In our work on the genealogy of vital systems security, Stephen and I have noted the importance of “imaginative enactment” as a form of VSS knowledge-production. Among other things, imaginative enactment is a method for determining infrastructural vulnerabilities in the absence of archival data on the historical incidence of what are termed “low probability, high consequence” events – such as a virulent influenza pandemic, a dirty bomb attack on a major city, a catastrophic earthquake, etc. One form of imaginative enactment that I’ve been looking at is the scenario-based exercise. These are role-playing games in which decision makers are faced with an urgent crisis sparked by an event (a terrorist attack, an outbreak of an infectious disease, etc), take action to intervene, and watch the results of their interventions unfold. In this post, I want to begin to explore the structure and history of this type of imaginative enactment – which was originally developed in the 1950s at RAND (along with everything else), and called the “political exercise.”

A recent example is the “Dark Winter” exercise held at Andrews Air Force base in June 2001, which simulated a smallpox attack on the United States. It was the product of a collaboration between the Center for Strategic and International Studies (CSIS) and the Johns Hopkins Center for Civilian Biodefense. The exercise was, its designers wrote, “intended to increase awareness of the scope and character of the threat posed by biological weapons and to catalyze actions that would improve prevention and response strategies.” Experienced political “decision makers” such as Sam Nunn and James Woolsey deliberated in a series of National Security Council meetings as the smallpox epidemic unfolded. Given a lack of sufficient vaccine supply, unclear lines of authority, and information breakdowns, the leaders did not have the means to halt the spread of the disease. A national catastrophe was the result.

In the wake of Dark Winter, participants engaged in a series of briefings to policy-makers in the executive branch and congress. Although its direct influence is hard to estimate, the exercise is often cited as a significant event – before the attacks of 9/11 and the anthrax letters – in galvanizing the US government to increase its biopreparedness activities.

My question here is: how does the exercise achieve its effects? It produces ‘experiential’ knowledge about vulnerability – that is, leaders’ experience of their own lack of knowledge and experience, which combines with the feeling of responsibility to produce a sense of helplessness in crisis. It targets this experience at the act of decision. To do this effectively, exercise designers must construct a plausible, realistic event in which the affect and judgment of decision-makers is invested. How does the method work? Where does it come from? I want to focus here on the role of what were called, in Dark Winter, the “exercise controllers.” These somewhat shadowy figures provide the briefings of facts and policy options that control the apparently contingent outcome of the scenario.

Creating a “Twilight Zone” 

CSIS began conducting “crisis games” in the 1980s, under the leadership of Robert Kupperman, a security policy intellectual with a background in operations research. Kupperman had been concerned about government readiness for crisis situations since his time in the Office of Emergency Preparedness (OEP) in the Nixon Administration, where he worked on various natural disasters, the energy crisis, the wage price freeze, and terrorist events such as Black September. In this context, he developed an interest in the common structure of crisis situations, and in the development of techniques that could be used to prepare for them in advance. He argued that crises, however diverse, shared a number of common problems: the paucity of accurate information, the difficulty of communication among decision-makers, and a confusing array of authorities seeking to take charge of the situation. Such situations involved uncertainty about what was unfolding, coupled with an urgent demand for immediate action to alleviate the crisis. Flexibility for decision-makers depended on the extent to which the crisis manager had forecast the situation and invested in preparation for it. “As we begin to recognize the complex problems that threaten every nation with disaster,” he and two colleagues from OEP asked in 1975, “can we continue to trust the ad hoc processes of instant reaction to muddle through? (Kupperman, Wilcox and Smith 1975: 229)”

In the 1980s, after a stint in the US Arms Control and Disarmament Agency, Kupperman joined the Center for Strategic and International Studies (CSIS), a Washington DC think tank. There he was co-author, with R. James Woolsey, of a 1984 Report on “crisis management in a society of networks” called America’s Hidden Vulnerabilities. The report argued that the U.S. relied for its well-being on a sophisticated and intricate set of systems, or networks, for energy distribution, communication, and transportation. It noted recent disruptions of these systems, and warned: “A serious potential exists … for much more serious disabling of networks crucial to life support, economic stability, and national defense.” (Woolsey and Kupperman 1985: 2)

