Abstract
This article has a dual purpose. First, it looks at the transfer of the methodology of systems analysis from the RAND Corporation to the International Institute for Applied Systems Analysis (IIASA) in the wake of an East–West bridge-building effort during the Cold War. Second, it draws out a more general argument about how the institutional structures of these research organizations condition their methodological orientations. Acknowledging the complexity of factors influencing methodological choices at RAND and IIASA, the article concentrates on the centrality of institutional purpose, institutional environments and internal organizational structure, and demonstrates how, when taken together, these factors led to a methodological diversification at IIASA that is best summarized as the internationalization of systems analysis.
Introduction
Systems analysis (SA) is a showcase example of a post-Second World War ‘techno-social science’, which sought to improve policy procedures by detaching decision-making from the flawed human decision-maker, instead relying on ‘rational’ decision techniques (Heyck, 2012). It has its roots in American military research during the Second World War and into the early Cold War, which shaped this methodology’s intellectual character both culturally and institutionally (Jardini, 2000). It is based on a specific understanding of rationality oriented towards algorithmic procedures that can, potentially, be automated (Erickson et al., 2013). In line with the overall thrust of the ‘rationalization’ of politics, early SA saw opposition in any form of experience-based intuitive decision-making as well as in those psychological or social effects impeding genuine rationality.
Despite these principles, the nature of systems analysis has remained intriguingly opaque. Its proponents did not produce any stable definition, not of basic concepts, such as rationality, nor of the concrete methods and techniques required for doing SA. Moreover, it seems that this opacity – or rather the openness it engendered – partly explains why, after its invention at the RAND Corporation in the 1940s and 1950s, SA has traveled to several other contexts, from government to social policy to industry and management and from the USA to other countries and international organizations. It is unsurprising that each of these transfers confronted systems analysts with different types of problem to which they sought different solutions. This article reconstructs one such transfer of SA from RAND to the International Institute for Applied Systems Analysis (IIASA), an internationally financed research institute where scientists from over a dozen countries, including representatives from both the Western and the eastern blocs of the Cold War, were expected to cooperate on issues of global relevance. In particular, the article explores the intellectual mutations of SA in the course of this contextual transfer.
The article will continue on the assumption that the institutional set-up of a research organization influences the methodological style of its research. In order to explore the relation between institutions and methodologies, I will proceed in three steps.
First, the invention of SA at the RAND Corporation will be briefly reviewed in order to establish the ideal-typical version of SA characterized by the largely unabated beliefs of scientists in rational human behavior, predictability and planning through formal quantitative methods against which various extensions and critiques become visible. Within the institutional structure of RAND, I will examine: (1) its institutional purpose or the type of problems it was concerned with; (2) the institutional environment in which the problems appear and through which solutions could have been sought; and (3) its internal organizational structure. With regard to the methodological orientation of systems analysis, I will concentrate on the simple dichotomy between quantitative and qualitative methods, while also distinguishing between explicit and implicit usages of both.
My second step concerns the historical background of the transfer of SA from RAND to IIASA, which will be outlined.
Third, and finally, I explore the same analytical questions used to describe the connection of institutional structure and methodological orientation of SA at RAND, which I will also address for IIASA.
My central argument is that in both research institutions there is a connection between certain institutional features and methodological preferences. To put it bluntly, the military context of RAND and the global problem-solving context of IIASA engendered different kinds of thinking that become most visible in their methodological debates. Although two institutions are portrayed, the evidence for this connection is not a comparative argument but rather processual, which is highlighted here for two connected historical instances.
There is a large literature on the RAND Corporation. Literature on IIASA is still small despite some important recent publications (Rindzevičiūtė, 2010, 2015, 2016; Schwartz, 2014). Thus in what follows, analysis of RAND will be concise and rely only on secondary sources. IIASA is the focus of this study, using published studies, original archival research and insights from interviews carried out by the author.
It is now possible to be more precise about the dual purpose pursued in this article. It both tells the story of the transfer of SA from RAND to IIASA and introduces the latter as a case for transnational exchange of scientific ideas across the East–West divide to the history and sociology of the social sciences in the Cold War. It also consolidates a more general argument about the institutional conditions of methodological change, asking specifically about the impact of the political context in which RAND and IIASA decided on conceptual and methodological issues. The article is thus intended as a contribution to explanatory approaches in historical sociology of science (Camic, Gross and Lamont, 2011; Camic and Gross, 2001).
