Additional endorsements for Foreign Policy 2e

Foreign Policy
Theories, Actors, Cases
SECOND EDITION

'The editors have mobilized an outstanding group of scholars and practitioners to explore through literature reviews and case studies how theories of international relations, such as realism, liberalism, and constructivism, can help us to understand foreign policy behaviour. They also demnstrate how the choice of analytical level-the states system, national and organizational characteristics, and personality traits-affect the explanations that emerge. The original studies are senstive to the role of non-state actors in accounting for foreign policy choices, and they also nclude importan! examples of middle powers' influence in certain global issue reas. The editors'theoretical visin of the project assures readers of a comprehensiveand enduring effort. Ths volume san authoritative last word in thefield of foreign policy analysis.' K.j. Holsti, University Killam Professor, Emertus, University of British Columbia

'A unique and indispensable resource. Its coverage is remarkably comprehensive and' provides ajudiciousblendoftheory and illustration. The theoretical chapters are clear and accessible, and the case materials and topical chapters offer a rich array of pedaggica! possibilities. Like The GlobaUsation of World Politics, this book deserves to be widely adopted.' Stephen M. Walt, Harvard University The editors have filled a long-neglected gap by producing a volume that authoritatively covers the state of the art in the study of foreign policy. The book looks set to become a definitive textfortheteachingand study of foreign policy.' Richard G. Whitman, University of Kent 'The book combines od and new perspectives with discerningcare. In-depth explorations of empirical examples present a geographically diverse set of cases for teaching. Highly recommended.' Olav F. Knudsen, Swedish Institute of International Affairs

Edited by

Steve Smith Amelia Hadfield Tim Dunne

OXFORD
UNIVERSITY PRESS

Section 1

Foreign Policy Analysis: Theoretical and Histrica! Perspectives

The hssory and evolution of foreign policy analysis

VALERIE M. HUDSON

Chapter conens
Introduction: three paradigmatic works Classic FPA scholarship (1954-1993) The psychological and societal milieux of foreign policy decisin making FPA self-reflection n the late 1970s and 1980s Conciusion: contemporary FPA's research agenda 13 17 23 27 30

Reader's guide
This chapter traces the evolution of Foreign Policy Analysis (FPA) as a subfield of International Relations (IR) from its beginnings n the 1950s through its classical period until 1993; t then sketches the research agenda of contemporary FPA, which s represented by the other chapters n this volume. Three paradigmatic works, by Richard Snyder and colleagues, James Rosenau, and Harold and Margare! Sprout, laid the foundation of this subfield. In turn, these works created three main threads of research in FPA, focusing on the decisin making of small/large groups, comparative foreign policy, and psychological/sociological explanations of foreign policy. These three primary reas of research have waxed and waned n.importance to the subfield over the years. Current FPA scholarship explores linkages between these literatures, seeking both greater cross-level integration of explanation and new methodologies more appropriate to cross-level analysis.1

Introduction: hree paradigmac works

What are the origins of foreign pocy analysis (FPA)? In one sense, FPA-style work-that is, scholarship whose theoretical ground is human decisin makers, acting singly or within groupshas been around as long as there have been historians and others who have sought to understand why national governments have made the choices they did regarding interstate relations. (See Box 1.1). But FPA-style work within the field of International Relations per se s best dated back to the late 1950s and early 1960s. Three paradigmatic works arguably built the foundation of FPA. e Decisin Making as an Approach to the Study of International Politics by Richard C Snyder, H.W. Bruck, and Burton Sapin (1954: see a/so Snyder et al. 1963; reprinted n 2002).

14,

VALERIEM. HUDSON

CHAPTER 1 THE HiSTORY AND EVOLUTGN OF FOREIGN FQUCY ANALYSIS

FOR'EIGN POLICY The strategy.orapproach chosen by the.natioh'i govemmentt'achieve-ts'gcjasJn.its;:. -':;:; '.reiationswith-exernalentities. Tnismcidesdecisipns .torio hdthing. - = -vv^^Jj^'^M'iK^SJv^' poucY:BEHAVipUR Th'jjserv'bl rtefacts of fcreign pblicy-speefic actibns 'andwqrds sedto ' "Influenc pthers n the realrh of foreign plicyiTnay inclde the ctegorization sfsuch hehav!&ur,'sucrras.' : along conflict-cooperatph continua, which categorizatibns cud be .used to.constrct data nclding . ; event data. FPB mayincludebehaviour that was accidental brumntendedby.thegpyernrrient, and in ."-.'-''. addition decisidns todo nottingmay not leye^anyBhaviourartfactTn'ifhrelsippag'-V^^enV--' thconceptof foreign policy and the cncptof foreign policy behaviour. : : -:;' :;' : , ' W.;" "/V--":j ..--::: '. FOREIGN POLICY ANALvsis The subfield of ntrnatiortaS refatios hat seeks te expiain. foreign poiicyipr, - : ; alternatiyely, foreign policy behaviour, with reference to the theo retica! grunc .of.haitiari'ciecision. i!,rL; .: rnakers, actingsinglyand ihgrpups.fhesubfield hasse\/erai hllmarks: V ' '" ::/ -> "; "i :
: 9 ; acommitmenttO:lobkblowthnatiori-stt-|ev6fanal\&toartor-^

RcHaM Snyder

James Rosenau. 'Pre-theones and Theones of Foreign Policy' byjames N Rosenau (a book chapter wntten in 1964 and published in Farrell 1966)

Harold and Margaret Sprout Mi Context of Internationa! Politics byjHarotd and Margaret Sprout (1956 expanded and revised in artide form in 1957 and their 1965 book The Ecohgcal Penpecttve on Human Affairs with Special Reference to International Politics)

.Eron Sapin;"-' . i se'!sp. Snyi.r et-a:

.'"-"": -/

.'Contabutedafocuson --*''*

_ Development of actor-specific theory that would leadlo the development of generlizable propostions at the level of middle-range theory

Foreign policy can only be explamed with reference to the psycho-milieu (the psychological, situational, political, and social contexts) of the individuis mvolved in decisin making

-a a"cpMmitmentto\buldactof-specificheo^asthelnt^aceoe&eeri.actp!;-gen
: !

TI, rather than just

cpmplexity'pf the real vyorld; . .'.- . '":'''- ::"--::~.'i'":=:' ''"' : " ".. ,- ' :''.'::i^:^:-'^>'.^:-^\,---

;;. a commitmentto prs rrulticausal exp'anaiions sptnning'mijlipjs Isvels cf anlysis;; Wf: \"-JM:--'V-' , a commitment to utilize theory and'findings frprn cross the specirum-bf social scjencg:' - '^ "::: :^-. - : - ' . comrnitmenttovievying the prqcess of foreign pciicy dsdsion making as mpcrtnt as the butputthrpf. ACTOS-GENERAL THEORY Theory that sxplains the behayjuf.of actprs in genera', such sgame 'theory/'^: ;:'.'' ACTOR-spEciFic THEORY Theory that xplains the behavipr ofspecific actprsrsiiqhVas FPA theory. Thii -\: , type of theory may be generalizable.'but hder specific scope cbnditions fr appca'bliy.'cpr-sp ific !..'
!

theory isa forrh of middle-range -thory, in that itis mor generlizable than'msightsdrivdfrpnicase '

studies'but, pri tri. otherhahd, has more severe scope coditions constrainng is generazsbit*/ than

"'ctor-generiripry. Hwever, given te nture.actor-spcific theory aiiows for richerexpiantibhand- :;;..even'prediction of the foreign policy behaviour of particular e'ntities than doesacbr-genera'thsory. "t:. .J

o 'Pre-theories and Theories of Foreign Policy' byjames N. Rosenau (a book chapter written in 1964 and published in Farrell 1966). s Mon-MHieu Relationship Hypotheses in the Context of International Politics by Harold and Margaret Sprout (1956: expanded and revised in article form n 1957) and their 1965 book The Ecological Perspective on Human Affairs with Spedal Reference to International Politics). (SeeBoxl.2.) The work of Richard Snyder and his colleagues nspired researchers to look below the nation-state level of analysis to the players involved: We adhere to the nation-state as the fundamental level of analysis, yet we have discarded the state as a metaphysical abstraction. By emphasizing decisin making as a central focus we have provided a way of organizing the determinants of action around those officials who act for the political society. Decisin makers are viewed as operating in dual-aspect setting so that apparently unrelated nternal and external factors become related in the actions of the decisin makers. Hitherto, precise ways of relating domestic factors have not been adequately developed. (Snyder et al. 1954: 53) Richard Snyder leading a foreign policy seminar.
Source: Princeton University Library. Princeton University Archives, Department of Rare Books and Special Collections, Princeton University Library.

