On Being Rigorous and Relevant: A Tribute to Tom Schelling Open Access

I arrived in Cambridge from the small city of Norwalk, Connecticut toward the end of the 1950s to begin my studies as a Harvard undergraduate. I was a kid from a largely working‐class background who had attended ordinary public schools and who had had none of the advanced placement courses, much less the cultural experiences, that are common among those entering college today.

Nevertheless, I did know some things about how the world works. I was fascinated by American politics from an early age. Although my working‐class background predisposed me to favor Democratic candidates, I fully expected Dwight D. Eisenhower to beat Adlai Stevenson in the presidential elections of 1952 and 1956. Early on, I also acquired an interest in world affairs and became intrigued with the operations of the United Nations. The development of the idea of peacekeeping under the inspired leadership of Dag Hammarskjöld as the U.N.’s secretary general from 1953 to 1961 sparked an interest that stayed with me throughout my education and figured prominently in my early years as a social scientist.

At Harvard or through connections made at Harvard, I had the good fortune to work with a galaxy of rising stars, including McGeorge Bundy, Herman Kahn, Henry Kissinger, and Thomas Schelling. During my first year at Harvard, I managed to gain a prized place in Bundy’s freshman seminar, which meant that he also became my tutor throughout that first year. Bundy, who had become dean of the faculty at Harvard at the age of 34 and who had just turned 40 at the time, set a high standard and became a mentor until he left for Washington in 1961 to become the national security advisor in the administration of President John F. Kennedy. It was a heady experience to visit him in his office in the West Wing of the White House.

Kahn, Kissinger, and Schelling, all under 40 at the time, were launched already on extraordinary careers. Two of these men ultimately won the Nobel Prize: Kissinger won the Peace Prize in 1973 for his role in negotiating a ceasefire in the Vietnam War, and Schelling won the Economics Prize in 2005 for his fundamental contributions to the discipline of economics. Kahn, who had recently established the Hudson Institute located in a former mansion in the community of Harmon‐on‐Hudson, New York, was busy thinking and talking about the role of nuclear weapons in foreign policy, an activity that gave him public notoriety and turned him into one of the models for the role of Dr. Strangelove in the 1964 movie of the same name. Those who worked at the institute at the time routinely described him as “Herman‐on‐Hudson,” a reference both to his physical bulk and to the size of his ego.

Although I did not appreciate it fully at the time, the opportunity to work with these people was a remarkable privilege. I learned many things from my association with each of them. Without a doubt, however, the one who made the greatest and most lasting impact on my thinking was Tom Schelling. Kissinger, who had become known as an influential foreign policy analyst following the publication of his book Nuclear Weapons and Foreign Policy in 1957, was busy showing the policy establishment how to integrate the advent of nuclear weapons into the mode of thought known to political scientists and international relations scholars as “political realism” (Kissinger 1957). Kahn, a physicist by training who was working then on what became a book entitled Thinking about the Unthinkable, was well‐endowed with sangfroid, although he lacked a sophisticated grasp of both the domestic and international policy processes (Kahn 1962). Both achieved a high level of public visibility during the 1960s.

Tom Schelling, by contrast, played a different game. He did not aspire to become an influential public official or to achieve name recognition as a public intellectual. He chose instead to devote his time and energy to developing a powerful model for analyzing human interactions ranging from the interpersonal to the intergovernmental, generating in the process provocative insights that were not available using more conventional types of analysis such as political realism. What is more, Tom demonstrated the value of linking this mode of thought to important policy issues. In the midst of the Cold War, he chose to direct attention to opportunities and options for arms control, bringing his analytic perspective on human interactions to bear on this crucial issue and, vice versa, showing how thinking about arms control could contribute new insights to our understanding of human interactions.

I learned from his example not only that powerful ideas provide an extraordinary route to influence over the long haul but also that one can make a compelling case for working back‐and‐forth between theory and practice. These lessons have stuck with me throughout the decades since I encountered Tom, becoming an essential part of my professional activities throughout a long career.

