Hypotheses on Misperception
®
Robert Jervis
World Politics, Vol. 20, No. 3 (Apr., 1968), 454-479. Stable URL:
http://links.jstoi\org/sici?sici=0043-8871%28196804%2920%3A3%3C454%3AHOM%3E2.0.CO%3B2-4 World Politics is currently published by The Johns Hopkins University Press.
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/aboutyterms.html. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/joumals/jhup.html.
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.
JSTOR is an independent not-for-profit organization dedicated to creating and preserving a digital archive of scholarly journals. For more information regarding JSTOR, please contact suppoit@jstor.org.
http: // w w w j s tor. org/ Thu May 13 01:38:04 2004
HYPOTHESES ON MISPERCEPTION
By ROBERT JERVrS*
IN determining how he will behave, an actor must try to predict how others will act and how their actions will affect his values. The actor must therefore develop an image of others and of their intentions. This image may, however, turn out to be an inaccurate one; the actor may, for a number of reasons, misperceive both others' actions and their intentions. In this research note I wish to discuss the types of misper-ceptions of other states' intentions which states tend to make. The concept of intention is complex, but here we can consider it to comprise the ways in which the state feels it will act in a wide range of future contingencies. These ways of acting usually are not specific and well-developed plans. For many reasons a national or individual actor may not know how he will act under given conditions, but this problem cannot be dealt with here.
I. Previous Treatments of Perception in International Relations
Although diplomatic historians have discussed misperception in their treatments of specific events, students of international relations have generally ignored this topic. However, two sets of scholars have applied content analysis to the documents that flowed within and between governments in the six weeks preceding World War I. But the data have been put into quantitative form in a way that does not produce accurate measures of perceptions and intentions and that makes it impossible to gather useful evidence on misperception.1
The second group of theorists who have explicitly dealt with general questions of misperception in international relations consists of those, like Charles Osgood, Amitai Etzioni, and, to a lesser extent, Kenneth Boulding and J. David Singer, who have analyzed the cold war in
*I am grateful to the Harvard Center for International Affairs for research support. An earlier version, of this research note was presented at the International Studies Association panel of the New England Political Science Association in April 1967. I have benefited from comments by Rohm Art, Alexander George, Paul Kecskemeti., Paul Leary, Thomas Schilling, James Sehlesinger, Morton Schwartz, and Aaron. Wildavsky.
1 See, for example, Ole Holsri, Robert North, and Richard Brody, "Perception and Action, in. the 1914 Crisis," in I- David Singer, ed., Quantitative International Politics (New York. ro,68). For a fuller discussion of the Stanford content analysis studies and the general problems of quantification, see my "The Costs of the Quantitative Study of International Relations," in Klaus Knorr and James N. Rosenau, eds., Contending Approaches to International Politics (forthcoming).
HYPOTHESES ON MISPERCEPTION
455
terms of a spiral of misperception.2 This approach grows partly out of the mathematical theories of L. F, Richardson5 and partly out of findings of social and cognitive psychology, many of which will be discussed in this research note.
These authors state their case in general, if not universal, terms, but do not provide many historical cases that are satisfactorily explained by their theories. Furthermore., they do not deal with any of the numerous instances that contradict their notion of the self-defeating aspects of the use of power. They ignore the fact that states are not individuals and that the findings of psychology can be applied to organizations only with great care. Most important, their theoretical analysis is for the most part of reduced value because it seems largely to be a product of their assumption that the Soviet Union is a basically status-quo power whose apparently aggressive behavior is a product of fear of the West. Yet they supply little or no evidence to support this view. Indeed, the explanation for the differences of opinion between the spiral theorists and the proponents of deterrence lies not in differing general views of international relations, differing values and morality,* or differing methods of analysis/ but in differing perceptions of Soviet intentions.
II. Theories—Necessary and Dangerous
Despite the limitations of their approach, these writers have touched on a vital problem that has not been given systematic treatment by theorists of international relations. The evidence from both psychology and history overwhelmingly supports the view (which may be labeled Hypothesis i) that decision-makers tend to fit incoming information into their existing theories and images. Indeed, their theories and images play a large part in determining what they notice. In other words, actors tend to perceive what they expect. Furthermore (Hypothesis la), a theory will have greater impact on an actor's interpretation of data (a) the greater the ambiguity of the data and (b) the
2 See, for example, Osgood, An Alternative to War or Surrender (Urbana 1962); Etzioni, The Hard Way to Peace (New York 1962); Boulding, "National Images and International Systems/' Journal of Conflict Resolution, m (June 1959), 120-31; and Singer, Deterrence, Arms Control, and Disarmament (Columbus ra.62).
3 Statistics of Deadly Quarrels (Pittsburgh i960) and Arms and Insecurity (Chicago 1960). For nonmathemacieians a fine summary of Richardson's work is Anatol Rapo-port's "L. F. Richardson's Mathematical Theory of War," Journal of Conflict Resolution, 1 (September 1957), 249-99-
4 See Philip Green, Deadly Logic (Columbus ig66) • Green, "Method and Substance in the Arms Debate," World Politics, xvi (July 1964), 642-67; and Robert A. Levine^ "Fact and Morals in the Arms Debate," World Politics, xiv (January 1962), 239-58,
B See Anatol Rapoport, Strategy arid Conscience (New York 1964).
456
WORLD POLITICS
higher the degree of confidence with which the actor holds the theory.* For many purposes we can use the concept of differing levels of perceptual thresholds to deal with the fact that it takes more, and more unambiguous, information for an actor to recognize an unexpected phenomenon than an expected one. An experiment by Bruner and Postman determined "that the recognition threshold for . . . incongruous playing cards (those with suits and color reversed) is significantly higher than the threshold for normal cards."7 Not only are people able to identify normal (and therefore expected) cards more quickly and easily than incongruous (and therefore unexpected) ones, but also they may at first take incongruous cards for normal ones.
However, we should not assume, as the spiral theorists often do, that it is necessarily irrational for actors to adjust incoming information to fit more closely their existing beliefs and images. ("Irrational" here describes acting under pressures that the actor would not admit as legitimate if he were conscious of them.) Abelson and Rosenberg label as "psycho-logic" the pressure to create a "balanced" cognitive structure—i.e., one in which "all relations among 'good elements1 [in one's attitude structure] are positive (or null), all relations among 'bad elements' are positive (or null)^ and all relations between good and bad elements are negative (or null)." They correctly show that the "reasoning [this involves] would mortify a logician."8 But those who have tried to apply this and similar cognitive theories to international relations have usually overlooked the fact that in many cases there are important logical links between the elements and the processes they describe which cannot be called "psycho-logic." (I am here using the term "logical" not in the narrow sense of drawing only those conclusions that follow necessarily from the premises, but rather in the sense of conforming to generally agreed-upon rules for the treating of evidence,) For example, Osgood claims that psycho-logic is displayed when the Soviets praise a man or a proposal and people in the West react by distrusting the object of this praise.9 But if a person believes that the Russians are aggressive, it is logical for him to be suspicious of their moves. When we say that a decision-maker "dislikes" another
d Floyd Allport, Theories of Perception and the Concept of Structure (New York 1955), 3825 Ole Holsti, "Cognitive Dynamics and Images ai the Enemy," in David Finlay, Ote Holsti, and Richard Fagen, Enemies in Politics (Chicago 1967), 70.
7 Jerome Bruner and Leo Postman, "On the Perceptions of Incongruity: A Paradigm," in Jerome Bruner and David Krech, eds., Perception and Personality (Durham, N.C., 1949)* 210.
8 Robert Abelson and Milton Rosenberg, "Symbolic Psycho-logic," Behavioral Science, rn (January 1958), 4-5,
flP. 27.