At CSIS, Kupperman and his colleagues sought to persuade national security officials of the problem of network-vulnerability, and the need to develop techniques of contingency planning. One of their approaches was to hold scenario-based simulations of crisis situations, and invite officials to participate. The emergency exercise was a tool for demonstrating to leaders the vulnerabilities of critical systems. As he and Woolsey wrote: “If planning has involved the operating teams and managers (as it always should) these critical personnel gain an increased understanding of how the system works and, particularly valuable, how it is likely to behave under abnormal conditions. Training with crisis games and emergency exercises will augment this benefit significantly.” (Woolsey and Kupperman 1985: 16)

In a 1987 New York Times article on the CSIS simulations, Kupperman argued that successful exercises had four key elements. First, a plausible scenario; second, a rapid sequence of events, leading to a feeling of pressure, a demand for immediate decision. This realism was linked to creating a sense of responsibility in participants: “We try to make the players feel personally responsible…We create a twilight zone; they know it’s not real, but they’re not quite sure”; third, participants: choosing experienced people; and fourth, having a “control staff” to simulate the real world (Halloran 1987). The Times noted the widespread use of the practice of simulation: “today, simulations have gone beyond military strategy to include politics, diplomacy, economic leverage, public opinion and the psychology of decision makers under the pressures of time, confusion and demands from every direction.”

The emphasis in designing such exercises was not on predicting or preventing the occurrence of a crisis, but on the decision-making process of leaders once a crisis was underway. In his forward to the CSIS volume, Admiral Thomas Moorer wrote: “The CSIS crisis simulations are not designed to be predictive. Rather, they are intended to provide insight into policy dilemmas likely to plague national leaders during real crises and to identify key decision-making pathologies that could lead to unwanted escalation.” (Kupperman and Goldberg, viii)

Plausibility, Not Probability

What was crucial to a successful scenario was that the players take their decisions in the exercise seriously. One had to somehow persuade them to behave as if the simulation were the real thing. As Kupperman and his colleagues wrote: “One of the greatest challenges for game designers is to induce players to take their actions seriously without having any actual ability to force them to accept responsibility for their actions the way the president, Congress, or the Soviet Union might” (15). Here realism was a critical factor: “The more realistic the game design, however, the more absorbed the players become.”

The point was to create a plausible – rather than a likely – scenario. As the CSIS authors write: “In developing the scenario, the main criterion was that of plausibility – rather than high probability – in what might occur rather than what would occur” (3). How was plausibility constructed? “A plausible crisis game must, therefore, realistically simulate a political environment characterized by intense time constraints, crosscutting political demands, and a high level of risk” (11). The experience of the realism of the event and its aftermath led to the absorption of responsibility: “Team players, therefore, bore the consequences of their acts in the domestic or global arena. NSC players experienced the threats, penalties, and opportunities posed by environmental factors through the control of informational input” (12). The reality-effect of the exercise depended not only on the plausibility of the scenario, but also on the interventions – during the event itself – of the “control group” – that is, the behind-the-scenes figures who provided the “results” of the official players’ interventions.

Let me turn now to the history of the role of the “control group” in creating the sense of realism necessary for absorbing the players, for making them feel responsible for their decisions, and therefore for generating the experience of vulnerability necessary to a successful exercise.

Designing Lack-of-Control

Kupperman and his collaborators at CSIS named the RAND and MIT “political exercises” of the 1950s and 1960s as an important precursor to their simulations. The political exercise was invented in the 1950s at RAND by members of the social science group, led by Herbert Goldhamer. The political exercise differed from classical war games in that it involved the strategic calculations of political decision-makers rather than military planners. In the context of the Cold War and the catastrophic consequences of escalation, a key issue was of course how to avoid going to war. The focus of the RAND political exercise was thus political decision in crisis. Its developers, Goldhamer and Hans Speier, also distinguished the exercise from more formal, mathematical games. According to Goldhamer and Speier, the attempt to formalize political decision-making processes in crisis “was abandoned when it became clear that the simplification imposed in order to permit quantification made the game of doubtful value for the assessment of political strategies and tactics in the real world.” In contrast to such simplification of the international situation, the political exercise made it possible “to simulate as faithfully as possible much of its complexity” (Goldhamer and Speier 1959: 72-3).

Goldhamer and Speier decided not to depict the present, but to design scenarios as projections into the future, in order to avoid entwinement with current events. The scenario, they wrote, was an “effort to describe how the world of January 1, 1957, would look. It provided the players with a common state of affairs from which to begin. The scenario rid them of the intrusion of current news into the game and served to focus it on problems of special analytical interest” (74).

