Intelligenceand Foreign Policy Jervis Robert A Review Essay ChristopherAndrew and David Dilks, eds. The Missing Dimension:Governin theTwentieth Century.Urbana: University mentsand Intelligence Communities of Illinois Press, 1984. AssessmentBeforethe Ernest R. May, ed. KnowingOne's Enemies:Intelligence Two WorldWars.Princeton:PrincetonUniversityPress, 1984. General Eisenhower pointed out that the U.S. Army treated intelligenceas a "stepchild."1 This characterizationapplied to most countries,at least until the development of surveillance satellites and other technologicalmarvels. Within the military, intelligencehas been the "slow track";aside froma few mavericks,officers were generallyshunted into the area when theywere deemed unfitformore importanttasks. This is not surprising:the job of the militaryis to fight,and so positions that directlyrepresentthis functionwill have the greatestprestige. Foreign ministries,of course, carryout diplomacy, and so one might thinkthat they would value intelligencemore highly because of its closer links to this mission. But most diplomats have prided themselves on being generalistsand have tended to believe, oftencorrectly,that they can understand other countries better than can specialized intelligenceofficers.Furthermore,intelligenceoperatorsare oftenan unrulybunch, and intelligence operations, if discovered, create frictionsbetween governmentswhich complicate the lives of diplomats, as the recent"Pollard affair"has done between the United States and Israel. Top decision-makersare more likely to value intelligence.In World War II, Churchillreferredto code breakersas his hens because theybroughthim golden eggs. But he, like most leaders, cared more about raw informationthan about the interpretationand analysis. RobertJervisis Professorof PoliticalScienceand memberof theInstituteof War and Peace Studiesat ColumbiaUniversity. 1. Dwight D. Eisenhower, Crusadein Europe(New York: Doubleday, 1948), as quoted in David Kahn, "United States Views of Germany and Japan in 1941," in Ernest R. May, ed., Knowing AssessmentBeforetheTwo WorldWars(Princeton:PrincetonUniversity One's Enemies:Intelligence Press, 1984), p. 501. International Security,Winter1986-87(Vol. 11, No. 3) ? 1986 by the Presidentand Fellows of Harvard College and of the Massachusetts Instituteof Technology. 141 This content downloaded from 128.59.161.126 on Fri, 06 Mar 2015 15:27:41 UTC All use subject to JSTOR Terms and Conditions International Security1142 Thus, until recently,intelligenceoperated on the fringeof governments. Not only were many of the people involved ratherpeculiar characters,but effortsat both collection and analysis were unsophisticated and underfunded. To take the lattercharacteristicfirst,WilliamFuller notes thatbefore World War I, the Russian militaryespionage officehad "a staffof only four to fiveofficers. . . [who] were so ill-paid thattheyall had to take extrajobs in the St. Petersburg district."2(The Russians, however, led the world in code breakingbeforeWorld War I, and Britishsuccesses in this area afterthe war owed much to Russian refugees.) A bit more support was available in European countries in the interwarperiod, but the Britishmajor who was appointed the chief intelligenceinstructoron the German army and POW interrogationat the start of World War II found the staffcollege library "scanty and out of date" and had to go to bookstores in Belgium fortexts. he planned his tripto Holland forMay 1940.3 Unfortunately, Many intelligenceoperations and much intelligenceanalysis was amateurish, especially in Britainand to a lesser extentin Germany-national stereotypes are not without some foundation.4When the new Britishrecruitwas sent to Prague in the 1930s to set up an intelligencenetwork,"he was given no trainingin the fieldcraftrequired forrecruitingand runningagents or in otherformsof intelligencegathering.When [he] asked fortips on howto be a spy, he was told he could spend two or threeweeks en route to Prague at Vienna where . .. one of the most experiencedstationchiefs. .. would give him on the spot training."5Of course amateurismis not automaticallybad, and professionalismdoes not always lead to effectiveness.The Britishcode breaking establishmentin World War II worked so well in part because it was able to assimilate eccentricintellectualswho could not have thrivedin a normal professionalbureaucracy.The otherside of this coin is that several of the intelligenceorganizations in Nazi Germany were ineffectivebecause 2. William Fuller, "The Russian Empire," in May, KnowingOne's Enemies,p. 107. For similar commentsabout France, see ChristopherAndrew, "France and the German Menace," in ibid., p. 135. For England, see Paul M. Kennedy, "Great Britainbefore 1914," in ibid., p. 186; and ChristopherAndrew, Her Majesty'sSecretService:TheMakingoftheBritishIntelligence Community (New York: Viking, 1986). 3. Andrew, Her Majesty'sSecretService,pp. 457-458. 4. Intelligence organizations concentratingon domestic subversion in the autocracies were generallymore efficientand professional;this subject was crucial to the governments,and the underminingbiases mentioned did not apply. 5. Ibid., p. 347. This content downloaded from 128.59.161.126 on Fri, 06 Mar 2015 15:27:41 UTC All use subject to JSTOR Terms and Conditions and ForeignPolicy1143 Intelligence their approaches and values clashed with those of the irrationalpolitical systemwithinwhich they had to operate.6 For all its glamour, intelligencehas also been a stepchild to academics. In a way, thisis odd: much of social science consistsof analyzingothercountries and so parallels the general mission of intelligenceorganizations, and it is no accident that many professorsserved in the Officeof StrategicServices (OSS). But academic neglect there has been, in large part because of the difficulty of gatheringreliable information.The most importantintelligence coup of the twentiethcentury-the breakingof German codes during World War 1I-was kept secret until the publication of F.W. Winterbotham'sThe UltraSecretin 1974. Recently,a great deal has been published about current Westernintelligencemethods and operations,but of course we do not know which published accounts are correctnor which mattersremain secret. Even the pre-WorldWar II record remains spotty,with a great many documents missing or destroyed and many archives still sealed. Indeed, Christopher Andrew notes thatwhile all the interceptedGerman messages duringWorld War II are open to public inspection,the Britishinterceptsof Soviet messages in the 1920s remain secret.7 In the United States, attentionhas concentratedon the legitimacyof covert operations and on intelligencefailures,especially those culminatingin surprise attack such as occurredat Pearl Harbor.8The books under review here show that these are not the only interestingquestions, that sufficientinformationis available to allow analysis ofhow intelligencehas functionedduring many periods and in many countries,and that this analysis is both intellectuallylegitimateand sheds a great deal of lighton internationalpolitics. Both the subject and the books under review cover so much ground that summaries are difficult.Indeed, strategicintelligencetouches on so many facetsof policy that it is difficultto generalize withoutconstructinga theory of foreignpolicy, if not of internationalpolitics. So I will discuss only a few of the questions that these books raise. Afterbrieflynotingthe nature of the 6. Michael Geyer, "National Socialist Germany:The Politics of Information,"in May, Knowing One's Enemies,pp. 321-322. 