HYPOTHESES ON MlSPERCEPTION
457
state this usually means that he believes that that other state has policies conflicting with those of his nation. Reasoning and experience indicate to the decision-maker that the "disliked" state is apt to harm his state's interests. Thus in these cases there is no need to invoke "psychologic," and it cannot be claimed that the cases demonstrate the substitution of "emotional consistency for rational consistency."10
The question of the relations among particular beliefs and cognitions can often be seen as part of the general topic of the relation of incoming bits of information to the receivers1 already established images. The need to fit data into a wider framework of beliefs, even if doing so does not seem to do justice to individual facts, is not, or at least is not only, a psychological drive that decreases the accuracy of our perceptions of the world, but is "essential to the logic of inquiry."11 Facts can be interpreted, and indeed identified, only with the aid of hypotheses and theories. Pure empiricism is impossible, and it would be unwise to revise theories in the light of every bit of information that does not easily conform to them.11 No hypothesis can be expected to account for all the evidence, and if a prevailing view is supported by many theories and by a large pool of findings it should not be quickly altered. Too little rigidity can be as bad as too much.13
This is as true in the building of social and physical science as it is
10 Ibid., 26.
111 have borrowed this phrase from Abraham Kaplan, who uses it in a different but related context in The Conduct of Inquiry (San Francisco 1964), 86.
12 The spiral theorises are not the only ones to ignore the limits of empiricism, Roger Hilsman found that most consumers and producers of intelligence felt that intelligence should not deal with hypotheses, but should only provide the policy-makers with "all the facts" (Strategic Intelligence and National Decision 1 [Glencoe 1956], 46). The close interdependence between hypotheses and facts is overlooked partly because of the tendency to identify "hypotheses" with "policy preferences."
13 Karl Deutsch interestingly discusses a related question when he argues, "Autonomy . . . requires both intake from the present and recall from memory, and selfhood can be seen in just this continuous balancing of a limited present and a limited past. . . . No further self-determination is possible if either openness or memory is lost, . . . To the extent that [systems cease to be able to take in new information], they approach the behavior of a bullet or torpedo: their future action becomes almost completely determined by their past. On the other hand, a person without memory, an organization without values or policy . . . —all these no longer steer, but drift: their behavior depends little on their past and almost wholly on their present, Driftwood and the bullet are thus each the epitome of another kind of loss of self-control . . ." (National-imt and Social Communication [Cambridge, Mass., 1954], 167-68). Also see Deutsch's The Nerves of Government {New York 1963), 98-109, 200-256. A physicist makes a similar argument; "It is clear that if one is too attached to one's preconceived model, one will miss all radical discoveries. It is amazing to what degree one may fail to register mentally an observation which does not fit the initial image. . . , On the other hand, if one is too open-minded and pursues every hitherto unknown phenomenon, one is almost certain to lose oneself in trivia" (Martin Deutsch, "Evidence and Inference in Nuclear Research," in Daniel Lerner, ed., Evidence and Inference [Glencoe 1958], 102).
458
WORLD POLITICS
in policy-making.1* While it is terribly difficult to know when a finding throws serious doubt on accepted theories and should be followed up and when instead it was caused by experimental mistakes or minor errors in the theory, it is clear that scientists would make no progress if they followed Thomas Huxley's injunction to "sit down before fact as a mere child, be prepared to give up every preconceived notion, follow humbly wherever nature leads, or you will learn nothing.'116
As Michael Polanyi explains, "It is true enough that the scientist must be prepared to submit at any moment to the adverse verdict of observational evidence. But not blindly. . . . There is always the possibility that, as in [the cases of the periodic system of elements and the quantum theory of light], a deviation may not affect the essential correctness of a proposition. . . . The process of explaining away deviations is in fact quite indispensable to the daily routine of research," even though this may lead to the missing of a great discovery.18 For example, in 1795, the astronomer Lalande did not follow up observations that contradicted the prevailing hypotheses and could have led him to discover the planet Neptune.1 T
Yet we should not be too quick to condemn such behavior. As Thomas Kuhn has noted, "There is no such thing as research without counter-instances."18 If a set of basic theories—what Kuhn calls a paradigm—has been able to account for a mass of data, it should not be lightly trifled with. As Kuhn puts it: "Lifelong resistance, particularly from those whose productive careers have committed them to an older tradition of normal science [i.e., science within the accepted paradigm], is not a violation of scientific standards but an index to the nature of scientific research itself. The source of resistance is the assurance that the older paradigm will ultimately solve all its problems, that nature
" Raymond Bauer, "Problems of Perception and the Relations Between the U.S. and the Soviet Union," journal of Conflict Resolution, v (September 1961), 223-29.
15 Quoted in W. L B. Beveridge> The Art of Scientific Investigation) 3rd ed. (London *957)> 50-
16 Scienee, Faith, and Society (Chicago 1964), 31. For a farther discussion of this problem, see ihid^ r.6, 26-41, go-94; Polanyi, Personal Knowledge (London 1958), 8-15, jo, 143-68, 269-93, 310-n; Thomas Kuhn, The Structure of Scientific Revolution (Chicago 1964); Kuhn, "The Function of Dogma in Scientific Research," in A. C. Cromhie, ed., Scientific Change (New York r.963), 344-69; the comments on Kuhn's paper by Hall, Polanyi, and Toulmin, and Kuhn's reply, ibid.t 370-95. For a related discussion of these points from a different perspective, see Norman Storer, The Social System of Science (New York 1960), 116-22.
17 "He found that the position of one star relative to others . . . had shifted. Lalande was a good astronomer and knew that such a shift was unreasonable, He crossed out his first observation, put a question mark next to the second observation, and let the matter go" (Jerome Bruner, Jacqueline Goodnow, and George Austin, A Study of Thinking [New York 1962], 105).
13 The Structure of Scientific Revolution, jg.
HYPOTHESES ON MISPERCEPTION
459
can be shoved into the box the paradigm provides. Inevitably, at times of revolution, that assurance seems stubborn and pig-headed as indeed it sometimes becomes. But it is also something more. That same assurance is what makes normal science or puzzle-solving science possible."19
Thus It is important to see that the dilemma of how "open" to be to new information is one that inevitably plagues any attempt at understanding in any field. Instances in which evidence seems to be ignored or twisted to fit the existing theory can often be explained by this dilemma instead of by illogical or nonlogical psychological pressures toward consistency. This is especially true of decision-makers' attempts to estimate the intentions of other states, since they must constantly take account of the danger that the other state is trying to deceive them.
The theoretical framework discussed thus far, together with an examination of many cases, suggests Hypothesis a: scholars and decision-makers are apt to err by being too wedded to the established view and too closed to new information, as opposed to being too willing to alter their theories.20 Another way of making this point is to argue that actors tend to establish their theories and expectations prematurely. In politics, of course, this is often necessary because of the need for action. But experimental evidence indicates that the same tendency also occurs on the unconscious level. Bruner and Postman found that "perhaps the greatest single barrier to the recognition of incongruous stimuli is the tendency for perceptual hypotheses to fixate after receiving a minimum of confirmation. . . . Once there had occurred in these cases a partial confirmation of the hypothesis... it seemed that nothing could change the subject's report."21
ia/AW., 150-51.
20 Requirements of effective political leadership may lead decision-makers to voice fewer doubts than they have about existing policies and images, but this constraint can only partially explain this phenomenon. Similar calculations of political strategy may contribute to several of the hypotheses discussed below.