The origins of systems analysis at RAND
RAND was founded in 1946 as Project RAND by the US Air Force and the Douglas Aircraft Company. It was an organization with which to continue the successful engagement of scientists in the service of military and government agencies in the Second World War. 1 The centerpiece of the wider political and military organization was that RAND was established as a not-for-profit research institute, being entirely financially dependent on the US Air Force, but also with privileged access to the military and governmental administration and information.
RAND’s research environment was consciously created to encourage creativity: ‘thinking out of the box’, the free flow of ideas, and interdisciplinary exchange between scholars from different academic backgrounds. United by the common objective to develop a comprehensive ‘science of warfare’ (Hounshell, 2000: 257), which includes research and development of weapons systems and political strategy, as well as all those organizational problems related with military and public administration, the economy of armament and technological development and the psychology of warfare, RAND is often portrayed as the prototypical interdisciplinary research organization of the postwar era. However, as Daniel Bessner has shown, the degree of integration of conflicting research philosophies had limitations when it came to the social sciences. The vast majority of RAND scholars were from the natural sciences. During the institution’s first 15 years, by and large they adhered to a vision of quantification in all aspects of their studies, ignoring concepts such as politics, culture, or human behavior that are palpable only through qualitative methods (Bessner, 2015: 32). While RAND’s Social Science Division (SSD) actively pursued qualitative methods, the philosophical differences proved to be too deep to engender a truly interdisciplinary integration of the two perspectives. 2
Systems analysis is a case in point. It is usually referred to as an interdisciplinary problem solving methodology, in which policy-makers are provided with more or less exact outlooks on the consequences of alternative scenarios depending on their actions. The problems to which systems analysts seek solutions are typically of a complex nature, involving high degrees of uncertainty. The intertwining of technical, natural and social elements within the problems studied requires knowledge and skills from various disciplines. 3 The core of the systems analytical approach, then, consists of the integration of these inputs through formal models. Therefore, the use of a formal language, strong reliance on quantitative analyses and the use of computer-based automation are typical characteristics that tend to include an aversion towards qualitatively oriented social science methods and concepts.
Beyond these very general characteristics, definitions of systems analysis usually remain vague and contain considerable ambiguities regarding its methodological preferences.
‘Systems analysis is little more than systematic thinking mixed with some attempt at problem quantification’ (Sapolsky, 2003: 72). Roger Levien was systems analyst at RAND and second director of IIASA (1975–81); his reaction to the frequent inquiries as to what system analysis really was is most telling. In his office at IIASA, he had a framed quotation from Lewis Carroll’s Through the Looking-Glass, which read: ‘“When I use a word”, Humpty Dumpty said, in a rather scornful tone, “it means just what I choose it mean – neither more nor less”’ (Hordijk, 2007: 14).
What Leen Hordijk, himself director of IIASA in 2002–8, highlights with this tale is that systems analysis has always defined itself as both science and art, or, in the words of others, emphasized the ‘craft’ aspects of this activity (Majone, 1980). For the historian of science, it follows that changes in the methodological preferences in systems analysis become a question of empirical interest.
Although the thrust of systems analysis was essentially to rationalize complex long-range strategic decisions through the use of exact scientific methods, its proponents and practitioners were very much aware of its limitations. As Charles Hitch candidly declared in a 1955 RAND paper, the problems systems analysts study ‘are not particularly appropriate for scientific method and would never be selected for the application of scientific method by a truly “academic” researcher’ (Hitch, 1973: 20). While the goal of systems analysis was clearly to drive back decision-making based on the intuition and judgement of the decider, its proponents openly admitted that they often had to resort to their own intuition and judgement in their analyses. There is, however, still an important difference between conceding the limitations of scientific method for practical decision-making – which makes total eradication of intuition and judgement impossible – and the conviction that the study of military strategy must genuinely address these ambiguous elements of decision through the qualitative methods of the social sciences.
Further characteristics of RAND’s early systems analysis can be highlighted here. First, the idea of goal variables is central to what is called the ‘rationalization of administrative processes’ and the role of quantitative analyses. Once it is possible to express a policy goal in a single quantifiable variable, defining the optimal solution becomes a matter of applying the mathematical techniques of operations research, while implementing it becomes a question of persuasion. Second, because SA was developed with the claim to be of help in real policy situations, systems analysts have to deal with the concrete institutional context in which a problem is located, including material, cultural and social aspects. Though systems analyses have placed much emphasis on this in both phases of analysis and implementation, the actual process of dissemination has often turned out to be very complicated. Third, RAND’s analysts, although civilians, had an intimate understanding of how the US Air Force functioned as an organization and had privileged access to information and people (Levien, 2000: 444). Therefore, the principally mathematically trained analysts could handle the analysis of the social systems elements in a rather intuitive way. Within the explicit methodological work, primary attention was directed at statistical techniques, linear and dynamic programming, game theory and econometrics, while social and institutional analyses were typically assigned to the ‘craft’ part of systems analysis. Social and psychological aspects of decision-making – in particular, experience-based decisions, intuition, rivalries, idle routines and distorting group effects – were by no means ignored. On the contrary, these were precisely the problems against which systems analysis stepped up and which it sought to replace by rigid analysis. The sociality of decision-making, in other words, was essentially seen as a source of disturbance for determining what would be a rational course of action.
