Cognitive Maelstroms, Nested Negotiation Networks, and Cascading Decision Effects: Modeling and Teaching Negotiation Complexity with Systemic Multiconstituency Exercises

In this article, we describe a method for training negotiators to be prepared for an increasingly complex world. The gap that has always existed between real‐world negotiations and the quality and scope of experiential learning exercises has only grown as the world that negotiators confront has grown more complex in multiple ways.

First, the number of formal and informal stakeholders that influence collective decisions across the public, private, and nonprofit sectors has increased. Because of globalization and the telecommunications revolution, actors now function across coalitions and networks. Each network operates within its own ecosystem of behaviors, assumptions, needs, and interests.

Second, relationships between and among the global, regional, and local environments are rapidly changing. The distinctions between these three environments blur together. Negotiators are often unaware of the interconnectedness of conflicts and unable to decode how individuals and groups react to each other not just within but across conflicts and environments.

Third, leaders around the world are hindered by the growing ineffectiveness of their conflict management tool kit, which challenges their ability to deliver effective governance and their legitimacy with their constituents. Examples include the financial crisis, the Middle East conflict system, and ongoing refugee crises around the world. These crises have, in turn, generated a volatile environment that makes it easier for “spoilers” to disrupt legitimate negotiation efforts.

As negotiation and conflict resolution instructors, we realize that no currently available experiential training tool captures these complexity features.¹ Confronted with a “VUCA” world characterized by volatility, uncertainty, complexity, and ambiguity (Yarger 2006: 18), and with the constant “tension between globalism, regionalism, and sectoralism” (Aggarwal 1998: 195), negotiators and negotiation instructors must not be deterred by chaos and complexity. Instead, they must embrace it to generate strategic action.

In this article, we first examine the meaning of “complexity” in negotiations and negotiation exercises. We then outline our understanding of negotiation complexity informed by the analysis of complex systems and system effects. Based on this diagnosis, we develop a model for operationalizing negotiation complexity in the context of experiential learning with the goal of advanced negotiation skill‐building. We explain the design of The Transition, a systemic multiconstituency exercise (SMCE) for seventy‐two people, which we developed based on this new design model. We conclude with observations, caveats, and recommendations that we have developed after conducting the exercise several times.²

No scholarly consensus on how to define complexity in negotiations has yet emerged, nor for what exactly complexity means in the context of negotiation simulation design and for what it implies for negotiation training.

Michael Watkins argued that “complexity is the rule in negotiation, not the exception” (Watkins 1999: 247). Larry Crump defined a complex negotiation as one that “includes two parties in which at least one party is not monolithic (e.g., individuals or groups within the party are not unified in their goals and/or strategies), a negotiation with more than two parties, or a negotiation with linkages” (Crump 2015: 135). Knut Midgaard and Arild Underdal argued that a fundamental consequence of adding more parties to a negotiation situation is that it “tends to become less lucid, more complex, and therefore, in some respects, more demanding” (Midgaard and Underdal 1977: 332). Roy J. Lewicki and his colleagues agreed that multiparty negotiations are “more complex, challenging, and difficult to manage” than two‐party negotiations (Lewicki, Saunders, and Minton 1997: 167). They distinguished between informational and computational complexity, social complexity, procedural complexity, and strategic complexity in multiparty negotiations (Lewicki, Saunders, and Minton 1997: 168). David Lax and James K. Sebenius suggested that complex agreements often require sequencing smaller, individual deals to achieve the desired outcome. They argued that “negotiation campaigns,” in which relevant parties are grouped into “negotiation fronts,” work as a strategy to forge complex deals (Lax and Sebenius 2012: 95). Generally, authors agree that more parties or more issues make negotiations more complex. On the other hand, other authors have argued that adding more issues can sometimes make an agreement more achievable because interests become tradeable (Menkel‐Meadow 2009: 424).

Over the past thirty years, negotiation scholars and teachers have produced a range of negotiation exercises that capture this “cumulative complication” (more actors and more issues), but that have not yet captured the complexity of a volatile, globalized world and its dynamism, turbulence, and uncertainty. Various forms of simulations (also known as role‐plays, games, serious games, etc.)³ remain extremely popular in negotiation courses (Druckman and Ebner 2013). Lawrence Susskind and his colleagues argued that simulations are crucial for “highlighting theoretical lessons and developing skills” (Susskind et al. 2005: 407). More recently, Daniel Druckman and Noam Ebner developed systematic recommendations for simulation instructors (Druckman and Ebner 2013).

For the most part, however, the exercise pedagogy has presumed that shifting from bilateral to multiparty exercises simply means doing more of what you do well with a single counterparty but with more actors. Susskind and his colleagues wrote that “to convey the complexity of the multiparty context, we would recommend that instructors include one long (all‐day) and complex simulation” (Susskind et al. 2005: 407). Leonard Greenhalgh and Lewicki (2015: 475) argued that “the increasing complexity that students need to understand has required that teaching materials advance beyond the robust, replicable studies of staged transactional negotiations” such as “prisoner’s dilemma games, buyer–seller transactions, or multiparty simulation scenarios with a few, well‐defined agenda items” (2015: 467). John Fayerweather and Ashok Kapoor examined the “adjustment of strategy to varied environmental conditions” in simulations (Fayerweather and Kapoor 1972: 19), and Gilbert R. Winham examined the use of simulations in the context of international relations (Winham 1991).

