Roundtable 10-4 on Perception and Misperception in International Politics and on How Statesmen Think: The Psychology of International Politics

Robert Jervis is at the same time a giant and a gadfly, a leader and a subversive in the field of international relations. In his career, Jervis has often been very much a theorist in the mainstream political science tradition. In some of his most famous works—including “Cooperation Under the Security Dilemma” and The Meaning of the Nuclear Revolution—Jervis has shown his skill at creating deductively derived theories about how states should respond to structural and technological changes in international security affairs.[1] Those works are extraordinary and have made an enormous contribution to the literature. But Jervis often notes with frustration that actual policy makers often diverged from the expectations and prescriptions of his theories. He laments that, despite the inescapable background condition of mutual nuclear vulnerability, the two Cold-War superpowers still developed destabilizing offensive nuclear weapons designed to target their enemies’ arsenals, planned to fight ‘limited’ nuclear wars of various levels of intensity, and obsessed about local conventional balances of power around the world. Jervis thought it would have been safer and less fiscally burdensome if Washington and Moscow had fully accepted the condition of mutually assured destruction and properly understood the stabilizing effects that condition should produce at all levels of potential military conflict.[2]

H-Diplo | ISSF Roundtable, Volume X, No. 4 (2017)

H-Diplo/ISSF Editors: Thomas Maddux and Diane Labrosse
H-Diplo/ISSF Roundtable and Web/Production Editor: George Fujii

Robert Jervis. Perception and Misperception in International Politics, New Edition. Princeton University Press, 2017. ISBN: 9780691177434 (hardcover, $99.50); 9780691175850 (paperback, $29.95)

Robert Jervis. How Statesmen Think: The Psychology of International Politics.Princeton University Press, 2017. ISBN: 9780691175058 (hardcover, $95.00); 9780691176444 (paperback, $29.95).

Published on 8 December 2017



© Copyright 2017 The Authors

Introduction by Thomas Christensen, Princeton University

Robert Jervis is at the same time a giant and a gadfly, a leader and a subversive in the field of international relations. In his career, Jervis has often been very much a theorist in the mainstream political science tradition. In some of his most famous works—including “Cooperation Under the Security Dilemma” and The Meaning of the Nuclear Revolution—Jervis has shown his skill at creating deductively derived theories about how states should respond to structural and technological changes in international security affairs.[1] Those works are extraordinary and have made an enormous contribution to the literature. But Jervis often notes with frustration that actual policy makers often diverged from the expectations and prescriptions of his theories. He laments that, despite the inescapable background condition of mutual nuclear vulnerability, the two Cold-War superpowers still developed destabilizing offensive nuclear weapons designed to target their enemies’ arsenals, planned to fight ‘limited’ nuclear wars of various levels of intensity, and obsessed about local conventional balances of power around the world. Jervis thought it would have been safer and less fiscally burdensome if Washington and Moscow had fully accepted the condition of mutually assured destruction and properly understood the stabilizing effects that condition should produce at all levels of potential military conflict.[2]

The two books reviewed in this roundtable demonstrate that Jervis is, however, much more than a mainstream IR theorist. Jervis is also an honorary diplomatic historian (and not coincidentally, he has been a major force in both creating and sustaining the ISSF/H-Diplo website). As such, in these books Jervis seems interested in explaining how leaders actually behaved, rather than how they should have behaved according to a pure, context-free theoretical logic. His real rebellion against mainstream political science is his insistence that decision makers, at the end of the day, are human: they suffer from cognitive limitations, biases, and personality quirks. Those individual characteristics often make them poor subjects for deductively derived, structural explanations for how rational actors should interact given objective changes in the environment in which they operate. That lesson is strongest and most clearly laid out in Perception and Misperception. As Jervis recognizes in the new preface, the book itself does not have a single clear theoretical take. This is true, unless, of course, one considers intelligent and historically rooted skepticism about clear theoretical takes themselves to be a strong theoretical position.

How Statesman Think, an updated compilation of previously published works, continues in this tradition, but in a very real sense brings together Jervis’s two skills as a deductively oriented social-science theorist and an inductive diplomatic historian. Jervis is enamored of general theories of coercive diplomacy, like Thomas Schelling’s Nobel Prize–winning game theories of conflict, and general theories of human psychology, like Tversky and Kahneman’s Nobel Prize–winning work on prospect theory.[3] But Jervis is a master at demonstrating that while the theories themselves may be general, it takes a tremendous amount of detective work to apply them to real-world cases. To illuminate the parsimonious power of the elegant theories, we need to get into the particular psychological makeup and perceptions of the leaders in question.

In Schelling’s theoretical work on coercive diplomacy, a core concept is the perceived status quo that can either be preserved through deterrence or changed through compellence. Since Schelling deems successful compellence much more difficult to achieve than successful deterrence, the distinction could hardly be more important. So the need to understand leaders’ varying perceptions of the status quo is built into the theory in a way that strongly privileges scholars like Jervis, who are steeped in diplomatic history, over the vast majority of game theorists in political science, who have focused almost exclusively on mathematics in their intellectual training and simply assume as given many things that in the real world vary wildly and consequentially. The same can be said for one of the most important lessons of Schelling’s game theoretic work, which is repeated often in the essays in How Statesmen Think: successful deterrence requires credible threats of punishment if proscribed behavior is adopted; but it also requires credible assurances that the punishment will be withheld if the perceived status quo is preserved. Without such assurances, the target has no reason to comply with the demands attached to the threat. There is always tension between these two equally important missions in coercive diplomacy, and that tension is captured by the concept of the security dilemma: a country’s individual efforts to secure itself through defense buildups and deterrence can be misread by another state as fundamentally hostile and aggressive, leading to a countering effort that leaves both sides less secure. To understand successful and failed instances of deterrence (or compellence), we need to comprehend not only the threatening and reassuring signals sent but how those signals are perceived by the target. In his qualitative research, Jervis is careful. Such care, however, is rarely reflected in the coding of cases for large n databases in the mainstream security studies literature, which ironically prides itself on superior scientific rigor.

Just as Schelling’s theories are broadly applicable but difficult to apply in every case, so is Tversky and Kahneman’s prospect theory. Since humans behave very differently when protecting what they have than they do when gaining new things, how issues are framed by individuals as being in the realm of gains or the realm of losses is all important. Just as one can determine the perceived legitimate status quo in Schelling’s models only through careful empirical research into leaders’ psychology in every case, we need to know a lot about the psychology of individual actors in the political world to determine what they themselves would consider a gain from the perceived pre-crisis status quo and what they would consider a loss.

The impact of Jervis’s work in the academy and beyond is reflected well in the essays below, which I will allow to speak for themselves. Scholars of coercive diplomacy like Professor Dianne Pfundstein Chamberlain find Jervis’s work an essential starting point for their own research because they understand how signals can be misread and how even much more powerful countries can fail to successfully coerce weaker ones. Jervis’s work has inspired political psychologists like Professor Rose McDermott, who believes that we need to extend that psychological research agenda to include factors that are even harder to consciously manipulate and correct through policy adjustments than leaders’ misperceptions and biases. These include human emotions and biology, avenues of inquiry that Jervis himself begins to wrestle with in How Statesmen Think. Scholars who are former foreign policy practitioners, like Professors James Steinberg and Philip Zelikow, find Jervis’s work very useful for both policy making and intelligence analysis.

As a former official myself, I agree with them. Jervis’s theoretical toolbox is much more useful to policy makers than most theories in international relations because contingency is built into his generalizable approaches, so there is plenty of room to allow for consideration of what policy makers know from experience to be important: individual leaders matter; context matters; diplomacy needs to be carefully crafted to demonstrate both resolve and restraint; and how the other side thinks about an international crisis or problem is as important, and sometimes more important, than how one’s own side thinks about such issues.