7. Andrew, Her Majesty'sSecretService,p. 506. 8. The classic study of Pearl Harbor is Roberta Wohlstetter,Pearl Harbor:Warningand Decision (Stanford:StanfordUniversityPress, 1962). But see also Kahn's remarksin "United States Views of Germanyand Japan," p. 500. An interestingrecentstudy by an intelligenceofficerwho was active at the time is Edwin Layton, "And I Was There"(New York: Morrow, 1985). On surprise attacksin general, see Richard Betts, SurpriseAttack(Washington,D.C.: Brookings,1982). This content downloaded from 128.59.161.126 on Fri, 06 Mar 2015 15:27:41 UTC All use subject to JSTOR Terms and Conditions International Security 1144 two collections,I will turn to Ernest May's argumentsabout the criteriafor evaluating intelligence performanceand examine how the powers judged each other's capabilities and intentionsbeforethe world wars. This leads to the broader topic of the relationshipbetween intelligenceand policy, and more particularlythe limits on the impact of the formeron the latter. In many cases, intelligencecannot, or at least does not, guide action because commitmentsto a policy are driven by powerful forces that are engaged before analysis of other countries can make itselffelt.Furthermore,to understand each other's behavior, decision-makersusually have to understand how theirown state is actingand how otherssee them. Although thiswould seem easy, in factit is not. States have powerful and idealized, if not selfserving,self-images;they follow double standards and rarelyappreciate the extentto which they menace others' interests. For an edited volume, KnowingOne's Enemiesis remarkablycoherent. In addition to excellent introductoryand concluding essays by Ernest May, it contains studies of how each of the main adversaries saw each other in the years preceding the two world wars. Without being squeezed into a procrusteanbed, each author discusses how his countrycarriedout what today is called "net assessment." How did each countryweigh the capabilitiesand intentionsof others? How accurate were theirviews? What distortionsand biases entered in? What was the role of bureaucraticinterestsand competition?How, ifat all, was intelligenceused by decision-makers?All the essays are of remarkablyhigh quality.Although some readers will findmore details about individuals and organizationalstructuresthan theyreallywant, anyone interestedin securitypolicy will learn a great deal. While May's focus lends coherence to the volume, it has one drawback: the sample of cases is biased. By looking at assessments before the world wars, the authors only examine cases in which the states did turnout to be adversaries. While this does not affectgeneralizationsabout perceptions of capabilities,it does limit the extentto which we can extend what we have learned about perceptionsof intentions.A naive lesson fromthe book would be that states should always take a pessimisticview of others' policies. This is clearlythe case forthe 1930s and is arguably true in the pre-WorldWar I period as well. While the authors do note some false alarms-the "May crisis"in 1938 and the false rumorsof impending German attackson Poland and Rumania in the springof 1939-a fullybalanced account requiresstudies of cases in which states made assessments of countries that the historical recordlaterrevealed were not in facttheirenemies. Questions of how, when, This content downloaded from 128.59.161.126 on Fri, 06 Mar 2015 15:27:41 UTC All use subject to JSTOR Terms and Conditions Intelligence and ForeignPolicy1145 and why states perceive threatscan only be answered by analyzing the full range of cases, including those in which wars resulted because one or both sides overestimated the other's hostilityand those less dramatic ones in which peaceful outcomes were permittedby the recognitionthat neither menaced the other's vital interests. The Missing Dimensionby Andrew and Dilks is more like a typicaledited volume. Although the essays are excellent,they are united only in dealing with intelligence in the twentiethcentury. Subjects as diverse as British intelligencein Ireland from1914 to 1921, the CIA's search forlegitimacy,and an intriguingif impressionisticaccount of "the Cambridge Comintern" (the Britishspies forthe Soviet Union) are covered. There is no concluding chapter, and, given this collection of papers, no such chapter could have been written. Several of the essays concentrateon code breaking, however, a subject that gets lost in the broader essays in May's book. One issue that has received a great deal of attentionrecently,especially frompoliticalscientists,is only touched on in the May essays, even though it bridges the subjects of policy and intelligence. This is the issue of the existence,causes, and consequences of defensive or offensivebiases in militaryperceptions and planning. Offensivebiases have been seen as an immediate cause of World War I, creating what we would now call "crisis but still adversely instability."9Western policy in the 1930s was differently affectedboth by the overestimateof the immediate havoc that bombing of citieswould bring10and the underestimateof the power ofa well-coordinated offensive. Criteria forJudgingIntelligence One obvious question to be asked of past intelligenceis how accurate it was. Of course one must be careful, as May is, not to oversimplify.Keeping a 9. The argumentis an old one. The recent literatureincludes: JackSnyder, The Ideologyof the Offensive: MilitaryDecision-Making and theDisastersof1914 (Ithaca: Cornell UniversityPress, 1984), and the articleson "The Great War and the Nuclear Age," in International Vol. 9, No. Secuirity, 1 (Summer 1984). For a good critique, see Scott Sagan, "1914 Revisited: Allies, Offense, and Instability,"International Security,Vol. 11, No. 2 (Fall 1986). Beliefsabout the relativeefficacyof the offenseand defense in the interwarperiod are discussed in Barry Posen, The Sourcesof MilitaryDoctrine:France,Britain,and GermanyBetweentheWorldWars(Ithaca: Cornell University Press, 1984); also see Williamson Murray,TheChangein theEuropeanBalanceofPower,1938-1939 (Princeton:PrincetonUniversityPress, 1984). 