COURSE MATERIAL REVIEW

JervisPerceptionandMisperceptioninInternationalPolitics.pdf

Home >Political Science homework help >COURSE MATERIAL REVIEW

R O B E R T J E R V I S / Perception and Misperception in International Politics 257

Perception and Misperception in International Politics

ROBERT JERVIS

In a presentation based on psychology and social psychology, Professor Jervis dis- cusses the relations among different levels of analysis used by students of inter- national relations. He also makes a case for the importance of understanding individual perception ( and misperception) in the study of foreign policy and world politics. In a "two-step" model, actor images and perceptions are related to the actor's understanding of reality based on available information.

Do Perceptions Matter?

Before discussing the causes and consequences of the ways in which decision-makers draw inferences from the world, we must ask a pre- liminary question: do the decision-makers' per- ceptions matter? This is an empirical question. Logic permits us to distinguish between the "psychological milieu" ( the world as the actor sees it) and the "operational milieu" (the world in which the policy will be carried out) and to argue that policies and decisions must be medi- ated by statesmen's goals, calculations, and per- ceptions.1 But it does not follow from this that we must deal with these intervening variables in order to understand and predict behavior. This is not an uncommon claim:

One may describe particular events, conditions, and interactions between states without neces- sarily probing the nature and outcome of the processes through which state action evolves. However, and the qualification is crucial, if one wishes to probe the "why" questions underlying the events, conditions, and interaction patterns which rest upon state action, then decision- making analysis is certainly necessary. We would

go so far as to say that the "why" questions cannot be answered without analysis of decision-making.2

But theory and explanation need not fill in all the links between cause and effect. Indeed, this is impossible. One can always ask for the links between the links. High density theories have no privileged status; they are not automati- cally illuminating or fruitful.' It is true that re- creating a decision-maker's goals, calculations, and perceptions is a satisfying method of explaining his behavior because the scholar, sharing as he does the decision-maker's charac- teristics ofbeing a thinking, goal-seeking person, can easily say: "If that is the way the statesman saw the situation, it is no wonder that he acted as he did." But the comfort we feel with this form of explanation should not blind us to the fact that, unless there are significant variations in the ways people see the world that affect how they act, we need not explore decision-making in order to explain foreign policy. Most case studies assume that the details presented significantly affected the outcomes. This may not be true, however. "Pleikus are streetcars," McGeorge Bundy said in explaining that the Viet Cong attack on the American installation in February 1965 had

258 R o B E R T J E R v I s / Perception and Misperception in International Politics

affected only the timing of the American bomb- ing of North Vietnam.' If you are waiting for one, it will come along. The specifics of the trig- gering event cannot explain the outcome because so many probable events could have substituted for it. To understand the American policy of bombing the North we should not examine the attack on Pleiku. Had it not occurred, something else would have happened that would have per- mitted the same response. Logic alone cannot tell us that a similar claim about the decision-mak- ing process is invalid: the way people perceive others and make decisions only marginally influ- ences outcomes. So we must seek empirical evidence on the question: do the important ex- planatory variables in international relations involve decision-making? In terms of percep- tions this can be separated into two subsidiary questions: "are important differences in policy preferences traceable to differences in decision- makers' perceptions of their environments?" and "are there important differences between reality and shared or common perceptions?"5

These questions raise the familiar level of analysis problem. Although it has been much debated, agreement is lacking not only on the substantive dispute but even on the number of levels. Arnold Wolfers proposes two, Kenneth Waltz three, and James Rosenau five.6 To fill in the sequence, we will discuss four. One is the level of decision-making, the second is the level of the bureaucracy, the third is that of the nature of the state and the workings of domestic politics, and the fourth focuses on the international environment. 7 Which level one focuses on is not arbitrary and is not a matter of taste-it is the product of beliefs ( or often hunches) about the nature of the variables that influence the phenomena that concern one. To restate the first question in terms of the level of analysis problem, we need not adopt a decision- making approach if all states behave the same way in the same objective situation, if all states of the same kind (i.e., with the same internal

characteristics and politics) behave the same way in the same objective situation, or if state behavior is determined by bureaucratic rou- tines and interests.

Although the empirical questions are cen- tral here, we should also note that the level of analysis problem has important moral implica- tions. When all people would respond the same way to a given situation, it is hard to praise or blame the decision-maker. Thus, those accused of war crimes will argue that their behavior did not differ from others who found themselves in the same circumstances. And the prosecution will charge, as it did against Tojo and his col- leagues, that, "These defendants were not automatons; they were not replaceable cogs in a machine .... It was theirs to choose whether their nation would lead an honored life ... or ... would become a symbol of evil throughout the world. They made their choice. For this choice they must bear the guilt." Similarly, if all nations follow similar courses of action, one cannot argue that some deserve to be branded as immorally aggressive. Thus in 1918 Bethmann- Hollweg rebutted those who blamed Germany for the war by pointing to the "general disposi- tion towards war in the world . . . how else explain the senseless and impassioned zeal which allowed countries like Italy, Rumania and even America, not originally involved in the war, no rest until they too had immersed themselves in the bloodbath ?"8