It seems fitting that these two books were published in the same year that Richard Thaler won the Nobel Prize for his work in behavioral economics, which, like the works of Jervis and Tversky and Kahneman, treats economic actors as full humans, rather than robotic utility maximizers. It may also be fitting that Jervis’s books were published in the first year of Donald Trump’s presidency. Many scholars who suggest in their work that what really matters in domestic and international politics are broad structural pressures on political actors, not those actors’ individual characteristics, now express uncharacteristic worry over this particular individual’s presence in the Oval Office. Perhaps deep down, they think more like Robert Jervis than their published works might suggest.


Robert Jervis (Ph.D., California at Berkeley, 1968) is the Adlai E. Stevenson Professor of International Politics and has been a member of the Columbia political science department since 1980. In 2000-2001, he served as President of the American Political Science Association. Professor Jervis is co-editor of the “Cornell Studies in Security Affairs,” a series published by Cornell University Press, and a member of numerous editorial review boards for scholarly journals. His publications include Perception and Misperception in International PoliticsThe Meaning of the Nuclear RevolutionSystem Effects: Complexity in Political and Social LifeAmerican Foreign Policy in a New Era, and Why Intelligence Fails: Lessons from the fall of the Shah and Iraqi WMD, and several edited volumes and numerous articles in scholarly journals. His latest book is How Statesmen Think: The Psychology of International Politics.

Thomas J. Christensen is William P. Boswell Professor of World Politics of Peace and War and Director of the China and the World Program at Princeton University. From 2006-2008 he served as Deputy Assistant Secretary of State for East Asian and Pacific Affairs with responsibility for relations with China, Taiwan, and Mongolia. His research and teaching focus on international security, China’s foreign relations, and the international relations of East Asia. His most recent book is The China Challenge: Shaping the Choices of a Rising Power (New York: WW Norton, 2016).

Dianne Pfundstein Chamberlain, Ph.D., is an Associate Research Fellow with the Arnold A. Saltzman Institute of War and Peace Studies, School of International and Public Affairs, Columbia University. She is the author of Cheap Threats: Why the United States Struggles to Coerce Weak States (Washington, D.C.: Georgetown University Press, 2016). Her policy work has recently appeared online with Foreign Affairs, The National Interest, and War on the Rocks. Pfundstein Chamberlain works on signaling and interstate coercion, and she is currently writing a book about British decision-making in 1940.

Rose McDermott is the David and Mariana Fisher University Professor of International Relations at Brown University and a Fellow in the American Academy of Arts and Sciences. She received her Ph.D. (Political Science) and M.A. (Experimental Social Psychology) from Stanford University and has taught at Cornell, UCSB, and Harvard. She has held fellowships at the Radcliffe Institute for Advanced Study, the Olin Institute for Strategic Studies and the Women and Public Policy Program, all at Harvard University. She has been a fellow at the Stanford Center for Advanced Studies in the Behavioral Sciences twice. She is the author of four books, a co-editor of two additional volumes, and author of over two hundred academic articles across a wide variety of disciplines encompassing topics such as experimentation, emotion and decision making, and the biological and genetic bases of political behavior.

James B. Steinberg is University Professor of Social Science, International Affairs and Law at Syracuse University, where he was Dean of the Maxwell School from July 2011 until June 2016. Prior to becoming Dean he served as Deputy Secretary of State (2009-2011) From 2005-2008 Steinberg was Dean of the Lyndon B. Johnson School of Public Affairs. From 2001 to 2005, Mr. Steinberg was vice president and director of Foreign Policy Studies at the Brookings Institution. Mr. Steinberg served in a number of senior positons under President Clinton, including deputy national security advisor and director of the State Department’s Policy Planning Staff. He is the recipient of the Joseph J. Kruzel Memorial Award, American Political Science Association (2014), the CIA Director’s Medal (2011) and the Secretary of State’s Distinguished Service Award (2011). Steinberg’s most recent books are A Glass Half Full: Rebalance, Reassurance and Resolve in the US-China Relationship (Brookings Institution Press 2017) and Strategic Reassurance and Resolve: US-China Relations in the 21st Century (Princeton University Press: 2014) both with Michael O’Hanlon. Professor Steinberg has an A.B from Harvard University (1973) and a J.D. from Yale Law School (1978).

Philip Zelikow is the White Burkett Miller Professor of History at the University of Virginia, where he has also served as dean of the Graduate School and director of the Miller Center of Public Affairs. His scholarly work has focused on critical episodes in American and world history. His books include Germany Unified and Europe Transformed; The Kennedy Tapes; and Essence of Decision. Before and during his academic career he has served at all levels of American government. His federal service has included positions in the White House, State Department, and the Pentagon. His most recent full-time position was as the Counselor of the Department of State, a deputy to Secretary Rice. He also directed a small and short-lived federal agency, the 9/11 Commission, as well as an earlier bipartisan commission on election reform, chaired by former presidents Jimmy Carter and Gerald Ford, that led to the Help America Vote Act of 2002. A former member of the Intelligence Advisory Boards for President George W. Bush and for President Barack Obama, he is currently a member of the Defense Policy Board.


Review by Rose McDermott, Brown University

Perception and Misperception has been a classic volume since its original publication in 1976, and the new edition provides a major new preface that stands as an independent bookend to the iconic work. It provides a brief overview of new areas of inquiry including evolutionary psychology, genetics, and neuroscience, as well as noting the important ways in which the study of emotions has increased since its original publication. How Statesmen Think provides a lightly edited compilation of previously published journal articles which all fall loosely under the umbrella of political psychology. It is a true testimony to the reach of Robert Jervis’s intellectual scope that while this volume alone exceeds what most scholars accomplish in both quality and quantity over a lifetime, it encompasses only part of his scholarly productivity over the course of his career, with equally influential work existing in many other areas, including nuclear security and strategy. Taken together, these two works provide a fascinating and substantive journey through the sophisticated, subtle and nuanced mind of one of the most important scholars in the history of international relations.

It is no small exaggeration to say that Perception and Misperception largely created and defined the individual level of analysis in international relations, setting out the scope of all the work that followed. It was, to be blunt, both seminal and definitive and, as a result, may have actually limited further work in the field because everyone who read it felt like there was little that could be added. As a result, it is nothing if not both daunting and humbling to try to comment on the new version, or to say something that has not already been said before. Although earlier work had been done applying psychological concepts to political phenomena, most notably the work of Harold Lasswell in the 1930s, none of that work really seeped into the subfield of international relations. [4] Certainly there are both theoretical and empirical reasons why this earlier work did not exert more of an impact. Theoretically, most of this work grew out of the fascination with Freudian psychoanalysis which permeated the intellectual environment of the social sciences at the time. The best of this work, such as Alexander and Juliette George’s master psychobiography of Woodrow Wilson, applied these concepts to individual leaders; this connection proved natural because psychodynamic notions had originally emerged from individual cases and are best suited methodologically to such applications.[5] But aside from the study of leaders, it proved seemingly insurmountable to apply concepts developed around individual drives, motives and feelings to understand the more complicated, interactive and structurally constrained actions of nation-states. All of this changed with the publication of Perception and Misperception, which opened up an entirely new way of studying the influence of the individual on state action. This work clearly benefitted from ongoing developments in social and cognitive psychology, which introduced an examination of human decision-making that was neither grounded in Freudian notions of sex and death, nor restricted to the simple stimulus-response paradigms which dominated the early days of the behavioral revolution. Drawing on this work, Jervis was able to examine the ways systematic and predictable biases in the human decision-making apparatus could influence leaders and enlighten our understanding of international relations.

