Scott Bennett and Allan C. Stam, The Behavioral Origins of War (Ann Arbor: University of Michigan, 2004), 145–147.

556

Notes

150. Bennett and Stam, The Behavioral Origins of War, 146–47. 151. Wayman, “Bipolarity, Multipolarity and the Threat of War,” 138–39. However, Wayman’s research is unable to confirm the relationship between cluster multipolarity and peace for the nineteenth century. 152. Stoll and Champion, “Capability Concentration, Alliance Bonding, and Conflict Among the Major Powers.” 153. Mearsheimer, The Tragedy of Great Power Politics, 40–41. 154. Mearsheimer, The Tragedy of Great Power Politics, 44–45, 337–346. 155. See Mearsheimer, The Tragedy of Great Power Politics, 337–38. 156. Mearsheimer, The Tragedy of Great Power Politics, 357, table 9.3. 157. Mearsheimer, The Tragedy of Great Power Politics, 155–62, 267–72. 158. On this point see John Vasquez, “The Steps to War,” 128.

11. THE INTERNATIONAL SYSTEM LEVEL OF ANALYSIS, PART II 1. Melvin Tumin, Social Stratification (Englewood Cliffs, NJ: Prentice-Hall, 1967), 12. 2. Johann Galtung, “A Structural Theory of Aggression,” Journal of Peace Research 1, no. 2 (1964): 96. 3. Maurice East, “Status Discrepancy and Violence in the International System: An Empirical Analysis,” in The Analysis of International Politics, ed. James Rosenau, Vincent Davis, and Maurice East (New York: Free Press, 1972), 300. 4. Galtung, “A Structural Theory of Aggression”; East, “Status Discrepancy and Violence in the International System.” 5. Galtung, “A Structural Theory of Aggression,” 96, 99. It seems, however, on the basis of data from individual level studies, that a considerable amount of discrepancy needs to be present before aggression takes place. See Galtung (102) on this point. 6. Galtung, “A Structural Theory of Aggression,” 103. Examples of overachievers would be Japan in the 1920s and 1930s and China in the late twentieth century. Examples of underachievers are Austria-Hungary in the immediate pre-WWI period and Russia in the immediate post-Soviet era. 7. East, “Status Discrepancy and Violence in the International System,” 303. 8. East, “Status Discrepancy and Violence in the International System,” 303. 9. An interesting aspect of the theory is that it postulates an exact parallel between several levels of analysis— individual, state, and system—with rank disequilibrium and aggression being related at each level. Thus there exists the possibility of a causal link between levels. Disequilibrium at one level can lead to disequilibrium at another level. See Galtung, “A Structural Theory of Aggression,” 106. 10. Michael Sullivan, International Relations: Theories and Evidence (Englewood Cliffs, NJ: Prentice-Hall, 1976), 177. 11. East, “Status Discrepancy and Violence in the International System.” 12. Michael D. Wallace, “Status, Formal Organization, and Arms Levels as Factors Leading to the Onset of War, 1820–1964,” in Peace, War and Numbers, ed. Bruce Russett (Beverly Hills, CA: Sage, 1972), 49–69; and Wallace, “Power, Status, and International War,” Journal of Peace Research 8, no. 1 (1971): 23–36. See also Wallace, War and Rank among Nations (Lexington, MA: Heath, 1973). 13. James Lee Ray, “Status Inconsistency and War Involvement in Europe, 1816–1970,” Peace Science Society (International) Papers 23 (1974): 69–80; and Charles S. Gochman, “Status, Capabilities, and Major Power Conflict,” in The Correlates of War: II, ed. J. David Singer (New York: Free Press, 1980), 83–123. 14. A. F. K. Organski, World Politics (New York: Knopf, 1958, 1968). In addition, major works on power transition theory include A. F. K. Organski and Jacek Kugler, The War Ledger (Chicago: University of Chicago Press, 1980); Jacek Kugler and A. F. K. Organski, “The Power Transition: A Retrospective and Prospective Evaluation,” in Handbook of War Studies, ed. Manus Midlarsky (Ann Arbor: University of Michigan Press, 1989), 171–94; Jacek Kugler and Douglas Lemke (eds.), Parity and War: Evaluations and Extension of the War Ledger (Ann Arbor: University of Michigan Press, 1996); Jacek Kugler and Douglas Lemke, “The Power Transition Research Program: Assessing Theoretical and Empirical Advances,” in Handbook of War Studies II, ed. Manus Midlarsky (Ann Arbor: University of Michigan, 2000), 129–63; and Ronald Tammen et al., Power Transitions: Strategies for the 21st Century (New York: Chatham House, 2000). 15. Inis Claude, Power and International Relations (New York: Random House, 1962), 56.

Notes

557

16. Organski argued that it is the industrial revolution that has made rapid increases in power possible; prior to the age of industrialization, such large and rapid shifts were unlikely before this time. The coming of industrialization created three kinds of states: preindustrial, industrializing, and industrialized—each growing at significantly different rates. States that industrialized later were likely to grow more rapidly than those that had already achieved industrialization and their relatively rapid growth rate would permit them to catch up with and surpass the early industrializers. Thus, power transition theory—at least in its original formulation—was applicable only to great power relations after roughly 1750. Organski, World Politics, 307. 17. Actually, the theory suggests the crucial condition is both static parity of power between the dominant state and contender and a dynamic shift in power—leaving open the question of which one is most important. Power transition theorists have gone back and forth on this issue. 18. James Morrow, “The Logic of Overtaking,” in Parity and War, 313–330; see 316. 19. J. Kugler and D. Lemke, “The Power Transition Research Program: Assessing Theoretical and Empirical Advances,” 132. 20. Harvey Starr, “ ‘Opportunity’ and ‘Willingness’ as Ordering Concepts in the Study of War,” International Interactions 4 (1978): 363–87. On this point see Lemke and Kugler, “The Evolution of the Power Transition Perspective,” in Parity and War, 3–33; see 21. 21. Kugler and Organski, “The Power Transition,” 186. 22. Note also that Organski’s theory suggests that war is rank related; states with higher power rankings are presumed to be more highly involved in war. On this point see Henk Houweling and Jan Siccama, “Power Transitions as a Cause of War,” Journal of Conflict Resolution 32, no. 1 (March 1988): 87–102. 23. Woosang Kim’s measure of satisfaction with the status quo is the degree of difference between alliance portfolios of each state with that of the dominant state in the system: the more similar the alliance links, the more satisfied the state. In several studies, which encompass the period from 1648 to 1975, Kim finds that dissatisfaction has a significant impact on the probability of major war. Lemke and Werner use extraordinary arms buildups by a challenger that exceed those of the dominant state during periods of parity as evidence of commitment to changing the status quo (a concept they prefer to “dissatisfaction”). They analyze relationships between dominant states and their challengers in both the global hierarchy and in South American regional hierarchies. Their findings indicate that the combination of parity and commitment to change greatly increase the probability of war for both major powers and for minor powers. Woosang Kim, “Alliance Transitions and Great Power War,” American Journal of Political Science 35 (1991): 833–50; Kim, “Power Transitions and Great Power War from 1648 to 1975,” in Parity and War, 93–105; and Kim and J. Morrow, “When Do Power Shifts Lead to War?” American Journal of Political Science 36 (1992): 896–922. Douglas Lemke and Suzanne Werner, “Power Parity, Commitment to Change, and War,” International Studies Quarterly 40 (1996): 235–60. Siverson and Miller note that using alliance portfolios to measure satisfaction makes the United States a dissatisfied state between 1900 and World War II since it had no alliance ties to any of the European great powers. Siverson and Miller, “The Power Transition: Problems and Prospects.” 24. See William R. Thompson, “Succession Crises in the Global Political System: A Test of the Transition Model,” in Crises in the World-System, ed. A. L. Bergeson (Beverly Hills, CA: Sage, 1983), 93–116. 25. Dale Copeland, The Origins of Major War (Ithaca, NY: Cornell University Press, 2000), 13. 26. Organski and Kugler, The War Ledger, 59–61. Presumably, the still-rising challenger can afford to wait to see if some of the dominant state’s allies can be split away from the pro–status quo coalition. Once parity with the dominant state is actually achieved, this consideration becomes less important. 27. Tammen et al., Power Transitions. 28. Kugler and Organski, “The Power Transition”; J. Kugler and F. Zagare, “The Long Term Stability of Deterrence,” International Interactions 15 (1990): 255–78. See also Jonathan DiCicco and Jack Levy, “Power Shifts and Problem Shifts: The Evolution of the Power Transition Research Program,” Journal of Conflict Resolution 43, no. 6 (1999): 675–704. Tammen et al., Power Transitions, 30–31, agree with this position—at least with regard to world wars. However, they argue that the probability of war actually increases prior to the overtaking, but that wars that occur at this early point are generally smaller in size. Conversely, the probability of war decreases after the point of overtaking, but wars that break out at this later point are much more severe, such as the two world wars of the twentieth century. Tammen et al., Power Transitions, 28–29. 29. Tammen et al., Power Transitions, 12. 30. Jacek Kugler and Frank Zagare, “The Long-Term Stability of Deterrence,” International Interactions 15 (1990): 255–78. This discussion of tendencies toward risk is essentially a response to a logical criticism of the theory

558

Notes

by Bruce Bueno de Mesquita and others that it would be irrational for a rising but still weaker challenger to initiate war against a stronger dominant state. Such a war would only make sense if the challenger was risk acceptant. 31. Morrow, “The Logic of Overtaking,” 314–321. 32. Morrow, “The Logic of Overtaking,” 327. See also Kim and Morrow, “When Do Power Shifts Lead to War?” Morrow explicitly refers to aversion to high war costs as a condition that poses a “deep challenge” to power transition theory, since by definition power transition wars are costly wars. 33. It is interesting that in Stuart Bremer’s study of the war capabilities of states, he identifies the number two ranked country in the system as the one with the highest relative frequency of war initiation. In power transition theory, as originally formulated, the rising challenger would essentially be the number two ranked state in terms of power capabilities. Stuart Bremer, “National Capabilities and War Proneness,” in The Correlates of War II: Testing Some Realpolitik Models, ed. J. David Singer (New York: Free Press, 1980), 57–82. 34. See DiCicco and Levy, “Power Shifts and Problem Shifts.” 35. Daniel Geller, “Power Transition and Conflict Initiation,” Conflict Management and Peace Science 12 (1992): 1–16. 36. Daniel Geller, “Relative Power, Rationality, and International Conflict,” in Parity and War, 127–44. 37. Lemke and Kugler, “The Evolution of the Power Transition Perspective,” 18. 38. Organski admits that balance of power theory may have had some validity in the preindustrial era, but that in the twentieth century, when rapid industrialization fueled rapid shifts in power, balances are too unstable to keep the peace. More recent statements by power transition theorists muddies the water somewhat regarding Organski’s original hypothesis about the speed of the transition. Tammen et al., Power Transitions, 30–31, say “Faster overtakings appear to lead to a lower probability of war and slower overtakings increase the probability of war . . . longer, slower transitions temporarily postpone conflict but may eventually result in the most severe wars. Shorter, faster transitions increase the likelihood but lower the severity of war . . . With fast overtakings, the two countries are in parity for a short period of time and there is relatively less opportunity for policy disputes to develop into war. But with slower transitions, those policy disputes have time to develop irreversible momentum leading to war . . . The longer nations are in relative parity, the higher the probability of war and the longer the duration of any war they may fight.” 39. Morrow, “The Logic of Overtaking,” 327. 40. It may actually describe an even more limited set of cases since its emphasis on industrialization limits it to the period after the mid-nineteenth century and since Organski and Kugler’s list of appropriate wars seems confined almost entirely to the wars of attempted German expansion. On this point see Randolph Siverson and Michael Sullivan, “The Distribution of Power and the Onset of War,” Journal of Conflict Resolution 27, no. 3(1983): 473–94, and Randolph Siverson and Ross Miller, “The Power Transition: Problems and Prospects,” in Parity and War, 57–73. 41. On this point see Lemke and Kugler, “The Evolution of the Power Transition Perspective,” 11. 42. Organski and Kugler, The War Ledger, 49–61; see especially table 1.7 on p. 52. 43. Lemke and Kugler, “The Evolution of the Power Transition Perspective,” 4. 44. John Vasquez, “When Are Power Transitions Dangerous? An Appraisal and Reformulation of Power Transition Theory,” in Parity and War, 35–56. Vasquez (41) notes that there is a German/British power transition only because the U.S. is not included. Germany is deemed to catch up with Britain in GNP in 1905 and to overtake her in 1913, and—after WWI—again early in the 1920s. 45. Houweling and Siccama, “Power Transitions as a Cause of War.” See also Houweling and Siccama, “A TwoLevel Explanation of World War,” in Kugler and Lemke, Parity and War, 107–16. Houweling and Siccama’s study has, of course, come in for some criticism of its own. Decisions about when to include countries like the U.S. and Japan as great powers has an effect on the relative power positions of the rest of the countries in the analysis. Siverson and Miller determine that by adding the U.S. as a great power in 1900, with 22 percent of the total system power (more than any other single state), singlehandedly produces approximately seven separate power transitions, two of which ultimately lead to war. Randall Siverson and Ross Miller, “The Power Transition: Problems and Prospects,” in Parity and War, 57–73. Thompson notes that the use of capability indexes that include “bulk indicators” such as population size and armed force size, tend to favor large land powers. According to this index, when China is included as a major power in 1950, it automatically passes Britain, France, Germany, the Soviet Union, and Japan; and China passes the U.S. in 1960. The Soviet Union also passes the U.S. in 1970. He concludes that if these reputed power transitions lack face validity, then the statistical findings based on them cannot be taken at full value. Thompson, “Balances of Power, Transitions, and Long Cycles,” 170–71.

Notes

559

46. Richard J. Stoll and Michael Champion, “Capability Concentration, Alliance Bonding, and Conflict Among the Great Powers,” in Polarity and War, ed. A. N. Sabrosky (Boulder, CO: Westview, 1985). 47. Indra de Soysa, John Oneal, and Yong-Hee Park, “Testing Power-Transition Theory Using Alternative Measures of National Capabilities,” Journal of Conflict Resolution 41, no. 4 (1997): 509–28. 48. Charles Gochman, “Capability-Driven Disputes,” in Prisoners of War? ed. C. Gochman and A. N. Sabrosky (Lexington, MA: Lexington Books, 1990), 141–59. 49. Gochman, “Capability-Driven Disputes,” 157. 50. Daniel Geller, “Power Differentials and War in Rival Dyads,” International Studies Quarterly 37 (1993): 173–93. 51. Lemke and Kugler, “The Evolution of the Power Transition Perspective,” 13, 19. The references are to Bruce Bueno de Mesquita, “Pride of Place: The Origins of German Hegemony,” World Politics 43 (1990): 28–52; J. Kugler and M. Arbetman, “The Phoenix Factor Revisited,” paper presented to International Studies Association, 1989; and Organski and Kugler, The War Ledger, chapter 2. 52. Douglas Lemke and William Reed, “Regime Type and Status Quo Evaluations: Power Transition Theory and the Democratic Peace,” International Interactions 22, no. 2 (1996): 143–64. 53. Douglas Lemke, “Small States and War: An Expansion of Power Transition Theory,” in Parity and War, 77–91. See also Tammen et al., Power Transitions, chapter 3. 54. The members of these regional hierarchies are a small number of states whose politically relevant neighborhoods overlap and who can plausibly rate each other as potential security threats. Their membership is essentially limited by Boulding’s concept of the loss-of-strength gradient; their ability to exert military power over each other diminishes with distance. Lemke constructs metrics to include states that can plausibly engage with other members of the region militarily and exclude those who cannot. Lemke, “Small States and Wars,” 83–86, and Tammen et al., Power Transitions, 66–68. 55. Tammen et al., Power Transitions, 8. 56. Tammen et al., Power Transitions, chapter 3. 57. By their figures, joint parity and dissatisfaction increases the dyadic probability of war from a baseline of 7.9 percent to 48.6 percent for great powers, but for the highest regional hierarchy, the Far East, the probability increases from the baseline of 2.2 percent to 19.4 percent. Tammen et al., Power Transitions, 72, figure 3.2. 58. Frank Waylon Wayman, “Power Shifts and the Onset of War,” in Parity and War, 145–163. 59. Wayman, “Power Shifts and the Onset of War,” 147. 60. Organski, World Politics, 1968, 364–367, and Kugler and Organski, “The Power Transition: A Retrospective and Prospective Evaluation,” 175, hint at possible dominant state motivations for war initiation, but do not follow up. 61. Wayman, “Power Shifts and the Onset of War,” 156, table 9.1. 62. Vasquez, “When Are Power Transitions Dangerous?” 49. 63. Wayman, “Power Shifts and the Onset of War,” 158, table 9.3. 64. Paul Huth and Bruce Russett, “General Deterrence between Enduring Rivals: Testing Three Competing Models,” American Political Science Review 87, no. 1 (1993): 61–73. 65. Peter Wallensteen, “Incompatibility, Confrontation, and War: Four Models and Three Historical Systems, 1816–1976,” Journal of Peace Research 18, no. 1 (1981): 57–90. 66. Woosang Kim, “Power, Alliance, and Major Wars, 1816–1975,” Journal of Conflict Resolution 32, no. 2 (June 1989): 255–73. 67. Kim, “Power Transitions and Great Power War from Westphalia to Waterloo,” and Kim and Morrow, “When Do Power Shifts Lead to War?” 68. D. Scott Bennett and Allan C. Stam, The Behavioral Origins of War (Ann Arbor: University of Michigan Press, 2004), 115, 137–41. 69. See Lemke and Kugler, “The Evolution of the Power Transition Perspective,” 12. Emphasis in the original. The problem is that reducing power transition theory to a theory of power parity, simply reduces it to an argument that states go to war when they have an opportunity to do so—which must then devolve into attempts to explain foreign policy decision-making under conditions of opportunity. Thus, systemic (or dyadic) structures are merely a starting point for what just necessarily become a decision-making theory found at the individual, state, or substate level of analysis. 70. Woosang Kim, “Power, Alliance, and Major Wars, 1816–1975”; Kim, “Alliance Transitions and Great Power War”; Kim, “Power Transitions and Great Power War from Westphalia to Waterloo”; Kim, “Power Parity, Alliance, and War from 1648–1975”; and Kim and James Morrow, “When Do Power Shifts Lead to War?”

560

Notes

71. Vasquez, “When Are Power Transitions Dangerous?” 47. See also John Vasquez, The War Puzzle (Cambridge: Cambridge University Press, 1993). 72. John Vasquez, “When Are Power Transitions Dangerous?” 50. Vasquez cites evidence from Wallensteen, “Incompatibility, Confrontation and War”; Geller, “Relative Power, Rationality and International Conflict”; and a personal communication from Wayman. 73. One complication here is that the traditional balance of power theory argues that an important aspect of inequality of power is found in alliance bonds. The dyadic balance is not the only consideration of states contemplating the use of force; they will seek to assess how alliance commitments affect the balance. In other words, peace is kept not by the dyadic balance alone, but by the balance between alliances. Therefore, looking solely at the dyadic balance may be misleading. Two studies give some support to this thesis, though their results produce as many disagreements as agreements: Randolph Siverson and Michael Tennefoss, “Power, Alliance, and the Escalation of International Conflict, 1815–1965,” American Political Science Review 78 (1984): 1057–69; and Woosang Kim, “Power, Alliance, and Major Wars, 1816–1975.” 74. Daniel Geller, “Capability Concentration, Power Transition and War,” International Interactions 17 (1992): 269–84; Geller, “Relative Power, Rationality, and International Conflict,” in Parity and War, 127–143. 75. Henk Houweling and Jan Siccama, “A Two-Level Explanation of World War,” in Parity and War, 107–116; Houweling and Siccama, “Power Transitions and Critical Points as Predictors of Great Power War,” Journal of Conflict Resolution 35 (1992): 642–58. 76. Dale Copeland, The Origins of Major War (Ithaca, NY: Cornell University Press), 27. 77. Copeland, The Origins of Major War, 15–20, 50. 78. Copeland, The Origins of Major War, 20. 79. Copeland, The Origins of Major War, 15. 80. Copeland, The Origins of Major War, 22–23, 52. 81. Copeland, The Origins of Major War, 16–20, 23–27. In situations of bipolarity, it is possible that a power transition has actually occurred prior to the outbreak of war; the challenger may have achieved parity. 82. Copeland, The Origins of Major War, 42–46. The fifth pathway, which is not consistent with the theory, is that states simply initiate major war due to “unit level” motives such as greed and glory. Copeland says that such unit level factor may play “an important and independent role in causing major war” (44). 83. Robert Gilpin, War and Change in World Politics (Cambridge: Cambridge University Press, 1981). See also Gilpin, “The Theory of Hegemonic War,” in The Origin and Prevention of Major Wars, ed. Robert Rotberg and Theodore Rabb (Cambridge: Cambridge University Press, 1988), 15–37. 84. Other theorists of hegemonic stability emphasize the construction and maintenance of a liberal, free-trading regime by the hegemon, which contributes to peace and stability in the system. See, for instance, Stephen Krasner, “State Power and the Structure of International Trade,” World Politics 28 (April 1976): 317–47; Robert Keohane, “The Theory of Hegemonic Stability and Changes in International Economic Regimes, 1967–77,” in Change in the International System, ed. Ole Holsti, Randolph Siverson, and Alexander George (Boulder, CO: Westview, 1980), 131–62; and Robert Keohane and Joseph Nye, Power and Interdependence (Boston: Little, Brown, 1977). Gilpin mentions this, but prefers not to emphasize it; he argues that when Britain was unable to contain the imperial ambitions of its continental rivals owing to its declining hegemonic position, it began to practice a kind of “preclusive” imperialism to minimize losses to its rivals. Thus, hegemonic powers do not always practice free trade. 85. Gilpin, War and Change in World Politics, xi–xii. 86. Gilpin, War and Change in World Politics, 197. 87. Gilpin, War and Change in World Politics, 199–202. 88. Gilpin, War and Change in World Politics, 200. In his later “The Theory of Hegemonic War” (30–34), Gilpin only mentions three modern hegemonic wars, omitting the wars of Louis XIV. 89. Gilpin, War and Change in World Politics, 208–9. Other options are, of course, open to the hegemon as well, such as an increase in the resources dedicated to hegemony or a reduction in commitments. 90. Gilpin, War and Change in World Politics, 92–93. 91. On this point see Robert North, War, Peace, Survival (Boulder, CO: Westview, 1990), 222. 92. Gilpin, War and Change in World Politics, 31. Like Blainey, Gilpin argues that the more decisive the initial military victory by the hegemon, the more clear-cut the postwar hierarchy, and therefore, the less likely war is to occur. 93. For an assessment of historical-structural perspectives and their view of the future, see Jack Levy, “Long Cycles, Hegemonic Transitions and the Long Peace,” in The Long Postwar Peace: Contending Explanations and Projections, ed. Charles W. Kegley (New York: HarperCollins), 1, 147–76.

Notes

561

94. K. Edward Spiezio, “British Hegemony and Major Power War, 1815–1939: An Empirical Test of Gilpin’s Model of Hegemonic Governance,” International Studies Quarterly 34, no. 2 (June 1990): 165–81. 95. Although some historical-structural theorists see Portugal and the United Provinces as hegemonic states (Modelski, for instance) or at least the United Provinces (Wallerstein), Gilpin prefers not to classify either as having achieved hegemonic status. The 1648–1815 period was classified by the European balance of power rather than by cycles of hegemonic succession, according to Gilpin, and the premodern age was characterized by a succession of empires (116). 96. Terry Boswell and Mike Sweat, “Hegemony, Long Waves, and Major Wars: A Time Series Analysis of Systemic Dynamics, 1496–1967,” International Studies Quarterly 35 (1991): 123–49. 97. Brian Pollins, “Global Political Order, Economic Change, and Armed Conflict: Coevolving Systems and the Use of Force,” American Political Science Review 90, no. 1 (March 1996): 103–117. 98. Bennett and Stam, The Behavioral Origins of War, 145–47. 99. Bennett and Stam, The Behavioral Origins of War, 147–50. 100. Robert Jervis, “Unipolarity: A Structural Perspective,” World Politics 61, no. 1 (2009): 188–213; see 193. 101. George Modelski, “The Long Cycle of Global Politics and the Nation-State,” Comparative Studies in Society and History 20, no. 2 (April 1978): 214–35; and William R. Thompson, On Global War: Historical-Structural Approaches to World Politics (Columbia: University of South Carolina, 1988). The long-cycle perspective does not claim to be an extensive theory; it is still evolving as its proponents continue to refine it. Thompson prefers to call it a “perspective” or an “analytical framework.” William R. Thompson, “Uneven Economic Growth, Systemic Challenges, and Global Wars,” International Studies Quarterly 27 (1983): 341–55. 102. Thompson, On Global War, 45, 118; Modelski, “The Long Cycle of Global Politics and the Nation-State,” 215; William Thompson, “Balances of Power, Transitions, and Long Cycles,” in Parity and War, 163–85. 103. William R. Thompson, “Polarity, the Long Cycle, and Global Power Warfare,” Journal of Conflict Resolution 30, no. 4 (December 1986): 587–615. 104. Karen Rasler and William R. Thompson, The Great Powers and Global Struggle 1490–1990 (Lexington: University of Kentucky Press, 1994), 7–8. 105. While the cycle of world leadership connected to global wars can be traced to 1494 (the Italian and Indian Ocean Wars), the pattern of long-term economic growth driven by technological innovations that create lead economies (and the tendency toward concentration, deconcentration, and reconcentration of power) actually date back farther—to the tenth and eleventh centuries. Rasler and Thompson identify first Northern Sung China, then Southern Sung, then Genoa, then Venice as the lead economies. During this early stage, successive changes in the lead economies were associated with intense conflict, but not with global wars. After 1494 the global system and the European regional system became fused, leading to the cycle of global wars centering on Europe. Karen Rasler and William R. Thompson, “Global War and the Political Economy of Structural Change,” in Handbook of War Studies II, ed. Manus Midlarsky (Ann Arbor: University of Michigan Press, 2000), 301–331. 106. Karen Rasler and William R. Thompson, “Global Wars, Public Debts, and the Long Cycle,” World Politics 35, no. 4 (July 1983): 489–516. See also Paul Kennedy, The Rise and Fall of Great Powers: Economic Change and Military Conflict from 1500 to 2000 (New York: Random House, 1988). 107. Thompson, On Global War, 49–50. Rasler and Thompson, The Great Powers and Global Struggle 1490–1990, 159. While others emphasize the hegemon’s tendency to become overburdened with protection costs and other consumptions costs, compared with investment (Gilpin) or to become globally overcommitted (Paul Kennedy), long-cycle theorists consider overextension to be the result of relative decline rather than its reason. Rasler and Thompson, “Global War and the Political Economy of Structural Change,” and Rasler and Thompson, The Great Powers and Global Struggle 1490–1990, 98–156. 108. George Modelski and Patrick Morgan, “Understanding Global War,” Journal of Conflict Resolution 29, no. 3 (September 1985): 399. 109. On this point see William R. Thompson, “Succession Crises in the Global Political System: A Test of the Transition Model,” in Crises in the World-System, ed. Albert Bergesen (Beverly Hills, CA: Sage, 1983), 109. 110. See George Modelski and William R. Thompson, “Long Cycles and Global War,” in Handbook of War Studies, ed. Manus Midlarsky (Boston: Unwin Hyman, 1989), 39; also Thompson, “Succession Crises in the Global Political System.” 111. Thompson, “Uneven Economic Growth, Systemic Challenges, and Global Wars,” 349. 112. See Levy’s comments on this point: Jack Levy, “Theories of General War,” World Politics 37, no. 3 (1985): 361–63.