Addressing analytical negotiation models and not simulations, Watkins complained about models that are “simplistic, sterile, and‐or static.” His insightful diagnosis (1999: 246) has useful implications for simulation design as well:

We realized that what was missing was a comprehensive application of the conditions that Watkins identified to the experiential learning environment. In the world of volatility, uncertainty, complexity, and ambiguity (VUCA), the playbook and the rules are not well defined. This places a premium on improvisation, creativity, and flexibility in negotiations. Without the proper training tools, our students will be ill equipped to face the VUCA world head‐on. Effective negotiation pedagogy needs to expand from the standard repertoire of negotiation training exercise to include advanced skill‐building and training platforms for managing complexity in negotiations. This involves reflecting on the kind of skills negotiators must employ to stay relevant and generate credible interventions in situations of turbulence and continuous disruption.

We diagnosed a direct connection between the complexities of today’s world and widespread “leadership paralysis.” While negotiation is increasingly becoming a core competence in political, economic, and military decision making, the stakes for “not getting it right” continue to increase. We seek to meet this challenge and to reduce the experiential learning and training gap on negotiation complexity to help negotiators get a better handle on difficult, evolving, complex conflicts.

The central proposition of this article is that complexity and turbulence complicate international negotiations, but they have not been sufficiently addressed by existing experiential learning tools. We define complex negotiations as not just those that involve multiple actors and issues, but as negotiations that feature multiple layers of interlocking and interacting system effects. These include linked sets of negotiations occurring simultaneously in a world of continuous disruptions, regime changes, and leadership transitions. We characterize negotiation processes that meet these criteria as “complex systems.”

A complex system comprises “diverse, interdependent, interacting entities whose aggregate behaviors can often transcend the characteristics of the parts” (Jones‐Rooy and Page 2013: 316). The behavior of a complex system is not directly determined by the actions of individuals, but “by the sum of these interactions” (Susskind 2010: 368). Thus, we conclude that events at the macro level of a complex negotiation are not always a linear consequence of events at the micro level. As Robert Jervis put it in his monumental work on system effects in international politics, “the whole is different from, not greater than, the sum of the parts” (Jervis 1997: 12). The overall system has its own character, but this character cannot be deduced from looking at its parts in isolation.

The properties of the complex system differ from the properties of the parts of the system and are called “emergent properties.” The system’s parts are heavily interconnected and “the fates of the units and their relations with others are strongly influenced by interactions at other places and at earlier periods of time” (Jervis 1997: 17). The actors’ choices affect other actors’ choices, but often in unpredictable ways. Parties’ abilities to anticipate each other’s behaviors and base decisions and perceived choices on these assumptions – whether correct, incorrect, or partly incorrect – exemplify this phenomenon.

This interconnectedness has important implications for the effects of individual and collective action, and for the acquisition of skills needed to anticipate and manage these kinds of effects. “Because of interconnections, many effects are indirect, mediated, and delayed. … When what happens influences events in places or times far removed, prediction and even post‐hoc explanation are very difficult …. Indirect effects may be more important than direct ones.” (Jervis 1997: 29) In a complex negotiation, these effects can be described as “negotiation system effects.”

We believe that few, if any, negotiation scholars and instructors have determined how to operationalize this complexity and convert it into a workable training opportunity for negotiation practitioners. What is missing is a training tool on negotiation complexity that meets Melissa Manwaring and her colleagues’ definition of an authentic learning experience: “one that reflects the richness, complexity, and unpredictability of actual negotiations … outside of the classroom” (Manwaring, McAdoo, and Cheldelin 2010: 122). We seek to close this pedagogical complexity gap to ensure that the next generation of decision makers will be capable of acting as “all‐terrain negotiators” who are agile and adaptable in both prevailing and emerging international environments.

Models, as realistic as their creators claim they are, are still “approximations for the real world” (Banks 2009: 3). They do not capture reality in all its facets. Looking beyond negotiation science, we find training tools that may, to some extent, qualify as useful models of complex systems that we could build on as we set out to design something new. For example, military war games first became common in the 1800s (Loftin 2009: 248). Role‐plays and computer‐based simulations are now used to educate audiences about the epidemiology of infectious disease (Verran et al. 2014). Likewise, an irrigation management simulation has been used in the context of engineering (Burton 1989), and a fictional city based on a simulated social system has been used to educate welfare workers (Hughes 1992). What these simulations have in common is that they aim to reconstruct patterns of real‐world systems instead of providing a tool for individual skill building at the cost of suppressing structural forces and negotiation system effects.

Stephen E. Weiss has written about an impressive “mega‐simulation” for MBA students in which students are assigned the roles of company representatives. The public package for this comprehensive business negotiation is eighty‐six pages long, in addition to role‐specific information. According to Weiss, such a mega‐simulation can help students to “develop a deeper understanding of negotiation subject matter and complex processes than they would by conducting standard role plays” (Weiss 2008: 1). He considered three learning objectives to be uniquely achievable through a mega‐simulation: “combining negotiation concepts and skills, appreciating the nature of complex negotiation dynamics and tasks, and obtaining multidimensional, multisource performance feedback” (Weiss 2008: 328). Jonathan M. DiCicco has written about a semester‐long simulation of the United States National Security Council. Student feedback indicated that simulation participants believed that they “improved their ability to present ideas, speak concisely, prepare options for others to consider, be prepared for questions, remain open to criticism, and see different facets of a complex problem” (DiCicco 2014: 455).

These skills are without doubt useful, and the use of simulations such as these should be encouraged. But we concluded that the inclusion of many parties, a lot of material, and/or an extensive time frame per se do not guarantee that simulation participants will experience negotiation system effects. Equally important, we concluded that lengthy simulations involving many parties do not per se provide a learning environment that trains people effectively in advanced negotiation skills. We realized that, to build a negotiation exercise that models negotiation complexity with the goal of teaching a new and expanded set of skills, we must shift our way of thinking from coalitional multiparty negotiations to dynamic and networked multiconstituency negotiations.

