International Organization and the Study of World Politics

International Organization and the Study of World Politics Author(s): Peter J. Katzenstein, Robert O. Keohane, Stephen D. Krasner Reviewed work(s): Source: International Organization, Vol. 52, No. 4, International Organization at Fifty: Exploration and Contestation in the Study of World Politics (Autumn, 1998), pp. 645-685 Published by: The MIT Press Stable URL: http://www.jstor.org/stable/2601354 . Accessed: 03/01/2012 11:34 Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

The MIT Press is collaborating with JSTOR to digitize, preserve and extend access to International Organization.

http://www.jstor.org

International Organization and theStudyofWorldPolitics PeterJ.Katzenstein, Robert0. Keohane, andStephenD. Krasner

In thisarticlewe tellthestoryofthecreationandevolutionof a subfield, popularly knownas "IPE," thathas beencloselyassociatedwithInternational Organization IPE was definedby thetopicsthatit investi(IO) foralmostthirty years.Initially, gated,suchas trade,finance, rawmaterials politics,andmultinational corporations. Scholarsassociatedwiththefielddrewon economicsand on a varietyof existing and Marxismat the systemic theoretical orientations, notablyrealism,liberalism, level,andMarxism,statism, andpluralism atthedomesticlevel. Overtimetheboundariesof thissubfield, as we defineit,have been setless by subjectmatterthanby theoretical perspectives. Whereassome researchprograms havereliedheavilyon economics,othershave distancedthemselves bothfromthe substantive concerns ofthatdisciplineandtherationalism itrepresents. Sincewe are seekingin thisarticletodescribehowtheorizing aboutworldpoliticsas represented inIO evolved,we focuson IPE, rather thanon thesubstantive issuesofinternational politicaleconomywithwhichit began.We use the terminternational political whenwe referto real-world connections economy betweenpoliticsandeconomics; we usethetermIPE whenwe refer tothesubfield ofwork,centered inIO since1971, thatevolvedfromthestudyofinternational of politicaleconomytoanalyzea variety aspectsofworldpolitics. Like anynarrative, ourstoryreflects theviewpoints of itsauthors;and sincewe a in it reflects ourownexperiences andbiases.As played role theseevents, inevitably notedinthepreface, as a perspective, notas a canonical we putouraccountforward of whatis mostimportant. we makeno claimthatthe representation Furthermore, of thisarticlewe receivedlong,thoughtful, and detailedcommentson previous Duringthewriting drafts. and equilibrating thehistorical narrative thatwe are Theywereenormously helpfulin clarifying here.Severalcolleaguesoffered valuableoralcomments as well.WewouldliketothankEmanuel offering PeterGoureAdler,David Baldwin,MarcL. Busch,BenjaminJ.Cohen,David Dessler,Judith Goldstein, Miles Kahler,CharlesKupchan,David vitch,StephanHaggard,ErnstB. Haas, PeterHall,RobertJervis, SimonReich,ThomasRisse,JohnGerard Lake,RobertLieber,HenryNau,JosephNye,M. J.Peterson, ArthurStein,JaniceStein,AlexanderE. Wendt,and Oran R. Ruggie,Bruce Russett,KironSkinner, Young. International Organization 52, 4, Autumn1998,pp. 645-685 ? 1998byThe 10 Foundation andtheMassachusetts ofTechnology Institute

646 Intemational Organization evolvingsubfield ofIPE encompassesthemostimportant workthathasbeendonein international politicsoverthelastthirty years.Majorresearch hasbeencarriedouton subjectssuch as war initiation, the "democraticpeace," and alliancepolitics,to mentiononlya few.Some theoretical orientations thathavebeenhighlysalientfor thestudyof security, especiallyorganization theory andcognitive psychology, have been muchless consequential forIPE. Studiesgenerated by large-Nstatistical researchprograms suchas theCorrelates ofWarprojecthavenotbeenprominent inthe pages of IO. Since thisarticlewas written foran anniversary issue of IO, it seems appropriate to groundit in themajorlinesof workforwhichthejournalbecame known. ofina variety ofways.In thisarticlewe focus Theoryinourfieldhasbeenthought on twoof its meanings:generaltheoretical orientations and specificresearchprograms.General theoreticalorientationsprovideheuristics-they suggestrelevant variablesandcausalpatterns fordevelopingspecificresearch thatprovideguidelines programs. Politicalscienceis an eclecticdisciplinethatfindsmanyofitsmostfiuitfulideas elsewhere, andmanyofthegeneraltheoretical thathavebeen orientations relevantforIPE wereborrowedfromotherdisciplines,especiallyeconomicsand sociology. Rationalist theories derivedfromeconomics, forinstance, offer thefollowifyouhavea puzzle,formulate itas a problemforrationalactorswith ingheuristic: unproblematically specified interests, competing ina situation characterized byscarce resources.Constructivist in contrast, lookto thehumanities theories, and sociology forinsightsintohow "reality,"includingtheinterests thatpartiallyconstitute the ofactors,is sociallyconstructed. We argueinthefourth identity sectionofthearticle thatrationalism bothliberalarguments (encompassing groundedin economicsthat emphasizevoluntary agreement andrealistarguments thatfocuson powerandcoernow providethemajorpointsof contestation cion) and constructivism forinternationalrelations scholarship. variablesto a set of outcomes,or Specific researchprograms linkexplanatory variables.Whatare theeffectsof variousdistributions of power,or of dependent on thepropensity of statesto fightwars?Underwhatconditionsdo democracy, international institutions Whatinstitutional features actuallypromotecooperation? of state-society in theeffectiveness of foreign ecorelationships explainvariations nomicpolicy?Whensuchtheories aretestedwithevidence,answersareproposedanswersthatarevirtually becauseofthedifficulties oftheory alwayscontested speciandcontrolled statistical fication, testing, analysisthatbedevilthenonexperimental sciencesin general,andfieldssuchas oursinparticular. General theoreticalorientations-such as realism,Marxism,liberalism, statism, historical rational choice and construcpluralism, institutionalism, institutionalism, inthestudyofintemational IPE tivism-havebeenparticularly relations. prominent is no exception. Suchgenericapproachesdo notdisappeareasily.Theyprovidesugbutareconsisgestionsaboutrelevantvariablesandtheirpossibleinterrelationships tentwithmanyspecificresearch tasksandclusters oftestablehypotheses. Werefer to as specificresearchprograms,without thesesetsoftasksandhypotheses necessarily

The StudyofWorldPolitics 647 acceptingImreLakatos'sphilosophy ofscienceas applicableto ourfield.,The connectionbetweengenericorientations andresearch programs meansthattheintellecof genericorientations tualstanding is affected, thoughnotentirely determined, by withevidence,andalso dueto shifts empiricalevidence.As a resultofconfrontation in worldpoliticsitself,somehypotheses, and theresearchprograms in whichthey wereembedded,have receivedmoresupportthanothers.Genericolientations that sheltered productive programs benefited at theexpenseof competitors; thosethat seemedtoilluminate newdevelopments intheworldalso gainedadherents. Dissatisfactionwithexistingorientations and researchprograms, coupledwithchangesin theworld,has createdopeningsforalternative Thisevolutionary conceptualizations. andimperfect: processis oftenindirect thereareno "decisiveexperiments" ininternationalrelations a research muchless a genericapproachthat thatdiscredit program has spawnedresearchprograms. Fromits inceptionIPE has evolvedtwo relatedthoughdistinctive sitesforresearch:theinternational betweendomesticpoliticsand systemand theinteractions international politicaleconomy. The first sitefocusesattheleveloftheinternational on theinternational system.In the 1970s systemicscholarship politicaleconomy ofthedistribution ofpoweramong drewon realistarguments abouttheimportance in international orientation relations. states,thentheprevailinggeneraltheoretical ofliberalresearch IPE scholarsalso developeda variety thatwerereflected programs in discussionsof Europeanintegration, thegrowingroleof multinational corporaIn the 1980s thedifferences tions,and increasesin intemational interdependence. betweenrealismand liberalismweresharpenedas neorealistsand neoliberalsdebatedtherelativemerits oftheircontrasting analytical programs. Ourstoryaboutthesystemic variantofIPE goes as follows.In thelate 1960sand international relationsseizedan mid-1970ssomeyoungpoliticalscientists studying openingcreatedbothby eventsin theworldand in thesocial sciences.Growing levelsofinternational interdependence pointedtobya feweconomists helpedin the of transnational relationstheory. This analyticalapproachproconceptualization thestudyof videdan alternative tothestate-centric, realistapproachthendominating It was thefirst formulation thatone international relations moregenerally. analytical field.Transnationalism was challenged couldclearlyassociatewithIPE as a distinct whichhad bothliberaland realistvariants.These by hegemonicstability theory, werereformulated as a resultoftheirinteraction witheach initialresearchprograms otherand withsomevariantsof Marxism.Hegemonicstability theoryencountered was difficult to relations bothlogicalandempiricalanomalies.Transnational theory In the1980sneoliberal institutionalism anda specificrealistformulaoperationalize. tionknownas neorealism becametheprincipal interlocutors-institutionalists emphatheimportance sizedthepotential forinterstate whereasrealistsstressed cooperation, andratioof statepower.Sincethelate 1980sa newdebatebetweenconstructivism has becomemoreprominent as connalism(includingbothrealismandliberalism)

1. Lakatos1970.

648 International Organization tructivists have builton epistemological challengesrootedin sociologicalperspecsharednormsandvalues. tivesemphasizing The secondsiteforresearchhas beenthelinkagebetweendomesticpoliticsand theinternational politicaleconomy.Scholarsworkingon theseissuesinquiredinto oftheforeign thedeterminants economicpoliciesof statesandcorporate strategies, whichtheyhaveinvestigated withclose attention to empiricaldetail.The questions andautonomy askedinthislineofresearch focusedontherelativeinfluence ofsocial forcesandpoliticalinstitutions. Researchwas in generalmoreempirically oliented. Metatheoretical concernsloomedless largeforscholarsin thisarea thanforthose leveltheories, indomesticas wellas systemic withsystemic working although analyresearch choice sesgeneral orientations-such as pluralism, statism, Marxism, rational institutionalism, and historicalinstitutionalism-informed specificresearchprothefocuson state-society relationsstarted withconcreteanalygrams.Empirically, ses oftheOECD countries andLatinAmericanstates.It spreadfromtheretoencomtheanalysis pass thepoliticaleconomyof statesin otherworldregions,specifically of varietiesof state-leddevelopment outof socialismin in Asia and thetransition eastemEuropeandthestatesoftheformer SovietUnion. Scholarsof comparative wereat best politicsinsistedthatsystemicarguments in domesticstructures incomplete; variations (definedin termsof socialstructure as nationaland corporate well as groupand partyalignments) wouldlead to different responsesto the same externalpressuresand opportunities. Some of theearliest domesticpoliticsformulations associatedwithIPE drewon pluralism, thenthepreof approaches, hisvailingapproachto U.S. politics.Overtime,a variety including toricalinstitutionalism rationalchoiceinstitutionalism, and,morerecently, wereapnumber ofpoliticaleconomiesworldwide. pliedtoa growing Therewas a close interaction and international relations betweencomparative theboundaliesbetweenthesetwo scholars;indeed,some individualstranscended fields.But dulingthe1970sand 1980sthedifference betweentheinternational and domesticstrands oftheIPE fieldremained That distinction somesalient. hasblulred bothrealism whatinthe1990sbecause,as we arguelater,rationalism (incorporating and liberalism)and constructivism have establisheda new pointof contestation. Someofthespecific research programs generated bygeneraltheoretical orientations, suchas rationalchoiceinstitutionalism, aremorereadilyapplicableacrossdifferent and domesticorientalevelsof analysisthanwas thecase formostof thesystemic in earlierperiods. thatwereprominent tions,suchas realismandpluralism, The difference betweeninternational and domesticperspectives does notovershadow,however,a commonresearchpracticethathas markedempiricalworkin ofIPE. Scholarshavespecified research bothofthemainbranches ina way problems thatwas empirically and politicallysignificant. tractable, intellectually interesting, whichoutvariablesin causal mechanismsthrough They embeddedexplanatory comesseemedtotakeplace.Theyconducted research thatsoughttocontrol forsome otherexplanatory factors. Andtheysoughttodetermine whether theselectedexplanaeffecton specifiedoutcomes.Fruitful research toryvariableexerteda discernible combineddeductiveandinductive Butinthe workindifferent programs proportions.

The StudyofWorldPolitics 649 finalanalysis,anyproductive researchprogram had to suggestwaysin whichvariablesorconceptscouldbe operationalized; anditspropositions hadtobe,at leastin principle, empirically falsifiable. Generalorientations andtheirassociatedspecificresearchprograms typically enterintocomplexmutualinterrogations withtheirpreviouslyestablishedcounterparts.As a resultof suchdialogues,orientations arereconfigured ormerged;orthey remainseparate,withone approachlosingadherents; ortheycoexist,competing on relatively equal terms.ThomasKuhndoesnotprovidean accuratedescription ofthe studyof IPE or,forthatmatter, international relations moregenerally.2 IPE has not beencharacterized byscientific revolutions succeededbya periodofnormalscience inwhicha particular generaltheoretical orientation is uncontested. Therehas always beenvigorousdebatebetweencompeting generaltheoretical orientations andassociatedresearchprograms. Thelinesofthinking originally nurtured inIO, andlargelylimitedtointernational politicaleconomyand its intersections withcomparative politicsand international seemin the1990sto be merging institutions, intoa broaderandrichlyreconfigured fieldof worldpolitics.Insightsoriginallygeneratedin studying the international politicaleconomyhavebeenextendedtootherissueareas,suchas environmental or securityaffairs.Linkagesamongissues meanthatmanyimportant phenomenain worldpoliticscan no longerbe neatlyclassifiedbyissuearea,suchas economicsor security affairs. Andold boundaries betweeninternational relations andcomparative politics,orbetweenIPE andtherestofinternational relations, havebecomeincreasinglyfluid.Respondingto thesechanges,scholarshave modifiedexistinglinesof researchand initiated new ones.Theyhave sharpened someanalyticaldistinctions and erasedothers.Theyhave soughtto advancereformulated synthetic interpretationsandfocusedon newpointsofintellectual contestation. andrationalism Sincethemid-1980sa newdebatebetweenconstructivism (includhasbecomemoreprominent. Newtheoretical develingbothrealismandliberalism) institutional opmentsin rationalist theory, open-economy economics,andcomparativepoliticsprovidedscholarswithnewintellectual openingsas theColdWarended. influenced in thehumanities, Conventional andcriticalcontructivists, bynewtrends thatemphasizedsharednormsandvaluesbut putforward sociologicalperspectives frompostmodemism. termssharplydifferentiated whichwere in epistemological In thefieldof nationalsecuritythediscussionbetweenrationalism (in its realist andliberalvariants)andconstructivism has beenmorefullyjoinedthanin thefield ofIPE. of eitherworldpoliticsor international We do notpresumeto predictthefuture ofthesubfield of relations Butwe do insiston twopoints.First,theevolution theory. andhypothesis IPE is better describedintermsoffocalpointsofcontestation testing overanother. thanas an all-outwarleadingto thevictory ofone generalorientation intellectual in IPE canbe mademorecomprehensible Second,current developments andpoliticalopenings themwiththelastcycle-fromtheintellectual bycomparing 2. Kuhn1962.