Jervis began the new preface as a talk, in response to an invitation of mine to come to Brown as part of a year-long speaker series considering the influence of psychology on other disciplines. In so doing, he took the opportunity to re-visit the ways that he might have written Perception and Misperception if he were to write it now. Jervis writes that the reason he is not re-writing the entire book in wake of recent developments now is that although new findings and methodological advances in psychology have emerged since the original publication of the book, he does not think that such advances have “fundamentally undermined what I said. Although the book is no longer completely up-to-date, I do not think it is misleading” (xxi). As with most of his observations, he is entirely correct with this one. Indeed, if anything, many of the new findings have only served to offer additional empirical strength for many of the interpretations and implications he presciently raised in the original volume.

In the new preface, Jervis writes that the main way in which his thinking around these issues has changed since the publication of the original edition is that he has become “more convinced…of the power of our preexisting beliefs to shape the way we see new information. Indeed, I think expectations and political and psychological needs….are the two strongest and most general influences on us.” (xxx) As he notes in the introduction to How Statesmen Think, the first of these considerations used to be considered unmotivated while the latter were considered motivated. Modern theorizing in psychology rejects such dichotomies because it tends to see such mechanisms as so intricately intertwined as to be fundamentally indistinguishable, but the labels and categories matter less than the importance of the concepts, which often benefit from independent consideration of their influence on policy makers. Of course, any comprehensive analysis of the first of these considerations must draw heavily on the work of Amos Tversky and Daniel Kahneman,[6] as indeed Jervis does in the second pairing of articles in How Statesmen Think. But the second half of Jervis’ depiction of what drives us harkens back to the importance of emotions that Freud originally emphasized, albeit in a more modern instantiation. But of course Freud was not wrong in his argument that feelings can prove driving factors not only for beliefs but behavior as well, even if we now understand the repertoire of emotions to extend far beyond the desire and destruction Freud originally identified. And indeed, the work of Tversky and Kahneman depends in a fundamental if unacknowledged way on the insights provided by Freud because of his creation of the notion of the unconscious. Although the later psychologists would never characterize the content or operation of the unconscious in any way resembling Freud’s notion, the entire field of neuroscience depends on an acceptance of the idea that a lot, if not most, of what we think we know about ourselves, and why we do what we do, is below the level of our conscious awareness and perhaps beyond our capacity to understand.

Among the topics that Jervis points to as new developments since the original publication of Perception and Misperception are the areas I have spent the last decade primarily exploring in my own work, including the influence of evolutionary psychology and genetics on political preferences, attitudes and behavior.[7] In Jervis’s categorization in the new preface, he includes this discussion under the topic of biology, which along with emotions, falls under the rubric of unconscious processes. This is analytically useful if, like the earlier division into motivated and unmotivated biases, such separation is certainly artificial in practice. Emotion and biology are intricately intertwined, just as both processes can exist in the conscious as well as unconscious realms. Anyone who has been sick, even with a bad case of the flu, knows that biology can interfere with conscious processes in an explicit way, just as we have all had experiences of believing, however falsely, that we know why we feel the way we do about a particular person, event or policy. These forces are reciprocal and integrated.

But, in addition to a more extensive inclusion of emotional impetuses which Jervis discusses, there are a couple of aspects of his argument I would push perhaps a bit farther than he does. He begins by arguing that Perception and Misperception was not atheoretical, but that it was not a theory. He then goes on to say that to develop a more unified theory, scholars would either have to unify deviations from rationality or build a new theory from the ground up without regard to rationality. Perhaps this is splitting hairs, but I would argue that one can accomplish the latter, assuming a non-economic model of rationality. Specifically, evolutionary theory provides a parsimonious theory (which Jervis notes he prefers on grounds of both elegance and memory in the introduction to How Statesmen Think) from which scholars can derive testable hypotheses. Drawing on evolutionary theory can unify seeming deviations from economic rationality through an explicit recognition of the kind of rationality driven by the goal of biological survival, which of course often privileges children over oneself. Such a theory can provide powerful predicative power while nonetheless frequently expecting outcomes which appear to be at odds with our economically truncated notion of rationality. Models resting on evolutionary theory have cohered a remarkable number of disciplines, from medicine through biological anthropology, biology, and psychology, around a small and simple set of hypotheses which can be derived and empirically tested. There is no reason why such a goal cannot be attempted in the realm of international relations. Such a model would rest on empirical validation that is much broader and more extensive, not to mention more explicitly multidisciplinary, than any of the existing ‘isms’ can claim.

Jervis argues that the limitations of evolutionary models lie in their lack of variance, because they rest primarily on the search for human universals in the psychological mechanisms that shape decision making. While this is true, some of those universals, for example emotions such as fear, appear in quite variant forms and circumstances, depending not only on the environmental triggers but also on individual dispositional variance. To take a more concrete example, all typical humans have hair and eyes. But some eyes are blue while others are brown, just as some hair is blond while others is black. One need not abandon a notion of variance in order to appreciate that difference sits on a foundational physiological platform of commonality. Human nature may be constant, but one of its constants lies in its sheer variance across individuals. Those variants may exist within a series of structural constraints delimited by biological reality, but that is no different than states existing in an anarchic world, and that reality did not stop Realists from developing a parsimonious theory to characterize state behavior. In other words, employing evolutionary models in concert with the methods and strategies used by geneticists to uncover the predispositions for disease offers a potential glimmer of the way such a unified theory of the psychological foundations of decision making around politics might emerge and proceed. Such a model might help us explain otherwise seemingly inexplicable or contradictory behaviors. To take one example, Jervis notes how often people want to oversell policies they like and seek to sublimate negative aspects of them in an attempt to avoid value trade-offs. Citing Trivers,[8] Jervis argues that such proclivity helps people build self confidence that allows them to go into dangerous and prolonged ventures, including wars. Richard Wrangham has argued that military incompetence can be seen as advantageous from an evolutionary perspective; overconfidence can be over-selected if bluffing allows for victory without battle enough times to provide a reproductive advantage to those possessing such traits. [9] This may seem odd, but evolutionary theory depends on the traction provided by the sheer neurocomputational power provided by billions of people acting over millions of years. Even tiny advantages can accumulate over long periods of time to shape human psychological mechanisms, having profound implications on traits as ubiquitous and important as out-group discrimination and warfare. And while Jervis is most certainly right that leaders are not normal, warfare certainly has been. Combat has existed in all times and all societies and it requires no stretch whatsoever to assume that small comparative advantages in performance would have exerted an enormous influence over time, particularly in the age before modern weapons or medicine.

For me, the most existentially troubling aspect of the work on the profound influence of unconscious biases of whatever sort on human decision-making lies in a question that Jervis raises in passing but does not address systematically; this conundrum surrounds the extent to which any of us can exert any real free will independent of forces that are beyond our control, and more often than not are driven by chance, such as when and where and to whom one is born. In emphasizing that he would place more emphasis on unconscious bias now than he did when he originally wrote Perception and Misperception, Jervis writes, “Whether people are aware of the stimulus or not, the implications are disturbing not only for rigorous notions of rationality, but for ideas about responsibility, self-control and even our notions of ourselves” (vi-vii). And he ends this section with the astute observation that, “Most of us believe that we have arrived at our most deeply held beliefs through careful thought and do not want to recognize that if we had been brought up by different parents, let alone in a different society, our fundamental values and attitudes would be very different” (ix). As Jervis notes in his discussion of how liberals and conservatives differ in their acceptance of genetic determinism, by highlighting distinctions in their relative responses to arguments that attribute intelligence and homosexuality to such forces, few want to believe that biology drives destiny across all domains. And yet if personal and professional needs, often instantiated in childhood before conscious awareness even arises, drives so many policy and other preferences unconsciously, it becomes more challenging to find a clear way to argue for the possibility of positive change through the intervention of enlightened civil deliberation. But if anyone’s work can shine light in the darkness, Jervis’s masterwork provides the brightest lamp to help us find the keys hiding in the darkness of our collective drunkard’s search for enlightenment.