10. See Uri Bialer, The Shadowof theBomber:The Fear ofAir Attackand BritishPolities,1932-1939 (London: Royal HistoricalSociety,1980). This content downloaded from 128.59.161.126 on Fri, 06 Mar 2015 15:27:41 UTC All use subject to JSTOR Terms and Conditions International Security1146 scorecard, while fun, may not be analyticallyuseful. If one wants to judge whether intelligencedid well or badly, one must consider the information that was available at the time, the difficultyof the task, and whether the outcome was stronglydeterminedor could have easily turnedout differently. For example, informationthat could in principlehave been available but in fact was not helps explain the Western belief that the Shah would not be overthrown.One reason why the Shah did not use all the forceat his disposal in 1978 was that he knew he was dying of cancer and realized that even if repression put down the revolt, it would defeat his goal of turningover power to his son. If forcewere successful,the armywould probablyinherit his power, and his son was not capable of runningas authoritariana governmentas he was. But the Shah's illness was a deep secret,unknown not only to the United States, but also to the French,whose doctorswere treating him. Sometimes the informationis unavailable even in principle,therebyrendering the task of intelligenceparticularlydifficult.Thus Donald Watt notes that, in many cases, "the time gap between Hitler's decisions and the subsequent action was very short.""1Of course, on occasion, intelligencecan predictwhat another countrywill do before that country'sdecision-makers have reached a judgment-just as individuals can sometimes predict what theirfriendswill do beforethey themselvesknow. But we cannot ordinarily expect such a profound understandingof the otherside or the situationit is in. May notes that a thirdreason why success is not the only criterionis that some crucial decisions could have easily gone the other way: intelligence often"made close calls which turned out not to be right"(p. 508). Thus it is not surprisingthat it was hard for Germany to predictwhen Russia would mobilizein 1914: "afterall, the Russian general staffremainedon tenterhooks for two days as the Tsar alternatelyissued orders for mobilization and rescinded them" (p. 507). Finally, intelligence may be right for the wrong reasons or wrong forthe rightreasons.12 11. Donald Cameron Watt, "BritishIntelligenceand the Coming of the Second World War in Europe," in May, KnowingOne's Enemies,p. 246. 12. See, forexample, the instances cited in ibid., pp. 251, 498. For a furtherdiscussion of the inherent limitationson intelligence, see Richard Betts, "Analysis, War, and Decision: Why IntelligenceFailures are Inevitable," WorldPolitics,Vol. 31 (October 1978), pp. 61-89; and Robert Jervis,"What's Wrong with the Intelligence Process?," International Journalof Intelligence and Counterintelligence, Vol. 1 (Spring 1986), pp. 28-30. This content downloaded from 128.59.161.126 on Fri, 06 Mar 2015 15:27:41 UTC All use subject to JSTOR Terms and Conditions Intelligence and ForeignPolicy1147 More fundamentally,May notes, "accuracy is not enough. BritishForeign Officeapocrypha tells of a retireesaying thatforfifty years, year in and year out, he had assured Foreign Secretariesthat therewould be no major European war. In all that time, he boasted, he had been wrong only twice.... Nor is accuracy about importantquestions sufficient,for intelligenceestimates are useful only if acceptable to the people who have to act on them" (pp. 503-504). But of course acceptabilitymay mean merelycateringto the preconceived notions of the policymakers,and so May argues, "a bettertest than either accuracy or acceptabilitymay be simply whether assessments address the right questions: that is, the questions rightanswers to which could be useful guides to action" (p. 504). JUDGING CAPABILITIES BEFORE THE WORLD WARS May's conclusion is interestingbut not entirelysatisfactory:"By the test of whether the rightquestions were pursued, no governmentdid well either before1914 or in the 1930s. . . . Broadlyspeaking, governmentsof the earlier period were at theirworst when estimatingcapabilities. Governmentsof the 1930s, on the other hand, made theirgreatesterrorswhen judging proclivities" (p. 504). (May prefers the term "proclivities" to the more common "intentions" because the latter "suggests a single brain: does the United States ever really have intentions?"[p. 503]. The point is a good one, but I do not think the term "intentions" needs to carry this connotation, and because the term is a common one, I will continue to use it here.)13"In gauging capabilities, pre-1914intelligencebureaus got littlethings rightbut big thingswrong" (p. 504). As the essays in the book indicate,each side was quite accurate in judging the size of the other's forces,theirmilitaryskills, and with a few major exceptions, the characteristicsof the weapons that would be used: "The facts were right,but the assumptions to which they were fittedwere fantastic,supposing thatsoldierscould marchthroughfields of fire and that politicians answerable to capricious publics would coolly negotiatetermsas soon as theircash ran out" (p. 507). Thus the capabilities that were misjudged were not the narrow militaryones of what one side could do to the other. 13. I have discussed some of the conceptual issues involved in Perception and Misperception in International Politics(Princeton:PrincetonUniversityPress, 1976), pp. 48-54. This content downloaded from 128.59.161.126 on Fri, 06 Mar 2015 15:27:41 UTC All use subject to JSTOR Terms and Conditions International Security1148 It seems to me thattherewere threecrucialerrors,all linkedto the common illusion that the war would be short.14First,as both a cause and a consequence of thisbelief,all the continentalpowers had a greatfaithin the power of the offensive.If the offensivewould carrythe day, then the war would be short; if the war were to be short, only an offensivestrategycould be effective.Second, the French misjudged where the Germans would attack, disregarding a large amount of evidence that they would move through Belgium northof the Meuse. Third,it was generallybelieved thatstates and societies could not mobilize the resources and stamina to fighta protracted war. What was misjudged, then, were characteristicsthat militaryintelligence was ill-suitedto estimate:the capabilitiesof the individual soldier on the one hand and societies as a whole on the other. Intelligenceofficersand commanders alike vastly overestimatedthe abilityof the fightingmen to attack in the face of witheringfirepower.Partly,of course, this was a misjudgment of the latter-the machine gun was particularlyunderestimated.But at least as importantwas the faith in the braveryand discipline of the men who were expected to keep chargingeven though those around them were dropping. Morale was centralbecause the other side was expected to retreatnot so much because of the physical damage it suffered,but because its fighting spiritwas to be broken. But this,ofcourse, requiredthatthe attacker'smorale remainintact.How thiswas to be done (and how intelligencewas to estimate whetherit would be possible) was never systematicallyanalyzed and, given the nature of the question, perhaps could not be. The second misjudgment of capabilities was the French failureto judge the Germans' militaryplans correctly.This has been discussed at length elsewhere,15and here I just want to note that it also rested on a failureto judge individual capabilities. The SchlieffenPlan was possible only because the Germans used reservistsin the frontlines. But this the French believed they could not do. Although much of their resistance to this idea was its implicationsforthe standing of the militarywithin the French political and social system,in part they failed to appreciate the abilityof nonprofessional soldiers. The most general misjudgmentconcerned not individuals, but the society and the state. Almost unanimously, people underestimatedthe willingness 14. The literaturehere is very large. See, for example, L.L. Farrar,Jr.,The Short-War Illusion (Santa Barbara, Calif.: ABC-Cleo, 1973). 