650 International Organization of the 1970s to therelativesyntheses of thelate 1980s,whichpresagedanother the politicaland intellectual upheavalbeginningaround1990. Thus,we interpret pastinordertounderstand thepresent. In thefirst sectionofthearticlewe briefly reviewsomeaspectsofthefieldin the yearsbetweenthelate 1940s(whenIO was founded)and 1968.We thendiscussthe intellectual openingforIPE, beginning in thelate 1960s.In thesecondsectionwe betweenrealismandliberalism duranalyzethesymbiotic yetcontested relationship ingthelastthirty years.In thethirdsectionwe tracetheevolutionofdomesticpoliticsresearchanditseventualdifferentiation intohistorical-institutionalist andrationaliststylesofresearch. In thefourth sectionwe analyzethenewintellectual opening createdby theend of theCold Warand arguethatefforts fromvariousquartersto understand actorpreferences in thestudyof arecreating newpointsofcontestation worldpoliticswhileblurring someestablished faultlines.In theconclusionwe argue is integrating IPE evenmorefullyintothebroaderdiscipline thatcurrent scholarship of international relationsand intooverarching debatesin thenaturaland social sciencesandinthehumanities.

An IntellectualOpening,1968-78 an established tradition in the IO existedfortwodecadesbefore1968,representing studyof international organizations. Duringthe 1960s IO publishedsome of the leadingworkon Europeanintegration, workthatchallengedrealismand provided As discussed conceptsfortheanalysisofthepoliticsofeconomicinterdependence. inthissection,after1968an intellectual openingforthestudyofinternational political economyemerged. IO's FirstTwoDecades, 1947-67 10 was founded relations. An at a timeofprofound andrapidchangeininternational

extensiveanduntried setofinternational had beenestablished organizations during butalso to andjustafterWorldWarII. Seekingnotonlyto describetheiractivities, and whytheyhaveor promote"a comparative studyof international organizations havenotworkedinvarying the trustees of the WorldPeace Foundacircumstances," issueappeared tiondecidedinthespringof 1946toestablishthisjournal,whosefirst in February1947.3By thattime,relationsbetweentheSovietUnionand itsformer Western allieshadbecomehighlystrained, andwhatlaterbecameknownas theCold Warwas beginning, markedbycrisesin IranandGreeceand,in thespringof 1947, bytheTrumanDoctrineandtheMarshallPlan. boththeformation ofnewintemational IO hadthetaskofanalyzing organizations andthesuperpower thatthreatened to killor maimthemat birth.In thelead rivalry 3. TheWorldPeace Foundation was established in 1910byEdwinGinn,headofthepublishing companybearinghisname.

The StudyofWorldPolitics 651 articlein thefirst issue,LelandM. GoodricharguedthatthenascentUN was uncomfortably similarto theLeague of Nationsin its design:"Like theLeague,butfor somewhat different technical reasons,theUnitedNations,inso faras itsenforcement activitiesare concerned, is an organization fortheenforcement ofpeace amongthe smallerstates."4Otherarticlesin thefirst threeyearsof thejournal'sexistenceofferedgeneralarguments on theroleofinternational inworldpoliticsas organizations wellas articleson specifictopicssuchas thepoliticsofinternational airtransport and theoperationof theUN SecurityCouncil.5Laterthejournalpublishedtrenchant analysesof theUN's adaptation to theCold War.Some of theoutstanding workin thisvein,by scholarssuchas Inis L. Claude,Jr.,and StanleyHoffmann, shrewdly commented on thepoliticsof international organizations and theconsequentlimits on theirpotential rangeofsuccessful operation. UN peacekeeping operations inSuez andtheCongowerea particular focusofattention. Thisworkwas notnaive.Neither legalismnormoralism-thoseallegedbugaboos ofAmericans-obscured are theauthors' recognition thatinternational organizations transformative profoundly affected by worldpolitics,and thattheseorganizations' is modest,atleastintheshortrun.However,despiteitsemphasison realispotential tic descriptive informed analysis,researchwas notparticularly by generalsocial sciencetheory, anditwas less concemedwithtesting alternative arguments thanwas thecase forsubsequent worksuchas RobertCox andHaroldJacobson'sanalysisof decisionmakingininternational organizations.6 Perhapsforthesereasons,workpublishedin 10 in the1950sand 1960swas closerto policyanalysisand commentary thanworkpublished inthejournalsincethemid-1970s.Thejournalremained sharply focusedon formalinternational organization, withsubstantial spacedevotedto summariesofactivities invariousUN agenciesuntilthelate1960s.Evenwhenbehavioralism enteredthepages of thejournal,it tookthe formof statistical analysisof GeneralAssemblyvoting. If theUN had continued to be as significant in worldpoliticsas itbriefly became underSecretary-General Dag Hammerskj6ld (1953-61), boththeshrewdpolitical behavioralstudyofpolitics-withintheUN andin analysisandthemoresystematic tootherintergovernmental havecontinued tocomcomparison organizations-might audience.But Hammerskj6ld's manda significant deathin a planecrashwas followed by thecollapseof theUN operationin theCongo and by the "Article19 Crisis"overwhether theSovietUnioncouldbe deprivedof itsvotein theGeneral forUN peacekeeping ofassessments Assemblyinresponsetoitsnonpayment operain Vietnamtookplace without UN involvement. tions.U.S. intervention significant Andin 1967 theUN peacekeeping forcewas withdrawn fromSuez rightbeforethe JuneWar.ActionsagainstIsrael,including thenotorious "Zionismas Racism"resofortheUN, including reducedU.S. support severalofthespeciallution,drastically irrelevant tomajorquestionsofworld ized agencies.As theUN becameincreasingly ofworldpoliticslostinterest init. politics,students 4. Goodrich1947,17. 5. See Rothwell1949;Little1949;andDennett1949. 6. Cox andJacobson1973.

652 International Organization TheIntellectual RootsofIPE Realismhas beenat thecenterof thetheoretical debatesof U.S. international relationsscholarship fora longtime.Historically, realismwas a breedalientotheliberal and progressive intellectual and politicalsensibilities of theUnitedStates.At its inception in theearlytwentieth century, thedisciplineofinternational relations was partoftheprogressive movement thatsoughttobuilda systematic socialsciencefor thebetterment ofmankind intheUnitedStatesand,byimplication, worldwide. World WarII and theHolocaust,experienced and interpreted by a generation of brilliant intellectuals closelylinkedto Europe,changedthis.And so did theprotracted Cold Warthatheldtheworldin its gripforfourdecades.Prudentstatecraft, realismargued,requireda space fordiplomacyand strategy thatwas uncontested by normal domesticpolitics.The granddebatesinthefield-idealismandrealismin the1930s, neoliberalism andneorealism inthe1980s-are products ofthisdistinctive historical legacy. No independently fieldforstudying theinternational recognized politicaleconomy existedin the1950sand 1960s.Specialistsof international relations paid littlesystematicattention to thepoliticalanalysisof economicissues.For publicpolicyas wellas fortheacademy,thefocuswas on security issuesand "highpolitics."Communismwas seenas an omnipresent threat. The SovietUnion,armedwithnuclear weaponsandrockets-as thesuccessfullaunchof Sputnikin 1957 indicated-was perceivedas a seriousmilitary, technological, economic,andideologicalrivalofthe UnitedStates.Leadingstudents ofworldpoliticsanalyzedtheroleofnuclearweapofdeterrence, andtheoperation ofU.S.-ledalliancessuchas NATO. ons,techniques ThomasSchellingintroduced rationalchoiceanalysisto international affairs; Henry andU.S. nuclearandalliancestrategies in KissingerwroteaboutEuropeanstatecraft the Europe;GrahamAllisonusedtheCubanMissileCrisisas a case studyto sharpen theoretical lensesofthediscipline.7 Innovative analysesalso dealtwithissuesofwar and peace,includingmajorconceptualcontributions by KennethN. Waltzand the ofWarProjectledbyJ.David ofanimpressive datasetbytheCorrelates development Singer.8

stockofintellecForthestudyoftheinternational politicaleconomya substantial tualcapitalcamefromclassicworkson politicaleconomy,suchas thoseofAlbert andJacobViner.9 for Hirschman, EugeneStaley,CharlesKindleberger, Hirschman, ofpoliticalpressure tradewas usedas an instrument example,explainedhowforeign in central-eastern EuwhichGermany builtitspoliticalsphereofinfluence through thecomplementarity ofpowerandwealth illuminated rope.VinerandKindleberger in theseventeenth and eighteenth and centuriesand in theinternational monetary financialsystem.Staleyexaminedhow barriersagainsttradewithstatessuch as conflict.These authorsoffered Japancreatedconditionsforpolitical-military 7. See Schelling1960,1966;Kissinger1957a,b,1965;andAllison1971. 8. See Waltz1959;andSinger1972. 9. See Hirschman [1945] 1980;Staley1939;Viner1948;andKindleberger 1973.See also D. Baldwin 1985.

The StudyofWorldPolitics 653 insights intothewaypowerprovidedthefoundations fortheinternational economy. Withinthefieldof international relations, EdwardHallettCarrused his blendof Marxismand realismto analyzethecollapseof theinternational order,including theinternational economy,in theinterwar period.10 The analyticalfocusof these books on materialcapabilitiesand asymmetric bargainingpowerprovidedthe basis forparsimonious accountsof developments in theinternational economythat had a profoundeffecton the studyof theinternational politicaleconomyin the 1970s. Studentsof comparative politicswho soughtto linktheirworkto international relations drewon a richtradition ofworkthatemphasizedtheimportance ofinstitutionalfactorsin theshapingof politicalregimes,notably, in the1960s,J.P. Nettl's articleon thestateand SamuelHuntington's analysisofpoliticaldecay.11 Theyalso drewon economicsand economichistory, suchas workson technology by David Landesand on tariffs byCharlesKindleberger.12 Some bookson U.S. politicswere politics.13Furthermore, widelyreadby scholarsof comparative thedomesticdeterminants offoreign policyreceivedextended attention byinternational relations scholarsinthe1960s,a factthatreceivedmorethanpassingnoticebycomparative politics specialistswhowereinterested ininstitutions. In 1959Kenneth Waltzinsistedon the primacy oftheinternational system inshapingstatepolicy,foreshadowing hisTheory 14 A decadelater, in a Politicspublishedtwenty ofInternational yearslater. however, brilliant bookon democratic foreign policyhe offered a theory offoreign policy.15In the1960sandearly1970sWaltzwas notalonein pointing to theimportance ofthe domesticdeterminants (such as politicalleadership, institutions, and bureaucratic organizations) of stateaction.HenryKissinger, GrahamAllison,MortonHalperin, andJohnSteinbruner, amongothers, publishedimportant articlesandbookson the subjectthatwerewidelyread,discussed,andcited.16 forstudents Yet,themainsourceof inspiration seekingto connectinternational relations withdomesticpoliticsarosefromwithin thefieldofcomparative politics.In to explain the 1960s discussionsin comparative analysisrevolvedaroundefforts in thepathstakenbymodernstates-towardliberaldemocracy, differences fascism, orstatesocialism,forexample.Interactions anddomesticfacbetweeninternational torshad beenlargelyignoredby thestructural-functional approach,whichwas the focalpointofworkincomparative politicsduringthe1960s.Butforscholarssuchas and subsequently Barrington Moore,ReinhardBendix,AlexanderGerschenkron, weretheresultnotonlyof a CharlesTillyand Theda Skocpol,thesetrajectories intotheinternational characteristics butalso ofhowitwas inserted polity'sinternal of comparative systembotheconomicallyand politically.17A youngergeneration 10. Carr[1946] 1962. 11. See Nettl1968;andHuntington 1968. 12. See Kindleberger 1962,1978;andLandes1969. 13. Bauer,Pool,andDexter1972. 14. Waltz1959. 15. Waltz1967. 16. See Kissinger1969;Allison1971;Halperin1974;andSteinbruner 1974. 17. See Bendix1964;Gerschenkron 1962;Moore1966;Tilly1975,1990;andSkocpol1979.

654 International Organization scholarswouldsoonadapttheinsights oftheseseminalstudiestotheanalysisofthe domesticdeterminants oftheinternational politicaleconomy. Despitethisrichintellectual legacy,the1960slackeda coherent bodyofpoliticaleconomyliterature in international relations,and thescattered worksthatexisted werefarfromthemainstream ofinternational relations. couldgo Graduatestudents through international relations programs atmajorAmerican universities, focusingon diplomatichistory, theoriesof war and peace, and policyissuesrevolvingaround deterrence, andremainlargelyinnocent ofeconomicsoritslinkstoworldpolitics.It didnothelpthateconomists werepayinglittleattention tohowpoliticalandinstitutionalcontexts affected theirsubjectofstudy. EconomicsPh.D. students wereincreasinglyrewarded fortheirmathematical acumenrather thantheiranalysisofhoworganizationsactuallyoperated;economistscould not fashionsuccessfulcareersby investigating thelendingpractices oftheInternational Monetary Fund.Hence,inthe analysisofinternational relations neither politicalsciencenoreconomicswereconductingsustained research on subjectsthatwouldlaterbe termed IPE. European Integrationas an IntellectualOpening

Fromtheearly1960s untiltheearly1970s,theliveliestdebateson international organizations focusedon thesignificance ofpoliticalintegration forthenation-state: "obstinateor obsolete?"as StanleyHoffmann putit in a famousdebatewithErnst Haas.18Haas and some youngerstudentsand colleaguesdevelopeda specificresearchprogram aboutpoliticalintegration, drawingon modernization theory andon thetheory offunctionalism as articulated byDavidMitrany.'9 Thisresearch program a generalliberaltheoretical of reflected orientation thatpointedto thepossibility of voluntary agreemultipleactorsin theinternational systemand theimportance ment.Theirworkcomplemented theanalysisofnationalandsupranational communityformation thatKarl Deutschand his students had pioneeredin the 1950s and 1960s,a research program thatforeshadowed whatlatercametobe labeledconstrucofidentity formation measuredbysocialtranstheimportance tivism.Ithighlighted a substantial hadspecified actionsandcommunications.20 By 1970integration theory a seriesofpolitinumber ofeconomicandsocialbackground factors thatconditioned and cal processes.These,in turn,shapedhowpoliticalactorsdefinedtheirinterests states.A nascentfieldof thusthepolicystrategies thatelitesadoptedin different studieswas formed, basedon fieldwork comparative regionalintegration byscholars who wentto East Africaand LatinAmericaas well as to Europe.2'In 1970 10 editedbyLeon LindbergandStuart publisheda specialissueon regionalintegration Thearticlesinthisvolumerepresented touse behavioralsocial Scheingold. attempts inthesuccessofefforts atregionalintegration.22 scienceto accountforvariation 18. Hoffmann 1966. 19. See Haas 1958,1961,1964b;andMitrany1966. 1966. 20. See Deutsch1953;Deutschetal. 1957;Russett1963;andMerritt includeLindbergand 21. Outstanding bookssynthesizing thisliterature andpresenting newfindings Scheingold1970;andNye 1971. 22. LindbergandScheingold1971.