Fourth and lastly, the general strategy to solve complex policy issues thus consisted of: (1) determining the optimal solution by use of analytical quantitative methods; and (2) designing hierarchical institutions that would enable the enlightened decision-maker to implement the solution against the cumulative resistance from those bound in their vortices of habits and self-interest. The essence of systems analysis management can therefore be seen in the centralization of power, intimately linked with the hierarchical structure and the culture of giving and taking orders in the military institutions from where it originated (Jardini, 2000: 341). With the methodology coming to maturity, analysis of social institutions was increasingly neglected to the point where familiarity with the situation ‘on the ground’ was considered unnecessary or even unfavorable. The understanding of the problem by the actors is itself part of the problem. Rational analysis had to maintain a critical distance from them and rely as much as possible on non-corruptible methods and procedures. 4
The transfer of systems analysis from RAND to IIASA
Systems analysis has been applied on various occasions in American governance since the early 1960s. The most well-known example is the ‘McNamara revolution’ in the Department of Defense after McNamara’s hiring of Charles J. Hitch as Assistant Secretary of Defense (Jardini, 2000: 326). Hitch was head of RAND’s economic division and co-author of the highly influential The Economics of Defense in the Nuclear Age (C. J. Hitch and McKean, 1960). Following this appointment, systems analysis was applied to non-military affairs, like urban planning and social policy programs (Jardini, 2000; Light, 2005). Because of its association with scientism and social engineering, SA disappeared from the social sciences by the 1970s, after which its principal proponents came from engineering. One of the main criticisms of systems analysis is the inadequacy of the inherently hierarchical approach for social policy situations (Boguslaw, 1965, 1981; Hoos, 1984).
I will now turn my attention to the second major topic of this article; the transfer of systems analysis regarding IIASA. In October 1972, an international research institute came into being where scientists from 12 (later 17) countries from the eastern and the Western blocs of the Cold War cooperatively applied and developed systems analysis as a research perspective on global change: the International Institute for Applied Systems Analysis (IIASA) located in Laxenburg, Austria.
The first push for this institute came from US President Lyndon B. Johnson late in 1966, who sought to advance his bridge-building initiatives with the eastern bloc by assessing the possibility of a scientific East–West research center. Johnson assigned his former National Security Advisor – and by then president of the Ford Foundation – McGeorge Bundy (1919–96), to lead the investigations. With the close involvement of the RAND Corporation, Bundy quickly took the issue to Moscow, where he and Roger Levien – a RAND systems analyst and expert on Soviet cybernetics – presented their proposal in March 1967 to Moscow officials (Riska-Campbell, 2011: 130).
A key figure on the Soviet side was Dzhermen Gvishiani (1928–2003), an expert on Western sociology of organizations and American management science in particular, vice-chairman of the Soviet State Committee on Science and Technology and son-in-law of the Soviet Prime Minister Aleksey Kosygin. Gvishiani was considered to be one of the ‘Americanizers’ in the Soviet administration and was immediately receptive of the idea. Indeed, for several years he was one of the central figures in the intensifying exchange on management education between the Soviet Union and the United States (Riska-Campbell, 2011: 58).
A key challenge for the negotiators was to build an institute that could work in an environment where it was protected enough not to fall prey to political turbulences but yet be scientifically and politically relevant enough to justify the high financial and political investment necessary for such a strenuous diplomatic effort. This tension was felt throughout IIASA’s Cold War years and translated into a continued attempt to balance methodological and applied work. As we will see, this also affected the disciplinary composition of IIASA’s staff and therefore the methodological orientation of the institute as a whole.
Why was systems analysis chosen as the central plank on which collaboration between scientists from East and West should be built? Howard Raiffa (1924–2016), pioneer in the field of decision science from Harvard and first director of IIASA from 1972 to 1975, recalls the decision to call the East–West institute the International Institute for Applied Systems Analysis as a viable compromise because not many people knew what Applied Systems Analysis was supposed to mean (Raiffa, 2011: 94). 5
Leena Riska-Campbell’s work clearly shows, however, that systems analysis was on Bundy’s agenda from the outset of the negotiations. One reason may well have been that Bundy and some analysts at RAND, who were commissioned to plan the new institute’s structure in early 1967 (Riska-Campbell, 2011: 127), thought of it as a promising approach. Finally, US foreign policy institutions have had a pronounced interest in Soviet cybernetics, which was seen by many as quasi-identical with what was called systems analysis in the USA.