15. See the literaturecited in footnote9 above. This content downloaded from 128.59.161.126 on Fri, 06 Mar 2015 15:27:41 UTC All use subject to JSTOR Terms and Conditions Intelligence and ForeignPolicy| 149 of the formerto endure hardships and of the latter to mobilize national resources. It was believed that modern countries simply could not fighta long war because theywould run out of resources. In part,this was a failure of imagination. Previous European wars had been short; the offensivehad quickly broken the opposing army; it was hard to see how societies could withstand the strain and disruption of prolonged mobilization. The war indeed would have had to have been shortif peacetime practiceshad been followed.16To see that the state apparatus could direct civil society to an extentpreviouslyunknown would have required rare intellectualand moral boldness. Furthermore,two complicatingfactorsinhibitedsuch ideas. First, no part of the governmentwas tryingto imagine how the role of the state mightbe enlarged to permita long war to be fought.Even in a society like Germany in which the militaryhad high standing and prestige,it probably would have been seen as impertinentfor the militaryto have undertaken such analysis. Second, and more basic, the beliefthatgovernmentscould do what in factthey did do during the war would have been directlycontrary to the prevailingliberalideology, especially strongin England but presentin othercountriesas well, which sharplyrestrictedthe role of the government in civil society. Thoughts that the governments could, let alone should, manage the economy would have been quite literallysubversivein thatthey would have attacked the normativeand empiricalbeliefs that underpinned the prevailingdomestic systems.To have concluded thatthe statecould fight a long and totalwar by expanding its power would have implied thatlaissezfairewas not the only possible systemand that the state could play a dominant role during peacetime as well. The implicationsforthe possibilitiesof social change and reformwould have been considerable. JUDGING INTENTIONS BEFORE THE WORLD WARS Analyses of others' intentionsbeforeWorld War I were more accurate, May argues: "Analysts and statesmen . . . identifiedand addressed most of the key questions" (p. 507), and "[flor the most part, [decision-makers]understood the likelytermsof debate in othercapitals" (p. 508) and grasped what others'interestswere. But neitherMay's conclusion nor the individual essays discuss thisissue in sufficientdetail to support thisconclusion. The question, 16. See David French,BritishEconomicand Strategic Planning,1905-1915(London: Allen & Unwin, 1982); and Paul M. Kennedy, "Strategyversus Finance in Twentieth-CenturyBritain,"in Kennedy, Strategy and Diplomacy,1870-1945 (London: Allen & Unwin, 1983). This content downloaded from 128.59.161.126 on Fri, 06 Mar 2015 15:27:41 UTC All use subject to JSTOR Terms and Conditions International Security| 150 ofcourse, is a large one, coveringmany of the centralissues about the origins of the war, and to have concentratedon it would have blurredthe focus of the volume. But judgments of the other side's intentionsare oftena crucial element in a state's foreignpolicy, and many debates hinge on diverging analyses of what the other is up to. I thinkit is fairto say that in the years preceding World War I, leaders only rarelypaused to carefullyanalyze the questions of whether the differenceswith theiradversaries were unbridgeable and whether the other could be best influencedby threatsor inducements. Documents like the Crowe memorandum and the replies to it were exceptional,ifnot unique. If World War I were avoidable-and of course this question is as controversialas it is fascinating-then there were crucial failures of intelligenceas well as policy. In any event, it is hard to argue that the analysis of other states' intentions(or of the state's own interests)was of high enough quality to justifythe momentous decisions that were made. Thus May's conclusion about perceptions of capabilities may apply to perceptions of intentionsas well: "pre-1914intelligencebureaus got littlethings rightbut big thingswrong" (p. 504). While theirshort-runpredictionswere generallyadequate, theirgrasp of other states' goals, motives, and perspectives were deficient.As a result,they may have missed chances foramelioratingconflictand avoiding the war that engulfed them. Interestinglyenough, in the 1930s, Britishand French officialspaid more attentionto these issues, but theirconclusions were even less accurate. Why this was so has been endlessly debated, and May's three arguments are worth repeating. First,one should not exaggerate the sharpness of the distinctionbetween "appeasers" and "anti-appeasers" and attributenothingbut blindness to the formerand nothing but perspicacityto the latter. Many people held changing or inconsistentviews, and the records "show almost everyone . .. making some statementsthatin retrospectseem stupid, others that seem wise" (p. 522). Furthermore,ifthe "appeasers" were wrong about Hitler's goals, the "anti-appeasers" had their own illusions, which were almost equally distantfromreality:theybelieved thatHitlercould be deterred by the threat of war (p. 520). Third, "concerning processes of decision, 'appeasers' and 'anti-appeasers' alike tended to imagine debates in Berlin comparable to those in London or at least Paris, with an adversary system pittingwhat Halifax's private secretarycalled the 'forward school' against 'moderates"' (p. 521). I would add threepoints of my own. First,analyses of Hitlerwere deeply influenced by the "lessons of the past." Most statesmen in London, and This content downloaded from 128.59.161.126 on Fri, 06 Mar 2015 15:27:41 UTC All use subject to JSTOR Terms and Conditions and ForeignPolicy| 151 Intelligence some in Paris, believed that World War I could have been avoided by conciliatorydiplomacy. This predisposed them to avoid the previous errorand to perceive Hitler as appeasable. Second, this beliefwas reasonable, indeed more reasonable than the view of Hitlerthatwe now see to be correct.Given the enormous destructionthat a war would entail, only a most unusual, if not insane, ruler would wage a major war to seek dominance. While all German leaders wanted to revise the Treatyof Versailles,none of the earlier leaders and few of Hitler's Nazi confidantswould have been willingto fight Britainand France. Common sense indicated that while Hitler was an evil tyrant,he would be content with regaining Germany's pre-1914 position. Indeed until he took over the non-German portions of Czechoslovakia in March 1939, all his behavior could be explained by his drive for this goal. Third,neitherthe "appeasers" nor the "anti-appeasers" acknowledged what was in facta sharp trade-offbetween securing Britishand French interests and avoiding war. As May notes, the "anti-appeasers" believed thata policy of firmnessbacked by strengthwould preserve the peace; the "appeasers" believed thatconcessions would reduce the dangers of war withoutallowing Hitlerto dominate. To see thatthe actual choices were more distastefulthan this would have been extremelypainful,and so it is not surprisingthat the perceptionswere inaccurate. Intelligence and Policy On these broad issues, intelligencecommunitiesare not likelyto have a great deal to say or to have much influence.Why should intelligenceanalysts be betterthan journalists or decision-makersthemselves-not to mention academics-at determiningwhetherthe other can be conciliated,the degree to which it is driven by unusual