Th~ three non-decision-making levels assert the importance of various aspects of the objective situation or the actor's role. 9 They say that if we know enough about the setting-international, national, or bureaucratic-we can explain and predict the actor's behavior. An interesting side- light is that if other actors believed that the set- ting is crucial they would not need to scrutinize the details of the state's recent behavior or try to understand the goals and beliefs held by the state's decision-makers. It would be fruitless and pointless to ask what the state's intentions are if

its behavio, which it fin< to predict h this will tell be. Decisio1 their powen ply ask the, were in the c worry abou1 have values It is interes1 that decisior usingthism others may the decision amongtherr deal of infor nificant pers

Ofcour answer to t1 important. I variables th, ]ems, the imJ one issue ar level of anal determined i we are seeki1 the general o specific resp while decisi explain why 1914, the ex- states to figh of variables a of a decision. dictate that , for a chang, bureaucracy sented to tl maker's pre, choice that v. tines of the b the decision variable may

RO B E RT J E RV I S I Perception and Misperception in International Politics 259

its behavior is determined by the situation in which it finds itself. Instead, observers would try to predict how the context will change because this will tell them what the state's response will be. Decision-makers could then freely employ their powers of vicarious identification and sim- ply ask themselves how they would act if they were in the other's shoes. They would not have to worry about the possibility that the other might have values and beliefs that differed from theirs. It is interesting, although not decisive, to note that decision-makers rarely feel confident about using this method. They usually believe both that others may not behave as they would and that the decision-makers within the other state differ among themselves. So they generally seek a great deal of information about the views of each sig- nificant person in the other country.

Of course it is unlikely that there is a single answer to the question of which level is most important. Rather than one level containing the variables that are most significant for all prob- lems, the importance of each level may vary from one issue area to another. Furthermore, which level of analysis is the most important may be determined by how rich and detailed an answer we are seeking. The environment may influence the general outline of the state's policy but not its specific responses. Thus it can be argued that, while decision- making analysis is needed to explain why World War I broke out in August 1914, the external situation would have led the states to fight sooner or later. Or the importance of variables at each level may vary with the stages of a decision. For example, domestic politics may dictate that a given event be made the occasion for a change in policy; bargaining within the bureaucracy may explain what options are pre- sented to the national leaders; the decision- maker's predisposition could account for the choice that was made; and the interests and rou- tines of the bureaucracies could explain the way the decision was implemented. And the same variable may have different effects at different

stages of the decision-making process-for ex- ample, conflicts among subordinates may in- crease the variety of information and the number of opportunities for decision that the top decision-maker gets, but may simultaneously decrease his ability to see that his decisions are faithfully implemented.

The importance of variables at one level can also vary with the state of variables at other lev- els. Rosenau suggests that the international envi- ronment is more important in determining the policy of small states than it is oflarge ones and Stanley Hoffmann argues that nuclear weapons and bipolarity have reversed this relationship. 10

More generally, the importance of the other lev- els decreases if the variables in one level are in extreme states. Thus, maneuvering within the bureaucracy may be more important when the top decision-makers are inexperienced or pre- occupied with other matters. And Wolfers ar- gues that states tend to behave the same way when they are faced with extreme danger or extreme opportunity, but that when environ- mental constraints are less severe there will be differences in behavior that must be explained at the decision-making level. More complex inter- actions among the levels are also possible. For example, the effect of internal instability on expansionism could vary with the opportunities for success in war. Unstable states may be more prone to aggression than stable ones when the chances of victory are high but might be more cautious than others when their leaders perceive a significant probability of defeat or even of temporary setback. Or the stability of the regime might influence its propensity for aggression, but the nature of the regime ( e.g., whether it is democratic or dictatorial) might be more im- portant in explaining how it makes peace.

To deal with all these questions would re- quire another book. Here all I will try to do is to outline the kinds of evidence necessary to estab- lish the validity of simple propositions about the importance of the various levels. In doing so, I

260 Ro BERT J E RV I S / Perception and Misperception in International Politics

will sketch the most extreme arguments for the importance of each level. It is obvious that the questions and arguments could be rephrased in more subtle terms but since I am concerned with the kinds of evidence that the propositions call for the gain in analytical clarity is worth the sacrifice involved in ignoring more complete explanations that combine a multitude of vari- ables at many levels.

The International Environment To argue that the international environment determines a state's behavior is to assert that all states react similarly to the same objective exter- nal situation. Changes in a state's domestic regime, its bureaucratic structure, and the per- sonalities and opinions of its leaders do not lead to changes of policies. Changes in the external situation, however, do alter behavior, even when variables on the other levels remain constant. To test these claims, we would need good measures of all the variables, especially the nature of the objective situation and the state's policies. Even if we had such indicators, we would have to cope with the paucity of the most desired kinds of comparisons. This is easily understood by glanc- ing at the similar issue in the study of individual behavior-the debate over the relative impor- tance of situation and role versus idiosyncratic variables in determining individual behavior. Because so many people of widely differing backgrounds, personalities, and opinions fill the same role and because the same person fills many different roles, we can try to determine the relative impact of situational and idiosyncratic variables by examining how a person's behavior varies as his role changes and how people of widely differing characteristics perform in simi- lar situations.