16

VALERSE M. HUDSOM

CHAPTER 1 THE HISTC'RY AMD EVGLUTIGN OF FOREGN PGUCY ANALYSIS

17

In taking this approach, Snyder and his coileagues bequeathed to FPA its characteristic emphasis on forign policy decisin making (FPDM) as versus foreign policy outcomes. Decisin making was best viewed as 'organizational behaviour', by which the basic determinants would be spheres of competence of the actors nvolved, communication and information flow;and motivationsofthevariousplayers. Desirableexplanations would thus be both multicausal and interdisciplinary. As explored in the Foreword, James Rosenau's pre-theorizingencouraged scholars to systematically and scientifically tease out cross-nationally applicable generalizations about nation-state behaviour. As Rosenau put it, To identify factors s not to trace their influence. To understand processes that affect externa! behavior is notto explain howand why they areoperative undercertain circumstances and not under others. To recognize that foreign policy is shaped by internal as well as externa! factors is nottocomprehend how the twointermix orto indcate the conditions under which one predomnales overthe other.... Foreign policy analysis lacks comprehenslve systems of testable generalizations Foreign policy analysis s devoid of general theory. (Rosenau 1966:98-9) General testable theory was needed, and the intent of Rosenau's article was to point in the direction it iay. However, the general theory Rosenau advcales is not the grand theory of Cold War IR: the metaphor Rosenau used in this work is instructive n this regard-FPA researchers should emulate Gregor Mendel, the father of modern genetics, who was able to discern genotypefrom phenotypein plantsthrough careful observation and comparison. Are there genotypes of nation-states, knowledge of which would confer explanatory and predictive power on our models of foreign policy interaction? What Rosenau was encouraging was the development of middle-range theory-theory that mediated between grand principies and the complexity of reality. At the time Rosenau wrote this article, he felt that the best way to uncover such mid-range generalizations was through aggregate statistical exploration and confirmation. Rosenau also underscored the need to intgrate information atseveral levis of analysis-from individual leaders to the international system-in understanding foreign policy. As with Snyder, the best explanations would be multlevel and multicausal, integrating information from a variety of social science knowledge systems.

Harold and MargaretSprout contrbuted to the formation of the field by suggestingthat understanding foreign policy outputs, which they associated with the analysis of power capabilities within an interstate system, without reference to foreign policy undertakings, which they associated with strategies, decisions, and intentions, was misguided: 'Explanations of achievement and estimations of capabilities for achievement invariably and necessarily presuppose anteceden! undertakings or assumptions regarding undertakings. Unless there s an undertaking, there can be no achievement-and nothng to explain or estmate' (1965: 225). To explain undertakings, one needs to look at the psycho-milieu of the individuis and groups making the foreign policy decisin. The psycho-milieu is the international and operational environment or context as it is perceived and nterpreted by these decisin makers. Incongruities between the perceived and the real operational environments can occur, leading to less than satisfactory choices n foreign policy. The sources of these incongruities were diverse, requiring once again multicausal explanations drawing from a variety of f ields. Even in these early years, the Sprouts saw a clear difference between foreign policy anaiysis and what we have called actor-general theory: Instead of drawing conclusions regarding an individual's probable motlvations and purposes, his environmental knowledge, and his ntellectua! processes linkng purposes and knowledge, on the basis of assumptions as to the way people are likely on the average to behave in a given social context, the cognitive behavioralist-be he narrative historian orsystematic social scientist-undertakes to find out as precisely as possible how specific persons actualiy did perceiveand respond n particularcontingencies. (Sprout and Sprout 1965:118) The message of these three works was powerful in its appeal to certain scholars: the particularities of the human beings making national foreign policy were vitally mportant to understanding foreign policy choice. Such particularitis should not remain as undigested diosyncracies (as n traditional single-country studies), but rather be incorporated as nstances of largercategories of variation in the process of cross-national middle-range theorybuilding. Mltiple levis of analysis, rangingfrom the most micro to the most macro, should ideally be ntegrated in the sen/ice of such theory. The stores of knowledge of all the social sciences must be drawn upon in this endeavour. The process of foreign policy making was at least as mportant, if not more mportant, than foreign policy as an output. The substance of this message was and contines to be the 'hard core' of FPA. (See Box 1.3.) The second wave of theorizing built upon the foundational paradigmatic works. Between 1974 and 1993 FPA developed a numberof parallel research pathways. (See Box 1.4.) Other parts of the message were more temporally bounded. As we shall see, certain methodological stances that perhaps seemed self-evident in the early 1960s would not stand the test of time. These would engender troubling paradoxes, which would plague the field and lead to a temporary decline in some reas in the mid to late 1980s untl they were satlsfactorily resolved. Despite these paradoxes, the first bloom of FPA, lasting from the late 1960s to the aforementioned decline, was a time of great intellectual effort and excitement.

Classic FPA scholarship (1954-1993)

The energy and enthusiasm of the first generation ofworkin FPA (1954-1973) was tremendous. Great strides n conceptualization, along with parallel efforts in data collection and methodologcal experimentaron, were the contributions of this time period. The second

James Rosenau, pioneer of foreign policy analysis.

18.

VALERIE M. HUDSON

CHAPTER 1 THE HiSTORY AND EVOLUTSON OF FOREIGN POLiCY AN.ALYSIS

19

B&x f ;f erssicalTpJign polcy^arajy?is4the secona genertioiif:

^i?efers:tqthg;prc2ssandstfuctrefgrops'maj<ingto ^sdicl'-anged n sze rcrri Very sni'l jrq'ups to isrge.drganizaions'and b-ra=ucracis.:insightfrorn. ._ .' r^bcial.iisychlogywereir.corported^ nio^RA^ltwasdiscteredthatthe^nctivaignto.iTTainfainjroDp, JonsenXsriSpersonaiaccpancebythe'graupcuId res^^ '-;.'-'.