562

Notes

113. Modelski and Morgan, “Understanding Global War,” 400–401, and Rasler and Thompson, “Global Wars, Public Debts, and the Long Cycle,” 500. 114. Thompson, “Uneven Economic Growth, Systemic Challenges, and Global Wars,” 349–51. 115. On this last point see William Thompson, “Systemic Leadership, Evolutionary Processes, and International Relations Theory: The Unipolarity Question,” International Studies Review 8 (2006): 1–22. 116. Thompson, On Global War. On this point see also William R. Thompson and Karen Rasler, “War and Systemic Capability Reconcentration,” Journal of Conflict Resolution 32 (June 1988): 335–66. Global wars also tend to facilitate an increase in the new world power’s leading economic sectors. 117. This, of course, conflicts with Singer, Bremer, and Stuckey’s finding that in the nineteenth century there is a positive correlation between capability concentration and ongoing war. Thompson argues that the Correlates of War findings are dependent on the classification of states as great powers, the selection of the 1820 starting point, the use of a capability index giving equal weight to demographic, military, and industrial measures, and the use of nationmonths as the measure of war. William R. Thompson (ed.), Contending Approaches of World System Analysis (Beverly Hills, CA: Sage, 1983). 118. Thompson, “Polarity, the Long Cycle, and Global Power Warfare,” and Thompson, On Global War, chapter 9. 119. Boswell and Sweat, “Hegemony, Long Waves, and Major Wars,” 146. 120. Thompson, “Succession Crises in the Global Political System,” 112. 121. These are summarized in Jack Levy, “Declining Power and the Preventive Motivation for War,” World Politics 40, no. 1 (October 1987): 85–88, and in Levy, “Long Cycles, Hegemonic Transitions and the Long Peace.” Levy draws on Organski, World Politics (1968), 376; Stephen Van Evera, “The Cult of the Offensive and the Origins of World War I,” International Security 9 (Summer 1984): 58–107; and Jack L. Snyder, “Perceptions of the Security Dilemma in 1914,” in Psychology and Deterrence, ed. Robert Jervis, Richard Ned Lebow, and Janice Gross Stein (Baltimore: Johns Hopkins Press, 1985), 153–79. 122. Brian Pollins, “Global Political Order, Economic Change, and Armed Conflict: Coevolving Systems and the Use of Force,” American Political Science Review 90, no. 1 (1996): 103–117. 123. Levy, “Long Cycles, Hegemonic Transitions, and the Long Peace,” 158. Long-cycle theorists also omit the Seven Years’ War (1755–1763) and the Dutch War of Louis XIV (1672–1678) from their list of global wars. 124. Levy, “Long Cycles, Hegemonic Transitions, and the Long Peace,” 159. 125. Rasler and Thompson, The Great Powers and Global Struggle, 1490–1990, and Rasler and Thompson, “Global War and the Political Economy of Structural Change.” Their framework is based on the earlier work of Ludwig Dehio, The Precarious Balance (New York: Vantage, 1962). 126. Thompson, “Balances of Power, Transitions, and Long Cycles,” 174. 127. Rasler and Thompson, The Great Powers and Global Struggle, 1490–1990, 42, 158. 128. Rasler and Thompson, The Great Powers and Global Struggle, 1490–1990, 12. 129. On this point see William R. Thompson, “Balances of Power, Transitions, and Long Cycles,” in Parity and War, 163–85. 130. Moreover, Rasler and Thompson argue that animosity between global powers and regional powers appears to be generated by the fact that they are organized according to different political principles and pursue dissimilar strategies. While regional land powers tend to be autocracies with large armies and large bureaucracies and expansionist foreign policies, global sea powers tend to have more representative and less autocratic polities and tend to be more oriented toward long-distance trading than toward territorial expansion. Rasler and Thompson, “Global War and the Political Economy of Structural Change.” 131. To be classified as a global power, a state must have a minimum of at least 10 percent of the global pool of naval capabilities and have a navy that operates beyond its local region. To qualify as a world power, a state must be on the winning side in a global war and possess a minimum of 50 percent of the global pool of naval capabilities. Ottoman Turkey and Russia/USSR are not included in the membership of European regional powers (and thus not added to the pool of regional land army capabilities). They are treated as external, Eastern balancers. Rasler and Thompson, The Great Powers and Global Struggle, 1490–1990, 17, 29. 132. Rasler and Thompson, The Great Powers and Global Struggle, 1490–1990, 69. 133. Rasler and Thompson, The Great Powers and Global Struggle, 1490–1990, 38–57, table 9.1 on 161. See also Thompson, “Balances of Power, Transitions, and Long Cycles.” 134. Rasler and Thompson, The Great Powers and Global Struggle, 1490–1990, chapter 3, and Thompson, “Balances of Power, Transitions, and Long Cycles.” Since there have only been three regional leaders in Europe during the time period, the cases are limited to France overtaking Spain and Germany overtaking France.

Notes

563

135. Rasler and Thompson, The Great Powers and Global Struggle, 1490–1990, 38–57. 136. William R. Thompson, “Systemic Leadership, Evolutionary Processes, and International Relations Theory,” International Studies Review 8 (2006): 1–22; see 16. In fact, Levy and Thompson find that historically, the leading sea power is less likely to provoke a counter-balancing coalition than the leading land power. Jack Levy and William Thompson, “Balancing at Sea: Do States Coalesce against the Leading Sea Powers?” paper presented at the annual conference of the American Political Science Association, Philadelphia, August, 2003. 137. Stephen Walt, The Origins of Alliances (Ithaca. NY: Cornell University Press, 1987). Quotes are from Levy and Thompson, “Balancing on Land and at Sea,” 38. 138. Jack S. Levy and William R. Thompson “Hegemonic Threats and Great-Power Balancing in Europe, 1495–1999,” Security Studies 14, no. 1 (January-March, 2005): 1–33, deals with balancing against land powers; Jack Levy and William Thompson, “Balancing on Land and at Sea: Do States Ally against the Leading Global Power?” International Security 35, no. 1 (Summer 2010): 7–43, deals with balancing against global maritime leaders. This summary deals with information reported in the latter study. 139. Levy and Thompson, “Balancing on Land and at Sea,” 33. 140. Levy and Thompson, “Balancing on Land and at Sea,” 34. 141. Thompson, On Global War, 277–80. 142. Modelski and Thompson, “Long Cycles and Global War,” 50. 143. Modelski and Morgan, “Understanding Global War,” 403; Modelski and Thompson, “Long Cycles and Global War,” 42. 144. Wallerstein’s major works include The Modern World-System (New York: Academic Press, 1974); The Capitalist World-Economy (New York: Cambridge University Press, 1979); and The Modern World System II: Mercantilism and the Coordination and the Consolidation of the European World Economy, 1600–1750 (New York: Free Press, 1980). For a short summary of his ideas see his Historical Capitalism (London: Verso, 1983). 145. Christopher Chase-Dunn, “Interstate System and Capitalist World-Economy: One Logic or Two?” International Studies Quarterly 25, no. 1 (March 1981): 27, 31. 146. World Wars are defined as a conflict in which one state seeks to take over, and thus destroy, the (multicentric) interstate system or as a struggle that determines leading power status. Christopher Chase-Dunn, Global Formation: Structure of the World-Economy (Cambridge, MA: Basil Blackwell, 1989), 159. 147. Wallerstein, Historical Capitalism, 62. 148. Wallerstein, Historical Capitalism, 64. 149. Chase-Dunn, “Interstate System and Capitalist World,” 23. 150. Christopher Chase-Dunn and Joan Sokolovsky, “Interstate System, World-Empires and the Capitalist WorldEconomy: A Response to Thompson,” International Studies Quarterly 27 (September 1983): 364. 151. Research Group on Cyclical Rhythms and Secular Trends, “Cyclical Rhythms and Secular Trends of the Capitalist World-Economy,” Review 2 (Spring 1979): 483–500, cited in Thompson, On Global War, 72. The dating of the phases is based on economic dominance in production, commerce, and finance. 152. Chase-Dunn, Global Formation, 183–84. 153. Chase-Dunn and Sokolovsky, “Interstate System, World-Empires and the Capitalist World-Economy: A Reply to Thompson,” 361. 154. Chase-Dunn, “Interstate System and Capitalist World Economy,” 38–40. 155. Wallerstein, Historical Capitalism, 58. 156. Chase-Dunn, Global Formation, 84–85, 343–45. 157. See the analyses in Thompson, On Global War, 167–195; and Rasler and Thompson, The Great Powers and Global Struggle, 1490–1990, 171. 158. Pollins, “Global Political Order, Economic Change, and Armed Conflict.” 159. Boswell and Sweat, “Hegemony, Long Waves, and Major Wars.” 160. North, War, Peace, Survival, 222–23. 161. N. D. Kondratieff, The Long Wave Cycle (New York: Richardson and Snyder, 1984; original edition, 1928). 162. Modelski, “The Long Cycle of Global Politics and the Nation-State,” 227–30; Modelski and Morgan, “Understanding Global War,” 402; and Rasler and Thompson, “Global Wars, Public Debts and the Long Cycle.” 163. William R. Thompson and Gary Zuk, “War, Inflation, and the Kondratieff Long Wave,” Journal of Conflict Resolution 26, no. 4 (December 1982): 621–44. 164. Thompson, On Global War, 53–54. 165. Edward Mansfield, “Distribution of Wars over Time,” World Politics 41 (1988): 21–51.

564

Notes

166. Joshua S. Goldstein, Long Cycles: Prosperity and War in the Modern Age (New Haven, CT: Yale University Press, 1988). 167. Joshua S. Goldstein, “Kondratieff Waves as War Cycles,” International Studies Quarterly 29, no. 4 (December 1985): 411–44; Goldstein, “Long Waves in War, Production, Prices, and Wages,” Journal of Conflict Resolution 31, no. 4 (December 1987): 573–600; and Goldstein, Long Cycles. In a later work, Goldstein identifies four longwave phases: expansion, war, stagnation, and rebirth. Joshua Goldstein, “A War-Economy Theory of the Long Wave,” in Business Cycles: Theories, Evidence, and Analysis, ed. N. Thygeses, K. Velupillai, and S. Zambell (London: Macmillan, 1991). 168. Boswell and Sweat, “Hegemony, Long Waves, and Major Wars,” 134–35. 169. Levy notes, however, that since K-waves are systemic phenomena, all great powers should benefit from the upswings and thus the balance of power should remain generally unchanged. Levy, “Long Cycles, Hegemonic Transitions, and the Long Peace,” 165. 170. Goldstein, Long Cycles, chapter 15 and 350–57. 171. Pollins, “Global Political Order, Economic Change, and Armed Conflict.” 172. Bennett and Stam, The Behavioral Origins of War, 144–45. 173. Levy, “Long Cycles, Hegemonic Transitions, and the Long Peace.” 174. World systems theorists see the structure of the modern world system as characterized by three constants, three cycles, and four secular trends. The three constants are (1) commodity production, (2) the core-periphery division of labor, and (3) the state system with relatively strong core states and relatively weak periphery states. The three cycles are (1) long waves (K-waves) of increases and decreases in the rate of capital accumulation and overall economic activity in the system, (2) a unicentric-multicentric cycle of competition among core states, and (3) swings from free trade to controlled trade in core-periphery economic relations. The four secular trends are (1) the expansion of the system over time, (2) the intensification and deepening of commodity relations, (3) the increasing power and control of states, and (4) the increasing size of economic enterprises. Christopher Chase-Dunn, “Comparative Research on World-System Characteristics,” International Studies Quarterly 23, no. 4 (December 1979): 607–8. 175. Goldstein, Long Cycles, 287–88. 176. Chase-Dunn, Global Formation, 164. Wallerstein’s view is slightly different. He argues that core states step up their exploitation of peripheral areas during the contraction phase as a way of minimizing the effects of the stagnation of the world economy. This is done primarily through attempts to seek out lower-wage workforces and incorporate them into the world-economy rather than through a search for new markets. Wallerstein, Historical Capitalism, 39. 177. Though Doran’s writings on this subject are numerous, see especially Charles F. Doran, “War and Power Dynamics: Economic Underpinnings,” International Studies Quarterly 27 (1983): 419–44; Doran, “Systemic Disequilibrium, Foreign Policy Role, and the Power Cycle: Challenges for Research Design,” Journal of Conflict Resolution 33, no. 3 (September 1989): 371–401; Doran, “Power Cycle Theory of Systems Structure and Stability: Commonalities and Complementarities,” in Midlarsky, Handbook of War Studies, 83–110; Doran, “Confronting the Principles of the Power Cycle: Changing Systems Structure, Expectations, and War,” in Midlarsky, Handbook of War Studies II, 332–368; and Doran and Wes Parsons, “War and the Cycle of Relative Power,” American Political Science Review 74 (1980): 947–65. 178. Doran, “Power Cycle Theory of Systems Structure,” 88. 179. Doran, “War and Power Dynamics: Economic Underpinnings,” 420; Doran and Parsons, “War and the Cycle of Relative Power,” 952. 180. Doran, “War and Power Dynamics: Economic Underpinnings,” 421–22, and Doran, “Power Cycle Theory of Systems Structure and Stability,” 104. 181. For a good summary of the issue of roles and the gap between role and power see Doran, “Confronting the Principles of the Power Cycle,” 338–40. As Doran notes (338), foreign policy role involves the concerns and ends of statecraft. It encompasses a state’s interests, position (place) in the system and its status. Role is a foreign policy behavior that the system has allowed a state to achieve. 182. Doran and Parsons, “War and the Cycle of Relative Power,” 949–50. 183. Doran, “War and Power Dynamics: Economic Underpinnings,” 422–26, and Doran, “Power Cycle Theory of Systems Structure and Stability,” 89. 184. Doran, “Power Cycle Theory of Systems Structure and Stability,” 90. 185. Doran, “Power Cycle Theory of Systems Structure and Stability,” 84. The passage of a major state through a critical point can also increase the importance of other, more longstanding systemic structures and processes such as

Notes

565

status discrepancy, hegemonic decline, or power concentration/deconcentration. Doran, “Power Cycle Theory of Systems Structure and Stability,” 90. 186. Doran, “Power Cycle Theory of Systems Structure and Stability,” 92. See also Doran, System in Crisis, 133, table 5.1. 187. Doran, “War and Power Dynamics: Economic Underpinnings,” 429–30. 188. Though leaders may correctly perceive critical points, there is probably a lag between the actual arrival at a critical point and the recognition of this fact, and the recognition takes place in spite of an array of phenomena that might prevent leaders from actually recognizing the fact. For possible reasons why leaders might not perceive that their states have arrived at a critical point see Greg Cashman, “Theoretical Buttresses for Power Cycle Theory,” paper presented to International Political Science Association XVIIth World Congress, Seoul, Korea, August 17–21, 1997. 189. Doran, Systems in Crisis, 122–23, 134. 190. See Doran, Systems in Crisis, chapter 1, for a detailed examination of the effects of critical points on decision makers. 191. Doran and Parsons, “War and the Cycle of Relative Power,” 460–62. 192. Doran, “War and Power Dynamics: Economic Underpinnings,” 430. 193. Charles Doran, Systems in Crisis (Cambridge: Cambridge University Press, 1991). The continued recalculation of critical points has led to some inconsistency. Doran and Parsons, “War and the Cycle of Relative Power” (1980), report thirteen critical points; Doran, “Systemic Disequilibrium, Foreign Policy Role, and the Power Cycle” (1989) reports ten critical points; and Doran, System in Crisis (1991) reports fourteen. As Hebron and James note, only six of the critical points are identical in the three studies. Lui Hebron and Patrick James, “Great Powers, Cycles of Relative Capability and Crises in World Politics,” International Interactions 23, no. 2 (1997): 145–173; see 151–52. 194. Doran, “Confronting the Principles of the Power Cycle,” 359. 195. Doran, “Confronting the Principles of the Power Cycle,” 350. 196. Doran, “System Disequilibrium, Foreign Policy Role, and the Power Cycle,” 384–87. For instance, there is a power transition preceding World War I, as Germany passes Britain in absolute capabilities. But the causal significance of this dyadic transition is essentially overwhelmed by the more important structural shift in which Germany’s power peaks relative to its system rivals in the aggregate. In absolute terms her power continues to grow, but other states in the system are now growing at a faster rate. Thus, war is not the result of expectations about absolute trends in dyadic power (Germany passing Britain); rather war is the result of unexpected changes in the nonlinear cycle of relative power involving all system members. See Doran, “Confronting the Principles of the Power Cycle,” 363–64. 197. Henk W. Houweling and Jan G. Siccama, “Power Transitions and Critical Points as Predictors of Great Power War,” Journal of Conflict Resolution 35, no. 4 (1991): 642–58. 198. Houweling and Siccama, “Power Transitions and Critical Points as Predictors of Great Power War,” 655–56. 199. Andrew Parasiliti, “The Causes and Timing of Iraq’s Wars: A Power Cycle Assessment,” International Political Science Review 24, no. 1 (2003): 151–65. 200. Sushil Kumar, “Power Cycle Analysis of India, China, and Pakistan in Regional and Global Politics,” International Political Science Review 24, no. 1 (2003): 113–22. No actual data about critical points and their dates are included in this article. 201. Hebron and James, “Great Powers, Cycles of Relative Capability and Crises in World Politics.” 202. Lui Hebron and Patrick James, “Research Note: Power Cycles and Interstate Disputes,” paper presented at International Studies Association Convention, March 18–22, 1997, Toronto, Canada. 203. Daniel Y. Chiu, “International Alliances in the Power Cycle Theory of State Behavior,” International Political Science Review 24, no. 1 (2003): 123–36. 204. Brock F. Tessman and Steve Chan, “Power Cycles, Risk Propensity, and Great-Power Deterrence,” Journal of Conflict Resolution 48, no. 2 (2004): 131–53. 205. Daniel Geller, “Toward a Unified Theory of War,” paper delivered at International Studies Association conference, Washington, DC, April 1990; Geller, “Capability Concentration, Power Transition and War,” International Interactions 17 (1992): 269–84; Geller “Relative Power, Rationality and International Conflict,” in Parity and War, 127–43. 206. G. John Ikenberry, Michael Mastanduno, and William Wohlforth, “Unipolarity, State Behavior, and Systemic Consequences,” World Politics 61, no. 1 (January 2009): 1–27, especially 22. 207. John Mueller, “Is War Still Becoming Obsolete?” paper presented at annual meeting of the American Political Science Association, Washington, DC, August 1991.

566

Notes

208. Charles Gochman, “Capability-Driven Disputes,” in Prisoners of War? Nation-States in the Modern Era, ed. Charles Gochman and A. N. Sabrosky (Lexington, MA: Lexington Books, 1990), 141–59. 209. G. J. Ikenberry, M. Mastanduno, and W. Wohlforth, “Unipolarity, State Behavior, and Systemic Consequences,” World Politics 61, no. 1 (2009): 1–27; see 6. Ikenberry et al. define unipolarity as a situation in which there is one state that (1) commands an especially large share of the system’s resources or capabilities and (2) excels in all components of state capability—size, population, territory, resource endowment, economic capacity, military might, and organizational competence; see 4. In other words, the unipole is unambiguously in a class by itself. 210. These statistics are from Ikenberry et al., “Unipolarity, State Behavior, and Systemic Consequences,” 6–10. 211. Thompson, “Systemic Leadership, Evolutionary Processes, and International Relations Theory,” 14. 212. Keir Lieber and Daryl Press, “The End of MAD? The Nuclear Dimension of U.S. Primacy,” International Security 30, no. 4 (Spring 2006): 7–44. 213. Jervis, “Unipolarity: A Structural Perspective,” 194. 214. See, for instance, David Wilkinson, “Unipolarity Without Hegemony,” International Studies Review 1, no. 2 (1999): 141–172. 215. See, for instance, Jervis, “Unipolarity: A Structural Perspective,” 201–204. 216. Thompson also makes this point. Whether the systemic leader is a land power that depends on territorial expansion or a sea power that relies on commercial and industrial expansion through access to markets should make a difference. William Thompson, “Systemic Leadership, Evolutionary Processes, and International Relations Theory.” 217. David Wilkinson, “Unipolarity without Hegemony,” 160. On China’s integration into various multilateral organizations in the region, see David Shambaugh, “China Engages Asia: Reshaping the Regional Order,” International Security 24, no. 3 (2004–2005): 64–99.

12. CONSTRUCTIVISM 1. Nicholas Onuf, World of Our Making: Rules and Rule in Social Theory and International Relations (Columbia: University of South Carolina Press, 1989). 2. Ronald L. Jepperson, Alexander Wendt, and Peter J. Katzenstein, “Norms, Identity, and Culture in National Security,” in The Culture of National Security, ed. Peter. J. Katzenstein (New York: Columbia University Press, 1996), 33–75; quote from 59. 3. Iver B. Neumann, “Identity and the Outbreak of War,” International Journal of Peace Studies 3, no. 1 (1998). 4. Alexander Wendt, “Anarchy Is What States Make of It: The Social Construction of Power Politics,” International Organization 46, no. 2 (1992): 391–425; quote from 398. 5. Emmanuel Adler, “Seizing the Middle Ground: Constructivism in World Politics,” European Journal of International Relations 3, no. 3 (1997): 319–63; quote from 337. 6. Ted Hopf, “The Promise of Constructivism in International Relations Theory,” International Security 23, no. 1 (1998): 171–200; quote from 175. 7. See Cynthia Weber, International Relations Theory: A Critical Introduction, 3rd ed. (New York: Routledge, 2010), 66–67. 8. J. G. March and J. P. Olsen, Rediscovering Institutions (New York: Free Press, 1989). 9. Alice Ba and Matthew Hoffmann, “Making and Remaking the World for IR 101: A Resource for Teaching Social Constructivism in Introductory Classes,” International Studies Perspectives 4 (2003): 15–33. 10. K. M. Fierke, “Constructivism,” in International Relations Theories: Discipline and Diversity, 2nd ed., ed. Tim Dunne, Milja Kurki, and Steve Smith (New York: Oxford University Press, 2010), 177–94; quote from 181. 11. Jeffrey Checkel, “The Constructivist Turn in International Relations Theory,” World Politics 50, no. 2 (1998): 324–48; see 326. 12. Wendt, “Anarchy Is What States Make of It,” 395. 13. Hopf, “The Promise of Constructivism,” 174. 14. Weber, International Relations Theory, 68. 15. Wendt, “Constructing International Politics,” International Security 20, no. 1 (1995): 71–81; quote from 77–78. 16. See for instance, Emmanuel Adler, “Imagined (Security) Communities: Cognitive Regions in International Relations,” Millennium: Journal of International Studies 26, no. 2 (1997): 249–77.

Notes

567

17. My thanks to Eric Rittenger for this insight. 18. Martha Finnemore and Kathryn Sikkink, “International Norm Dynamics and Political Change,” in International Politics: Classical and Contemporary Readings, ed. Scott Handler (Los Angeles: Sage, 2013), 108–16; see 109. 19. Wendt, “Constructing International Politics,” 71–73. 20. Hopf, “The Promise of Constructivism,” 177. 21. Finnemore and Sikkink, “International Norm Dynamics and Political Change,” 109. 22. See Iver B. Neumann, “Identity and the Outbreak of War,” International Journal of Peace Studies 3, no. 1 (1998): 7–22. 23. Paul D’Anieri, International Politics: Power and Purpose in Global Affairs, 2nd ed. (Boston: Wadsworth, 2012), 99. 24. Wendt, “Constructing International Politics,” 76. 25. Hopf, “The Promise of Constructivism,” 193. 26. Wendt, “Constructing International Politics,” 72 and n6. 27. Wendt, “Constructing International Politics,” 72. 28. Jeffrey Checkel, “The Constructivist Turn in International Relations Theory,” World Politics 50, no. 2 (1998): 324–48; see 341. 29. Theo Farrell, “Constructivist Security Studies: Portrait of a Research Program,” International Studies Review 4, no. 1 (2002): 49–72; quote from 60. 30. See Elizabeth Kier, “Culture and French Military Doctrine before World War II,” in The Culture of National Security, ed. Peter Katzenstein (New York: Columbia University Press, 1996), 186–215; Thomas U. Berger, “Norms, Identity, and National Security in Germany and Japan,” in Katzenstein, The Culture of National Security, 317–56. 31. Peter M. Haas, “Introduction: Epistemic Communities and International Policy Coordination,” International Organization 46, no. 1 (1992): 1–35. 32. Robert G. Herman, “Identity, Norms, and National Security: The Soviet Foreign Policy Revolution and the End of the Cold War,” in Katzenstein, The Culture of National Security, 271–316. See also Jeffrey Checkel, “Ideas, Institutions, and the Gorbachev Foreign Policy,” World Politics 45 (1993): 271–300; and Thomas Risse-Kappen, “Ideas Do Not Float Freely: Transnational Coalitions, Domestic Structures, and the End of the Cold War,” International Organization 48, no. 2 (Spring 1994): 185–214. 33. Other think tanks were also involved: the Institute for the United States and Canada (ISKAN) and its heads, Georgi Arbatov and Andrey Kokoshin, played a vital role in developing security policy, as did the Institute of Economics of the World Socialist System (IEMSS). 34. Checkel, “Ideas, Institutions, and the Gorbachev Foreign Policy Revolution,” 273. 35. Herman, “Identity, Norms and National Security,” 304. 36. Wendt, “Constructing International Politics,” 77. 37. Stephen Walt, The Origin of Alliances (Ithaca, NY: Cornell University Press, 1987), 5. 38. Wendt, “Constructing International Politics,” 73. 39. Hopf, “The Promise of Constructivism,” 188, 199. 40. See Farrell, “Constructivist Security Studies,” 65–67, and Hopf, “The Promise of Constructivism,” 192. 41. Ba and Hoffmann, “Making and Remaking the World of IR 101.” 42. See, for instance, Gregory Raymond, “International Norms: Normative Orders and Peace,” in What Do We Know About War? ed. John Vasquez (Lanham, MD: Rowman & Littlefield, 2000), 282–95; Charles W. Kegley Jr., and Gregory Raymond, “International Legal Norms and the Preservation of Peace, 1820–1964: Some Evidence and Bivariate Relationships,” International Interactions 8 (1981): 171–87; Charles W. Kegley Jr. and Gregory Raymond, “Normative Constraints on the Use of Force Short of War,” Journal of Peace Research 23 (1986): 213–27; Charles W. Kegley Jr. and Gregory Raymond, When Trust Breaks Down: Alliance Norms and World Politics (Columbia: University of South Carolina Press, 1990); and Charles W. Kegley Jr. and Gregory Raymond, “Preventive War and Permissive Normative Order,” International Studies Perspectives 4 (2003): 385–94. 43. Raymond, “International Norms,” 283. 44. Raymond, “International Norms,” 282. 45. Raymond, “International Norms,” 292. 46. Raymond, “International Norms,” 292–93; Kegley and Raymond, “International Legal Norms and the Preservation of Peace, 1820–1964”; Kegley and Raymond, When Trust Breaks Down; Kegley and Raymond, “Normative Constraints on the Use of Force Short of War.”