It is much harder to make the analogous comparisons in international relations. In only a few international systems do we find many cases in which states play, either simultaneously or consecutively, several roles and in which each role is filled by states that are otherwise quite

different. This would occur in a long-lasting sys- tem where there were frequent changes in the relations among the actors. Thus each state might at one time be a neutral, a "holder of the balance," a state with aggressive designs, a de- fender faced by a state whose intentions are dif- ficult to determine, and so on. To a limited degree this test is possible in a balance-of-power system. But it is not available for most other sys- tems, for example the one prevailing since World War IL Most nations have not changed roles) and indeed cannot do so because of such permanent factors as size and geography. The United States can never play the role of a sec- ond-ranking state caught between two blocs of greater powers. France can never be the leader of one of two dominant blocs. And while the United States and France may have played roles similar to these in the past, the extensive differ- ences in the situation mean that any differences in response that might be found would not show that roles are unimportant.

Compulsion in Extreme Circumstances. It is worthwhile to look at cases of the kind that are supposed to show most strongly the influence of external conditions. If there are differences of behavior here, the argument for not ignoring the other levels of analysis will apply a fortiori to a wider domain of cases. Arnold Wolfers argues that, the greater the external compulsion, the greater the homogeneity of behavior and there- fore the less the need to study decision-making. In a well-known passage he says: "Imagine a number of individuals, varying widely in their predispositions, who find themselves inside a house on fire. It would be perfectly realistic to expect that these individuals, with rare excep- tions, would feel compelled to run toward the exits . ... Surely, therefore, for an explanation of the rush for the exits, there is no need to analyze the individual decisions that produced it." 11

But the case is not as clear as this analogy suggests. If a situation were so compelling that all people would act alike, decision-makers would

not he: tive p, debate: key de those i1 shall P For des tion th, danger, fire, in tested. perceiv1 us bac Church Hitler t lain, thi, for othe decision which t spond. · Again, t< existed. much gr Europe, of our a: that thei1 agree abi true, for the turn Federalis that they the Rep1 menace. i ment was ues and analyses c

In ex certainty, sion-mak1 armed att that the st. tive situat the state's exits from the extent

,e te ,e ,_

f- :d

" s- :e :d :h 1e c- of er 1e es ·r- es IW

re of of ng to ,es he ~e- 1g. :a eir : a to :p- :he of rze

,gy 1at .t!d

.,. I

R O B E RT J E RV I S I Perception and Misperception in International Politics 261

not hesitate nor feel torn among several alterna- tive policies, nor would there be significant debates within the decision-making elite. In fact, key decisions that are easily reached, such as those involving the Truman Doctrine and Mar- shall Plan, stand out because they are so rare. For despite the implication of Wolfer's proposi- tion that we know when we are faced by extreme danger, just as we can tell when the house is on fire, in fact this question is often bitterly con- tested. (To say that once decision-makers perceive the fire they will head for the exits leads us back to decision-making analysis.) For Churchi!l, the house was burning soon after Hitler took power in Germany; for Chamber- lain, this was the case only after March 1939; and for others there never was a fire at all. To some decision-makers, the Soviet Union is a threat to which the United States is compelled to re- spond. To others the threat passed years ago. Again, to a growing number of scholars it never existed. Similarly, American statesmen saw a much greater threat from communism in both Europe and Southeast Asia than did the leaders of our allies. Decision-makers may even agree that their state's existence is threatened but dis- agree about the source of the threat. This was true, for example, in the United States around the turn of the nineteenth century, when the Federalists believed France so much a menace that they favored war with her. At the same time, the Republicans believed England an equal menace. (It should be noted that this disagree- ment was rooted as much in differences in val- ues and interests as in divergent empirical analyses of the situation.)

In extreme cases we can specify with some certainty an indicator of the "fire" that all deci- sion-makers will recognize-for example a large armed attack-and we can be relatively certain that the state wi!l react. But even then the objec- tive situation will not determine all aspects of the state's response. There are apt to be several exits from the burning house. Will the state limit the extent of its involvement? What will its war

aims be? While the United States may have had no choice but to declare war on Japan after Pearl Harbor, the major decisions that followed were less compelled and require further explanation. For example: the United States decided not to concentrate its energies on the country that had attacked it but to fight Germany first; the war was to be fought with few considerations for the shape of the postwar world; and no compromise with the enemies would be accepted (had the Japanese realized this would be the case, they almost certainly would not have attacked).

Even if all states and all statesmen responded similarly to similar high threats, we have to explain how the threat arose-i.e., why the adversary was so aggressive. In some cases we may be able to do this by reference to the other's objective situation, for example by focusing on the anarchic nature of the international system and the resulting security dilemma. . . . But when this analysis is insufficient, the state (and later scholars) must examine variables at other levels of analysis to establish some of the most important facts about the objective situation that the state faces.