In the fall of 1960, I enrolled in Tom’s undergraduate economics course. I was looking for an economics course that would highlight policy relevance rather than formal modeling of the sort that was pervasive in neoclassical microeconomics. But I knew little about the subject matter of Tom’s course.

As it turned out, 1960 was a watershed year for the development of the ideas that made Tom influential on a global scale. His newly published book, The Strategy of Conflict, became the key text for the course (Schelling 1960). In the class, he referred also to another book published the same year, Anatol Rapoport’s Fights, Games, and Debates (Rapoport 1960). Not only were these books central to the ideas Tom presented in his course; they made such a lasting impression that they have traveled with me wherever I have gone for almost sixty years—they are on my desk as I compose this tribute.

Tom observed at the outset that a large proportion of human interactions take the form of what he called mixed‐motive or competitive‐cooperative situations. In essence, human actors find themselves engaged in situations characterized by a complex combination of incentives to compete for advantages without allowing the resultant competition to undermine the cooperation needed to avoid negative and in some cases highly destructive results. Although Tom is often seen as someone concerned with ways to maximize the gains for individual participants in such situations, it is fair to say that he was equally interested in avoiding outcomes undesirable from the perspectives of all concerned. He was as concerned about avoiding “collective bads” as he was about pursuing the interests of individuals.

To think rigorously about the dynamics of such situations, Tom adopted the apparatus of game theory and made use of a range of game‐theoretic examples including such well‐known models as prisoner’s dilemma, the game of chicken, and battle of the sexes. He did not, however, become a game theorist seeking to derive theorems from relatively sparse sets of assumptions about human behavior. Unlike such analysts as John von Neumann, Oskar Morgenstern, Duncan Luce, and John Nash, who focused on what game theorists know as non‐cooperative games, Tom chose to focus on cooperative games or, in other words, situations in which communication (sometimes tacit rather than explicit) between or among the participants is possible. In so doing, he sacrificed the prospect of arriving at powerful theorems through processes of deductive reasoning. Instead, he developed an innovative and rigorous method of analyzing a wide swath of situations ranging from interactions among family members to interactions between and among powerful nation states.

Starting from this point of departure, Tom (Schelling 1960: 5) observed that “[to] study the strategy of conflict is to take the view that most conflict situations are essentially bargaining situations” (emphasis in original). Because both (all) sides prefer outcomes anywhere in the contract zone (i.e., zone of possible agreement or ZOPA) to an outcome of no agreement, the ability to stake out desired outcomes, and to commit oneself to such positions in a manner that is credible to the other side(s) becomes the focus of bargaining processes. If a participant can convince the other party that he/she is irrevocably committed to a certain outcome, the other party must agree or accept an inferior outcome. So, Tom began to think in a creative way about what he characterized as “committal tactics.”

The result was a brilliant analysis of the sources of credibility in competitive–cooperative interactions. Although some sources of credibility are relatively familiar, Tom opened up new lines of thinking that seem strange to some observers but that expose important facets of communication in such engagements. The idea of the rationality of irrationality, for example, turns out to be fundamental to the operation of nuclear deterrence based on the notion of mutually assured destruction and the doubtful value of retaliating once deterrence fails. Similarly, Tom identified the role of the threat that leaves something to chance or, in other words, a commitment to take actions that are no longer controllable by the party issuing the threat. The idea of a so‐called “doomsday machine” owes much to this insight. He also pointed out the strategy in which one party place bets with a third party that the first party will lose if the bargaining process ends in failure or in serious compromises on the part of the first party. In effect, the bargainer deliberately incurs potential costs (e.g., a loss of face or reputation) to enhance the credibility of his/her commitment in the primary arena.

Tom noticed as well that there are fundamental differences between one‐shot interactions and ongoing engagements, especially in situations that have no natural end points. Formal theory suggests that iterative but finite games do not differ fundamentally from one‐shot interactions because they logically regress starting with the last round and working back to the first round. But in real‐world situations, reputation can play an important role in ongoing engagements. A powerful reason for sticking to commitments made early on is to gain a reputation in subsequent interactions not only as a hard bargainer but also as someone who makes good on the terms of agreements once they are finalized. Interestingly, the phenomenon of reputation may carry over from interactions with a single opponent/partner to others, a condition that reinforces the role of reputation and that may well justify risking no agreement in early interactions to solidify a reputation that can prove beneficial during later encounters.