Review by Dianne Pfundstein Chamberlain, School of International and Public Affairs, Columbia University

It is both an honor and a terror to review Robert Jervis’s new publications. It is an honor because Perception and Misperception was one of the first books on international politics that I read as an undergraduate and the one that helped to convince me that I had made the right choice in attending graduate school. It is a terror for exactly the same reasons. It would be difficult to overstate the extent to which Perception and Misperception has served as a foundational work in the field of international politics or the extent to which it has influenced the thinking of generations of students, scholars, and (one hopes) political leaders. The key insight from the book—that our preexisting beliefs affect the way we process new information and the conclusions we draw from it—remains as relevant now, as the United States struggles to respond to the nuclear posturing of the North Korean regime and to Russian incursions into Eastern Europe, as it did during the Cold War.

This second edition of Perception and Misperception includes a new 85-page preface in which Jervis reflects on the original text and on developments in the field of political psychology. Although the research on political psychology has exploded in the decades since Jervis wrote the original book, he concludes that these new developments do not fundamentally alter his original insights. Thus the text itself remains unchanged from the original version released in 1976. Jervis admits that Perception and Misperception does not provide a unified theory of perception, and he does not attempt to provide one in the preface (xxi-xxii). He concedes that if he were to write the book again, he would place much greater emphasis on the extent to which emotions and motivated biases affect our ability to process information (lxxxviii).

Although the text itself remains unchanged, this new preface does provide new historical examples that tie Jervis’s original insights to more recent events. These include examples from the end of the Cold War, from Jervis’s analysis of the intelligence failures that preceded the 2003 Iraq invasion, and from the Obama administration. This will no doubt prove useful when teaching the book to undergraduates for whom the invasion of Iraq is at best a hazy memory of evening news broadcasts from early childhood. It also demonstrates that the text is as useful and important for understanding world politics as it was when it was first published forty years ago.

One of the insights from the new preface to Perception and Misperception that poses the greatest challenge to current research on international politics is Jervis’s critique of Bayesian logic, which is derived from a theory of probability attributed to the Reverend Thomas Bayes and published in 1763, two years after his death. Although Bayesian logic is often specified and modeled in formal terms, Jervis explains the basic insight of this approach quite clearly and succinctly: “the direction and degree to which I should change my beliefs depends on how certain I am of the validity of my original views as modified by the extent to which the new information is diagnostic (i.e., discriminates between my view and some alternative)” (xlvii). In essence, Bayesianism focuses on the ability of new information to alter our beliefs about the type of actor or situation we are facing, given what we previously believed about the state of the world. For example, according to standard Bayesian logic, we should be willing to change our assessment of an opponent in the face of new information that discriminates between different types of actors (aggressive or security-seeking). Jervis’s key insight, however, is that one’s beliefs (or ‘priors’ in Bayesian terms) are not independent of the information that would enable one to discriminate between different actors or situations: “In most political cases, the prior beliefs and judgments of the diagnosticity of the evidence are not independent, contrary to the assumptions of standard Bayesianism, and this severely limits the applicability of this approach” (l). Because the way we process and assess new pieces of evidence is affected by our preexisting beliefs, we cannot assume that actors will update their assessments in the face of new information in the way that Bayesianism would suggest. In other words, we generally see what we expect to see, and continue to see it in the face of much new information. This is a convincing and important critique of Bayesian logic, and it must be taken seriously by anyone who finds this approach compelling.

How Statesmen Think is intended to serve as a companion volume to Perception and Misperception. This new book presents a collection of twelve essays in which Jervis explores various topics in political psychology and their implications for international politics. The essays have all been published previously as articles or book chapters and have been edited for presentation in this volume. They are divided into four sections according to their content. The first section provides a comprehensive and relatively concise introduction to the major topics and concepts in political psychology. Part II focuses on heuristics and biases, while Part III presents more specific arguments about the implications of political psychology for international politics. Part IV focuses on topics related to national security including deterrence, crisis stability, the domino theory, and the end of the Cold War.

These essays are rich with insights about international conflict that can be readily employed to assess current events. For example, Chapter Ten, “Psychology and Crisis Stability,” suggests that psychological biases affect a leader’s ability to manage and defuse crises. One such bias is that “the way people process information is likely to lead them to overestimate the likelihood that the other is about to attack,” (219), which can augment pressures to strike first rather than back down in a crisis. Jervis’s arguments about perceptual biases and crisis stability suggest that it may be more difficult than we imagine for any leader to keep a crisis from escalating. One cannot help but think about the United States’ current struggles to communicate with the North Korean regime in the midst of its latest round of missile tests. Similarly, I was struck by an important insight in Chapter 5, “Signaling and Perception: Projecting Images and Drawing Inferences.” Jervis argues that signaling and perception must be studied together. Most scholars treat them as separate issues (and I too have been guilty of this), but Jervis correctly points out that we cannot assess whether our own signals will be effective without considering how the other actor perceives our signal, with all the distortions in perception that go along with that (107-108). We spend a lot of time debating how President Donald Trump can carefully calibrate a response to North Korea’s latest missile test in the hope that the response will communicate clearly and precisely to the North Korean regime. We often forget, however, that the receiver of U.S. signals may interpret them far differently than we imagine. Studying signaling in isolation from perception can never yield more than an incomplete theory of interstate behavior.

There is some overlap between the material covered in Perception and Misperception and the essays in How Statesmen Think, and even among the twelve essays in the latter volume. This is understandable given the fact that the twelve essays were originally published as stand-alone pieces. Likewise, the reader encounters the same historical illustrations on a few occasions—for example, the failures of perception prior to the Japanese attack on Pearl Harbor, and the intelligence failures that preceded the 2003 invasion of Iraq. These examples are important and effective, but this does have the effect of producing an occasional feeling of déjà-vu for the reader. On the other hand, this may prove very useful for readers who are dipping their toes into the waters of political psychology for the first time.

In the end, I found that revisiting the ideas in Perception and Misperception and How Statesmen Think was both familiar and troubling. Familiar, because the insights about how our beliefs and biases affect our ability to process new information and the conclusions we draw from it are just as relevant today as they were in 1976. No doubt Perception and Misperception will retain its status as a foundational work for many years to come. It should serve as the starting point for anyone who wishes to understand the role that perception and beliefs play in international politics. How Statesmen Think augments these insights very effectively, and each of the chapters could serve independently as an authoritative source on political psychology and international politics.

It was troubling, however, because these books force one to confront a series of unsettling realities. Evidence does not speak for itself, and people process information in such a way that they are unlikely to update their assessments in the face of most new information. They will also work hard to try to avoid confronting painful tradeoffs at the expense of more informed decision-making. People do not behave in the way that straightforward rational choice or Bayesian perspectives would like us to believe. Given the evidence that people and leaders will tend to misperceive the world around them, what do we do about this? Hope that more self-awareness will lead to better outcomes? Or does Jervis’s work suggest something along the lines of classical realist arguments about human nature—that the human brain is flawed, and thus we cannot hope that people will be able to process and respond to the world in anything other than a deeply flawed manner? I do not have the answers, but I do know that Perception and Misperception and How Statesmen Think have gotten us pointed in the right direction.