Finally, one cannot prove that the external environment determines the response by simply showing that the decision-makers believed this to be the cause. It is not enough to say with Kecskemeti that "In tense war situations, the decision-maker is likely to feel that he is acting from necessity rather than from deliberate choice." Nor is it sufficient to cite Holsti's find- ing that the decision-makers on both sides in July 1914 felt that they had no choice but to make the decision they did, or to show that when "Mr. Acheson was advised not to favor the production of the first thermonuclear bomb, he is reported to have declared that its production was a matter of necessity and not of choice: in other words, that he was experiencing 'compul- sion.' " 12 The subjective feeling of determinancy is interesting and may lead decision-makers unnecessarily to restrict their search for alterna- tives, but it does not show that other decision-

262 R O B E R T J E R V I S / Perception and Misperception in International Politics

makers in the same situation would have felt the same compulsion and acted in the same way. Indeed the theory of cognitive dissonance ... and other theories of irrational cognitive consis- tency ... lead us to expect that decision-makers may avoid psychological conflict by thinking that they had no choice but to act as they did. This also means that, when scholars claim that a situation permitted no policy other than the one that was adopted, it may be that at least part of the reason why the circumstances appear over- whelming in retrospect is that they were claimed to be so by the decision-makers.

These arguments arei of course, far from conclusive. The necessary comparisons have merely been mentioned, not made. But, as we have seen, there are many points at which people can disagree about what the objective situation is or what policies will best cope with it, and there is little evidence for the existence of the homo- geneity of behavior that would allow us to ignore everything except the international setting.

Domestic Determinants Even if all states do not behave similarly in sim- ilar situations, the details of decision-making and images may not be significant. Instead, the state may be the appropriate level of analysis- i.e., variations in decision-makers' policies may be accounted for by variations in social and eco- nomic structure and domestic politics of the states they are serving. Wilsonian and Marxist theories are examples of this position. Other theories at this level of analysis argue for the importance of a state's geographical position, its traditions, its national style, or the conse- quences, often unintended, of domestic con- flicts. Extreme formulations hold that the state's internal system determines its foreign policy, while weaker versions claim that foreign policies are a product of both domestic politics and international circumstances.

The forms of the assertions correspond to those discussed in the previous section. States with the same critical internal attributes behave the same way in similar situations-and often

behave the same way in the face of significant variations in the environment-and this behav- ior is different from that displayed by other states with different attributes even when the setting is the same. The latter claim denies the overriding importance of the international environment. Thus while Cold War revisionists stress the importance of America's domestic political and economic needs, others reply that American actions were heavily influenced by external con- straints and that her behavior was not peculiarly American or peculiarly capitalist but rather was typical of any great power in such a situation. Because we are concerned -with examining the importance of decision- making, we will not treat this part of the argument that deals with conflicts between claims for two other levels of analysis.

If states of the same type behave in the same way, then changes in a state's leadership will not produce significant changes in foreign policy, and we need not examine the images, values, and calculations of individual decision-makers. Unfortunately, claims about continuity in a state's foreign policy are notoriously difficult to judge. We might try to see whether we could deduce changes in the identities of the state's decision-makers from the course of its foreign policy. Could we tell when Democrats replaced Republicans or Conservatives replaced Labour governments? Scholars used to agree that Stalin's death led to major foreign policy changes, but now even this is in doubt. 13 Before taking office, decision-makers often claim they will introduce new policies. But these promises are often neglected. Eisenhower's foreign policy more closely resembled that of his predecessor than it did his campaign rhetoric. Gladstone pledged himself to avoid immoral and wasteful imperialism, and, although he successfully extri- cated Britain from some entanglements, he was eventually drawn into commitments similar to those made by Disraeli. And while in 1937 Clement Atlee said that "the foreign policy of a Government is the reflection of its internal pol- icy," when his party took power the foreign sec- retary declared that "Revolutions do not change

geography, and graphical needs

Manyargm des can be und, the autonomy Those who prai that, had he car been able to r avoiding the en to Russia, being recklessly antag, tries. The rejoin1 man domestic s system would ha work no matter ing skill ofBisma the underlying fc origins of the Co ilar question o changed Ameri, accounts argue t anti-Soviet positi as Truman did. revisionists who and economic S} policy but is disp1 War as avoidabl, President Kenne, nam argue that Johnson did. The or opposed not American foreign of these-and ot tent. While those determinants as criticize the gene1 policy see the ca, that few significar the identity of the

These questic partly because the ernment always C fronted the old. I choose between de major war. F. D. consolidate her b

s g t. e d n ,- y lS

,e 1t

,t y, s, s. a

:o Id 's :n ,d 1r at cy re ey es cy or ae itl ri-

·as to 37 fa ,l- :c- ge

R O B E RT J E R V I S / Perception and Misperception in International Politics 263

geography, and revolutions do not change geo- graphical needs." 14

Many arguments about the wisdom of poli- cies can be understood in terms of claims about the autonomy of the decision-making level. Those who praise Bismarck's diplomacy claim that, had he continued in office, he would have been able to maintain German security by avoiding the errors of severing Germany's ties to Russia, being forced to rely on Austria, and recklessly antagonizing several powerful coun- tries. The rejoinder is that the dynamics of Ger- man domestic society and of the international system would have destroyed Bismarck's handi- work no matter who was in power. The glitter- ing skill of Bismarck's diplomacy could not alter the underlying forces at work. Debates about the origins of the Cold War must deal with the sim- ilar question of whether Roosevelt's death changed American policy. Most traditional accounts argue that F. D.R. was coming _to an anti-Soviet position and would have acted much as Truman did. This view is shared by those revisionists who look to the American political and economic system for the roots of foreign policy but is disputed by those who see the Cold War as avoidable. Similarly, those who defend President Kennedy but opposed the war in Viet- nam argue that he would not have acted as Johnson did. Those who either favored the war or opposed not only it but also most recent American foreign policies argue that the policies of these-and other-presidents were consis- tent. While those who supported the war see the determinants as international and those who criticize the general lines of America's postwar policy see the causes as domestic, both argue that few significant differences can be traced to the identity of the president.