Building on this analytic foundation, Tom was able to add more insights over the years that not only deepened our understanding of bargaining processes but also shed light on several other dynamics of mixed‐motive situations. A few examples will suffice to indicate the range of Tom’s thinking about mixed‐motive situations and the generative character of many of his insights. As these examples suggest, a particularly interesting feature of these contributions is the extent to which they adumbrated insights that have become prominent in more recent years in the field we now know as behavioral economics (Thaler 2015).

When I took his course, Schelling was thinking intently about the idea of salience or prominence and how norms influence outcomes in mixed‐motive situations. Salience refers to the fact that some outcomes located within the contract zone have features that make them particularly prominent as solutions, whether or not they produce results that conform to any known conception of fairness. Those of us who had grown up in the northeastern United States generally had no trouble “agreeing” to meet in New York City at noon on a specified day at the clock in Grand Central Station, even when no hour or place had been determined in advance and when we had no formal means of communication (this was before the days of cell phones!). But everyone realized immediately that the salience of this solution was culturally limited; strangers to New York would have no way of understanding the prominence of this outcome. So, an obvious question is how to think about salience more universally. The answer is not straightforward, but clearly some solutions are widely regarded as salient, which becomes even more important when we scale up the analysis to the level of interactions between or among nation‐states.

The role of norms is also interesting in this regard. Consider the game known as “Divide the Dollar” in which the first mover is asked to propose a division of a dollar with a second player whose only option is to accept or reject that first mover’s offer. In game‐theoretic terms, the second player should accept any proposed division that yields more than zero, but this is rarely how subjects asked to play this game actually behave. Although first movers seldom offer a fifty‐fifty split, they generally do offer the second player some amount well above zero, and second players sometimes reject the offers even when this means they end up worse off than if they had accepted the offers. What is going on here? Do we hold to a widespread norm that suggests that overexploitation is inappropriate in such situations, and does behavior in such situations vary across time, places, and types of situations?

Having analyzed threats and promises in some detail in The Strategy of Conflict, Tom moved on soon after in a book entitled Arms and Influence to draw a distinction between deterrence and what he called “compellence” (Schelling 1967). Whereas deterrence involves persuading someone not to do something, compellence requires persuading the target to take a positive action. Although they are logically equivalent in terms of conventional utility calculations, it is generally understood that compellence is more difficult than deterrence. Deterrence works when the relevant subject simply does nothing; compellence by contrast requires the subject to behave in some well‐defined way.

An interesting issue arising in this connection is the relationship between the use of threats and promises on the one hand and deterrence and compellence on the other. Most accounts of deterrence center on threats of retaliation; rewards of one kind or another are commonly used in situations featuring compellence. But it is easy to exaggerate this difference. Threats may well play a role in effective compellence, as in (admittedly dicey) initiatives in which sanctions are imposed to persuade an actor to alter some behavior deemed undesirable (e.g., discontinuing nuclear weapons programs).

It would also be interesting to extend Tom’s analysis to consider the possible role of promises in conjunction with deterrence. To take the example of nuclear weapons again, it certainly seems relevant to think about rewarding those who agree to eschew nuclear weapons rather than threatening to punish them if they do take steps toward the acquisition of such weapons.

Moving on, Tom directed his attention explicitly to what we have come to know as collective‐action problems in a book aptly titled Micromotives and Macrobehavior (Schelling 1987). In this book, he extended his earlier work by focusing on recurrent situations in which groups of actors select options that seem rational from an individualistic perspective only to find that the resultant outcome is “Pareto inferior” or, in other words, worse for everyone than one or more of the feasible alternatives. A form of this problem particularly relevant in today’s world is known as the “free rider problem” in which everyone hopes that others will fund the supply of a collective or public good that he/she can enjoy without making a contribution (Olson 1965).