Slava Gerovitch reports on fears in US intelligence and policy circles during the 1960s that the Soviet Union might have made significant advances in scientific management through cybernetics that potentially gave it a significant competitive advantage over the USA (Gerovitch, 2009). John J. Ford headed a study group on Soviet cybernetics put together by the CIA in 1962. He gave a lecture on the issue to Attorney-General Robert Kennedy in the home of Secretary of Defense Robert McNamara on 15 October 1962, which was interrupted by the message that the Soviet Union had placed nuclear weapons in Cuba. The worries about an imminent cybernetic revolution in the Soviet Union were taken seriously enough that, immediately after the Cuban Missile Crisis was overcome, President John F. Kennedy asked his scientific advisor, Jerome Wiesner, to put together a panel of experts on the matter. The panel included, among others, two later Nobel laureates in economics, Kenneth Arrow and the Moscow-born Leonid Hurwicz, who left Russia after the 1917 revolution and moved to the USA at the beginning of the Second World War. The panel met a few times in 1963 and was generally less impressed by the prospects of cybernetic programming than the warnings of the CIA suggested (Riska-Campbell, 2011: 65). The CIA, however, kept spreading visions of a potential leap in rationalizing social and economic planning in the Soviet Union and urged government and military officials to pay close attention to what was happening there (Gerovitch, 2009: 51).
RAND, too, was concerned with cybernetics. Roger Levien, mentioned above, who was a senior systems analyst and later the second director of IIASA (1975–81), wrote a RAND report on Soviet cybernetics in 1964 (Maron and Levien, 1964). In this report, he highlights a necessity of closely observing Soviet progress in cybernetics for its vast topical scope in civil and military domains, its high technical quality and, most importantly, the potential ideological implications that the import of universalistic, scientized thinking might entail for the Soviet bureaucracy. Reviewing philosophical debates in the Soviet Union about the contradictions between the cybernetic approach and Marxism–Leninism fueled hopes that embracing cybernetics might eventually lead to an ideological erosion of the Soviet system (ibid.: 24).
From the Soviet perspective, interest in further developing these techniques was naturally high, given the fact that cybernetics and automation have been increasingly embraced by the highest ranks of politics. By 1957, the Soviet Academy of Science had already emphasized the importance of computer technology as having ‘an absolutely exceptional significance’ (Gerovitch, 2008: 337). In 1964, the USSR’s Council of Ministers decided to develop a unified ‘automatic system of management’ (Rindzevičiūtė, 2010: 303).
IIASA’s institutional structure
While general agreement between the USA and the Soviet Union to engage in the endeavor of a common research institute was quickly reached, the institute’s concrete outlook was the matter of a prolonged negotiation process. 6 Only the most relevant outcomes of this process are reviewed here. First, IIASA was founded as a non-governmental institute, largely to avoid signing an international treaty with the German Democratic Republic (GDR), whose statehood was not recognized by the Federal Republic of Germany (FRG) and its allies. 7 Members of IIASA were not governments but scientific organizations like the national academies of sciences or similar bodies. However, governments financed these National Member Organizations (NMOs), which is why IIASA might best be described as a semi-governmental institution.
The second cornerstone was multilateralism. Only through the involvement of a larger number of countries could (1) political stability be achieved and (2) the pool for qualified scientists be large enough to achieve the scientific excellence which was considered crucial for an institution of this kind both to gain and to maintain research credibility. When the IIASA Charter was signed on 4 October 1972, the institute consisted of 12 member countries: Bulgaria, Canada, Czechoslovakia, the Federal Republic of Germany, France, the German Democratic Republic, Italy, Japan, Poland, the Soviet Union, the United Kingdom and the United States of America. Within a few years, Austria (1973), Hungary (1974), Sweden and Finland (both 1976), and the Netherlands (1977) had joined, making 17 member countries, 6 of which were members of the Warsaw Pact, and 7 of NATO, with 3 having neutral status.