S1P. 221. Similarly, in experiments dealing with his subjects' perception of other people, Charles Dai ley found that "premature judgment appears to make new data harder to assimilate than when the observer withholds judgment until all data are seen. It seems probable , . . that the observer mistakes his own inferences for facts" ("The Effects of Premature Conclusion Upon the Acquisition of Understanding of a Person," journal of Psychology, xxx [January 1952], 149-50). For other theory and evidence on this point, see Bruner, "On Perceptual Readiness," Psychological Review, lxiv (March 1957), j 23-53; Gerald Davidson, "The Negative Effects of Early Exposure to Suboptimal Visual Stimuli," Journal of Personality, xxxn (June 1964), 278-95; Albert Myers, "An Experimental Analysis of a Tactical Blunder,'1 Journal of Abnormal and Social Psychology, lxix (November 1964), 493-98; and Dale Wyacc and Donald Campbell, "On the Liability of Stereotype or Hypothesis," Journal of Abnormal and Sacid Psychology, XLtv (October 1950), 496-500. It should be noted that this tendency makes "incremental" decision-making more likely (David Braybrooke and Charles Lindblom, A Strategy of Deciiian [New York 1963]), but the results of this process may lead the actor further from his goals.
460
WORLD POLITICS
However, when we apply these and other findings to politics and discuss kinds of misperception3 we should not quickly apply the label of cognitive distortion. We should proceed cautiously for two related reasons. The first is that the evidence available to decision-makers almost always permits several interpretations. It should be noted that there are cases of visual perception in which different stimuli can produce exactly the same pattern on an observer's retina. Thus, for an observer using one eye the same pattern would be produced by a sphere the size of a golf ball which was quite close to the observer, by a baseball-sized sphere that was further away, or by a basketball-sized sphere still further away. Without other clues, the observer cannot possibly determine which of these stimuli he is presented with, and we would not want to call his incorrect perceptions examples of distortion. Such cases, relatively rare in visual perception, are frequent in international relations. The evidence available to decision-makers is almost always very ambiguous since accurate clues to others' intentions are surrounded by noise22 and deception. In most cases, no matter how long, deeply, and "objectively" the evidence is analyzed, people can differ in their interpretations, and there are no general rules to indicate who is correct.
The second reason to avoid the label of cognitive distortion is that the distinction between perception and judgment, obscure enough in individual psychology, is almost absent in the making of inferences in international politics. Decision-makers who reject information that contradicts their views—or who develop complex interpretations of it— often do so consciously and explicitly. Since the evidence available contains contradictory information, to make any inferences requires that much information be ignored or given interpretations that will seem tortuous to those who hold a different position.
Indeed, if we consider only the evidence available to a decision-maker at the time of decision, the view later proved incorrect may be supported by as much evidence as the correct one—or even by more. Scholars have often been too unsympathetic with the people who were proved wrong. On closer examination, it is frequently difficult to point to differences between those who were right and those who were wrong with respect to their openness to new information and willingness to modify their views. Winston Churchill, for example; did not open-mindedly view each Nazi action to see if the explanations provided by the appcasers accounted for the data better than his own beliefs. Instead,
41 For a use of this concept in political communication,, see Roberta Wohlstetter, Pearl Harbor (Stanford 1962).
HYPOTHESES ON MlSPERCEPTlON
461
like Chamberlain, he fitted each bit of ambiguous information into his own hypotheses. That he was correct should not lead us to overlook the fact that his methods of analysis and use of theory to produce cognitive consistency did not basically differ from those of the appeasers.13
A consideration of the importance of expectations in influencing perception also indicates that the widespread belief in the prevalence of "wishful thinking" may be incorrect^ or at least may be based on inadequate data. The psychological literature on the interaction between affect and perception is immense and cannot be treated here, but it should be noted that phenomena that at first were considered strong evidence for the impact of affect on perception often can be better treated as demonstrating the influence of expectations.1* Thus} in international relations, cases like the United States' misestimation of the political climate in Cuba in April 1961, which may seem at first glance to have been instances of wishful thinking, may instead be more adequately explained by the theories held by the decision-makers (e.g., Communist governments are unpopular). Of courscj desires may have an impact on perception by influencing expectations, but since so many other factors affect expectations, the net influence of desires may not be great.
There is evidence from both psychology55 and international relations that when expectations and desires clash3 expectations seem to be more important. The United States would like to believe that North Vietnam is about to negotiate or that the USSR is ready to give up what the United States believes is its goal of world domination, but ambiguous
73 Similarly, Robert Caulondre, the French ambassador to Berlin in 1939, was one of the few diplomats to appreciate, the Nazi threat. Partly because of his earlier service in the USSR, "he was painfully sensitive to the threat of a Berlin-Moscow agreement, He noted with foreboding that Hitler had not attacked Russia in his Reichstag address of April 28. . . . So it went all spring and summer, the ambassador relaying each new evidence of the impending diplomatic revolution and adding to his admonitions his pleas for decisive counteraction" (Franklin Ford and Carl Schorske, "The Voice in the Wilderness: Robert Coulondre," in Gordon Craig and Felix Gilbert, eds., The Diplomats, Vol III [New York 1563] 573-74)- His hypotheses were correct, but it is difficult to detect differences between the way he and those ambassadors who were incorrect, like Neville Henderson, selectively noted and interpreted information. However, to the extent that the fear of war influenced the ippeasers' perceptions of. Hitler's intentions., the appeasers' views did have an element of psycho-logic that was not present in their opponents' position.
"See, for example, Donald Campbell, "Systematic Error on the Part of Human Links in Communications Systems," Information and Control, 1 (iggS), 346-50; and Leo Postman, 'The Experimental Analysis of Motivational Factors in Perception," in Judson S, Brown, ed., Current Theory and Research in Motivation (Lincoln, Neb.,
2B Dale Wyatt and Donald Campbell, "A Study of Interviewer Bias as Related to Interviewer's Expectations and Own Opinions," International Journal of Opinion and Attitude Research, iv (Spring 1950), 77-83.
462
WORLD POLITICS
evidence is seen to confirm the opposite conclusion, which conforms to the United States' expectations. Actors are apt to be especially sensitive to evidence of grave danger if they think they can take action to protect themselves against the menace once it has been detected.
III. Safeguards
Can anything then be said to scholars and decision-makers other than "Avoid being either too open or too closed, but be especially aware of the latter danger1'? Although decision-makers will always be faced with ambiguous and confusing evidence and will be forced to make inferences about others which will often be inaccurate, a number of safeguards may be suggested which could enable them to minimize their errors. First, and most obvious, decision-makers should be aware that they do not make "unbiased" interpretations of each new bit of incoming information, but rather are inevitably heavily influenced by the theories they expect to be verified. They should know that what may appear to them as a self-evident and unambiguous inference often seems so only because of their preexisting beliefs. To someone with a different theory the same data may appear to be unimportant or to support another explanation. Thus many events provide less independent support for the decision-makers1 images than they may at first realize. Knowledge of this should lead decision-makers to examine more closely evidence that others believe contradicts their views.