While realists were busy arguing about the necessity of finding a dominant player or hegemon (i.e., an actor who values the collective good more than the cost of supplying it) as the critical factor in such situations, Tom came up with the idea of a “k group” as a means of solving problems of this sort. A “k group” is a small subset of a larger group willing to band together for a period of time to supply a collective good, even though other members of the group get a free ride. Logically, the members of such a subset may be tempted to defect at various points along the way. In terms of coalition building, however, it is clearly easier to forge strong bonds among a small group of players than to persuade all members of a larger group to join forces. As Schelling himself noted in later years, this line of reasoning is particularly relevant to the dilemma of reducing greenhouse gas emissions to limit climate change. This explains the importance many analysts attached to the agreements President Barack Obama and President Xi Jinping reached in the run‐up to the 2015 Paris Agreement.

One thing that made Tom’s analysis easily accessible and broadly persuasive was his remarkable facility in introducing concrete – even homely – examples of various types of behavior likely to occur in mixed‐motive interactions. The behavior of children in interactions with their parents provides striking examples of the rationality of irrationality. Consider the case of a child who throws a tantrum in order to get his/her way in dealing with a parent. The selection of meeting places, even when they impose asymmetric burdens on the relevant players, is a tribute to the power of salience. And the difficulties housemates encounter in overcoming free rider behavior relating to household chores such as taking out the trash make it clear that collective‐action problems occur in almost all walks of life.

At the same time, Tom was alert to opportunities to use his analysis to shed light on much larger public policy issues. During the 1960s, he directed his attention to arms control in particular. President Kennedy had focused during the 1960 presidential election on the so‐called “missile gap” between the Soviet Union and the United States; he promised if elected to place high priority on closing it. For their part, such “realists” as Kissinger were busy assimilating the influence of nuclear weapons into their analyses of the exercise of power in international affairs. A favorite topic of conversation among them was the question of the relative stability of bipolar and multipolar political systems in the nuclear age.

By contrast, Tom started from the premise that even in the depths of the Cold War the relationship between the superpowers had all the hallmarks of a mixed‐motive situation. Several important insights arose from this observation. First, he realized there was a contract zone in this situation because some outcomes were better for both the United States and the Soviet Union than others, while others were worse for both sides. In addition, he noted a counterintuitive feature of the doctrine of nuclear deterrence: should deterrence fail in the sense that one side actually launched a first strike, it was far from clear whether it would make sense for the target of the strike to go through with launching a retaliatory attack. In short, the relationship between the Soviet Union and the United States was simultaneously more dangerous than many observers realized but also more susceptible to mutually beneficial agreements.

The result was a small volume published in 1961 called Strategy and Arms Control written with Morton Halperin, a brilliant young scholar then at the Harvard Center for International Studies who became, among other things, my advisor during the time I prepared my senior thesis (Schelling and Halperin 1961). Actual events – the disastrous 1961 Bay of Pigs invasion, the tensions arising from the 1961 Berlin crisis, and, especially, the sobering experience of the Cuban missile crisis in 1962 – made the logic that Tom and Mort outlined even more persuasive. I do not want to overemphasize the influence of Tom’s mode of analysis on the actual course of public policy, but it seems clear to me that the dramatic shift in strategic thinking that led initially to the Partial Nuclear Test Ban Treaty in 1963 and later to broader strategic arms limitation agreements between the Soviet Union and the United States owes relatively more to the insights flowing from Tom’s thinking about mixed‐motive interactions than to the strictures of the realists about strategies for wielding political power in the nuclear age.

One major lesson I drew from this experience concerns the value of linking analysis to praxis and working back‐and‐forth between the two in a search for deeper insights. There can be no doubt about the value of Tom’s analysis of the dynamics of competitive‐cooperative interactions as a way to frame some of the great issues in international politics, but, vice‐versa, applying his analysis to the case of arms control helped deepen the underlying analysis itself. Thinking about what U.S. Secretary of Defense Robert McNamara famously called the “mad momentum” of the arms race made it clear how tenuous game‐theoretic assumptions about rationality and utility maximization are in most real‐world settings. That is why monitoring, reporting, and verification (enabled by new technologies that make it possible for nations to monitor each other’s compliance without intrusive on‐site inspections) have a key role to play in ensuring success even in cases when the parties should know better than to think about violating the terms of the agreements.