These questions are so difficult to resolve partly because the situation facing the new gov- ernment always differs from that which con- fronted the old. Kennedy was never forced to choose between defeat in Vietnam and fighting a major war. F. D. R. did not live to see Russia consolidate her hold over East Europe. The

questions must then be hypothetical, and the comparisons that underlie our judgments are often strained. This problem can be avoided by using alternative comparisons-by examining the views of members of the elite to see whether they favor the policy that was adopted. Of course disagreement with a policy does not prove that a person would have acted on his views were he in office. His opposition might be rooted in his role in the government, lack of information, freedom from the pressures that accompany holding power, or the belief that opposition is politically expedient. But when these explanations are not satisfactory, internal elite disagreement reveals the limits of the impact of both domestic politics and the inter- national situation.

The Bureaucracy Even if state behavior cannot be explained by the state's internal politics and external environ- ment, we still may not need to examine the perceptions and calculations of the top decision- makers. The workings of the bureaucracy may determine policy. It is not enough for propo- nents of this position to show that the state's course of action appears inconsistent and lacks value integration. Such inadequacies can be the product of individual decision-making. As we will show later, normal human behavior often does not fit even a loose definition of rationality. Individuals as well as organizations fail to coor- dinate their actions and to develop carefully designed strategies. The fact that people must reach decisions in the face of the burdens of multiple goals and highly ambiguous informa- tion means that policies are often contradictory, incoherent, and badly suited to the information at hand. Unless we understand this, puzzling state behavior will automatically be seen as the product of either internal bargaining or the autonomous operation of different parts of the government. Thus if we did not know better it would be tempting to argue that the contradic- tory and erratic behavior displayed by Richard Nixon on Watergate and related matters shows

264 Ro B E RT J E RV I S J Perception and Misperception in International Politics

that "Nixon" is not a single individual at all, but rather a title for the set ofbehaviors that are pro- duced by the interaction of conflicting entities, each pursuing its own narrow interests in igno- rance of or indifference to any wider goal of the "general Nixon interest." Similarly, if we were to insist that theories of individual behavior apply only when the actor is following a coher- ent path guided by his self-interest, we would have to say that Spiro Agnew was an uncoordi- nated bureaucracy rather than a person because he simultaneously accepted kickbacks and sought the presidency.

Because incoherent policy is insufficient evidence for the importance of bureaucracies, the (<pure" theories of this type must make two basic assertions. First, bureaucrats' policy pref- erences are determined by their positions in the government: "Where you stand is determined by where you sit." The external environment and the nature of the state and domestic politics have only limited and indirect impact on these preferences. Of course if the concept of bureau- cratic interest is to be more useful than the con- cept of national interest, we must be able to specify in advance what the bureaucratic posi- tion will be. Even if we cannot do this, it would still be significant if everyone in each unit of the government had the same position on a given issue. If, on the other hand, there is a good deal of disagreement within the organization about the goals to be sought or the means that are appropriate, then we would have to apply deci- sion-making analysis to the bureaucratic level, and so this approach would lose much of its dis- tinctiveness. More importantly, if people in dif- ferent units share the same policy preferences or if preferences are distributed at random throughout the government, then the first asser- tion would be undermined.

The second basic claim of theories on this level of analysis is that the state's policies are formed by bureaucratic bargains and routines. Bureaucratic actions either determine the states- man's decision or else implement it in a way that

renders the decision largely irrelevant to what is actually done. This point is vital because, even if bureaucrats' policy preferences were linked to their positions within the government, this would be relatively unimportant unless these preferences explain policy outcomes. But we should note at the start that even if this were true we would have to explore the sources of power of parts of the bureaucracy. If we find, for exam- ple, that the military often prevails in conflicts with the organization in charge of arms control, this may be because over a period of years the state's leaders have supported the former more than the latter. Sometimes we can go back some years to find a decisive action that set the guide- lines for both the policy and the distribution of power within the bureaucracy. In less dramatic cases the relative strength$ of interests represent the standing decision of the decision-makers- and often of wider publics-and their choices among competing policies and values. To the extent that this distribution of power is both important and accounted for by factors outside the bureaucracy, an explanation of specific out- comes in terms of bureaucratic maneuvering will be superficial at best.

