Roundtable 11-16 on Diplomacy:  Communication and the Origins of the International Order49 min read

Robert Trager’s Diplomacy: Communication and the Origins of the International Order focuses on the role of communication in diplomacy with emphasis on the role of costless exchanges such as private discussions between two foreign policy ministers versus costly signaling such as moving troops to the frontier of an adversary or a drone strike on a hostile paramilitary force. Trager makes use of two related datasets from the Confidential Print of the British Foreign Office’s communications between 1855 and July of 1914 with emphasis on detailed case studies on significant historical events including the negotiations leading to the outbreak of World War I.

H-Diplo | ISSF Roundtable XI-16

Robert F. Trager.  Diplomacy:  Communication and the Origins of the International Order Cambridge:  Cambridge University Press, 2017.  ISBN:  978-1-107-62712-3 (paperback, $32.99).

18 May 2020 \
Editor: Diane Labrosse | Commissioning Editor and Chair: Thomas Maddux | Production Editor: George Fujii


PDF Version

Introduction by Thomas Maddux, California State University Northridge, Emeritus

Robert Trager’s Diplomacy: Communication and the Origins of the International Order focuses on the role of communication in diplomacy with emphasis on the role of costless exchanges such as private discussions between two foreign policy ministers versus costly signaling such as moving troops to the frontier of an adversary or a drone strike on a hostile paramilitary force. Trager makes use of two related datasets from the Confidential Print of the British Foreign Office’s communications between 1855 and July of 1914 with emphasis on detailed case studies on significant historical events including the negotiations leading to the outbreak of World War I.

The reviewers are unanimous in their praise of Trager’s study with respect to its research, careful assessment of conflicting theories on communication, and the author’s recognition that further study is necessary on issues such as the role of emotion in diplomatic exchanges.  Brian Rathbun applauds Diplomacy as a “landmark in the study of diplomacy specifically, and international communication and signalling more broadly.”  Admittedly a critic of rationalist studies, Rathbun concludes that Trager’s work is “a model of rigorous empirical analysis and bold theory that should set the standard going forward for formal approaches to international relations.”  Todd Hall agrees with Rathbun that Trager’s “arguments are original, the painstaking work of amassing the data impressive, and the case studies thought-provoking.  It truly sets the bar for future work in this area.” What Hall emphasizes as central to the study is its argument that “private diplomacy can convey meaningful information about intentions” which contradicts rationalist studies that dismiss “private messages as worthless when it comes to providing rational acts with information about a sender’s intentions.”  Marcus Holmes endorses Hall’s emphasis on the central contribution of the book and emphasizes its impact as an “exceptional addition to literatures on diplomacy, signaling, bargaining, and IR theory more generally.”  For Anne Sartori, Trager’s study is a major contribution on the “effectiveness of costless diplomacy” with emphasis in the case studies on how states communicate and “when and why diplomacy leads policy-makers to change their minds, focusing in particular on the origins of the two World Wars.”

The reviewers do raise some issues that merit further study and Trager’s responds to their observations in his response.  Sartori, for example, raises the challenging issue that the data in Trager’s study, the writing by diplomats, “may not tell us exactly what they were thinking for several reasons,” including the desire to impress superiors and advance their preferred response to issues, although she notes that Trager brings in other evidence to support his assessments.  Rathbun raises the issue of how irrational emotions affect diplomatic communication and Trager’s thesis on costless negotiations. In his response Trager elaborates on the centrality of the expectations of the negotiators and suggests that “we need to go beyond rationalism to understand the sources of expectations and rational responses to the strategic context.”  Hall suggests that constructivist and psychological scholarship on diplomacy raises questions about the “implications for Trager’s arguments” and that he may need to “relax his assumption of rationality to better reflect the human element of diplomacy.”  Holmes raises a number of questions on the central role of intentions in costless communications, how intentions may be changed, how potential cost is very subjective, and that both rational and psychological factors need to be considered.

The reviewers, however, agree with Holmes that the “questions and criticism are opportunities for future research and further analysis” and “do not detract from Trager’s important work” which “provides yet another reason to take diplomacy, even the private type, seriously.”