,'- Rgsesfchers began .siuay the influence qf rgaizatipn'prpcss'pd bureajcratic pciltics on fqfign>Vj;:-"-. :: plicv'ecision rnaking/prgnizations nt.bureaucracies'put tfteir own:sL'r.''ivaTaf the tc.pf 'th'sir ist cf.-'':; :. "priorities; often.th'ey viil! ;eekto1ncreasethe!rreativstrength..!1:v/asfq"d that the itenorqbjetyp'Hfv '

-:Cornpafailv.fQre!gn poiicy^^^'i^-r-'-;^:^";/:"':^;-.: " : ,:^>?::'-;,;{;"x;->^^>. ; V-;

Harold and MargaretSproutwho emphasized the psychological milieu of individual and group decisin making.
Source: Priceton University Library. Princeton University Archives, Department of Rare Books and Special Collections, Princeton University Library.

vijfiubfisld fcompartive foregn p'oiicy CFP}devetoec tS^arespbnse.to-'anissScsena's challenge : v.-> ^:tobi[d:a1ro3ifhatidnaiandrriutfie^ '.'as vyrfre;treatv'.fria;!g,-o'r diplornacy-these avents cQ'yd.be crrparec a^d:fgresasc.Da&Vas f^ ': .fcllscted-ona'vriety.cf possib!ecxpianatpryfactrstO:deterrnnepatrn:'y.wh!chthesa!ndspsndent' :":. variables Viere corrktu. Researchrs hpedtp emerge wiih:agrnd.n;f:sdinso]'/::5ff3rignpolcy:-.-:V-'

Cpgnitiye processes Cognition.lernihg, heurisiic aHades, emoticn, etc.; : y/; -"f; " ', "-'-. '-:._ Leader personality and orientation Qperationl cdds, roiivations, psychpbigraphy, etc. Smal-groupdynamcsGrouptr.irk,hai\'grop..ccaiitiors,:etc. '-';; !;..''.':' .'.-' '.---.'-'?';' :'?".-;-'-; :

rPsyehlgical infiuencs npregn pciicy decisin rnaking.S *: ^^ -;.S /v

:'-' lncreasng.ttentQri'vs directet tb themnaof the fbraigh policy cecisip" n~ksr. nder.certafn stressfui--:;cpncitipris,' incividai charac'erlstics vvuid.bsccme crcian'hdrsandirg hsw decsons are. pade'.-' - : -Aisc, th.-prqblern of rnisperceptionvv identified.' with coerttiai disstrous ccnscgnces .n rgaticn to;" . . : .qustionsf'A'arnd peace.- -;'";; :i:v>: : ^V^ -:';: ''.'.'-, :v'-^^;;j:':s"' % '.,":.' ".'.v.;r :":> :' .- J;:!. ";.',"' '-:

Organiztion prcess incremental iearning, standard operating procedures, impsmentation issues, etc. BureaucrticpoltcsTurf,.morEle,'bdget,jnfnce, intr-gncy group plics, aic. . , ; . : ' '_ }.

Culture and fcrslgn policy Identityand naiionaism, herdichistories.'roiehapry, etc. : ;.,;-'. ; '. ; ;V;_: ^ Domestic plitical cpntestation Regifne type, media, pdiiticl intefestgrops, organizad party ,: contestafion and electoral poltcs,;ec. \:'^ ':. ;' ;. - -''.: .: : v !.-. _<. ";.'" '! :: . "/ ; ;: ^; . :; : - .' ; : -:";" National attributes and foreign policy Geography, rssources,-econc'micfactors, etc. '-,-.....
: System

!effectson foreign policy Anarchy,;dstribution of povvr, rgicna; 'balances'f pqwr etc. ; .:

.Tnscdeta ccbnpxt aiso canie to the "tora Resarqfi'ers exarnned':howfe hational atfibtss,.such as ' :'.: ' culjure, Kistory.'gogrphy/ecpnorh'ics.'-pplitic! instituticns, miiitary pows: idablcgv1, and derngraphics, .' determined policy making. The hature'of regirae type alsp rose n prominence;particularly with the '-.: : ';:re[zat!on'tht"derrbcracies'tended not tp figh with, pneanothef.'^v; :': ,-: _ :-.-( .,:, - '.i ';'^^''~ ;-;'>: - '"'

generation of workfrom about 1974 to 1993 expressly built upon those foundations. Though t is always difficult to set the boundaries of a field of thought, the overview which follows includes a representative sampling of classic works n the first and second generation which both examined how the 'spedfics' of nations lead to differences n foreign policy choice/behaviour, and put forward propositions in this regard that at least have the potential to be generalizable and applicable cross-nationally.

Group decisin making Snyder and colleagues had emphasized the process and structure of groups making foreign policy decisions (Snyder extended his work with case studies n collaboration with Glenn Paige; see Snyder and Paige 1958; Paige 1959; Paige 1968). Numerous scholars echoed this theme n their work, which ranged from the study of foreign policy making in very small groups to the studyof foreign policy making in very large organizations and bureaucracies.

20

VALERIEM. HUDSON Small group dY Some of the mosttheoretically long-lived work produced duringthis period centred on the consequences of making foreign policy decisions n small groups. Social psychoiogists had explored the unique dynamics of such a decisin setting before, but never in relation to foreign policy decisin making, where the stakes might be much higher. The most important work is that of Irvingjanis, whose seminal Victims of Gmupthink almost single-handedly began this research tradition. In that volume, and using studies drawn specifically from the realm of foreign policy, Janis shows convincingly that the motivation to maintain group consensus and personal acceptance by the group can cause deterioration of decision-making quality. The empirical research of Leana (1975), Semmel (1982), Semmel and Minix (1979), Tetlock (1979), and others extended this research using aggregate analysis of experimental data as well as case studies. Groupthink becomes one of several possible outcomes in the work of C.F. Hermann (1978). Hermann categorizes groups along several dimensions (size, role of leader, rules for decisin, autonomy of group partid pants), and is able to make general predictions aboutthe likely outcome of deliberations in each type of group. The work of the second wave moved 'beyond groupthink' to both refine and extend our understandingofsmall-groupprocesses. Representativeworkincludes Hereketa/. (1987, 1989), McCauley (1 989), Ripley (1 989), Stewart et al. (1 989), Hart (1 990), Gaenslen (1 992), and Hart et al. (1 997). The second wave also brought with t a new research issue. How does a group come to understand, represen!, and frameagiven foreign policy situation? Works indudethosebyGeorge Breslauer, Charles F. Hermann, Donald Sylvan, Philip Tetlock, and James Voss (Vertzberger 1 990; Breslauer and Tetlock 1 991 ; Voss et al. 1 991 ; Billings and Hermann 1 994). Turning to efforts by individual scholars, we wiil highlight the work of Khong (1 992) and Boynton (1 991 ). Boynton wishes to understand how human agents in groups come to agreement on the nature of a foreign policy situation. In his 1991 paper(cited above), he uses the officiai record of Congressional Committee hearings to investgate how committee members make sense of current events and policies. By viewing the questions and responses in the hearing as an unfolding narrative, Boynton is able to chart how 'meaning' crystallizes for each committee member, and how they attempt to share that meaning with other members and with those who are testifying. Boynton posits the concept of 'interpretive triple' as a way to understand how connections between facts are made through plausible interpretation-in effect, ascertaining which nterpretations are plausible within the social context created by the hearings. Khong's 1 992 book, Analogies at War, has a similar aim but with a different focus: the use of analogies to guide problem framing by foreign policy makers. In this particular work, Khong demnstrales how the use of conflicting analoges to frame the problem of Vietnam led to conceptual difficulties in group reasoning about policy options. The 'Korea' analogy gained ascendance in framing the Vietnam problem, without sufficient attention being paid to the incongruities between the two sets of circumstances.