ambitions,or the extentto which it is reacting to perceived threats?The painstakingcollectionof detailed informationsimply will not yield answers to these questions. Perhaps a spy withinthe other side's circle of top decision-makers might provide real insight, but such instancesare largelyconfinedto fiction.17Furthermore,questions ofthe other 17. Two modern instances come to mind. First,the U.S. may have received informationfrom a memberof Indira Gandhi's cabinet during the 1971 war. But Kissinger seems to have used these reportsto bolster his preconceived and incorrectestimates of India's intentionsand so U.S. policy was not improved. Second, Donald Maclean was privyto many aspects of Western policytoward the Soviet Union when he served at the Britishembassy in Washingtonfrom1944 to 1948. Of course while we know what was said at many of the top secretmeetingshe attended, This content downloaded from 128.59.161.126 on Fri, 06 Mar 2015 15:27:41 UTC All use subject to JSTOR Terms and Conditions International Security1152 side's intentionsinvolve general politicaljudgments which decision-makers are likely to want to make for themselves.18 We do not know a great deal about what determinesthese judgments. Personalitymay be important:for example, some people may be prone to see adversaries where others see only resolvable differences.General political beliefs, if not ideologies, are likelyto play a role: general attitudestowardleft-winggovernmentsinfluence the degree to which Americans see Nicaragua as a threat.Not only are these factorsbeyond the reach of intelligence,but in most cases statesmen will have developed views about other countriesbefore taking office.Thus it is not surprisingthat when people enter the governmentand gain access to classified intelligence reports, they rarely change their basic beliefs about other countries. People may change, but only in response to traumaticexternal events, not because of new or betteranalysis. For example, can we imagine any study of the Soviet Union and its foreignpolicy that would convince Ronald Reagan that his ideas about thatcountryare incorrect? The basic outlines of threat assessment then are rarely the province of intelligence,even though one could argue that this should be its most important function.David Kahn's conclusion about the United States before World War II may be extreme but could be applied with modificationsto other countries as well: "Intelligencehad littleto do with American assessments of Germanyand JapanbeforeDecember 1941. The actions of the Nazi government convinced Americans, high and low, that the United States might someday be the target,if not the victim, of its aggression. On the other hand, American racism and rationalismkept the United States from thinkingthatJapan would attackit."19 In cases that did not appear to be so obvious, decision-makersusually employed large doses of intuition,withoutrespectforthe generallyuninteresting debate over whether intelligenceis an art, a craft,or a science.20In we do not know what he reported,let alone what conclusions were drawn in Moscow. He and the Russians may have dismissed as rationalizationsthe fears of Soviet expansion, which he constantlyheard, and Stalin might have been emboldened by the obvious inhibitionsin the West against using its vast superiorityin militaryand economic power. For a briefdiscussion, see Robert Cecil, "The Cambridge Comintern,"in Andrew and Dilks, The Missing Dimension: Governments and Intelligence Communities in theTwentieth Century(Urbana: Universityof Illinois Press, 1984), pp. 185-187, 196. 18. See, for example, Michael Barnhart,"Japanese Intelligencebefore the Second World War: Best Case Analysis," in May, KnowingOne's Enemies,p. 424; and May's discussion of the different judgmentsreached by Hitler and his generals, ibid., pp. 514-519. 19. Kahn, "United States Views of Germany and Japan," p. 476. 20. However, a few relevant comments on this debate can be found in May, KnowingOne's Enemies,pp. 3, 158, 191-192, 296, 424, 518. This content downloaded from 128.59.161.126 on Fri, 06 Mar 2015 15:27:41 UTC All use subject to JSTOR Terms and Conditions and ForeignPolicy| 153 Intelligence his discussion of French militaryintelligencein the 1930s, Robert Young points to "the tension between knowledge and understanding,"21a tension thatexistedin othercountriesand otherperiods as well. Accurateconclusions and disastrously wrong ones as well often ran far ahead of the available information.Thus as earlyas February1933,RobertVansittart,the permanent undersecretaryof the Foreign Office who was to become a leading "antiappeaser," said thatthe Germans were "likelyto relyfortheirmilitarypower ... on the mechanical weapons of the future,such as tanks, big guns and aboveall militaryaircraft."Eighteenmonthslater,when criticizingthe military forbeing slow to appreciate the rise of German power, he said: "Prophecy is largely a matterof insight. I do not thinkthe Service Departments have enough. On the otherhand theymightsay thatI have too much. The answer is that I know the Germans better."22But arguments about who knows anothercountrybetteror who has the giftofprophecyare not easy to resolve. COMMITMENTS AND POLICY PREFERENCES The impact of intelligenceon the broad outlines of the state's foreignpolicy is furtherlimitedby two otherfactors.First,the statemay have littlefreedom of maneuver. While knowledge can multiplythe effectivenessof national power, it cannot compensate for all weaknesses. As Norman Stone notes, "the heart of the matter[forAustria-Hungarybefore 1914] was simply that [the Empire] was tryingto act the part of a great power with the resources of a second-rank one."23 Similarly,RobertYoung argues that because of its dependence on Britainin the 1930s, even the best intelligenceabout Germany could not be harnessed by France to an effectivepolicy.24 More interestingly,states sacrificeor limit their freedom of action in a psychologicalsense when they become committedto a policy, which often 21. RobertJ. Young, "French MilitaryIntelligenceand Nazi Germany,1938-1939," in ibid., p. 306. 22. Quoted by Watt, "BritishIntelligence,"p. 268. 23. Norman Stone, "Austria-Hungary,"in May, KnowingOne's Enemies,p. 52. 24. Young, "French MilitaryIntelligence," pp. 287, 310, 307-308. A parallel conclusion was reachedby the Commander-in-Chiefof the Britisharmyin the Far East afterthe fallofSingapore: "There's no doubt that we underestimatedthe Jap. ... But suppose we'd made a bettershot and had got the Jap at his true worth,would it have made any difference?I very much doubt it. Our policy was to avoid a war with Japan as long as we could (or to make America cause it, ifitwas to happen) and we gambled on thatpolicy succeeding (or ifit didn't succeed on America bearing the brunt). With all our other commitmentsI don't believe that, however highly we had rated the Japs as fighters,we would have been caused therebyto improve the conditionof Sir our Servicesin the Far East." Brian Bond, ed., ChiefofStaff:The DiariesofLieuitenant-General HenryPownall,1933-1944 (London: Cooper, 1974), Vol. 2, p. 92. This content downloaded from 128.59.161.126 on Fri, 06 Mar 2015 15:27:41 UTC All use subject to JSTOR Terms and Conditions International Security1154 happens without the benefitof adequate intelligenceassessments. As a result, policy often drives intelligenceas