Are policy preferences determined by one's role within the government? With the impor- tant exception of questions of military hardware and doctrine, the evidence is limited and ambiguous. It is not hard to find examples of units taking consistent and unified stands and political appointees adopting their units' views and thus expressing different opinions depend- ing upon their positions in the government. "General Marshall, while Chief of Staff, opposed the State Department's idea of using aid to pro- mote reforms in the Chinese government. Then, when he became Secretary of State, he defended this very idea against challenges voiced by the new chiefs of Staff." In "1910, Winston Churchill, as Home Secretary, led the attack upon the demand of McKenna, First Lord of the Admiralty, for more ships; by 1913 they had exchanged offices and each, with equal convic-

ROB

tion, maintaine, Samuel Hoare w strongly fought a Air Arm; when h Admiralty he toe Theophile Delea, in France before ported an expedi France a lever to secretary, he soui

But not all 1 able to roles. 01 loyalties are !es: unusual rather t tively quick deci and generally sh most important : involving war an, to roles. When w American foreig terms for future c era! framework · conducted-it de for the top decisi, determines wher1

In several i1 striking is the deg of I 94 7 there wa, government that needed. Three ye foreign policy co scale rearmamer agreement-whic with bureaucra1 domestic politic made such a pol War removed thi: cials were again i important cases t key to them is no ests. Doves and I found in all parts Views on wheth, Japan before Wo: the crucial issue

R O B E R T J E RV I S / Perception and Misperception in International Politics 265

tion, maintained the opposite view.» When Samuel Hoare was secretary of state for air, he strongly fought against naval control of the Fleet Air Arm; when he later served as first lord of the Admiralty he took the opposite position. When Theophile Delcasse was the minister of colonies in France before the turn of the century, he sup- ported an expedition to the Nile that would give France a lever to use against Britain. As foreign secretary, he sought to recall the adventure.

But not all policy disagreements are trace- able to roles. Organizational perspectives and loyalties are less important when issues are unusual rather than routine, necessitate rela- tively quick decisions, and involve important and generally shared values. Beliefs about the most important issues of foreign policy-those involving war and peace-are usually unrelated to roles. When we look at the major decisions of American foreign policy-those that set the terms for future debates and established the gen- eral framework within which policy was then conducted-it does not seem to be true, at least for the top decision-makers, that "where you sit determines where you stand."

In several important cases what is most striking is the degree of unanimity. In the spring of 1947 there was general agreement within the government that massive aid for Europe was needed. Three years later most officials felt that foreign policy considerations argued for large- scale rearmament, although there was a dis- agreement-which was not tightly connected with bureaucratic interests-over whether domestic political and economic constraints made such a policy feasible. Once the Korean War removed this opposition, government offi- cials were again in general agreement. In other important cases there are basic disputes, but the key to them is not divergent bureaucratic inter- ests. Doves and hawks on Vietnam were to be found in all parts of the American government. Views on whether to take a hard line toward Japan before World War II, and specifically on the crucial issue of embargoing oil and other

vital raw materials, were only loosely related to organizational affiliations. The advice that Tru- man received at the start of the Berlin blockade and the Korean War and most of the differences that emerged in the discussions during the Cuban missile crisis were not predictable by the participants' roles.

In the missile crisis none of the leading offi- cials espoused views that were linked to his posi- tion within the government. The Republican secretary of the treasury was concerned about the effects of a "soft" response on the fortunes of the Democratic party in the coming elections; the secretary of defense at first argued that the missiles did not present a major military threat; the secretary of state did not take a strong posi- tion and did not pay special attention to the political consequences of various moves; and the attorney general opposed an air strike. (It should also be noted that his view carried great weight not because of his governmental position or independent political resources, but because he was thought to speak for the president.)

The other claim-that policies can be explained by bureaucratic maneuvering-could be supported in either of two ways. First, it could be shown that different parts of the gov- ernment carry out, or fail to carry out, policies in ways that are consistent with their preferences and routines rather than with the decisions of the national leaders. But the other possible link- age in the second point-the argument that authoritative decisions can be explained by the interaction of bureaucratic stands-raises diffi- culties that go deeper than the temporary absence of evidence. To verify this claim we must be able to specify the expected relationship between the variety of bureaucratic positions on the one hand and policy outcomes on the other. It is not enough to show that the outcome can be seen as a compromise among views that have been advocated by various parts of the govern- ment. Almost any decision could fit this descrip- tion. The theory must provide generalizations that tell us more exactly what the outcome will

266 RO B E RT J ERV I S I Perception and Misperception in International Politics

be. If the goals of different parts of the bureau- cracy are complementary, then presumably each agency will give up its position on the part of the program it cares least about in order to gain a larger voice on those issues that are more important to it. Presumably the success of an organization in conflicts with others is related to its strength ( determined independently of outcomes), although as we noted this raises fur- ther questions. Still another likely pattern is that the symbols will be given to one side in a bureaucratic conflict and the substance to the other. But much more detail is needed. Further- more, these generalizations must not involve the values and beliefs that vary with the identity of the top decision-makers, and they must be able to explain how policies change. The latter task poses great problems since bureaucratic struc- tures and interests often remain constant over periods in which policies shift.