IIASA was to have a director responsible for the daily operational business. The director had a far-reaching profile that included control of the recruitment of scientists and many content-related decisions regarding the scientific work itself. It was informally agreed that the director would be assigned by the USA. Additionally, a council was established as …the governing body of the Institute and shall be responsible for establishing relations with governments and multi-national bodies; for determining financial and managerial policies, and subject areas for research of the Institute; and for ensuring that the activities of the Institute are in line with its objectives, the provisions of this Charter and the interest of member institutions. (International Institute for Applied Systems Analysis, 1972: 4)
Composed of representatives from all NMOs, the council usually came together twice a year and, in addition to its other duties, was responsible for appointing the institute’s director and deputy director. Although not specified in the charter, it was agreed that the council be chaired by a Soviet representative – a role filled by Gvishiani until 1987.
The most controversial debate regarding IIASA’s scientific orientation revolved around whether the institute would be primarily concerned with abstract methodological work – a position taken by Gvishiani and the Soviet team – or with applied projects of direct policy impact – most vigorously demanded by Sir Solly Zuckerman who had led the negotiations for the United Kingdom. This issue remained unsettled; the compromise consisted of the simple statement that both perspectives would be pursued. The best characterization of IIASA’s research agenda is probably its focus on two kinds of problem that were called ‘global’ and ‘universal’ (Levien, 2000: 454). The former – global problems – are those that transcend national borders and thus require to be handled by a number of nations in mutual coordination. Among the most important issues on which IIASA worked within this class of problems are: air pollution; acid rain; and the management of large international rivers. Universal problems, on the other hand, are problems that can be analysed and solved within national borders, but that appear in a similar way in every industrialized society. Examples include health management, waste management and urban planning.
To summarize, IIASA was established as a residential scientific research institute, where scholars from over a dozen countries from East and West cooperatively engaged in research projects under the roof of an 18th-century Habsburg castle in Lower Austria where in the past central European high aristocrats had enjoyed hunting. Both the initial drive that led to its foundation and its institutional design were set up to foster political rapprochement between the opposing blocs of the Cold War during a phase of détente. A common denominator for this cooperation was found in the institute’s methodological commitment to systems analysis. As well as being a diplomatic tool, the methodology had to bear fruit in new domains of cooperative problem-solving on a transnational scale. This radical shift of context would eventually lead to a number of methodological adaptations of SA, which will be analysed in what follows.
Institutional features and methodological change
Returning to the institutional features identified above as relevant in order for systems analysis to be developed in the form it had at RAND, we can now ask similar questions about IIASA’s institutional features and their consequences for IIASA’s methodological orientation. As before, we will consider: (1) IIASA’s institutional purpose visible through the type of research problems it was predominantly concerned with; (2) the institutional environments in which the problems appear and through which solutions can be sought; and (3) the disciplinary and national structure of IIASA’s staff. As will become clear, none of these factors alone is sufficient to cause sustainable methodological change. Taken together, however, they amounted to a particular constellation that led to a diversification of the methodological foundations of SA.
Institutional purpose: Multi-objective decisions
RAND was concerned with policy situations in which a client – in most cases the US Air Force – pursued a clearly defined goal. The intellectual challenge, then, was to reformulate the political goal into a single quantifiable variable, often called the ‘index of performance’; i.e. a utility measure consisting of several indices that depict the various objectives that are followed in a given system (Gibson, Scherer and Gibson, 2007: 40). Methodological emphasis was thus placed onto the techniques that allow the analyst to link together the various, logically diverse variables and build an integrated model that allows for meaningful conclusions in the real policy situation.
At IIASA, an important difference lay in its very nature as an international multilateral institute. Because of this, it usually did not have a single client who defined its goals, but its institutional purpose was to touch primarily on policy problems of international, and especially transnational, scale. Using only two paradigmatic examples – air pollution and the management of large international rivers – it becomes obvious that the process of goal definition is complex because it always includes various and, most importantly, conflicting interests. In other words, these decision situations included multiple and often contradicting objectives that had to be reconciled. While maximization of a given goal variable appears as the logical mode in RAND’s context, IIASA more often than not was confronted with situations in which a consensus about the goal had to be reached.
Institutional environment: Lack of hierarchical institutions
The second aspect concerns the institutions through which solutions can be sought to the problems studied at RAND and IIASA and extends the argument made in the previous section on multi-objectivity. RAND’s early systems analysts designed solutions for military and governmental bureaucracies that were able to govern by decree. Inasmuch as they had to deal with competing forces within these bureaucracies, the predominant managerial strategy was centralization of power in order to ease institutional streamlining and to diminish interference from intermediary actors. Within IIASA, no institutional mechanisms were available that allowed for similar hierarchical approaches to decision-making. With less control over the actors populating the sites of IIASA’s problems, researchers had to be more receptive to the heterogeneity of how the actors understood themselves and how they defined their situations.