Second, decision-makers should see if their attitudes contain consistent or supporting beliefs that are not logically linked. These may be examples of true psycho-logic. While it is not logically surprising nor is it evidence of psychological pressures to find that people who believe that Russia is aggressive are very suspicious of any Soviet move, other kinds of consistency are more suspect. For example, most people who feel that it is important for the United States to win the war in Vietnam also feel that a meaningful victory is possible. And most people who feel defeat would neither endanger U.S. national security nor be costly in terms of other values also feel that we cannot win. Although there are important logical linkages between the two parts of each of these views (especially through theories of guerrilla warfare), they do not seem strong enough to explain the degree to which the opinions are correlated. Similarly, in Finland in the winter of 1939, those who felt that grave consequences would follow Finnish agreement to give Russia a military base also believed that the Soviets would withdraw their demand if Finland stood firm. And those who felt that concessions would not lead to loss of major values also believed that Russia would
HYPOTHESES ON MISPERCEPTION
463
fight if need be.29 In this country, those who favored a nuclear test ban tended to argue that fallout was very harmful, that only limited improvements in technology would flow from further testing, and that a test ban would increase the chances for peace and security. Those who opposed the test ban were apt to disagree on all three points. This does not mean, of course, that the people holding such sets of supporting views were necessarily wrong in any one element. The Finns who wanted to make concessions to the USSR were probably correct in both parts of their argument. But decision-makers should be suspicious if they hold a position in which elements that are not logically connected support the same conclusion. This condition is psychologically comfortable and makes decisions easier to reach (since competing values do not have to be balanced off against each other). The chances are thus considerable that at least part of the reason why a person holds some of these views is related to psychology and not to the substance of the evidence.
Decision-makers should also be aware that actors who suddenly find themselves having an important shared interest with other actors have a tendency to overestimate the degree of common interest involved. This tendency is especially strong for those actors (e.g., the United States, at least before 1950) whose beliefs about international relations and morality imply that they can cooperate only with "good" states and that with those states there will be no major conflicts. On the other hand, states that have either a tradition of limited cooperation with others (e.g., Britain) or a strongly held theory that differentiates occasional from permanent allies27 (e.g., the Soviet Union) find it easier to resist this tendency and need not devote special efforts to combating its danger,
A third safeguard for decisionmakers would be to make their assumptions, beliefs, and the predictions that follow from them as explicit as possible. An actor should try to determine, before events occur, what evidence would count for and against his theories. By knowing what to expect he would know what to be surprised by, and surprise could indicate to that actor that his beliefs needed reevaluation.28
A fourth safeguard is more complex. The decision-maker should try
29 Max Jacobson, The Diplomacy of the Winter War (Cambridge, Mass., igfii}, 27 Raymond Aron., Peace and War (Garden City 1966), 29.
aa Cf. Rutin,, The Structure of Scientific Revolution, 65. A fairly high degree of knowledge is needed before one can state precise expectations. One indication of the lack of international relations theory is that most of us are not sure what "naturally" flows from our theories and what constitutes either "puzzles" to be further explored with the paradigm or "anomalies" that cast doubt on the basic theories.
464 WORLD POLITICS
to prevent individuals and organizations from letting their main task, political future, and identity become tied to specific theories and images of other actors.29 If this occurs, subgoals originally sought for their contribution to higher ends will take on value of their own, and information indicating possible alternative routes to the original goals will not be carefully considered. For example, the U.S. Forest Service was unable to carry out its original purpose as effectively when it began to see its distinctive competence not in promoting the best use of lands and forests but rather in preventing all types of forest fires.30
Organizations that claim to be unbiased may not realize the extent to which their definition of their role has become involved with certain beliefs about the world. Allen Dulles is a victim of this lack of understanding when he says, "I grant that we are all creatures of prejudice, including CIA officials, but by entrusting intelligence coordination to our central intelligence service, which is excluded from policy-making and is married to no particular military hardware, we can avoid, to the greatest possible extent, the bending of facts obtained through intelligence to suit a particular occupational viewpoint "31 This statement overlooks the fact that the CIA has developed a certain view of international relations and of the cold war which maximizes the importance of its information-gathering, espionage, and subversive activities. Since the CIA would lose its unique place in the government if it were decided that the "back alleys" of world politics were no longer vital to U.S. security, it is not surprising that the organization interprets information in a way that stresses the continued need for its techniques.
Fifth, decision-makers should realize the validity and implications of Roberta Wohlstetter's argument that "a willingness to play with material from different angles and in the context of unpopular as well as popular hypotheses is an essential ingredient of a good detective, whether the end is the solution of a crime or an intelligence estimate."32 However, it is often difficult, psychologically and politically, for any one person to do this. Since a decision-maker usually cannot get "unbiased" treatments of data, he should instead seek to structure conflicting biases into the decision-making process. The decision-maker, in other words, should have devil's advocates around. Just as, as Neustadt points out,38 the decision-maker will want to create conflicts among his
29 See Philip Selzmck, Leadership in Administration (Evanston 1957).
30 Ashley Schiff, Fire and Water: Scientific Heresy in the Forest Service (Cambridge, Mass.j 1962). Despite its title, this book is a fascinating and valuable study.
31 The Craft of Intelligence (New York 1963), 53.
32 P. 302. See Beveridge, 93, for a discussion of the idea that the scientist should keep in mind as many hypotheses as possible when conducting and analyzing experiments.
33 Presidential Power (New York i$6a).
HYPOTHESES ON MISPERCEPTjON
465
subordinates in order to make appropriate choices, so he will also want to ensure that incoming information is examined from many different perspectives with many different hypotheses in mind. To some extent this kind of examination will be done automatically through the divergence of goals, training, experience, and information that exists in any large organization. But in many cases this divergence will not be sufficient. The views of those analyzing the data will still be too homogeneous, and the decision-maker will have to go out of his way not only to cultivate but to create differing viewpoints.
While all that would be needed would be to have some people examining the data trying to validate unpopular hypotheses, it would probably be more effective if they actually believed and had a stake in the views they were trying to support. If in 1941 someone had had the task of proving the view that Japan would attack Pearl Harbor, the government might have been less surprised by the attack. And only a person who was out to show that Russia would take objectively great risks would have been apt to note that several ships with especially large hatches going to Cuba were riding high in the water, indicating the presence of a bulky but light cargo that was not likely to be anything other than strategic missiles. And many people who doubt the wisdom of the administration's Vietnam policy would be somewhat reassured if there were people in the government who searched the statements and actions of both sides in an effort to prove that North Vietnam was willing to negotiate and that the official interpretation of such moves as the Communist activities during the Tet truce of 1967 was incorrect.
Of course all these safeguards involve costs. They would divert resources from other tasks and would increase internal dissension. Determining whether these costs would be worth the gains would depend on a detailed analysis of how the suggested safeguards might be implemented. Even if they were adopted by a government, of course, they would not eliminate the chance of misperceptlon. However, the safeguards would make it more likely that national decision-makers would make conscious choices about the way data were interpreted rather than merely assuming that they can be seen in only one way and can mean only one thing. Statesmen would thus be reminded of alternative images of others just as they are constantly reminded of alternative policies.
These safeguards are partly based on Hypothesis 3: actors can more easily sssimilatc into their established image of another actor information contradicting that image if the information is transmitted and
466
WORLD POLITICS
considered bit by bit than if it comes all at once. In the former case, each piece of discrepant data can be coped with as it arrives and each of the conflicts with the prevailing view will be small enough to go unnoticed, to be dismissed as unimportant, or to necessitate at most a slight modification of the image (e.g., addition of exceptions to the rule). When the information arrives in a block, the contradiction between it and the prevailing view is apt to be much clearer and the probability of major cognitive reorganization will be higher.