Although the paucity of research on this level makes conclusions especially tentative, it is hard to see how any of the major decisions of American foreign policy in recent years could meet this test. The Marshall Plan, the establish- ment of NATO, the crucial decisions in Korea, the rearmament that followed, the decision to integrate West Germany into West Europe, the New Look in defense, American policy in the Suez crisis, Kennedy's attempt to increase con- ventional forces in Europe, the major decisions to fight and later withdraw from Vietnam, and crucial choices in the Cuban missile crisis can- not be explained as the outcome of intrabureau- cratic conflict. That these decisions combined major elements of positions held within the bureaucracy is hardly surprising because differ- ent parts of the bureaucracy serve and represent divergent values that the president seeks to fur- ther. Thus what seems to be a clash of bureau- cratic interests and stands can often be more fruitfully viewed as a clash among values that are widely held in both the society and the decision- makers' own minds. What embarrasses the the- ories under consideration here is that, while the decisions listed above did embody some of the

preferences that had been articulated by parts of the bureaucracy, they did not combine them in a way that can be predicted by rules of bureau- cratic politics. Or, to put the argument more exactly, until we have a theory that specifies how policy is formed out of conflicting bureaucratic perspectives and preferences, we cannot tell whether any given outcome can be explained by this level of analysis. As things stand now, there is no way to explore the extent to which bureau- cratic factors cause the outcome because we have no ground for claiming that a different constellation of bureaucratic interests and forces would have produced a different result or that the outcome would have been different were there no bureaucracies at all.

Perceptions, Reality, and a Two-Step Model

Our discussion thus far leads to the conclusion that it is often impossible to explain crucial deci- sions and policies without reference to the deci- sion-makers' beliefs about the world and their images of others. That is to say) these cognitions are part of the proximate cause of the relevant behavior and other levels of analysis cannot immediately tell us what they will be. And even if we found that people in the same situation- be it international) domestic, or bureaucratic- behave in the same way, it is useful to examine decision-making if there are constant differ- ences between the decision-makers' perceptions and reality. In this case all people might react in the same way to the same situation, but this behavior would puzzle an observer because it was self-defeating, based on incorrect beliefs about the world, or generally lacking in a high degree of rationality. Many of the propositions advanced in here fit in this category: they are generalizations about how decision-makers perceive others' behavior and form judgments about their intentions. These patterns are ex- plained by the general ways in which people draw inferences from ambiguous evidence and, in turn, help explain many seemingly incompre- hensible policies. They show how, why, and

-':!

whenhigt men misf fied ways.

Othe1 an analys because p differently differ in ti era! and of it will be u but often i pie differe, as they did

Thee: and the de sitions dis context ol Rather tha the direct, levels of ai examine ti immediate and France geredbyG, may have correct in grouped tc causes of tt

Ourur beliefs affo about that, of the actor was believe, ers had the quick victo tions conce1 ways by wh the stakes o believed tha ing that the that the deci view of the Papers and that the cm was conside high expect<

R O BERT J E RV I s / Perception and Misperception in International Politics 267

when highiy intelligent and conscientious states- men misperceive their environments in speci- fied ways and reach inappropriate decisions.

Other propositions deal with cases in which an analysis of decision-making is necessary because people in the same situations behave differently. This is often the case because people differ in their perceptions of the world in gen- eral and of other actors in particular. Sometimes it will be useful to ask who, if anyone, was right; but often it will be more fruitful to ask why peo- ple differed and how they came to see the world as they did.

The exploration of the images actors hold and the development of the two kinds of propo- sitions discussed above should be seen in the context of a mediated or two-step model. 15

Rather than trying to explain foreign policies as the direct consequence of variables at the three levels of analysis previously discussed, we will examine the actor's perceptions as one of the immediate causes of his behavior. Thus Britain and France felt that their security was endan- gered by Germany before both world wars. They may have been mistaken in the first case and correct in the second, but both cases can be grouped together in discussing the immediate causes of their responses.

Our understanding of the actor's images and beliefs affects the further question that we ask about that event and the behavior that we expect of the actor in other cases. For example, when it was believed that most American decision-mak- ers had thought that escalation would bring a quick victory in Vietnam, the interesting ques- tions concerned the reasons for the error and the ways by which successive small steps increased the stakes of the conflict. If the decision-makers believed that victory was cheap, it is not surpris- ing that they acted as they did. But by revealing that the decision-makers had a relatively accurate view of the chances of success, the Pentagon Papers and related commentaries have shown that the crucial question is why saving Vietnam was considered important enough to justify the high expected price. This then leads us to look at

this and other American actions in terms of beliefs about "domino effects" rather than direct- ing our attention to commitments that develop inadvertently and "quagmires" that trap unwary statesmen. Similarly, the question about Russian behavior raised by the Cuban missile crisis prob- ably is not "What Soviet calculus and risk-taking propensity could explain this bold and danger- ous step?" but rather "How could they have so badly misestimated the probable American response?"16 And previous Soviet behavior can be re-examined to see if it could be explained by similar misperceptions. . . . Actors as well as scholars must engage in these kinds of analyses.

Of course perceptions, and more specifi- cally perceptions of other actors, are not the only decision-making variables that are impor- tant. That two actors have the same perceptions does not guarantee that they will adopt the same response. But their responses will often be the same and, when they are not, it is usually rela- tively easy to find the causes of the differences. Although people with different images of an adversary may agree on the appropriate re- sponse, just as people may favor the same policy for different reasons, this agreement is apt to be short-lived .... The roots of many important disputes about policies lie in differing percep- tions. And in the frequent cases when the actors do not realize this, they will misunderstand their disagreement and engage in a debate that is unenlightening.