568

Notes

47. See Martha Finnemore, “Constructing Norms of Humanitarian Intervention,” in The Culture of National Security, ed. Peter Katzenstein (New York: Columbia University Press, 1996), 153–85. 48. Richard Price, “A Genealogy of the Chemical Weapons Taboo,” International Organization 49, no. 1 (1995): 73–103, and Price, The Chemical Weapons Taboo (Ithaca, NY: Cornell University Press, 1997). 49. Nina Tannenwald, The Nuclear Taboo (Cambridge: Cambridge University Press, 2007); T. V. Paul, The Tradition of Non-Use of Nuclear Weapons (Stanford, CA: Stanford University Press, 2009). 50. Paul, The Tradition of Non-Use of Nuclear Weapons, 21. 51. Paul, The Tradition of Non-Use of Nuclear Weapons, 26. 52. John Mueller, Retreat from Doomsday: The Obsolescence of Major War (New York: Basic Books, 1989); Mueller, “Changing Attitudes towards War: The Impact of the First World War,” British Journal of Political Science 21 (1991): 1–28; and Mueller, “Is War Still Becoming Obsolete?” paper presented to American Political Science Association conference, Washington, DC, 1991. 53. Mueller, “Is War Still Becoming Obsolete?” 54–55. 54. Peter Howard, “Why Not Invade North Korea? Threats, Language Games, and U.S. Foreign Policy,” International Studies Quarterly 48 (2004): 805–28. 55. Howard, “Why Not Invade North Korea?” 812.

13. CONCLUSION 1. D. Scott Bennett and Alan C. Stam, The Behavioral Origins of War (Ann Arbor: University of Michigan Press, 2004). 2. John Vasquez, “What Do We Know About War?” in What Do We Know About War? 2nd ed., ed. John Vasquez (Lanham, MD: Rowman & Littlefield, 2012), 329. 3. It is probably true that certain cultures tend to be more accepting of martial values than others and this affects the content of its leaders’ operational codes. In these cultures, bullying tactics are more generally acceptable, and political elites who represent these values tend to gain and retain a substantial degree of legitimacy. 4. Jack Levy, “The ‘Paths-to-War’ Concept,” in Vasquez, What Do We Know about War? 2nd ed., 281–90. 5. Jack Levy and William R. Thompson, Causes of War (Chichester, UK: Wiley-Blackwell, 2010), 213. 6. See, for instance, Charles W. Kegley Jr. and Gregory A. Raymond, The Global Future: A Brief Introduction to World Politics, 2nd ed. (Belmont, CA: Wadsworth, 2007), 164–66. 7. Other theorists would add the rising costliness of war, accelerated tremendously by the creation and proliferation of nuclear weapons. See, for instance, Kenneth Waltz, “Nuclear Myths and Political Realities,” American Political Science Review 84, no. 3 (September 1990): 731–45; John Mearsheimer, “Back to the Future: Instability in Europe After the Cold War,” International Security 15, no. 1 (Summer 1990): 5–56; Bruce Bueno de Mesquita and William Riker, “An Assessment of the Merits of Selective Nuclear Proliferation,” Journal of Conflict Resolution 26 (1982): 287–306; and Carl Kaysen, “Is War Obsolete? A Review Essay,” International Security 14, no. 4 (Spring 1990): 42–64. 8. See the discussion in John A. Vasquez, The War Puzzle Revisited (Cambridge: Cambridge University Press, 2009), 362–71. 9. Bruce Russett and John Oneal, Triangulating Peace: Democracy, Interdependence, and International Organizations (New York: Norton, 2001).

Selected Bibliography

Adler, E. (1997) “Seizing the Middle Ground: Constructivism in World Politics.” European Journal of International Relations 3 (3): 319–63. Allee, T., and P. Huth (2006) “The Pursuit of Legal Settlements in Territorial Disputes.” Conflict Management and Peace Science 23 (Winter): 285–307. Allison, G. (1971) Essence of Decision: Explaining the Cuban Missile Crisis. Boston: Little, Brown. Allison, G., and M. Halperin (1972) “Bureaucratic Politics: A Paradigm and Some Policy Implications,” 40–79 in R. Tanter and R. Ullman (eds.), Theory and Policy in International Relations. Princeton, NJ: Princeton University Press. Allison, G., and P. Zelikow (1999) Essence of Decision: Explaining the Cuban Missile Crisis, 2nd ed. New York: Longman. Anderson, P. A. (1987) “What Do Decision Makers Do When They Make Foreign Policy?” In New Directions in the Study of Foreign Policy, ed. C. F. Hermann, C. W. Kegley, and J. N. Rosenau, 285–308. Boston: Allen and Unwin. Ardrey, R. (1966) The Territorial Imperative. New York: Atheneum. Art, R. (1974) “Bureaucratic Politics and American Foreign Policy: A Critique.” Policy Sciences (Summer): 467–90. Axelrod, R. (1980) “Effective Choice in the Prisoners’ Dilemma.” Journal of Conflict Resolution 24: 3–25. ——— (1984) The Evolution of Cooperation. New York: Basic Books. Ba, A., and M. Hoffmann (2003) “Making and Remaking the World for IR 101: A Resource for Teaching Social Constructivism in Introductory Classes.” International Studies Perspectives 4: 15–33. Babst, D. V. (1964) “Elected Governments: A Force for Peace.” The Wisconsin Sociologist 3: 9–14. Badie, D. (2010) “Groupthink, Iraq, and the War on Terror: Explaining US Policy Shift toward Iraq.” Foreign Policy Analysis 6 (4): 277–96. Bandura, A. (1980) “The Social Learning Theory of Aggression.” In The War System, ed. R. Falk and S.S. Kim, 141–56. Boulder, CO: Westview. Baram, A. (2012) “Deterrence Lessons from Iraq: Rationality Is Not the Only Key to Containment.” Foreign Affairs 91 (4) (July/August): 76–90. Bar-Joseph, U., and R. McDermott (2008) “Personal Functioning Under Stress: Accountability and Social Support of Israeli Leaders in the Yom Kippur War.” Journal of Conflict Resolution 25 (1): 144–70. Beardsley, K., and V. Asal (2009) “Nuclear Weapons as Shields.” Conflict Management and Peace Science 26: 235–55. Behr, R. (1981) “Nice Guys Finish Last—Sometimes.” Journal of Conflict Resolution 25: 289–300. Bendor, J., and T. Hammond (1992) “Rethinking Allison’s Models.” American Political Science Review 86 (2) (June): 301–22. Bennett, D. S. (1997) “Democracy, Regime Change, and Rivalry Termination.” International Interactions 22 (4): 367–97. ——— (2006) “Towards a Continuous Specification of the Democracy-Autocracy Connection.” International Studies Quarterly 50 (2): 313–38. Bennett, D. S., and A. C. Stam (2004) The Behavioral Origins of War. Ann Arbor: University of Michigan Press.

569

570

Selected Bibliography

Benoit, K. (1996) “Democracies Really Are More Pacific (in General).” Journal of Conflict Resolution 40 (4): 636–57. Benson, B. (2012) “Alliances: ATOP Data and Deterrence.” In What Do We Know About War? 2nd ed., ed. J. Vasquez, 45–62. Lanham, MD: Rowman & Littlefield. Ben-Yehuda, H. (2004) “Territoriality and War in International Crises: Theory and Findings, 1918–2001.” International Studies Review 5: 85–105. Betts, R. K. (1977) Soldiers, Statesmen and Cold War Crises. Cambridge, MA: Harvard University Press. ——— (1999) “Must War Find a Way?” International Security 24 (2): 166–98. Birt, R. (1993) “Personality and Foreign Policy: The Case of Stalin.” Political Psychology 14, no. 4, 607–25. Blainey, G. (1973) The Causes of War. New York: Free Press. Boulding, K. (1956) The Image. Ann Arbor: University of Michigan Press. ——— (1962) Conflict and Defense: A General Theory. New York: Harper & Row. ——— (1967) “The Learning and Reality Testing Process in the International System.” In Image and Reality in World Politics, ed. J. C. Farrell and A. P. Smith, 1–15. New York: Columbia University Press. Braybrooke, D., and C. Lindblom (1969) “Types of Decision-Making.” In International Politics and Foreign Policy, ed. J. Rosenau, 207–16. New York: Free Press. Brecher, M., and J. Wilkenfeld (1989) Crisis, Conflict and Instability. Oxford: Pergamon Press. ——— (2000) A Study of Crisis. Ann Arbor: University of Michigan Press. Bremer, S. (1980) “National Capabilities and War Proneness.” In The Correlates of War II: Testing Some Realpolitik Models, ed. J. D. Singer, 57–82. New York: Free Press. ——— (1992) “Dangerous Dyads: Conditions Affecting the Likelihood of Interstate War, 1816–1965.” Journal of Conflict Resolution 36 (2): 309–41. Bremer, S., J. D. Singer, and U. Luterbacher (1973) “The Population Density and War Proneness of European Nations, 1816–1965.” Comparative Political Studies 6: 329–48. Bueno de Mesquita, B. (1981) The War Trap. New Haven, CT: Yale University Press. ——— (1981) “Risk, Power Distribution and the Likelihood of War.” International Studies Quarterly 25 (4): 541–68. Bueno de Mesquita, B., and D. Lalman (1992) War and Reason. New Haven, CT: Yale University Press. Bueno de Mesquita, B., R. Siverson, and Gary Woller (1992). “War and the Fate of Regimes,” American Political Science Review 86 (3): 638–46. Bueno de Mesquita, B., A. Smith, R. Siverson, and J. Morrow (2003) The Logic of Political Survival. Cambridge, MA: MIT Press. Bundy, McG. (1988) Danger and Survival: Choices about the Bomb in the First Fifty Years. New York: Random House. Burr, W., and J. Richelson (2000/01) “Whether to ‘Strangle the Baby in the Cradle’: The United States and the Chinese Nuclear Program, 1960–64,” International Security 25: 54–99. Byman, D., and K. Pollack (2001) “Now Let Us Praise Great Men: Bringing the Statesmen Back In.” International Security 25 (4): 107–46. Caldwell, D. (1977) “Bureaucratic Foreign Policy Making.” American Behavioral Scientist 21 (2): 87–110. Caprioli, M., and P. F. Trumbore (2005) “Rhetoric versus Reality: Rogue States in International Conflict.” Journal of Conflict Resolution 49 (5) (October): 770–91. Carniero, R. (1994) “War and Peace: Alternating Realities in Human History.” In Studying War: Anthropological Perspectives, ed. S. P. Reyna and R. E. Downs, 3–27. Langhorne, PA: Gordon and Breach. Cartwright, D. (1971) “Risk-taking by Individuals and Groups: An Assessment of Research Employing Choice Dilemmas.” Journal of Personality and Social Psychology 20: 261–78. Cashman, G. (1997) “Theoretical Buttresses for Power Cycle Theory.” Paper presented at the International Political Science Association XVIIth World Congress, Seoul, Korea, August 17–21. Cashman, G., and L. Robinson (2007) An Introduction to the Causes of War: Patterns of Interstate Conflict from World War I to Iraq. Lanham, MD: Rowman & Littlefield. Chan, S. (1984) “Mirror, Mirror on the Wall . . . Are the Freer Countries More Pacific?” Journal of Conflict Resolution 28 (4): 617–48. ——— (1997) “In Search of Democratic Peace: Problems and Promise.” Mershon International Studies Review 41 (1): 59–91. Chase-Dunn, C. (1981) “Interstate System and Capitalist World-Economy: One Logic or Two? International Studies Quarterly 25 (1): 119–42.

Selected Bibliography

571

Checkel, J. (1993) “Ideas, Institutions, and the Gorbachev Foreign Policy Revolution.” World Politics 45: 271–300. ——— (1998) “The Constructivist Turn in International Relations Theory.” World Politics 50 (2): 324–48. Chojnacki, S. (2006) “Democratic Wars and Military Interventions, 1946–2002: The Monadic Level Reconsidered.” In Democratic Wars: Looking at the Dark Side of Democratic Peace, ed. A. Geis, L. Brock, and H. Muller, 13–37. New York: Palgrave Macmillan. Choucri, N., and R. North (1975) Nations in Conflict: National Growth and International Violence. San Francisco: W. H. Freeman. ——— (1989) “Lateral Pressure in International Relations: Concept and Theory.” In Handbook of War Studies, ed. M. Midlarsky, 289–326. Boston: Unwin Hyman. Colaresi, M. P., K. Rasler, and W. R. Thompson (2007) Strategic Rivalries in World Politics: Position, Space and Conflict Escalation. Cambridge: Cambridge University Press. Colaresi, M., and W. Thompson (2005) “Alliances, Arms Buildups and Recurrent Conflict: Testing a Steps-to-War Model.” Journal of Politics 67 (May): 345–64. Copeland, Dale C. (2000) The Origins of Major War. Ithaca, NY: Cornell University Press. Crawford, N. (2000) “The Passion of World Politics: Propositions on Emotion and Emotional Relationships.” International Security 24 (4): 116–56. Crawford, T. (2009) “The Endurance of Extended Deterrence: Continuity, Change, and Complexity in Theory and Policy.” In Complex Deterrence, ed. T. V. Paul, P. Morgan, and J. Wirtz, 277–303. Chicago: University of Chicago Press. Crescenzi, M. J. C., and A. J. Enterline (1999) “Ripples from the Waves? A Systemic, Time-Series Analysis of Democracy, Democratization, and Interstate War.” Journal of Peace Research 36 (1): 75–94. Darcey, R., and N. Pendegraft (1988) “The Optimality of TIT-FOR-TAT.” International Interactions 15 (1): 45–57. Dassel, K. (1998) “Civilians, Soldiers, and Strife: Domestic Sources of International Aggression.” International Security 23 (1): 107–40. Dassel, K., and E. Reinhardt (1999) “Domestic Strife and the Initiation of Violence at Home and Abroad.” American Journal of Political Science 43 (1): 56–85. Davies, G. (2002) “Domestic Strife and the Initiation of International Conflicts: A Directed Dyad Analysis, 1950–1982.” Journal of Conflict Resolution 46 (5): 672–92. Dawisha, K. (1980) “The Limits of the Bureaucratic Politics Model: Observations on the Soviet Case.” Studies in Comparative Communism 13 (4) (Winter): 300–346. DeRouen, K. (2003) “The Decision Not to Use Force at Dien Bien Phu.” In Integrating Cognitive and Rational Theories of Foreign Policy Decision Making, ed. A. Mintz, 11–28. New York: Palgrave Macmillan. Desch, M. C. (2002) “Democracy and Victory: Why Regime Type Hardly Matters.” International Security 27 (2): 5–47. Deutsch, K., and R. Merritt (1965) “Effects of Events on National and International Images,” 132–87. In International Behavior, ed. H. Kelman. New York: Holt, Rinehart & Winston. Deutsch, K., and J. D. Singer (1964) “Multipolar Power Systems and International Stability.” World Politics 16 (3): 390–406. DiCicco, J., and J. S. Levy (1999) “Power Shifts and Problem Shifts: The Evolution of the Power Transition Research Program.” Journal of Conflict Resolution 43 (6): 675–704. Diehl, P. F. (1983) “Arms Races and Escalation: A Closer Look.” Journal of Peace Research 20 (3): 205–12. ——— (1985a) “Contiguity and Military Escalation in Major Power Rivalries, 1816–1980.” Journal of Politics 47 (4): 1203–11. ——— (1985b) “Arms Races to War; An Analysis of Some Underlying Effects.” Sociological Quarterly 26: 331–49. ———,ed. (1988) The Dynamics of Enduring Rivalries. Urbana: University of Illinois Press. Diehl, P. F., and G. Goertz (1988) “Territorial Changes and Militarized Conflict.” Journal of Conflict Resolution 32 (l): 103–22. ——— (2000) War and Peace in International Rivalry. Ann Arbor: University of Michigan Press. ——— (2012) “The Rivalry Process: How Rivalries are Sustained and Terminated.” In What Do We Know About War? 2nd ed., ed. John A Vasquez, 83–109. Lanham, MD: Rowman & Littlefield. Diehl, P. F., and J. Kingston (1987) “Messenger or Message? Military Buildups and the Initiation of Conflict.” Journal of Politics 49: 789–99. Dixon, W. J. (1986) “Reciprocity in United States-Soviet Relations: Multiple Symmetry or Issue Linkage?” American Journal of Political Science 30: 421–45.

572

Selected Bibliography

——— (1993) “Democracy and the Management of International Conflict.” Journal of Conflict Resolution 37 (1): 42–68. ——— (1994) “Democracy and the Peaceful Settlement of International Conflict.” American Political Science Review 88 (1): 14–32. Dixon, W. J., and P. D. Senese (2002) “Democracy, Disputes, and Negotiated Settlements.” Journal of Conflict Resolution 46 (4): 547–71. Doran, C. F. (1983) “War and Power Dynamics: Economic Underpinnings.” International Studies Quarterly 27: 419–44. ——— (1989) “Systemic Disequilibrium, Foreign Policy Role, and the Power Cycle: Challenges for Research Design.” Journal of Conflict Resolution 33 (3): 371–401. ——— (1991) System in Crisis: New Imperatives of High Politics at Century’s End. Cambridge: Cambridge University Press. ——— (2000) “Confronting the Principles of the Power Cycle: Changing Systems Structure, Expectations, and War.” In Handbook of War Studies II, 332–68. Ann Arbor: University of Michigan Press. Doran, C. F., and W. Parsons (1980) “War and the Cycle of Relative Power.” American Political Science Review 74: 947–65. Doyle, M. W. (1986) “Liberalism and World Politics.” American Political Science Review 80 (4): 1151–69. ——— (1996) “Kant, Liberal Legacies, and Foreign Affairs.” In Debating the Democratic Peace, ed. M. Brown, S. Lynn-Jones, and S. Miller, 3–57. Cambridge, MA: MIT Press. ——— (1997) Ways of War and Peace. New York: W. W. Norton. Downes, A. (2009) “How Smart and Tough Are Democracies? Reassessing Theories of Democratic Victory in War.” International Security 33 (4) (Spring): 9–51. Duelfer, C. C., and S. B. Dyson (2011) “Chronic Misperception and International Conflict: The U.S.-Iraq Experience.” International Security 36 (1) (Summer): 73–100. Dyer, G. (1985) War. New York: Dorsey. Dyson, S. B. (2006) “Personality and Foreign Policy: Tony Blair’s Iraq Decisions.” Foreign Policy Analysis 2 (3): 289–306. ——— (2009) “‘Stuff Happens’: Donald Rumsfeld and the Iraq War.” Foreign Policy Analysis 5 (4): 327–47. East, M. A. (1972) “Status Discrepancy and Violence in the International System. An Empirical Analysis.” In The Analysis of International Politics, ed. J. N. Rosenau, V. Davis, and M. A. East, 299–319. New York: Free Press. Ember, M., and C. R. Ember (1994) “Cross-Cultural Studies of War and Peace: Recent Achievements and Future Possibilities.” In Studying War: Anthropological Perspectives, ed. S. P. Reyna and R. E. Downs, 185–208. Langhorne, PA: Gordon and Breach. Enterline, A. (1996) “Driving While Democratizing.” International Security 20 (4): 183–96. ——— (1998) “Regime Changes, Neighborhoods, and Interstate Conflict, 1816–1992.” Journal of Conflict Resolution 42 (6): 804–29. Esser, J. K. (1998) “Alive and Well after 25 Years: A Review of Groupthink Research.” Organizational Behavior and Human Decision Processes 73 (2/3): 116–141. Etheredge. L. (1975) “Personality and Foreign Policy.” Psychology Today (March): 37–42. ——— (1978) “Personality Effects on American Foreign Policy, 1898–1968.” American Political Science Review 72: 434–51. ——— (1978) A World of Men: The Private Sources of American Foreign Policy. Cambridge, MA: MIT Press. ——— (1979) “Hard Ball Politics: A Model.” Political Psychology I: 3–26. Fabbro, D. (1980) “Peaceful Societies” In The War System, ed. R. Falk and S. S. Kim, 180–203. Boulder, CO: Westview. Farber, H. S., and J. Gowa (1995) “Polities and Peace.” International Security 20 (2): 123–46. Farrell, T. (2002) “Constructivist Security Studies: Portrait of a Research Program.” International Studies Review 4 (1): 49–72. Fearon, J. (1994) “Domestic Political Audiences and the Escalation of International Disputes.” American Political Science Review 88: 577–92. ——— (1995) “Rationalist Explanations for War.” International Organization 49: 379–414. Feng, H. (2005) “The Operational Code of Mao Zedong: Defensive or Offensive Realist?” Security Studies 14 (4): 637–62. Festinger, L. (1957) A Theory of Cognitive Dissonance. Evanston, IL: Row, Patterson.

Selected Bibliography

573

Filson, D., and S. Werner (2002) “A Bargaining Model of War and Peace: Anticipating the Onset, Duration, and Outcome of War.” American Journal of Political Science 46 (4): 819–38. Finnemore, M., and K. Sikkink (2013) “International Norm Dynamics and Political Change.” In International Politics: Classic and Contemporary Readings, ed. S. Handler, 108–16. Los Angeles: Sage. Fodor, E. M., and T. Smith (1982) “The Power Motive as an Influence on Group Decision Making.” Journal of Personality and Social Psychology 42: 178–85. Fossey, D. (1983) Gorillas in the Mist. Boston: Houghton Mifflin. Freud, S. (1985) “Why War?” In International War: An Anthology, ed. M. Small and J. D. Singer, 158–63. Homewood, IL: Dorsey Press. Fukuyama F. (1998) “Women and the Evolution of World Politics.” Foreign Affairs 77 (5): 24–40. Gallucci, R. (1975) Neither Peace nor Honor. Baltimore: Johns Hopkins University Press. Galtung, J. (1964) “A Structural Theory of Aggression.” Journal of Peace Research 1: 95–119. Gamson, W. A., and A. Modigliani (1971) Untangling the Cold War: A Strategy for Testing Rival Theories. Boston: Little, Brown. Garnham, D. (1976) “Dyadic International War, 1816–1965: The Role of Power Parity and Geographic Proximity.” Western Political Quarterly 29: 231–42. ——— (1985) “The Causes of War: Systemic Findings,” In Polarity and War, ed. A. N. Sabrosky, 7–23. Boulder, CO: Westview. ——— (1986) “War-Proneness, War-Weariness, and Regime Type: 1816–1980.” Journal of Peace Research 23 (3): 279–89. Gat, A. (1999) “The Pattern of Fighting in Simple, Small-Scale, Prestate Societies.” Journal of Anthropological Research 55: 563–83. ——— (2006) War in Human Civilization. Oxford: Oxford University Press. Gaubatz, K. T. (1991) “Election Cycles and War.” Journal of Conflict Resolution 35: 212–44. Gelb, L., and R. Betts (1979) The Irony of Vietnam: The System Worked. Washington, DC: Brookings Institution. Geller, D. S. (1992) “Power Transition and Conflict Initiation.” Conflict Management and Peace Science 12: 1–16. ——— (1996) “Relative Power, Rationality, and International Conflict.” In Parity and War, ed. J. Kugler and D. Lemke, 127–44. Ann Arbor: University of Michigan Press. ——— (2000) “Status Quo Orientation, Capabilities, and Patterns of War Initiation in Dyadic Rivalries,” Conflict Management and Peace Science 18 (1): 73–96. ——— (2000) “Material Capabilities: Power and International Conflict.” In What Do We Know About War?, ed. John Vasquez, 259–77. Lanham, MD: Rowman & Littlefield. ——— (2012) “Nuclear Weapons and War.” In What Do We Know about War? ed. J. Vasquez, 139–63. Lanham, MD: Rowman & Littlefield. Geller, D. S., and J. David Singer (1998) Nations at War: A Scientific Study of International Conflict. Cambridge: Cambridge University Press. Gelpi, C. (1997) “Democratic Diversions: Government Structure and the Externalization of Domestic Conflict.” Journal of Conflict Resolution 41 (2): 255–82. George, A. L. (1972) “The Case for Multiple Advocacy in Making Foreign Policy.” American Political Science Review 66: 751–85. ——— (1980) “The Operational Code”: A Neglected Approach to the Study of Political Leaders and Decision Making,” In The Conduct of Soviet Foreign Policy, ed. E. Hoffman and F. Fleron, 165–90. New York: Aldine. George, A. L., and R. Smoke (1974) Deterrence in American Foreign Policy: Theory and Practice. New York: Columbia University Press. Geva, N., and A. Mintz, eds. (1997) Decision Making on War and Peace: The Cognitive-Rational Debate. Boulder: Lynne Rienner. Gibler, D. (1997) “Control the Issues, Control the Conflict: The Effects of Alliances That Settle Territorial Issues on Interstate Rivalries.” International Interactions 22: 341–68. _____ (2000) “Alliances: Why Some Cause War and Why Others Cause Peace.” In What Do We Know About War?, ed. J. Vasquez. Lanham, MD: Rowman & Littlefield. ——— (2007) “Bordering on Peace: Democracy, Territorial Issues, and Conflict.” International Studies Quarterly 51: 509–32. ——— (2012) “The Implications of a Territorial Peace.” In What Do We Know About War?, 2nd ed., ed. J. Vasquez, 211–34. Lanham, MD: Rowman & Littlefield.