IV. Sources of Concepts
An actor's perceptual thresholds—and thus the images that ambiguous information is apt to produce—are influenced by what he has experienced and learned about.34 If one actor is to perceive that another fits in a given category he must first have, or develop, a concept for that category. We can usefully distinguish three levels at which a concept can be present or absent. First, the concept can be completely missing. The actor's cognitive structure may not include anything corresponding to the phenomenon he is encountering. This situation can occur not only in science fiction, but also in a world of rapid change or in the meeting of two dissimilar systems. Thus China's image of the Western world was extremely inaccurate in the mid-nineteenth century, her learning was very slow, and her responses were woefully inadequate. The West was spared a similar struggle only because it had the power to reshape the system it encountered. Once the actor clearly sees one instance of the new phenomenon, he is apt to recognize it much more quickly in the future.35 Second, the actor can know about a concept but not believe that it reflects an actual phenomenon. Thus Communist and Western decision-makers are each aware of the other's explanation of how his system functions, but do not think that the concept cor-
34Most psychologists argue that this influence also holds for perception of shapes. For data showing that people in different societies differ in respect to their predisposition to experience certain optical illusions and for a convincing argument that this difference can be explained by the societies' different physical environments, which have led their people to develop different patterns of drawing inferences from ambiguous visual cues, see Marshall Sega 11, Donald Campbell, and Melville Herskovits, The Influence of Culture ok Visual Perceptions (Indianapolis 1966).
35 Thus when Brimer and Postman's subjects first were presented with incongruous playing cards (Le., cards in which symbols and colors of the suits were not marching, producing red spades or black diamonds), long exposure times were necessary for correct identification. Bur once a subject correctly perceived the card and added this type o£ card to his repertoire of categories, he was able to identify other incongruous cards much more quickly. For an analogous example—in this case, changes in the analysis of aerial reconnaissance photographs of an enemy's secret weapons-testing facilities produced by the belief that a previously unknown object may be present—see David Irving, The Mare's Nest (Boston 1064), 66-6j, 274-75.
HYPOTHESES ON MlSPERCEPTiON
467
responds to reality. Communist elites, furthermore, deny that anything could correspond to the democracies' description of themselves. Third, the actor may hold a concept, but not believe that another actor fills it at the present moment Thus the British and French statesmen of the 1930's held a concept of states with unlimited ambitions. They realized that Napoleons were possible, but they did not think Hitler belonged in that category. Hypothesis 4 distinguishes these three cases: mispercep-tion is most difficult to correct in the case of a missing concept and least difficult to correct in the case of a recognized but presumably unfilled concept. All other things being equal (e.g., the degree to which the concept is central to the actor's cognitive structure), the first case requires more cognitive reorganization than does the second, and the second requires more reorganization than the third.
However, this hypothesis does not mean that learning will necessarily be slowest in the first case, for if the phenomena are totally new the actor may make such grossly inappropriate responses that he will quickly acquire information clearly indicating that he is faced with something he does not understand. And the sooner the actor realizes that things are not—or may not be—what they seem, the sooner he is apt to correct his image.36
Three main sources contribute to decision-makers' concepts of international relations and of other states and influence the level of their perceptual thresholds for various phenomena. First, an actor's beliefs about his own domestic political system are apt to be important. In some cases, like that of the USSR, the decision-makers' concepts are tied to an ideology that explicitly provides a frame of reference for viewing foreign affairs. Even where this is not the case, experience with his own system will partly determine what the actor is familiar with and what he is apt to perceive in others. Louis Hartz claims, "It is the absence of the experience of social revolution which is at the heart of the whole American dilemma.... In a whole series of specific ways it enters into our difficulty of communication with the rest of the world. We find it difficult to understand Europe's 'social question'. . . . We are not familiar with the deeper social struggles of Asia and hence tend to interpret even reactionary regimes as 'democratic.'"37 Similarly, George Kennan argues that in World War I the Allied powers, and especially America, could not understand the bitterness and violence of others' internal conflicts: ". . . The inability of the Allied statesmen to picture to themselves the passions of the Russian civil war [was partly caused
M Bruner and Postmark 220.
37 The Liberal Tradition in America (New York 1955), 306.
468
WORLD POLITICS
by the fact that] we represent... a society in which the manifestations of evil have been carefully buried and sublimated in the social behavior of people, as in their very consciousness. For this reason, probably, despite our widely traveled and outwardly cosmopolitan livesj the mainsprings of political behavior in such a country as Russia tend to remain concealed from our vision."38
Second, concepts will be supplied by the actor's previous experiences. An experiment from another field illustrates this. Dearborn and Simon presented business executives from various divisions (e.g., sales, accounting, production) with the same hypothetical data and asked them for an analysis and recommendations from the standpoint of what would be best for the company as a whole. The executives' views heavily reflected their departmental perspectives.38 William W. Kaufmann shows how the perceptions of Ambassador Joseph Kennedy were affected by his past: "As befitted a former chairman of the Securities Exchange and Maritime Commissions, his primary interest lay in economic matters*. . . The revolutionary character of the Nazi regime
was not a phenomenon that he could easily grasp____It was far simpler,
and more in accord with his own premises, to explain German aggressiveness in economic terms. The Third Reich was dissatisfied, authoritarian, and expansive largely because her economy was unsound."*0 Similarly it has been argued that Chamberlain was slow to recognize Hitler's intentions partly because of the limiting nature of his personal background and business experiences.41 The impact of training and ex-
38 Russia and the West Under Lenin and Stalin (New York 1962), 142-43.
38 DeWitt Dearborn, and Herbert Simon, "Selective Perception: A Note on the Departmental Identification of Executives," Sociometry, xxt (June 1958), 140-44.
"Two American Ambassadors: Bullitt and Kennedy/' in Craig and Gilbert, 358-59.
41 Hugh Trevor-Roper puts this point well: "Brought up as a business man, successful in municipal politics, [Chamberlain's] outlook was entirely parochial. Educated Conservative aristocrats like Churchill, Eden, and Cran borne, whose families had long been used to political responsibility, had seen revolution and revolutionary leaders before, in their own history, and understood them cor reedy; but the Chamberlains, who had run from radical imperialism to timid conservatism in a generation of life in Birmingham, had no such understanding of history or the world: to them the scope of human politics was limited by their own parochial horizons, and Neville Chamberlain could not believe that Hitler was fundamentally different from himself. If Chamberlain wanted peace, so must Hitler" ("Munich—Its Lessons Ten Years Later," in Francis Loewenheim, ed., Peace or Appeasement? [Boston 1965], 152-53). For a similar view see A. L. Rowse, Appeasement (New York 1963), 117.
But Donald Lammers points out that the views of many prominent British public figures in the ig3o's do not fit this generalization (Explaining Munich [Stanford 1966], 13-140). Furthermore, arguments that stress the importance of the experiences and views of the actors' ancestors do not explain the links by which these influence the actors themselves. Presumably Churchill and Chamberlain read the same history books in school and had the same basic information about Britain's past role in the world. Thus what has to be demonstrated is that in their homes aristocrats like Churchill learned different things about politics and human nature than did middle-class people
HYPOTHESES ON MISPERCEPTION
469
perience seems to be demonstrated when the background of the ap-peasers is compared to that of their opponents. One difference stands out: "A substantially higher percentage of the anti-appeasers (irrespective of class origins) had the kind of knowledge which comes from close acquaintance, mainly professional, with foreign affairs."43 Since members of the diplomatic corps are responsible for meeting threats to the nation's security before these grow to major proportions and since they have learned about cases in which aggressive states were not recognized as such until very late, they may be prone to interpret ambiguous data as showing that others are aggressive. It should be stressed that we cannot say that the professionals of the 1930*3 were more apt to make accurate judgments of other states. Rather, they may have been more sensitive to the chance that others were aggressive. They would then rarely take an aggressor for a status-quo power, but would more often make the opposite error.43 Thus in the years before World War I the permanent officials in the British Foreign Office overestimated German aggressiveness.44
A parallel demonstration in psychology of the impact of training on perception is presented by an experiment in which ambiguous pictures were shown to both advanced and beginning police-administration students. The advanced group perceived more violence in the pictures than did the beginners. The probable explanation is that "the law enforcer may come to accept crime as a familiar personal experience, one which he himself is not surprised to encounter. The acceptance of crime as a familiar experience in turn increases the ability or readiness to perceive violence where clues to it are potentially available."45 This
like Chamberlain and that these experiences had a significant impact. Alternative I y, it could be argued that the patterns of child-rearing prevalent among the aristocracy influenced the children's personalities in a way that made them more likely to see others as aggressive. *2IMd., 15.