much as intelligencedrives policy. This can take place both with specificpolicies and with the broad outlines of the state's goals and approach, which are especially likelyto be quite powerfuland to predate analysis of what others will do. Holger Herwig's appraisal is not unique to WilhelmianGermany: In the finalanalysis, Germany'sleaders based theirgrand strategiesnot upon available intelligenceconcerningpotentialadversariesbut ratherupon certain inflexibleideological convictions and militaryprinciples. One needs above all to understand and to appreciate that strangecomposite of historicaldeterminism,racism,and autism which constitutedthe mentalite of Wilhelmian Germany. It cannot be stressed enough that German planners dealt not so much in day-to-day evaluations as in grand designs, whose sweep often encompassed the lives of a generationor two. The strengthsand weaknesses of potentialenemies were taken into account only afterinflexibleblueprints had been drafted.25 As Herwig concludes, "it is likelythat only a total realignmentof the social fabricof Wilhelmian Germany could have permittedmore realisticassessments or action in 1914. '26 Leaders who have become committedto a policy of expansionism will rarelyperceive the obstacles in theirway accurately.Michael Barnhartmakes it clear that this was true forJapan: that country's "reluctance to appraise rationallyAmerica's vast economic predominance came from a feeling of indifference.It was futile,and perhaps treasonous, to suggest that any war with the United States was unwinnable.... Since the only alternative[to fighting],by 1941, was instant surrender,Japan's policy-makerselected to ignorecontraryindicationsand believed thata limitedand thereforewinnable conflictwas possible."27As a section chiefof the General Staff'sintelligence division put it when asked if he expected victoryin a war with the United States: "A Japan-America war is necessary: it is no longer a question of victoryor defeat."28Once a course of action is seen as necessary, great psychological pressures are generated to believe that it is likelyto succeed; intelligenceofficersas well as decision-makersusually feel these pressures 25. 26. 27. 28. Holger H. Herwig, "Imperial Germany,"in May, KnowingOne's Enemies,p. 95. Ibid., p. 97. Barnhart,"Japanese Intelligence,"pp. 454-455. Ibid., p. 452. This content downloaded from 128.59.161.126 on Fri, 06 Mar 2015 15:27:41 UTC All use subject to JSTOR Terms and Conditions Intelligence and ForeignPolicy| 155 and, if they do not, are likelyto be ignored.29Thus it is not surprisingthat WesleyWark's thoroughstudyofBritishintelligenceon Nazi Germanyshows that, by and large, militaryestimates could be predicted by Britishpolicy ratherthan vice versa. Neither the swing to appeasement in the mid-1930s nor the swing away fromit in the earlyspringof 1939 was based on changing analyses of German strength.Rather,estimatesbecame more pessimisticas the appeasement policy gatheredforceand shiftedquite dramaticallytoward optimism six months before the startof the war when it became clear that Britainwould have to confront,and probably fight,Germany.30The British overestimateof German airpower and the misreading of how its air force would be used can similarlybe explained in large part by the policy preferences of the decision-makers.31 The commitmentto a policy may also operate at lower levels both of generalityand of the bureaucracy. Various groups and sectors within the governmentneed to hold a certainpictureof the other side's likelystrategy and tacticsin order to justifythe course of action or forcestructurethatthey prefer.For example, one reason why the French rejected informationindicating that the Germans would use reservistson the frontline in 1914 was that the French army was committedto keeping its own reserves in a secondary role in order to maintain its corporate professionalism,which was under attackfromdomestic politicalopponents.32Thus in many instances, a state's militarypolicy or doctrineinfluencesits estimates of how the other side will fightmore than being determinedby such estimates.Paul Kennedy notes that"since [pre-1914]Britishdoctrinerequireda decisive clash of battle fleets,German plans had to be construed as leading to such an outcome," even though there was littleevidence pointingin that direction.33Similarly, 29. See RobertNorth, "Perceptionand Action in the 1914 Crisis," JournalofInternational Affairs, Vol. 21 (1967). For a furtherdiscussion of this phenomenon, see IrvingJanisand Leon Mann, DecisionMaking (New York: Free Press, 1977); Richard Cottam, ForeignPolicyMotivation(Pittsburgh: Universityof PittsburghPress, 1977); Paul Schroeder, Austria, Great Britain,and the CrimeanWar(Ithaca: Cornell UniversityPress, 1972); Snyder,TheIdeologyoftheDefensive; Richard Ned Lebow, BetweenPeaceand War(Baltimore:JohnsHopkins UniversityPress, 1981); and Robert and Deterrence Jervis,Richard Ned Lebow, and JaniceGross Stein, Psychology (Baltimore:Johns Hopkins UniversityPress, 1986). 30. Wesley Wark, The UltimateEnemy:BritishIntelligence and Nazi Germany,1933-1939 (Ithaca: Cornell UniversityPress, 1985). Also see Wark, "In Search of a Suitable Japan: BritishNaval Intelligencein the PacificBefore the Second World War," Intelligence and NationalSecurity,Vol. 1 (May 1986), pp. 189-211. 31. For a furtherdiscussion of the self-deterrence involved, see Jervis,"Deterrenceand Perception,"International Security,Vol. 7, No. 3 (Winter1982-83), pp. 14-19. 32. Snyder,The IdeologyoftheOffensive. 33. Kennedy, "Great Britainbefore 1914," p. 185. This content downloaded from 128.59.161.126 on Fri, 06 Mar 2015 15:27:41 UTC All use subject to JSTOR Terms and Conditions International Security1156 the size of the army, which the British believed would be necessary to stronglyinfluencethe outcome of a continentalwar, was determinedmore by the size of the standing army which could be created without political disruptionat home than it was by an analysis of the militarybalance between the two sides.34The same effectoccurs with narrowerquestions, such as the efficacyof various weapons and tactics.Thus Young points out that,in early 1939, French militaryintelligence: suggested that the Spanish experience had made many German officers wonder if a tank was really worth its constructionand maintenance cost. It is interestingto note that this particularexample of faultyguesswork came only a few months after Gamelin's trainingdirectivefor 1939, a directive which forbadexposing Frenchtroops-though not theirofficers-toGermanstyle mobile maneuvers, for fear of confusingthem as to what was appropriate doctrineand what was not.35 Net assessment is oftendrivenby bureaucraticand domesticneeds at least as much as by disinterestedanalysis of the other side. What Christopher Andrew says about the French debate in 1913 over whetherto lengthen the termof militaryserviceby one year in order to match similarGerman moves holds for many other issues as well: as the debate proceeded, "assessment of the German army played only a secondary role and sometimes almost disappeared fromview."36The kinds of argumentsthat are made, the evidence that is deemed important,and the methods that are applied are often determinedby the answers thatare desired. Today, forexample, the logic of net assessment cannot explain why doves look primarilyat the numbers of warheads, and hawks look at throwweightor megatonnage. Indeed, if the strategictheories held by these two groups were drivingtheirassessments, then those who were concerned about war-fighting would look to warheads, and those who emphasized the importance of assured destructionwould stress gross explosive power. That this is not the case is primarilyexplained by the fact that the latter measure shows the U.S. to be way behind the Soviet Union while the formerreveals parity or a slight U.S. lead. Net 34. Ibid., pp. 193-195. At least some Britishgenerals were aware that they were being disingenuous: General Henry Wilson, the Directorof MilitaryOperations, told a colleague that the six divisions thatwere claimed to be decisive were probably"fiftytoo few." Quoted in Nicholas in PeacetimeBritain,1902-1914 d'Ombrain, WarMachineryand High Policy:DefenseAdministration (London: Oxford UniversityPress, 1973), p. 104. For another example, see Young, "French MilitaryIntelligence,"p. 297. 