We did not start from the premise that “complexity” is a phenomenon inherent to every negotiation nor inherent to every simulated negotiation exercise. We also did not assume that simulations automatically become more “complex” as time, parties, or issues are added. Instead, we sought to model a complex negotiation system that moves beyond “simplistic, sterile, and‐or static” (Watkins 1999: 246) simulation design and confronts negotiators with three key negotiation complexity challenges that capture the progression toward networked, multiconstituency dynamics:

• “cognitive maelstroms” in high‐stakes, high‐consequence situations;

• nested negotiation networks instead of conventional coalitions; and

• cascading decision effects in a volatile, uncertain, complex, and ambiguous world.

We use the term “cognitive maelstroms” to describe the emotional and psychological implications of the cognitive flooding of individuals and groups that occurs in high‐stakes, high‐consequence situations in a complex negotiation system. Negotiators often perceive the amount of incoming information as so massive, the speed of events so fast, and the time to make a decision so brief, but the stakes so high that they feel overwhelmed. Groupthink and conformity bias, polarization, and misperceptions are some examples of the cognitive biases and psychological traps that arise in this context.

Negotiators must be able to manage conflict in situations in which this cognitive flooding occurs; they must know how and when to share and withhold information strategically and how to manage their own fatigue after sleepless nights, while maintaining a “zoom‐in/zoom‐out” perspective: they struggle to see the “whole game,” although they may only be able to grasp small pieces of the overall puzzle. Knowing the difference between sufficient and unanimous consensus, managing information leaks and misinformation, knowing when to be risk‐accepting and when to be risk‐averse, being able to persevere and to stay focused for long periods of time even if (or especially when) one’s issues are not being currently discussed are examples of key skills that can help the negotiator cope with decision making in high‐stakes, high‐consequences situations. Complex negotiations create cognitive maelstroms that can rattle, overpower, or “devour” even the toughest, hard‐nosed bargainer. Complex negotiations are psychologically exhausting and can exact an emotional toll on insufficiently resilient negotiators.

A complex negotiation system does not comprise static negotiation coalitions, but networks of fluid relationships. We know from conventional negotiation analysis that negotiation coalitions (and how they are built, managed, and broken) can play important roles. The coalitions in a complex negotiation system, however, are less static and more dynamic, and are often so contradictory in their objectives and overlapping in their ties that it is difficult to classify them as “coalitions” at all. Negotiators must manage these shifting and interlocking alliances, which are often subject to repeated breaches of trust. They must be able to identify sources of leverage that may not be visible or apparent within specific sub‐networks of stakeholders and they must know how to exploit these sources at different stages and phases in the negotiation.

Negotiators often exert influence indirectly and across networks, not directly, which makes it even more crucial to identify leverage opportunities. If a negotiator misses the window of opportunity for a specific move, the window may never open again. It may be too late to act opportunistically, even if the planned strategy is brilliant. Negotiation constituencies are extremely important in complex negotiation systems, but they are often fragmented. As Eileen Babbitt put it, “the most notable feature of international diplomatic negotiation is that it is carried out by agents who are working for a multiplicity of principals” (Babbitt 1999: 136). Failure to recognize these constituency dynamics and to decode the multilayered web of relationships can quickly result in self‐inflicted paralysis.

A complex negotiation system is always evolving and always in motion. Jervis’s (1997: 29) description of complex systems provides guidance for complex negotiations as well: “Many crucial effects are delayed and indirect; the relations between two actors are determined by each one’s relations with others; interactions are central and cannot be understood by additive operations; many outcomes are unintended; regulation is difficult.” Confronted with these uncertainties, a negotiator must still make decisions.

But no decision is without consequences in a volatile, uncertain, complex, and ambiguous world: failure to achieve a negotiated agreement has consequences, and reaching agreement also has consequences. And the agreement is not the end in itself: the negotiator may not see and feel its consequences right now, but they exist. Negotiators must learn how to cope with the wide‐ranging cascading effects of their decisions. They must master the art of “firm/flexible commitments,” that is, contingent deals that allow for enough planning reliability to move forward while keeping the window open for changes if the unforeseen happens – which, most likely, it will. Negotiators must confront the ethical dilemmas that those decisions generate and also manage the range and the scope of commitments and concessions over a long time frame in a multistage scenario, while leading from the front, from the middle, or from behind; all the while actively managing spoilers, hold‐outs, and issue‐hostage‐takers who are willing to manipulate the game and determined to ignore every rule if such behavior advances their agenda.

How could we incorporate these three negotiation complexity challenges into a simulated environment? Luckily, negotiations scholars have already identified many valuable design lessons that we were able to build on, but we need to go further.⁴ To build something truly new, “it’s important to become comfortable with the process of trial and error and iterative prototyping, or you’ll be tempted to focus on the less risky mode of mastery, to the exclusion of originality” (Martin 2007: 183). We realized that, despite the richness of the field, it would be necessary to innovate and to create a new, interdisciplinary negotiation architecture different from that of existing simulations.

To build a prototype for a systemic multiconstituency exercise, we had to grapple with opposing principles of exercise architecture in particular and of negotiation teaching in general. For instance, simulations can vary significantly in their levels of realism and with regard to the degree of psychological stress they impose on participants. As Ebner and Yael Efron wrote (2005: 382), “reality can mediate between the participants and the scenario. This situational familiarity lends the participant a degree of comfort and security in her role and in the scenario. … However, this familiarity, when taken too far, can cause the participant to lose touch with the skills being practiced, with her role, and often with the trainer.” Laurence de Carlo (2012: 352) drew on the psychoanalytical approach developed by Donald Winnicott when he described the paradoxes in teaching negotiation and simulation design between “caring for the students while simultaneously causing them frustration” and “helping the students be more autonomous while manipulating them.”