It should be remembered that despite the mathematical bias in the methodological writings on systems analysis, the familiarity of the analyst with the practical details of the case has always been prominently discussed. It was, however, usually considered part of the ‘craft’ side of systems analysis, rather than being a subject of explicit methodological concern. This nonchalant handling of qualitative research aspects was not problematic, as long as the analysts were familiar with the situations in which they operated. For example, many RAND analysts had sufficient familiarity with the US military administration to avoid major misunderstandings. Similarly, applications in industrial settings – where systems analysis has found its most fertile field of application – are usually concerned with designing solutions on a rather technical level that need not really engage with fine-grained analyses of diverging definitions of the situation.
IIASA’s research agenda, however, set out to understand issues where, in addition to not having control over the actors, the analysts could typically not rely on habitual intimacy. One way to meet the challenge of non-intimacy with local particularities in global research situations was IIASA’s habit of cooperating with local research organizations in the member countries. Beyond that, however, IIASA’s leadership understood that serious efforts were necessary in order to assign a much stronger role to social scientists. We find evidence for this awareness even in the first Annual Report of the founding director Howard Raiffa, who lists the over-representation of economists and ‘narrower mathematical fields of systems analysis’ at the expense of the social sciences as one of the major concerns with which the institute will have to deal better in the future (International Institute for Applied Systems Analysis, 1974: 9).
This view was not unusual in the systems analysis community. Under Roger Levien’s guidance, in the early 1970s IIASA pursued a survey project to assess the current state-of-the-art in systems analysis, in order to prepare the publication of a handbook of systems analysis. 8 The survey project distributed questionnaires through IIASA’s NMOs in order to collect the professionals’ views on what topics and problems were seen as fundamental to current and future approaches in systems analysis. The results were published in two IIASA research reports in the summer of 1976 (Quade et al., 1976a, 1976b). The respondents particularly emphasized the need to engage more thoroughly with the ‘social, political and institutional aspects of systems analysis’ and warned IIASA of an ‘excessive emphasis on technology’ (Quade et al., 1976b). Nevertheless, as can be seen from IIASA’s staff statistics, the number of social scientists remained low during both Raiffa’s (1972–75) and Levien’s (1975–81) terms as directors (see Table 1).
Composition of IIASA staff scientists according to disciplines, 1973–81.
Sources: IIASA Annual Reports 1979, 1980, 1981; percentages: own calculations.
Numbers and notations are reproduced here as they were published in IIASA’s Annual Reports. Later than 1981, IIASA did not publish such statistics any more.
Staff structure: Disciplinary and cultural diversity
IIASA was an environment in which cultural diversity was a very salient daily experience. In fact, as several interviews suggest, the prospect of being part of a research team of such exceptional internationality was a most attractive feature and among the principal motivations to become an IIASA scholar. International contacts across the Iron Curtain were generally rare and, even if they were available, were usually restricted to short-term study visits. While RAND very consciously designed its institutional structure in order to encourage maximum exchange of ideas between scientists of different disciplinary orientation, culturally its research staff was highly homogenous.
IIASA aspired to a similar kind of interdisciplinary exchange but additionally, and certainly no less purposively, it was designed to foster intercultural dialogue between scientists from different nations. Regular internal seminars in which lectures of general interest or work in progress were presented to the entire research staff were well-attended. Cultural events were organized and national holidays, such as the Soviet celebration of the Russian Revolution on 7 November or the US Independence Day on 4 July, were celebrated with lively participation. Travels to various member countries were frequent tasks, particularly of the project leaders who were expected to establish research relations between the member countries in their fields of interest. Workshops and conferences were regularly organized at the institute, but also in other places such as the Hungarian city of Sopron, some 50 km [30 miles] south of Laxenburg. Despite certain restrictions over maintaining private contacts for some of the eastern scientists, mutual curiosity and a sense of adventure prevailed during the early IIASA years. Together with Raiffa’s preference for hiring young enthusiasts, along with established professionals, the ‘IIASA spirit’ was supportive of daring innovation and open to failure.
The 1980s: The culturization of systems analysis
One way to assess the methodological consequences of the institutional features discussed above is to look at the work of IIASA’s Methodology Group. This research area was one of IIASA’s earliest study groups and had two functions. On the one hand, the group’s members were supposed to serve as in-house consultants to the various applied research projects wherever methodological or technical support was needed. On the other, the group was to engage in self-standing research on particular conceptual and technical problems of applied systems analysis. Therefore, reviewing the methodology group’s work over time provides a good indicator of what were understood as the major challenges IIASA had to face in the course of its research.