Review by James B. Steinberg, University Professor of Social Science, International Affairs and Law, Syracuse University

Few would question the extraordinary contribution Robert Jervis has made to both the study and to the practice of international relations. His books and articles are required fare in the graduate school syllabi and among the most frequently cited in the field. His ground-breaking insights in the application of psychology to the study of international relations represented an important alternative to the structural IR theorists who came to dominate political science with their focus on the nature of the international system rather than the nature of the human beings who make up the system–and most importantly, who make the decisions that drive the system.[10] For this reason, Jervis’s work has enormous appeal for two very important audiences–the diplomatic historian and the diplomatic practitioner whom the historian studies. The timely reissue of his seminal Perception and Misperception in International Politics (first published in 1976) with an extended new preface and an updated collection of some of his most influential essays, How Statesman Think, offer an opportunity to revisit and re-appreciate the significance of Jervis’s body of work for both communities.[11]

Jervis’s point of departure is what seems to be a straightforward observation – “the broadest justification for looking at how leaders perceive, judge and choose is that international politics depends on national actions, even if the international results are not the simple sum of how each country behaves” (Perception, xvii). In contrast with the structural theorists, Jervis emphasizes the role of the individual in the study of international relations. But also in contrast with the game theoreticians like Thomas Schelling[12], among others with whom he studied, Jervis argues that the person who makes the decisions is not just a generic ‘rational’ actor, but an individual located in a particular time and place, whose understanding of her world and the motivations and intentions of others are shaped by history, culture, class and upbringing. Indeed, his work forces us to re-evaluate the very utility of the concept of ‘rational’ actor: “in trying to explain how states see the world, it may be better to proceed without constant reference to whether what we find can be labeled rational or not” (xxv). By marrying the emerging insights from psychology (his debt to innovative thinkers such as Erving Goffman, Albert Hirschman, Amos Tversky and Daniel Kahneman,[13] for example, is copiously acknowledged) with a deep knowledge of the history and practice of international relations (particularly the Cold War), Jervis has over four decades created a valuable perspective that made it possible to build a synthesis between the world of parsimonious theory and the complexity of historical inquiry. As he makes clear at the outset of Perception and Misperception, he is interested in explaining not just the recurring patterns of why policy makers make the choices they do, but also why different people in the same situation make different choices (28-29). Thus, he notes, “Perception and Misperception did not present a unified theory, but rather was inductive in drawing on several strands of cognitive and social psychology and looking for congruence with patterns that I saw from my reading of diplomatic history (xxi).

Jervis’s work played an important role in encouraging the application of insights from modern psychology into the training of intelligence analysts and policy makers. Shortly after the initial public of Perception, Robert Bowie, the Director of the CIA’s National Foreign Assessment Center, engaged Jervis as a consultant.[14] His initial engagement led over time to two important studies of intelligence ‘failures’–the failure to anticipate he fall of the Shah of Iran in 1977-78, and the inaccurate assessment of Iraqi WMD during the George W. Bush Administration–as well as numerous unpublished studies for the intelligence community. But the engagement was not simply a one-way street–Jervis provides a compelling account of how his direct experience of the intelligence and policy community contributed to his understanding of how government works and the “necessary tensions between good intelligence and policymaking.” “In return [for access to the process] I gained some understanding of how the system worked and was disabused of several of my naïve notions.”[15] Jervis’s life work is testament to the importance of “bridging the gap” between academia and practice to both communities.[16]

Jervis was not the only analyst who sought to apply insights from psychology to intelligence analysis. Richards Heuer’s influential The Psychology of Intelligence Analysis[17]) was similarly inspired by Kahneman and Tversky, and Heuer’s early articles that formed the basis for the book were written during this same period (between 1978 and 1986). But what makes Jervis’s writings so compelling both for the scholar and the practitioner is the richness of his knowledge of diplomatic history. Jervis goes beyond the insights of the psychology lab (he notes that in writing Perception he spent a year “look[ing] through all the relevant journals in psychology” (xvii), to find support for his conclusions in the ‘real’ world. In what is perhaps one of the most influential chapters of Perception, “Deterrence, the Spiral Model and Intentions of the Adversary,” Jervis illustrates the power of Schelling’s insight on why policy makers might rationally refuse to make even modest concessions by drawing not just on the familiar stories of Munich, the Cuban Missile Crisis, and Vietnam, but on the British decision not to offer Vichy France a conciliatory proposal before attacking its fleet in Oran harbor and on Japan spurning British calls for mediation in the Russo-Japanese War (Perception 58-59).

Many of Jervis’s constructs shape the contemporary debate about international security, both in the academy and the world of policy. It is impossible, for example, to imagine the discussion about the future of U.S.-China relations without considering the relevance of the spiral model[18], nor of the U.S. engagement in Afghanistan absent as assessment of how ‘prospect theory’ influences the continued U.S. involvement. It is no accident that Jervis is one of a handful of scholars judged as most influential both by policymakers and by scholars.[19]

Jervis’s approach not only helps us understand why policymakers do what they do, but perhaps even more important, how to do a better job in the future – ‘How Statesmen Should Think.’ Thus Perception concludes with the chapter “Minimizing Misperception.” Jervis argues that “although prescription is not the purpose of this book, our analysis suggests a number of measures that would decrease misperception. If decision makers become aware of perceptual errors, they may be able to avoid or compensate for them” (409). These ideas are later fleshed out in Chapter 8 of How Statesmen Think, “Political Psychology Research and Theory: Bridges and Barriers”,[20] in which Jervis follows in the path of Alexander George’s work that culminated in Bridging the Gap. Jervis observes:

Political psychology, at least as it deals with international politics, tends to be normatively inflected and reformist. Like the social psychology of the 1930s in which it has its roots, the objective, to slightly alter Marx’s phrase, is not only to understand the word, but to change it (Statesmen 125).

In Why Intelligence Fails Jervis modestly remarks “It remains unclear how much good my reports did for intelligence, but at least I trust they did no harm.”[21] These two works stand as clear evidence of what generations of students, scholars, and practitioners will attest: over a lifetime, Bob Jervis has done remarkable good for the study and practice of international relations.


Review by Philip Zelikow, University of Virginia

“The Magic Mirror Maze”

The publication of a new edition of Perception and Misperception in International Politics along with the collection of essays about How Statesmen Think is an occasion for reflection about the stature and impact of the scholarship of Robert Jervis. The stature is deservedly great. As for the impact, I am not so sure.

The field of international relations theory does not quite know what to do with Jervis. The work is unassailably important. But it is kept in a case off to the side. It is dangerous stuff. To ingest more than 50 milligrams of it could kill off most of the field.

Jervis is a kind of analytic historian. He has actually worked as a historian, in his book Why Intelligence Fails, which did a very good job of unpacking and diagnosing assessments of Iran before the 1979 revolution and of Iraq before the 2003 U.S.-led invasion.[22] For the rest, he has ably analyzed available history in order to build theories about varieties of political psychology.

If Jervis prefers to identify himself as a political scientist, fair enough. Adam Smith thought of himself as a natural philosopher before society gained enough wisdom to invent economists.

In building his theories, Jervis is doing what historians themselves should do more of. He is using the history to extract, catalogue, and label possibilities. These possible behaviors are recurrent patterns or tendencies–mirror-imaging, the ‘drunkard’s search’ for evidence, and many more. Statesmen may do this; they may do that.

Jervis is not saying that these behaviors are probable, just that they are quite possible, possibilities that therefore should be taken into account. He is not building scientific generalizations that invite predictive probability estimates. He actively discourages a priori reliance on theoretical preconceptions.

Jervis notices that most of his colleagues who work on political psychology prefer to stick with mass behavior, large n samples of structured and measurable preferences like ‘attitudes and voting behavior.’ It is easier for them to try for scientific generalizations.

He writes that, like them, he wants to find not only “common patterns” but also “generalizations that apply only to some groups and that separate them from others, and also for individual idiosyncrasies.” But he then qualifies this desire with the warning that “individual and group differences are important,” and then the warning “not to overgeneralize from Western experience or experimental subjects,” and then the further warning that even within the same cultural group there are wide differences, and then the yet further warning that international politics are interactive, so that beliefs are still more dynamic and individualized (2-4).