Images, however, are not first causes, and so we will try to find the causes both of common misperceptions and of differences in perceptions. Thus the second step in the model involves relat- ing the images held, if not to reality, then at least to the information available to the actor. How, for example, do statesmen come to develop their images of other actors? What evidence do they pay most attention to? What makes them per- ceive threat? Under what conditions do they think that the other, although hostile, has only limited objectives? What differentiates legitimate inducements from bribes? What kinds of behav- ior are most apt to change an established image?

268 R O B ER T J E RV I S I Perception and Misperception in International Politics

This is not to claim that we will be able to explain nearly all state behavior. Propositions about both the causes and the effects of images can only be probabilistic. There are too many variables at work to claim more. In the cases in which W-e are interested, decision-makers are faced with a large number of competing values, highly complex situations, and very ambiguous information. The possibilities and reasons for misperceptions and disagreements are legion. For these reasons, generalizations in this area are dif- ficult to develop, exceptions are common, and in many instances the outcomes will be influenced by factors that, from the standpoint of most the- ories, must be considered accidental. Important perceptual predispositions can be discovered, but often they will not be controlling.

Notes

1. See especially the following works by Harold and Margaret Sprout: Man-Milieu Relationship Hypotheses in the Context of International Politics (Princeton: Center of International Studies, 1956); The Ecological Perspective on Human Affairs (Princeton: Princeton University Press, 1965); and An Ecological Paradigm for the Study of International Politics (Princeton: Center of International Studies, Princeton University, Research Monograph No. 30, March, 1968).

2. "Decision-Making As an Approach to the Study of International Politics," in Richard Snyder, H. W. Bruck, and Burton Sapin, eds., Foreign Policy Decision-Making (New York: Free Press, 1962), p. 33 ....

3. Hirschi and Selvin,Delinquency Research, p. 38 .... 4. Quoted in Townsend Hoopes, The Limits of Inter-

vention (New York: McKay, 1969), p. 30. 5. The question of the existence and nature of reality

need not be treated here in its profound sense. For our purposes the consensus of later observers usually pro- vides an adequate operational definition of reality.

6. Arnold Wolfers, "The Actors in International Poli- tics," in Discord and Collaboration (Baltimore, Md: Johns Hopkins Press, 1962), pp. 3-24; Kenneth Waltz, Man, the State and War (New York: Columbia Univer-

sity Press, 1959); James Rosenau, "Pre-Theories and Theories of Foreign Policy," in R. Barry Farrell, ed.

1 Approaches to Comparative and International Polities (Evanston, IL: Northwestern University Press, 1966), pp. 29-92.

7. We refer to the international environment rather ~an the internatio~al system because _we ~re not dealing with systems theories. Our concern IS with explaining specific foreign policies rather than finding general pat- terns of interaction.

8. Quoted in Robert Butow, Tojo and the Coming of the War (Princeton: Princeton University Press, 1961), p. 506; quoted in Egmont Zechlin, "Cabinet versus Eco- nomic Warfare in Germany," in H. W. Koch, ed., The Origins of the First World War (London: Macmillan & Co., 1972), p. 165.

9. See K. Holsti, "National Role Conceptions in the Study of Foreign Policy," International Studies Quarterly 14 (September 1970), 233-309. 10, James Rosenau, "Pre-Theories and Theories of For-

eign Policy," pp. 47-48; Stanley Hoffmann, "Restraints and Choices in American Foreign Policy," Daedalus (Fall 1962), 692-94. 11. Wolfers, Discord and Collaboration, p. 13. 12. Paul Kecskemeti, Strategic Surrender (New York:

Atheneum, 1964), pp. 19-20; Ole Holsti, "The 1914 Case," The American Political Science Review 59 (June 1965), 365-78; Wolfers, Discord and Collaboration, p. 14. 13. Marshall Shulman, Stalin's Foreign Policy Reap-

praised (Cambridge, MA: Harvard University Press, 1963). 14. Michael Gordon, Conflict and Consensus in

Labour's Foreign Policy, 1914-1965 (Stanford, CA: Stan- ford University Press, 1969), p. 6; M.A. Fitzsimons, The Foreign Policy of the British Labour Government, 1945-1951 (Notre Dame, IN: University of Notre Dame Press, 1953), p. 26. 15. See Charles Osgood, "Behavior Theory and the

Social Sciences," in Roland Young, ed., Approaches to the Study of Politics (Evanston, IL: Northwestern University Press, 1958), pp. 217-44 .... 16. Daniel Ellsberg, "The Quagmire Myth and the

Stalemate Machine," in Papers on the War (New York: Simon and Schuster, 1972), pp. 42-135; Leslie Gelb, "Vietnam: The System Worked," Foreign Policy No. 3 (Summer 1971), 140-67; Klaus Knorr, "Failures in National Intelligence Estimates: The Case of the Cuban Missiles," World Polirics 16 (April 1967), 455--<i7 ....

OLE

Thea Worlt ity, ti, ityof

One of the more on crisis behavio individual stress. previously, focus induced stress or performance that ing. In assessing on cognitive perf, so against realisti on rationality incl from psychology capacity to rece: information abou generate the enti fragmentary knc quences of each order preferences on a single utilit straints exist in all making situations the impact on crit optic rationality. 1 of criteria might i to do the followin:

Identify adequa

Survey the

tion;

Estimate t well as the posi1

From Paul Gordon La reprinted by permissio