In line with IIASA’s focus on multi-player and multi-objective decision problems, since its first research plan in the mid-1970s, the methodology group defined ‘fair division’ and ‘multi-objective decision-making’ as its major research foci (International Institute for Applied Systems Analysis, 1976: 45).
From autumn 1973, the first leader of the Methodology Program was the famous US statistician George B. Dantzig, who approached these issues in the common mathematical tradition of systems analysis. Dantzig defined 4 main work areas for the program: linear programming; dynamic programming; decision analysis; and econometrics. All of these areas were linked to applied research strands, in particular the Energy Program and the Program on Water Resources. They remained, however, mathematically abstract and largely restricted to technical aspects of the systems studied in these programs.
A major reason for the dominance of abstract mathematical research during those early years can be found in the politics of IIASA. Probably the most defining feature of IIASA was to maintain a precarious balance between two potentially contradictory tasks: to touch upon the great global challenges of the time in order to justify the political and financial investment in this project; and simultaneously to stay a safe distance away from any activities that might be called ‘too political’ in the council or in the home countries. In this context, mathematical abstraction had the very important function of being, allegedly, immune to any ideological adhesiveness and thus served as the discursive legitimation for scientific cooperation between the blocs (cf. Rindzevičiūtė, 2010: 293).
All member countries were not convinced by this argument, however. After 10 years, IIASA’s existence was seriously threatened for the first time when anti-communist governments in the UK and the USA declared their unwillingness to provide further funding for an institute that was seen – not completely without justification – as a nest of spies. 9 The UK, under Margaret Thatcher, pulled out in 1982 and never resumed its membership. In the USA, and also in 1982, IIASA membership switched from the National Academy of Sciences to the American Academy of Arts and Sciences after the Reagan administration decided to withdraw government funding.
IIASA’s emphasis on multi-objective decision analysis did, however, eventually lead to a methodological diversification of IIASA’s work in the 1980s. After George Dantzig resigned as the leader of the Methodology Program, the position was, after two short terms under Tjalling Koopmans (July–December 1974) and William Jewell (January–June 1975), again given to US scientists. Michel Balinski, a reputable professor of mathematics from the City University of New York (1975–8), was the first. He was followed by Peter de Janosi (1978–9), an experienced foundation officer and trained economist (Wierzbicki and Young, 1981). 10 Both continued Dantzig’s research streams in what was renamed the Systems and Decision Sciences (SDS) Area. 11 Next to decision analysis the fields in which the SDS Area invested its energies were optimization theory, control theory, resource allocation and economic modeling and planning. All of these continued the mathematical tradition in SA research.
In 1979 Andrzej P. Wierzbicki, dean of the Faculty of Electronic Engineering of the Technical University in Warsaw, accepted the position of the SDS area chairman. He held degrees in automatic control and mathematical programing and optimization and initiated a major methodological diversification. His first annual report of 1980 explicitly mentions the integration of qualitative research methods in decision analysis and defines the overall objective of this study section as ‘Towards a Fair Share for All People’ (International Institute for Applied Systems Analysis, 1981: 27).
Wierzbicki’s most innovative work during those years centered on the question of how alternative conceptualizations to the goal maximization approach in decision analysis can be mathematically modeled. Real negotiators, in particular in multi-player and multi-objective situations, apparently followed strategies of ‘satisficing’ minimum requirements in different objectives, rather than maximizing an abstract index of overall performance (International Institute for Applied Systems Analysis, 1983: 26). Accordingly, Wierzbicki worked on ways to translate more realistic understandings of human decision-making into mathematical models – a line of research that has become known as ‘Soft Computing’ (Makowski and Wierzbicki, 2003).
The increased attention to more complex and empirically informed conceptualizations of the human decision-maker was even more visible in the methodology group’s focus on the question of ‘what effects cultural factors and institutional structures have on the decision process’ and its deciding to declare the ‘use of theories of cultural anthropology’ to approach these hitherto under-researched aspects of decision processes (International Institute for Applied Systems Analysis, 1982: 42). IIASA’s new director, Canadian ecologist C. S. Holling (1981–4) had been an expressive advocate of engaging with various affected groups of actors in any given policy situation. Under his leadership, IIASA’s Ecology Group had already worked in this way during the mid-1970s, placing much emphasis on risk perception.