574

Selected Bibliography

Gibler, D., T. Rider, and M. Hutchison (2005) “Taking Arms against a Sea of Troubles: Conventional Arms Races during Periods of Rivalry.” Journal of Peace Research 42: 131–47. Gibler, D., and J. Tir (2010) “Settled Borders and Regime Type: Democratic Transitions as Consequences of Peaceful Territorial Transfers.” American Journal of Political Science 54 (4): 951–68. Gibler, D., and J. A. Vasquez (1998) “Uncovering Dangerous Alliances, 1495–1980.” International Studies Quarterly 42: 785–807. Gilpin, R. (1981) War and Change in World Politics. Cambridge: Cambridge University Press. ——— (1988) “The Theory of Hegemonic War.” In The Origin and Prevention of Major Wars, ed. R. Rotberg and T. Rabb, 15–37. Cambridge: Cambridge University Press. Glaser, C. L. (1997) “The Security Dilemma Revisited.” World Politics 50 (October): 171–201. Gleditsch, K. S., I. Salehyan, and K. Schultz (2008) “Fighting at Home, Fighting Abroad: How Civil Wars Lead to International Disputes.” Journal of Conflict Resolution 52 (4): 479–506. Gleditsch, N. P. (1995) “Geography, Democracy and Peace.” International Interactions 20: 297–323. ——— (1998) “Armed Conflict and the Environment: A Critique of the Literature.” Journal of Peace Research 35 (3): 381–400. Gleditsch, N. P., and H. Hegre (1997) “Peace and Democracy: Three Levels of Analysis.” Journal of Conflict Resolution 41 (2): 283–310. Gochman, C. (1990) “Capability-Driven Disputes,” 141–59 in C. Gochman and A. N. Sabrosky (eds.), Prisoners of War? Nation-States in the Modern Era. Lexington, MA: Lexington Books. Gochman, C., and Z. Maoz (1984) “Militarized Interstate Disputes, 1816–1976: Procedures, Patterns and Insights.” Journal of Conflict Resolution 28: 585–616. Goertz, G., and P. F. Diehl (2000) “(Enduring) Rivalries.” In Handbook of War Studies II, ed. M. I. Midlarsky. Ann Arbor: University of Michigan Press. ——— (2000) “Rivalries: The Conflict Process.” In What Do We Know About War?, ed. J. A. Vasquez. Lanham, MD: Rowman & Littlefield. Goldstein, J. (1988) Long Cycles: Prosperity and War in the Modern Era. New Haven, CT: Yale University Press. ——— (1991) “Reciprocity in Superpower Relations: An Empirical Analysis.” International Studies Quarterly 35 (2): 195–209. ——— (2001) War and Gender. Cambridge: Cambridge University Press. Goldstein, J. and J. R. Freeman (1990) Three-Way Street: Strategic Reciprocity and World Politics. Chicago: University of Chicago Press. Goncharov, S. N., J. W. Lewis, and X. Litai (1993) Uncertain Partners: Stalin, Mao, and the Korean War. Stanford, CA: Stanford University Press. Goodall, J. (1990) Through a Window: My Thirty Years with the Chimpanzees of Gombe. Boston: Houghton Mifflin. Haas, J. (2001) “Warfare and the Evolution of Culture.” Archaeology at the Millennium, ed. G. M. Feinman and T. D. Price, 329–50. New York: Kluwer/Plenum. Haas, P. (1992) “Introduction: Epistemic Communities and International Policy Coordination.” International Organization 46 (1): 1–35. Hagan, J. (2001) “Does Decision Making Matter?” International Studies Review 3 (2): 5–46. Halberstam, D. (1972) The Best and the Brightest. Greenwich, CT: Fawcett. Halperin, M. (1974) Bureaucratic Politics and Foreign Policy. Washington, DC: Brookings Institution. Hart, P. ‘t (1990) Groupthink in Government: A Study of Small Groups and Policy Failure. Baltimore: Johns Hopkins University Press. ——— (1991) “Irving L. Janis’ Victims of Groupthink.” Political Psychology 12: 247–78. Hart, P. ‘t, E. Stern, and B. Sundelius, eds. (2008) Beyond Groupthink: Political Group Dynamics and Foreign Policy-making. Ann Arbor: University of Michigan Press. ——— (2008) “Foreign Policy-making at the Top: Political Group Dynamics.” In Beyond Groupthink, ed. P. ‘t Hart, E. Stern, and B. Sundelius, 3–34. Ann Arbor: University of Michigan Press. Hart, P. ‘t, and U. Rosenthal (1998) “Reappraising Bureaucratic Politics.” Mershon International Studies Review 42 (2): 233–40. Harvey, F. (1997) “Deterrence and Compellence in Protracted Crises: Methodology and Preliminary Findings.” International Studies Notes 22 (1): 12–23. Harvey, F., and P. James (2009) “Deterrence and Compellence in Iraq, 1991–2003: Lessons for a Complex Paradigm.” In Complex Deterrence, ed. T. V. Paul, P. Morgan, and J. Wirtz, 222–55. Chicago: University of Chicago Press.

Selected Bibliography

575

Hastings, M., and S. Jenkins (1983) The Battle for the Falklands. New York: Norton. Hebron, L., and P. James (1997) “Great Powers, Cycles of Relative Capability and Crises in World Politics.” International Interactions 23 (2): 145–73. Henderson, E. (2002) Democracy and War: The End of an Illusion? Boulder: Rienner. Hensel, P. (1996) “Charting a Course to Conflict: Territorial Issues and Interstate Conflict, 1816–1992.” Conflict Management and Peace Science 15 (Spring): 43–73. ——— (2000) “Territory: Theory and Evidence on Geography and Conflict.” In What Do We Know About War? ed. J. Vasquez, 57–84. Lanham, MD: Rowman & Littlefield. ——— (2012) “Territory: Geography, Contentious Issues, and World Politics.” In What Do We Know About War? 2nd ed., ed. J. Vasquez, 3–26. Lanham, MD: Rowman & Littlefield. Hensel, P., and P. Diehl (1994) “It Takes Two to Tango: Non-Militarized Response in International Disputes.” Journal of Conflict Resolution 38 (September): 479–506. Hensel, P., G. Goertz, and P. F. Diehl (2000) “The Democratic Peace and Rivalries.” Journal of Politics 62 (4): 1173–88. Herek, M., I. L. Janis, and P. Huth (1987) “Decision Making During International Crises: Is Quality of Process Related to Outcome?” Journal of Conflict Resolution 31 (2): 203–26. Herman, R. G. (1996) “Identity, Norms, and National Security: The Soviet Foreign Policy Revolution and the End of the Cold War.” In The Culture of National Security: Norms and Identity in World Politics, ed. P. J. Katzenstein, 271–316. New York: Columbia University Press. Hermann, M. G. (ed.) (1977) A Psychological Examination of Political Leaders. New York: Free Press. ——— (1978) “Effects of Personal Characteristics of Political Leaders on Foreign Policy.” In Why Nations Act: Theoretical Perspectives for Comparative Foreign Policy, ed. M. East, S. Salmore, and C. F. Hermann, 49–68. Beverley Hills, CA: Sage. ——— (2005) “Assessing Leadership Style: Trait Analysis.” In The Psychological Assessment of Political Leaders, ed. J. Post, 178–212. Ann Arbor: University of Michigan Press. ——— (2005) “Saddam Hussein’s Leadership Style.” In The Psychological Assessment of Political Leaders, ed. J. Post, 375–86. Ann Arbor: University of Michigan. Hermann, M. G., T. Preston, B. Korany, and T. Shaw (2001) “Who Leads Matters: The Effects of Powerful Individuals.” International Studies Review 3 (2): 83–131. Hobson, J. A. (1965) Imperialism: A Study. Ann Arbor: University of Michigan Press. Hollis, M., and S. Smith (1986) “Roles and Reasons in Foreign Policy Decision Making.” British Journal of Political Science 16 (3) (July): 269–86. Holsti, K. J. (1970) “National Role Conceptions in the Study of Foreign Policy.” International Studies Quarterly 14 (3): 233–309. ——— (1991) Peace and War: Armed Conflicts and International Order 1648–1989. Cambridge: Cambridge University Press. Holsti, O. (1967) “Cognitive Dynamics and Images of the Enemy.” In Image and Reality in World Politics, ed. J. C. Farrell and A. P. Smith, 16–39. New York: Columbia University Press. ——— (1969) “The Belief System and National Images: A Case Study.” In International Politics and Foreign Policy, rev. ed., ed. J. Rosenau, 543–50. New York: Free Press. ——— (1970) “The ‘Operational Code’ Approach to the Study of Political Leaders: John Foster Dulles’ Philosophical and Instrumental Beliefs.” Canadian Journal of Political Science 3 (1): 123–57. ——— (1972) Crisis, Escalation, War. Montreal: McGill-Queens University Press. ——— (1987) “Theories of Crisis Decision Making.” In International Relations Theory, ed. P. Viotti and M. Kauppi, 244–81. New York: Macmillan. ——— (1990) “Crisis Management.” In Psychological Dimensions of War, ed. Betty Glad, 116–42. Newbury Park, CA: Sage. Holsti, O., R. Brody, and R. North (1965) “Measuring Affect and Action in International Reaction Models: Empirical Materials from the 1962 Cuban Crisis.” Peace Research Society (International) 2: 170–90. Holsti, O., R. North, and R. Brody (1968) “Perception and Action in the 1914 Crisis.” In Quantitative International Politics, ed. J. D. Singer, 123–59. New York: Free Press. Homer-Dixon, T. (1999) Environment, Scarcity, and Violence. Princeton, NJ: Princeton University Press. Hopf, T. (1991) “Polarity, the Offense-Defense Balance, and War.” American Political Science Review 85: 475–94. ——— (1994) Peripheral Visions: Deterrence Theory and American Foreign Policy in the Third World, 1965–1990. Ann Arbor: University of Michigan Press.

576

Selected Bibliography

——— (1998) “The Promise of Constructivism in International Relations Theory.” International Security 23 (1): 171–200. Houweling, H., and J. Siccama (1988) “Power Transitions as a Cause of War.” Journal of Conflict Resolution 32 (1): 87–102. ——— (1991) “Power Transitions and Critical Points as Predictors of Great Power War: Toward a Synthesis.” Journal of Conflict Resolution 35 (4): 642–58. Howard, P. (2004) “Why Not Invade North Korea? Threats, Language Games, and U.S. Foreign Policy.” International Studies Quarterly 48: 805–28. Hoyt, P. D., and J. A. Garrison (2008) “Political Manipulation with the Small Group: Foreign Policy Advisers in the Carter Administration.” In Beyond Groupthink, ed. P. ‘t Hart, E. Stern, and B. Sundelius, 249–74. Ann Arbor: University of Michigan Press. Huntington, S. P. (1958) “Arms Races: Prerequisites and Results,” In Public Policy, vol. 8, ed. C. J. Friedrich and S. E. Harris, 41–86. Cambridge, MA: Graduate School of Public Administration, Harvard University. Huth, P. (1988) Extended Deterrence and the Prevention of War. New Haven, CT: Yale University Press. ——— (1996) Standing Your Ground: Territorial Disputes and International Conflict. Ann Arbor: University of Michigan. ——— (1999) “Deterrence and International Conflict: Empirical Findings and Theoretical Debates.” Annual Review of Political Science 2: 25–48. Huth, P., and T. Allee (2002) The Democratic Peace and Territorial Conflict in the Twentieth Century. New York: Cambridge University Press. Huth, P., D. S. Bennett, and C. Gelpi (1992) “System Uncertainty, Risk Propensity, and International Conflict among the Great Powers.” Journal of Conflict Resolution 36: 478–517. Huth, P., and B. Russett (1984) “What Makes Deterrence Work? Cases from 1900–1980.” World Politics 36: 496–526. ——— (1988) “Deterrence Failure and Crisis Escalation.” International Studies Quarterly 32: 29–45. ——— (1990) “Testing Deterrence Theories: Rigor Makes a Difference.” World Politics 42: 466–501. ——— (1993) “General Deterrence between Enduring Rivals: Testing Three Competing Models.” American Political Science Review 87 (1): 61–73. Ikenberry, J., M. Mastanduno, and W. Wohlforth (2009) “Unipolarity, State Behavior, and Systemic Consequences.” World Politics 61 (1): 1–27. James, P. (1988) Crisis and War. Montreal: McGill-Queens University Press. James, P., and J. R. Oneal (1991) “Influences on the President’s Use of Force.” Journal of Conflict Resolution 35: 307–34. James, W. (1968) “The Moral Equivalent of War.” In War: Studies from Psychology, Sociology, Anthropology, rev. ed., ed. L. Bramson and G. Goethals, 21–31. New York: Basic Books. Janis, I. L. (1982) Groupthink, 2nd ed. Boston: Houghton Mifflin. Janis. I. L., and L. Mann (1977) Decision-Making: A Psychological Analysis of Conflict, Choice and Commitment. New York: Free Press. Jepperson, R., A. Wendt, and P. Katzenstein (1996) “Norms, Identity, and Culture in National Security.” In The Culture of National Security, ed. P. Katzenstein, 33–75. New York: Columbia University Press. Jervis, R. (1976) Perception and Misperception in International Politics. Princeton, NJ: Princeton University Press. ——— (1978) “Cooperation under the Security Dilemma.” World Politics 30 (2): 167–214. ——— (1989) “Rational Deterrence: Theory and Evidence.” World Politics 41 (2): 183–207. ——— (1994) “Political Implications of Loss Aversion.” In Avoiding Losses/Taking Risks: Prospect Theory and International Conflict, ed. Barbara Farnham, 23–40. Ann Arbor: University of Michigan Press. ——— (1999) “Realism, Neoliberalism, and Cooperation: Understanding the Debate.” International Security 24 (1): 42–63. ——— (2006) “Understanding Beliefs.” Political Psychology 27 (5): 641–63. ——— (2009) “Deterrence, Rogue States, and the U. S. Policy.” In Complex Deterrence, ed. T. V. Paul, P. Morgan, and J. Wirtz, 133–57. Chicago: University of Chicago Press. Jervis, R., R. N. Lebow, and J. G. Stein (1985) Psychology and Deterrence. Baltimore: Johns Hopkins University Press. Johnson, D. D. P. (2004) Overconfidence and War: The Havoc and Glory of Positive Illusions. Cambridge, MA: Harvard University Press.

Selected Bibliography

577

——— and D. Tierney (2011) “The Rubicon Theory of War: How the Path to Conflict Reaches the Point of No Return.” International Security 36, no. 1 (Summer): 7–40. Kacowicz, A. M. (1995) “Explaining Zones of Peace: Democracies as Satisfied Powers?” Journal of Peace Research 32 (3): 265–76. Kagan, D. (1995) On the Origins of War. New York: Doubleday. Kahneman, D. (2011) Thinking, Fast and Slow. New York: Farrar, Straus and Giroux. Kahneman, D., and A. Tversky (1979) “Prospect Theory: An Analysis of Decision Under Risk.” Econometrica 47: 263–91. Kalinovsky, A. (2009) “Decision-Making and the Soviet War in Afghanistan.” Journal of Cold War Studies 11 (4): 46–73. Kang, D. (2003) “International Relations Theory and the Second Korean War.” International Studies Quarterly 47: 301–24. Kaufmann, C. (2004) “Threat Inflation and the Failure of the Marketplace for Ideas: The Selling of the Iraq War.” International Security 29 (4): 5–48. Kaysen, C. (1990) “Is War Obsolete? International Security 14 (4): 42–64. Keeley, L. (1996) War Before Civilization. New York: Oxford University Press. Kegley, C. W. (1991) The Long Postwar Peace: Contending Explanations and Projections. New York: Harper Collins. Kegley, C. W., and M. G. Hermann (1995) “Rethinking Democracy and International Peace: Perspectives from Political Psychology.” International Studies Quarterly 39: 511–34. ——— (1995) “Military Intervention and the Democratic Peace.” International Interactions 21 (1): 1–21. Kegley, C. W., and G. Raymond (1982) “Alliance Norms and War: A New Piece in an Old Puzzle.” International Studies Quarterly 26: 572–95. ——— (1994) A Multipolar Peace? Great Power Politics in the Twenty-First Century. New York: St. Martins. Kelly, R. C. (2000) Warless Societies and the Origins of War. Ann Arbor: University of Michigan Press. Kelman, H. C. (1965) “Social-Psychological Approaches to the Study of International Relations.” In International Behavior: A Social-Psychological Analysis, ed. H. Kelman, 3–39. New York: Holt, Rinehart & Winston. Kennan, G. F. (1952) American Diplomacy, 1900–1950. New York: Mentor Books. Kennedy, P. (1988) The Rise and Fall of Great Powers: Economic Change and Military Conflict from 1500 to 2000. New York: Random House. Keohane, R. O. (1980) “The Theory of Hegemonic Stability and Changes in International Economic Regimes, 1967–77.” In Change in the International System, ed. O. Holsti, R. Siverson, and A. George, 317–47. Boulder, CO: Westview. Khong, Y. F. (1992) Analogies at War: Korea, Munich, Dien Bien Phu, and the Vietnam Decision of 1965. Princeton, NJ: Princeton University Press. Kim, W. (1989) “Power, Alliance, and Major Wars, 1816–1975.” Journal of Conflict Resolution 32 (2): 255–73. ——— (1991) “Alliance Transitions and Great Power War.” American Journal of Political Science 35: 833–50. ——— (1992) “Power Transitions and Great Power War from Westphalia to Waterloo.” World Politics 45: 153–72. ——— (1996) “Power Parity, Alliance, and War from 1648–1975.” In Parity and War: Evaluations and Extensions of the War Ledger, ed. J. Kugler and D. Lemke, 93–105. Ann Arbor: University of Michigan Press. Kinder, D., and J. Weiss, (1978) “In Lieu of Rationality.” Journal of Conflict Resolution 22 (4): 707–35. Kocs, S. (1995) “Territorial Disputes and Interstate War, 1945–1987.” Journal of Politics 57 (1): 159–75. Kowert, P., and M. G. Hermann (1997) “Who Takes Risks?” Journal of Conflict Resolution 41 (5): 611–37. Krasner, S. (1972) “Are Bureaucracies Important? (Or Allison Wonderland).” Foreign Policy 7: 159–79. Kugler, J. (1984) “Terror without Deterrence: Reassessing the Role of Nuclear Weapons.” Journal of Conflict Resolution 28: 470–506. Kugler, J., and D. Lemke (1989) “The Power Transition Research Program: Assessing Theoretical and Empirical Advances.” In Handbook of War Studies, ed. M. Midlarsky, 129–63. Ann Arbor: University of Michigan Press. ——— (eds.) (1996) Parity and War: Evaluations and Extensions of “The War Ledger.” Ann Arbor: University of Michigan Press. Kugler, J., and A. F. K. Organski (1989) “The Power Transition: A Retrospective and Prospective Evaluation.” In Handbook of War Studies, ed. M. Midlarsky, 171–94. Boston: Unwin Hyman. Kumar, S. (2003) “Power Cycle Analysis of India, China, and Pakistan in Regional and Global Politics.” International Political Science Review 24 (1): 113–22.

578

Selected Bibliography

Lake, D. A. (1992) “Powerful Pacifists: Democratic States and War.” American Political Science Review 86 (1): 24–37. ——— (2010/11) “Two Cheers for Bargaining Theory: Assessing Rationalist Explanations of the Iraq War.” International Security 35 (3): 7–52. Langer, W. (1969) “The Origin of the Russo-Japanese War.” In Explorations in Crisis, ed. C. E. Schorske and E. Schorske, 3–45. Cambridge, MA: Harvard University Press. Larson, D. W. (1994) “The Role of Belief Systems and Schemas in Foreign Policy Decision-Making.” Political Psychology 15 (2): 17–33. Layne, C. (1994) “Kant or Cant: The Myth of the Democratic Peace.” International Security 19 (2): 5–49. Leaky, R. (1981) The Making of Mankind. New York: Dutton. Lebow, R. N. (1981) Between Peace and War: The Nature of International Crises. Baltimore: Johns Hopkins University Press. ——— (1984) “Windows of Opportunity: Do States Jump Through Them?” International Security 9 (1) (Summer): 147–86. ——— (1985) “Miscalculations in the South Atlantic: The Origins of the Falklands War.” In Psychology and Deterrence, ed. R. Jervis, R. N. Lebow, and J. G. Stein, 89–124. Baltimore: Johns Hopkins University Press. Lebow, R. N., and J. G. Stein (1990) “Deterrence: The Elusive Dependent Variable.” World Politics 42: 336–69. Leeds, B. B. (2003) “Do Alliances Deter Aggression? The Influence of Military Alliances on the Initiation of Militarized Interstate Disputes.” American Journal of Political Science 47 (3): 427–39. ——— (2005) “Alliances and the Expansion and Escalation of Militarized Interstate Disputes.” In New Directions in International Relations, ed. A. Mintz and B. Russett, 117–34. Lanham, MD: Lexington Books. Leeds, B. A., and D. R. Davis (1997) “Domestic Political Vulnerability and International Disputes.” Journal of Conflict Resolution 41: 814–34. Legro, J. W., and A. Moravcsik, “Is Anybody Still a Realist?” International Security 24 (2) (Fall): 5–55. Lemke, D. (1995) “The Tyranny of Distance: Redefining Relevant Dyads.” International Interactions 21: 23–38. ——— (1996) “Small States and War: An Expansion of Power Transition Theory.” In Parity and War, ed. J. Kugler and D. Lemke, 77–92. Ann Arbor: University of Michigan Press. Lemke, D., and W. Reed (1996) “Regime Type and Status Quo Evaluations: Power Transition Theory and the Democratic Peace.” International Interactions 22 (2): 143–64. Leng, R. J. (1983) “When Will They Ever Learn? Coercive Bargaining in Recurrent Crises.” Journal of Conflict Resolution 27: 379–419. ——— (1984) “Reagan and the Russians: Crisis Bargaining Beliefs and the Historical Record.” American Political Science Review 78: 338–55. ——— (1988) “Crisis Learning Games.” American Political Science Review 82: 179–94. ——— (1993) Interstate Crisis Behavior, 1816–1980: Realism versus Reciprocity. Cambridge: Cambridge University Press. ——— (2000) “Escalation: Crisis Behavior and War.” In What Do We Know About War? ed. J. A. Vasquez, 235–58. Lanham, MD: Rowman & Littlefield. ——— (2000) Bargaining and Learning in Recurring Crises: The Soviet-American, Egyptian-Israeli, and IndoPakistani Rivalries. Ann Arbor: University of Michigan Press. ——— (2004) “Escalation: Competing Perspectives and Empirical Evidence.” International Studies Review 6: 51–64. Leng, R. J., and C. S. Gochman (1982) “Dangerous Disputes: A Study of Conflict Behavior and War.” American Journal of Political Science 26: 664–87. Leng, R. J., and H. B. Wheeler (1979) “Influence Strategies, Success and War.” Journal of Conflict Resolution 23: 655–84. Lenin, V. I. (1939) Imperialism: the Highest Stage of Capitalism. New York: International Publishers. Levy, J. S. (1983) War in the Modern Great Power System, 1495–1975. Lexington: University Press of Kentucky. ——— (1983) “Misperception and the Causes of War: Theoretical Linkages and Analytical Problems.” World Politics 36 (1): 76–99. ——— (1985) “Theories of General War.” World Politics 37 (3): 344–74. ——— (1986) “Organizational Routines and the Causes of War.” International Studies Quarterly 30 (2): 193–222. ——— (1987) “Declining Power and the Protective Motivation for War.” World Politics 40 (l): 82–107. ——— (1988) “Domestic Politics and War.” In The Origin and Prevention of Major Wars, ed. R. Rotberg and A. Rabb, 79–99. Cambridge: Cambridge University Press.

Selected Bibliography

579

——— (1989) “The Diversionary Theory of War: A Critique.” In Handbook of War Studies, ed. M. Midlarsky, 259–88. Boston: Unwin Hyman. ——— (1992) “Prospect Theory and International Relations: Theoretical Applications and Analytical Problems.” Political Psychology 13 (2): 283–310. Reprinted in Avoiding Losses/Taking Risks: Prospect Theory and International Conflict, ed. B. Farnham, 119–46. Ann Arbor: University of Michigan Press (1994). ——— (1994) “An Introduction to Prospect Theory.” In Avoiding Losses/Taking Risks: Prospect Theory and International Conflict, ed. Barbara Farnham, 7–22. Ann Arbor: University of Michigan Press. ——— (1997) “Prospect Theory, Rational Choice, and International Relations.” International Studies Quarterly 41: 87–112. ——— (2000) “Loss Aversion, Framing Effects, and International Conflict.” In Handbook of War Studies II, ed. M. Midlarsky, 193–221. Ann Arbor: University of Michigan Press. ——— (2008) “Preventive War and Democratic Politics.” International Studies Quarterly 52 (1) (March): 1–24. ——— (2012) “The ‘Paths-to-War’ Concept.” In What Do We Know about War? ed. J. Vasquez, 281–90. Lanham, MD: Rowman & Littlefield. Levy, J. S., and T. C. Morgan (1986) “The War Weariness Hypothesis: An Empirical Test.” American Journal of Political Science 30: 26–50. Levy, J. S., and W. R. Thompson (2010) “Balancing on Land and at Sea: Do States Ally against the Leading Global Power? International Security 35 (1) (Summer): 7–43. ——— (2010) Causes of War. Walden, MA: Wiley-Blackwell. Levy, J. S., and L. Vakili (1993) “Diversionary Action by Authoritarian Regimes: Argentina and the Falklands/ Malvinas Case.” In The Internationalization of Communal Strife, ed. M. Midlarsky, 118–46. New York: Routledge. Lieber, K. (2007) “The New History of World War I and What It Means for International Relations Theory.” International Security 32 (2) (Fall): 155–91. Lindblom, C. (1965) The Intelligence of Democracy. New York: Free Press. Lobell, S. E., N. M. Ripsman, and J. W. Taliaferro (eds.) (2009) Neoclassical Realism, the State and Foreign Policy. Cambridge: Cambridge University Press. Lorenz, K. (1966) On Aggression. New York: Bantam. Luard, E. (1986) War in International Society. New Haven, CT: Yale University Press. Macfie, A. L. (1938) “The Outbreak of War and the Trade Cycle.” Economic History 3: 89–97. Mansfield, E. and J. Snyder (1995) “Democratization and the Danger of War.” International Security 20 (1): 5–38. ——— (2002) “Incomplete Democratization and the Outbreak of Militarized Disputes.” International Studies Quarterly 46 (4): 529–49. ——— (2005) Electing to Fight. Cambridge, MA: MIT Press. ——— (2009) “Pathways to War in Democratic Transition.” International Organization 63 (Spring): 381–90. Maoz, Z. (1989) “Joining the Club of Nations: Political Development and International Conflict, 1816–1976.” International Studies Quarterly 32 (2): 199–231. ——— (1997) “The Controversy over the Democratic Peace: Rearguard Action or Cracks in the Wall?” International Security 22 (1): 162–98. ——— (2004) “Pacifism and Fightaholism in International Politics: A Structural History of National and Dyadic Conflict, 1816–1992.” International Studies Review 6: 107–33. Maoz, Z., and N. Abdolali (1989) “Regime Type and International Conflict, 1816–1976.” Journal of Conflict Resolution 33 (1): 3–35. Maoz, Z. and B. Russett (1992) “Alliance, Contiguity, Wealth, and Political Stability: Is Lack of Conflict Among Democracies a Statistical Artifact?” International Interactions 17: 245–67. ——— (1993) “Normative and Structural Causes of Democratic Peace, 1946–1986.” American Political Science Review 87: 624–38. March, J. G., and H. Simon (1958) Organizations. New York: Wiley. Maslow, A. (1954) Motivation and Personality. New York: Harper & Row. May, E. (1973) “Lessons” of the Past: The Use and Misuse of History in American Foreign Policy. New York: Oxford University Press. Mayer, A. J. (1969) “Internal Causes and Purposes of War in Europe, 1870–1956.” Journal of Modern History 41: 291–303. McCauley, C. (1989) “The Nature of Social Influence in Groupthink: Compliance and Internalization.” Journal of Personality and Social Psychology 57: 250–60.