The StudyofWorldPolitics 655 Comparative regionalintegration studiesfoundered on theoverlyoptimistic assumptionof neofunctionalism thatsuch efforts wouldsucceedaroundtheworld. Non-European efforts at regionalintegration in the1960sfailed,and theEuropean effort afterCharlesDe Gaulle'smaneuvers ofthemid-1960sandtheLuxstagnated embourgCompromise of 1966.The fieldofintegration studiesatrophied. Formerly optimistic theorists becamepessimistic. Theirarguments becamemoreindeterminateas an increasing number ofexplanatory variableswas neededto accountforthe andretardation acceleration ofintegration processes. The difficulty encountered after1966byneofunctionalism was ironically similar tothatfacedbyneorealism(whoseerrorwas excessivepessimism)after1989.Both approachesinvolvedspecificresearchprograms withtestablepropositions, someof whichwereinvalidatedby events.Theirarguments in waysthat wereformulated couldbe falsified. Whendisillusionment setin,manyyounwithintegration theory gerscholarsturnedto newerthemessuchas thoseofinternational andcomparative politicaleconomy.Integration theory, however,providedmuchof theexistinganalyticalcapitalforinvestigating theconnections betweenpoliticsandeconomics.Severalyearslater,ErnstB. Haas publishedan articleon the"obsolescence"ofregional integration theory.23 However,empiricalworkon theEuropeanCommunity continued in Europe,and whenU.S. scholarsregainedinterest in thesubjectin thelate 1980s,theyreliedheavilyon thisEuropeanworkand wereless inclinedto putforwarda generaltheory ofintegration processes.24 The Emergenceof IPE

As longas security concernsdominated theattention of academicsandpolicymakvacuumin thestudyoftheinternational ers,theintellectual politicaleconomycould in thelate 1960s,however, a seriesofdevelopments be ignored.Beginning brought economicissues:therevivalof theEuropeanand politicalsalienceto international intheUnitedStates,theabandonment of Japaneseeconomies,inflationary pressures Bretton Woodsin 1971,andtheOPEC oil embargoin 1973-74all combinedto end the era of overwhelming U.S. dominanceof theworldeconomyand theregime constructed inthelate1940s. A fewpoliticalscientists beganattheendofthe1960stoobservethesechangesin theinternational stimulated politicaleconomy, especiallybyworkoneconomicinterdependenceby RichardCooperand on multinational enterprises by RaymondVerIn 197110 publisheda specialissueon transnational relations.26 Thisvolume non.25 doneby Vernonand others: was inspiredby theworkon multinational enterprises in theeditors'introduction overa thirdofthethirty-one worksreferenced scholarly to theactivities ofthesefirms. The editorssoughtto broadenthestudyof pertained 23. Haas 1975,1976. 24. See Moravcsik1991, 1998; and thecollectionseditedbyWallace 1990; Keohaneand Hoffmann 1991;andSbragia1992. 25. See Cooper1968;andVernon1971. witha twenty-year perspective, 26. KeohaneandNye 1972.For a reviewofthesubsequent literature see Risse-Kappen1995b.

656 International Organization worldpoliticsto include"transnational relations," involvingtheactivitiesof nonstateactors,and "transgovemmental relations," in whichsubunitsof governments dealtdirectlywiththeircounterparts The transnational relationsspecial abroad.27 activity issuewas chieflya pointingexercisethatmadeclearhowmuchinteresting paradigm. No hadescapedtheattention ofanalystsimprisoned inthe"state-centric" testabletheorywas presented. The editorsfocusedon theimpactof changingecoassumptions were nomicand technological forceson politics.Theirfundamental borrowed frompluralism, whichwas thereigning theoretical orientation ondomestic politics,atleastintheUnitedStates.Actorswerethestarting point;theirpreferences werenotexplained.The factthatactorshad different createdopportunipreferences tiesformutually beneficial trade-offs. Beginning withtheperceivedrealityoftransdepennationalrelations, Robert0. KeohaneandJosephS. Nyediscussedinterstate thetrade-offs andpolitical dence,andtheyhighlighted betweentheeconomicbenefits oftransnational relations costsofinterdependence. theimplications By considering to classic issues in forstateautonomy, theylinkedthestudyof interdependence politicalscience. RobertGilpinchallengedtheeditors'liberalorientation witha powerfulessay inthesamevolume.28 relations couldonly published Gilpinarguedthattransnational A central be understood within thecontext ofinterstate andcontinuing issuepolitics. theinteractions economicchanges,on theonehand,andstate betweentransnational power,on theother-was embeddedin theinternational politicaleconomydebate Gilpindevelfromtheoutset.In a subsequent bookon multinational corporations, andself-consciously, a realistanalysisoftheinternational politioped,systematically cal economy.His trenchant and corporations analysisof therole of multinational U.S. foreign andpolicyconcernsintoa policyfusedrealism'sanalytical, historical, powerful synthesis.29 Gilpinshowedhow realism'semphasison powercouldprovidebotha politicalexplanation oftheemergence ofliberalprinciples andpractices in U.S. foreign economicpolicyand a parsimonious critiqueof liberalscholarship. whichinterpreted His dynamicmodelof changeparalleledsomeMarxistwritings, thatdrovepower-seeking theVietnam Waras a reflection ofeconomiccontradictions a counterrevolutionary elitesto devise,undertheflagof international liberalism, imperialist strategy.30 inthe a debateabouttheroleofstatesandmarkets Gilpin'sworkhelpedcrystallize international politicaleconomy,whichbecamea majorthemeduringthenextdeandtotry whattheymeantbyinterdependence cade.Scholarsbegantoconceptualize

relations."TheyquoteArnoldWolfers 27. KeohaneandNyedidnotinventthephrase"transnational publishedin 1959.See KeohaneandNye 1972,x. Wolferstreatsthe"transnaas usingitin an essayfirst toidentify a set stating that"thereis no lackofa suitablevocabulary tional"terminology as conventional, called thatall thetermsreferto something significance ofnon-state corporate actors,butitis notwithout Wolfers1962,20. feature ofthenation-state." 'national,'whichis thecharacteristic 28. Gilpin1972. orpowerful as workwitha realistcast,albeitnotso systematic 29. Gilpin1975.Forotheroutstanding Gilpin's,see Knorr1975;andStrange1976. 30. See BaranandSweezy1966;Magdoff1969;Mandel1976;andGilpin1975,1981.

The StudyofWorldPolitics 657 to measureit.3IThe writings ofSusanStrangewereparticularly in focusimportant ingattention on markets andhow statesinteracted withthem.32 On one side of the debatein the1970s wereliberalanalysesthatemphasizedeconomicprocessesand technological changein theprocessofmodernization.33 Thesearguments havetheir counterparts in some contemporary workon globalization, which,in thewordsof Strange,contendsthat"the authority of thegovernments of all states,largeand small,strongandweak,has beenweakenedas a resultoftechnological andfinancial changeandoftheacceleratedintegration ofnationaleconomiesintoone singleglobal market economy."34 In contrast to thesemarket-oriented is thestatearguments structural orientation ofGilpinorofStephenKrasner, whoarguedin 1976thatvariationsinthedegreeofopennessorclosureinforeign tradewereinexplicable without understanding configurations of stateinterests andpower."In recentyears,"he declared,"students ofinternational relations havemultinationalized, transnationalized, bureaucratized, and transgovernmentalized thestateuntilit has virtually ceased to exist as an analytic construct.

.

. This perspectiveis at best profoundlymislead-

ing."35 Whether one emphasizedtheroleof markets or of states,it becameclearduring the1970sthatneither couldbe adequatelyanalyzedwithout phenomenon accounting fortheother.As KeohaneandNyeexplicitly recognizedin theirbookpublishedin 1977,powerandinterdependence wouldhavetobe analyzedtogether.36

IPE: Liberal Challengesto Realism Duringthedecadeafterthepublication of Gilpin'sbookthevariousapproachesto and cooperation problemsof bargaining drewtheirinspiration fromlong-standing andMarxism.Eachperspective genericorientations, notably realism, liberalism, gave prideofplace to a different explanatory variable:thedistribution ofpowerforrealof different of theeconomyor, ists,theinterests groupsforliberals,thestructure moresimplistically, theinterests ofcapitalists forMarxists. Each perspective emphasized different causal relations:powerand coercionforrealism,mutualagreement and contracting forliberalism, mechanisms of exploitation forMarxism.Thesegenericorientations createdspecificresearchprograms thatweresubjectto empirical anddepenliberalregimetheory, verification, including hegemonicstability theory, variants The mostprominent debateswerethosebetweendifferent of dencytheory. realismand liberalism, and increasingly betweensystemic-level analysis(particuand analysisrootedin domestic-level interests and larlyneoliberalinstitutionalism) 31. See Waltz1970;RosecranceandStein1973;andKatzenstein 1975. 32. Strange1979,1988. between 33. Morse1970and 1976.In an investigation ofwar,Gilpinfurther exploredtherelationships technological changeandworldpolitics.Gilpin1981. 34. Strange1996,14. 35. Krasner1976,317. 36. See KeohaneandNye 1977;andBaldwin1979.

658 International Organization institutions, whichsoughttounderstand "stateinterests," rather thantakingthemfor granted. Realism and Its Critics in the 1970s

For mostAmericanstudentsof international politics,at least through the 1980s, realismwas theperspective againstwhichnewideashadtobe tested;andduringthe 1960s,whenIPE emergedas an independent fieldofstudy, realismwas intellectually hegemonic. HansMorgenthau's PoliticsAmongNationswasthemostimportant text.37 Kenneth Waltzhadpublishedhisfirst bookin 1959inwhichhe initiated theintellectualprojectthatculminated in his influential expositionof whatcameto be termed neorealismtwentyyearslater.38 Systemstheorywas guidedby a realistrhetoric, as Waltzlaterpointedout,thesestudiesweremorereductionist although, thaneither theirauthors orreadersrealized.39 But realismhas alwaysbeen vulnerable, because some of its core assumptions wereproblematic, andbecauseitsempiricalvalidationwas nevercompelling. Realism's coreassumptions can be variouslyclassified, butfourareparticularly important:(1) statesarethekeyactorsin worldpolitics;(2) statescan be treated as homogeneousunitsactingon thebasis of self-interest; (3) analysiscan proceedon the basis of theassumption thatstatesactas iftheywererational;and (4) intemational in the international anarchy-theabsence of any legitimateauthority systemmeansthatconflict betweenself-interested statesentailsthedangerof warand the possibilityof coercion.The state-centric assumption was challengedby workon transnational thehomogeneity of domesticstrucrelations, assumption by students tureand bureaucratic politics,therationality assumption by analystsof cognitive of andgroupdecisionmaking,andtheanarchyassumption psychology bytheorists intemational societyandinstitutions. their Duringthe1960sand1970sthreemajorliberalchallengestorealismdirected fireat theassumption thatstatescouldbe treated rationalactors:neofuncas unified andlinkagepolitics.40 All bureaucratic relations tionalism, politics,andtransnational threeweregroundedin a pluralistconceptionof civil societyand thestate.Public interests. policywas theresultof clashesamongdifferent groupswithconflicting whichwouldvaryfromone Groupscouldoftenonlysucceedbybuildingcoalitions, issueareatoanother. RobertDahl,themostinfluential Americanexponent ofpluralof thatthesecross-cutting ism,emphasized cleavageswouldprecludethedominance about anyone specificgroup,an observation designedto rebutMarxistarguments theformulation theabilityofmajorcapitalists todominate ofpublicpolicy.4' of Neofunctionalism thatinstitutional stipulated changewouldaltertheincentives groupsin civil society,leadingthemto supportpoliciesthatwouldpromotestill 37. 38. 39. 40. 41.

Morgenthau 1948. Waltz1959,1979. See Kaplan1957;Rosecrance1963;andWaltz1979,1986. See Haas 1958,1964a,b;Allison1971;KeohaneandNye 1972;andRosenau1969a. Dahl 1961.

The StudyofWorldPolitics 659 moreintegration ina processthatwouldspilloverfromoneissueareatoanother. But as we have seen,thisargument apparently failedto predictor explainthedirection takenbytheEuropeanCommunity after1966.42Bureaucratic politicsextendedpluralistinterest grouparguments intothegovernment itself:"whereyoustanddepends uponwhereyousit."43Policy,including foreign policy,is a productofcompromise amongdifferent bureaucratic actors,reflecting thepowerandintensity ofinterest of bureaus,whichvaryfromone issueareato another. Coherent foreign policyis elusive,sincedecisionsin different issueareasarebeingtakenbydifferent actorswith different Theoriesabouttransnational preferences. relations, whichwerecentral when IPE emergedas a distinct fieldin the1970s,canied thepluralist argument one step further: therecouldbe manydifferent actorsintheinternational includenvironment, inggroupsfromcivilsociety.Thesenongovernmental actorscouldbe directly linked in othercountries, withtheircounterparts in tiesthatwouldnotnecessarily be controlledbystates.Transgovernmental relations werealso possible:situations inwhich therewouldbe directrelationsamonginterested bureaucracies in different countries.44 For IPE, bureaucratic politicsand transnational relationsorientedearlyworkin thefield,buttheseresearchprogramsfaltereddespitethefactthattheirpolitical a nuancedlandscapecomposedof manydifferent ontology, elements,was much richerthantheblack-boxviewof thestateprovidedby realism.Transnational relainterests and tionspositeda worldcomposedofmanydifferent actorswithdifferent Buttheoperationalization capabilities.Sucha modelcan providea richdescription. ofcauseandeffect is complexbecauseitis difficult to specifyinterests relationships andcapabilitiesex ante.The largerthenumberofactors,thegreater thediversity of and thewiderthenumberof theirresources(ideas,money,access, organization); suchspecification possiblealliances,themoredifficult becomes,especiallyifthere areinteraction effects amongdifferent groups. Bureaucratic and politicscaptured thecomplexity involvedin policyformulation in any advancedpolity-complexity thateludedrealismwithits implementation the aboutthenatureofthestate.In someinstances radicallysimplifying assumptions interests ofbureaucracies wereclear,butin otherstheyweremoreelusive.Some of ofbureaucratic thatthemostobviousspecitheinitialproponents politicsrecognized fication ofbureaucratic interests, maximizing budgetsize,didnotalwayswork.For WorldWarII didnottrytokeepcontroloftheairforce example,theU.S. Armyafter thecreation servicethatwouldbe coequal butrather ofa third supported independent withthearmyandthenavy.45 Armyleadersrejectedkeepingtheairforceas partof WorldWarII) becauseitwouldhave thear-my (itwas theU.S. ArmyAirForceduring changedthe "bureaucratic essence,"or whatwouldbe called todaythecollective andtheinfantry. Bureaucratic ofthearmy, whichwas rootedinlandwarfare identity through theEuropeanjudicialprocess,over42. BurleyandMattlisuggestedthatspillovercontinued lookedbymanyanalysts.BurleyandMattli1993. thisaphorism toDon K. Price.Allison1971,316,n. 71. 43. Allisonattributes 44. KeohaneandNye 1974.See also Rosenau1969a. 45. Halperin1974.