Probably the most significant figure in strengthening the role of empirical social research at IIASA was British anthropologist Michael Thompson, who joined IIASA’s methodology group in 1979. While methodologically important, Thompson was also quite a character – a former soldier with the British army and a professional mountaineer. Thompson’s work is most remarkable because of his efforts to integrate the grid-group-theoretic framework of Mary Douglas (Douglas, 1970) into IIASA’s research designs, which he adopted under the label Theory of Plural Rationality (Thompson, 2008). The bottom line of this approach is the insight that in any policy situation, different groups of actors hold different and contradicting convictions of what the problem is and what solutions are at hand. Instead of imposing the logic of one of these rationalities against the others (because of its analytical coherence this is called an elegant solution), he claims that solutions are effective and viable only if they make sense for each group and each corresponding rationality; later called ‘clumsy solutions’ (Verweij and Thompson, 2011). A precondition for any such work is the availability of ethnographic knowledge about the population affected by the policy problem and the integration of results into the abstract modeling of systems analysis.
Although resistance to this methodological revolt was at times intense, cultural extensions of the traditional systems analytic approaches continued during the 1980s. One final example illustrates this movement. The Processes of International Negotiations (PIN) Program officially started in 1986 (Kremenyuk, 2002). Howard Raiffa, a pioneer in the field of decision analysis from a game theory background, had pursued the idea of studying international negotiations at IIASA for many years but met with resistance from council members who considered the project too close to actual policy-making. In an atmosphere of growing mutual trust, however, the council eventually approved Raiffa’s ideas and gave him the green light for a project whose objective was to seek better communication between practitioners and theoreticians of negotiations in international settings. Given an awareness of the limited practical relevance of much of the academic research for real decision-making, the project was eventually successfully integrated with IIASA’s older work on the management of large international rivers (LIR). Using case studies on the Danube and the Zambezi rivers, the PIN project found real-world applications for engaging with conceptual and empirical discussions of the ways in which cultural differences affect international negotiations (Faure and Rubin, 1993).
Conclusions
The article has analysed the ways in which various institutional structures shaped the methodological orientation of systems analysis. At the RAND Corporation, excessive emphasis was placed on developing quantitative analytical techniques in order to rationalize decision processes in large military bureaucracies. While not neglecting the importance of the social institutions in which the military problems were situated, the ‘human factor’ appeared primarily as a source of disturbance, be it through psychological mechanisms that affect the single decider (biases, self-interest, cognitive fallacies, etc.) or through sociological group mechanisms. In other words, the sociality of decision-making was essentially seen as a source of irrationality and thus had to be circumvented by rigidly formalized analysis. Therefore the dominant strategy to rationalize decision-making in large military bureaucracies consisted of: (1) determining the most rational solution to a given problem by detached mathematical analysis; and (2) creating hierarchical institutions that are in the position to implement these solutions against the resistance of those in the bureaucracy who are following their own biased interests and assumptions.
Continuing this analysis, the article moved to provide the historical background for the establishment of the International Institute for Applied Systems Analysis (IIASA) and the rationale that underpinned the choices behind building this unique endeavor of East–West cooperation upon the methodological commitment to systems analysis. Besides being of mutual interest to both the USA and the Soviet Union, it was precisely the alleged ideological neutrality of this abstractly mathematical approach to problem-solving that served as both justification of, and reassurance towards, critics of political rapprochement between the blocs. In itself, however, IIASA held a number of institutional features that posed a constant source of disturbance of the scientistic orientation in systems analytic research. In particular, its institutional purpose of touching upon complex multi-party policy problems on a global scale, the lack of hierarchical institutions through which these problems could have been handled and the daily experience of international, intercultural and interdisciplinary encounters opened up countless opportunities for researchers to pursue new, and sometimes fruitful, ways of thinking about the limits of their approaches and possible innovations. Fundamentally, more complex conceptualizations of human deciders and the affected populations of the policy problems that were studied at IIASA enriched the excessive emphasis on quantitative analytical techniques in systems analysis. Citing exemplary research projects at IIASA reveals that, during the 1980s, it was the concept of culture that most profoundly challenged the traditional methodological orientations of systems analysis and led to an increased usage of qualitative methods.
Besides the historical interest in IIASA, the theoretical purpose of this article was to show how institutional factors condition methodological preferences at two of the principal research organizations in systems analysis. Future work will have to deepen this argument in two ways. First, in-depth case studies of IIASA research programs will provide a more detailed understanding of how exactly the identified institutional conditions impacted methodological choices. Second, an attempt must be made to strengthen this hypothesis using a larger number of cases including those where the quantitative bias has prevailed.
Footnotes
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The author’s research was supported by the Austrian Science Fund (FWF) as part of the project ‘Cold War Inventions in Social Research Methodology and their Trajectories’ (project no. P 24694).