580

Selected Bibliography

McDermott, R. (1998) Risk-Taking in International Politics: Prospect Theory in American Foreign Policy. Ann Arbor: University of Michigan Press. ——— (2004) “The Feeling of Rationality: The Meaning of Neuroscientific Advances for Political Science.” Perspectives on Politics 2 (4): 691–706. ——— (2009) “Mutual Interests: The Case for Increasing Dialogue between Political Science and Neuroscience.” Political Research Quarterly 62 (3): 571–83. McInerney, A. (1994) “Prospect Theory and Soviet Policy Towards Syria, 1966–1967.” In Avoiding Losses/Taking Risks, ed. B. Farnham, 101–18. Ann Arbor: University of Michigan Press. Mead, M. (1973) “Warfare Is Only an Invention—Not Biological Necessity.” In Peace and War, ed. C. Beitz and T. Herman, 112–18. San Francisco: W. H. Freeman. Mearsheimer, J. (1990) “Back to the Future: Instability in Europe after the Cold War.” International Security 15 (l): 5–56. ——— (2001) The Tragedy of Great Power Politics. New York: Norton. Meernik, J. (1994) “Presidential Decision Making and the Political Use of Force.” International Studies Quarterly 38: 121–38. Meernik, J., and P. Waterman (1996) “The Myth of Diversionary Use of Force by American Presidents.” Political Research Quarterly 49: 573–90. Mercer, J. (1996) Reputation and International Politics. Ithaca: Cornell University Press. ——— (2005) “Rationality and Psychology in International Politics.” International Organization 59: 77–106. ——— (2005) “Prospect Theory and Political Science.” Annual Review of Political Science 8: 1–21. Midlarsky, M. (1975) On War. New York: Free Press. Miller, B. (2007) States, Nations and the Great Powers: The Sources of Regional War and Peace. Cambridge: Cambridge University Press. Miller, G. (2012) The Shadow of the Past: Reputation and Military Alliances before the First World War. Ithaca, NY: Cornell University Press. Miller, R. (1995) “Domestic Structures and the Diversionary Use of Force.” American Journal of Political Science 39: 760–85. ——— (1999) “Regime Type, Strategic Interaction, and the Diversionary Use of Force.” Journal of Conflict Resolution 43 (3): 388–402. Mintz, A. (1993) “The Decision to Attack Iraq: A Noncompensatory Theory of Decision Making.” Journal of Conflict Resolution 37 (4): 595–618. ——— (2004) “How Do Leaders Make Decisions? A Poliheuristic Perspective.” Journal of Conflict Resolution 48 (1): 3–13. Mintz, A., and K. DeRouen, Jr. (2010) Understanding Foreign Policy Decision Making. Cambridge: Cambridge University Press. Mitani, J. C., D. P. Watts, and S. J. Amsler (2010) “Lethal Intergroup Aggression Leads to Territorial Expansion in Wild Chimpanzees.” Current Biology 20 (12): 507–8. Mitchell, D., and T. G. Massoud (2009) “Anatomy of Failure: Bush’s Decision-Making Process and the Iraq War.” Foreign Policy Analysis 5 (3): 265–86. Mitchell, S. M. (2012) “Norms and the Democratic Peace.” In What Do We Know About War? 2nd ed., ed. J. Vasquez, 167–88. Lanham, MD: Rowman & Littlefield. Mitchell, S. M., and B. C. Prins (2004) “Rivalry and Diversionary Uses of Force.” Journal of Conflict Resolution 48 (6): 937–61. Modelski, G. (1978) “The Long Cycle of Global Politics and the Nation-State.” Comparative Studies in Society and History 20 (2): 214–35. Modelski, G., and P. Morgan (1985) “Understanding Global War.” Journal of Conflict Resolution 29 (3): 391–417. Modelski, G., and W. R. Thompson (1989) “Long Cycles and Global War.” In Handbook of War Studies, ed. M. Midlarsky, 23–54. Boston: Unwin Hyman. Morgan, P. M. (1985) “Saving Face for the Sake of Deterrence.” In Psychology and Deterrence, ed. R. Jervis, R. N. Lebow, and J. G. Stein, 125–52. Baltimore: Johns Hopkins University Press. ——— (2003) Deterrence Now. Cambridge: Cambridge University Press. Morgan, P. M., and T. V. Paul (2009) “Deterrence among Great Powers in an Era of Globalization.” In Complex Deterrence, ed. T. V. Paul, P. Morgan, and J. Wirtz, 259–76. Chicago: University of Chicago Press. Morgan, T. C., and K. N. Bickers (1992) “Domestic Discontent and the External Use of Force.” Journal of Conflict Resolution 36 (1): 25–52.

Selected Bibliography

581

Morgan, T. C., and V. L. Schwebach (1992) “Take Two Democracies and Call Me in the Morning: A Prescription for Peace?” International Interactions 17 (4): 305–20. Morrow, J. D. (1989) “A Twist of Truth: A Reexamination of the Effects of Arms Races on the Occurrence of War.” Journal of Conflict Resolution 33 (3): 500–29. Most, B., P. Schrodt, R. Siverson, and H. Starr (1990) “Border and Alliance Effects in the Diffusion of Major Power Conflict, 1816–1965.” In Prisoners of War? Nation-States in the Modern Era, ed. C. Gochman and A. N. Sabrosky, 209–29. Lexington, MA: Lexington Books. Most, B., and H. Starr (1980) “Diffusion, Reinforcement, Geopolitics and the Spread of War.” American Political Science Review 74: 932–46. Moul, W. (1988) “Balance of Power and the Escalation of Serious Disputes among European Great Powers, 1815–1939: Some Evidence.” American Journal of Political Science 32: 241–75. ——— (1988) “Great Power Nondefense Alliances and Escalation to War of Conflicts between Unequals, 1815–1939.” International Interactions 15 (1): 25–43. Mousseau, M. (2012) “A Market-Capitalist or a Democratic Peace?” In What Do We Know About War?, 2nd ed., ed. J. Vasquez, 189–210. Lanham, MD: Rowman & Littlefield. Mueller, J. (1989) Retreat from Doomsday: The Obsolescence of Major War. New York: Basic Books. ——— (1991) “Changing Attitudes Towards War: The Impact of the First World War.” British Journal of Political Science 21: 1–28. Muller, H., and J. Wolff (2006) “Democratic Peace: Many Data, Little Explanation?” In Democratic Wars: Looking at the Dark Side of Democratic Peace, ed. A. Geis, L. Brock, and H. Muller, 41–73. New York: Palgrave Macmillan. Myers, D. G., and H. Lamm (1976) “The Group Polarization Phenomenon.” Psychological Bulletin 83: 602–27. ——— (1977) “The Polarizing Effect of Group Discussion.” In Current Trends in Psychology: Readings from the American Scientist, ed. I. Janis. Los Altos, CA: Kaufmann. Narang, V., and R. M. Nelson (2009) “Who are These Belligerent Democratizers? Reassessing the Impact of Democratization on War.” International Organization 63 (Spring): 357–79. Neumann, I. (1998) “Identity and the Outbreak of War.” The International Journal of Peace Studies 3 (1): 7–23. Nevin, J. A. (1996) “War Initiation and Selection by Consequences.” Journal of Peace Research 33 (1): 99–108. Nish, I. (1985) The Origins of the Russo-Japanese War. London: Longman. North, R. C. (1967) “Perception and Action in the 1914 Crisis” In Image and Reality in World Politics, ed. J. C. Farrell and A. P. Smith, 103–22. New York: Columbia University Press. North, R. C., R. Brody, and O. Holsti (1964) “Some Empirical Data on the Conflict Spiral.” Peace Research Society (International) 1: 1–15. Oneal, J., and B. Russett (1997) “The Classic Liberals Were Right: Democracy, Interdependence, and Conflict, 1950–1985.” International Studies Quarterly 41: 267–94. Onuf, N. (1989) World of Our Making: Rules and Rule in Social Theory and International Relations. Columbia: University of South Carolina Press. Organski, A. F. K. (1958) World Politics. New York: Knopf. Organski, A. F. K., and J. Kugler (1980) The War Ledger. Chicago: University of Chicago Press. Orme, J. (1986–1987) “Deterrence Failures: A Second Look.” International Security 11: 96–124. Osgood, C. E. (1962) An Alternative to War or Surrender. Urbana: University of Illinois Press. ——— (1971) “Graduated Unilateral Initiatives for Peace.” In Conflict Resolution: Contributions from the Behavioral Sciences, ed. C. G. Smith, 515–25. Notre Dame, IN: Notre Dame University Press. Ostrom, C. W., and B. L. Job (1986) “The President and the Political Use of Force.” American Political Science Review 80: 554–66. Otterbein, K. F. (2009) The Anthropology of War. Long Grove, IL: Waveland Press. Owen, J. M. (1994) “How Liberalism Produces Democratic Peace.” International Security 19 (2): 87–125. ——— (1997) “Perceptions and the Limits of Liberal Peace: The Mexican-American and Spanish-American Wars.” In Paths to Peace: Is Democracy the Answer? ed. Miriam Fendius Elman, 153–89. Cambridge, MA: MIT Press. Owsiak, A. P. (2012) “Signing Up for Peace: International Boundary Agreements, Democracy, and Militarized Interstate Conflict.” International Studies Quarterly 56 (1) (March): 51–66. Oye, K. (1985) “Explaining Cooperation under Anarchy: Hypotheses and Strategies.” World Politics 38 (1): 1–24. Parasiliti, A. T. (2003) “The Causes and Timing of Iraq’s Wars: A Power Cycle Assessment.” International Political Science Review 24 (1): 151–65. Park, W. (1990) “A Review of Research on Groupthink.” Journal of Behavioral Decision Making 3: 229–45.

582

Selected Bibliography

Patchen, M. (1987) “Strategies for Eliciting Cooperation from an Adversary: Laboratory and International Findings.” Journal of Conflict Resolution 31: 164–85. Paul, T. V. (1994) Asymmetric Conflicts: War Initiation by Weaker Powers. Cambridge: Cambridge University Press. ——— (1995) “Nuclear Taboo and War Initiation in Regional Conflicts.” Journal of Conflict Resolution 39 (4): 696–717. ——— (2005) “Causes of the India-Pakistan Enduring Rivalry.” In The India-Pakistan Conflict: An Enduring Rivalry, ed. T. V. Paul, 3–26. Cambridge: Cambridge University Press. ——— (2009) The Tradition of Non-Use of Nuclear Weapons. Stanford, CA: Stanford University Press. Peceny, M. (1997) “A Constructivist Interpretation of the Liberal Peace: The Ambiguous Case of the Spanish American War.” Journal of Peace Research 34 (4): 415–30. Perlmutter, A. (1974) “The Presidential Political Center and Foreign Policy: A Critique of the Revisionist and Bureaucratic-Political Orientations.” World Politics 27 (1): 87–106. Pickering, J., and E. Kisangani (2005) “Democracy and Diversionary Military Intervention: Reassessing Regime Type and the Diversionary Hypothesis.” International Studies Quarterly 49: 23–43. Pilisuk, M., and P. Skolnick (1968) “Inducing Trust: A Test of the Osgood Proposal.” Journal of Personality and Social Psychology 8: 122–33. Pollard, K. S. (2009) “What Makes Us Human?” Scientific American (May 2009): 44–49. Post, J. (1991) “The Impact of Crisis-Induced Stress on Policy Makers.” In Avoiding War: Problems of Crisis Management, 471–94. Boulder: Westview. ——— (2005) “Assessing Leaders at a Distance: The Political Personality Profile.” In The Psychological Assessment of Political Leaders, ed. J. Post, 69–104. Ann Arbor: University of Michigan Press. ——— (2005) “Saddam Hussein of Iraq: A Political Psychology Profile.” In The Psychological Assessment of Political Leaders, ed. J. Post, 335–66. Ann Arbor: University of Michigan Press. Powell, R. (2002) “Bargaining Theory and International Conflict.” Annual Review of Political Science 5: 1–30. ——— (2006) “War as a Commitment Problem.” International Organization 60: 169–203. Press, D. (2205) Calculating Credibility: How Leaders Assess Military Threats. Ithaca, NY: Cornell University Press. Preston, T. (2008) “‘Following the Leader’: The Impact of U.S. Presidential Style upon Advisory Group Dynamics, Structure, and Decision.” In Beyond Groupthink, ed. P. ‘t Hart, E. Stern, and B. Sundelius, 191–248. Ann Arbor: University of Michigan Press. Price, R. (1995) “A Genealogy of the Chemical Weapons Taboo.” International Organization 49 (1): 73–103. ——— (1997) The Chemical Weapons Taboo. Ithaca, NY: Cornell University Press. Putnam, R. (1988) “Diplomacy and Domestic Politics: The Logic of Two-Level Games.” International Organization 42 (3): 427–60. Rasler, K., and W. R. Thompson (1983) “Global Wars, Public Debts, and the Long Cycle.” World Politics 35 (4): 489–516. ——— (1994) The Great Powers and Global Struggle 1490–1990. Lexington: University Press of Kentucky. ——— (2000) “Global War and the Political Economy of Structural Change.” In Handbook of War Studies II, ed. Manus Midlarsky, 301–31. Ann Arbor: University of Michigan Press. ——— (2006) “Contested Territory, Strategic Rivalries, and Conflict Escalation.” International Studies Quarterly 50 (1): 145–67. Rauchhaus, R. (2009) “Evaluating the Nuclear Peace Hypothesis: A Quantitative Approach.” Journal of Conflict Resolution 53 (2): 258–77. Ray, J. L. (1974) “Status Inconsistency and War Involvement in Europe, 1816–1970.” Peace Science Society (International) Papers 23: 69–80. ——— (1989) “The Abolition of Slavery and the End of International War.” International Organization 43: 405–39. ——— (1993) “Wars Between Democracies: Rare or Nonexistent?” International Interactions 18: 251–76. ——— (1995) Democracy and International Conflict: An Evaluation of the Democratic Peace Proposition. Columbia: University of South Carolina Press. ——— (1998) “Does Democracy Cause Peace?” Annual Review of Political Science 1: 27–46. ——— (2000) “Democracy: On the Level (s), Does Democracy Correlate with Peace?” in What Do We Know About War? ed. J. Vasquez, 299–316. Lanham, MD: Rowman & Littlefield.

Selected Bibliography

583

Raymond, G. (1994) “Democracies, Disputes, and Third-Party Intermediaries.” Journal of Conflict Resolution 38 (1): 24–42. ——— (2000) “International Norms: Normative Orders and Peace.” In What Do We Know About War? ed. J. Vasquez, 281–97. Lanham, MD: Rowman & Littlefield. Redd, S. B. (2005) “The Influence of Advisers and Decision Strategies on Foreign Policy Choices: President Clinton’s Decision to Use Force in Kosovo.” International Studies Perspective 6: 129–50. Regan, P. M. (1998) “Choosing to Intervene: Outside Interventions in Internal Conflicts.” Journal of Politics 60 (3): 754–79. Reiter, D. (1995) “Exploding the Powder Keg Myth: Preemptive Wars Almost Never Happen.” International Security 20 (2): 5–34. Reiter, D., and A. Stam (1998) “Democracy and Battlefield Military Effectiveness.” Journal of Conflict Resolution 42: 259–77. ——— (1998) “Democracy, War Initiation and Victory.” American Political Science Review 92 (2): 377–90. ——— (2002) Democracies at War. Princeton, NJ: Princeton University Press. Renshon, J. (2006) Why Leaders Choose War: The Psychology of Prevention. Westport, CT: Praeger. ——— (2008) “Stability and Change in Belief Systems: The Operational Code of George W. Bush.” Journal of Conflict Resolution 52 (6): 820–49. Renshon, J., and S. Renshon (2008) “The Theory and Practice of Foreign Policy Decision Making.” Political Psychology 29 (4): 509–36. Renshon, S. (1993) “The Gulf War Revisited.” In The Political Psychology of the Gulf War, ed. Stanley Renshon, 329–57. Pittsburgh: University of Pittsburgh Press. Reyna, S. P., and R. E. Downs, eds. (1994) Studying War: Anthropological Perspectives. Langhorne, PA: Gordon and Breach. Rhodes, E. (1994) “Do Bureaucratic Politics Matter? Some Disconfirming Findings from the Case of the U.S. Navy.” World Politics 47: 1–41. Richardson, L. F. (1960) Statistics of Deadly Quarrels. Chicago: Quadrangle. ——— (1960) Arms and Insecurity. Pittsburgh: Boxwood Press. Rider, T., M. Findley, and P. F. Diehl (2011) “Just Part of a Game? Arms Races, Rivalry, and War.” Journal of Peace Research 48: 85–100. Rohl, J. C. G. (1987). The Kaiser and His Court. Cambridge: Cambridge University Press. Rokeach, M. (1960) The Open and Closed Mind. New York: Basic Books. Rosati, J. (1981) “Developing a Systematic Decision-Making Framework: Bureaucratic Politics in Perspective.” World Politics 33 (2): 234–52. ——— (2000) “The Power of Human Cognition in the Study of World Politics.” International Studies Review 2 (3): 45–75. Rosato, S. (2003) “The Flawed Logic of Democratic Peace Theory.” American Political Science Review 97 (4): 585–602. Rosecrance, R. (1963) Action and Reaction in World Politics. Boston: Little, Brown. ——— (1969) “Bipolarity, Multipolarity, and the Future.” In International Politics and Foreign Policy, rev. ed., ed. J. Rosenau, 325–35. New York: Free Press. Ross, D. (1980) “Coalition Maintenance in the Soviet Union.” World Politics 32 (2): 258–80. ——— (1984) “Risk Aversion in Soviet Decisionmaking.” In Soviet Decisionmaking for National Security, ed. J. Valenta and W. Potter, 237–51. Boston: Allen and Unwin. Rousseau, D., C. Gelpi, D. Reiter, and P. Huth (1996) “Assessing the Dyadic Nature of the Democratic Peace, 1918–88.” American Political Science Review 90 (3): 512–533. Rummel, R. J. (1979) Understanding Conflict and War, Volume 4: War, Power and Peace. Beverly Hills, CA: Sage. ——— (1983) “Libertarianism and International Violence.” Journal of Conflict Resolution 27 (l): 27–71. ——— (1985) “Libertarian Propositions on Violence Within and Between Nations: A Test Against Published Research Results.” Journal of Conflict Resolution 29 (l): 419–55. Russett, B. (1969) “The Calculus of Deterrence.” In International Politics and Foreign Policy, rev. ed., ed. J. Rosenau, 359–69. New York: Free Press. ——— (1990) “Economic Decline, Electoral Pressure and the Initiation of Interstate Conflict.” In Prisoners of War? Nation-States in the Modern Era, ed. C. Gochman and A. N. Sabrosky, 123–40. Lexington, MA: Lexington Books. ——— (1993) Grasping the Democratic Peace. Princeton, NJ: Princeton University Press.

584

Selected Bibliography

——— (1996) “The Fact of Democratic Peace.” In Debating the Democratic Peace, ed. M. Brown, S. Lynn-Jones, and S. Miller, 58–81. Cambridge, MA: MIT Press. Russett, B., and J. Oneal (2001) Triangulating Peace: Democracy, Interdependence, and International Organizations. New York: Norton. Russett, B., and H. Starr (2000) “From Democratic Peace to Kantian Peace: Democracy and Conflict in the International system.” In Handbook of War Studies II, ed. M. Midlarsky, 93–128. Ann Arbor: University of Michigan. Sabrosky, A. N. (ed.) (1985) Polarity and War. Boulder, CO: Westview. Sagan, S. (1988) “The Origins of the Pacific War.” In The Origin and Prevention of Major War, ed. R. Rotberg and T. Rabb, 323–52. Cambridge: Cambridge University Press. Sahlins, M. (1976) The Use and Abuse of Biology: An Anthropological Critique of Sociobiology. Ann Arbor: University of Michigan Press. Sample, S. G. (1997) “Arms Races and Dispute Escalation: Resolving the Debate.” Journal of Peace Research 34: 7–22. ——— (2000) “Military Buildups: Arming and War.” In What Do We Know About War? ed. J. Vasquez, 165–95. Lanham, MD: Rowman & Littlefield. ——— (2002) “The Outcomes of Military Buildups: Minor States vs. Major Powers.” Journal of Peace Research 39: 669–92. ——— (2012) “Arms Races: A Cause or a Symptom?” In What Do We Know About War? ed. J. Vasquez, 111–38. Lanham, MD: Rowman & Littlefield. Sathasivam, K. (2003) “‘No Other Choice’: Pakistan’s Decision to Test the Bomb.” In Integrating Cognitive and Rational Theories of Foreign Policy Decision Making, ed. A. Mintz, 55–76. New York: Palgrave Macmillan. Schelling, T. (1963) The Strategy of Conflict. New York: Oxford University Press: Galaxy Books. Schmookler, A. B. (1984) The Parable of the Tribes: The Problem of Power in Social Evolution. Boston: Houghton Mifflin. Schweller, R. (1992) “Domestic Structure and Preventive War: Are Democracies More Pacific?” World Politics 44: 235–69. ——— (1994) “Bandwagoning for Profit: Bringing the Revisionist State Back In.” International Security 19 (1): 72–107. ——— (1996) “Neorealism’s Status-Quo Bias: What Security Dilemma?” Security Studies 5 (3): 90–121. Sechser, T. S. (2004) “Are Soldiers Less War-Prone Than Statesmen?” Journal of Conflict Resolution 48 (5) (October): 746–74. Semmel, A. K. (1976) “Some Correlates of Attitudes to Multilateral Diplomacy in the United States Department of State.” International Studies Quarterly 20 (2): 301–24. ——— (1982) “Small Group Dynamics in Foreign Policymaking: A Comparative Analysis.” In Biopolitics, Political Psychology, and International Politics, ed. G. Hopple, 94–113. New York: St. Martin’s. Senese, P. D., and J. A. Vasquez (2003) “A Unified Explanation of Territorial Conflict: Testing the Impact of Sampling Bias, 1919–1992.” International Studies Quarterly 47 (2): 275–98. ——— (2008) The Steps to War: An Empirical Study. Princeton, NJ: Princeton University Press. Shepard, G. H. (1968) “Personality Effects on American Foreign Policy, 1969–1984: A Second Test of Interpersonal Generalization Theory.” International Studies Quarterly 32 (l): 91–123. Singer, J. D. (1972) “The Correlates of War Project: An Interim Report and Rationale.” World Politics 24: 243–70. ———, ed. (1980) The Correlates of War II: Testing Some Realpolitik Models. New York: Free Press. Singer, J. D., S. Bremer, and J. Stuckey (1972) “Capability Distribution, Uncertainty, and Major Power War, 1820–1965.” In Peace, War and Numbers, ed. B. Russett, 19–48. Beverly Hills, CA: Sage. Singer, J. D., and M. Small (1966) “Formal Alliances, 1815–1939: A Quantitative Description.” Journal of Peace Research 3: 1–32. ——— (1966) “National Alliance Commitments and War Involvement, 1815–1945.” Peace Research Society (International) Papers 5: 109–40. ——— (1967) “Alliance Aggregation and the Onset of War, 1815–1945” In Quantitative International Politics, ed. J. D. Singer, 246–86. New York: Free Press. ——— (1972) The Wages of War, 1816–1965: A Statistical Handbook. New York: Wiley. Siverson, R. M. (1995) “Democracies and War Participation: In Defense of the Institutional Constraints Argument.” European Journal of International Relations 1 (4): 481–89. Siverson, R. M., and P. Diehl (1989) “Arms Races, the Conflict Spiral, and the Onset of War.” In Handbook of War Studies, ed. M. Midlarsky, 195–218. Boston: Unwin Hyman.