660 International Organization politics,like transnational relations, had greatdifficulties in operationalizing varitorealistperspectives. ables.It didnotbecomea compelling alternative In themid-1970sa newliberalchallengeto realismbeganto emerge.This challengefocusedon theconceptof "international regime,"drawnfroma long-standing tradition ofinternational law andfirst usedin thepoliticalscienceliterature byJohn Ruggieand subsequently elaboratedby Keohaneand Nye.Ruggiedefinedregimes as setsof "mutualexpectations, rulesandregulations, plans,organizational energies and financial commitments, whichhave beenacceptedby a groupof states."KeohaneandNyetreatedthemsimplyas "governing arrangements thataffect relationwas sociologicalorconstructivshipsofinterdependence."46 Ruggie'sunderstanding ist,emphasizing theimportance ofintersubjective, sharedunderstanding thatdefines rather thanjust reflects thepreferences of actors;Keohaneand Nye understood regimesas devicesforenhancingtheutilityof actorswhoseinterests weretakenas given.Studentsof international regimesdid notchallengeone of themeaningsof "anarchy":thatno institutional hierarchy capableofenforcing rulesexistsin world thatanarchy inthissenseimplied politics.Theydidquestionthefrequent implication theabsenceofinstitutions basedon rules. However,regimescouldalso be explainedinrealistterms-as GilpinandKrasner had,ineffect, soughttodo. By theendofthe1970sstudents ofinternational regimes had introduced a potentially important new dependent variableintotheanalysisof worldpolitics.At thattime,however,thisnew dependent variablewas notlinked setof explanatory causal mechawitha distinctive variablesthrough an articulated nismand,therefore, didnotseriously thewell-articulated threaten explanatory project ofrealismbasedon interests, power,andanarchy. Supportforrealisttheorywas enhancedby thefactthatthebest-operationalized oftheseyears,hegemonic consisnewresearch program stability theory, was entirely itsfirst tentwithrealistpremises-although, articulation ironically, (withsomewhat different wasbya liberaleconomist, terminology) Kindleberger.47 Hegemonicstabilthatan open international itytheorymaintained systemwas mostlikelyto occur Kindlewhentherewasa singledominant economicsystem. powerintheinternational bergerarguedthattheGreatDepression,a marketfailureof monumental proporfinantions,was causedbytheabsenceofa lenderoflastresortin theinternational cial system. iftherewerea singledominant Therecouldonlybe a lenderoflastresort powerin theinternational system.Onlya hegemonwouldhave thecapacityand interest used a to providethepublicgood of financialstability. Thus,Kindleberger realistontology (theactorswerestates)anda halfrealistcausality(theoutcomewas thevolundetermined of poweramongstates,althoughthrough by thedistribution theprovision tarychoiceofthehegemonnotcoercion)to analyzea liberalproblem, ofcollectivegoodsin theinternational system. GilpinandKrasnersuggestedthata similaranalysiscouldbe appliedto multinationalcorporations andtrade.48 Notonlyweretheontology andcausalityfullyrealist, 46. See Ruggie1975,570; andKeohaneandNye 1977,19. 47. Kindleberger 1973. 48. See Gilpin1975;andKrasner1976.

The StudyofWorldPolitics 661 butso was thefundamental problem. Stateswereinterested inmaximizing theirown interests. Andthepromotion oftheseinterests couldinvolverelativegainsanddistributionalconflicts. In theiranalyses,thedistribution ofpowerwas thekeyexplanatoryvariable,accounting fortherulesgoverning multinational corporations, andfor tradeopennessor closure.All stateshad a few simplegoals thattheysoughtto in theinternational promote economicsystem-economic utility, growth, socialstability, andpoliticalleverage.Becauseitprovided economicutility withlimitedsocial instability and enhancedthepoliticalleverageof thedominant power,fora hegemonicstatean opensystemwas themostattractive wayto securetheseobjectives. The firmsof thehegemonicpowerwouldmostbenefit froman opensystem,since theywerethemostcompetitive andhadtheeasiestaccesstocapital. Hegemonicstability theory-or,moregenerally, arguments abouttherelationship betweenthedistribution of powerand thecharacteristics of international economic behavior-operationalized theexplanatory variable,statepower,in termsof some overallmeasureof size (such as GNP or shareof worldtrade)and thedependent economicbehavior, interms ofopennessintheworldeconomy variable,intemational as indicated bothbyrulesandthepattern ofexchange.Thesesystemic theories could ofinternational be provenwrong:eitherthepattern economicbehaviorchangedwith thedistribution of poweramongstatesor it did not.The operationalization of variof causal relationships, and falsifiability ables,specification made systemicpower researchprogram. Its findings couldbe elaboratedand systematitheorya fruitful callycriticized. However,theveryclarityofhegemonicstability theory anditsease ofoperationalizationmadeitvulnerable topersuasivecritiques. Timothy McKeownshowedthat withthetheory, GreatBritain'sbehaviorin thenineteenth was inconsistent century andArthur Steinarguedthatthetradeliberalization measuresofthatcentury didnot thatpermitted discrimiemergefromhegemonic policybutfromasymmetric bargains nationagainstthehegemon.49 David Lake and DuncanSnidaldemonstrated thata of powerwas nottheonlyone thatwas consistent withan hegemonicdistribution open economicsystem.Reasoningfromontologicaland epistemological premises that thatwereentirely consistent withhegemonicstability theory, theydemonstrated an a smallnumber ofleadingstateswouldhavetheinterest andcapability to support of a realistapEven thoughsome arresting open system.50 reconceptualizations tradewerelaterputforward,51 proachtointernational bythemiddleofthe1980sthe of thefieldwas thathegemonicstability generalassumption theoryhad been seriouslyundermined. At theendofthe1970s,theserefutations tohegemonicstability theory layin the WhenRonaldReaganwonthepresidency, realistanalysisstillheldprideof future. beforea place in theUnitedStates;it was stillthetheorythathad to be refuted itsdominant Realismmaintained intellectual convincing challengecouldbe offered. todescribeacthatappearedmoreaccurately positiondespitealternative arguments 49. See McKeown1983;andStein1984. 50. See Lake 1984;andSnidal1985b.See also,on tradepolicy,Milner1987,1988;andOye 1992. 51. Gowa 1989,1994.

662 International Organization tors,and despitethefactthatitsempiricalvalidation had alwaysbeenproblematic. Realismcontinued tobe primus interpares becauseliberalism didnotofferan alternativeresearch program thatspecified causalityandoperationalized variablesclearly The renewalof theCold WaraftertheSovietinvasionof enoughto be falsifiable. Afghanistan attheendof 1979seemedtoreinforce realism'sintellectual triumph. Neoliberal Institutionalism

The development ofneoliberalinstitutionalism poseda seriouschallengeforrealist in 1982.In hisintroduction analysis.A specialissueof10 laid thefoundations Stea definition phenKrasnerpresented of regimes,developedby thegroupof authors forthisissue:"setsofimplicit orexplicit writing principles, norms, rules,anddecisionaroundwhichactors'expectations makingprocedures convergein a givenarea of international relations."52 This agreeddefinition was ambiguous;butit identified as advoregimesas socialinstitutions andavoideddebilitating definitional struggles, catesof theregimesresearchprogram soughtto showthattheirworkcouldilluminatesubstantive issuesofinternational relations. In hisarticleinthatvolume,andmorecomprehensively twoyearslaterina book, toexplaintheexistenceofinternaRobertKeohanedevelopeda rationalist argument tionalinstitutions.53 Drawingan analogytoproblems ofmarket failureineconomics, he arguedthathightransaction costsandasymmetrical uncertainty couldlead,under conditions suchas thosemodeledbyPrisoners' Dilemma(PD) games,tosuboptimal to actors(notby enforcing outcomes.Chieflyby providing information rulesin a centralized institutions couldenablestatesto achievetheirownobjectives manner), moreefficiently. Institutions wouldalterstatestrategies by changingthecostsof institutionalization could thuspromotecooperation.Keohaneargued alternatives; thatinstitutions mattered becausetheycouldprovideinformation, monitor complifacilitate issue linkages,definecheating, and offersalient ance,increaseiterations, solutions.Keohanedidnotdenytheimportance ofpower,butwithintheconstraints statescouldreapgains imposedby theabsenceof hierarchical globalgovernance, fromcooperation institutions. bydesigning appropriate forthislineofargument camefromnewworkineconomics Theinitialinspiration andfromtherenewedattention beingpaid to PD games.Economistshad begunto of institutions.54 RobertAxelrodsuggestedthatPD could recognizetheimportance be resolvedifthepayoffmatrixwerenotskewedtoo muchin favorof thesucker's ifthecostsofmonitoring ifgameswereiterated andindefinitely, payoff, frequently others'behaviorandofretaliating weresufficiently low,andifactorsdidnotdiscount thefuture at too higha rate.55 Institutions could,it was argued,affectthevaluesof theseparameters, forinstance, bynestingparticular gamesin durablerules,provid52. Krasner1983b,1. 53. See Keohane1982and 1984. 54. See Olson 1965;Akerlof1970;North1981and 1990;andWilliamson1975. 55. See Axelrod1981, 1984; and Oye 1986. See also Rapoportand Chamnah1965. For a further development ofthislineofwork,see Axelrod1997;andCederman1997.

The StudyofWorldPolitics 663 inginformation aboutotherstates'activities, andfurnishing standards forevaluating whether cheatingwas takingplace.In the1990sU.S. andEuropeanscholarsdevelopeda numberofdifferent thoughcomplementary approachesto analyzinginternationalregimes.56 Wheretheneoliberalinstitutionalism researchprogram differed withrealistargumentswas noton itsassumptions aboutactors,butrather on thenatureoftheexemplaryproblemintheinternational system:werestatesprimarily concerned withmarketfailureorwithrelative gainsanddistributional conflicts? Couldissuesbe resolved through thevoluntary acceptanceofinstitutions thatleftall actorsbetter off,orwould coercionand powerbe moreimportant fordetermining outcomes?Krasnersugconflicts thanmarket failureorrelativegainsarethe gestedthatdistributional rather centralconcernforstatesin theinternational system.The issueis notjust reaching theParetofrontier, butthepointon thefrontier thatis chosen;an issuethatcan be resolvedonlythrough bargaining andpower,notjustoptimalinstitutional design.57 JosephGriecoarguedthatstateswere,infact,concerned withrelativegainsevenin theEuropeanCommunity, whichseemedto be designedto enhanceabsolutewellRobertPowellclarified thisrelative thatevenwithin being.58 gainsdiscussion, arguing iftheycompromised a state'sfuture a realistlogicrelativegainsonlymattered ability A numberof important to secureabsolutebenefits.59 studiesin theearly1990s exploredtheconnections betweenpowerandpotential gainsfrompoliticalexchange.60 Neoliberalinstitutionalism has offered a set of heuristically powerfuldeductive be mademoreprecise.Indeed,sucharguments can arguments thatcouldeventually be formalized as HelenMilnerarguesin herarticlein thisissue. usinggametheory, Hypotheses generated byneoliberalinstitutionalism wereappliedto a widerangeof suchas bargaining betweentheUnitedKingdomandtheother empiricalproblems, members oftheEuropeanCommunity overtheFalklandsortheevolution ofinternationalregimesfordebtrescheduling.61 The appealofneoliberalinstitutionalism was enhancedby itsaffinity withthereigning kingof thesocial sciencesin theUnited States-economics.

DomesticPoliticsand IPE andinternational FromtheoutsetIPE blurredtheboundaries betweencomparative politics.Even analystswho tookthestate-as-actor approachdid so explicitlyfor theories-rather thanon phenomconvenience-toenablethemtodevelopcoherent 1993. 56. See, forexample,Young1989,1994,1997;Lipson1991;andRittberger 57. Krasner1991. 58. Grieco1988,1990. in Baldwin1993.See also Keohane1993.The relativegainsdiscussionis 59. Powell 1991,reprinted ofneorealism "isms"-the assumptions thatbeganatthelevelofcompeting an exampleofa controversy research andeventually yielded generated someempirical respectively-that andneoliberal institutionalism, to an analytical solution. 60. See Stein1990;Martin1992a;andOye 1992. 61. See Martin1992c;andAggarwal1996.

664 International Organization enologicalgrounds.In fact,realismand liberalismas generalresearchorientations forrealism,andpluhad specificdomesticandsystemic researchprograms: statism of for liberalism. The actorswere ralismand varioustheories interest aggregation at the domestic and but the causal voluninternational mechanisms, different levels, as remained the same. Usingclass taryexchange opposedto powerand coercion, anddomesticpolitical an integrated viewofinternational analysis,Marxismoffered economy.Analystsof theinternational politicaleconomycontinuedto emphasize in domesticpoliticsaffected how variations foreignpolicyand to suggestwaysin could affect domestic and interwhichtheinternational system politicalstructures ests.WebeginthissectionbydiscussingwhathappenedtoMarxism;we thenturnto statist ofdomesticstructure arguments andto a variety approaches. Marxism

ForMarxists, theorganization ofcapitalismdetermined politicalandeconomicoutlevels.Marxismoffered a structural or comesatboththedomesticandinternational rather an integrated picture institutional thanan actor-oriented argument, providing ofbothdomesticandinternational politics. instrumenArguing thatthestatewas simplythehandmaiden ofleadingcapitalists, talMarxismoffered clearcausal statements thatprovedto be empirically problematic.Manyof themajorpublicpolicyinitiatives of thetwentieth suchas century, social security and therecognition of laborunions,had been opposedby leading thisproblem andproposedstruccapitalists. Sophisticated Marxistanalysts recognized turalMarxismas a morepersuasive framework withMarx's thatwas moreconsistent ownformulations. Structural Marxistsarguedthatcapitaliststateswouldact in the autonomous interest of preserving capitalismas a whole.The statewas relatively inanempirifromitsowneconomyandsociety. thisorientation However,specifying Ifthestatewererelatively autonomous, exactlyhow callytractable waywas difficult. autonomous coulditbe? Whatpolicywouldbe inconsistent withthisperspective? Bothpoliciesthatreflected thepreferences of thecapitalistclass and thosethatdid notcouldbe accountedforbystructural Marxistanalyses. Scholarsinfluenced by Marxismalso emphasizedsocial forcesand production influential research as in theworkof RobertW. Cox.62One particularly relations, In thelate basedon a genericMarxistorientation was dependency theory. program anArgentinian attheUN EconomicComeconomist 1940sRaoulPrebisch, working missionforLatinAmerica,arguedthattheworldeconomyenmeshedpoorercountriesexporting rawmaterials in relationships ofunequalexchange.Prebisch'sargufrom weredevelopedby social scientists ments,and olderones aboutimperialism, developingareas,especiallyLatinAmerica,as well as fromNorthAmericaand ofthestatesofthesouth thatexplainedthepoverty Europe,intoa researchprogram in termsof theirpositionin theworldeconomy.63 The worldsystemsresearchproand his colleaguespresenteda similar gramelaboratedby ImmanuelWallerstein 62. R. Cox 1981,1987. 63. See Prebisch1959;Cohen1973;andCardosoandFalleto1979.