Robert F. Trager is Associate Professor of Political Science at the University of California, Los Angeles. He has also taught at Yale and Oxford Universities, held an Olin Fellowship at Harvard University, and worked in investment banking in New York. His book, Diplomacy: Communication and the Origins of International Order, was published in 2017 by Cambridge University Press.  His articles have appeared in such journals as the American Political Science Review, the American Journal of Political Science, the Journal of Politics, International Organization, International Security, and Security Studies.  Current research projects focus on the moral and emotional bases of foreign policy preferences, the strategic implications of artificial intelligence developments, the role of women in making democracies more pacific and the dynamics of affinity networks.

Todd H. Hall is Tutor in Politics, Saint Anne’s College & Associate Professor, Department of Politics and International Relations, University of Oxford. Hall is author of Emotional Diplomacy: Official Emotion on the International Stage, published by Cornell University Press in 2015.  He is currently working on a project examining Sino-Japanese relations.

Marcus Holmes is Associate Professor of Government at The College of William & Mary, where he also serves as Director of the Political Psychology and International Relations lab and Co-Director of the Social Science Research Methods Center. His most recent book is Face-to-Face Diplomacy: Social Neuroscience and International Relations (Cambridge University Press, 2018) and he is currently working on a co-authored book project with Nicholas J. Wheeler on building trust in cyber space. His research has been published in International OrganizationInternational Studies QuarterlyJournal of Theoretical Politics, and other journals.

Brian Rathbun is Professor of International Relations at the University of Southern California. He received his PhD from the University of California, Berkeley. He is the author of Partisan Interventions (Cornell University Press, 2004), Trust in International Cooperation (Cambridge University Press, 2012), Diplomacy’s Value (Cornell University Press, 2014) and Reasoning of State (Cambridge University Press, forthcoming). He is currently writing a book on ethics in international relations.

Anne Sartori is Research Scientist at the Sloan School of Management and Visitor in the Department of Political Science at MIT.  Her research interests include interstate communication and nuclear strategy and stability. Her current project studies the impact of civil-military relations on international conflict, and her previous work includes the book Deterrence by Diplomacy (Princeton: Princeton University Press, 2005).



Review by Todd H. Hall, University of Oxford

‘Tour de force.’  That is the most fitting description I can find for Robert Trager’s monograph.  I admit that this characterization is not original—James Fearon is cited using the same words on the back cover—but it nevertheless is very apt.  In this work, Trager offers the field of international relations an analogue to the triple threats of theatre that can sing, dance, and act: he formally models his arguments, examines them in light of quantitative evidence generated by the coding of sixty years’ worth of British diplomatic documents, and also explores their explanatory power through detailed historical case studies.  The arguments are original, the painstaking work of amassing the data impressive, and the case studies thought-provoking.  It truly sets the bar for future work in this area.

The reason I predict it will be remembered, however, is the proposition at the heart of book: private diplomacy can convey meaningful information about intentions.  This runs counter to a large contingent of rationalist work within the field of international relations on signalling that treats private messages as worthless when it comes to providing rational actors with information about a sender’s intentions.  The logic of this latter position is as follows: where there are no costs to bluffing, rational actors have the incentive to misrepresent private information in order to achieve a better deal, making it difficult to distinguish between actors that are lying and telling the truth.   Public diplomacy can, theoretically, help solve this dilemma by creating audience costs—that is, penalties imposed on policymakers by their constituents for staking a position and then retreating from it.[1]  An actor witnessing a public statement can thus gauge its sincerity on the basis of the potential political damage the speaker would suffer from subsequently backing down; simply, a policymaker who was bluffing would not want to generate such costs.  But statements made in private—exchanged unbeknownst to constituents capable of exacting a political price for reneging—remain ‘cheap talk,’ and therefore their recipients cannot trust their veracity.