As if these warnings were not enough, one of Jervis’s later theoretical contributions is to warn of how easy and misleading it can be to assume “representativeness” in theory-driven perceptions about others. (Chapter 3) Jervis is conservative in his critique of “representativeness.” He does not go where some philosophers have gone is questioning the representations of reality in many of our synthetic constructs for theory-building (dictatorships, democracies, etc.). Even so, this chapter is one of those where, the more you think about it, the more destabilizing it is. Careful, only 50 milligrams.

All these caveats end up leading Jervis to “an appreciation of the limits of a priori reasoning and a deep commitment to empirical research” (5). He uses theorizing the way Ernest May and I used to talk about the value of analogies. His theories suggest better questions in considering evidence about the case at hand. They are cautionary reminders about possible misperceptions. They are not shortcuts pointing at the likely answers to these questions.

Yet the overall net effect of going through hundreds of pages of examples of How Statesmen Think with Jervis as your guide can be really bewildering. There are so many examples of deadly confusion, so many kinds of ways that people go wrong. Reading the book, with its combination of intricate stories and constant misperception, kept bringing to my mind a scene at the end of the Orson Welles movie, “The Lady from Shanghai” (1948).

The plot is convoluted even by film noir standards. The femme fatale, played by Rita Hayworth, has involved the Welles character in dizzying plots to kill her wealthy and despicable husband. The climactic scene is set in a deserted ‘funhouse’ in San Francisco.

The Welles character and the Hayworth character meet in an attraction called the “Magic Mirror Maze.”

[Welles]: I thought it was only your husband you wanted to kill.

[Hayworth]: Why don’t you try to understand? George was supposed to take care of Arthur. But he lost his silly head and shot Broome! After that I knew I couldn’t trust him. He was mad. He had to be shot.

[Welles]: And what about me?

At this point the husband also arrives in the mirror maze. Husband and wife see many images of each other in the mirrors. They pull out their guns and blaze away, trying to shoot the real thing. Mirrors shatter; reflections fragment. They end up killing each other.

On stage and screen Welles loved to do magic tricks and illusions to disorient his audience. Jervis is not a showman. But as he guides us again and again through his own mirror mazes, his ‘odditorium’ of wars and near-wars, and his catalogues of illusions, his goal, however, is like that of Welles: to disorient the audience.

The rest of the field of IR theory seems to have the opposite goal. So be careful, only 50 milligrams.

For those who work on international policy, on the other hand, much more consumption of Jervis is advised. In that realm, however, he has not yet had enough impact. I will give two practical examples.

First, consider the character of government assessments and the problem of rationality. Jervis has much to say about the danger of assumed rationality, in which people who are presumed to behave in ways that seem rational to the observer. In my own work on this, in the rewritten edition of Essence of Decision (1999), Graham Allison and I replaced the old straw-man version of ‘objective’ rationality with Herbert Simon’s concept of ‘bounded rationality,’ bounded, that is, by the cognitive limitations and preferences of the person being assessed.[23]

A model of ‘bounded rationality’ is just a model of explanation that assumes that the behavior reflects the ‘bounded’ purposes of a state’s leader or leaders, whatever others might think of those purposes or what informs them. Jervis adds much depth and detail to the factors and processes that the observer might consider.

Imagine, then, an intelligence assessment that took Jervis’s ideas entirely to heart. Imagine such an assessment right now on the very important issue of possible North Korean intentions if the government was armed with certain nuclear capabilities. Quite a lot of public commentary about this already reflects the kind of errors Jervis has trained us to notice. For instance, observers generally assume that the North Korean leaders are just doing more to aid regime survival, even as their policies–at least during the last five years–are having the foreseeable effect of reducing the odds of regime survival.

But what would better observers, learning from Jervis, do instead? They would be very modest, of course, about what they really know and can predict. This humility is not necessarily reassuring.

Above all, such assessments would probably not foreground a single point estimate of any kind, whatever ‘probability’ hedge was placed on it. Instead, I imagine that a Jervis-like approach would invite a set of almost novelistic recreations of possible North Korean perceptions, decision processes, and outcomes–grounded in whatever evidence was available. Rather than an overwhelming focus on Kim Jong-un in isolation, he would be placed in a context–of how he gets information or how he makes and conveys decisions.

There could be several such illustrative scenarios, which might illuminate key variables or telltale clues. In September 1938, during the ongoing Czech crisis, the British Foreign Office was debating how to assess the intentions of the German dictator, Adolf Hitler. Was he just a greedy nationalist with limited aims? Or were his aims much more radical and dangerous? After British Prime Minister Neville Chamberlain’s first summit with Hitler in Berchtesgaden, plausible evidence and arguments could still be assembled on either side of the Hitler assessment. The Foreign Secretary, Lord Halifax, and his Permanent Secretary, Alexander Cadogan, looked at Hitler’s responses to Chamberlain’s offers of Czech dismemberment at the next summit, in Bad Godesberg, as a test of the competing theories.

Hitler rejected Chamberlain’s offers. That evidence provided a decisive clue in favor of the ‘radical’ assessment. Both Halifax and Cadogan concluded that their prior assessments of Hitler had been wrong. They assessed that his behavior could only be reconciled with the hypotheses of the Office’s hawk, Robert Vansittart, who had argued the ‘radical’ case. Halifax thereupon informed Chamberlain that Britain had to be prepared to go to war against Germany. The British Cabinet was pushing that way when Italian dictator Benito Mussolini invited the powers to a further summit in Munich, at which Hitler believed he backed down by settling for Chamberlain’s earlier offers.[24]

Another example of using hypotheses to set up an evidentiary test, a ‘Godesberg’ test, can be drawn from the Cuban Missile Crisis. Jervis argues (163) that CIA director John McCone believed, unlike his analysts, that the Soviets would put missiles in Cuba in 1962 because of his general worldview, not expert knowledge. That is not quite right. In fact, McCone was looking at a particular evidentiary clue, which seemed like a key variable to him, his version of a ‘Godesberg’ test.

Decisive to McCone was intelligence that the Soviets were installing their latest surface-to-air missile systems in Cuba. “Difficult for me to rationalize extensive costly defenses being established in Cuba,” he argued, unless “MRBMs to be installed by Soviets after present phase completed and country secured from [U-2] overflights.” This reconstructive hypothesis turned out to be right.

McCone could not be sure that his hypothesis was right. But this was a reason why he advocated risky U-2 overflights over Cuba that could check it out.[25]

Although many intelligence officers have studied Jervis’s work, his insights have not yet worked themselves well into the way intelligence estimates are prepared, to the extent that I can judge from the many estimates I have seen in recent years. Particular analysts have the knowledge and gifts to do the kind of fact-based yet also speculative analyses that his method would require. Much better work is possible. It is sad to say, though, that with very rare exceptions, the process–for reasons understandable in that world–discourages preparation of products that have the necessary detail, color, variety and individual craft to embrace Jervis’s cautions.

My second practical example also relates to habits of assessment. It arises from prospect theory. Jervis hits prospect theory hard, emphasizing its tendency toward loss-aversion: “More than the hope of gains, the specter of losses activates, energizes, and drives actors, producing great (and often misguided) efforts that risk–and frequently lead to–greater losses.” The chapter on this in How Statesmen Think (chapter 4) republishes a 1992 essay which he could have revised in order to note the Iraq war of 2003 as a rather powerful illustration.