The StudyofWorldPolitics 665 In a 1978 specialissueofIO JamesCapoanalysisfora muchlongertimeperiod.64 rasoemphasizedthedistinction anddependency. In contrast to betweendependence theory offered a disguised theintemationalism of traditional Marxism,dependency in whichtheroleofthestateloomedlarge,especiallyforweak formofnationalism and weak,but polities.It was notjust thatthepolitiesof thesouthweredependent thatundermined theirautonomy and thattheywerein a relationship ofdependency ofexploitation includedbothspecificecoexploitedtheirwealth.Thesemechanisms ofdeveloping statesbymorepowernomicarrangements andthegeneralpenetration fulandbetter statesofthenorth.65 organizedcapitalist and as As theexperienceof thedevelopingworldbecamemoredifferentiated thanothers, somestatesandworldregionsdidbetter dependency theorists suggested thattherecouldbe a pattern ofdependent Some groupswithindeveldevelopment. wouldallythemselves with opingstates,suchas largercapitalistsand themilitary, miliactorsfromthenorth, suchas multinational corporations andnorthern powerful taries.Statesin thesouthcouldprosper, buttheiroptionswouldalwaysbe truncated intotheworldcapitalist system.66 bythewayinwhichtheywereinserted theoryencountered seriouscriticism Fromtheearly1980s onward,dependency andanomalies.It was criticized forfailingtoclearlyspelloutcausalregularities that it had greatdifficulty Furthermore, could be empirically supportedor falsified.67 in whatwas knownas theThirdWorld,espetheunevenratesofgrowth explaining ofEastAsiancountries. economicdevelopment ofa number Crossciallythestunning inendowments, andpoliciesseemedtoprovidemore nationalvariations institutions, promising explanations. Politicaleconomyworkon developingcountriesincreasof economicsandcomparative thanon politicsrather inglyreliedon a combination FerOne oftheleadingexponents ofdependent development, dependency theory.68 ofBrazil. president nandoHenriqueCardoso,evenbecametheliberalreformist theprofound The collapseoftheSovietempireand,moreimportantly, corruption thatdrewon revealedby its demisedealta heavyblow to theresearchprograms orientation. orientaMarxismfortheirtheoretical However,Marxismas a theoretical tiondidnotdisappear. Marxistshavegrappledwithsomeofthesameissuesthathave ofMarxistscholarship hasbuilton strand engagedliberalscholars.A neo-Gramscian and sourcesof legitimating a tradition of analysisthatemphasizestheimportance moreconsistent withconstructivThisscholarship is therefore ideasandideologies.69 of sectionofthisarticle)thatstressestheimportance istwork(reviewedinthefourth versionsofMarxism. ideasandculturethanarematerialist ofMarxistanalysessimplyin termsof It wouldbe a mistaketojudgetheinsights Marxistanalysisposes probingquestionsabout anyone specificresearchprogram. 64. See Wallerstein 1974,1979,1991;andArrighi1994. 65. Caporaso1978a. 66. Evans 1979. 67. Packenham1992. 68. See, forexample,Haggard1990;Wade 1990;Evans 1995;andHaggardandKaufman1992,1995. 69. Fora collectionofessayson Gramsciandinternational relations, see Gill 1993,whichreprints the seminalessaybyCox 1983.

666 International Organization therelationship a conceptualapparatus betweenpowerandwealth.It offers thatcan be appliedbothtointemational anddomesticdevelopments andtotheirconnections. Andit addressesenduiingmoralconcernsconcerning equalityandjustice.Specific researchprograms generated by a genericMarxistorientation can,likespecificliberaland realistresearchprograms, be calledintoquestionby problemsof variable andempiricalevidence.Thisdoes notmean,however, specification thatMarxismas a generalorientation willnecessarily be discarded. Rather, itis precisely theheuristic richnessofthemajorgeneralresearchorientations thatallowsthemto be reformulatedto addresschangedempiricalandpoliticalcontexts. The fateofa generaltheoreticalorientation doesnotdependonthesuccessofanyonespecific research program. Statism: ReactionAgainstLiberalismandMarxism Statismis a generaltheoretical orientation thathas generatedseveralspecificresearchprograms, all ofwhichasserttheautonomy of stateinstitutions. Statismthus tothesocietally standsincontrast oriented domesticpoliticalperspectives thatdominatedmuchofliberalandMarxistpoliticalanalysisinthe1970s.Statismgavegreater thestability attention to stateinstitutions, especiallythosechargedwithmaintaining ofthepolityas a whole.The statecouldbe conceivedofas an actor, andwell-being notsimplyan arenain whichconflicting to securetheir societalinterests struggled prefeiTed policyobjectives.Statescould be strongor weak,relativeto theirown societies.70 Statistarguments didnothavea particularly ofthe sophisticated conceptualization Statescouldbe strongin someissueareasbut relationship amongstateinstitutions. weakinothers.Specifying thetrade-off amongdifferent issueareaswas problematic. Statismhaddifficulty thenuancesofstate-society Itdetachedthe relations. capturing statenotonlyfromparticular grouppressures butalso fromthelargerpolityinwhich itwas embedded.71 These empiricaldifficulties led notto thedisappearance of statismas a general theoretical orientation buttothereformulation ofspecificresearch Instead programs. of a narrowfocuson the state,whichwas itselfa reactionto an overlysocietal of additionalworkanalyzedinteractions betweendifferent perspective, components thepolity.In 1988 JohnIkenbeiTy, David Lake, and MichaelMastandunoediteda theeffect on U.S. foreign economic specialissueofIO inwhichauthors investigated of interest, theabilityof stateleadersto mobilize policyof different configurations societalsupport, andtheconsequencesofideas,as wellas thediscretionary powerof relationsemployinga theexecutive.72 More recently, discussionsof state-society

and Skocpol 1985; and 70. Krasner1977, 1978. See also Katzenstein1976; Evans,Rueschemeyer, Evans 1995. 71. In a prescient articlethatdrewondiscoursetheory, BruceAndrewssuggested thatthestatecouldbe seenas a rule-governed social actor.This lineof analysislay dormant formorethana decade.Andrews 1975. 72. Ikenberry, Lake,andMastanduno1988.

The StudyofWorldPolitics 667 rationalchoiceperspective on institutions, especiallytheimportance ofcommitment, have shownthatsomeof thefactorsthatstatismidentified as weaknesseswere,in fact,sourcesof strength. Democraticstatesareoftenable to extract moreresources fromtheirown societiesthanare autocratic states,preciselybecause membersof civilsocietybelievethatwhatwouldhavebeenconsidered a weakstatewillkeepits promises.In thisanalytic jujitsu,thenotionthatstateswerestrong becausetheywere independent oftheirsocietiesis turned on itshead.Statescan secureresourcesfrom theirown societiesonlyiftheyareconstrained.73 Reformulated versionsof statism focuson stateguidanceoftheeconomy, thelinksbetweenpoliticalpartiesandstate howstateinstitutions bureaucracies, relateto socialmovements, andtheroleoflaw. As a generaltheoretical orientation, statism has beenrefurbished butnotabandoned as someofitsresearchprograms encountered empirical anomalies. Domestic Structuresand TheirRelation to theInternationalSystem

between ofcomparative on theconnections Students politicsfocusedtheirattention domesticstructures andinternational whichwerebracketed relations, byneoliberalinformed taxism and realism.Katzenstein, forexample,reliedon a historically ofstateandsocietyindifferent politionomythatemphasized different constellations cal settings. and Moore,Katzenstein argued DrawinginsightsfromGerschenkron thatearlyindustrializers likeBritaindiffered fromlateindustrializers systematically of thedominantsocial coalitionand in thedegreeof like Japanin thecharacter In sharpcontrast ofstateandsociety.74 tothestatist centralization anddifferentiation thatviewedstatesas actors,theanalysisofdomesticstructures privileged literature Different social coalitionsdefinethecontentof policy. state-society relationships. indomestic ontheformulaAnddifferences policynetworks havediscernible effects tionand implementation of foreigneconomicpoliciesin different economicissues to thepoliticsof suchas moneyand trade.JohnZysmanextendedthisperspective andfinance.75 industry scholarsanalyzedcontrasts In variouspolicydomains,bothforeign anddomestic, welfarestatecapibrandofAnglo-American betweentheliberalmarket capitalism, and developmental statecapitalismin Japanand talismon theEuropeancontinent, EastAsia. EllenComissoandLauraTysonediteda specialissueofIO on comparativesocialism.76 Subsequently, comparative politicaleconomyspreadto encompass fromsocialismin thesuccessorstatesof theSoviet LatinAmerica,thetransitions in the Unionand east-central Europe,and even theLeninistcapitalismemerging People's Republicof China.77This body of researchspecified,in contextualand

73. See NorthandWeingast1989;Fearon1994b;andSchultz1996. 74. Katzenstein 1978. 75. Zysman1977,1983. 76. ComissoandTyson1986. 77. The literature is too largeto be listedhere.Since theearly1980s,CornellStudiesin Political volumeson thissubject. Econony,forexample,comprisemorethanseventy

668 International Organization or corporateactorsto choose historicaldetail,incentivesforstates,governments, specificstrategies. thattheinternational state PeterGourevitch emphasizedthepervasiveinfluences and systemandtheintemational politicaleconomycan haveon domesticstructures lead,analystspointedto thepolicypreferences ofgroups.78 FollowingGourevitch's intheworldeconomycouldaffect twodifferent different waysinwhichenmeshment structures of polipolities:first, involvement couldinfluence thebasic institutional ties,includingtheirgoverning norms;second,it could affectthecapabilitiesand of different interest included strategic opportunities groups.This researchprogram oftheinternational corporatism of analysesoftheeffects systemon thedemocratic thesmallEuropeanwelfarestates,on societalgroupsor economicsectors,and on coalitions, institutions, ideologies,andeconomicstructures.79 to The conceptof two-levelgameselaboratedby RobertPutnamwas one effort andconstraints, domesticstructures, opportunities systematically integrate systemic andforeign economicpolicy.80 mustsatisfy bothother Anyintemational agreements statesand domesticconstituencies. The bargaining powerof a statecould be enwouldonlyaccept hancedifitsrulerscan demonstrate thattheirdomesticsupporters a narrow rangeofoutcomes.In morerecentwork,AndrewMoravcsikhaselaborated a relatedperspective on domestic-international interactions thatemphasizeshowsocietalinterests shapethepoliciesofstates.8' One difficulty encountered withthislineofresearchwas theabsenceofa general In a boldandimaginaandsystematic forclassifying domesticstructures. taxonomy tivebookRonaldRogowskioffered one answerto thistaxonomic issue.He applied theelegantreasoningoftheStolper-Samuelson modelofinternational tradeto show how,in general,tradepoliciesand practiceswouldaffectsocial cleavages.Exportoriented thefactorsofproduction withwhicha country was goodsusedintensively Iftradebecamemoreopen,theabundant factorwouldbenwell-endowed. relatively In efit.If itbecamemoreclosed,therelatively scarcefactorwouldbe advantaged. by,andwould generala moreopeninternational trading systemwouldbe supported therelatively abundantfactorin different states.These changesin the strengthen, suchas laborand theownership of landand domesticpositionof different factors, Friedenemcapital,couldinfluence policy.82 Usingbasicallythesamelogic,Jeffry such as labor of factorspecificity; factorsof production phasizedtheimportance werenothomogenousbutratherwereassociatedwithspecificeconomicsectors. andbe affected macroeconomic Moreover, policy,notjusttradepolicy,couldaffect bya country's opennesstotheworldeconomy.83

78. Gourevitch 1978. 1984,1985; Snyder1991; Kupchan1994; andSim79. See Cameron1978; Kurth1979; Katzenstein mons1994. 80. See Putnam1988;andEvans,Jacobson, andPutnam1993. 81. Moravcsik1997. 82. Rogowski1989. 83. Frieden1991.

The StudyofWorldPolitics 669 Rogowski,Frieden,andotherswhofollowedtheirlead werefullyawarethattheir perspective offered no easywayofincorporating variations in institutional arrangethatHelenMilnerandDavid Lake,amongothers, havesought ments,a shortcoming to correct.84 But theyprovideda parsimonious, generalexplanation thatlinksthe worldeconomytodomesticinterests andpolicies.Theiranalyticframework requires no specificknowledgeaboutcontext. Itcanbe appliedto anypoliticalsystem. In the framework's sparseformulation thepreferences of actorscan be read directly off materialstructures, bypassingtheanalysisofpoliticalprocessesandideationalpheitrequiredinformation nomena.Although aboutwhichfactors wererelatively abundant,itdidnotrequireanyinstitutional knowledgeaboutspecificpolities.However, suchgeneralstructural theories encounter numerous empirical anomalies,as Rogowski'shistorical discussionrevealed. and DomesticPolitics Globalization levelsoftransboundary andtheirassociatedeffects, Increasing movements whathas come to be termedglobalization, encouragea moreintimateanalyticrelationship betweeninternational flowsare and domesticpolitics.Highlevelsof cross-border notan unprecedented development. Labormigration reacheditshighestlevelsinthe nineteenth century. By somemeasuresinternational capitalmarkets weremoreintegratedat theendof thenineteenth century thantheyare now,sincefinancial flows felldramatically withthetwo worldwarsand theGreatDepression.Tradeflows have followeda similarpattem,increasingsharplyduringthenineteenth century, thenfallingin thefirst andreachingunprecedented levelsfor partof thetwentieth, somecountries, mostnotablytheUnitedStates,inthelasttwodecadesofthetwentiethcentury. has dramatically reducedthecostsofcommunications. SoTechnology cial movements havebeenmobilizedin specificlocalesforglobalissuessuchas the Illicitactivities humanrights, andfeminism. crime environment, including organized and tradein drugshave becomemoresalient.All of thesedevelopments, lumped underthelabelofglobalization, haveaffected bothnationalpolitiesandtheinternationalsystem.85 totheincreasing Globalization drawsourattention politicalsalienceoftransboundof politiesin theinternational aryactivities.Is thegrowingenmeshment political andpolicies,groupsandindividuals morealike,or are economymakinginstitutions mostof theirdifferences? Is globalizationaltering"inter"national theyretaining or is it fundamentally themto "trans"national relationsmarginally, transforming relations? Thosewhoemphasizehowglobalization is remaking worldpoliticsstress howpolicypreferences andpoliticalcoalitionsat homechangeas a resultofchangs conGarrett' inginternational pressures. Conversely, analyses,including Geoffrey tothisissue,thatemphasizethepersistence ofdistinctive nationalpractices tribution 84. See FriedenandRogowski1996,42-47; Milner1997;andLake 1996andforthcoming. 85. See KeohaneandMilner1996,10-14;FeldsteinandHorioka1980;Oman1994;Shaw 1994;Jones 1995;Rodrik1997;andObstfeldandTaylor1997.