The overall learning environment forces participants to confront their own strengths and weaknesses. Properly set up, de Carlo argued (2012: 362), the classroom “can function as a ‘transitional space’ (Winnicott 2001) that seeks to help student negotiators situate themselves between extreme self‐perceptions of impotence and omnipotence, and develop a realistic understanding of negotiation settings and their place within them.” We sought to apply all these principles in our architectural process choices. We did not seek to make our students feel “stuck” in an endless and confusing scenario in which our learning objectives might get lost. So, we recognized that we might have to remove some of the complexity elements, if we hoped to avoid building an “awesome negotiation rollercoaster ride” that lacked a clear teaching purpose.

The Transition is the outcome of our design process: a systemic multiconstituency negotiation exercise for seventy‐two participants, which we prototyped and then ran several times at Harvard University. With this tool, we seek to extend the limits of negotiation experiential learning. We define systemic multiconstituency exercises (SMCEs) as role‐play simulation exercises that move from a multiparty negotiation scenario to a networked multiconstituency negotiation scenario characterized by an immersive environment, emergent properties, and a dynamic architecture. While one or two of these qualities can also be observed in conventional multiparty exercises, the unique contribution of the SMCE prototype lies in the integration of all three characteristics into a functioning teaching tool that models a complex negotiation system. Each of the three simulation design components corresponds to one of the negotiation complexity challenges that we sought to model (see Table One).

To create a functional learning environment, we had to translate these design components into specific scenario features. We decided to model our systemic multiconstituency exercise on a real‐world conflict system featuring sufficient turbulence and chaos as well as the negotiation complexity challenges listed in Table One so that it would serve as a blueprint for a simulation with multiple negotiation system effects.⁵

We based our scenario on the specific conflict in Afghanistan and the broader political dynamics at work in Central Asia today. This system’s elements include the following :

• an ongoing civil war,

• the presence of North Atlantic Treaty Organization (NATO) forces,

• the presence of varied militant Islamist groups,

• conflicts over societal modernization,

• ethnic tensions,

• socioeconomic problems, and

• the involvement of both great powers (the United States, China, Russia) as well as regional powers (Iran, Pakistan, India).

The conflict system in this vast region is as fascinating as it is puzzling. Conceiving of the Afghanistan conflict as a problem heavily intertwined with regional politics reveals a web of relationships in which many state and non‐state actors are involved. These interconnections pose extraordinary challenges to negotiators (Bell 2015).

We applied what Ebner and Efron called the “pseudo‐reality method” (Ebner and Efron 2005: 382), creating a simulation based on a real‐world scenario and inspired by historic events, but adding in fictional events and distortions of reality to highlight specific lessons and sustain high levels of participant engagement. The simulation scenario should feel familiar, well‐known, and not entirely fictional to participants, but it is not a perfect account of reality. Instead, facts and information are “added or altered in ways that aid the role‐playing process without damaging the general framework” (Ebner and Efron 2005: 382).

We knew that if we wanted to recreate the cognitive maelstroms of complex negotiations, we had to make the scenario “feel real” in ways that the participants would not expect. We thought we could accomplish this by triggering their curiosity to then lure them into a turbulent, fictional, and immersive world. We hoped to compel students to “stay focused for deep and lasting learning” (Brazeau and Brazeau 2009: 1) despite the constant multitasking distractions of cell phones, the Internet, and social media. This is not to say that negotiators do not or should not multitask – multitasking is often necessary. But we sought to make sure students would multitask exclusively within the context of a fictional universe that would not disrupt the reality of the simulation.

As Chris Guthrie observed (2009: 66), people “appear to become more curious when they are in a good mood (Murray et al. 1990; Hirt et al. 1996), when working with others (Isaac, Sansone, and Smith 1999; Sansone and Thoman 2005), and when participating in novel or complex activities that they nonetheless find comprehensible (Silvia 2005, 2006, 2008).” The Transition incorporates these insights: over the course of the first day, participants are gradually introduced to the simulation scenario and walked through the technical details of how the fictional universe works. This approach is designed to get participants excited about the promise of a highly realistic exercise and to allow the instructors to answer questions outside of the simulation environment. Participants are then provided with general information as well as with unique role‐specific instructions and asked to prepare individually, as well as in relevant teams such as country delegations. After they are given an opportunity to ask questions of the instructors, the simulation is launched. From this point onward, until the simulation concludes the next day, the simulation environment is never broken. Participants remain in character and fully immersed in the scenario. The simulation itself runs over roughly twenty‐four hours which includes a night during which no official negotiation sessions are scheduled. However, with some exceptions, informal negotiations may take place at the discretion of the participants. The Transition concludes with a multistage debrief (See section “Observations, Caveats, and Recommendations” for suggestions for an effective SMCE debrief). Throughout these three days (see Table Two), all participants must stay in close proximity to each other – such as on or near a university campus or in a hotel or conference venue – which guarantees full immersion, reduces distractions, and allows for overnight negotiations in person. With a few exceptions, we provide participants with meals throughout the three‐day program.

With The Transition we built a negotiation simulation with an immersive environment relying on a radical interpretation of what Jean Poitras, Arnaud Stimec, and Kevin Hill called “mundane realism” (Poitras, Stimec, and Hill 2013: 445) They observed that many role‐plays do not feel realistic because the setting and the norms of interaction do not adequately reflect real‐world patterns. Hence, we decided to literally build a world by turning several floors in a number of buildings into spaces ranging from the United Nations (UN) Security Council to the Oval Office in the White House to the Afghan presidential palace. We tried to remove everything from the set that reminded participants of a classroom and instead brought in flags, carpets, pictures, constitutional documents, and other “stage props,” and we asked participants to adhere to a dress code that reflected their assigned identity.