Uncertainty and contingency must be experienced – this gives players a sense of responsibility for their decisions. Thus the exercise provided players with “new insight into the pressures, the uncertainties, and the moral and intellectual difficulties under which foreign policy decisions are made. This, of course, is part a tribute to the earnestness and sense of responsibility with which the participants played their roles, since otherwise these pressures and perplexities would not have made themselves felt” (79).

How then to generate the experience of uncertainty? A key requirement of the game, for Goldhamer and Speier, was the “simulation of contingent factors” – what they called “nature”. As they wrote: “In political life many events are beyond the control of the most powerful actors, a fact designated in political theories by such terms as fortuna, ‘chance,’ ‘God’s will,’ ‘changes in the natural environment,’ etc. We tried to simulate this by the moves of ‘Nature.’”

Referees played the role of Nature: “This arrangement…. Permitted the referees to make certain non-governmental moves which constituted indirect, partial evaluations of the state of affairs that had been reached at any chosen point of the game” – “the referees could introduce such evaluations in the form of press roundups, trade union resolutions, intelligence reports, speeches made in the United Nations, etc” (73-4).

“The role of ‘Nature’ was to provide for events of the type that happen in the real world but are not under the control of any government: certain technological developments, the death of important people, non-governmental political action, famines, popular disturbances, etc.” (73)

Dissemination

The method developed at RAND was then disseminated in academic and policy arenas, as the field of “strategic and international studies” was institutionalized (see Kuklick). Goldhamer and his fellow developers collaborated with colleagues at SSRC, the Center for Advanced Study in the Behavioral Science at Stanford, Yale, the Carnegie Endowment for International Peace at Princeton, the Brookings Institution, Northwestern and MIT (80).

Political scientist Lincoln Bloomfield became an enthusiastic developer of the methodology at the MIT Center for International Studies in the 1960s – he saw the political exercise as a possible solution to the search for “ways of bringing to foreign policy planning some of the imaginative analytical techniques employed by military planners and operational analysts.” The purpose of the games was four-fold, he wrote: to throw light on hypotheses about foreign policy and strategy; to pre-test strategies of action; to “discover unanticipated contingencies, alternatives or possible outcomes as a consequence of the interaction between conflicting strategies in the simulation”; and to “examine closely one line of policy action that illustrates vividly what a single plausible outcome might resemble in detail” (Bloomfield and Whaley, 1965: 887).

Bloomfield and his group took up from the RAND design the practice of having a control group enact “nature” as the source of contingency. As they wrote, the control group “represents ‘nature,” introducing unexpected events; it is umpire, ruling on the plausibility and outcomes of moves; it is, as it were, ‘god,’ requiring the players to live with the implications of their chosen strategies” (858).

Although they did not name it “nature” or “god,” CSIS emphasized the central role of the “control strategy” in creating the realistic situation of crisis in which unpredictable events are unfolding in real-time, and demand immediate response. “In formulating control strategy, the research group sought to pose to the team a number of functional problems, which would reflect key dimensions of crisis dynamics. This was done by simulating organizational impediments, domestic political impediments, problems of allies and regional actors and, finally, issues invoking U.S.-Soviet coercive diplomacy” (Kupperman and Goldberg, 1987: 12). In an exercise simulating a crisis on the Korean peninsula, the “the control group deliberately structured a leaky news environment to heighten the tension, as well as the realism, of the exercise” (18). The control group’s input demonstrates the lack-of-control of the decision-makers in a crisis situation for which they are not prepared – and generates strong affect (“tension”) among participants.

References

Bloomfield, Lincoln P. and Barton Whaley, “The Political-Military Exercise: A Progress Report,” in Orbis VIII: 4 (1965), 854.

Goldhamer, Herbert and Hans Speier, “Research Note: Some Observations on Political Gaming,” in World Politics, October 1959.

Halloran, Richard. “The Game is War, and it’s for Keeps,” New York Times, June 1, 1987.

Kupperman, Wilcox and Smith (1975). “Crisis Management: Some Opportunities,” in Science 187.

Kupperman, Robert H. and Andrew Goldberg, Leaders and Crisis: The CSIS Crisis Simulations: a report of the Arms Control and Crisis Management Program (Washington, 1987).

Woolsey, James R. and Robert H. Kupperman, America’s Hidden Vulnerabilities: Crisis Management in a Society of Networks (1985).