CHAP7E3 1 THE HISTORY AMD EVQLU71QN OF FORSIGN PQLICY ANALYSIS

Economic Organizations (from the 1 920s). First-period research showed how 'rational' foreign policy making can be upended by the attempt to work with and through large organized governmental groups. Organizations and bureaucracies put their own survival at the top of their list of priorities, and this survival s measured by relative nfluence vis vis other organizations ('turf'), bytheorganization's budget, and bythe moraleof its personnel. The organization will jealously guard and seek to increase ts turf and strength, as well as to preserve undiluted what itfeels to be its 'essence'or'mission'. Large Organizations also develop standard operating procedures (SOPs) which, while allowlng them to react reflexively despite their inherentunwieldiness, permitlittleflexibilityorcreativity.TheseSOPsmay be the undoing of more innovative solutions of decisin makers operating at levis higher than the organizaron, but there s little alternative to the implementation of policy by bureaucracy. The interface between objectives and implementation s directly metatthis point, and there may be substantial slippage between the two because of the incompatibility of the players' perspectives. Although the articulation ofthis research agenda can befound in works such as Huntington (1960), Schilling et a!. (1962), Hilsman (1967), and Neustadt (1970), probably the most cited works are Allison (1 971 ) and Halperin (1 974) (additional works co-authored by Halperin include Allison and Halperin (1 972) and Halperin and Kanter (1 973)). In his famous Essence of Decisin, Graham Allison offers three cuts at explaining one episode in foreign policy-the Cuban Missile Crisis of 1 962. Investigating both the US and the Soviet sides ofthis case, Allison shows that the unitary rational-actor model of foreign policy making does not suffice to explain the curiosities of the crisis. Offering two additional models as successive 'cuts' at explanation, the Organizational Process Model and the Bureaucratic Politics Model (one of intra-organizational factors, and one of inter-organizational factors), allows Allison to explain more fully what transpired. His use of three levis of analysis also points to the desire to intgrate rather than segregate explanations at different levis. Halperin's book Bureaucratic Politics and Foreign Policy (1974) s an extremely detailed amalgam of generalizations about bureaucratic behaviour, accompanied by unforgettable examples from American defence policy making of the Elsenhower, Kennedy, and Johnson years. Itshould be noted that bureaucratic politics research gained mpetus from the Vietnam War ongoing duringthis period, because the warwasseen bythe publicas defence policy run amokdue, in part, to bureaucratic imperatives(e.g. Krasner1971).

Comparativa foreign policy

Those who took up James Rosenau's challenge to build a cross-national and multilevel theory of foreign policy and subject that theory to rigorous aggregate emprica' testing created the subfield known as comparativo foreign policy (CFP). it is in CFP that we see most directly the legacy ofscientism/behaviouralism in FPA's genealogy. Foreign policy could notbestudied n aggregate-foreign policy behaviour could. Searching for an analogue to the Vote' as the fundamental explanandum n behaviouralist American political studies, CFPers proposed the foreign policy eventthe tangible artefact of the influence attempt that s foreign policy, alternatively viewed as 'who does what to whom, how' n nternational affairs. Events could be compared along behavioural dimensions, such as whether positive or negative effect was

Organizalional procsss and bureaucratic poiitics

This first period also saw the emergence of a strong research agenda that examined the influence of organizational process and bureaucratic politics on foreign policy decisin making. The foundations of this approach can be traced back to Weber's The Theory of Social and

22

VALERIE M. KUDSON

CHAFTER 1 THE HiSTORY AND EVOLUTiOM OF FOREIGN PGUCY ANALYSIS

?23:

'n that the Kansas team has developed machine coding of events, leadingto much morereliable and capacious data collection and coding than was possible n the first wave of events data (Schrodt 1995).

Integrated explanations In contrast to the other two types of FPA scholarship being discussed, CFP research aimed explicitly at ntegrated multilevel explanations. The four most ambitious of these projects were those of Brecher (1972) and his associates n the IBA Project (Wilkenfeld et al. 1980), DON (Rummel 1972,1977), CREON (East et al. 1978; Callahan eta/. 1982), and Harold Guetzkow's INS (Guetzkovv 1963). Independent variables at several levis of analysis were linked by theoretical propositions (sometimes instantiated n statistical or mathematical equations) to properties ortypes of foreign policy behaviour. At least three of the four attempted to confirm ordisconfirm the propositions byaggregateempirical testing. Unfortunately, the fact that the empirical results were not all that had been hoped for ushered in a period of disenchantment with all things CFP, as we shall see in a later section.

beingdisplayed, or what instruments of statecraft(diplomatic, military, economcs, etc.) were used n the influenceattempt.or what levelofcommitmentofresources was evident. Behaviour as disparate as a war, a treaty, or a state visit could now be compared and aggregated n a theoretically meaningful fashion. This conceptualization of the dependen! variable was essential to the theory-building enterprise in CFP. To uncover law-like generalizations, one would have to conduct emprica! testing across nations and across time; case studies were not an efficient methodology from thisstandpoint. However, with the conceptual breakthrough of the'evenf, it was now possible to collect data on a variety of possible explanatory factors and determine (by analysing the variance in the event's behavioural dimensions) the patterns by which these independent variables were correlated with foreign policy behaviour (see McGowan and Shapiro 1973). Indeed, to talk to some scholars involved in CFP research, t seemed that their goal was nothing less than a GUT (grand unified theory) of all foreign policy behaviour for all nations for all time. Some set of master equations would link all the relevant variables, independen! and dependent, together, and when apped to massive databases providing vales for these variables, would yield R-squares approaching 1.0. Although the goal was perhaps naive n its ambition, the sheer size of the task called forth mmense efforts in theory building, data collection, and methodological innovation that have few parallels in IR. Events data

The psychological and societal milieuxof foreign policy decisin making

The mind of a foreign policy maker s not a tabula rasa: it contains complex and intrcately related information and patterns, such as beliefs, altitudes, vales, experiences, emotions, Iraits, style, memory, national, and self-conceptions. Each decisin maker's mind is a microcosm of the variety possible in a given society. Culture, history, geography, economics, political nstitulions, deology, demographics, and innumerable other factors shape the societal context in which the decisin maker oprales. The Sprouts (1956, 1957, 1965) referred to these as the milieu of decisin making, and scholarly efforts to explore that milieu were both innovalive and impressive during Ihis first period. Michael Brecher's work cited above (Brecher 1972) belongs in this genotype as well. Brecher's The Foreign Policy System of Israel explores that nation's psycho-cultural environment and ils effecls on Israei's foreign policy. Unlike Brecher's inlegrative approach to the psycho-social milieu, most works in this genotype examined eilher Ihe psychological aspects of FPDM, or ils broader socielal aspects.

The collection of events data was funded to a significant degree by the US government. Andriole and Hopple (1981) estmate that the government (primarily Defense Advanced Research Projects Agency and the National Science Foundation) provided over $5 million for the development of events dataseis during the time period 1967-1981. Generally speaking, the collection effort went like this: students were employed to comb through newspapers, chronologies, and other sources for foreign poiicy events, which they would then code according to rules usted in their coding manuals, have their coding periodically checked for intercoder reliability, and finally punch their codings up on computer cards. For example, if we wanted to code an event such as 'The USA invaded Afghanistan', we would code a date (DDMMYYYY), the actor (USA), the subject (Afghanistan), and some code or series of codes that would indcate 'invasin'. A series of codes mght work like this: the code for invasin mght be '317', the '3' indicatng that this was a hostile act, the '1' indicating t was a military act, and the'7'indicatingin morespecificfashion invasin. Many other variables could also be coded; for example, we might code that the United Nations facilitated the act by sponsoring a Security Council Resolution, we might link in previous events such as Mullah Omar's refusal to turn in Osama bin Laden, and so forth. Events data sets, then, contain thousands or even millions of unes of code, each of which isa foreign policy'event. The acronymsof some of these events data projects live on: some because the data arestill being collected (e.g. Gerner et al. 1994) (some collection is funded by the DDIR (Data Development for International Research) Project of the NSF), and others because the data are stl useful as a testing ground for hypotheses-WEIS (the World Event/lnteraction Survey),'CORDA B (the Conflict and Peace Data Bank), CREON (Comparative Research on the Events of Nations), and so forth. The Kansas Event Data System (KEDS) is more of a second-wave effort,

Individual characteristics Would Ihere be a distinct field of foreign policy analysis wilhout this most micro of all explanatory levis? Arguabiy not. It s in the cognition and informalion processing of an aclua! human agent that all the explanatory levis of FPA are in reality ntegrated. What sets FPA apart from more mainstream IR s this nsistence that, as Hermann and Kegley put t, '[a] compelling explanation (of foreign policy) cannot treat the decider exogenously' (1994:4). Political psychology can assist us n understanding the decider. Under certain conditionshigh stress, high uncertainty, dominant position of the head of state in FPDM-the personal characteristics of the individual will become crucial in understanding foreign policy choice.