Trager argues that the value of private diplomacy has been wrongfully discounted.  Acknowledging the contributions of existing work on the incentives states perceive in maintaining a reputation for honesty,[2] he outlines four additional mechanisms by which messages relayed in private can nonetheless shape recipient’s beliefs about a sender’s intentions.  The first entails the scope of demands levelled in a private communication.  Under certain circumstances, an extreme demand can communicate resolve, for it forecloses the possibility of a middle-range compromise and risks conflict; alternately, the offer of a middle-range compromise, by already ceding the sender’s most desired outcome, can convince the recipient that the sender is willing to make a concession, although it may also elicit the expectation of further capitulation.  The second involves communication that signals a willingness to frustrate the recipient’s designs and risk a breach in the relationship.  Such messages, because they may elicit from the recipient counter-measures to the detriment of the sender, can also carry informational value.  The third is messages of support for a protégé, for the newly emboldened protégé may consequently entangle the sender in a war.  And last is the private approach to communicate the risk of conflict with third parties or improve relations in advance of the same; this message can be credible due to the fact that the approaching state makes itself therewith vulnerable to exploitation of its predicament by the recipient.

Granted, the above outline does not do justice to the full complexity of the arguments and their formal expressions.  But to simplify these even further, the core, basic insight threaded throughout the theoretical chapters is that private diplomacy can in certain circumstances generate costly risks—be they of war, deteriorating relations, or exploitation—that render its messages credible.  The argument appears, on its own rationalist grounds, intuitive.  And the evidence as mustered—both quantitative and qualitative—appears to offer clear support for the claim that these logics are at work.

And yet, that said, how do we square this work with writings on diplomacy outside the rationalist vein?  Trager himself notes emerging constructivist and psychological scholarship on diplomacy and does not challenge the potential contributions these offer (15, 41-42).  On the psychological side, he grants that perceptions can be altered by framing and other non-rational factors and recommends this as a future area for research (42).  But, at the same time, he suggests that regardless the psychological factors at work, it is the explicit reasons actors provide—given that individuals also need to “carry the day in arguments among decision makers”—that may be most important for analysis (42).  The problem here, though, is not simply the ways in which psychological factors may result in rationalisations being presented in the costume of reasons, but also the manner in which psychological dynamics may shape which signals and indices are granted salience in the first place.  By looking to only events recorded and reasons given, we do not know what potential communications, signals, behaviours, or scraps of ­evidence were omitted, overlooked, or ignored.  It is the potential problem of the dogs that did not bark, or—perhaps better said—the dogs whose barking went unheard as a result of pre-existing dispositions, assumptions, and biases or—alternately—substantively irrelevant aspects of the messenger or means of delivery.  Nor can we rule out the ways in which prejudices, affinities, and animosities towards various actors shaped the identification and interpretation of evidence, particularly when these were widely shared by both the authors and readers of the diplomatic correspondence.  Given, in particular, the rise of behavioural approaches not just within economics, but also now international relations as well,[3] one wonders what the implications for Trager’s arguments would be were he to relax his assumption of rationality to better reflect the human element of diplomacy.

Additionally, as Trager himself concedes, there is the matter of the socio-historical environment within which the actors he studies were located: “There is, however, a sense in which the game theoretic presentation conceals the dependence of the findings on particular international cultures” (221).  European Great Power diplomacy in the era Trager examines was a game that its protagonists had developed to an art; its practices reflected a relatively clear, mutually understood set of aims, rules, and expectations.  And other literature has also suggested that certain, more specific shared beliefs—be they social Darwinism or the cult of the offensive—played a crucial role in how intentions were assessed and behavior interpreted in key cases Trager explores, particularly the outbreak of World War I.[4]  This is further complicated by the massive subsequent growth in institutions—both formal and informal—that may have changed the nature of the game today significantly.  The danger thus arises that the findings hold for a specific game of diplomatic chess (or Risk?) by sophisticated, experienced players with shared beliefs, but may face limitations when the players move to a different form of recreation, or even—as initially with the Bolsheviks—do not agree on the game being played.  Again, Trager is not unaware of these issues.  But in leaving these issues to “future research” (222) he sidesteps important questions with direct implications for the empirical relevance of his models.