Every member of the nine‐person teaching team conducts administrative tasks throughout the simulation “in character,” for instance as a U.S. Congressional staffer, bodyguard, or journalist. During the simulation, the staff avoids any references to the world outside of the scenario, but also avoids sterile “simulation language” – instead, administrators address all negotiators in character (“Sir, Madam, Mr. President,” etc.), and expect all participants to do the same. Most of the simulation material is conveyed in the form of a realistic real‐world equivalent, not in the form of “meta‐information” about the scenario. For instance, instead of a document called “role‐specific instructions” a foreign minister may receive a comprehensive memo from her prime minister, an intelligence official a “top secret” briefing folder from his agency, and a human rights activist two urgent petition letters from local non‐governmental organizations (NGOs).

While this mechanism increased the realism of the simulation, it also made it easier for negotiators to stay engaged. When negotiators need to “look up” information they would never have to look up in the real world in the middle of a conversation (i.e. consulting a character sheet for guidance on how they think or what they are supposed to be interested in), it breaks the realism of the simulation environment. Even in the real world, however, it is realistic that a negotiator briefly consult a memo that a staffer has prepared or ask a team member to remind him of a specific detail. The counterpart may perceive this as unprofessional, but not as unrealistic. Thus, we tried to convey the most crucial parts of a negotiator’s identity on character sheets, while “outsourcing” technical information and background reading to memos, letters, articles, and other realistically designed and formatted documents. We almost completely restrained from telling any negotiator the exact “feelings” of their assigned character. Instead, we aimed to create an immersive environment that would make it likely that specific emotional dynamics would arise as a natural part of the simulation.

While we provided extensive, deep, and detailed briefing materials to bring people up to speed, we did not encourage negotiators to explore the real‐world conflict beyond the exercise documents. Instead, we explained to them that their fictional universe was based on real events, but that they would get to write the next chapter of Afghanistan and Central Asia based on their decisions in these negotiations. This made many negotiators curious about the backstories of other characters and encouraged them to actively explore the interests and concerns of others. We encourage their curiosity, both because it would fully unleash the potential of the immersive environment and because it is a useful negotiation skill. As Robert Mnookin, Peppet, and Tulumello (2000: 58) advised, “Don’t assume you know the other side’s story. If you think you do, you’re probably wrong. Even if you turn out to be substantially right, you will still be more effective if you begin with an attitude of curiosity about how the other side sees the world.” A deep, immersive environment rewards those negotiators who are curious and creative, because its boundaries are not set by one clear‐cut script that everyone will have read before the simulation starts. Instead, an immersive environment is full of surprises, secrets, and backstories that will feel surprisingly real once it comes to life.

To create nested negotiation networks, we mapped out the actors, interests, and relationships in the Central Asia conflict system, identified the most important constituencies, and developed different “small worlds” that we connected in a networked simulation universe. We modeled a variety of intra‐Afghan societal conflicts, the armed conflict with the Taliban, the phenomenon of terrorism in the wider region, the geopolitics of Central Asia, as well as the internal conflicts and the behind‐the‐table dynamics of important stakeholders.

Characters. The Transition⁶ has seventy‐two individual roles with unique role‐specific instructions. Additional, optional roles are also available. Roles include government officials, legislators, diplomats, insurgents, terrorists, military commanders, NGO activists, clerics, intelligence officers, and others (including, in an optional add‐on, journalists) who represent nineteen nations, five international organizations, eight militant groups, and a diversity of overlapping national, ethnic, political, and other communities.

Rounds. The exercise comprises three separate rounds that each represent one year. By the end of Year Two, NATO forces will leave Afghanistan and the government of Afghanistan will be fully responsible for security. The exercise simulates the negotiation challenges arising in the context of this transition. The outcome may be a negotiated settlement that integrates the Afghan insurgency into the political system while taking the interests of all political groups in Afghanistan as well as those of the states of the region and other stakeholders into account. The outcome could also be an ongoing civil war that would represent a political victory for the insurgent hardliners and other militant Islamists as well as for other actors that would benefit from such a situation. Another possible outcome is a preliminary agreement in which the main parties agree on a “roadmap” to peace.

Camps. The exercise is structured around four different “camps” in which negotiators deal with different aspects of the Afghanistan conflict and with a variety of internal disputes: Afghan government and opposition, the United States of America, the Afghan insurgency and other militant Islamists, and the international community. Several real‐world bodies⁷ meet at given times throughout the exercise to internally negotiate their position on the Afghanistan conflict as well as other issues and to agree on an external negotiation strategy.

For each constituency, we implemented a unique setup and discreet set of rules, developed distinct, conflicting but connected conflict narratives, and provided negotiators with specialized information that disguised other negotiators’ resources and sources of leverage, so they had to ask smart and probing questions to get an idea of the overall conflict system.

Time line. The simulation is situated on a time line of events, such as scheduled meetings, elections, and sessions. This forces actors to identify sources of power depending on the moment and to incorporate deadlines, time pressure, and new opportunities for coalition‐building into their negotiation strategy. Not only does this architecture recreate the idea of different organizational cultures, but it also allows for a realistic information flow over time. No one has the full picture of the interlocking puzzles; negotiations take place under constant uncertainty, and while it is difficult enough to keep track of “your own world” (Afghanistan, United States, etc.), it is close to impossible to understand where the entire negotiation universe is going. Negotiations involving multiple issues and stakeholders are always and constantly unfolding simultaneously (e.g., while the American Secretary of State briefs the U.S. Congress, Afghan leaders will gather in the Grand Assembly, the NATO defense ministers will meet at the North Atlantic Council, and informal sidebars will be conducted between those who are not engaged in formal sessions, etc.).

Negotiators must realize that all these processes are (directly or indirectly, loosely or tightly) connected and that the stakeholders and constituencies their counterparts answer to are real. If the multiple and interlinked nature of the processes generates problems, their solutions will likely be linked as well (Penetrante 2012). The overall conflict system, however, displays a behavior that is different from the actions of its parts. By connecting four worlds in this interlocking manner, we built a negotiation exercise with emergent properties that confronted negotiators with nested negotiation networks similar to their real‐world equivalents.