24

VAJ.ERIE M. KE5SQM

CHAPTER 1 THE HIS7O8Y AND EVQLUTION OF FOREIGN POLICV ANALVSSS

.25

The work of Harold Lasswell on political leadership wasasignificant nfluenceon many early pioneers of political psychology with reference to foreign policy (see Lasswell, 1930, 1948). Joseph de Rivera's The Psychological Dimensin of Foreign Policy (1968) is an excellent survey and integration of early attempts to apply psychologica! and social psychologica! theory to foreign policy cases. Another early effort at a systematic study of leader personality effects is the concept of operational code, an idea originating with Leites (1951), and refined and extended by one of the most mportant figures in this rea of research-Alexander George (1969). Defining an operational code involves identifying the core political beliefs of the leader about the inevitability of conflict in the world, the leader's estimation of his or her own powerto change events, and so forth, as well as an exploration of the preferred means and style of pursuing goals (see also O. Holsti 1977; Johnson 1977; Walker 1977). It should be noted that George's influence on the field is by no means confined to his work on operational codes; he has offered useful suggestions on methodological issues (see George on process tracing (George 1979), on the demerits of abstract theorizing versus actor-specific theory (George and Smoke 1974; George 1993), and on the need to bridge the gap between theory and practice in foreign policy (George 1993,1994). The work of Margaret G. Hermann is likewise an attempt to typologize leaders with specific reference to foreign policy dispositions. A psychologist by training, she was also Envolved in a CFP project (CREON). However, the core of her research is leaders' personal characteristics (Hermann 1970, 1978). Using a modified operational code framework n conjunction with content analysis, she is able to compare and contrast leaders' beliefs, motivations, decisional styles, and interpersonal styles. Furthermore, Hermann integrales this Information into a more holistic picture of the leader, who may belong to one of six distinct 'foreign policy orientations'. Orientation allows her to make more specific projections about a leader's behaviour n a variety of circumstances. In the second wave of research, scholars began to explicitly compare and contrast the findings of different personality assessment schemes (Winter et al. 1991; Singer and Hudson 1992; Snare 1992; see also Winter 1973; Rost 990). The role of perceptions and images in foreign policy was a very important research agenda n this first generation of FPA. The work of both Robert Jervis and Richard Cottam deserves special mention here.Jervis's Perception and Misperception in International Politcs (1976) and Cottam's Foreign Policy Motivaton: A General Theory and a Case Study (1977) both explcate the potentially grave consequences of misperception in foreign policy situations by explorng its roots. Deterrence strategies can fail catastrophically f misperception of the other's intentions or motivations occur (see also the stimulus-response models of Holsti et al. (1968)). Likejanis, Halperin, and others, the work of Jervis and Cottam is consciously prescriptive: both include advice and suggestions for policy makers. Work in the late 1980s continuingthistradition included scholarship byJaniceGrossStein, Richard Ned Lebow, Ole Holsti, Alexander George, Deborah Welch Larson, Betty Glad, and Stephen Walt (as well as Jervis et al. 1985; Larson 1985,1993; M. Cottam 1986; Glad 1989; George and Smoke 1989; O. Holsti 1989; Lebow and Stein 1990; Walt 1992). An excellent example of work in this period is that of Richard Herrmann (1985,1986,1993), who developed a typology of stereotypcal images with reference to Soviet perceptions (the other as 'child', as 'degenerate', etc.) and began to extend his analysis to the images held by other nations, ncluding American and Islamic images.

The work on cognitive constraints was nformed by the work of scholars in other fields, including that of Herbert Simn (1985) on bounded rationality, Heuer (1999, but written between 1978 and 1986) oncognitivebias, and Kahneman etal. (1982)on heuristic error. Many other important cognitive and psychological studies which appeared during the 1970s and early 1980s dealt with a diversity of factors: motivations of leaders (Barber 1972; Winter 1973; Etheredge 1978), cognitive maps, scripts, and schemas (Shapiro and Bonham 1973; Axelrod 1976, Carbonell 1978), cognitive style (Suedfeld and Tetlock 1977); life experience of leaders (Stewart 1977), and others. Good edited collections of the time include Hermann (1977) and Falkowski(1979).

National and societal characteristics Kal Holsti's (1970) elucidation of national role conception s'pans both the psychological and the social milieu. With this concept, Holsti seeks to capture how a nation views itself and its role n the international arena. Operationally, Holsti turns to lite perceptions of national role, arguingthatthese perceptions are arguably more salientto foreign policy choice. Perception of national role is also nfluenced by societal character, a product of the nation's socialization process. Differences here can lead to differences in national behaviour as well (e.g. Broderson 1961; Hess 1963; Merelman 1969; Renshon 1977; Bobrowef a/. 1979). The methodology of national role conception was continued n the 1980s by Walker (1987) and others (Wish 1980; Cottam and Shih 1992; Shih 1993). The study of culture asan independen! variable affecting foreign policy was just beginning to be redeveloped near the end of the 1980s after petering out n the 1960s (Almond and Verba 1963; Pye and Verba 1965). Culture might have an effect on cognition (Motokawa 1989); t might have ramifications for structuration of institutions such as bureaucracies (Sampson 1987). Conflict resolutlon techniques might be different for different cultures as well (Cushman and King 1985; Pye 1986; Gaenslen 1989). Indeed, the very processes of policy making might be stamped by one's cultural heritage and socialization (Holland 1984; Etheredge 1985; Lampin 1986; Merelman 1986; Leung 1987; Banerjee 1991; Voss and Dorsey1992). The study of the role of societal groups n foreign policy making can be seen as an outgrowth of the more advanced study of societal groups in American domestic politics. Sometimes an individual scholar used theory developed for the American case to explore the more diverse universe of the international system. For example, Robert Dahl's volume, Regimes and Oppositions (1973), provided the key theoretical concepts necessary to analyse the relationship between domestic political pressure by societal groups and foreign policy choice by the government. Other more country- and region-specific case studies were also developed: see Deutsch etal. (1967), Hellman (1969), Dallin (1969), Chittick(1970), Hughes (1978), and Ogata(1977), among others. In the late 1980s, a new wave ofthinking began to explore the limits of state autonomy in relation to other societal groups n the course of policy making. The work of Putnam (1988) on the two-level game of foreign and domestic policy was paradigmaticforestablishingthe majorquestionsofthis research subfield. Other excellent work includes Evans et al. (1985), Hagan (1987), Levy(1988), Lambornand Mumme (1989), Levy and Vakili (1989), and Mastanduno et al. (1989). A second wave of research n this rea can be seen in the work of Kaarbo (1993), Skidmore and Hudson (1993), and Van

26

VALERSE M. HUDSON

CKAPTES 1 THE HISTORY AND EVOLUTION OF FOREIGN POLICY ANALYSIS

.27;

Selle (1993) (see also Bueno de Mesqulta and Lalman (1992) for an interesting comblnatlon of game theory and FPA to understand domestic poltica! imperatives and their effect on foreign policy).