670 International Organization have shownhow domesticinstitutions blockpricesignals,freezeexistingpolitical coalitionsandpolicies,andshapethenationalresponsetoglobalchange.86 Bothof thesearguments focuson domesticpoliticalinstitutions, firms, interests groups,and economicsectorsas unitsof analyses.An altemative conceptualization on therelationships focusesnoton theunitsthemselves, butrather amongthemand makesproblematic thenatureoftheseunitsinthefirst place.Households,communiin a gloreconstitute themselves ties,regions,andsocialmovements, amongothers, bal setting. This conceptualization pointsto processesof "glocalization"thatare theidentities, and strategies of actorsthrough a combination transforming interests, ofglobalandlocal processesandarethusaddingnewpoliticalactorsandprocesses toan increasingly globalpolitics.87 of a phenomenon howeverconceived,is a reflection thatscholars Globalization, associatedwithIPE have recognizedsincetheinceptionof thefield:international anddomesticpoliticscannotbe isolatedfromeachother.Neorealists andneoliberals did notincorporate domesticpoliticsintotheirtheoretical butthey formulations, neverdenieditsimportance. Statesdidnotall respondinthesamewaytotheopportunitiesand constraints presented by theinternational system.Studiesof domestic of whatneorealistsand neoliberalstookfor politicsenhancedour understanding intheirtheories inthe1980s:"statepreferences." analygranted Domesticstructure in twopossiblycomplementary sis suggestedthatpreferences couldbe understood of the normsor as theaggregations ways: eitheras theresultof institutionalized preferences ofindividuals, firms, andgroups.Historical-institutional researchon the ofdomesticstructures andtheintemational reciprocaleffects politicaleconomyhas beencomplemented bya decidedlyeconomicandmaterialist variantthatpaysvirtutotheroleofideas,norms, andinstitutions. allyno attention Different interpretations oftheprocessofglobalization reflect thisdifference in orientation.

A Post-Cold War Opening:Rationalism and SociologyRevisited EvenduringtheCold War,therewas substantial withreigning realist dissatisfaction andliberalapproachestointernational relations, especiallyoutsidetheUnitedStates and in therelatedfieldof comparative politics.The end of the Cold War was a inseveralways.Itraisednewissuesfortheongoingrationalist catalyst debate,which pittedrealists,who stressedtheroleof coercion,againstliberals,who emphasized contractual The endoftheCold Waralso openedup spaceforcultural relationships. to as "constructivist," and sociologicalperspectives, oftenrefeffed thathad been neglectedbybothrealistsand liberals.Andthediscussionsthatensuedhighlighted of rationalism betweenpossiblepointsof complementarity conceptualdifferences andconstructivism. 86. See KeohaneandMilner1996; Garrett andLange 1995 (reprinted in KeohaneandMilner1996); PaulyandReich1997;andSamuels1994. 87. See Rosenau1990,1997;andAppadurai1996.

The StudyofWorldPolitics 671 Rationalism: Realismand Liberalism AftertheCold War Realismhas beennotonlya salientgeneraltheoretical orientation butalso partofa moreenduringnormative discourse,like liberalismand constructivism, aboutthe mostappropriate way to securepeace, stability, andjusticein humansociety.Its self-conscious intellectual pedigreeis longandimpressive. It willnotdisappear. However,recentdevelopments in worldpoliticsandwithinspecificresearchprogramshaveconfronted realismwithmuchgreater challengesthanithas facedsince ofIO. Forrealism,powerandconflict thefounding areinherent aspectsof internationalpolitics.Theinterests ofstateswilldiffer. Forceandcoercionarealwaysavailable options.The astonishingly peacefulendoftheCold Warandthecollapseofthe SovietUnionare notwhata realistwouldhave expected.88 Realismhas notbeen fortheendoftheCold Waris thatSoviet silent,ofcourse.The simplest explanation powerdeclined;theSovietUnionwas a challenger thatcouldno longerchallenge. Predictions aboutrelativechangesin statecapability haverarelybeenincorporated intorealistresearchprograms, and realismdid notpredictthisdecline.89 Realists, especiallyWaltz,haveemphasizedtheimportance ofnuclearweaponsinaltering the likelihood ofwar.Withsecuresecond-strike itis moreevidentnowthanat capability, anyothertimein humanhistory thata conflict amongthemajorpowerswouldreduce thewell-beingof all states.At least some observersview thissituationas a in the changein thenatureof theinternational systemitself,notjust an alteration ofindividual Froma realistperspective, ina nonnuclear world characteristics states.90 itwouldhavebeenmuchriskier fortheSovietUniontoabandonitsempireineastern Europeandforanyleaderto breakup theSovietUnionitself,actsthatwouldhave leftevenRussia'scoreterritory toinvasion. morevulnerable inthe1980sanalystsworking withina realistframework werearguNevertheless, ingthatbipolarity wouldcontinue. Andtheyassumedthatneither pole coulddisapWhentheSovietUniondid collapse,realistswereskepticalabout pearpeacefully. therobustness of international institutions, especiallythoserelatedto international in theinternasuchas NATO,andtheprospects forcontinued security, cooperation outmuchbetter thanrealists tionaleconomy. Overthelastdecadethingshaveturned hadanyrighttoexpect.91 The challengesto realismpresented by thepeacefulend of theCold Warwere salienceofneorealism as a specificresearchprogram. aggravated bytheintellectual was to forceanalyststo makea fundamental Waltz'smostimportant contribution distinction betweenwhathe calledstructural andreductionist Waltzwas, arguments. ofdomesticpoliticalfactors, ofcourse,awareoftheimportance buthe insistedthat neorealism was concernedonlywiththedistribution of poweramongstates.States 88. See Lebow 1994;andStein1994. 89. Wohlforth 1994. 90. See Waltz1990;andSnyder1996. 91. See Mearsheimer 1990;andWaltz1993.See also Mearsheimer 1994andthesubsequent exchange inInternational Security.

672 International Organization wereinterested in security, notin expandingtheirpower.Treating all statesas ifthe international relations goalswerethesamewas a departure fromearlierrealistdiscussion,muchofwhichhadbeenbasedon theassertion thatitwas necessaryto distinThe conflict guishbetweenrevisionist and statusquo states.92 betweentheSoviet withbotha strictly neorealist Unionand theUnitedStateswas consistent analysis that (thepoles in a bipolarworldwouldbe in conflict)and withrealistarguments positedtheimportance ofexogenously givenvariations instateobjectives(a revisionistSovietUnionwouldbe inconflict witha statusquo UnitedStates). Withtheend of theCold War,neorealismoffered less purchaseon international invariations conflict thatappearedtobe embeddedprimarily inthegoalsofstatesor, inthecase ofethnicconflict, substate actors.Manyscholarsarguedthatthematerialistassumptions ofneorealist itfromexplaining analysisprevented therapidchanges observedin core nationalsecurity issues.Accordingto theseauthors,historically constructed norms,ideas,and discoursesneededto be analyzedbeforeone could make sense of patterns and changein worldpolitics.93 of stability Some realists thatall responded tothischallengebyrejecting thesparseassumption ofneorealism stateswouldseeksecurity andembracing insteadtherecognition thatstateobjectives notsystemic couldvarybecauseofdomestic, factors. States'goalscouldbe aggresTheextent siveorpassive,revolutionary ortolerant. to orstatusquo,ethnonationalist whichthepresenceof revisionist stateswouldresultin confrontations, especially of materialpowerin theinternational war,would stilldependon thedistribution a manisystem.The rulersof greedystateswouldnotcommitsuicideby attacking festlystronger enemy.Butknowledgeof thedistribution ofpoweralonewouldnot allowanalyststounderstand ofinternational conflict patterns andcooperation.94 The burgeoning of ethnicconflictin thelast decade has presented thekindsof thatrealistperspectives weredesignedtoanalyze,although notnecessarily problems withstatesas themostsalientactors.Ethnicconflicts havearisenbetweenstatesand inenvironments whereauthority amonggroupswithinstates,albeitgroupsoperating structures have erodedor disappearedand wheresecurity dilemmasoperateat the of variations in substatelevel.95However,thisreturn to a focuson theimportance stateobjectives(ofwhichethnicconflict is onlyone example),as a resultoffactors realist of powerin thesystemas a whole,confronts exogenousto thedistribution shouldexist-a chalanalysiswiththechallengeofexplaining whysuchvariations a moresystematic of realismwith lengethatcan onlybe metthrough integration domesticpolitics. In international politicaleconomythespecificresearchprogramsthatreflecta ofcoercionthat orientation havefocusedon thepossibility generalrealisttheoretical can leave someactorsworseoff,on theconsequencesof bargaining asymmetries, and on theproblemsof commitment in an anarchicenvironment. Some of these of constructivist More theinfluence or liberalperspectives. specificprojectsreflect 92. 93. 94. 95.

See Morgenthau 1948,chap.2, 3; andSchweller1996. See Katzenstein 1996b;andLapidandKratochwil 1996. Forexample,Frankel1996. See FearonandLaitin1996;andPosen 1993a.

The StudyofWorldPolitics 673 thatweakeractorshave powerful statesmaybe in a positiontoaltertheconceptions oftheirownself-interest, especiallywheneconomicandmilitary powerhas delegitimatedideologicalconvictions in weakerordefeatedpolities.The UnitedStates,for instance,pressedfora particular visionof how international societyshouldbe orthisprojectaftertheendof deredafterWorldWarII andrenewedandreinvigorated but theCold War.The goal was notsimplytopromote a particular setofobjectives, toalterhowothersocietiesconceivedoftheirowngoals.ThisemphasisonwhatNye hascalledsoftpowerengagesbothconventional realistconcernsaboutrelativecapabilitiesandconstructivism's focuson beliefsandidentity.96 in some Powerful statescan also alterstrategic optionsin waysthatskewpayoffs cases by unilaterally changingtheirown policies.97 Theymaybe able to establish institutional arrangements thatprecludecertaininitiatives or facilitate thestrategic to engagein use ofinformation.98 Commitment problemsmaymakestatesreluctant intheshortterm,especiallyif thatprovidethemwithabsolutebenefits arrangements theymaybe relatively worseoffin the long term.If a state'sfuturebargaining leveragewouldbe compromised, itmightrejectimmediate gainsbecauseofitsanxiofagenda etyaboutpotential future losses.99 Arguments emphasizing theimportance andstrategic arebasedonthesamegame-theoretic setting, uncertainty, manipulation emformulations thathaveguidedmuchrecentworkfrommoreliberalperspectives phasizingthemutualbenefits of cooperation. In thepost-ColdWar worldrealist in stateobjectivesand to a more projectshavebecomemoresensitiveto variations setofrelationships betweenabsolutegainsanddistributional conflicts. complicated Neoliberalinstitutionalism thattheendoftheColdWarwould correctly anticipated notundermine suchinstitutions as NATO andtheEuropeanUnion,so it didnotgo an "agonizingreappraisal"suchas thatexperienced through by somerealists.Insuch as deed, institutionalists began to applytheirtheoryto securityinstitutions terms.100 alliancesandto interpret And,as post-ColdWarpoliticsin institutionalist thearticlesin thisissue by Lisa Martinand BethSimmonsand by Helen Milner fromotherfields,notably workdrewheavilyon scholarship indicate,institutionalist morerigorous. workis so U.S. politics,tobecometheoretically Sinceinstitutionalist thereis no needforus toreviewitindetailhere. welldiscussedinthesetwoarticles, as indicating Brevityshouldnotbe interpreted insignificance. years, Institutionalist thinking hasmadea bigimpactonIPE duringthelastfifteen a set of researchprogramsthathave illuminated relationships among stimulating But it was of less valuein understanding shifting interests, power,and institutions. hadforsevAdvocatesofdomesticstructure identity politicsafterwards. approaches relationsresearch, eraldecadescriticizedinternational includingneoliberalinstituor identities of theactorswhomit fortakingforgrantedthepreferences tionalism, forexample,to the no attention, studied.Neoliberalinstitutionalism paid virtually 96. See Nye 1990;andIkenberry andKupchan1990. 97. See ScottandLake 1989;andNau 1990. 98. Morrow1994b. 99. Fearon1995. 100. See Keohane,Nye,andHoffmann 1993;andWallander1998.

674 International Organization phenomenon ofnationalism. Anditcouldnotcapturethefactthatduringthe1980s increasedinterest in humanrightsand environmental issuesseemeddrivenlargely After1989,somerationalists bynormative concerns. begantothinkofideasas variables thataffected thesolutionsto games-forinstance, or byreducinguncertainty '0' Ideas couldbe incorporated focalpoints. framework providing intoaninstitutional howparticular becomeinstitutionalized byemphasizing conceptions and,therefore, Sinceitwas notweddedexclusively persistovertime.'02 to a materialist conception of structure, neoliberalism could engagesomeof theissuesof changingbeliefsor identities posedbyendoftheCold War. TheRevivalofSociologicaland CulturalPerspectives Sociologicalperspectives have alwaysbeenimportant forcomparative politicsand haveneverbeencompletely absentfrominternational studies.In Europe,wherethe boundary betweeninternational anddomesticpoliticswas neverparticularly salient, thesociologicalbentofscholarship differed fromprevailing Americanperspectives. This was true,forexample,of Scandinavianand Germanpeace research,which intheUnitedStatesandwas oftenregarded remained largelyunnoticed as politically suspectwhenread.The theoretical contributions of theFrenchSchool,represented byStanleyHoffmann's writings andthoseofRaymond Aron,especiallyinthe1950s and 1960s,andof theBritishSchool,clustering in the1970sand 1980saroundthe of HedleyBull and MartinWight,remaineduninterested in the debates writings betweenthegeneraltheoretical orientations thatdominated Americanscholarship, suchas realismand liberalism.103 These schoolsof thought wereat odds withthe emphasisin Americaninternational relationsscholarship on clearlystatedcausal propositions andtheirsystematic exploration inmethodologically rigorous ways. In theUnitedStatesRuggiepublisheda seriesofpapersdemonstrating thevalue of a sociologicalorientation. He arguedthatthepostwarinternational economicrewhathe termed gimereflected embeddedliberalism, identified bya sharedintersubthatopeninternational markets wouldbe tempered jectiveunderstanding bytheneed to maintainsocial stability. He criticized Waltz'stheoryforitslack of sociological withFriedrich contentand forfailingto explainsystemicchange.'04 And,together he pointedoutthatanalystshad failedto investigate thesharedunderKratochwil, thatled to theconvergence of actorexpectations on which,by someacstandings ofprinciples counts,regimestability depended.Theyarguedalso thatthetreatment andnormsas "independent" or "intervening" variables,linkingmaterialstructures to outcomes,was noteasilyaccommodated foundations withintheepistemological ofinstitutional andnormative Kratochwil andNicholasOnuf analysis.Subsequently, 101. See GoldsteinandKeohane1993b;andGarrett andWeingast1993. 102. Goldstein1993,1996.See also Hall 1989b;andSikkink1991,1993a,b. 103. Raymond Aron'smassivesociologicalstudyofwarandpeacedidnotplaya majorroleingraduate in theUnitedStates,andHedleyBull's majorbookon intemational training societyhadlittleimpactuntil abouta decadeafteritspublication. See Aron1966;Bull 1977;andHoffmann 1986,1987. 104. Ruggie1983a,b,1998.