My own work has long since moved on from a focus on “the politics of force” or more generally what analysts commonly call the realm of high politics. But I have never forgotten what Tom’s work taught me about the virtues of joining analysis and praxis. For several decades, I have explored the role of social institutions in addressing needs for environmental governance, especially at the international level where there is no overarching government in the ordinary sense of the term. Mixed‐motive interactions are prominent in such issue areas and the dangers of falling prey to collective‐action problems are pervasive. I have invested much time in applying my thinking about governance to such large‐scale environmental issues as climate change and the loss of biological diversity.

Today, we think increasingly of the Earth as a system and seek to identify what some have described as a “safe operating space for humanity” (Rockström et al. 2009). Natural scientists understandably play a critical role in efforts to address this concern. But the essence of the problem is to find ways to guide mixed‐motive interactions among human actors in the Anthropocene, a new era in which the actions of humans have emerged as a major driving force on a planetary scale. Much like the dangers of the arms race during the second half of the twentieth century, the problems of governing human–environment interactions on a planetary scale have emerged as a central challenge of the first half of the twenty‐first century (Young 2017). They constitute a preeminent contemporary challenge for all those concerned with the dynamics of mixed‐motive interactions.

In 1989, the National Science Foundation asked the National Research Council (the operating arm of the U.S. National Academy of Sciences) to establish a Committee on the Human Dimensions of Global Change. The idea was to supplement the work of natural scientists in the field of Earth system science and to develop a systematic understanding of the role of human actions as drivers of large‐scale changes in biophysical systems, the impacts of such changes on human well‐being, and the responses of humans to major environmental changes. The NRC invited me to serve as the founding chair of the committee, a challenge that I accepted at once. It quickly became apparent that determining the composition of the committee would be a real challenge. The NRC staff and I began an extensive process of consultation as we sought to put together a group that would both function effectively as a team and comprise articulate representatives of most of the social sciences and several of the natural sciences. Early on, we agreed to invite Tom to join the committee. Tom, who was then in the process of retiring from Harvard University and moving to the University of Maryland, was close at hand, liberated from some of the conventional duties of senior faculty members, and as intellectually creative as ever.

I found the interplay between Tom’s perspectives and those of marine ecologist Elliott Norse memorable and instructive. Norse, a pioneer in developing the concept of biological diversity, had an encyclopedic knowledge of ecosystems and a passionate desire to protect them from the destructive force of human actions. He was critical of the seemingly cold‐blooded way in which Tom dissected the dynamics of human actions and analyzed how human actions could lead to undesirable macrobehavior. Tom responded by suggesting innovative ways to steer human behavior rather than holding out hope for transformative change in human nature. In my view, the dialectic between these approaches – a passionate concern for the integrity of the planet’s ecosystems and a calculated assessment of how to steer human behavior without expecting to transform it – was an important source of creativity in the work of the committee.

The committee’s first major product was a book entitled Global Environmental Change: Understanding the Human Dimensions, which I co‐edited with Paul Stern and Daniel Druckman (Stern, Young, and Druckman 1992). Since the book’s publication, the human domination of the planet has become more apparent as has the need to manage human actions that generate large‐scale environmental changes (e.g., climate change). The NRC committee continues to play an important role, advising the U.S. Global Change Research Program and participating in the Future Earth Program, an umbrella organization that represents the international global change research community.

By engaging in theoretically significant research that at the same time was clearly relevant to the great issues of the day, Tom Schelling landed squarely in what those who think about the uses of knowledge call “Pasteur’s Quadrant” (Stokes 1997). He developed a mode of analysis, articulated initially in The Strategy of Conflict, that provides useful tools for understanding human behavior in a wide range of contexts. He used these tools to shed light on a variety of public issues from the effort to address questions of arms control in the depth of the Cold War to the effort to come to terms with global change in more recent times. He is less well‐known to members of my generation than Henry Kissinger or Herman Kahn, for example, but the continuing influence of his thinking on a wide range of important issues is enormous. To my way of thinking, there is something both personally satisfying and socially significant about this sort of career success.