However, n the second-wave years, one notable exception to all the analysis of the previous years burst forth upon the scene-democratic peace theory. Democracies, it was noted, tend not to fight one another, though they fight non-democratic countries as often as other non-democracies do. This appeared to be an example of how a difference in polity type produced a difference in foreign policy behaviour (Russett 1993a,b). This has been a particularly interesting bridging question for FPA and IR (and is examined further in ChapterThree). Why do democracies not fight one another? Here we find more abstract theorists of war (Merritt and Zinnes 1991; Morgan 1992; Bremer 1993; Dixon 1993; Ray 1993; Maoz and Russett 1993) wrestling with a question that leads them into FPA waters and into conversaron with FPA scholars (Hagan 1994; Hermann and Kegley 1995). Finally, f it is possible to see the international system as part of the psycho-social milieu in which foreign policy decisin making takes place, then the work of much of mainstream IRat this time can be seen as contributing to the Foreign Policy Analysis research agenda. The effects of system type, as elucidated by Kaplan (1957,1972), may depend on the number of poles in the system, the distribution of power among poles, and the rules of the system game that permit its maintenance. This structure may then determine to a large extentthe range of permissible foreign policy behaviour of nations. The work of Waltz was extremely influential in its description of the effects of an anarchical world system on the behaviour of its member states (see also Hoffmanr, 1961; Rosecrance 1963; Singer e.1 a!. 1972). FPA seemed not to emphasize this type of explanation, primarily because the variation n behaviour during the time when a certain system is maintained cannot be explained by reference to system structure because the structure has not changed. Explanation of that variation must be found at lower levis of analysis, where variation in the explanations can be dentified. Here, then, is one of several sources for the notable lack of integration between actor-general systems theory in IR and FPA.

FPA self-refiection in the late 1970s and 1980s

A period of critical self-reflection n FPAbegan in the late 1970s and continued until the mid1980s. The effects were felt unevenly across FPA, with CFP being affected the most; it is here that we see the most pruning, both theoretical and methodological, which will be discussed later. In decision-making studies there was a period of rather slow growth because of methodologicai considerations. The information requirements for conducting a high-quality group or bureaucratic analysis of a foreign policy choice are tremendous. If one were not part of the group or bureaucracy in question, detailed accounts of what transpired, preferably from a variety of primary source viewpoints, would be necessary. Because of security considerations in foreign policy, such information is not usually available for many years (e.g. until declassified). The question facing decision-making scholars became; Is it possible to be theoretically and policy relevant f one s relegated to doingcase studies of events twenty years or more od? If so, how? If not, how is it possible to manoeuvre around the high data requirements to say something meaningful about more recent events? (See P. Anderson 1987.) Scholars wrestling with this issue carne up with two basic responses: (a) patterns in group/ bureaucratic processes can be isolated through historical case studies, on the basis of which both general predictions of and general recommendations for present-day foreign policy

The second-wave work of Joe Hagan desenes special note. He has compiled an extensive datbase on the fragmentation and vulnerability of political regimes, with special reference to executive/legislative structures (Hagan 1993). The set ncludes 94 regimes for 38 nations over a 10-year period. His purpose is to explore the effects of political opposition on foreign policy cholee. Using aggregate statistical analysis, Hagan is able to show, for example, that the internal fragmentation of a regime has substantially less effect on foreign policy behaviour than military or party opposition to the regime. Domestic poiitical imperatives could also beascertained byprobing lite and massopinion (again, piggy-backing onto the sophisticated voter-attitude studies of American politics). Though usually confined to studies of democratic nations (especially America, where survey research results were abundant), these analyses were used to investgate the limits of the socalled Almond-Lippman consensus-that s, that public opinin is incoherent and lacking unity on foreign policy issues, and thus that public opinin does not have a large impact on the nation's conduct of foreign policy (see Bailey 1948; Almond 1950; Lippman 1955; Campbell et al. 1964; Converse 1964; Lipset 1966). Opinin data collected during the Vietnam War period appears to have served as a catalyst to re-examine this question. Caspar/ (1970) and Achen (1975) found more stability in American public opinin concerningforeign policy and international involvement than their predecessors. Mueller (1973) used the Vietnam War to show that although the public may change their opinions on international issues, they do so for rational reasons. Holsti and Rosenau (1979) and Mandelbaum and Schneider (1979) used survey data to identify recognizable ideological positions to which the public subscribes on foreign policy issues. A large amount of research was undertaken to show that public and lite opinin does affect governmental foreign policy decisin making (see Cantril 1967; Verba et a!. 1967; Graber 1968; Verba and Brody 1970; Hughes 1978; Yankelovich 1979; Beal and Hinckley1984). The study of the effect of national attributes (size, wealth, political accountability, economic system, etc.) on foreign policy was certainly, in a theoretical sense, in the Sprout genotype, but was carried out by scholars and with methods more to be placed in the Rosenau genotype (fyou exclude pre-Rosenau writers such as Lenin). The propensity to be involved in war was usually the foreign policy dependent variable of cholee in this work (see Rummel 1972,1977,1979; Kean and McGowan 1973; East and Hermann 1974; East 1978; Salmore and Salmore 1978; for a more holistic treatment, see Korany 1986). The questions raised by these theorists are fascinating. Are large nations more likely to go to war than small nations? Are rich nations more likely to goto war than poorones? Areauthoritarian regimes more bellicose than democracies? Statistical manipulation of aggregate data, at best a blunt nstrument, was unable to uncover any lawlike generalizations on this score (though for an interesting and hard-to-classify treatment of the multilevel causes and effects of war, see Beer (1981)). Political economy research on the effects of economic structures and conditions on foreign policy choice are fairly rare: the 'culture' of PE and the 'culture' of FPA did not mix well, for reasons explored below. However, the works of Neil Richardson and Charles Kegley (e.g. Richardson and Kegley 1980) and Peter Katzenstein (e.g. Katzenstein 1985) are notable as exceptions to this generalization.

28

VALERSE M. HUDSCN

CHAPTER 1 THE HSTORY AND EVOtTSON OF FOREiGN POLCY ANALYSIS

29

decisin making can be made; (b) innovative at-a-distance indicators of closed group/bureaucracy process can be developed, which allow for more specific explanation/prediction of resultan! foreign policy choice. FPA work at the psychological level actually expanded during this tirne period, bui vvork at the societal level arguably contracted on some research fronts. The reason for this bifurcation in the genotype was methodological: psychology provided ready-made and effective tools forthestudy of political psychology, but political sciencedid notoffer the foreign policy analyst the same advantage. To understand how the broader socio-cultural-political context within a nation-state contributes to its governmental policy making (whether domestic or foreign) is, perforce, the domain of comparative politics. It is hopefully not controversial to aver that the theories and methods of comparative politics in this period were not quite as highly developed as those of psychology. The attempt to graft 'scientific' statistical analyses of variance onto the underdeveloped theory of comparative politics of the 1970s and 1980s was a failure. More successful were efforts to spin existing comparative politics work on a particular nation to the cause of explainingfactors that contribute to that nation's foreign policyfor example, borrowing techniques from American politics (such as public opinin surveys) to study domestic political imperatives in the USA on foreign policy issues. Still missing were the conceptual and methodological tools necessary to push past the artificial barrier between comparative poiitics and nternationalreiations that stymied theory development. One of the greatest leaps forward in the present period of FPA s the innovative work begun on conceptualzingthe'two-ievelgame'(Putnam 1988). As mentioned, CFP dwindled in the 1980s. Indeed, the very term 'comparative foreign policy'began to sound quaintand naive. Membershipof the Comparative Foreign Policy Section of the International Studies Association plummeted. Public vivisections took place, while Rosenau-genotype-style scholarship became scarce. Both sympathetic and unsympathetic criticism abounded (e.g. Ashley 1976,1987; Munton 1976; East 1978; Kegley 1980; Caporaso et al. 1987; Hermann and Peacock 1987; Smith 1987). At one point, n exasperation, Kegley (1980: 12), himself a CFPer, chides, 'CFP risks being labelled a cult of methodological flagellomaniacs'. This searing criticism and self-criticism revealed a number of inconsistencies in the CFP approach, which needed to be sorted out before any progress could be contemplated. The stumbling blocks included the following: 1. You cart have your parsimony and eat it, too. The tensin between the desir'e of some CFPers for a hard-science-like grand unified theory and the assumption that micro-level detail s necessary f one really wants to explain and predict foreign policy behaviour became unbearable. Rosenau's 'Pre-theories' article (Rosenau 1966), when reviewed from this vantage point, sets the genotype up for an inevitable dilemma about parsimony. To what should we aspire: richly detallad, comprehensively researched micro-analyses of afew cases, or conceptually abstract, parsimonious statistico-mathematical renderlngs of thousands of events? One can seethe problem n desiring richly detailed, comprehensively researched micro-analyses of thousands of events: a lifetime would be over before a theorist had collected enough datato do the first big 'run'! But many CFPers rejected the case study approach as unscientific and too much like the soft ancdota! research of the'traditionalists'(Kegley 1980).