*3 During a debate on appeasement in the House of Commons, Harold Nicobon declared, "I know that those of us who believe in. the traditions of our policy, . , . who believe that one great function of this country is to maintain moral standards in Europe, to maintain a settled pattern at international relations, not to make friends with people who are demonstrably evil , . . —I know that those who hold such beliefs are accused of possessing the Foreign Office mind, I thank God that I possess the Foreign Office mind" (quoted in Martin Gilbert, The Roots of Appeasement [New York 1966], 187), But the qualities Nicolson mentions and applauds may be related to a more basic attribute of "the Foreign OfEce mind"—suspiciousness.
** George Monger, The End of Isolation (London 1963). I am also indebted to Frederick Collignon for his unpublished manuscript and several conversations on this point.
*s Hans Toch and Richard Schulte, "Readiness to Perceive Violence as a Result of Police Training," British Journal of Psychology, mi (November 1961), 392 (original italics omitted). It should be stressed that one cannot say whether or not the advanced police students perceived the pictures "accurately." The point is that their training pre-
470
WORLD POLITICS
experiment lends weight to the view that the British diplomats' sensitivity to aggressive states was not totally a product of personnel selection procedures,
A third source of concepts, which frequently will be the most directly relevant to a decision-maker's perception of international relations, is international history. As Henry Kissinger points out, one reason why statesmen were so slow to recognize the threat posed by Napoleon was that previous events had accustomed them only to actors who wanted to modify the existing system, not overthrow it.w The other side of the coin is even more striking: historical traumas can heavily influence future perceptions. They can either establish a state's image of the other state involved or can be used as analogies. An example of the former case is provided by the fact that for at least ten years after the Franco-Prussian War most of Europe's statesmen felt that Bismarck had aggressive plans when in fact his main goal was to protect the status quo. Of course the evidence was ambiguous. The post-1871 Bismarckian maneuvers, which were designed to keep peace, looked not unlike the pre-1871 maneuvers designed to set the stage for war. But that the post-1871 maneuvers were seen as indicating aggressive plans is largely attributable to the impact of Bismarck's earlier actions on the statesmen's image of him.
A state's previous unfortunate experience with a type of danger can sensitize it to other examples of that danger. While this sensitivity may lead the state to avoid the mistake it committed in the past, it may also lead it mistakenly to believe that the present situation is like the past one. Santayana's maxim could be turned around: "Those who remember the past are condemned to make the opposite mistakes." As Paul Kecskemeti shows, both defenders and critics of the unconditional surrender plan of the Second World War thought in terms of the conditions of World War Li7 Annette Baker Fox found that the Scandinavian countries' neutrality policies in World War II were strongly influenced by their experiences in the previous war, even though vital aspects of the two situations were different. Thus "Norway's success [during the First World War] in remaining non-belligerent though
disposed them to see violence in ambiguous situations. Whether on balance they would make fewer perceptual errors and better decisions is very hard to determine. For an experiment showing that training can lead people to "recognize" an expected stimulus even when that stimulus is in fact not shown, see Israel Goldiamond and William F. Hawkins, "Vexierversuch: The Log Relationship Between Word-Frequency and Recognition Obtained in the Absence of Stimulus Words/1 Journal of Experimental Psychology, lvi (December 1058), 457-63.
46 A World Restored (New York 1964), 2-3.
"Strategic Surrender (New York 1964), 215-41.
HYPOTHESES ON MISPERCEPTION
471
pro-Allied gave the Norwegians confidence that their country could again stay out of war,11*5 And the lesson drawn from the unfortunate results of this policy was an important factor in Norway's decision to join NATO.
The application of the Munich analogy to various contemporary events has been much commented on, and I do not wish to argue the substantive points at stake. But it seems clear that the probabilities that any state is facing an aggressor who has to be met by force are not altered by the career of Hitler and the history of the io/jo's. Similarly the probability of an aggressor's announcing his plans is not increased (if anything, it is decreased) by the fact that Hitler wrote Mein Kampf. Yet decision-makers are more sensitive to these possibilities, and thus more apt to perceive ambiguous evidence as indicating they apply to a given case, than they would have been had there been no Nazi Germany.
Historical analogies often precede, rather than follow, a careful analysis of a situation (e.g., Truman's initial reaction to the news of the invasion of South Korea was to think of the Japanese invasion of Manchuria). Noting this precedence, however, does not show us which of many analogies will come to a decision-maker's mind. Truman could have thought of nineteenth-century European wars that were of no interest to the United States. Several factors having nothing to do with the event under consideration influence what analogies a decisionmaker is apt to make. One factor is the number of cases similar to the analogy with which the decision-maker is familiar. Another is the importance of the past event to the political system of which the decisionmaker is a part. The more times such an event occurred and the greater its consequences were, the more a decision-maker will be sensitive to the particular danger involved and the more he will be apt to see ambiguous stimuli as indicating another instance of this kind of event, A third factor is the degree of the decision-maker's personal involvement in the past case—in time, energy, ego, and position. The last-mentioned variable will affect not only the event's impact on the decision-maker's cognitive structure, but also the way he perceives the event and the lesson he draws. Someone who was involved in getting troops into South Korea after the attack will remember the Korean War differently from someone who was involved in considering the possible use of nuclear weapons or in deciding what messages should be sent to the Chinese. Greater personal involvement will usually give the event greater impact, especially if the decision-maker's own views
4fl The Power cf Small States (Chicago 1959), Si.
472
WORLD POLITICS
were validated by the event One need not accept a total application of learning theory to nations to believe that "nothing fails like success."*3 It also seems likely that if many critics argued at the time that the decision-maker was wrong, he will be even more apt to see other situations in terms of the original event. For example, because Anthony Eden left the government on account of his views and was later shown to have been correct, he probably was more apt to see as Hitlers other leaders with whom he had conflicts (e.g., Nasser). A fourth factor is the degree to which the analogy is compatible with the rest of his belief system. A fifth is the absence of alternative concepts and analogies. Individuals and states vary in the amount of direct or indirect political experience they have had which can provide different ways of interpreting data. Decision-makers who are aware of multiple possibilities of states* intentions may be less likely to seize on an analogy prematurely. The perception of citizens of nations like the United States which have relatively little history of international politics may be more apt to be heavily influenced by the few major international events that have been important to their country.
The first three factors indicate that an event is more apt to shape present perceptions if it occurred in the recent rather than the remote past. If it occurred recently, the statesman will then know about it at first hand even if he was not involved in the making of policy at the time. Thus if generals are prepared to fight the last war, diplomats may be prepared to avoid the last war. Part of the Anglo-French reaction to Hitler can be explained by the prevailing beliefs that the First World War was to a large extent caused by misunderstandings and could have been avoided by farsighted and nonbelligerent diplomacy. And part of the Western perception of Russia and China can be explained by the view that appeasement was an inappropriate response to Hitler.50
V. The Evoked Set
The way people perceive data is influenced not only by their cognitive structure and theories about other actors but also by what they are concerned with at the time they receive the information. Information
*9 William Inge, Outspoken Essays, First Series (London 1923), 88.
50 Of course, analogies themselves are not "unmoved movers." The interpretation of past events is not automatic and is informed by general views of international relations and complex judgments. And just as beliefs about the past influence the present, views about the present influence interpretations of history. It is difficult to determine the degree to which the United States' interpretation of the reasons it went to war in 10,17 influenced American foreign policy in the igao's and 1930^ and how much the isolationism of that period influenced the histories of the war.