In complex systems, as Andrea Jones‐Rooy and Scott E. Page (2013: 326) pointed out, “an individual action can have multiple consequences (pleiotrapy) and be influenced by multiple other actions (epistasis).” Most negotiation simulations, however, do not involve any “hard” actions. Participants engage in “soft” actions – asking, answering, presenting, pretending, lying, and so forth, but the simulations lack what Poitras, Stimec, and Hill (2013: 445) called experimental realism: “outcomes are only theoretical. Participants get a deal on paper, but it has little real impact for them — they have, as the saying goes, no ‘skin in the game.’”

To overcome this problem, Watkins (2007: 336) advocated for simulations that are “manageably dynamic:”

Complex negotiations are dynamically path‐dependent. Early decisions ripple throughout the entire system. Only by modeling this phenomenon is it possible to simulate the cascading decision effects in a volatile, uncertain, complex, and ambiguous world. Hence, a dynamic architecture with interconnected, but movable parts underpins The Transition. As negotiators, individually and in groups, make specific decisions, their counterparts’ options either contract or expand in response. Negotiators cannot create actions or resources out of nowhere, they must work with the resources they have, or with those they can unlock through negotiation. In the exercise, they execute decisions by completing and submitting simulation material that reflects a real‐world equivalent, such as resolutions, bills, or commands. For example,

• Negotiators representing militant groups can orchestrate fifteen different types of operations against military and civilian targets, ranging from car bomb attacks against Afghan government officials to attacks against NATO supply routes to military strikes against American forces. Each attack comes with an associated probability of success loosely based on real events and requires a different combination of resources (money, fighters, weapons, etc.) to be triggered. For the most part, insurgent and terrorist negotiators must acquire these resources by means of negotiations.

• The United States of America has an arsenal of military and intelligence options at its disposal (drone strikes, special forces operations, Central Intelligence Agency programs), but the use and effectiveness of these programs depends on internal negotiations within the United States such as negotiations to secure appropriate funding from the U.S. Congress or to coordinate among agencies.

• The United Nations Security Council can provide international troops with a U.N. mandate, impose and lift sanctions, employ election observers to Afghanistan, and pass resolutions on a variety of scenario‐related issues.

• Various states can influence events by providing resources to their nonstate proxies, which increase the leverage of these groups or individuals in their respective internal negotiations.

• Throughout the simulation, participants have the option to work on the three pillars of a comprehensive multilevel agreement for Afghanistan and Central Asia that the U.N. proposed. They also have the option, however, to pursue a variety of other potential (side) deals with various smaller groups of negotiators, and they can pursue these agreements either in addition to, as a prerequisite of, or at the costs of an overall deal.

When dynamic components are triggered – that is, a negotiator or a group of negotiators submits the necessary action card, result sheet, and/or resources to the administrative team – real consequences will be implemented. Terrorist attacks and military strikes may send negotiators to the “hospital” (they are removed from the exercise for a specific time frame) or even “kill” them (they are removed from the exercise without the option to return; their characters are dead), the transfer of money will expand the zone of possible agreement for specific subagreements, hacking attacks may give negotiators access to their opponents’ briefing materials, and so in.

As a result, both experimental and psychological realism (Poitras, Stimec, and Hill 2013) increase substantially. For example, negotiators do not have to be told to be angry – they will feel angry if one of their allies falls victim to an attack orchestrated by a third party and suddenly disappears from the exercise, or if an agreed‐upon deal falls apart in the implementation phase because an important stakeholder refuses to honor the signed contract.

This dynamic architecture distinguishes SMCEs from conventional multiparty exercises and challenges the negotiators’ creativity and adaptability. Events feel more real and tangible. As we tell the negotiators before The Transition is launched,

We ran the prototype of the exercise during the “Advanced Workshop in Multi‐Party Negotiation and Conflict Resolution” at the Harvard Kennedy School, an advanced graduate course for students with prior negotiation training. In the postclass exercise evaluations of its very first run, The Transition received the highest scores in all three evaluation categories of perceived learning value, perceived subjective enjoyment, and perceived usefulness for future negotiation skill‐building courses among the fifteen simulations used in the class. After a total of five runs of The Transition in different settings, however, we have a more nuanced understanding of what works and what does not, and we can draw some conclusions on the use of this simulation in particular and of SMCEs in general.

Systemic multiconstituency exercises are designed to push participants to their cognitive limits. As a result, certain phenomena, which are well known from conflict resolution research, can be expected to arise in such a context, such as in‐group/out‐group dynamics involving differences in perceived hierarchy, status, and gender. We also observed a phenomenon that one may call “emergent groups,” that is, small clusters of negotiators situated between coalitions and networks who were able to build a high degree of internal trust in spite of an overall hostile environment. We recommend having simulation observers in place who are part of the teaching team to keep an eye on these developments to bring them up in the debrief.⁸

We have learned that students are willing to do the extensive preparation work and to make the significant emotional investment it takes to engage faithfully in a time‐consuming and exhausting multiday exercise. Overall participant engagement remained high throughout the full experiment, even when the exercise was placed at the very end of a two‐week‐long, intensive negotiation workshop. At the same time, we learned that some roles in a huge multiparty exercise can not only be more demanding, but they are also perceived as being seemingly more “prestigious” than others. In some situations, participants with what seemed to be more “powerful” roles voiced more satisfaction than those with roles that were perceived as less influential. Some participants had challenges finding their own voice and staying relevant in various sessions and struggled to find their appropriate sources of leverage. The specific roles for which this was the case varied across the different runs of the exercise, which suggests that the simulation design is not the primary reason.