CFPers wanted to be behaviouralists and to be scientific, and a hallmark of this was aggregate empirical testing of cross-nationally applicable generalizations across large vales ofN.At the same time, they were fiercely committed to unpacking the black box of decisin making, so the detai! of their explanans grew, and with it their rejection of knee-jerk idealization of parsimony. Push had to come to shove at some point: CFP methods demanded parsimony in theory; CFP theory demanded nuance and detail in method. 2. To quantify or not to quantijy? A corollary of large-N-size testing s the need for more precise measurementof data; indeed, quantification of variables is essential to linear regression and correlation techniques, as well as to mathematical manipulations such as differential equations. However, the ndependent variables of CFP induded such non-quantifiables as perception, memory, emotion, culture, history, etc., all placed n a dynamic and evolvingstream of human action and reaction that might not be adequately captured byarithmetic-based relationships. To leavesuch non-quantifiable explanatory variables out seems to defeat the very purpose of micro-analysis; to leave them in by forcing the data into quasi-interval level pigeonholes seems to do violence to the substance that CFP sought to capture. CFPers began to ask whether their methods were aiding them inachieving their theoreticalgoalsorpreventing them from everachieving those goals. 3. A final nconsstency centred n policy relevance. As mentioned earlier, CFP had received a large amount of money from the government to crate events data sets. CFP researchers successfully argued that such an nvestment wouid yield information of use to foreign policy makers. Specifically, events data would be used to set up early warning systems that would alert policy makers to crises in the making around the world (as if they do not also read the same sources from which events data come!). Computerized decisin aids and analysis packages with telltale acronyms began to appear EWAMS (Early Warning and Monitoring System); CASCON (Computer-Aided Systems for Handling Information on Local Conflicts); CACIS (Computer-Aided Conflict Information System); XAIDS (Crisis Management Executive Decisin Aids) (see Andriole and Hopple 1981). Unfortunately, these could never live up to their promise: the collected events could be had from other sources and so were nothing without the theory to explain and predict their occurrence. The methodological paradoxes explicated above resulted n theory that was stuck, by and large, at the level of globally applicable but specifically vacuous bivariate generalizations such as that large nations particpate more n international interactions than small nations' (see McGowan and Shapiro 1973). Again, CFP found itself pulled in two opposed directions: Was the research goal to say something predictive about a specific nation at a specific time in a specific set of circumstances (which would be highly policy relevant, but which might closely resemble the output of a traditional country expert)? Or was the goal a grand unified theory (which would not be very policy relevant, but which would qualify you as a sdentist and a generalist)? Attempts to accomplish both with the same research led to producs that were unsatisfactory in a scholarly as well as a policy sense. Hindsight s always 20/20; it does seem clear n retrospect that change was necessary. Left behind were (1) the aim of a grand unified theory, and (2) the methodological straitjacket mposed by the requirementof aggregate empirical testing. In 1980, Kegley spokeof the need

VALERIE M. HUDSON

CHAPTES 1 THE HISTORY AND EVOLUTiON OF FORE1GM POLICY ANALYSiS

to come down from the rarified air of grand theory to middle-range theory, and to capture more of the particular:

s a commitment to build middle-range theory as the nterface between actor-general theory and the complexity ofthe real world; s a commitment to pursue multicausal explanations spanning mltiple levis of analysis; s a commitment to utilize theory and findings from across the spectrum of social science; o a commitment to viewing the process of foreign policy decisin making as important as the output thereof.

To succeed partially s not to fail completely. . . . Goals (should be) downgraded to better fit capadties.... This prescribes reduction in the level of generality sought, so that more contextually-qualified, circumstantially bounded, andtemporally/spatially-specified propositions are tested. More of the peculiar, unique, and particular can be captured at a reduced level of abstraction and generality. (Kegley 1980:12,19) To be fair, this was arguably Rosenau's original aim, and the CFP community had to reach a consensustoreturntoitsfounding visin.The conferenceon NewDirectionsintheStudyof Foreign Policy, held at Ohio State University in May 1985, probably represents a finalization of these changes for the CFP group (see the resulting voiume, Hermann et al. 1987; see also Cerner 1992).

Conclusin: contemporary FPA's research agenda

Nevertheless, FPA has evolved n the sophistication of the questions asked, and n the means of answering those questions. Indeed, FPA's ability to ask new questions is perhaps more promising in relation to itsfuture theoretical potential than any other indicator. Einstein and Infeld (1938) commented that, '(t)he formulation of a problem is often more essential than itssolution which may be merely a matterof K skill. To raise new questions, newpossibilities, to regard od problems from a new angle, requires creative imagination and works real advance in science'. ln order to see this advance, let us examine some of the new questions that have evolved fromtheold.AsadetailedoverviewofFPAscholarshipfrom1993tothepresentcanbefound n Hudson (2005,2007), let us nowturn tothe micro-levels of analysis and then movetoward rnacro-levels.

As FPA was being liberated from its inconsistencies in the late 1980s, the world was being liberated from the chess match of the Cold War. This was a felicitous coincidence for FPA, and was an added source of vigourfor its research agenda. The significance of this temporal coincidence can be understood by remembenngwhattypes of IR theory were in ascendanceat the time: neorealistsystemsstructure theory and rationalchoicemodelling. Indeed, the dominance was so overwhelming that on taklng an IR theory course during this time, one would think these two were the summum bonum of all thinkingin international relations (at least n the USA). This state of affairs was natural for American thinkers: America was one of two poles of power n the Cold War international system. A bipolar quasi-zero-sum rivalry lends itself relatively well to abstract actor-general analysis focused primarily on the macro-constraints mposed by the system. Furthermore, actor-general theory was more practical for scholars during the Cold War, becauseso little was known ofthe blackboxof theclosed Soviet, Chnese, and Eastern Bloc foreign policy decisin making bodies. However, when the bipolar system collapsed with the fall ofthe Soviet bloc regimes, an mportant theoretical discovery was made: t s mposs'tble to explain or predict system change on the basis of system-level variables alone. Along the same unes, in a period of great uncertainty and flux, lack of empirically grounded inputs to rational choice equations is deadly in terms ofthe usefulness of such analysis. Our ntuitive understanding ofthe collapse involves variables more to be found in FPA: the personalities of Gorbachev, Havel, and Walesa; the activities of transnational groups such as the Lutheran Church and the Creen Movement; the struggles between various domestic political players, such as the military, the Communist Party, the bureaucrats, etc.;the roleof economics and societal needs n sparkingthe desirefor change, etc. With the fall ofthe Iron Curtain, the need for an 'actor-specific' complement to mainstream IR theory became stark in its clarity. FPA n the post Cold War era retains the distinctive theoretical commitments that demarcated at ts inceptlon. Included among these are: e a commitment to look below the nation-state level of analysis to actor-specific information;