HYPOTHESES ON MISPERCEPTION
473
is evaluated in light of the small part of the person's memory that is presently active—the "evoked set." My perceptions of the dark streets I pass walking home from the movies will be different if the film I saw had dealt with spies than if it had been a comedy. If I am working on aiding a country's education system and I hear someone talk about the need for economic development in that state, I am apt to think he is concerned with education, whereas if I had been working on, say, trying to achieve political stability in that country, I would have placed his remarks in that framework.61
Thus Hypothesis 5 states that when messages are sent from a different background of concerns and information than is possessed by the receiver, misunderstanding is likely. Person A and person B will read the same message quite differently if A has seen several related messages that B does not know about. This difference will be compounded if, as is frequently the case, A and B each assume that the other has the same background he does. This means that misperception can occur even when deception is neither intended nor expected. Thus Roberta Wohlstetter found not only that different parts of the United States government had different perceptions of data about Japan's intentions and messages partly because they saw the incoming information in very different contexts, but also that officers in the field misunderstood warnings from Washington: "Washington advised General Short [in Pearl Harbor] on November 27 to expect 'hostile action' at any moment, by which it meant 'attack on American possessions from without/ but General Short understood this phrase to mean 'sabotage.' "51 Washington did not realize the extent to which Pearl Harbor considered the danger of sabotage to be primary, and furthermore it incorrectly believed that General Short had received the intercepts of the secret Japanese diplomatic messages available in Washington which indicated that surprise attack was a distinct possibility. Another implication of this hypothesis is that if important information is known to only part of the government of state A and part of the government of state B, international messages may be misunderstood by those parts of
51 For some psychological experiments on this subject see Jerome Eruner and A. Leigh Minturn, "Perceptual Identification and Perceptual Organization" Journal of General Psychology, nil (July 1955), 21-18; Seymour Feshbach and Robert Singer, "The Effects ot Fear Arousal and Suppression of Fear Upon Sodal Perception," Journal cf Abnormal mi Social Psychology, lv (November 1957), 283-88; and Elsa Sippoal, "A Group Study of Some Effects of Preparatory Sets," Psychology Monographs, xtvi, No, 210 (1935), 27-38. For a general discussion of the importance of the perceiver's evoked set, see Postman, 87.
"Pp. 73-74-
474
WORLD POLITICS
the receiver's government that do not match, in the information they have, the part of the sender's government that dispatched the message.83
Two additional hypotheses can be drawn from the problems of those sending messages. Hypothesis 6 states that when people spend a great deal of time drawing up a plan or making a decision, they tend to think that the message about it they wish to convey will be clear to the re-ceiver.Si Since they are aware of what is to them the important pattern in their actions, they often feel that the pattern will be equally obvious to others, and they overlook the degree to which the message is apparent to them only because they know what to look for. Those who have not participated in the endless meetings may not understand what information the sender is trying to convey. George Quester has shown how the German and, to a lesser extent, the British desire to maintain target limits on bombing in the first eighteen months of World War II was undermined partly by the fact that each side knew the limits it was seeking and its own reasons for any apparent "exceptions" (e.g., the German attack on Rotterdam) and incorrectly felt that these limits and reasons were equally clear to the other side.59
Hypothesis 7 holds that actors often do not realize that actions intended to project a given image may not have the desired effect because the actions themselves do not turn out as planned. Thus even without appreciable impact of different cognitive structures and backgrounds, an action may convey an unwanted message. For example, a country's representatives may not follow instructions and so may give others impressions contrary to those the home government wished to convey. The efforts of Washington and Berlin to settle their dispute over Samoa in the late i88q's were complicated by the provocative behavior of their agents on the spot. These agents not only increased the intensity of the local conflict, but led the decision-makers to become more suspicious of the other state because they tended to assume that their agents were obeying instructions and that the actions of the other side represented official policy. In such cases both sides will believe that the other is reading hostility into a policy of theirs which is friendly. Similarly,
53 For example, Roger Hilsman points out, "Those who knew of the peripheral reconnaissance flights that probed Soviet air defenses during the Eisenhower administration and the U-2 flights over the Soviet Union itself . . , were better able to understand some of the things the Soviets were saying and doing than people who did not know of these activities" (To Move a 'Nation [Garden City 1967], 66). But it is also possible that those who knew about the U-a flights at times misinterpreted Soviet messages by incorrectly believing that the sender was influenced by, or at least knew of, these flights.
541 am grateful to Thomas Schelling for discussion on this point. Deterrence Before Hiroshima (New York T966), r05-22,
HYPOTHESES ON MlSPERCEPTION
475
Quester's study shows that the attempt to limit bombing referred to above failed partly because neither side was able to bomb as accurately as it thought it could and thus did not realise the physical effects of its actions.56
VI. Further Hypotheses From the Perspective oe the Perceiver
From the perspective of the perceiver several other hypotheses seem to hold. Hypothesis 8 is that there is an overall tendency for decisionmakers to see other states as more hostile than they are.57 There seem to be more cases of statesmen incorrectly believing others are planning major acts against their interest than of statesmen being lulled by a potential aggressor. There are many reasons for this which are too complex to be treated here (e.g., some parts of the bureaucracy feel it is their responsibility to be suspicious of all other states; decisionmakers often feel they are "playing it safe" to believe and act as though the other state were hostile in questionable cases; and often, when people do not feel they are a threat to others, they find it difficult to believe that others may see them as a threat). It should be noted, however, that decision-makers whose perceptions are described by this hypothesis would not necessarily further their own values by trying to correct for this tendency. The values of possible outcomes as well as their probabilities must be considered, and it may be that the probability of an unnecessary arms-tension cycle arising out of mispercep-tions, multiplied by the costs of such a cycle, may seem less to decisionmakers than the probability of incorrectly believing another state is friendly, multiplied by the costs of this eventuality.
Hypothesis 9 states that actors tend to see the behavior of others as more centralized, disciplined, and coordinated than it is. This hypothesis holds true in related ways. Frequently, too many complex events are squeezed into a perceived pattern. Actors are hesitant to admit or even see that particular incidents cannot be explained by their theories.58 Those events not caused by factors that are important parts of the per-ceiver's image are often seen as though they were. Further, actors see others as more internally united than they in fact are and generally overestimate the degree to which others are following a coherent policy. The degree to which the other side's policies are the product of internal
Ee Ibid.
57 For a slightly di fie rent formulation of this view, see Holsti, 27.
58 The Soviets consciously hold an extreme version of this view and seem to believe that nothing is accidental. See the discussion in Nathan Leites, A Study of Bolshevism (Glencoe 1953), 67-73.
476
WORLD POLITICS
bargaining,53 internal misunderstandings, or subordinates' not following instructions is underestimated. This is the case partly because actors tend to be unfamiliar with the details of another state's policy-making processes. Seeing only the finished product, they find it simpler to try to construct a rational explanation for the policies, even though they know that such an analysis could not explain their own policies.60
Familiarity also accounts for Hypothesis 10: because a state gets most of its information about the other state's policies from the other's foreign office, it tends to take the foreign office's position for the stand of the other government as a whole. In many cases this perception will be an accurate one, but when the other government is divided or when the other foreign office is acting without specific authorization, misper-ception may result. For example, part of the reason why in 1918 Allied governments incorrectly thought "that the Japanese were preparing to take action [in Siberia], if need be, with agreement with the British and French alone, disregarding the absence of American consent,"61 was that Allied ambassadors had talked mostly with Foreign Minister Motono, who was among the minority of the Japanese favoring this policy. Similarly, America's NATO allies may have gained an inaccurate picture of the degree to which the American government was committed to the MLF because they had greatest contact with parts of the government that strongly favored the MLF. And states that tried to get information about Nazi foreign policy from German diplomats were often misled because these officials were generally ignorant of or out of sympathy with Hitler's plans. The Germans and the Japanese sometimes purposely misinformed their own ambassadors in order to deceive their enemies more effectively.