Selected Bibliography

585

Siverson, R. M., and H. Starr (1990) “Opportunity, Willingness and the Diffusion of War, 1816–1965.” American Political Science Review 84: 47–67. Small, M., and J. D. Singer (1970) “Patterns in International Warfare, 1816–1965.” Annals of the American Academy of Political and Social Sciences 391: 145–55. ——— (1976) “The War Proneness of Democratic Regimes.” Jerusalem Journal of International Relations 1: 49–69. ——— (1982) Resort to Arms: International and Civil Wars, 1816–1980. Thousand Oaks, CA: Sage. Smith, A. (1996) “Diversionary Foreign Policy in Democratic Systems.” International Studies Quarterly 40: 133–53. Smith, M. (2008) “US Bureaucratic Politics and the Decision to Invade Iraq.” Contemporary Politics 14 (1) (March): 91–105. Smith, S. (1984) “Groupthink and the Hostage Rescue Mission.” British Journal of Political Science 15: 453–58. ——— (1984–85) “Policy Preferences and Bureaucratic Position: The Case of the American Hostage Rescue Mission.” International Affairs 61 (1) (Winter): 9–25. Smith, T. C. (1980) “Arms Race Instability and War.” Journal of Conflict Resolution 24: 253–84. ——— (1988) “Curvature Change and War Risk in Arming Patterns.” International Interactions 14: 201–28. Snyder, G. H., and P. Diesing (1977) Conflict Among Nations: Bargaining, Decision-making, and System Structure in International Crises. Princeton, NJ: Princeton University Press. Snyder, J. L. (1985) “Perceptions of the Security Dilemma in 1914.” In Psychology and Deterrence, ed. R. Jervis, R. N. Lebow, and J. G. Stein, 153–79. Baltimore: Johns Hopkins University Press. ——— (1991) Myths of Empire: Domestic Politics and International Ambition. Ithaca, NY: Cornell University Press. Spiezio, K. E. (1990) “British Hegemony and Major Power War, 1815–1939: An Empirical Test of Gilpin’s Model of Hegemonic Governance.” International Studies Quarterly 34 (2): 165–81. Spiro, D. E. (1994) “The Insignificance of the Liberal Peace.” International Security 19 (2): 50–86. Sprout, H., and M. Sprout (1965) The Ecological Perspective on Human Affairs. Princeton, NJ: Princeton University Press. Stam, A. C. (1996) Win, Lose, or Draw: Domestic Politics and the Crucible of War. Ann Arbor: University of Michigan Press. Starr, H. (1978) “‘Opportunity’ and ‘Willingness’ as Ordering Concepts in the Study of Wars.” International Interactions 4: 363–87. ——— (1984) Henry Kissinger: Perceptions of International Politics. Lexington: University Press of Kentucky. Starr, H., and G. D. Thomas (2005) “The Nature of Borders and International Conflict: Revisiting Hypotheses on Territory.” International Studies Quarterly 49 (1): 123–39. Starr, H., and B. Most (1978) “A Return Journey: Richardson, ‘Frontiers,’ and Wars in the 1946–1965 Era.” Journal of Conflict Resolution 22: 441–67. ——— (1983) “Contagion and Border Effects on Contemporary African Conflict.” Comparative Political Studies 16: 92–117. Stein, J. G. (1985) “Calculation, Miscalculation, and Conventional Deterrence 1: The View from Cairo.” In Psychology and Deterrence, ed. R. Jervis, R. N. Lebow, and J. G. Stein, 34–59. Baltimore: Johns Hopkins University Press. ——— (1987) “Extended Deterrence in the Middle East: American Strategy Reconsidered.” World Politics 39 (3): 326–52. ——— (1991) “Deterrence and Reassurance.” In Behavior, Society, and Nuclear War, vol. II, ed. P. E. Tetlock et al. New York: Oxford University Press. ——— (2008) “Foreign Policy Decision-Making: Rational, Psychological, and Neurological Models.” In Foreign Policy: Theories-Actors-Cases, ed. Steve Smith, Amelia Hadfield, and Tim Dunne, 102–14. Oxford: Oxford University Press. ——— (2009) “Rational Deterrence against ‘Irrational’ Adversaries? No Common Knowledge.” In Complex Deterrence: Strategy in the Global Age, ed. T. V. Paul, P. M. Morgan, and J. J. Wirtz, 58–82. Chicago: University of Chicago Press. Steinbruner, J. (1974) The Cybernetic Theory of Decision. Princeton, NJ: Princeton University Press. Steiner, M. (1977) “The Elusive Essence of Decision.” International Studies Quarterly 21 (2): 389–422. Stern, E. (2008) “Probing the Plausibility of Newgroup Syndrome: Kennedy and the Bay of Pigs.” In Beyond Groupthink, ed. P. ‘t Hart, E. Stern, and B. Sundelius, 153–89. Ann Arbor: University of Michigan Press.

586

Selected Bibliography

Stern, E., and B. Sundelius (1994) “The Essence of Groupthink.” Mershon International Studies Review 38: 101–7. ——— (2008) “Understanding Small Group Decisions in Foreign Policy: Process Diagnosis and Research Procedure.” In Beyond Groupthink, ed. P. ‘t Hart, E. Stern, and B. Sundelius, 123–50. Sullivan, C. J. (2011) “The Kremlin and Kabul: The 1979 Soviet Invasion of Afghanistan in Retrospect.” Washington Review of Turkish and Eurasian Affairs. http: //www.thewashingtonreview.org/articles/the-kremlin-and-kabul-the-1979-soviet-invasion-in-retrospect.html. Taliaferro, J. W. (2000/01) “Security Seeking under Anarchy.” International Security 25 (3): 128–61. ——— (2004) Balancing Risks: Great Power Intervention in the Periphery. Ithaca, NY: Cornell University Press. Tammen, R., J. Kugler, D. Lemke, A. Stam, C. Alsharabati, M. Abdollahian, B. Efird, and A. F. K. Organski. (2000) Power Transitions: Strategies for the 21st Century. New York: Chatham House. Tannenwald, N. (2005) “Stigmatizing the Bomb: Origins of the Nuclear Taboo.” International Security 29 (4): 5–49. ——— (2007) The Nuclear Taboo. Cambridge: Cambridge University Press. Tessman, B. F., and S. Chan (2004) “Power Cycles, Risk Propensity, and Great-Power Deterrence.” Journal of Conflict Resolution 48 (2): 131–53. Tetlock, P. E. (2005) Expert Political Judgment: How Good Is It? How Can We Know? Princeton. NJ: Princeton University Press. Tetlock, P. E., R. S. Peterson, C. McClure, S. Chang, and P. Feld (1992) “Assessing Political Group Dynamics: A Test of the Groupthink Model.” Journal of Personality and Social Psychology 63: 403–25. Thayer, B. (2000) “Bringing in Darwin: Evolutionary Theory, Realism, and International Politics.” International Security 25 (2): 124–51. Thomas, E. (1959) The Harmless People. New York: Knopf. Thompson, W. R. (1982) “Phases of the Business Cycle and the Outbreak of War.” International Studies Quarterly 26: 301–11. ——— (1988) On Global War: Historical-Structural Approaches to World Politics. Columbia: University of South Carolina Press. ——— (1996) “Balances of Power, Transitions, and Long Cycles.” In Parity and War, ed. J. Kugler and D. Lemke, 163–85. Ann Arbor: University of Michigan Press. ——— (2001) “Identifying Rivals and Rivalries in World Politics,” International Studies Quarterly 45: 557–86. ——— (2006) “Systemic Leadership, Evolutionary Processes, and International Relations Theory: The Unipolarity Question.” International Studies Review 8: 1–22. Thompson, W. R., and K. A. Rasler (1988) “War and Systemic Capability Reconcentration.” Journal of Conflict Resolution 32: 335–66. Thompson, W. R., and R. M. Tucker (1997) “A Tale of Two Democratic Peace Critiques.” Journal of Conflict Resolution 4 (2): 428–54. Thomson, J. C. (1973) “How Could Vietnam Happen? An Autopsy.” In Readings in American Foreign Policy: A Bureaucratic Perspective, ed. M. Halperin and A. Kantor, 98–110. Boston: Little, Brown. Thorpe, I. J. N. (2003) “Anthropology, Archaeology, and the Origin of Warfare.” World Archaeology 35 (1): 145–65. ——— (2005) “The Ancient Origins of Warfare and Violence.” In Warfare, Violence and Slavery in Prehistory, ed. I. J. N. Thorpe, 1–18. Oxford: British Archaeological Reports. Tiger, L., and R. Fox (1971) The Imperial Animal. New York: Holt, Rinehart Winston. Tir, J. (2005) “Keeping the Peace after Secession.” Journal of Conflict Resolution 49 (5): 713–41. Tir, J., and P. F. Diehl (1998) “Demographic Pressure and Interstate Conflict: Linking Population Growth and Density to Militarized Disputes and Wars, 1930–1989.” Journal of Peace Research 35 (3): 319–39. ——— (2002) “Geographic Dimensions of Enduring Rivalries.” Political Geography 2 (February): 263–86. Tir, J., and M. Jasinksi (2008) “Domestic-Level Diversionary Theory of War: Targeting Ethnic Minorities.” Journal of Conflict Resolution 52 (5) (October): 641–64. Toynbee, A. (1954) A Study of History, vol. IX. London: Oxford University Press. Tuchman, B. (1962) The Guns of August. New York: Dell. Tucker, R. (1973) Stalin as Revolutionary: 1879–1929; A Study in History and Personality. New York: Norton. Valenta, J. (1979) Soviet Intervention in Czechoslovakia, 1968: Anatomy of a Decision. Baltimore: Johns Hopkins University Press. ——— (1984) “Soviet Decisionmaking on Afghanistan.” In Soviet Decisionmaking for National Security, ed. J. Valenta and W. Potter, 218–36. Boston. Allen and Unwin.

Selected Bibliography

587

Valeriano, B. (2012) “Becoming Rivals: The Process of Rivalry Development.” In What Do We Know About War? 2nd ed., ed. John A. Vasquez, 63–82. Lanham, MD: Rowman & Littlefield. Valeriano, B., and J. Vasquez (2010) “Identifying and Classifying Complex Interstate Wars.” International Studies Quarterly 54: 561–82. Van Evera, S. (1984) “The Cult of the Offensive and the Origins of World War I.” International Security 9: 58–107. ——— (1985) “Why Cooperation Failed in 1914.” World Politics 38: 80–117. ——— (1994) “Hypotheses on Nationalism and War.” International Security 18 (4): 5–39. ——— (1998) “Offense, Defense, and the Causes of War.” International Security 22 (4): 5–43. ——— (1999) Causes of War: Power and the Roots of Conflict. Ithaca, NY: Cornell University Press. Vasquez, J. A. (1987) “The Steps to War: Toward a Scientific Explanation of Correlates of War Findings.” World Politics 50 (l): 108–45. ——— (1993) The War Puzzle. Cambridge: Cambridge University Press. ——— (1996) “Distinguishing Rivals That Go to War from Those that Do Not: A Quantitative Comparative Case Study of the Two Paths to War,” International Studies Quarterly 40 (4): 531–58. ——— (2004) “The Probability of War, 1816–1992,” International Studies Quarterly 48: 1–27. ——— (2005) “The India-Pakistan Conflict in Light of General Theories of War, Rivalry, and Deterrence.” In The India-Pakistan Conflict: An Enduring Rivalry, 54–79. Cambridge: Cambridge University Press. ——— (2009) The War Puzzle Revisited. Cambridge: Cambridge University Press. Vasquez, J. A., and M. T. Henehan (2001) “Territorial Disputes and the Probability of War, 1816–1992,” Journal of Peace Research 38: 123–35. Vasquez, J. A., and B. Valeriano (2010) “Classification of Interstate Wars.” Journal of Politics 72: 292–309. Vertzberger, Y. (1998) Risk Taking and Decisionmaking. Stanford, CA: Stanford University Press. de Waal, F. (1989) Peacemaking among Primates. Cambridge, MA: Harvard University Pres. ——— (1996) Good Natured: The Origins of Right and Wrong in Humans and Other Animals. Cambridge, MA: Harvard University Press. Wagner, H. (2000) “Bargaining and War.” American Journal of Political Science 44 (3): 469–84. Waite, R. G. L. (1990) “Leadership Pathologies: The Kaiser and the Fuhrer and the Decisions for War in 1914 and 1939.” In Psychological Dimensions of War, ed. B. Glad, 143–68. Newbury Park, CA: Sage. Walker, S. G. (1977) “The Interface Between Beliefs and Behavior: Henry Kissinger’s Operational Code and the Vietnam War.” Journal of Conflict Resolution 21 (l): 129–68. ——— (1999) “Presidential Operational Codes and Foreign Policy Conflicts in the Post-Cold War World.” Journal of Conflict Resolution 43 (5): 610–25. ——— (2005) “Profiling the Operational Codes of Political Leaders.” In The Psychological Assessment of Political Leaders, ed. J. Post, 215–45. Ann Arbor: University of Michigan Press. ——— (2005) “Saddam Hussein: Beliefs and Integrative Complexity.” In The Psychological Assessment of Political Leaders, ed. J. Post, 387–90. Ann Arbor: University of Michigan Press. Wallace, M. D (1973a) War and Rank Among Nations. Lexington, MA: D.C. Heath. ——— (1973b) “Alliance Polarization, Cross-Cutting, and International War, 1815–1964.” Journal of Conflict Resolution 17: 576–604. ——— (1982) “Armaments and Escalation: Two Competing Hypotheses.” International Studies Quarterly 26: 37–56. Wallensteen, Peter. (1981) “Incompatibility, Confrontation, and War: Four Models and Three Historical Systems, 1816–1976,” Journal of Peace Research 18 (1): 57–90. Wallerstein, E. (1983) Historical Capitalism. London: Verso. Walt, S. M. (1987) The Origins of Alliances. Ithaca, NY: Cornell University Press. ——— (1996) Revolution and War. Ithaca, NY: Cornell University Press. Waltz, K. N. (1959) Man, the State and War. New York: Columbia University Press. ——— (1979) Theory of International Politics. Reading, MA: Addison-Wesley. ——— (1988) “The Origins of War in Neorealist Theory.” In The Origin and Prevention of Major Wars, ed. Robert I. Rotberg and Theodore K. Rabb, 39–52. Cambridge: Cambridge University Press. ——— (1990) “Nuclear Myths and Political Realities.” American Political Science Review 84 (3): 731–45. Wang, K., and J. L. Ray (1994) “Beginners and Winners: The Fate of Initiators of Interstate Wars Involving Great Powers Since 1945.” International Studies Quarterly 38 (1) (March): 139–54. Ward, M. D., and K. S. Gleditsch (1998) “Democratizing for Peace.” American Political Science Review 92 (1): 51–61.

588

Selected Bibliography

Wayman, F. (1985) “Bipolarity, Multipolarity, and the Threat of War,” In Polarity and War, ed. A. N. Sabrosky, 115–44. Boulder, CO: Westview. ——— (1996) “Power Shifts and the Onset of War.” In Parity and War, ed. J. Kugler and D. Lemke, 145–62. Ann Arbor: University of Michigan Press. ——— (2000) “Rivalries: Recurrent Disputes and Explaining War.” In What Do We Know About War? ed. J. A. Vasquez, 219–34. Lanham, MD: Rowman & Littlefield. Weart, S. (1998) Never at War: Why Democracies Will Not Fight One Another. New Haven, CT: Yale University Press. Weede, E. (1976) “Overwhelming Preponderance as a Pacifying Condition among Contiguous Asian Dyads, 1950–69.” Journal of Conflict Resolution 20: 395–411. ——— (1984) “Democracy and War Involvement.” Journal of Conflict Resolution 28 (4): 649–64. ——— (1992) “Some Simple Calculations on Democracy and War Involvement.” Journal of Peace Research 29: 377–83. Welch, D. (1992) “The Organizational Process and Bureaucratic Politics Paradigms: Retrospect and Prospect.” International Security 17 (2) (Fall): 112–46. ——— (1998) “A Positive Science of Bureaucratic Politics?” Mershon International Studies Review 42 (2) (November): 210–16. Wendt, A. (1992) “Anarchy Is What States Make of It: The Social Construction of Power Politics.” International Organization 46 (2): 391–425. ——— (1995) “Constructing International Politics.” International Security 20 (1): 71–81. ——— (1999) Social Theory of International Politics. New York: Cambridge University Press. Werner, S. (2000) “The Effects of Political Similarity on the Onset of Militarized Disputes, 1816–1985.” Political Research Quarterly 53 (2): 343–74. Werner, S., and D. Lemke. (1996) “Power Transitions and Military Buildups: Resolving the Relationship between Arms Buildups and War.” In Parity and War, ed. J. Kugler and D. Lemke, 187–210. Ann Arbor: University of Michigan Press. Whyte, G. (1989) “Groupthink Reconsidered.” Academy of Management Review 14 (1): 40–56. Wiegele, T. (1973) “Decision-Making in an International Crisis: Some Biological Factors.” International Studies Quarterly 17: 295–333. Wilkenfeld, J., V. L. Lussier, and D. Tahtinen (1972) “Conflict Interactions in the Middle East, 1949–1967.” Journal of Conflict Resolution 16: 135–54. Wilkinson, D. (1999) “Unipolarity Without Hegemony.” International Studies Review 1 (2): 141–72. Wilpert, B., P. Burger, J. Doktor, and R. Doctor (1976) “The Risky Shift in Policy Decision Making: A Comparative Analysis.” Policy Sciences 7: 365–70. Wilson, E. O. (1975) Sociobiology: The New Synthesis. Cambridge, MA: Harvard University Press. ——— (1978) On Human Nature. Cambridge, MA: Harvard University Press. Wilson, M. L., and R. W. Wrangham (2003) “Intergroup Relations in Chimpanzees.” Annual Review of Anthropology 32: 363–92. Winter, D. G. (1973) The Power Motive. New York: Free Press. ——— (2005) “Assessing Leaders’ Personalities: A Historical Survey of Academic Research Studies.” in The Psychological Assessment of Political Leaders, ed. J. Post, 11–38. Ann Arbor: University of Michigan Press. ——— (2005) “Saddam Hussein: Motivations and Mediation of Self-Other Relationships.” In The Psychological Assessment of Political Leaders, ed. J. Post, 370–74. Ann Arbor: University of Michigan Press. Woods, K., J. Lacey, and W. Murray (2006) “Saddam’s Delusions: The View from the Inside.” Foreign Affairs 85 (3) (May/June): 2–26. Woodwell, D. (2004) “Unwelcome Neighbors: Shared Ethnicity and International Conflict During the Cold War.” International Studies Quarterly 48: 197–223. Wright. Q. (1965) A Study of War, 2nd ed., 2 vols. Chicago: University of Chicago Press. Yetiv, S. (2003) “Groupthink and the Gulf Crisis.” British Journal of Political Science 33 (3) (July) : 419–42.

Index

Abdolali, Nasrin, 175, 262, 274 accommodation, of gains and losses, 77–78 accommodationists, and steps-to-war theory, 307 Acheson, Dean, 103, 137 achievement orientation, 53, 54 action-reaction pattern: arms races, 293–306; conflict spiral model, 280–292 Adams, John Quincy, 265 adaptiveness, 24 Adorno, T. W., 55 adrenaline, 108 affect heuristic, 64 affiliation orientation, 53, 54 Afghanistan, 80, 144, 219 agency, constructivism and, 463 aggression, 13–48; causes of, 14; deterrence theory and, 343; and human development, 16–17; intraspecific, 15; ritualized, 16; self-esteem and, 53; social learning theory and, 43–45; speciespreserving functions of, 15 Agricultural Revolution, 31–33, 36 Akhromeyev, Sergei, 138 Albright, Madeleine, 138, 358 Allee, T., 246 alliances: and deterrence theory, 351–352; and distribution of power, 389; empirical research on, 395–401; nuclear deterrence and, 365; types of, 351 alliance transition model, 422–423 Alliance Treaty Obligations and Provisions (ATOP) project, 351, 399 Allison, Graham, 128, 129, 132, 136, 137, 139, 150 alpha states, 197 Al Qaeda, 365 altruism, 24 Amarna tablets, 471 Amin, Hafizullah, 80 amygdala, 28, 62

analogies, historical, 70–76 analysis, levels of, 11–12 anarchy, 440; constructivism and, 463–464; and international system, 372, 373–374, 377; neoclassical realism and, 387 Anderson, Paul, 120, 137 Andropov, Yuri, 80, 144, 146 Angola, 219 anocracy, 175, 181 anomalies, theory and, 10 anticipatory compliance, 160 anxiety, 65; and risk taking, 83 appeasement, 71–72, 75; deterrence theory and, 343; groupthink and, 155; lessons from, 71 appeasement gestures, 16 appropriateness, logic of, 463 Arapesh people, 34 archaeology: and evidence for war, 41–42; timeline of, 31, 32 Argentina, 183; and contested institutions, 211–212; and deterrence theory, 357; diversionary war theory and, 201, 202–204 Aristotle, 1 Armenia, 183 Armitage, Richard, 150, 162 arms races, 293–306; as action-reaction pattern, 293–296; definition of, 293; empirical evidence on, 299–304, 301; game theory and, 325–326; and rivalry, 257; types of, 303–304; and war, 9–10, 296–299 Arrow, Kennety, 134 Arrow’s paradox, 134 ascending hegemony, 442 associative activation, 69 associative memory, 69 asymmetrical combat, in primitive societies, 38 asymmetrical escalation pattern, 281, 282 atlatl, 33 589

590 audience costs, 268, 288, 341 Augustine of Hippo, saint, 14 Austria, 202, 426 authoritarianism, 55 authoritarian regimes, and war, 174 autocracies: and diversionary war, 208–209; and dyadic relationships, 261–262 autocratic peace, 262 availability heuristic, 67, 76 Axelrod, Robert, 137, 325, 326 Ba, Alice, 463, 471 Babst, Dean, 257 Badie, Dina, 161–162 balanced systems, 404–405 balance-of-interest theory, 386 balance of power: classical realism on, 376; dyadic, 275–277; neorealism on, 378; and rivalry, 256–257 balance of risks theory, 81–82 Ball, George, 75, 115 Bandura, Albert, 43 bandwagoning, 380 Bangladesh, 49, 224 Baram, Amatzia, 355, 356 Barber, Benjamin, 251 bargaining theory, 330–338; critique of, 335–336; two-level, 336–338 Bar Joseph, Uri, 110 Barnett, Richard, 157 Bay of Pigs crisis, 143; groupthink and, 155; newgroup syndrome and, 163 Beckett, Thomas, 133 Begin, Menachem, 107, 110 behavioral ecology, 23 Behr, Roy, 328 beliefs: changes in, 94–97; and leadership, 107–108; role of, 84–97 belief system, 84 Bendor, J., 140 Bennett, D. S., 117, 184, 422, 478; and dyadic analysis, 263, 276; and hegemonic war theory, 430; and K-waves, 445–446; and polarity, 403; and territorial disputes, 238 Bennett, Scott, 261, 262 Benoit, K., 176 Benson, Brett, 351–352, 399 Ben-Yehuda, H., 248 Berlin, Isaiah, 95 beta states, 197 Betts, Richard, 124, 124–125, 138, 552n54 bias(es): and deterrence theory, 361; in information processing, 67–70; and misperception, 106–107; versus rational actor model, 119; role of, 66–70 Bickers, K. N., 205

Index biology, and war, 28–30 bipolarity, 390–394; arguments for, 393; critiques of, 393–394; long cycle theory and, 434 Birt, Raymond, 58 Blainey, Geoffrey, 101, 105, 199, 206, 213–214, 331; and business cycle, 190, 191; and dyadic analysis, 276; and war weariness, 231 Blair, Tony, 56 bonobos, 21–22 Boswell, Terry, 430, 434, 443 Boulding, Kenneth, 239 bounded rationality, definition of, 121 bounded rationality theory, 121–127 BPM. See bureaucratic politics model brains: and emotions, 62; and war, 28–29 Braybrooke, David, 122, 123 Brecher, M., 111, 248 Bremer, Stuart, 192, 195, 238, 401, 520n91; and conflict spiral model, 288; and dyadic analysis, 276; and polarity, 401 Brezhnev, Leonid, 80, 138, 144, 147, 149 Britain: and arms races, 294–295, 298; and democratic peace, 260; and democratization, 182; and deterrence theory, 342–343, 350; as hegemon, 429, 430, 432, 433, 434, 437–438, 441 Brown, Harold, 149 Brzezinski, Zbigniew, 92, 96, 149 Buchanan, James, 265 Bueno de Mesquita, Bruce, 116, 117, 118; and democratic peace, 267; and polarity, 402; and polarization, 394–395; and territorial disputes, 239 Bundy, McGeorge, 120 bureaucratic inertia, 129; and conflict spiral model, 285 bureaucratic politics, definition of, 132 bureaucratic politics model (BPM), 128, 132–152; applications of, 140–141; consequences of, 135–136; versus groupthink, 161, 165–166; and other governments, 142–143; as theory, 150–152; and U.S. decision making, 136–140 Burke, Edmund, 173 Bush, George H. W., 72, 127, 156, 173, 268, 316, 317, 356, 387 Bush, George W., 73–74, 79, 95–96, 107, 162, 173, 185, 205, 356, 387, 468–469, 474; operational code of, 87, 88–89 Bush Doctrine, 185 business cycle, and wars, 190–192 Byman, D., 56, 497n1 Cambodia, 217 Campbell, S. H., 175 Camus, Albert, 13 capitalism, 186–188; and democratic peace, 272–273

Index Caprioli, M., 186 Carneiro, Robert, 33 Carr, E. H., 375 cartelized systems, 145 Carter, Jimmy, 79, 92, 96–97, 149 Cashman, G., 95, 99, 161, 200, 277, 292 Cassandra’s advocate, 166 catecholamines, 108 causation, system 1 and, 69 causes of war, 477–489; aggression, 13–48; bureaucratic politics model and, 143–144; empirical theory and, 1–12; groupthink and, 165; liberalism and, 171; organizational process model and, 130; precautions with, 8; psychological factors, 49–113 Ceausescu, N., 147 Central Intelligence Agency, 93 certainty effect, 78 Chaco War, 419 Chad, 219 Champion, Michael, 403, 417 Chan, Steve, 174, 262, 453 change, constructivism and, 463 Chase-Dunn, Christopher, 441, 442, 446 Checkel, Jeffrey, 465, 468 chemical weapons, norms on, 472 Cheney, Dick, 73, 79, 136, 150, 161, 162 Chervonenko, S. V., 148 chicken games, 321–323, 321; solving, 327, 342 Chile, 183, 184 chimpanzees, 19, 20–21 China, 127; conflict spiral model and, 285; current status of, 457; deterrence theory and, 348; economy of, 455; misperception and, 102, 102–103; relative power cycle theory and, 452; and revolution, 217; rise of, 459; status discrepancy theory and, 409 choice shifts, 156–158 Chojnacki, Sven, 176 Choucri, Nazli, 195, 196–197, 294, 295 Christopher, Warren, 149 Churchill, Winston, 295, 375 civil wars. See internal conflicts class, 408 classical realism, 374–377; neoclassical realism, 386–388 Claude, Inis L., Jr., 371, 378, 411 Clausewitz, Karl von, 375 Clay, Henry, 265 Clinton, Bill, 92, 126, 268, 387 cluster polarity, 389; decrease in, 456. See also polarization coalition maintenance, 143 coercive diplomacy, 340 cognitive consistency, 90–91

591

cognitive dissonance, 90, 92–94, 504n181 cognitive psychology, 66; and bargaining theory, 335; and deterrence theory, 360–361; revolution in, 62–76 coherence, desire for, 69 Colaresi, Michael: and rivalry, 252–253, 254–255, 256; and steps-to-war theory, 310, 312 Colbert, Stephen, 477 Cold War: and alliances, 398–399; conflict spiral model and, 283, 284–286; and democratic peace, 271; and deterrence theory, 353–354; end of, 315–317, 455–456, 467–468; Mearsheimer on, 405; steps-to-war theory and, 309–310, 311 collective avoidance, 160 collective over-optimism, 160, 163 colonialism, 197 commitment problem, bargaining theory and, 332 compellence, 340 complex image, 89–90 complexity, organizational process model and, 128 compliance, 159 compromise, bureaucratic politics and, 134 concepts: formulation of definitions for, 3–4; nature of, 3 conditional deterrent alliances, 351–352 confirmation bias, 70 conflict: bureaucratic politics model and, 133–135; interaction patterns, 280–282 conflict-cohesion theory, 201 conflict spiral model, 280–282; conciliatory responses in, 283–289; critiques of, 291–293; and deterrence, 351; empirical evidence of, 286–287; GRIT model and, 312–317; and steps-to-war theory, 306 conflicts short of war: democracies and, 177; nuclear weapons and, 363; regime dyads and, 262–264; territorial disputes and, 248–249 conformity pressures: and groupthink, 153; and newgroup syndrome, 162; types of, 160 Congo, Democratic Republic of, 219 consequences, logic of, 463 conservatism: on conflict spiral, 283–289; term, 536n6 consistency theory, 90–91; examples of, 92–94 constitutional regimes, and war, 174 constitutive norms, 463 constructivism, 461–475; and bargaining theory, 335–336; and deterrence theory, 364, 368; and dyadic interaction level, 470–471; and individual level, 465–466; and international system level, 471–474; and Iraq War, 474–475; principles of, 461–464; and state level, 469; and substate level, 466–469; term, 461 content analysis, 86 contested institutions, 210–213; definition of, 210