New questions
When studying the effects of individual leaders on foreign policy decisin making, the key question s whether we can extend our understanding of how a leader's personality affects foreign policy through determining its effect on choice of advisors, preference for issues, preference for certaln group processes, and so forth? Moreover, can we intgrate different analytical schemes for analysing leader personality and its effects? What are the ramifications of new breakthroughs in neuroscience for FPA? How do various leader personality types shapethe structure and process of groups servingthem? At the group level, we mustthen ask how problems are actuallyrecognized by the group? How are situations 'framed' and 'represented'? How are options developed? How does a group come to share an interpretation ofthe situation? How does a group change an established interpretation? How does a group learn? How is the group's potential for creativity enhanced or dampened? How does group memory affect group action? How do groups become players n the 'two-level game'? How are group structure and process a function of societal culture? At the level of society and political competition, we explore whether we can uncover the societal sources of change in shared perceptions? For example, how do altitudes of leaders and publics change as context changes? Can national role conception be re-configured to serve as the theoretical nterface between a society and the individual members of that society who come to lead it and make ts foreign policy dedsions? Can we specify the effect on foreign policy of domestic political competition? Can we complete the theoretical arele and specify the effects on domestic politics ofthe implementation of a certain foreign policy choice? How can we discern culture's nfluence on foreign policy? Does type of political system impact on foreign policy? What s the effect of systemic change on foreign policy?

32

VALERSE M. HUDSON

CHAPTER 1 THE HiSTORY AMO EVOLUTION OF FOREiGE--! POLCY ANALYSIS

33

Methodologically speaking, there are just as many key questions to be considered. These nclude: Can evnts data be re-conceptualzed to be of use to contemporary FPA? Can FPA utilize methods created to simlate human decisin makingas a means of integratingcomplex non-quantifiable data? Can we think of non-arithmetic ways to relate variables? Can rational choice models be altered to accommodate actor-specific idiosyncracies with regard to utility, choice mechanisms, and choice constraints? Can we crate models that will allow us to use as inputs the actor-specific knowledge generated by country/region experts? When s the detail of actor-specifics necessary, and when is actor-general theory sufficient to explain and project FPDM? How could one nstantiate a model of the 'two-level game'? Can discursive analysis or interpretivism be used to introduce the dynamics of evolving understanding n FPDM?

An Atlantic divide?
While there are new efforts to both catalogue and promote the anaiytical study of foreign policy n the Global South (Brummer 2011; Giacalone 2011; Zhang 2011), atthis point such study is predominantly of Atlantic origins. This raises the question of whether there are mportant differences n the way such studies are conducted n the USA compared with European countries, to which an affirmative answer can be given. In a recent overview of such differences (Hadfield and Hudson 2011), the authors point out several distinctions between FPA (American) and what they term AFP (the analysis of foreign policy, European). This issue is of nterest to the readers of this volume, for about half of its authors are American and about half are European. Hadfield and Hudson note a greater emphasis on cognitively oriented theories n FPA than in AFP, as well as the more frequent use of quantitative methods by Americans and historical process-tracing by Europeans. They also note a clear preference for the use of American cases by American scholars, which, while not unexpected and also understandable, also bears predictable consequences for theory-building. For example, Zhang (2011) finds that in the Chnese case, being 'ideological' and being 'practica!' are not oxymoronic terms, whereas in American-inspired theory, such personality orientations are seen as precluding one another. Hadfield and Hudson also note a greater sense of community among American FPA scholars, which they attribute to the small number of gradate programmes training FPA scholars in that country; they are all likey to know one another within one or two degrees of separation. Turning to the European context, there are almost no gradate programmes that emphasize FPA/AFP, and so as yet scholars have not been able to crate a critical social mass that s the prerequisite for anepistemic community. Finally, Hadfield and Hudson suggest that AFP isfar more theoretically inclusive than FPA has been to date, embracing not only actor-specific theories, but also grand theory and constructivist approaches. For example, one is not likey to see a reference to the work of Roy Bhaskar in American FPA literature, while such a reference might be very likely n AFP work.

Neoclassical realism attempts to cross the divide from the IR side to the FPA side, while behavioural IR s moving in the opposite direction. Neoclassical realism is premised on the understanding that 'unit-level variables constrain or faciltate the ability of all types of states-great powers as well as lesser states-to respond to systemic mperatives' (Lobell et al. 2009: 4). While the power distribution within a system may bound grand strategy, the implementation ofthisstrategythrough decisions by foreign policy executivesconcerningthreatassessment, risk, and mobilizationof domesticresources, including public support, simply cannot be inferred from the 'grand' level. Historical processtracing s the preferred methodology of this school. Behavioral IR, on the other hand, takes as its touchstone thefield of'behavioural economics', associated with the work of such scholars as Richard Thaler. Walker and his co-authors define the approach as 'a social-psychological analysis of world politics, which emplys a general systems theory [i.e. role theory-ed.] to unify the understanding of actors, actions, and relations that constitute foreign policy and international.reiations' (Walker et al. 2011: 5). In other words, behavioural IR seeks to intgrate the external world of evnts with the internal world of beliefs by examining strategic moves within dyads. The operational codes of the two actors are deciphered by quantitative content analysis of leader texts, and then a theory of strategic game moves (TOM) is employed to determine (at an abstract level) the next move that each actor in the dyad will take and how their game will resolve over sequential moves. The use of a game-theoretic logic informed by unit-level characteristics is an interesting amalgam of actor-general and actor-specific theorizing which, like neoclassical realism, attempts to bridge the divide between FPA and IR. Noteworthy, however, s the fact that neoclassical realism and behavioural IR-research programmes with the very same bridging goalare working in such divergent methodological traditions that one wonders if they could possibly communicate with one another. But that is a topic for the future. These are all exciting new questions, issues, and approaches to be exploring. Doubtless someofyou will be nvolved n this work. Itisawonderfultimeto become engaged in FPA-a time of new horizons.

Key points
9 Foreign policy analysis takes as its theoretica! ground the human decisin makers, actingsingly and in groups.who make foreign policy. 9 Three paradigmatic works laid the foundation of FPARichard Snyder and colleagues on decisin making, James Rosenau on comparative foreign policy, and Haroldand Margare! Sprouton the psycho-social milieu of foreign policy decisin making. Several emphases, correspondingto levis of analysis in FPA, began to emerge from this foundation, including work on small/large groups, evnts data, political psychology of leaders, cultural effects on foreign pollcy, the effects of domestic political contestation on FPDM, and the influence of national attributes and systemic characteristics on FPB. 0 FPA retains its emphases on actor-specific theory, multicausal explanations, interdisciplinarity, and the explanations of foreign policy processes, as well as foreign policy outcomes.

New attempts to bridge IR and FPA

Also fairly recent in origin are sustained organized attempts to bridge the divide between FPA and IR. The two that we will mention in this section are neoclassical realism (exemplified by Lobell et al. 2009) and behavioural IR (exemplified by Walker et al. 2011; see also Mintz 2007).

<s Current FPA scholarship explores linkages between the levis of FPA analysis, and combines that with a search for new methodologies that are more appropriate for actor-specific theoretical investigation.