Participants who were assigned roles of militant negotiators often found it both rewarding and difficult to communicate and signal commitments when facing severe power imbalances. In the postexercise debrief, many participants reported their misperceptions about sources of power that did not equal the perceived status of negotiators as a major insight. In many situations, we observed a “one‐down effect” (Donohue and Taylor 2007: 322) emerge in the simulation, that is, individuals in lower‐power role assignments, who believed that they had fewer options on the table than their counterparts, embraced more aggressive strategies that ultimately made it even more difficult to come to an agreement.

To harness these and other phenomena productively and generate learning value for the participants, we recommend instructors who run The Transition or other SMCEs consider the following suggestions:

• build on prior training,

• frame and shape the exercise,

• monitor self‐fulfilling prophecies, and

• conduct a modularized debrief.

Our experience confirmed a finding from Druckman and Ebner: “should teachers prefer to use simulation as a multipurpose tool (for teaching as well as for enhancing motivation and retention), they should not use it as a stand‐alone” (Druckman and Ebner 2013: 76). When The Transition was used as the capstone exercise of an intensive, advanced negotiation workshop, participants made fewer technical negotiation mistakes (such as failing to switch to single‐text procedure to synchronize competing drafts) than the participants who engaged in it as a stand‐alone event. Although the outcome of the simulation scenario (i.e., the final “situation” in the simulated version of Afghanistan and Central Asia) has been vastly different from run to run, the risk of having the participants “set the world on fire” while burning important learning opportunities with it seems to be lower if people have practical negotiation training experience. If a systemic multiconstituency exercise is used as a stand‐alone event, we recommend spending additional time on preparatory sessions. These should introduce basic negotiation frameworks to level the playing field; for example, through shorter, conventional multiparty exercises. A special focus should be on preparing participants for the challenge of staying relevant in networked multiconstituency negotiations, especially in the position of a “below‐the‐line negotiator” with (seemingly) fewer sources of leverage.

Participants must understand that they will experience a learning environment blending role‐play simulation with a military war game and that they must take it seriously and refrain from dismissive jokes or half‐hearted engagement that can ruin the experience for others. A SMCE such as The Transition addresses serious, real‐life issues including terrorism, violence, and religious and cultural conflict; the exercise must be framed accordingly by the instructors. To maximize the pedagogical value, the teaching team should be aware that the exercise participants might confront issues that are “sacred” to them and be prepared to address such issues with sensitivity (Tenbrunsel et al. 2009). For instance, all participants may be asked before the simulation starts if they are comfortable with being assigned a “militant” role and be given the opportunity to opt out of these roles.

If the participants are graduate students, for example, the introductory sessions should be structured differently than if the participants are security practitioners, such as diplomats or senior officials. Preparatory lectures should be used to bring participants up to speed, and these should be tailored to the skill‐building needs and expectations of the respective group. In framing and shaping the exercise, the teaching team must acknowledge the moral and ethical dilemmas that arise within the simulation. The goal is to “challenge students to examine their moral values” without “pressure … to compromise [them]” (Kirgis 2012: 93–94).

Participants often fall into the trap of believing that exercise structures are not malleable and flexible. Many do not readily accept the idea that they should never accept the game as given. These tendencies can lead to self‐fulfilling prophecies. For example, in one run of the simulation, many militants perceived the United States to be extremely powerful, coordinated, and almost invincible. This gradually paralyzed them and made it even easier for the United States to exploit their weaknesses. Students with militant role assignments complained afterward about the simulated “military might” of the United States, which had drones, airplanes, special forces, and various intelligence assets at its disposal, and they expressed how relatively powerless they felt. But during an extensive debrief in which all sides told their stories, many of the same “militants” were surprised to hear about the degree of internal negotiations and domestic conflict the U.S. representatives had participated in and how much luck had been involved in what eventually turned out to be relatively “effective” American decisions. Although members of insurgent groups were usually aware of their own constraints (lack of resources, lack of free movement, internal divisions, outside pressure, lack of diplomatic recognition, etc.), they mostly did not realize the constraints put on their U.S. counterparts (bureaucratic politics, lack of NATO commitment, partisan politics, lack of Congressional support, information leaks, cyberattacks, etc.). Jervis described this phenomenon in a real‐world context: “A common misperception is to see the behavior of others as more centralized, planned, and coordinated than it is” (Jervis 1976: 319). Course instructors have to closely monitor these dynamics in a systemic multiconstituency exercise. Up to a certain point, such simulation phenomena generate useful lessons and can even enhance the post‐exercise debrief. But sometimes negotiators become frustrated, “give up,” and disengage: they perceive their own position as so weak and hopeless that it indeed becomes hopeless. The teaching team must anticipate this tipping point and intervene before it happens. In fact, this is one of the few scenarios for which we recommend intervening in the simulation environment to keep the exercise “on track.”

“Simply playing a game does not automatically result in learning or skill building” (Powers and Kirkpatrick 2012: 53). It is crucial that a systemic multiconstituency exercise comes with an extensive, modularized, and interactive debrief. It must be extensive enough to address all learning goals, most importantly in terms of skill building, that the instructor has selected, as well as the concerns voiced by the participants. It must be modularized because groups of negotiators need a chance to debrief issues that will only concern a small number of participants. It must be interactive because all participants need opportunities to speak and to react to each other. We hence recommend a three‐step debrief that consists of peer‐to‐peer feedback, camp debriefs, and a plenary debrief.

Peer‐to‐peer debrief. This refers to participant‐organized feedback discussions in small groups or teams of two to five negotiators, which provide an opportunity to share personal feedback among those who interacted most closely during the exercise.

Camp debriefs. These are group discussions within the main negotiation constituencies (here: Afghanistan, militants, international community, U.S.) to discuss dynamics that shaped this group’s experience.