It is possible that in so rigorously deploying his theoretical tools and methods of choice, Trager has also simultaneously demonstrated their limits.  One cannot account for signals and indices of which no account was recorded; one cannot capture the workings of psychological biases in models that assume their absence; and one cannot model the processes that generate the starting assumptions of the model.  And there is also the larger issue of how to reconcile Trager’s theoretical ontology with the psychological and constructivist approaches he does not outright reject.  Some might celebrate the shelving of such issues as a move away from the overwrought paradigm wars of the previous generation and a step towards a greater analytical eclecticism.[5]  And yet, in those cases where non-rationalist approaches would suggest alternative assumptions and competing answers, there exists the hazard that we are papering over significant differences.  While it may be unfair to expect Trager’s monograph to resolve these issues, it is important to note that by its nature, intentionally or not, it presents them.  That said, nothing should not take away from what this book is: an impressive piece of scholarship, a true tour de force.



Review by Marcus Holmes, William & Mary

As the United States (U.S.) engages in both very public and private forms of diplomacy with North Korea, South Korea, China, Japan, and other states in the region over the future of the Democratic People’s Republic of Korea’s (DPRK) nuclear program, the key questions revolve, as they often do, around intentions and trust. Does the DPRK have the intention to dismantle its nuclear program?  Does the U.S. have the resolve to go to war if they do not?  Can either side be trusted to uphold their commitments in either the short-term or the long-term?

Robert Trager has written a terrific book that provides useful insight into these critical questions.  It is theoretically innovative and empirically extremely well supported.  Trager’s combination of formal models, a large-n dataset, and in-depth case studies is complemented by intriguing anecdotes, some of which are well known, while others will be undoubtedly be new to many readers.  The use of anecdotes in such a compelling way is reminiscent of Robert Jervis’s masterful usage in The Logic of Images in International Relations.[6] Trager’s writing style is also very lucid, precise, and he writes in a way that is accessible to the lay reader. Readers who are not particularly comfortable interpreting formal models will nevertheless be able to follow the argument but will also appreciate the complexity therein.  Few books are able to combine so many diverse types of evidence and remain a compelling read, making this an exceptional addition to literatures on diplomacy, signaling, bargaining, and IR theory more generally.

The central contribution of the book is its argument that costless communication can shape perceptions about state intentions.  If states can signal their intentions through ‘cheap talk,’ then this serves as an important corrective to the view that ‘costly’ signals alone convey sincerity.  While I am broadly in agreement with the perspective provided in this book on the usefulness of private diplomacy, and am convinced by the evidence provided, as with any good book there are nevertheless a few areas where additional questions can be raised. These include the meaning of intentions, the problem of subjectivity of perceptions, and the future uncertainty problem.

The book’s central claim has to do with intentions.  More specifically, Trager asks

“[h]ow can leaders draw conclusions about their adversaries’ intentions from mere statements?” (3).  He argues that not only do diplomats and leaders draw perceptions about intentions from costless statements, they have likely been doing so “longer than recorded history” (5).  This sets up an important puzzle: if costless communication cannot reveal intentions, then what have diplomats and leaders been perceiving for all these millennia?  Trager’s answer is that, in fact, communication, even of the costless variety, such as speech acts delivered in private, affects perception of intentions.  There are a variety of mechanisms of inference through which this occurs.  In one of the mechanisms, referred to as ‘scope of demands,’ the use of threats in private can affect perceptions of intentions.  As Trager argues, “attempts at communication will change the threatened states beliefs about the threatening states intentions” (52).  The logic is straightforward: by demanding more, senders can tell the target that they are resolved to such an extent that they are willing to forgo a compromise that would be beneficial to them.

In one of the many compelling anecdotes in the book, Trager highlights how in 1859, before Italian unification, the kingdom of Piedmont-Sardinia attempted to obtain a secret alliance with France, and in so doing conducted military operations near the Austrian border, making Austria very nervous. A compromise was floated: the Austrians pull back from the border, and the Piedmontese would demobilize as well.  Through private diplomacy the Austrians rejected the compromise, insisting that the Piedmontese be the first mover.  As Trager argues, “[a]ll parties knew what the Austrians wanted; what they did not know initially was whether Austria was willing to fight rather than accept a compromise” (47).  The private diplomacy, in which Austria made a high demand, forgoing a compromise, led to a belief that “Austria is resolved to fight for the more substantial demand” (47).  And it worked.  The British ambassador to Austria noted that he had “not the smallest hope that the Austrian Government will agree to any such [compromise]” (47).  What this case shows is that states may be able to convey their level of resolve to adversaries through costless communication.  Austria successfully signaled that it was willing to fight by not taking the compromise that was on the table.