I am impressed by Jervis’s conclusion that Prospect Theory is not only the “best known” but also the most “intriguing” psychological theory developed since his Perception and Misperception was published more than forty years ago (How Statesmen Think, 8). I believe the most pervasive symptom of the U.S. government’s and society’s tendency toward loss-aversion evidences itself in the intelligence community’s ingrained mindset of ‘threat assessment.’

‘Threat assessment’ is a profoundly defensive, loss-focused form of assessment. It is hard to overstate the influence of this paradigm on the way Americans and American officialdom see the world. The intelligence community is massively oriented toward this paradigm. ‘Threat assessment’ naturally focuses American attention on the least important, but broken, areas of the world, at the expense of attention on the most important areas for potential gain.

An alternative, or at least complementary, paradigm could be a focus on ‘opportunity assessment.’ Where might the U.S. government positively advance its interests in some catalytic region or on some catalytic topic? In the private sector world of investment, investors are quite conscious of loss-aversion tendencies. They think hard about how to counter it. Hedging and diversification strategies are one example, although still with a defensive cast. But investment analysts also scout for possible wins, knowing that they may have to develop evidence and arguments to counter the reflexive tendencies that Jervis illustrates.

Jervis leads his readers into a disturbing maze of mirrors, where people sometimes really do shoot at fractured reflections of reality. But there are some practical counters that policymakers and intelligence officials can consider. Disorientation need not lead to hopelessness.


Author’s Response by Robert Jervis, Columbia University

I am grateful to my four colleagues for their reviews, to Tom Christensen for dubbing me (in slight paraphrase) ‘a giant gadfly,’ and for Tom Maddux for having organized this Roundtable (and for all the work he has done to make H-Diplo and ISSF a success). I am especially glad that he selected people with experience in the policy world as well as those who have made their careers as research scholars; while my primary target is the academic community, the work has policy implications. Finding all the comments acute, I want to take the opportunity to comment and expand upon a few of the reviewers’ remarks.

It is particularly valuable to try to look forward and see which areas political psychology should draw on and expand into in the future, and here Rose McDermott is astute in her comments, as she is in her own research. Relevant work in biology, neuroscience, and genetics has been especially fruitful in the past decade but has received little attention from political scientists (and is barely acknowledged by historians). Part of the reason is that these areas are extremely complex, and evaluating them outruns our competence. But I do not think that this is all there is to it. While almost all educated people accept the idea of evolution, the obvious achievements–both positive and negative–of humanity conduce toward us seeing human beings as unique in their abilities to control their environments and their fates. Although the commonalities in our bodies and bodily functions with other animals are hard to miss, it is easier to convince ourselves that our brains are somehow immune to biology and evolution. Furthermore, the mention of biology, let alone genetics, calls to mind the view that humans and societies cannot change, that what exists must be ordained if not by God then by evolution, and that the familiar hierarchies of race and gender are appropriate. Such associations are solidified by memories of the Nazis and fears that individuals, if not races, will be pigeonholed according to their supposed abilities and, somewhat contradictorily, by the concern that acting upon this science will lead to the development of ‘designer babies.’

These fears should not be dismissed, and the past associations are indeed disturbing. To shut our eyes to the possible role of biology in both grounding our generalizations and looking at individual differences makes no sense, however, particularly because the best work in this area is fully alive to the interactions between biological and social factors. Done well, this work does not dehumanize us; quite to the contrary, it brings out how we are humans in general and what distinguishes us as individuals. It is striking that many scholars who applaud ‘transgressive’ work, especially in the humanities, are quick to say that this research is beyond the pale. This is not to say that it is easy or the paths forward are clear. But there are indeed paths in the plural, rather than one best way to proceed, and the work will require difficult interdisciplinary conversations and movement among different levels of analysis. Difficult does not mean impossible.

I am glad that James Steinberg affirmed the significance of the security dilemma to both scholars and decision-makers, of which he was one. The security dilemma also is interesting because it combines structural and perceptual factors. Neither one alone fully accounts for the phenomenon, and in this regard I doubt that this is unusual. Of course I did not invent this concept, whose lineage goes back to Thucydides, but it was not highly salient when I wrote the first edition of Perception and Misperception. Indeed, although it was important for one of my mentors, Glenn Snyder, I did not fully understand it until I started putting it down on paper. As Steinberg’s co-authored book, Strategic Reassurance and Resolve: U.S – China Relations in the Twenty-First Century[26] shows, it is at the heart of debates about how the U.S. should cope with the rise of China. It leads us to ask not only about each side’s motives and intentions, but about how (and whether) a state can make itself tolerably secure without trenching excessively on the security of others.

As Steinberg notes, my discussion of deterrence versus the spiral model was based on arguments during the Cold War, ones in which the protagonists often failed to understand the central issues that divided them, which concerned not the validity of competing general models, but their applicability to the Soviet-American conflict. My own views on this changed over time and remain unstable. I started out as a hawk, believing that the deterrence approach was appropriate to the conflict, but became less certain as a result of my research. Years later, I asked, “Was the Cold War a Security Dilemma?” and came up with the somewhat questionable concept of a deep security dilemma in which each side sought security but believed that this required a drastic reduction of the other’s power, if not its elimination.[27] But as Shiping Tang argues, if there is no way for one side to increase its security without making the other less secure, then this is not a security dilemma at all and I may have been guilty of conceptual stretching.[28] Furthermore, I consider the question posed in the title to be very much an open one. After coming back from a dinner and re-reading the final draft, I tore out the conclusion and re-wrote it from scratch, something that I rarely do.

I want to pick up on two of Dianne Pfundstein Chamberlain’s points. First, she highlights the importance of my “Signaling and Perception” essay. As she explains, these two topics are normally treated separately, yet they are inherently joined. Teaching a graduate course on this subject reinforces my sense of how difficult this is and that I have only scratched the surface of it. Much as I like work on signaling, which was the topic of my dissertation and the book that proceeded Perception and Misperception,[29] it makes heroic assumptions about how signals are perceived and acted upon. States often see each other, they situation they are in, and specific behaviors very differently. We often talk about international politics as a game of poker; but a much better analogy is the Japanese short story and movie Rashomon, which tells the same tale from the vantage point of four participants who see it in wildly different ways.

Chamberlain closes by noting that “these books force one to confront a series of unsettling realities” about the difficulties in understanding our world that are created not only by the now-familiar cognitive and affective biases but also by the fact that particular events and bits of evidence cannot be interpreted with higher-order beliefs and theories. Limits on the pursuit of objectivity are produced not only by biases we can seek to reduce, but by the very nature of perception and understanding.

Phillip Zelikow, who like Steinburg is both an academic and a past policy-maker, brings the somewhat unconventional acuity for which he is known to his review. I join in his delight in “The Lady from Shanghai” and am glad to accept the label of being “a kind of analytic historian.” I certainly could not do my research or teaching without heavy doses of history, and greatly enjoy, as well as learn from, reading as much of it as I can. I frequently return to the work of Zelikow’s mentor, Ernest May.

I think he is right that it is hard for the IR field to fully absorb the role of misperception. To take just one example, even though it is obvious that deterrence depends on calculations made by the target state and that credibility and reputation matter or even exist only in the minds of perceivers, scholars are prone to talk otherwise, to analyze situations as though these factors are objective or, even worse, to portray them as properties of the actor, especially when it is the U.S. that is trying to do the deterring.

The barriers to intelligence analysts and, even more, policy-makers internalizing the notion that world politics resembles Rashomon are even greater. The very fact that this approach forces us to see the multiple possibilities implicit in most situations complicates their tasks. It also is threatening to realize that others may interpret our behavior in a very different way than we do, to see us as weak when we think we are strong or grasping and selfish when we think we are defensive and providing public goods. An understanding of the multiple possibilities for error also can undermine decision-makers’ self-confidence, and with it, their ability to provide effective leadership. One of my main arguments is that people shy away from perceiving painful value trade-offs and it is quite possible that a better understanding of the nature of international (and indeed human) interaction would bring with it high costs as well as real benefits.