592

Index

contiguity, 238–251; and rivalry, 255 contingent realism. See defensive realism controlled pressure, 354 cooperation, strategies for, 315 cooperation-hostility scale, 284 Copeland, Dale, 424–427 Copper Eskimo people, 34, 37, 38 core, 440–441 Correlates of War (COW) project, 3, 173, 176, 195, 238, 244, 258 correlation, versus causation, 8 Coser, Lewis, 201 cosmopolitan law, Kant on, 269 costly signals, 341 counting coup, 35 Cousteau, Jacques, 1 Crawford, Neta, 64, 64–65 credibility, deterrence theory and, 343–348 credibility trap, 345 Crescenzi, M. J. C., 274 Crichlow, Scott, 164 Cuba, 217 Cuban missile crisis: bureaucratic politics model and, 137, 139–140; game theory and, 323; groupthink and, 155; organizational process model and, 128, 129–130 culture: and democratic peace, 264–266; evolution of, 31–34; and nationalism, 223–226; social learning theory and, 43–45. See also nurture Cusack, Thomas, 234–235 custodian manager, 167 cybernetic procedures, 129 cyclical theories, 427–454; war weariness, 231–235 Cyert, Richard, 121 Czechoslovakia, 145–149, 194, 251 Damasio, Antonio, 62–63 D’Anieri, Paul, 68, 465 Dart, Raymond, 15 Dassel, Kurt, 210–211, 213 data, gathering, 6–9 Davies, Graeme, 206, 213 Davis, D. R., 208 Dawisha, Karen, 513n131 Dayan, Moshe, 110 deadlock preference, 426 death instinct, 14 death watch wars, 214 decision making: bureaucratic politics model and, 132–152; emotion and, 62–65; factors affecting, 50–51; group, 12, 115–168; before outbreak of war, 97–98; poliheuristic theory and, 125–127; prospect theory and, 76–79; rational, 51–52; small-group, recommendations for, 166–168 declining hegemony, 442

deduction, 2 de-escalation pattern, 281–282 default value, 75 defense: versus deterrence theory, 339; misperception and, 104 defensive avoidance, 106 defensive realism, 81, 378–380 definitions: of concepts, formulation of, 3–4; operational, 3; in psychology, 62 DeGaulle, Charles, 375 Demichev, P. N., 146 democracy(ies), 172–180; degree of, and peace, 259, 260; and diversionary war, 208–209; in dyadic relationships, 257–261; increased number of, 274, 488; indicators of, 4; and peace, 6–8, 6–7, 173–178, 179–180; and winning wars, 178–179 democratic peace: alternative explanations for, 270–273; constructivism and, 471; dyadic relationship and, 257–261; marginal cases and, 259–261, 260; and power transition theory, 418; term, 258; theory on, 264–270 democratization, 180–185; and contested institutions, 210; evidence for wars in, 182–185; pathway to war, 181–182 Deng Xiaoping, 127 dependent variable, definition of, 4 De Rivera, Joseph, 102 deterrence: case studies of, 353–359; definition of, 544n56; by denial, 339; statistical studies of, 349–352; types of, 341 deterrence theory, 339–369; and arms races, 304; cognitive psychology and, 360–361; empirical research on, 348–359; madman strategy, 322, 542n4; nuclear weapons and, 362–365 Deutsch, Karl, 240 de Waal, Frans, 22–23, 45, 493n37 Diamond, Jared, 46–47 Diehl, Paul, 196, 238; and arms races, 300–301, 303–304; and dysfunctional crisis learning, 288; and rivalry, 252, 253, 254, 256 Diesing, P., 11, 71, 120, 134, 138, 142; and deterrence theory, 363; and game theory, 322 direct deterrence, 341 dissatisfaction: and power transition theory, 412; and status discrepancy theory, 408 distribution of power: definition of, 389; empirical research on, 401–403; and international system, 374 diversionary wars, 78–79; internal, 207–208; and rivalry, 257; state-to-nation imbalance and, 229 diversionary war theory, 200–210 diversity, versus groupthink, 154 dogmatism, 54 domestic instability path, 486

Index domestic issues: and bargaining theory, 336; rivalry and, 253; territorial disputes and, 242 domestic unrest, and diversionary war theory, 205–206 dominance, 55 Doran, Charles F., 446–454 dorsal striatum, 28–29 Downes, Alexander, 179 Doyle, Arthur Conan, 319 Doyle, Michael, 174, 258, 261, 268–269 drive, 491n9 drive-discharge model, 17; critique of, 18 Dubcek, Alexander, 145, 147 Duelfer, C. C., 100 Dulles, John Foster, 93–94, 138 dyadic interaction level, 12, 237–277, 279–317, 319–369; constructivism and, 470–471; deterrence theory and, 342 Dyer, Gwynne, 35 dynamic differentials theory, 424–427 Dynasties, Age of, 243 dysfunctional crisis learning, 287–288 Dyson, Stephen, 56, 100 East, Maurice, 408, 409 economic cycle: global, 191; and wars, 190–192 economic development, and diversionary wars, 208 economic surplus, 35 economic systems, 186–188; and democratic peace, 272–273; K-waves and, 444–446; world systems theory and, 440–446 Eden, Anthony, 107, 110 egalitarian band societies, 35 egocentric bias, 68, 345 EGRIT policy, 315 Egypt: and arms races, 297; deterrence theory and, 364; and rivalry, 252 Ehrenreich, Barbara, 19 Eisenhower, Dwight D., 126, 187 elites: and capitalism, 187; and diversionary wars, 205; and regime transition, 180; and revolution, 215 Ember, Carol and Melvin, 36–37, 39 emotions, 62–65; definition of, 62; and framing, 84; and identity, 499n60 empirical theory, 1–12; definition of, 2 endowment effect, 77 enduring rivalries, 252; constructivism and, 470; and power transition theory, 418 England. See Britain Enterline, A. J., 274 epistemic communities, 467 equifinality, 486 Eritrea, 183

593

escalation/fight pattern, 280–281, 282; empirical evidence of, 286–287 Eskimo people, 34, 37, 38 Etheredge, Lloyd, 53, 56 Ethiopia, 217, 224 ethnic issues, 251–252 ethnocentrism, 25 ethology: critique of, 17–19; definition of, 15; early, 15–19; second-wave studies in, 20–30 European Union, 457, 489 event-interaction approach, 283–284; empirical evidence of, 284–286 evidence, theory and, 10 evoked sets, 97 evolution: and culture, 31–34; sociobiology and, 23–26 evolutionary psychology, 23 expansionist orientation, 56 expected utility theory, 116–119, 125 explanations: bureaucratic politics model and, 150–151; term, 8 extended deterrence, 341, 344–348, 349 extended GRIT policy, 315 extensive wars, 446 extroversion, 55–56 Fabbro, David, 34–35 Falklands Islands crisis, 156; and deterrence theory, 357; diversionary war theory and, 201, 202–204 Farber, H. S., 271 Farrell, Theo, 466 fear, 62; threats and, 64–65; and war, 485–486 Fearon, James, 330–332, 341–342 Fearon, Robert, 267 feedback, 372 Feith, Douglas, 87 Ferris, Wayne, 536n216 Fieldhouse, D. K., 189 fightaholic states, 170, 276; characteristics of, 170 fight/escalation pattern, 280–281, 282 Finland, 73, 258, 259–260 Finley, David, 284 Finnemore, Martha, 472 first inflection point, in relative power cycle theory, 447, 448 foreign aid, 221 Fore people, 34 Fox, Robin, 15, 19 fox approach, 92, 95; and deterrence theory, 360 framing, 77 France: and democratic peace, 258; and democratization, 182; deterrence theory and, 342; diversionary war theory and, 202; dynamic differentials theory and, 426; hegemonic war theory and, 429; long cycle theory and, 434; and

594

Index

revolution, 215, 217; and rivalry, 252; status discrepancy theory and, 409 Frank, Jerome, 60 Freedman, L., 203 Freeman, J. R., 285, 315, 316, 317 Freud, Sigmund, 14, 60 Frost, Robert, 237 F-scale, 55 fundamental attribution bias, 68; and deterrence theory, 346, 361 gains, valuation of, 77, 77–78 Galtieri, L. F., 201, 202, 203 Galtung, Johann, 408 Gamba-Stonehouse, V., 203 game matrix, 320 game theory, 319–330; critique of, 330; two-level, 336–338 gamma states, 197 Gamson, William, 284 Gandhi, Indira, 49 Garnham, David, 275, 389 Gat, Azar, 25, 26, 34, 37, 38 Gates, Robert, 93 Gelb, Leslie, 124, 124–125 Geller, Daniel, 200, 363; and deterrence theory, 365; and dyadic analysis, 275, 277; and power transition theory, 414, 418; and prisoners of war (POWs), 424; and rivalry, 257 Gelpi, Christopher, 206, 209 gender: and risk taking, 83; and war, 29–30 general deterrence, 341 generalization, 9 genetic fitness, 24 George, Alexander, 85, 86, 167, 544n56, 547n113; and deterrence theory, 353–354 Germany: and arms races, 294–295, 298; current status of, 457; and democratic peace, 260–261; and democratization, 182, 185; deterrence theory and, 342–343; diversionary war theory and, 202; dynamic differentials theory and, 425, 426; and lebensraum, 194; long cycle theory and, 434; and power transition theory, 417; relative power cycle theory and, 451; and rivalry, 252; and war weariness, 233 Gibler, Douglas, 302; and alliances, 397–398, 399; and territorial disputes, 249, 271–272 Gilpin, Robert, 427–431 Glaser, Charles, 378, 379 Glaspie, April, 105 Gleditsch, N. P., 176, 183, 195, 219, 220; and democratic peace, 259, 260; and dyadic analysis, 262 global economic cycles, 191 global power, 562n131

Goa crisis, 203 Gochman, Charles, 193, 286–287, 410, 418 Goddard, Stacie, 544n40 Goertz, Gary, 252, 253, 254, 256 Goldstein, Joshua, 29–30, 191, 285; and GRIT model, 315, 316, 317; and K-waves, 444–445 Gomulka, Wladyslaw, 146 Goodall, Jane, 18, 19, 20 Goodsell, Robert, 280–281, 285 Gorbachev, Mikhail, 92–93, 315, 315–317, 467, 468 governmental politics model, 128 Gowa, J., 271 graduated and reciprocated initiative in tension reduction. See GRIT model grand theories, 11 greedy states, 379, 386 Grey, Charles, 60, 97 Grishin, V., 146 GRIT model, 312–317; evaluation of, 314–317; guidelines for, 313–314; versus TIT-FOR-TAT (TFT) strategy, 329–330 Gromyko, Andrei, 80, 120, 144 group cohesion, 515n178; diversionary war and, 201; and groupthink, 153–154, 156, 160 group decision making, 12, 115–168; issues in, 120; rationality and, 115–116; recommendations for, 166–168 group polarization, 157 group selection, 493n54 groupthink, 90, 152–166, 515n178; avoiding, 155, 166–167; versus bureaucratic politics model, 161, 165–166; causal pathways to, 160; characteristics of, 152; conditions for, 153–154; and deterrence, 368; effects of, 152–153; evidence for, 154–156; integrated model of, 159–162; recent scholarship on, 158–163; term, 152; as theory, 163–165 Guatemala, 183 guerrilla warfare, 320–321, 320 Gulf War: misperception and, 105; organizational process model and, 131 Gurr, Robert, 207 Haas, Jonathan, 41–42 Haas, Michael, 174, 188, 193, 524n71 hairy chest syndrome, 157 halo effect, 69 Halperin, Morton, 139, 150 Hammond, T., 140 hardliners, and steps-to-war theory, 307 Harvey, Frank, 355, 359 Hastings, M., 202 Haushofer, Karl, 194 hawkish operational code, 86–87 Hebron, L., 452–453

Index hedgehog approach, 92, 95; and deterrence theory, 360 hegemon, term, 411, 427 hegemonic maturity, 442 hegemonic power, 441 hegemonic struggle path, 487 hegemonic victory, 442 hegemonic war theory, 427–431; empirical research on, 430–431; K-waves and, 446 Hegre, H., 176, 259, 260, 262 Henehan, M. T., 245 Henry II, king of England, 133 Hensel, Paul, 238, 248, 249, 250 Herek, M., 164 Herman, Robert, 467, 468 Hermann, C. F., 174 Hermann, M., 54, 56, 83 Hess, Rudolf, 194 heuristic device, game theory as, 330 heuristics, term, 67 heuristic theories, 66–70 Hilsman, Roger, 510n68 Hirohito, emperor of Japan, 223 historical-cyclical theories, 427–454; and power transition theory, 453–454 history, analogies and lessons from, 70–76 Hitler, Adolf, 194, 342–343 Hobbes, Thomas, 14, 375 Hobbesian world, 373 Hobson, John, 186–187 Hoffman, Matthew, 463, 471 Holsti, K. J., 230, 243–244, 525n75 Holsti, Ole, 84, 93–94, 109, 238, 284 honor cultures, deterrence theory and, 360 Hoole, Francis, 396 Hopf, Ted, 347, 402, 462, 465 Houweling, H., 200, 417, 451–452, 558n45 Howard, Peter, 474, 475 human nature: classical realism on, 375; Jervis on, 374 Human Resources Area Files (HRAF), 37 hunter-gatherer societies, 34, 37, 37–38, 39; archaeological evidence for war in, 41–42 Huntington, Samuel, 211, 293, 297–298 Hussein, Saddam, 127, 150, 201, 217–218, 334, 355–356; and misperception, 105; psychological characteristics of, 54, 59–60, 72, 100, 497n19, 498n43 Hutchison, M., 302 Huth, Paul, 164, 177, 525n92; and deterrence theory, 346, 349–350, 351, 352; and dyadic analysis, 276; and nationalism, 224; and territorial disputes, 242, 245–247, 249 Hutterite people, 34 hydraulic model, 17

595

hypotheses: construction of, 4–5; definition of, 4; testing, 2, 6–9 ICB. See International Crisis Behavior project idealist views of state, 170–171 identities, 462; emotion and, 499n60 Ideology, Age of, 243, 529n30 IGOs, and democratic peace, 270 images: changes in, 94–97, 96; content of, 85–87; definition of, 84; and leadership, 107–108; resistance to change, 89–94, 96; role of, 84–97; structure and dynamics of, 89–92 imaginability, 68 immediate deterrence, 341, 349 imperialism, 186–188 inclusive fitness, 24, 493n54 incoherent states, 227 incrementalism, 122–123; bureaucratic politics model and, 135; and Vietnam War, 124–125 independent variables, 5 India, 49–50, 102, 224; and democratic peace, 261; diversionary war theory and, 203; relative power cycle theory and, 452; and rivalry, 252; steps-towar theory and, 312 individual level, 12; aggression and, 13–48; constructivism and, 465–466; deterrence theory and, 342; effects of, 50–51; psychological factors and, 49–113 indivisible issues, bargaining theory and, 332–319 Indonesia, 213 induction, 2 industry, 557n16; long cycles and, 431; world systems theory and, 441, 443–444 inequality, wars of, 277, 487 inevitability of war, misperception of, 103–104 inflection points, in relative power cycle theory, 447, 448 information: democracies and, 179; private, bargaining theory and, 331, 338 information processing, biases in, 67–70 inherent bad faith model, 94, 107 instant endowment effect, 77–78 instinct, 15, 491n9; critique of, 18 institutional socialization, 133 integrated model of groupthink, 159–162 integrative complexity, 89–90 intensity, of war, 3 interests, 462 intergovernmental organizations (IGOs), and democratic peace, 270 internal conflicts, 199–221; contested institutions, 210–213; diversionary war theory, 200–210; “kick ’em while they’re down” wars, 213–214; outside intervention in, 218–220; revolution, 215–218

596

Index

International Crisis Behavior (ICB) project, 248 international norms, 371, 473–474; and democratic peace, 264–266; and nuclear weapons, 363–364, 473; spread of, 472, 488–489; on treaties, 472 international system: classification of, 390; definition of, 371 international system level, 12, 371–406, 407–459; constructivism and, 471–474; steps-to-war theory and, 307–308; theories on, implications of, 454–459 interstate war, definition of, 1 interventionism, 172–173 introversion, 55–56 Iran, 97; and death watch wars, 214; and revolution, 217–218 Iran-Contra affair, 156; groupthink and, 160 Iran hostage crisis, 156; bureaucratic politics model and, 149; groupthink and, 155 Iran-Iraq War, 217–218 Iraq, 127, 183; and death watch wars, 214; deterrence theory and, 355–356; diversionary war theory and, 201; relative power cycle theory and, 452; and revolution, 215, 217–218 Iraq War, 73–74, 514n135; bargaining theory and, 333–335; bureaucratic politics model and, 149–150; constructivism and, 474–475; groupthink and, 161–162; misperception and, 104; neoconservative operational code and, 87, 88–89 Iron and Rye coalition, 145 islands of theory(ies), 11 isolated rivals, 252, 254 isolationism, 172 Israel: and arms races, 297; and deterrence theory, 354, 358–359, 364; and rivalry, 252 iterated games, 324 Jakubovsky, I., 146, 148 James, Patrick, 204, 355, 452–453, 523n33 James, William, 14 Janis, Irving, 106, 106, 152, 153, 153, 154–155, 157, 163, 164, 166 Janis’s Law, 153–154 Japan: current status of, 457; and defensive realism, 382; and democratization, 182; embargo on, 134–135; and lebensraum, 194; military of, 144, 212–213; and nationalism, 223; Russo-Japanese War, 332; two-level bargaining theory and, 337–338 Jasinski, Michael, 207 Jenkins, S., 202 Jensen, Lloyd, 295 Jervis, Robert, 63, 70, 71, 102, 106, 106; and conflict spiral model, 290; and defensive realism, 378, 379, 380; and deterrence theory, 346, 358; and

international system, 374; and prospect theory, 77, 78, 79, 503n165 Job, B. L., 204 Johnson, Dominic, 97, 98 Johnson, Lyndon, 55, 75, 92, 124 joint freedom proposition, 258 Jones, Daniel, 288 Jones, David, 149 Jordan, Hamilton, 149 Kadar, Janos, 147 Kahneman, Daniel, 63, 64, 66–70, 67, 76, 77 Kaiser, David, 93 Kant, Immanuel, 170, 489; and democratic peace, 258, 268–270 Kanter, Arnold, 139 Kargil War, 261, 362 Karsten, P., 233 Kashmir, 224 Kazakov, general, 148 Keeley, Lawrence, 36, 38, 39 Kegley, C. W., 402, 471 Kelly, Raymond, 39–40, 41, 496n146 Kelman, Herbert, 113 Kennan, George F., 375 Kennedy, John F., 72, 140, 323, 375; groupthink and, 154–155; newgroup syndrome and, 163; organizational process model and, 129–130 Kennedy, Robert, 72, 137, 140 KGB, 146, 148 Khan, Yahya, 49 Khomeini, ayatollah, 217 Khong, Yuen Foong, 74, 76 Khrushchev, Nikita, 120, 142, 323, 508n16 “kick ’em while they’re down” wars, 213–214 Kim, Samuel, 19 Kim, Woosang, 276, 414, 421–422, 422–423, 557n23 kin selection, 24, 25 Kisangani, E., 209 Kissinger, Henry, 55, 87, 92, 94, 279, 375 Kjellen, Rudolf, 194 Klein, Gary, 167 Kocs, Stephen, 244–245, 249 Kondratieff, Nikolai, 444 Kondratieff waves, 444–446 Korean War, 71–72, 224, 505n208; balance of risks theory and, 81–82; groupthink and, 155; misperception and, 102–103; versus Vietnam War, 75–76 Kosovo, 126, 357–358 Kosygin, Anatoly, 80, 147, 149 Kowert, P., 83 Krasner, Stephen, 137

Index Kugler, J., 414, 416; and deterrence theory, 365; and power transition theory, 413–414, 414–415, 417 Kumar, Sushil, 452 !Kung San people, 34, 35, 37, 38 K-waves, 444–446 Lake, David, 178, 274, 333, 334, 335 Lalman, David, 117 land powers, long cycle theory and, 438–439 Lasswell, Harold, 53 lateral pressure, 196–198; definition of, 196 law, 9 Layne, Christopher, 270–271 LDCs. See less developed countries leaders: characteristics of, effects of, 51; and criticism, 168; and decision making, 50–51; personality traits of, 52–61; stress and, 108–112 leadership: beliefs and personality and, 107–108; directive, and groupthink, 154, 156, 161 lead industries, 431, 441 Leakey, Richard, 31 Lebanon, 218, 219 lebensraum, 194–196; definition of, 194 Lebow, Richard Ned, 60, 61, 100, 101, 103, 168, 202, 203, 544n60; and conflict spiral model, 288; and deterrence theory, 352, 357, 360; and misperception, 106–107 Leeds, B., 208, 399 Leites, Nathan, 85 Lemke, Douglas, 416, 418, 419 Lemnitzer, Lyman, 72 Leng, Russell, 280–281, 284; and conflict spiral model, 285, 286, 286–287, 288–289; and deterrence theory, 348; and dysfunctional crisis learning, 287–288; and game theory, 329; and war weariness, 234 Lenin, V. I., 187–188, 319, 354 Lepcha people, 34 Leroy, Marcel, 521n104 less developed countries (LDCs): and democratic peace, 273; and war, 195 lessons of history, 70–76 levels of analysis, 11–12 Levi, Werner, 36 Levy, Jack, 77, 78, 78–79, 84, 100, 101, 130, 143, 177, 486; and alliances, 396, 400; and democratic peace, 259; and diversionary war theory, 201, 203, 205; and K-waves, 446; and long cycle theory, 435, 438–439; and polarity, 402; and polarization, 395; and steps-to-war theory, 312 lexicographic choice, 125–126 Libby, I. Lewis “Scooter”, 87, 150 liberalism, 170–171; on conflict spiral, 283–289; term, 536n6 Lieber, Kier, 552n70

597

life instinct, 14 limbic system, 62 limited probe, 354 Lincoln, Abraham, 140 Lindblom, Charles, 122, 123 Lindskold, Svenn, 314 Lobell, S. E., 387, 388 Locke, John, 170 Lockhart, Charles, 342 logrolling, 134, 145 long cycles, 191, 432 long cycle theory, 431–440; critique of, 435; empirical research on, 434–435; implications of, 439–440; reformulated, 435–439 long waves, 444–446 Lorenz, Konrad, 15, 16, 16–17, 19 loss aversion, 77 losses, valuation of, 77, 77–78 loss of strength gradient, 239 lower turning point, in relative power cycle theory, 447, 448 Luard, Evan, 103, 242–243, 372 MacArthur, Douglas, 102 Macfie, A. L., 190 Machiavelli, N., 375, 377 madman strategy, 322, 542n4 magnitude, of war, 3 Magsaysay, Ramon, 72 majority, will of, bureaucratic politics model and, 134 Malaya analogy, 72–73, 75 Malaysia, 213 malignant narcissism, 59 Malinowski, Bronislaw, 30 Manichaean world view, 107–108 manipulation, 159 Mann, Leon, 106, 106 Mansfield, Edward, 180, 181–182, 182–183, 200, 261, 444 Maori people, 46–47 Maoz, Zeev, 169–170, 175, 193, 215; and conflict spiral model, 292; and dyadic analysis, 237, 262–263, 274, 276 Mao Zedong, 505n208 March, James, 121–122 Marshall Plan, 155, 221 Marx, Groucho, 461 Masherov, P. M., 146 Maslow, Abraham, 52 Massoud, Tansa, 514n135 May, Ernst, 71, 72 Mayaguez rescue, groupthink and, 155 Mbuti people, 34, 37 McCauley, C., 155, 159

598 McDermott, Rose, 62, 110 McInerney, Audrey, 79–80 McKinley, William, 266 McNamara, Robert, 72, 138, 139, 140 McNaughton, John, 345 Mearsheimer, John, 385, 404–405 Meir, Golda, 110 Mercer, Jonathan, 63, 346–347, 499n60 mere exposure effect, 69 Mesolithic Age, 32; evidence for war in, 41–42 Mexican-American War, 265 Mexico, 217, 265 Middle East: conflict spiral model and, 285; deterrence theory and, 358–359 middle range theories, 11 Midlarsky, Manus, 402 Mihalka, Michael, 275 Miles’ Law, 132 militarized disputes (MIDs): territorial disputes and, 248–249. See also war military: capabilities of, misperception of, 100–103; and contested institutions, 211–212; and risky shifts, 157–158; subcultures in, 137; and use of force, 138–139, 144; and war weariness, 233 military balance, and territorial disputes, 246 military-industrial complex, 187 Miller, Benjamin, 226–230 Miller, Ross, 208–209 Milosevich, Slobodan, 357–358 minimax solution, 321 Minix, Dean, 157 minorities at risk, 207 Mintz, Alex, 125 misperception: causes of, 106–107; role of, 98–108 Mitchell, David, 514n135 Mitchell, S. M., 207, 209 Mitchell, Sarah McLaughlin, 274 mixed-regime dyads, 262 Modelski, George, 431–440, 443 Modigliani, Andre, 284 Molotov-Ribbentrop Pact, 399 Moltke, Helmuth von, 131, 144 monadic peace, 172–180 Mondale, Walter, 149 monoamine oxidase A, 29 monocausal explanations, 8; critique of, 18 Monsen, J., 174 Montagu, Ashley, 18 mood theory, 500n70 More, Thomas, 49 Morgan, C., 175 Morgan, Patrick, 339, 340, 347–348, 353, 357–358, 361, 368; and war weariness, 234 Morgan, T. C., 175, 205, 237, 263 Morgenthau, Hans, 342, 375