The StudyofWorldPolitics 675 putforward a conception ofrulesinformed by,amongothers, philosophy, linguistics, 105And in severalpapersWendtsuggested and sociology. a social theory ofinternationalrelations thatengagedtheclaimsofneorealism headon.106 This sociologicalturnwas intellectually deeplyindebtedto fieldsof scholarship well beyondtheconfinesof IPE. Philosophy, structural linguistics, criticaltheory, geography, scienceand technology studies,postmodern politicaltheory, anthropolall had,in different ogy,mediastudies,andliterary criticism, amongothers, though relatedways,grappledwiththeprojectof modernity gone awryin thetwentieth Thereis a growing relations andin security century.'07 bodyofworkin intemational studiesbut,significantly, notyetinIPE thatis self-conscious inconducting empirical researchfroma constructivist perspective. Sociologicalworkfallsintothreebroad 108 clusters:conventional, critical, andpostmodern. Conventional constructivists insistthatsociologicalperspectives offera general thatcan rivalor complement theoretical orientation and specificresearchprograms In thisviewa fullunderstanding rationalism. ofpreferences requiresan analysisof areconstituted. thesocialprocessesbywhichnormsevolveandidentities Sincethey constitute structures emphasizehow ideationalor normative agentsand theirinterfromrationalists on questionsofonests,conventional constructivists differ sharply aremutually conconstructivists insistthatagentsandstructure tology.Furthermore, stitutive andthushopeto givesocial sciencea moredynamicconception ofchange no great ofsystemstructures. On issuesofepistemology andmethodology, however, differences divideconventional constructivists fromrationalists. is theresearchprogramof the Outstanding amongconstructivist contributions sociologistJohnMeyerand his colleagues.Informed by a cognitiveapproachto an astonishing worldculture, thisresearchprogram demonstrated degreeofsimilarto issuesas diverseas censuses,socialsecuityin formalnationalpractices relating in nationalsocioeconomicand rity,education,and sciencedespitegreatvariations For Meyerand his associates,thekeyto understanding ideologicalcharacteristics. thisstoryis thescriptof modernity, oftenpresented by international organizations, whichsuggested tonationalleaderswhatpoliciestheyoughttoadoptiftheywanted toappear,tothemselves tobe modernandprogressive.'09 andothers, a numSociologicallyinclinedscholarshaveinrecentyearsanalyzedempirically berof cases to bolsterthisresearchprogram in international relationsand security as acceptstudies.Slaveryandchildlabor,forexample,wereacceptedformillennia ablesocialpractices; inthecourseofbarelya century with theybecameincompatible

105. See Kratochwil andRuggie1986;Onuf1989;andKratochwil 1989,especiallychap.4. 106. Wendt1992andforthcoming. 107. For a brilliant discussionof theconstructivist perspective froma philosophicalstandpoint, see Searle1995. 108. We choosethethreecategorieshereforheuristic purposesand in fullawarenessof thefactthat considerabledifferences existwithineach of theseclusters,bordersbetweenclustersare porous,and scholarsmaychangepositionsin different publications. Ourdiscussionis influenced byPriceandReusSmit1998;andHopf1998. 109. See Meyer1980;Thomas,Meyer,Ramirez,andBoli 1987;Strang1991;andFinnemore 1996a,b.

676 International Organization civilizedsociety.110 International warmaypossiblybe relegated toa similarstatus.111 Halfa century ago it was normaland appropriate forJapaneseand Germanyoung mentovolunteer todie foremperor andfatherland. By the1990stheinstitutionalizationof identities and normsthathave markedJapaneseand Germanpoliticssince 1945 makesuchindividualchoicesand social practicesa rareexception.112In nationalsecurity studiesa growingnumberofmostlyyounger scholarsaddressedconventionaltopics-such as thespreadof weaponsof mass destruction, deterrence, armsraces,strategic or alliancepolitics-withunconventional culture, sociological and culturalapproaches.113And duringthelast two decades feminists have been successfully redefining themeaningof humanrightsto encompassgenderidentities.'14

Conventional constructivist studieshavefocusedbothon criticalhistorical juncturesfromwhichnew structural arrangements emergeand on interactions between existingstructures and agents.Ruggie'shistorical arguments aboutthereplacement offeudalism bythemodernstatesystemhavemadea majorimpacton international relations theory.115 Recentanalysesof how actorsand structures are reproduced in contemporary worldpoliticsincludestudieson Japanesenormsindicating thatdrasticchangesinJapan'ssecurity andstudieson changesin policiesarehighlyunlikely normsaboutweaponsindicating thatsomeclassesofweaponsmaybe seenas illegitimate,regardless oftheirstrategic usefulness.116 As a generaltheoretical orientation constructivist researchilluminates thesources forexample,has formulated ofbothconflict and cooperation. lain Johnston, a constructivist militant thatseeksto accountforChina'sconsistently argument security The balanceof materialcapabilitiesin theinternational strategy. systemchanged overthelastdecadesandcenturies; China'sparabellum culturedid greatly strategic a compellingexplanation of security not.Hence it is thelatterthatoffers policy.117 in hiswriting andrealistinsights Similarly, HenryNau has combinedconstructivist toaddresscentralelements inU.S. foreign policy.118 ofhumannature, criticalconstructivists In rejecting rationalist conceptions agree conon theissue of ontology. Like conventional withconventional constructivists in how actorsand systemsare constituted and costructivists, theyare interested issuesthatinclude,besidesnationevolve.Theirresearch focusesonidentity program andsexuality. Criticalconstructivists alism,subjectssuchas race,ethnicity, religion, ofsocialscientific basedonempirical research. also acceptthepossibility knowledge offormulating Theyare,however, deeplyskepticalofthepossibility generalcoveraboutappropriate researchmethodologies. Instituinglaws,andtheyarepluralistic 110. KeckandSikkink1998. 111. Mueller1989. 112. See Katzenstein 1996a;andBerger1996. in Katzenstein 1996b.See also Kier1997. 113. See thecontributions 114. See Tickner1992;andKeckandSikkink1998. 115. Ruggie1983a. 116. See Katzenstein 1996a;andPrice1997. 117. Johnston 1995a. 118. Nau 1990,1997.

The StudyofWorldPolitics 677 contionalarrangements, norms,and identities are embeddedin specifichistorical an textsthatcan varyso dramatically thattheycan onlybe investigated through ideographic ratherthana nomothetic approach.Emphasisis placedon thedetailed studyof textsto understand thesymbolicsystemsthatgovernactors'discourses, rather thanon an analysisofa largenumber ofcases.119 Criticalconstructivists insistthatscholars'workhas normative consequences.A scientist strainof maytrytofinda cureforcancerorinsteaddevelopa morevirulent anthrax. This choice,however,does notalterthemechanisms thatcause canceror anthrax. Criticalconstructivists, however, understand theirprojectnotsimplyas rethatexistindependent oftheinvestigator, butalso as havingthe vealingrelationships to altertheserelationships themselves. potential are developinga researchprogram thatis generating Criticalconstructivists new and significant issuesin worldpoliticsforwhichrationalist insightson important answers.Forexample,constructivists offer analysishas lackedcompelling analyses ofthetransformative shiftfromthemedievaltothemodemstatesystem;oftheend on the oftheCold Waras a recent, thatsystem;andeffects significant changewithin in themoralpurposesof statesand different international systemof variations sysfromthe temsofprocedural aboutissuesranging justice.120Constructivist arguments roleofnormsinsanctions againstSouthAfricatowhychemicalweaponsareviewed as "weaponsofmassdestruction" evenwhentheirdestructive powermaybe smaller thanthatof "conventional" andchallengerationalist acweaponsbothcomplement 121 counts. Thisresearchprogram is opento rationalist critiqueson theuse of evidenceand thelimitsofinterpretation, thepossibility andthestatusofgeneralizations, theuse of andproblemsofvariability altemative andcomparability. Rationalists explanations, This mayviewcriticalconstructivists working closelywithtextsas postmodernists. is a mistaken impression. Whatseparatescriticalconstructivism andpostmodernism is notthesharedfocuson discourse, buttheacknowledgment bycriticalconstructivistsofthepossibility ofa socialscienceanda willingness toengageopenlyin scholarlydebatewithrationalism. On bothscorespostmodemists differ whom sharplyfromcriticalconstructivists, andpositivism inrationalism thebackpostmodernists chargewithbringing through foranyknowledge. door.122 Postmodernists insistthatthereis no firmfoundation orethicaljudgments, Sincethereis nopositionfromwhichtopassscientific postmodernistanalysisis restricted to thetaskof unmasking thepowerrelationsthatare concealedinall knowledge theirown,andall formsofcommunicaclaims,including tiverationality. Througha close analysisof language,postmodernism pointsour totheinherent ofall symbolicandpoliticalorders.Sincesubjects attention instability theworldthrough onlyunderstand language,andcontroloflanguageimpliespower, 119. See Risse 1997;andAdlerandBamett1998. 120. See Ruggie1983a,1992,1998;KoslowskiandKratochwil 1994;Risse-Kappen1994;andReusSmitforthcoming. 121. See Klotz 1995c;andPrice1997. 122. George1994.

678 International Organization linguistic presentations arealwaysopentocognitive andpoliticalprocessesofdestabilization. Postmodernist instability. Itis analysisseeksoutthesesourcesofpotential in decentering interested establisheddiscourse,includingitsown,by payingattentiontowhatis marginal orsilent. bothinside Sincethemid-1980s,postmodernist analysishas grownsubstantially in thehumanities. RichardAshandoutsideoftheUnitedStates,although primarily ley was one of thefirstin international relationswho suggestedthatneorealism's visiondetermined notonlyinternational relations researchbutalso diplototalizing macy.Thus,it obscuredissues of humanagencyto whichrealismhad remained oftheanalytical andideologiresponsive.123 Forpostmodernism, reality is a creation cal categoriesthrough whichthattheoryperceivestheworldand in thenameof ofcommunicative whichitexercisesa coercivepowerthatprecludestheemergence rationality. to an Littleof thisdebatewas publishedin IO, since IO has been committed thatpostmodernism denies:theuse of evidenceto adjudicatebetween enterprise truth claims.In contrast to conventional andcriticalconstructivism, postmodernism relations fallsclearlyoutsideof thesocial scienceenterprise, and in intemational to research itrisksbecomingself-referential anddisengaged fromtheworld,protests thecontrary notwithstanding. ofthis influence Yetitis easyto underestimate thedirectimportance andindirect intellectual cuiTent. Postmodernism has foundmanyadherents bothin thebroader international studiesfieldin theUnitedStatesand in Europewheremajorjoumals andbookseriesarededicatedtopublishing theresultsofthiswork.Especiallyyounnotso mucha "turn"but haveexperienced gerscholarsofconstructivist persuasion an evolutionofviewsthatwas rootedin thepostmodernist challenge.The powerof ofsocial scienceresearchin theUnitedStates therationalist andempiricist currents makescriticalengagement withrationalism morecompelling thanisolation.Hence, indifferent andcriticalconstructivists havepositioned themwaysbothconventional such as selves quite self-consciously betweenrationalist theoretical orientations, orientations.124 realismorliberalism, andpostmodernist andResearchComplementarities Terminological Differences andrationalism Boththedifferences andcomplementarities betweenconstructivism orientations a producbetweenthesetwotheoretical promisetomaketheinteraction Bothare concernedwithwhatin ordinary tivepointof contestation. languageare thisconceptin different calledbeliefs,buttheyunderstand waysand use different arepreferences, termsin theiranalyses.The keytermsforrationalists information, areidentities, The keytermsforconstructivists andcommonknowledge. strategies, a waytounderandinterests. Rationalist orientations do notoffer norms, knowledge, do notprovidea waytoanalyze standcommonknowledge. Constructivist arguments 123. Ashley1984. 124. See Katzenstein 1996a,b;andAdler1997.

The StudyofWorldPolitics 679 strategies. Yetbothstrategy andcommonknowledgeareusuallynecessary tounderstandpoliticaloutcomes.We first discussterminology, thenturnto somedifferences and complementarities in how rationalists and constructivists analyzetherole of beliefs. Terminology. All rationalists relyon theassumption ofinstrumental rationality to providethecruciallinkbetweentheenvironment and actorbehavior.Gametheory providesa usefullanguageforrationalist analysismoregenerally. Forgametheory to offer a tractable analyticframework itmustassumethatactorshavecommonknowledge. Theyall sharethesame view of thegame,includingthepayoffmatrix, the choicepoints,thetypesofactorstheyareplayingagainst,andtheprobabilstrategic ityof each type.Playersknowtheoptionsfromwhichtheycan choose.If theyare uncertain aboutthenatureoftheiropponent, theymayhavetheopportunity toupdate theirprobability assessments as thegameprogresses becauseof information thatis revealedbythemovestakeninthegame.Givenpreferences, probabilities, andchoice points,itis possibleto derivea completesetof strategies, choicesthatplayerswill makeat everynodein thegame,and equilibrium outcomes,of whichtheremaybe many. All rationalists use theassumption of rationality to providethecruciallinkbetweenfeaturesof theenvironment-power, and institutional interests, rules-and actorbehavior.Buton theissueoftheimportance ofinformation, theyaredivided. Rationalists who subscribeto a materialist view of how to studytheinternational politicaleconomy, suchas RogowskiandFrieden, assumepreferences formorewealth and inferstrategies fromstructure, especiallythecompetitive positionsof factors, sectors,or firmsin theworldpoliticaleconomy.125Variationsin information are in theiranalysis.These authorsexpectactorsto understand unimportant theworld and theydo notconceptualizeactors'choicesin termsof gametheory accurately, involving interdependent decisions. In contrast, rationalists whosethinking is moreindebtedto gametheoryemphasize theimportance of imperfect information and strategic interaction. Theystress howchangesininformation can accountforvariations in strategies, evenifthepreferencesof actorsremainunchanged. Smallchangesin information can have a profoundimpacton equilibrium outcomes.Institutions or rulescan be consequential becausetheycan alterinformation and empowerplayersto set theagenda,make andacceptorrejectthefinalpackage.126 amendments, Constructivists insiston theprimacyof intersubjective structures thatgive the materialworldmeaning.127These structures havedifferent components thathelpin theinterests specifying thatmotivate action:norms, identity, andculture. knowledge, Normstypicallydescribecollectiveexpectations with"regulative"effectson the 125. See Rogowski1989;andFrieden1991. 126. See Shepsle 1986; Morrow1994b; and Fearon1995. It shouldbe noted,however,thatgametheoretic rationalism typically showsthatmultipleequilibriaexist,whichundermines anydeterministic predictions ofbehavior. 127. Katzenstein 1996b.