35. Ibid., p. 303. For another example, see p. 142. 36. Andrew, Her Majesty'sSecretService,p. 138. This content downloaded from 128.59.161.126 on Fri, 06 Mar 2015 15:27:41 UTC All use subject to JSTOR Terms and Conditions and ForeignPolicy1157 Intelligence assessment is highly political; we should not be surprised that the analysis is often explained by rather than explains the conclusions. All too often, then, intelligenceestimates tell us more about interestsand foreignpolicy preferencesof powerful groups in governmentthan it does about what the other side's intentionsand capabilitiesare. As May demonstratesin his discussion ofnet assessment in Britain,Russia, and GermanybeforeWorld War I, when key decision-makerscome together formeetingsat which the possibilityof war is discussed, each oftenhas good political and bureaucratic reasons to avoid a full and frank treatmentof militarystrategyand militarybalance. Bureaucraciesand the politicianswho lead them do not want othersmeddlingin theirdepartments'affairs.Sharing of informationand analysis oftenruns the riskof sacrificingpower. Exposing assumptions to criticalscrutinycan lead to divisive policy debates. The costs of honest and disinterestedevaluation are immediate and involve pressing if often narrow interests;the gains are postponed if not hypothetical.Furthermore,these gains will accrue only if the otherparticipantsare able and willingto put aside theirparochial blinders.In the familiardilemma of public goods, any person or departmenttaking the broader perspective runs the risk of sacrificinghis or her individual interestsand preferenceswithout the national interest. furthering BEST-CASE ANALYSIS AND WORST-CASE ANALYSIS The policies, interests,and needs of the actors can help explain when actors will make "worst-case" estimatesand when theywill expect the "best case." Both errorsare common; the argumentthat statesmenalways, or even usually, expect the best or expect the worst cannot be sustained.37Many of the errorsare no doubt random. Others are the products of cognitivepredispositions and general beliefs about how states are likelyto behave. But often the person sees what he or she needs to see in orderto support the personal or organizationalinterestsinvolved or the policy preferencesthat are held. Thus Donald Watt argues: Assessments . . . tended to reflecteitherthe expectation that the enemy would do what the branch of the armed forcesmaking the assessment most 37. For examples of worst-caseanalysis, see May, KnowingOne's Enemies,pp. 110, 114, 122, 125, 261, 301, 350-351; forthe opposite error,see pp. 125, 146, 424, 433, 437-439, 445, 449, 455. For a debate about whetherAmerican estimatesof Soviet militarystrengthhave more usually erred on the side of optimismor pessimism, see the articlesby Les Aspin and WilliamLee in Strategic Review,Vol. 8 (Summer 1980), pp. 22-59. This content downloaded from 128.59.161.126 on Fri, 06 Mar 2015 15:27:41 UTC All use subject to JSTOR Terms and Conditions International Security1158 fearedor to which theyfeltmost vulnerable,or the assertionthatthe enemy would follow a course of action to which the particularbranch of the armed forcesmaking the assessment would be the most effectivecounter-if only it had the lion's share of all present and futurebudgetaryappropriations.38 Similarforcesoperate on broader questions of net assessment. It is no accident thathawks and doves differnot only in theirviews of Soviet intentions but also in theirviews of Soviet militarystrength.While it is logicallypossible forthe latterdifferencesto have determinedthe former,in factthe opposite seems to be the case. General political judgments usually form firstand stronglyinfluenceestimates of capabilities in a way that reinforcesexisting policy preferences. IMAGES AND SELF-IMAGES In all too many cases, assessments are also stronglycolored by national or racial stereotypes.39Germans are seen as efficient,Austrians as undisciplined, Britishas soft,Americans as too preoccupied with business to fight, and Japanese as generallyinferior.In many areas of intelligence,improvement is hard to find;but governmentsand public opinion do seem to have become much more sophisticated on this point. Because nationalism has become less virulent and racial generalizations are discredited, distorting prejudice of thiskind no longer plays a major role. (I findquite unconvincing the argumentthat American policy in Vietnam was significantly influenced by racial stereotypes.) At least as important,but less remarked upon, is the influence of selfimages. The way states view othersis partlya productof the way theythink of themselves. Indeed, estimates of the militarybalance are often driven more by analyses of one's own strengthsand weaknesses than by analyses of those of the other side. Thus in December 1940, when General Marshall discussed "keeping the fleet in the Pacific until a major offensivebegan against the Axis in the Atlantic,. . . he did not mentionGerman or Japanese forcesor plans and, indeed, did not seem to be thinkingof them at all. He was concentratingon American capabilities."40Although this habit of losing sight of the adversary is unfortunate,it is not surprising.Decision-makers 38. Donald Cameron Watt, personal communicationto Ernest R. May, quoted in May, "Conclusions: Capabilities and Proclivities,"in May, KnowingOne's Enemies,p. 541. 39. See May, KnowingOne's Enemies,pp. 111, 216, 333, 356, 364, 446, 474, 477, 493, 505. 