Hypothesis 11 states that actors tend to overestimate the degree to which others are acting in response to what they themselves do when the others behave in accordance with the actor's desires; but when the behavior of the other is undesired, it is usually seen as derived from internal forces. If the effect of another's action is to injure or threaten
59 A. W. Marshall criticizes Western explanations of Soviet military posture for railing to take this into account. See his "Problems of Estimating Military Power," a paper presented at the i§66 Annual Meeting of the American Political Science Association, 16.
60 It has also been noted that in lahot-management disputes both sides may be apt to believe incorrectly that the other is controlled from above, either from the international union office or from the company's central headquarters (Robert Blake, Herbert Shepard, and Jane Mou ton, Managing Inter group Conflict in Industry [Houston 1964], 182). It has been further noted that both Democratic and Republican members of the House tend to see the other party as the one that is more disciplined and united (Charles Clapp, The Congressman [Washington 1063], 17-19).
ai George Kennan, Russia Leaves the War (New York 1967), 484.
HYPOTHESES ON MlSPERCEPTlON
477
the first side, the first side is apt to believe that such was the other's purpose. An example of the first part of the hypothesis is provided by Kennan's account of the activities of official and unofficial American representatives who protested to the new Bolshevik government against several of its actions. When the Soviets changed their position, these representatives felt it was largely because of their influence/2 This sort of interpretation can be explained not only by the fact that it is gratifying to the individual making it, but also, taking the other side of the coin mentioned in Hypothesis 9, by the fact that the actor is most familiar with his own input into the other's decision and has less knowledge of other influences. The second part of Hypothesis 11 is illustrated by the tendency of actors to believe that the hostile behavior of others is to be explained by the other side's motives and not by its reaction to the first side. Thus Chamberlain did not see that Hitler's behavior was related in part to his belief that the British were weak. More common is the failure to see that the other side is reacting out of fear of the first side, which can lead to self-fulfilling prophecies and spirals of misperception and hostility.
This difficulty is often compounded by an implication of Hypothesis 11: when actors have intentions that they do not try to conceal from others, they tend to assume that others accurately perceive these intentions. Only rarely do they believe that others may be reacting to a much less favorable image of themselves than they think they are projecting/3
For state A to understand how state B perceives A's policy is often difficult because such understanding may involve a conflict with A's image of itself. Raymond Sontag argues that Anglo-German relations before World War I deteriorated partly because "the British did not like to think of themselves as selfish, or unwilling to tolerate legitimate' German expansion. The Germans did not like to think of themselves as aggressive, or unwilling to recognize 'legitimate' British vested interest."6*
6 * lbid.t 404, 408, 500.
ea Herbert Butterfield notes that these assumptions can contribute to the spiral of "Hobbesian fear, . , , You yourself may vividly feel the terrible fear that you have of the other party, but you cannot enter into the other man's counter-fear, or even understand why he should be particularly nervous. For you know that you yourself mean him no harm,, and that you want nothing from him save guarantees for your own safety; and it is never passible for you to realize or remember properly that since he cannot see the inside of your mind, he can never have the same assurance of. your intentions that you have" (History and Human Conflict [London 1951]., 20),
84 European Diplomatic History iSyi-igp (New York 1933), 125. It takes great mental effort to realize that actions which seem only the natural consequence of defending your vital interests can look to others as though you are refusing them any
478
WORLD POLITICS
Hypothesis 13 suggests that if it is hard for an actor to believe that the other can see him as a menace^ it is often even harder for him to see that issues important to him are not important to others. While he may know that another actor is on an opposing team, it may be more difficult for him to realize that the other is playing an entirely different game. This is especially true when the game he is playing seems vital to him.85
The final hypothesis, Hypothesis 14, is as follows: actors tend to overlook the fact that evidence consistent with their theories may also be consistent with other views. When choosing between two theories we have to pay attention only to data that cannot be accounted for by one of the theories. But it is common to find people claiming as proof of their theories data that could also support alternative views. This phenomenon is related to the point made earlier that any single bit of information can be interpreted only within a framework of hypotheses and theories. And while it is true that "we may without a vicious circularity accept some datum as a fact because it conforms to the very law for which it counts as another confirming instance, and reject an allegation of fact because it is already excluded by law,"aa we should be careful lest we forget that a piece of information seems in many cases to confirm a certain hypothesis only because we already believe that hypothesis to be correct and that the information can with as much validity support a different hypothesis. For example; one of the reasons why the German attack on Norway took both that country and England by surprise, even though they had detected German ships moving toward Norway, was that they expected not an attack but an attempt by the Germans to break through the British blockade and reach the
chance at increasing their influence, In rebutting the famous Crowe "balance of pawer" memorandum of igojs which justified a policy of "containing1' Germany on the grounds that she was a threat to British national security, Sanderson, a former permanent undersecretary in the Foreign OfRce, wrote, "It has sometimes seemed to me chat to a foreigner reading our press the British Empire must appear in the light of some huge giant sprawling all over the globe, with gouty fingers and toes stretching in every direction, which cannot be approached without eliciting a scream" (quoted in Monger, 315). But few other Englishmen could be convinced that others might see them this way.
65 George Kennan makes clear that in 1918 this kind of difficulty was partly responsible for the inability of either the Allies or the new Bolshevik government to understand the motivations of the other side: 'There is . . . nothing in nature more egocentrical than the embattled democracy, , , . It , , , tends to attach to its own cause an absolute value which distorts its own vision of everything else. ... It will readily be seen that people who have got themselves into this frame of mind have little understanding for the issues of any contest other than the one in which they are involved. The idea of people wasting time and substance on any other issue seems to them preposterous" (Russia and the West, n-12}.
Ae Kaplan, 89.
HYPOTHESES ON MiSPERCEPTION
479
Atlantic. The initial course of the ships was consistent with either plan, but the British and Norwegians took this course to mean that their predictions were being borne out.aT This is not to imply that the interpretation made was foolish, but only that the decision-makers should have been aware that the evidence was also consistent with an invasion and should have had a bit less confidence in their views.
The longer the ships would have to travel the same route whether they were going to one or another of two destinations, the more information would be needed to determine their plans. Taken as a metaphor, this incident applies generally to the treatment of evidence. Thus as long as Hitler made demands for control only of ethnically German areas, his actions could be explained either by the hypothesis that he had unlimited ambitions or by the hypothesis that he wanted to unite all the Germans. But actions against non-Germans (e.g., the takeover of Czechoslovakia in March 1938) could not be accounted for by the latter hypothesis. And it was this action that convinced the appeasers that Hitler had to be stopped. It is interesting to speculate on what the British reaction would have been had Hitler left Czechoslovakia alone for a while and instead made demands on Poland similar to those he eventually made in the summer of 1939. The two paths would then still not have diverged, and further misperception could have occurred.
fl7Johan Jorgen Hoist, "Surprise, Signals, and Reaction; The Attack on Norway," Cooperation and Conflict, No. r (1066), 34. The Germans made a similar mistake in November 1942 when they interpreted the presence of an Allied convoy In the Mediterranean as confirming their belief that Malta would be resupplied. They thus were taken by surprise when landings took place in North Africa (William Langer, Our Vichy Gamble. [New York 1966], 365).