Index Moriori people, 46–47 Morrow, James, 303, 414 motivated biases, 106 motives, psychological, 52–54 Moul, W., 275 Mousseau, Michael, 272–273 Mueller, John, 362–363, 473–474, 488 multiple advocacy, 167–168 multiple causation, 478; and territorial disputes, 247–248 multiple hierarchy perspective, 419–420 multipolarity: arguments for, 391–392; critiques of, 392 multivariate explanations, 8 Munich analogy, 71, 71–72, 75, 204 Munich Crisis, deterrence theory and, 342–343 Narang, V., 184–185 narcissism, 57, 58–60; stress and, 111 narcissistic rage, 498n43 nation: imbalance with state, 226–230; term, 221 national identity, territory and, 241 nationalism, 221–230; nation-centric, 223–226; state-centric, 222–223; term, 221–222 Nationalism, Age of, 243 national strength, indicators of, 3–4 nature: early research on, 15–19; versus nurture, 14, 26–28; second-wave research on, 20–30 necessary condition, 8–9 needs: hierarchy of, 52; psychological, 52–54 negotiation. See bargaining theory Nelson, R. M., 184–185 neoclassical realism, 386–388 neoconservative operational code, 87, 88–89 Neolithic Age, 32; evidence for war in, 41–42 neorealism, 377–378, 550n25; and hegemonic war theory, 428 nesting, 372 Netherlands: as hegemon, 432, 437–438, 441; long cycle theory and, 434 Nevin, John, 235 newgroup syndrome, 162–163 Nicaragua, 217 Nicholas II, tsar of Russia, 102 Nietzsche, Friedrich, 115 Nixon, Richard, 55, 138, 375, 542n4 noncompensatory calculus, 125 noncontenders, power transition theory and, 418–421 nonstate actors, 371 normative orders: permissive, 471; restrictive, 472 normative theories, 2 norms, 462–463; and democratic peace, 264–266; and international system, 371; and nuclear weapons, 363–364, 473; types of, 463. See also international norms

Index norms entrepreneurs, 466–467 North, Robert, 100, 196–197, 372; and arms races, 294, 295; and conflict spiral model, 285 North Korea, constructivism and, 474–475 nuclear taboo, 363–364 nuclear weapons: and arms races, 299; and deterrence, 362–365; norms and, 363–364, 473 nurture, 30–47; versus nature, 14, 26–28 obsessive-compulsive personality, 57 offense, misperception and, 104 offense-defense balance theory, 216, 379, 380–385 offensive realism, 385–386; polarity and, 404–406 Ogaden War, 224 Ogarkov, Nikolai, 138, 144 Oneal, John, 177–178, 184, 204, 417, 489, 523n33; and autocratic peace, 262; and democratic peace, 261, 269–270 Onuf, Nicholas, 461, 462 openness, and risk taking, 83 operational code(s), 85–87, 505n208; definition of, 85; hawkish, 86–87; neoconservative, 87, 88–89 operational definitions, 3 operational milieu, 85 opportunistic attack, 382 optimal strategy, 116 optimism: collective over-optimism, 160, 163; economic recovery and, 191; and war, 105 Oren, Ido, 260–261 organizational process model, 128–131 Organski, A. F. K., 411–412, 416, 451, 557n16, 558n38; and deterrence theory, 365; and long cycle theory, 434; and power transition theory, 413–414, 414–415, 417 Orme, John, 347, 359 Osgood, Charles, 313, 313 Ostrom, Charles, 204, 396 Otterbein, Keith, 28–29, 33–34, 40, 496n140 Ottoman Empire, 184–185, 438 outside intervention, in internal conflicts, 218–220 overconfidence, 101 Owen, John, 179, 261, 265, 534n166 Oye, Kenneth, 327 pacifist states, 170; size and, 193 Pakistan, 49–50, 183, 224; and democratic peace, 261; and intervention, 219; and relative power cycle theory, 452; and rivalry, 252; and steps-towar theory, 312 Paleolithic Age, 32; evidence for war in, 41–42 Paleologue, Maurice, 134 paranoid personality, 57–58, 61 Parasaliti, Andrew, 452 parity, 376, 389, 422 Park, Yong-Hee, 417

599

parsimony, theory and, 10 Parsons, W., 450 Patchen, Martin, 543n19 paths to war: alternative, 486–487; dynamic differentials theory and, 426 Paul, T. V., 277; and constructivism, 473; and deterrence theory, 364, 365 Pavlovsky, general, 148 peace: democracy and, 6–8, 6–7; GRIT model and, 312–317; primates and, 22–23; systems theories and, 458; territorial issues and, 249–251; trend toward, 473–474, 488–489. See also democratic peace peaceful societies, 34–43; critiques of, 36–43 Pearl Harbor, preliminaries of: groupthink and, 155; two-level bargaining theory and, 337–338 Peloponnesian Wars, 426, 429 Pel’she, A., 146 Pennsylvania, 276 Pentagon Papers, 345 periphery, 440–441 Perkins, Dexter, 191 Perlmutter, Amos, 139 permissive normative orders, 471 perseverance effect, 76, 502n124 personality, 52–61; and beliefs, 107–108; complexes, 56–60; disorders of, 57; and resistance to change, 95; and risk taking, 82–84; stress and, 111; term, 56; traits, 54–60 Peru, 183, 184 PGRIT strategy, 315 Phoenix Factor, 415 Pickering, J., 209 plausibility, theory and, 10 Pogo, 13 Polanyi, Karl, 272 polarity, 388–390, 456; definition of, 389; dynamic differentials theory and, 425; empirical research on, 401–403; and offensive realism, 404–406; and polarization, 403–404 polarization, 388–390; decrease in, 456; definition of, 389; empirical research on, 394–395; and polarity, 403–404 poliheuristic theory (PH), 125–127 Politburo, 142–143 politically relevant dyads, 237, 239 political process model, 510n68 politics, bureaucratic politics model and, 133–135 Polk, James, 265 Pollack, Kenneth, 56, 497n1 Pollins, Brian, 430, 435, 443 Ponomarev, Boris, 147, 149 population: growth, 194–196; lateral pressure, 196–198; role of, 194–198 Portugal, as world power, 432

600

Index

Posen, Barry, 378 Post, Jerrold, 57, 59, 110, 111 post–Cold War period: constructivism and, 468–469; and deterrence theory, 369; neoclassical realism and, 387–388; steps-to-war theory and, 310 Powell, Colin, 138, 150, 161, 162 Powell, Jody, 149 Powell, Robert, 333 power, 408; indicators of, 3–4 power orientation, 53–54 power polarity. See distribution of power power shifts, 420–421 power transition: definition of, 412; hegemonic war theory and, 429 power transition theory, 411–424; critiques of, 421–423; empirical research on, 416–418; and historical-structural theories, 453–454; and noncontenders, 418–421 predatory states, 386 prediction, theories and, 11 preemptive war, 102; and conflict spiral model, 291; dynamic differentials theory and, 426; offensedefense realism and, 381, 384 premature cognitive closure, 90 premortem, 167 preponderance, 376, 389, 423 president, role of, 139–140 Press, Daryl, 347 Preston, T., 132, 140 preventive war, 382, 384, 507n262; dynamic differentials theory and, 424 Primakov, Yevgeny, 467 primatology: on aggression, 19–22; early research on, 15–19; experiments in learning and, 45–46; on peacemaking, 22–23; second-wave studies in, 20–30 principal-agent problem, 522n10 Prins, B. C., 207, 209 prisoners’ dilemma games, 323–325, 324; and international relations, 325–326; solving, 326–329 private goods, 118 private information, bargaining theory and, 331, 338 probabilistic hypotheses, 5 process tracing, 8 progressive GRIT strategy, 315 prospect polarization, 158–159 prospect theory, 76–79, 503n165; applications of, 79–82 proto-rivals, 252 protracted conflicts, definition of, 248 proximity, state-to-nation imbalance and, 229 Prussia. See Germany pseudocertainty effect, 79 psychohistory, 60–61

psychological explanations, 49–113; rivalry and, 253; war weariness, 231–235 psychological milieu, 85 public goods, 118; hegemon and, 427 pulling and hauling, 134, 150 Putin, Vladimir, 44–45 Putnam, Robert, 336–338 qualitative arms race, 298 qualitative methods, 8 qualitative threshold, 298–299 quantitative arms race, 298 Rabin, Yitzhak, 110 Radford, admiral, 138 rank disequilibrium theory, 407–411 rapid approaches, 420 Rapoport, Anatol, 280, 327 Rasler, Karen: and long cycle theory, 435–439; and rivalry, 252–253, 254–255, 256; and steps-to-war theory, 310, 312 rational actor model (RAM), 115–116, 128; government avoidance of, 119–121 rational choice model, and hegemonic war theory, 428 rationality: and bargaining theory, 335; bounded rationality theory, 121–127; and democratic peace, 267–268; and deterrence theory, 340, 368; expected utility theory and, 116–119; and game theory, 320 rational model, 51–52; emotions and, 62, 63 Rattinger, Hans, 294 Ratzel, Friedrich, 194 Rauchhaus, R., 365 Ravenal, Earl, 345 Ray, James Lee, 259, 270, 410 Raymond, Gregory, 402, 471, 472 Reagan, Ronald, 71, 160, 173, 319 realism: classical, 374–377; and conflict spiral model, 288–289; defensive, 81, 378–380; neoclassical, 386–388; neorealism, 377–378; offense-defense, 380–385; offensive, 385–386, 404–406; and polaritiy and polarization, 388–390 realist views of state, 170–171 realpolitik tactics: and deterrence theory, 364; and steps-to-war theory, 306, 307 reassurance, 351 reciprocity, empirical evidence of, 284–286 reciprocity/symmetry pattern, 280, 282 recovery, economic, and wars, 190–192 Redd, Steven, 126 red lines, 340 Reed, W., 263, 418 reference dependence, 77 regime instability, 185

Index regime transitions, and nationalism, 226 regime type: and conflicts short of war, 262–264; and diversionary war, 208–210; in dyads, 257–275; and war, 172–190 regional cycles, 435–439 regulative norms, 463 reinforcement, 43–44 Reinhardt, T., 213 Reiter, Dan, 178, 335; and conflict spiral model, 291–292; and defensive realism, 381 relative power cycle theory, 446–454; empirical research on, 450–453 releasing signals, 16 religion, and Iran-Iraq War, 217 Religion, Age of, 243 Renshon, Jonathan, 95, 107–53 repetition, and falsehood, 69 representativeness heuristic, 76, 501n102 republic, term, Kant on, 258, 268 reputation, deterrence theory and, 344–348, 364 research, value of, 477–478 resource scarcity, and war, 197–198 restrictive normative orders, 472 revisionist states, 227, 376; and arms races, 304 revolution, 215–218 Rhodes, Edward, 137 Rice, Condoleezza, 73, 161 Richardson, Lewis Fry, 232; and arms races, 284–285, 303; and territorial disputes, 238–239 Rider, T., 302 Ridgeway, general, 138 rigor, and testing hypotheses, 6 Ripsman, N. M., 387, 388 risk, misperception of, 102–103 risk bias, 76–84 risk factors, 478–479; model of, 479–486; and stepsto-war theory, 306–307 risk orientation, 56; and deterrence theory, 360; gains and losses and, 77; stress and, 111 risk taking, personality and, 82–84 risky shifts, 156–158 rivalries, constructivism and, 470 rivalry(ies), 252–257; definition of, 252; and diversionary wars, 206–207; evidence on, 254–257; and general deterrence, 341; and stepsto-war theory, 307; wars of, 277 Robinson, L., 95, 99, 161, 200, 277, 292 Roeder, Philip, 511n79 rogue states, 185–190; constructivism and, 474–475 Rohl, John, 60 Rokeach, Milton, 54 role expectations, 133 role theory, 132 Roosevelt, Franklin, 92, 134, 199 Roosevelt, Theodore, 99

601

Rosati, J., 271 Rosato, Sebastian, 265, 268 Rosecrance, Richard, 201 Ross, Dennis, 142 Rousseau, David, 177 Rousseau, Jean-Jacques, 36, 170, 290, 373, 373–374 Rubicon Theory, 97–98 Rummel, R. J., 174–175, 193, 199, 258 Rumsfeld, Donald, 73, 150, 161, 162 Rusk, Dean, 72, 75, 138 Russett, Bruce, 174, 177–178, 184, 204, 206, 489, 520n87, 523n33; and autocratic peace, 262; and democratic peace, 259, 260, 261, 269–270; and deterrence theory, 346, 349–350, 351, 352, 369; and dyadic analysis, 263 Russia, 438; current status of, 457; diversionary war theory and, 202; and intervention, 219; relative power cycle theory and, 451; and revolution, 217 Russo-Finnish War, 73 Russo-Japanese War, 332 Sabrosky, A. N., 399 Sadat, Anwar, 80–81, 360 saddlepoint, 321 Salehyan, I., 219 Salmore, S., 174 Sample, Susan, 301, 302, 304–305 Sapolsky, Robert, 45–46 satisficing, 121–122 scapegoat theory, 200–210 Schafer, Mark, 164 Schelling, Thomas, 330 schema theory, 74–75, 91–92 Schlieffen Plan, 130, 131, 145 Schmookler, Andrew Bard, 32 Schulz, K., 219 Schwebach, V. L., 175, 237, 263 Schweller, Randall, 386, 551n37 scientific method, 2–9 Scott, J. P., 17, 19, 23 Scott, P., 43 Scott, Walter, 199 sea power: arms races and, 294–295; long cycle theory and, 431, 433, 438–439; world systems theory and, 441 Sechser, Todd, 139 second inflection point, in relative power cycle theory, 447, 448 Second Punic War, 426, 429 secrecy, and decision making, 120 security dilemma: classical realism on, 376; conflict spiral model and, 289–291; constructivism and, 470–471; and rivalry, 254; state-to-nation imbalance and, 229 segmented societies, 40–41

602

Index

selective perception, 89 selectorate, 118; and democratic peace, 267 self-actualization, 52, 53 self-consciousness, and risk taking, 83 self-deterrence, 363–364, 473 self-esteem, 52, 53; and aggression, 53 Semai people, 34, 37 semiperiphery, 440–441 Semmel, Andrew, 137, 157 Senese, Paul: and alliances, 398, 401; and arms races, 302, 305; and dyadic analysis, 263; and steps-to-war theory, 306, 308–311, 312, 486; and territorial disputes, 240–241, 248–249 September 11, 2001, 73; and beliefs, 95 sequential search, 121–122 Serbia, 183, 224 serotonin, 29, 494n82 Seven Years’ War, 426 severity, of war, 3 Shelest, Pyotr, 146, 148 Shepard, Graham, 137 Shevardnadze, Eduard, 467 Shtemenko, general, 148 Shubik, Martin, 320 Shultz, George, 93 Siccama, J., 200, 417, 451–452, 558n45 signal-to-noise ratio, 119 Simmel, Georg, 201 Simon, Herbert, 121, 121–122 simple image, 89 simplicity, theory and, 10 Singer, J. David, 3, 8, 169, 173; and alliances, 396; and conflict spiral model, 280, 288; and democratic peace, 258; and polarity, 401; and war weariness, 234–235 Siriono people, 34 Six Day War, 247–248, 360 Slovic, Paul, 64 Small, Melvin, 3, 169, 173; and alliances, 396; and democratic peace, 258; and war weariness, 234 Smith, Adam, 170 Smith, Martin, 150 Smith, Theresa Clair, 303 Smoke, Richard, 353–354, 544n56, 547n113 Snyder, Glenn, 11, 71, 120, 134, 138, 142; and deterrence theory, 363; and game theory, 322 Snyder, Jack, 145, 180, 181–182, 182–183, 200, 261; and defensive realism, 378 Sobchack, Anatoli, 44 social evolution, 33 socialism, 187; and aggression, 189 social learning theory, 43–45 social-market capitalism, 272 social suitability, 40–41 sociobiology, 23–26; critiques of, 27–28

Somalia, 224 Sorenson, Ted, 137 South Vietnam, 219 Sovereignty, Age of, 243 Soviet Union: and Afghanistan, 80; and arms races, 294; breakup of, 194, 251, 455–456, 467–468; bureaucratic politics model and, 142–143, 145–149; general secretary, role of, 142; GRIT model and, 315–317; policy in Middle East, 79–80; reactions to glasnost policy, 92–93. See also Cold War Soysa, I., 417 Spain: and democratic peace, 265–266; and intervention, 219 Spanish-American War, 265, 266 Spiezio, Edward, 430 spiral model. See conflict spiral model Sprout, Harold and Margaret, 85 stability-instability paradox, 363 stag hunt analogy, 373 staghunt preference, 426 Stalin, Josef, 53, 58, 65, 71, 375 Stam, A. C., 117, 178, 184, 422, 478; and dyadic analysis, 263, 276; and hegemonic war theory, 430; and K-waves, 445–446; and polarity, 403; and territorial disputes, 238 standard operating procedures (SOPs), 128–129 state(s): idealist and realist views of, 170–171; imbalance with nation, 226–230; size and power of, and war, 192–194; term, 221; war involvement of, 169–170 state level, 12, 169–198, 199–236; constructivism and, 469 state-to-nation imbalance theory, 226–230, 487 status, 408 status discrepancy theory, 407–411; empirical evidence for, 410–411; implications of, 409–410 status quo states, 376 Stein, Janice Gross, 66, 70, 80–81, 94; and deterrence theory, 351, 352, 358, 360–361, 361; and mood theory, 500n70 Steinbruner, John, 66, 92, 129 steps-to-war theory, 144, 306–312; empirical evidence on, 308–311; and territorial disputes, 240–241 Stern, Eric, 161, 162, 163 Stoessinger, John, 49, 99 Stoll, Richard, 403, 417 Storr, Anthony, 17 strategic choice approach, and democratic peace, 267–268 strategic opportunity, 523n44 strategic rivalries, 252 stress: effects of, 109, 110; and groupthink, 153, 154; role of, 108–112

Index strong-congruent states, 227 strong-incongruent states, 227–228 structural-institutional argument, for democratic peace, 266 Stuckey, John, 401 substate/small-group level, 12, 115–168; constructivism and, 466–469 subsystems, 372 sufficient condition, 9 sufficient reasonableness, and deterrence, 368 Sukarno, 213 Sullivan, Charles, 80 Sundelius, Bengt, 161, 162, 163 sunken costs, 78 super-GRIT policy, 315 Suslov, Mikhail, 147, 149 Sweat, Mike, 430, 434, 443 Sweden, 189 Switzerland, 189 Syria, 79–80, 228 system 1, 63, 66–70, 67; operating characteristics of, 76–79 system 2, 63, 66–70, 67 Tahitian people, 34 Taiwan, 348 Talbott, Strobe, 169 Taliaferro, Jeffrey, 81–82, 387, 388 Tammen, R., 419, 558n38 Tannenwald, Nina, 364, 473 Taraki, Nur Mohammad, 80 Tasaday people, 34 Taylor, A. J. P., 173 technology, and population pressure, 196 Tenet, George, 161, 162 Tennyson, Alfred, 407 Terhune, K. W., 54 territorial disputes, 238–251, 250; evidence on, 242–248, 244; and nationalism, 224; as settled, 249–250, 488; special value of territory and, 241–242; state-to-nation imbalance and, 229; and steps-to-war theory, 306 territoriality, 17, 20–21, 241; agriculture and, 31; critique of, 18–19 territorial peace, 271–272 terrorism: rogue states and, 186; and system 1, 68 Tessman, B. F., 453 testosterone, 29–30 Tetlock, Philip, 92, 119, 155 Thailand, 183 thalamus, 62 ‘t Hart, Paul, 132, 140, 156, 159–162, 515n178 Thayer, Bradley, 24, 25 theory(ies): bureaucratic politics model as, 150–152; empirical, 1–12; evaluation and comparison of,

603

10–11; groupthink as, 163–165; islands of, 11; nature of, 2; testing, 2 third states, misperception and, 105 Thirty Years’ War, 402, 429 Thompson, William, 78, 84, 130, 143, 191–192, 431–440, 444, 486; and rivalry, 252–253, 254–255, 256; and steps-to-war theory, 310, 312 threats: and defensive realism, 380; effects of, 64–65; and escalation, 280–281; misperception of, 99–100. See also deterrence theory Thucydides, 169, 375, 388 Tiananmen Square crisis, 127 Tierney, Dominic, 97, 98 Tiger, Lionel, 15 timing of war, power transition theory and, 413 Tinbergen, Nikolaas, 15 tinderbox hypothesis, 300 Tir, Jaroslav, 196, 207, 247 Tirpitz, Alfred von, 144, 294 TIT-FOR-TAT (TFT) strategy, 315, 326–329, 350; versus GRIT, 329–330 Tito, Josip Broz, 147 topdog, 408 totalitarian regimes, and war, 174 Toynbee, Arnold, 231–235 trade: and democratic peace, 269–270; and lateral pressure, 196; liberalism and, 171 transitional democracies, 180; and democratic peace, 265 Trapeznikov, S. P., 146 treaties: Amarna tablets on, 471; international norms on, 472 trench warfare, game theory and, 325 tribalism, 251 Triska, Jan, 284 Tristan de Cunha, 34 Truman, Harry S., 71–72, 75, 82, 103, 155, 375 Trumbore, P. F., 186 Turkey, 183. See also Ottoman Empire Turner, Stansfield, 149 turning points, in relative power cycle theory, 447, 448 Tversky, Amos, 76 Twain, Mark, 407 two-level bargaining, 336–338 U-2 incident, 135 Ulbricht, Walter, 146 ultimatum game, 63–64 Ulyanov, V. I. See Lenin, V. I. unbalanced systems, 404–405 uncertainty, 116 uncommitted thinker, 147, 149 unconditional compellent alliances, 352 unconditional deterrent alliances, 351–352

604

Index

underdog, 408 uneven growth, law of, 428 unipolarity, 404, 456–458; hegemonic war theory and, 429 United Kingdom. See Britain United Nations Educational, Scientific and Cultural Organization (UNESCO), 49 United States: and arms races, 294; bureaucratic politics model and, 136–140; decline of, 457–458; and democratization, 182, 183; deterrence theory and, 348; diversionary war theory and, 202, 204–205, 208; economy of, 455, 456–457; GRIT model and, 315–317; as hegemon, 429, 430, 432, 433, 441; interventionism versus isolationism in, 172–173; two-level bargaining theory and, 337–338; world systems theory and, 443–444. See also Cold War unit-level variables, term, 387 universal hypotheses, 5 unmotivated biases, 106; versus rational actor model, 119 unpolarized system, 456 upper turning point, in relative power cycle theory, 447, 448 use of force decision, military and, 138–139 Ustinov, Dmitry, 80, 138, 144 Vakili, L., 203 Valenta, Jiri, 142, 145, 148 Valeriano, Brendan, 244, 288, 310 value complexity, 115 value trade-offs, 120 Vance, Cyrus, 149 Van Evera, Stephen: and conflict spiral model, 292; and defensive realism, 378, 380–385; and game theory, 328; and nationalism, 224–225, 225–226 variables, 4; types of, 4–5 Vasquez, John, 47, 144; and alliances, 397, 398, 400, 401; and arms races, 302, 305; and constructivism, 471; and contagion, 219; and dyadic analysis, 263, 277; and findings, 479; and power transition theory, 417, 421, 423; and rivalry, 253, 255; and steps-to-war theory, 306–312, 486; and territorial disputes, 238, 240–241, 244, 245, 248–249, 250, 271; and war weariness, 233 Vertzberger, Yaacov, 65, 83 Vietnam War, 71, 72, 89; incrementalism and, 124–125; Kissinger’s operational code and, 87; versus Korean War analogy, 75–76; versus Malaya analogy, 72–73 vigilant appraisal model, 155 Vincennes incident, 97, 108 Waite, Robert, 60

Walbiri people, 34, 37 Wallace, Michael, 299–301, 304, 395, 397 Wallensteen, Peter, 238, 253, 421 Wallerstein, Immanuel, 440–446 Walsh, William, 237 Walt, Stephen, 215, 216; and constructivism, 470; and defensive realism, 378, 380 Waltz, Kenneth, 189, 373, 377, 378 Wang, Kevin, 204 Waorani people, 38 war: arms races and, 9–10; biological bases for, 28–30; costs and duration of, misperception of, 104; definition of, 1, 37–38, 40; inevitability of, misperception of, 103–104; misperception and, 98–108; most violent, 176; nurture and, arguments for, 30–31; as obsolete, 488–489; origins of, 31–34, 36; as sport, 35–36. See also causes of war Ward, M. D., 183 Ward, Michael, 285, 296 war initiation: democracies and, 172; hegemonic war theory and, 429; long cycle theory and, 433; power transition theory and, 413–415; war weariness and, 232 War of 1812, 260, 265, 438 War of the Pacific, 419 war weariness, 231–235; empirical studies on, 234–235; reverse, 233 Watergate, groupthink and, 155 Wayman, Frank: and polarity, 402; and polarization, 394, 403; and power transition theory, 420–421; and rivalry, 256–257 weak-congruent states, 228 weak-incongruent states, 228 weapons of mass destruction (WMD), rogue states and, 186 Weart, Spencer, 258 Weber, Max, 408 Weede, Erich, 175, 275, 300 Weinberger, Caspar, 93, 133 Welch, D., 140 Welles, Orson, 169 Wendt, Alexander, 461, 463–464, 465, 470 Wheeler, H. B., 286, 329 Whyte, Glen, 158–159 Wilhelm II, kaiser of Germany, 53, 60–61, 97, 99, 131 Wilkenfeld, J., 111, 174, 248 Wilkinson, David, 459 Williams, William Appleman, 190 Wilson, E. O., 16, 23–24, 26, 31, 493n54; critiques of, 27 Wilson, Woodrow, 170, 173, 261 window theory, 382–383 winning, democracies and, 178–179, 274

Index win-sets, 336–337 Winter, D. G., 53–54, 55 wishful thinking, 103, 105 WMDs, rogue states and, 186 Wohlforth, William, 456–457 Wolfers, Arnold, 375 Wolfowitz, Paul, 87, 150 Woodwell, Douglass, 251–252 world economies, 440 world empires, 440 world power, 431, 562n131 world systems theory, 440–446; evidence and criticism on, 443–444 worldview, 84 World War I: game theory and, 328–329; misperception and, 104; organizational process model and, 130–131; relative power cycle theory and, 449 World War II, nationalism and, 224 world wars: hegemonic war theory and, 429, 442–443; long cycle theory and, 431–432; steps-

605

to-war theory and, 307; world systems theory and, 441 Wright, Quincy, 36, 173, 195, 201, 257 xenophobia, 25 Yakovlev, Alexander, 467 Yeltsin, Boris, 317 Yepishev, general, 146 Yetiv, S., 156, 514n151 Yom Kippur War, 80–81, 364 Yugoslavia, 224; breakup of, 194, 251 Zagare, Frank, 414 Zaire, 219 Zelikow, P., 129, 136 zero-sum games, 320–321; territorial disputes as, 242 Zinnes, Dina, 174 Zuk, Gary, 444 Zuni people, 34

About the Author

Greg Cashman is a professor of political science and the director of the international studies program at Salisbury University. His teaching and research interests include international conflict, international relations theory, American foreign policy, international law, and Russian politics. His previous books include What Causes War? and (with Leonard C. Robinson) An Introduction to the Causes of War: Patterns of Interstate Conflict from WWI to Iraq.

607