Plenary debrief. This is an opportunity for all negotiators to share insights from the two previous debrief stages. They discuss the connections between the different “small worlds,” and the instructors highlight the key teaching points (cognition, networks, decision effects) and connect them and the debrief survey data into comprehensive and applicable take‐aways.

We learned that negotiators greatly appreciate the involvement of negotiation practitioners as simulation observers who then share their comments during the debrief sessions. For instance, we worked with a former U.S. ambassador who had been responsible for international security negotiations similar to the ones featured in The Transition scenario. Students reported that they found a discussion of their “simulation history” in the context of real‐world politics while debriefing their own mistakes and successes helpful and enlightening.

We believe that systemic multiconstituency exercises are useful pedagogical tools that can help close the training and experiential learning gap between the complexity challenges that negotiators confront in the real world and those that students typically encounter in the simulated environment. They allow negotiators to move from transactional bargaining to managing negotiations across all three dimensions (Lax and Sebenius 2006) in a complex system in which not all interests and preferences are static, but constructed and deconstructed over time. They more accurately reflect important real‐world, structural patterns in multiactor/multi‐issue negotiation scenarios, which Sebenius (1992: 24) described as follows:

The complexity of The Transition compels participants to employ a variety of moves and microskills that transcend the “standard repertoire” of negotiation skills they know from conventional exercises. Confronted with cognitive maelstroms, nested negotiation networks, and cascading decision effects, the most skilled participants rely on a combination of leadership, negotiation, and decision‐making skills to generate strategic action. Systemic multiconstituency exercises, if properly employed, are teaching, training, and research platforms in advanced negotiation, leadership, and decision making (see Table Three).

One important contribution of systemic multiconstituency exercises to the field of negotiation science is that they integrate negotiation, leadership, and decision‐making training in one functioning experiential learning platform. We hope to build on this step by developing a taxonomy of negotiation skills that will increase our understanding of how specific skills should be employed differently when negotiations move from two‐party to multiparty to multiconstituency negotiations. (See the  Appendix for a first attempt at developing such a taxonomy.)

We believe that The Transition is uniquely suited to teach about cognitive maelstroms, nested negotiation networks, and cascading decision effects as key challenges in complex negotiations. Given the dynamic architecture of the simulation and the (hopefully unleashed) creativity of the participants, instructors can tailor the exercise frame to their needs and structure the broader teaching environment according to their pedagogical preferences.

Instructors should be clear, however, about the learning objectives. These may be, for instance, to focus exclusively on training expert negotiation skills such as dealing with heterogeneous preference intensities (Martin 1994), parasitic integration (Gillespie and Bazerman 1997), multiparty log‐rolling (Guggenbühl 2012), and negotiation set‐up moves (Sebenius 2009) for which adequate simulations are often missing from the classroom. In public policy classes in particular, the many “stories” that occur during the simulation and involve dozens of individuals and their constituencies can be used in the debrief to illustrate the side effects and unintended consequences of political decisions: “Being aware of the many cascading effects of a particular policy choice can mean more effective policymaking with fewer unintended outcomes” (Jones‐Rooy and Page 2013: 326).

A systemic multiconstituency exercise is a very labor‐ and resource‐intensive undertaking. For example, The Transition requires a nine‐person administrative team, an appropriate venue with seven separate rooms in addition to breakout spaces, a variety of “stage props,” and at least sixty‐eight participants who are fully committed to three days of preparation, negotiation, and debrief. As such, it is most suitable for advanced negotiation courses or training workshops led by experienced instructors who can provide the necessary infrastructure.

Crisis management knows simulations as a tool for improving interagency coordination and responsiveness in case of disasters (van Niekerk et al. 2015) and industrial organizations have used simulations as decision‐making support tools since at least the 1940s (Fontaine et al. 2009: 198). It is hence not surprising that The Transition has so far resonated well with negotiation practitioners from all sectors. Systemic multiconstituency exercises provide them with an interactive platform to strengthen leadership skills beyond at‐the‐table negotiations – skills that become increasingly important in an interconnected world. These exercises are also very valuable for training people in the difficult disciplines of process leadership and mediation: “One of the prime reasons that complex negotiations fail is that the parties never have a clear understanding of the process itself” (Cormick 1989: 126).

Systemic multiconstituency exercises allow participants to experiment with the interdependence of structural forces and individual behavior without running the risk of destroying a “real” complex system (company, government, NGO) in the process. In addition, they deliver the kind of organizational experience Schenk and Susskind (2014: 157) have observed in the context of a role play simulation (RPS) exercise about climate change: “Exercise participants do not just increase their own understanding. They increase their appreciation for the multifaceted nature of decision making in complex environments, and thus their ability to work with other stakeholders to enhance resilience and address vulnerabilities.”

Although the value of RPS exercises in general for teaching purposes is well known, “their use for research purposes is rare and not well understood” (Schenk 2014: 252). Hence, it is too early to comment in depth about the research value of systemic multiconstituency exercises in particular. We speculate that their value lies at the intersection of negotiation science, complexity theory, and network analysis. Given that they are “mini‐universes” that, when properly designed, model patterns of complex systems in the real world, we expect that they may be able to tell us something about how a social system influences behavior, and vice versa, and which negotiation strategies are more effective than others in order to shape the circumstances and expectations that in return shape the negotiator: “How people act and live are shaped, though in no way dictated or determined, by the larger circumstances in which they find themselves” (West 1993: 12). For us, one of the most exciting and gratifying aspects of the exercise has been observing students and practitioners wrestle with the world we created. We were impressed by their creativity and perseverance. We saw that some people are not frightened by complexity, but they thrive on it. We suggest that negotiation scholars and teachers explore fully the positive potential that negotiation complexity holds: “Embedded within the chaos that fosters conflict is the powerful energy of passion that, if properly harnessed, can lead to progress through actionable agreement” (Podziba 2003: 289).