680 International Organization properbehaviorofactorswitha givenidentity. In somesituations normsoperatelike rulesthatdefinetheidentity of actors;theyhave "constitutive" effects thatspecify theactionsthatwillcauserelevant otherstorecognizea particular identity. Epistemic knowledgeis also partof a social processby whichthematerialworldacquires meaning.Finally,cultureis a broadlabelthatdenotescollectivemodelsofauthority or identity, cairiedby customor law. Culturerefersto bothevaluativestandards (suchas normsand values)and cognitivestandards (suchas rulesandmodels)that definethesocialactorsthatexistin a system, howtheyoperate,andhowtheyrelate tooneanother. Constructivist researchis notcutfromone cloth,as Finnemore andSikkinkargue in thisissue.Conventional andcriticalconstructivist analysesoftenfocuson different componentsof a commonconstructivist researchprogram-norms, identity, andculture-and,inempirical accorddifferent knowledge, research, weightstoeach ofthem.Andtheyaredividedonwhether therelationship betweenthesecomponents is definitional, are reflected conceptual, causal,or empirical.These disagreements also intheinevitable tensions thataccompany their toengagerationalism, jointeffort in theapproacheach takesto issuesof epistemology and despitesomedifferences methodology. Commonknowledge:A pointofcomplementarity.Rationalism andconstructivismaregenerictheoretical thatarecomplementary onsomecrucialpoints. orientations Game-theoretic rationalists assumetheexistenceofactors,whohavepreextypically and who sharecommonknowledgeof thegame,whichenables istingpreferences themtoengageinstrategic Constructivist research focusesonthesources bargaining. ofactors'identities-ingame-theoretic theirpreferences-and terms, oftheirinterofthecontext oftheiraction:commonknowledge. pretations Hence,rationalism and constructivism sharean interest inbeliefsorknowledge. Game theoryprovidesa vocabularyand a visualimagethathighlight notonly whererationalist andconstructivist arguments partwaysbutalso wheretheymight cometogether. thenatureoftheactorsin the Anyrationalist analysismuststipulate theirpreferences senseof specifying and theircapabilities.Whatdo actorsdesire? Whatmovescan theymake?Moreover,foranyformalgametheoretic analysisto work,it is necessaryto assumecommonknowledge.The playershaveto sharethe sameknowledge aboutthegame.Theymustknowwhattheydo notknowbecauseof andtheymustsharethesameviewofthepayoff matrixand imperfect information, theavailablesetof strategies. Rationalist accountsmakeverylimitedclaimsabout theinsights intotheoriginsofsuchcommonknowledge. theycan offer accountssuggestthatnormative structures canbe generated from Somerationalist that been material interests. institutions have createdto promote David Kreps,for has arguedthatwhathe termscorporate cultureis developedbecauseitis instance, The impossibletofullyspecifythedutiesofemployeesinanycomplexenvironment. to be written. worldis too complicatedforcompletecontracts Corporateculture notcovprovidesa setofnormsor guidelinesthatcan guidebehaviorin situations

The StudyofWorldPolitics 681 eredby formalarrangements.'28 FollowingDavid Humeand Blaise Pascal,Robert Sugdenpointsoutthatpracticesinitially developedtopromotespecificinterests can acquirea normative elementif theyare widelyunderstood and practicedwithina givensocial setting.'29 An actorwhoviolatesexistingpracticeswillbe normatively Thisis a lineof analysisthatputs sanctioned by othermembersof thecommunity. rationalists at theborderof constructivism. Normsare based on materialinterests, buttheycan takeon an auraofauthority thattranscends theirinitialpurpose. Normscan also be consequential, becausetheycan providefocalpointsin situationsof multipleequilibria;thatis, wherethereare manypossiblesolutionsto the are verycommonin gamesof game.Game theoryhas shownthatsuchsituations A rationalist incompleteinformation. analysiscan stipulatethatone of theseoutcomescan be chosen,butit does nottellus whichone it will be. Sharedcultural normsoffer onewayofselectingwhichequilibrium willbe salientfortheplayers.'30 In thelate1950sThomasSchelling, thenteachingatYale,askedaboutforty acquaintanceswhereandwhentheywouldmeeta friendin New YorkCity;morethanhalf on theinformation coordinated boothat GrandCentralStation(wherethetrains fromNew Haven)atnoon.'3'Forprofessors in 1959, arrived atColumbiaUniversity thetimemighthave been thesame,butGrandCentralStationwouldhardlyhave been as salient.Commonknowledgeis contextualized withina specificsocial setting. Not all commonknowledgecan be explainedby practicesand institutions deforexample,by resolvingproblemsof mulsignedto maximizematerialinterests, tipleequilibria.Normsthatdefinetheoptionsavailableto players,and thatshape theirpreferences, areoftenpriorto theseinstrumental For practicesandinstitutions. thelaborofconqueredlands example,slaverywas a conventional optionforsecuring in earlierperiods;it is notan optionforcontemporary states.Capturing slaves as of spoilsof war is no longeran availablemove.The medievalguilds,theholders. theburghersof Amsterdam, Englishcapitalin the late seventeenth century, and MichaelMilkenwereall concernedaboutenhancing theirmaterialwealth,butthe optionsavailabletothem(themovestheycouldmake)werehardlythesame. seekto understand areformedandknowledgeis Constructivists howpreferences Constructivism generated, priorto theexerciseof instrumental rationality. analyzes referto as comdiscoursesandpracticesthatcontinuously recreatewhatrationalists monknowledge. Constructivists do notemphasizemisperception: cognitiveoremotionalbiasesthatdistort andcanbe corrected theanalysisofnew rationality through information.'32 in thecollectiveprocessesthatstudents of Theyaremoreinterested socialpsychology have identified.'33 Constructivists focuson discursiveand social 128. Kreps1990a. 129. Sugden1989.Sugdenmentions Hume'sargument intheeighteenth Pascalmadea similar century. a century argument earlier;see N. Keohane1980,278-81. 130. Ferejohn1991. 131. Schelling1960,55-56. 132. Jervis1976. 133. Janis1983.

682 International Organization practicesthatdefinetheidentity ofactorsandthenormative orderwithinwhichthey maketheirmoves.We can thinkoftheseprocessesintwodifferent ways:intermsof ideas aboutcause-effect relationsand regulatory norms'34 or as moreor less con135 testedprocessesofidentity formation. The differences and complementarities betweenrationalism and constructivism are illustrated ofpersuasion.136Rationalists by theirtreatments interpret persuasion inthelanguageofincentives, strategic bargaining, andinformation. Theyanalyzethe provisionof new information, sometimesthrough costlysignaling, and appealsto audiences.Fora consistent rationalist, itwouldbe anomalousto thinkofpersuasion in termsofchangingothers'deepestpreferences. Constructivists, by contrast, insist on theimportance of social processesthatgeneratechangesin normative beliefs, suchas thoseprompted of thenineteenth the bytheantislavery movement century, or nationalist contemporary campaignforwomen'srightsas humanrights, propainvolveschanging ganda.Forconstructivists, persuasion preferences byappealingto moralobligations, and normsconceivedof as standardsof appropriate identities, behavior. Thedifferent stylesofanalysis-"thin"information forrationalists versus"thick" normsandidentities forconstructivists-to someextent contestin reflect thefamiliar social sciencebetweeneconomicand sociologicaltraditions.137 is Constructivism is nomothetic. is adequateto ideographic, whereasrationalism Neitherperspective coverall aspectsof social reality.But at one criticalpointtheyare joined. Both as a backrecognize-constructivism as a centralresearchprojectand rationalism groundcondition-that humanbeingsoperateina sociallyconstructed environment, whichchangesovertime.Hence,bothanalyticalperspectives focusin one way or another on commonknowledge-constructivism onhowitis created, rationalism on howit affects strategic decisionmaking.The coreof theconstructivist projectis to inpreferences, availablestrategies, explicatevariations andthenatureoftheplayers, acrossspace and time.The core of therationalist projectis to explainstrategies, andcommonknowledge. Neither givenpreferences, information, projectcanbe comtheother. pletewithout

Conclusion Thehistory ofIPE as a fieldwerebuilton a richintellectual ofIO andtheemergence tradition thatdevelopedin the1940s,1950s,and 1960s.Manyofthemajorlinesof thathavepreoccupiedscholarsofIPE and international relationsforthe arguments 134. See Goldsteinand Keohane 1993b;Haas 1993, 1997; Sikkink1993a,b;Adlerand Haas 1992; Adler1987;andHaas 1992c. See also Marchand Olsen 1989; Powelland DiMaggio 1991; 135. Wendt1987, 1992,forthcoming. andJepperson, andKatzenstein 1996. Wendt, 136. On persuasion, see Finnemore1996a; anda longtradition oflegal scholarship, including Franck 1990;andChayesandChayes1995. 137. Swedberg1991.

The StudyofWorldPolitics 683 lastthreedecadesweredevelopedduringtheseyears.Thenas nowscholarscontinue betweenpower,wealth,andsocialpurpose. toanalyzetheinteraction The history ofIO andtheemergence ofIPE havebeenpartof theelaboration of morecomplexandself-consciously analytical formulations withwhichscholarshave narrative canreproduce analyzedenduring problemsofworldpolitics.No historical thecomplexandidiosyncratic evolutionofscholarship. Forease ofpresentation and because it reflectsimportant our aspectsof our own intellectual autobiographies, storyhas followedtwodistinctive strands:theinterplay betweenrealistand liberal currents oftheory intheevolutionoftheanalysisofIPE andtheanalysisofdomestic politicsandIPE. We havearguedthatinthe1990ssomeofthemajorpointsofcontestation shifted. Influenced in economicsand culturalstudies,debatesbetween by strongcurrents rationalism andconstructivism arebecomingmoreimportant. Theyoffer contrasting forresearchin thesocialsciencesatlargeandininternational orientations analytical relations andIPE. Thegreatest promiseintheintellectual debatebetweenproponents ofrationalism andconstructivism does notlie in theinsistence thatrealitycan only be analyzedin one conceptuallanguage-theone preferred bytheanalyst.Insisting on one's ownlanguageis a sterileintellectual exercise.Knowledgeandunderstandofdifferent researchorientations ingarepromoted bydebatesamongtheproponents andresearch Butoneshouldneverforget thatattheendofthedayorientaprograms. tionsand programsare onlyusefulif theyare deployedto specifyintellectually tractable andsubstantively important questions. Analyticalprogressin thestudyof IPE is possiblein researchprograms despite contestations betweengeneraltheoretical orientations. We believethat continuing thefieldhas becomeincreasingly we havebetterconceptualtoolsand sophisticated; richerinterpretations thanwe had in the1970s.Our substantive findings, however, remainmeager:counterintuitive, well-documented causal arguments are rare.And someanalytical advanceshavetoldus whywe cannotmakestrong rather predictions thanhowto go aboutdoingso, as is thecase fortherevelation fromgametheory of of multipleequilibria.Nevertheless, thefrequency we knowa greatdeal morethan we didthirty ofprocessesthatarecentraltohowtheworld yearsago abouta number politicaleconomyworks,suchas howpoweris deployedundervariousconditions of howinternational vulnerability, regimesaffect government policies,andhowdomesticinstitutions and worldpoliticsaffecteach otherthrough institutional processes. without theaccumulaManyofthearticlesin thisissuecouldnothavebeenwritten research inissuesranging tionofsubstantive fromtradeandindustry, findings money and energy, and financeand investment andhuto,amongothers,theenvironment manrights. In international well-established boundpoliticsand in theworldof scholarship ariesarebeingblurredandnewonesarebeingcreated.Worldpoliticsis witnessing enormous changein thewakeofthecollapseoftheBerlinWall,theendof strategic and thepeacefuldisintegration of theSovietUnion.We are observing bipolarity, different typesof democratization processesin different worldregions,ethnicconflictsoverthecontrolofterritory, andgrowingconflicts overthespreadofweapons

684 International Organization ofmassdestruction toregionalpowers.We arealso seeingfar-reaching experiments witheconomicintegration in someworldregionsandcontinuing marginalization in others, expandedtradeundertheauspicesofa revamped WorldTradeOrganization, andtheredefinition oftheroleoftheInternational Monetary Fundandothergovernancemechanisms in thewakeofAsia's financial crisis.At thesametime,religious fundamentalism is gaininggroundin muchoftheMuslimworld.Andstrongsecular socialmovements championing environmentalism, feminism, andhumanrightsare linksacrossdisactiveworldwide. Manyopportunities existforbuildinginnovative All of thesedevelopments the ciplinaryand subdisciplinary boundaries. illustrate ofcontemporary complexity worldpoliticsandtheopportunity todrawnewconnectionsacrossgenerictheoretical orientations andbetweenspecificresearchprograms. The rubricforthisera of increasingtransboundary activityis variouslycalled "internationalization" or "globalization."This is not an unprecedented developat theend of theninement.International financialmarkets werehighlyintegrated teenthcentury-perhaps evenmoreso thantheyhavebeensince,at leastuntilvery recently. Tradeflowswerealso muchhigheratthistimethanformostofthetwentiethcentury. New technologies have openednewopportunities, butit is notevident how theywill affectthecontroland authority structures in stateand society.How extensively nationalpoliciesand practiceswill be reshapedby globalization-and how effectively institutions will resistsuch pressures-remains to be seen. Such research willfocusagainon issuesoftherelationships thathavebeencentralto the literature of thelast thirty years:betweenwealthand power,statesand markets, interests andinstitutions, theinternational politicaleconomyand domesticpolitics. Constraints andopportunities willchangeand so maytheidentity ofthekeyactors andthenormstheyaccept.Butobservers ofthisnewrealitywillstillbe abletolearn fromtheaccomplishments, andthemistakes, ofpreviousgenerations ofIPE scholars. Thesophistication ofworkinIPE has notmadeitdirectly applicabletopolicy.As notedearlier,we cannotpointto clearscientific "findings"aboutcause and effect ofmorerigorthatpolicymakers canreadilyapply.Atthesametime,theapplication ous socialscientific forIO seemsto have discouraged standards by referees policy articleshavebecomefewandfarbetween.Hence,the commentary: policy-relevant andpracticein thegulfthathas developedbetweenscholarship pagesofIO reflect international relations. can Yetevenifthelinksbetweenscholarship andpolicyarenotclose,connections in theliterature be made.Moresignificant thanthespecificdebatesorevenfindings are theinterpretations of changingrealitythathave beenputforward by analysts: in whichthefieldhas engaged.A numberofideas conceptsandbroad"theorizing" are beingtakenforgrantedin policydiscusformulated in theliterature originally orsimply oftheinternational sions.Whether thisreflects an effect relations literature inthepolicyandacademicworldsis notclear.Atanyrate,it parallelunderstandings forpoweras well is conventional hasimplications wisdomnowthatinterdependence as forwealth,thatinternational institutions constitute a valuablesetof instruments forpromoting theinterests of statesthrough and thatunderstanding cooperation,

The StudyofWorldPolitics 685 butalso domesticinterests domesticpoliticaleconomyrequiresnotjust examining institutions as wellas thestrucdomesticandinternational takingintoconsideration betweennonof interactions tureof theworldpoliticaleconomy.The significance inthepolicyworld. understood stateandstateactorsis also increasingly The specificapproachthatscholarschoosetofollowin theirworkwilldependon viewpointor to to advancinga theoretical committed whether theyare principally theirmethpredispositions, theirownanalytical solvingspecificempirical problems, odologicaltools,thedatato whichtheyhaveaccess,theresourcesat theirdisposal, andthevaluestheyhold.Noneofus shouldbe toosurethatourownchoiceswillbe of Even thoughwe buildon worksof earliergenerations intellectually productive. shouldmake socialscientists, ourvisionis limited.Ournearsightedness outstanding theoretius skepticalthatthelatestturnofthescrewofa particular methodological, debatewill magicallybringour analyticalbinocularsinto cal, or epistemological debatesaboutaspectsofworldpoliticsthatchange, sharper focus.Yetnewintellectual of IPE froman increasing integration and thosethatdo not,pointto highreturns for scholarship intobroadersocial sciencedebates.This is a welcomeopportunity fortheeditorsandreadersofIO. excitingprospect anyauthorandan intellectually