40. Kahn, "United States Views of Germanyand Japan," p. 478. For otherexamples, see May, pp. 197-198. KnowingOne's Enemies,pp. 359, 379, 449; and Snyder,IdeologyoftheOffensive, This content downloaded from 128.59.161.126 on Fri, 06 Mar 2015 15:27:41 UTC All use subject to JSTOR Terms and Conditions Intelligence and ForeignPolicy1159 know more about their own forces than about those of potential enemies. Much of theirtime is spent listeningto theirgenerals and determiningthe size of the militarybudget and the shape of the armed forces.Furthermore, beliefin theirautonomy and efficacycan lead themto thinkthatthe outcome of any battlewill be more stronglyinfluencedby the conditionof theirarmies than by that of the armies of the other side. This would explain why many people judge proposed arms controlagreementsseverely;the restrictionson theirforcesloom larger than those on the adversary's. This is not to argue that decision-makersare prone to overratetheirown strength. Of course some cases when they do will receive our attention because they lead to disastrous wars, but underestimatesof strengthwill often pass unnoticed because they will usually lead to peaceful and less dramaticpolicies (although under some circumstances,theycan also produce preventivewars). But it does seem that variable moods of national revival and self-doubtstronglyinfluence the optimism or pessimism of the judgments.41When decision-makerslook abroad, what they see is oftenheavily colored by their views of the condition of their own society. More importantly,elites and general publics alike not only thinkabout themselves;they usually thinkwell of themselves.42When states are not consciously expansionistic,they tend to see theirown actions as purely defensive and cannot believe that others feel threatened by what they are doing. Even cynical leaders are self-righteous.They rarelysee theircountryas an unbiased observer would, let alone as others do. Indeed they often fail to grasp the natureof theirself-imagesor the factthatothersare not likelyto share them. As a result,states oftenunderestimatethe extentto which theyare following double standards or infringingon the vital interests of others. Any hostilityby the other side is more readilyattributedto the other's aggressive designs than to a reaction to the state's own behavior. The failureto accuratelyperceive the other side is then oftena product of the failureto grasp the nature of one's own policy. Thus Raymond Garthoffdemonstrateswith painstakingand disinterestedanalysis that a prime reason why Americans misjudged Soviet behavior during the period of detente was that theyfailed to understand that theircountrywas "waging a vigorous competitionalong with the Soviet Union."43With the possible exception of the SALT commit41. See, forexample, May, KnowingOne's Enemies,pp. 111, 147, 169, 539-540. 42. Ralph White,NobodyWantedWar (Garden City,N.Y.: Doubleday, 1968). 43. Raymond Garthoff,Detente and Confrontation (Washington,D.C.: Brookings,1985), p. 1083. This content downloaded from 128.59.161.126 on Fri, 06 Mar 2015 15:27:41 UTC All use subject to JSTOR Terms and Conditions International Security1160 ments, it is hard to find any instances in which the U.S. restraineditselfin orderto conformto the principlesof detente. Changes in public and Congressional attitudes (the "post-Vietnam syndrome") account for what internal inhibitionsoperated duringthe 1970s, and even these are easy to exaggerate. During detente, American political competitionwith the USSR in the Third World did not slacken, and the U.S. pushed the Soviet Union out of the Middle East and courted China. But Americans, leaders and public alike, saw theircountryas willing to cooperate on reasonable termsand as following a code of conduct that should have been acceptable to any partnerfitfor cooperation. Although we can know less about Soviet perceptions,they seem to have been similar. Even when the Soviets occupied Afghanistan,they could not understand that others were genuinelyoutraged and menaced. Therefore: the U.S. reaction. .. failed to carrythe intended message because the Soviet leaders could not believe thatthe Americanleaders meant it. They knew that Afghanistanwas not a vital interestof the United States, and yet the American reaction implied it was. Since the Soviet leaders knew they were not threateningreal and vital U.S. interestsin the Persian Gulf, they did not creditthe Americanexplanations forthe CarterDoctrineand the wide swath of sanctions.... The American action was seen not as a response to the Soviet move . . ., but as the line the U.S. leaders preferred,for which Afghanistanwas just a pretext.44 The problem is compounded for intelligenceanalysts because they are hampered not only by the general national self-images,but also by their ignorance of exactlywhat theircountryis doing. Because intelligenceis not supposed to be contaminated by or to meddle in the state's own foreign policy, analysts often know littlemore about this than does the informed newspaper reader. (ChristopherAndrew notes thatFrenchintelligencebefore World War I often learned about their country's secret policies through decoding theiradversaries' telegrams.)45But oftenother states' behavior is at least in part a response to the state's acts. When these acts are secret, analystswill not know of them and so cannot take them into account; even when they are open, unwrittennorms often inhibit analysts from giving themprominence.For example, one reason forthe Shah's lack of confidence in his abilityto control the situation in 1978 was that he believed that the 44. Ibid., pp. 964-965. 45. Andrew and Dilks, TheMissingDimension,pp. 36-37. This content downloaded from 128.59.161.126 on Fri, 06 Mar 2015 15:27:41 UTC All use subject to JSTOR Terms and Conditions Intelligence and ForeignPolicy1161 U.S. was not fullysupportinghim and indeed mightbe aiding the opposition. But it would not have been easy for intelligenceto analyze how the Shah saw the U.S. or what U.S. officialswere doing thatmightbe contributingto his fears. Thus regulationsand habits that are designed to keep intelligence and policy separate may badly skew the resultinganalysis. Conclusion Sun Tzu's words of 2,500 years ago have been much quoted: "Know the enemy and know yourself;in a hundred battlesyou will never be in peril." They are probably right, but both more worrisome and more accurate a descriptionis his accompanying comment: "If ignorantboth of your enemy and of yourself,you are certainin every battle to be in peril."46We should not expect too much of intelligence,especially because understanding the behavior of one's own countryis so importantand so difficult.We probably can do a much betterjob of determiningwhethera nuclear strikeis on the way-a crucialquestion, to be sure-than we can of determiningwhat others' foreignpolicies are likelyto be and what will influencethem. If the historical recordis a guide to the future(and thereis littlereason to expect otherwise), errorswill be common. Indeed, it is hard to find cases in which two states, even ifallies,47perceived each otheraccurately.The debates over the origins of World War I remind us that even afterthe fact,we usually argue about the causes of states' behavior and the alternativepaths they would have followedifothershad acted differently. Lack of information, cognitivebiases, commitmentsto established policies, and domestic interestsmake the task of intelligenceeven harder. The implications for decision-makingare not reassuring:intelligenceis likelyto be deficient,and policies thatwill succeed only if the images of others that underlie them are completelyaccurate are likelyto yield disaster. May's closing sentence is fullyjustified:"If just one exhortationwere to be pulled fromthis body of experience,it would be, to borrow Oliver Cromwell's words to the ScottishKirk: 'I beseech you in the bowels of Christthinkit possible you may be mistaken."'48 46. Sun Tzu, TheArtof War,trans. Samuel Griffith (Oxford: Clarendon Press, 1963), p. 84. 47. See the fascinatingstudy by Richard Neustadt, AlliancePolitics(New York: Columbia UniversityPress, 1970). 48. May, KnowingOne's Enemies,p. 542. This content downloaded from 128.59.161.126 on Fri, 06 Mar 2015 15:27:41 UTC All use subject to JSTOR Terms and Conditions