[1] Robert Jervis, “Cooperation under the Security Dilemma,” World Politics 30:2 (January 1978); and The Meaning of the Nuclear Revolution: Statecraft and the Prospect of Armageddon (Ithaca: Cornell University Press, 1989).

[2] For Jervis’s complaints about U.S. military doctrine under conditions of mutually assured destructions, see Jervis, The Meaning of the Nuclear Revolution; and Robert Jervis, The Illogic of American Nuclear Strategy (Ithaca: Cornell University Press, 1984). In How Statesmen Think, 186-187, Jervis recognizes that there was a long overdue acceptance of “security dilemma thinking” under Soviet leader Mikhail Gorbachev’s rule, which did not begin until 1985.

[3] Representative examples of their work include: Thomas Schelling, Strategy of Conflict (New York: Oxford University Press, 1963); Arms and Influence (New Haven: Yale University Pres, 1967); and Amos Tversky and Daniel Kanheman. “Prospect Theory: An Analysis of Decision under Risk,” Econometrica 47:2 (March 1979), 263-291.

[4] Harold Lasswell, Psychopathology and Politics. (Chicago: University of Chicago Press, 1930); Harold Lasswell, Power and personality. (New York: Norton, 1948).

[5] Alexander and Juliette George, Woodrow Wilson and Colonel House: A Personality Study, Vol. 1144. (New York: Courier Corporation, 1954).

[6] Daniel Kahneman and Amos Tversky, “Prospect theory: An analysis of decision under risk,” Econometrica: Journal of the Econometric Society (1979): 263-291; Tversky and Kahneman, “Judgment under Uncertainty: Heuristics and Biases,” in Tversky and Kahneman, eds., Utility, Probability, and Human Decision (Amsterdam: Springer Netherlands, 1982), 141-162.

[7] Peter K. Hatemi and Rose McDermott, “The Genetics of Politics: Discovery, Challenges, and Progress,” Trends in Genetics 28:10: 525-533; Hatemi and McDermott, Man is by Nature a Political Animal: Evolution, Biology, and Politics. (Chicago: University of Chicago Press, 2011); McDermott, James Fowler and Oleg Smirnov, “On the Evolutionary Origin of Prospect Theory Preferences,” The Journal of Politics 70:2 (2008): 335-350.

[8] This argument is made most extensively in Robert Trivers, The Folly of Fools: The Logic of Deceit and Self-Deception in Human Life (New York: Basic Books, 2008).

[9] Richard Wrangham, “Is Military Incompetence Adaptive?” Evolution and Human Behavior 20:1 (1999): 3-17.

[10] Kenneth Waltz’ influential Man, the State and War (New York: Columbia University Press 1959) and his later Theory of International Politics (McGraw-Hill 1979) fostered the powerful ‘neo-realist’ movement in political science, which prioritizes the impact of structural features on interstate relations, while acknowledging the (less consequential) impact both of individuals (‘first image’) and domestic factors (‘second image’). Jervis not only leads us to focus on the decisionmaker, but also helps us understand the inextricable links between each nominally distinct level of analysis (this is the focus of Chapter One of Perception, “Perception and the Level of Analysis Problem” and further treated in the new Preface (xix-xxi). Jervis himself characterizes his thinking as “defensive realism” arguing that “this form of Realism could easily accommodate a role for perceptual distortions” Perception xxvi.

[11] Jervis himself somewhat downplays the significance of Perception. The Logic of Images [his first book] …. was much, much harder book to write on than the one on misperception, and it is a more original book, because the misperception book is put together from work a lot of people had done….”. Nicholas J. Wheeler, “Interview with Robert Jervis” International Relations 28:4 (2014), 485. Perhaps, but the synthesis Jervis offers in Perception is in itself a distinct and powerful perspective.

[12] See, for example, Thomas J. Schelling, The Strategy of Conflict (Cambridge: Harvard University Press 1980). Schelling understood the limits of this approach: “If we confine our study to the theory of strategy, we seriously restrict ourselves by the assumption of rational behavior…. If our interest is the study of actual behavior, the results we reach under this constraint may prove to be either a good approximation of reality or a caricature.” 4.

[13] Erving Goffman, The Presentation of Self in Everyday Life (New York: Doubleday 1959); Albert O. Hirschman, Exit, Voice and Loyalty (Cambridge: Harvard University Press 1970); Amos Tversky and Daniel Kahneman, “Judgement Under Uncertainty: Heuristics and Bias”, Science, New Series, 185:4157. (27 September 1974), 1124-1131.

[14] Jervis’s account of his early involvement with the CIA appears in the first chapter of his book, Why Intelligence Fail: Lessons from the Iranian Revolution and the Iraq War (Ithaca: Cornell University Press 2010).

[15] Why Intelligence Fails, 13-14.

[16] The expression comes from Alexander George Bridging the Gap: Theory and Practice of Foreign Policy (Washington, D.C.: Institute of Peace Press 1993) which forms the point of departure for Jervis’s Chapter 6 in Statesmen, discussed below.

[17] Richards J. Heuer, The Psychology of Intelligence Analysis, (Center for the Study of Intelligence, 1999).

[18] Jervis’s approach deeply informs the analysis that Michael O’Hanlon and I offer in our book, Strategic Reassurance and Resolve (Princeton: Princeton University Press 2014).

[19] See Paul C. Avey and Michael C. Desch, “What do Policymakers Want from US? Results of a Survey of Current and Former Senior National Security Decisionmakers” International Studies Quarterly, 58:2 (June 2014): 227-246. The other two are Joseph Nye and Kenneth Waltz.

[20] Originally published as “Bridges, Barriers and Gaps: Research and Policy” Political Psychology 29 (August 2008):571-592.

[21] Why Intelligence Fails, 13.

[22] Robert Jervis, Why Intelligence Fails: Lessons from the Iranian Revolution and the Iraq War (Cornell University Press, 2010)

[23] Graham T. Allison and Philip Zelikow, Essence of Decision: Explaining the Cuban Missile Crisis, 2nd ed. (New York: Longman, 1999).

[24] Ernest R. May, Strange Victory: Hitler’s Conquest of France (New York: Farrar, Straus & Giroux, 2000), 177-78, 83-84.

[25] McCone to his deputy, Marshall Carter, 10 September 1962, reprinted in CIA History Staff, CIA Documents on the Cuban Missile Crisis 1962 (Washington, D.C.: CIA, 1992), 59. Although McCone’s hypothesis was right, it could also have been wrong. The Soviet government had decided to deploy the SA-2 missiles in April 1962, as part of the original Cuban military assistance package, before Soviet Premier Nikita Khrushchev made the separate decision to deploy nuclear-armed ballistic missiles there. That decision made, since the SA-2s were going, they went in a sequence that would allow them to protect Cuba from snooping overflights, as McCone guessed. Why the SA-2s were then not actually used when the U-2s first flew is another story, one that also can call on psychological theory.

[26] James Steinberg and Michael E. O’Hanlon, Strategic Reassurance and Resolve: U.S-China Relations in the Twenty-First Century (Princeton: Princeton University Press, 2014).

[27] Robert Jervis, “Was the Cold War a Security Dilemma?”, Journal of Cold War Studies 3 (Winter 2001): 36-60.

[28] Shiping Tang, A Theory of Security Strategies for our Time: Defensive Realism (New York: Palgrave Macmillan, 2010); Giovanni Sartori, “Concept Misformation in Comparative Politics,” American Political Science Review 64 (November 1970): 1033-53.

[29] The Logic of Images in International Relations (New York, Columbia University Press, 1989).