But is this the same as conveying one’s intentions?  For Trager, resolve and intention appear to be synonymous.  Indeed, they are used interchangeably at various times in the text.  I would argue that the two are conceptually distinct and this has important implications for identifying limitations to Trager’s argument.  Joshua Kertzer’s recent book on resolve, for example, explicitly disaggregates the two concepts.[7] For Kertzer, resolve is “a state of firmness or steadfastness of purpose, a ‘second-order volition’ that is neither reducible to an actor’s intentions, no isomorphic with its capabilities.”[8] One can imagine scenarios where resolve is decoupled from intention. In the Piedmont case, for example, by not taking the compromise Austria conveyed that it had resolve.  It was not going to back down.  But this does not necessarily mean it conveyed an intention to fight.  The steadfastness of resolve could have resulted in further negotiation, developing allied relationships, or linking the Piedmont issue to another issue.  For sure, it is likely case that in practice resolve and intention are often observationally synonymous, but a conflation of the two analytically reduces some of the explanatory leverage and may have difficulty when applied to cases where resolve and intentions are not linked. This raises an important question as to what receivers are actually perceiving: resolve, intentions, or both?

There is another way altogether to think about intentions that is touched on in the book, but not fully developed.  In the example used above, the aim was to communicate a level of resolve, or intention in Trager’s language, to an adversary.  We might call this an exogenous intention.  It developed outside of interaction and was communicated through interaction.  Another way that intentions can be communicated, and indeed created, is through an endogenous process.  Actors may enter into an interaction with only a vague sense of what they intend, or no sense at all, and it is through the interaction itself, an intersubjective process, that intentions are created. In Jennifer Mitzen’s work on the Concert of Europe, the aim of the communication that took place was not just to convey pre-formed individual intentions but rather to develop collective intentions.[9] Models of intention formation in practice theory, such as Vincent Pouliot’s work, are similar.[10]

Trager hints at the notion of developing intentions through intersubjectivity (115) but does not fully develop the concept, which is fair given the goals of the book.  Yet this different understanding of how intentions come about poses an intriguing question about how the mechanisms might operate in a collective intention, rather than individual intention, context.  Can states create and change intentions through costless dyadic communication in the same way that they can convey exogenous ones?  This has particularly important ramifications for Trager’s claim that costless communication can create international order.  International order can be constructed through diplomatic processes, but it is the type envisaged by Mitzen, Pouliot, Sharp[11], and others, where diplomacy is constitutive of the international system not in the sense of the geopolitical structure of states competing under anarchy communicating through diplomacy, but rather in the sense of a thick intersubjective structure where diplomacy is “an emergent phenomenon whose form changes over time.”[12] If diplomacy is conceptualized as speech acts that communicate intentions, it is not clear that we can get to a thick intersubjective international order from communication of exogenous intentions alone.

A second area to probe relates to the notion of the objectivity of threats and claims.  As alluded to above, in one of the mechanisms Trager outlines, it is the size of a demand that does the work in signaling intention.  “The magnitude of a demand conveys information because of the risks involved in larger demands.  The demanding state may end up worse off than if it had negotiated for a compromise outcome that is more likely to be forthcoming” (70).  The intuition is straightforward: by demanding something large, states run the risk that the other side will forego a concession, and perhaps not negotiate at all, thus conveying their level of resolve.  The question I want to raise here is whether it is obvious what constitutes a large demand to both parties.  One of the criticisms that has been leveled at the costly signaling literature is the idea that cost is subjective.  What is costly to one might be viewed as non-costly to another.  This is particularly salient in cases where actors operate with negative perceptions, assumptions, or images about the other and thus interpret their moves in an unsympathetic way.  Nicholas J. Wheeler’s recent book captures the logic of this succinctly: without what he terms “bonded trust,” it is difficult to know whether signals, even very costly ones, are sincere.[13]  The cost is in the eye of the beholder. While certainly not private diplomacy, Egyptian President Anwar Sadat learned this the hard way after his historic visit to Israel in 1977, a visit that cost Sadat in several respects: “his foreign minister resigned, Egypt was expelled from the Arab League, and eventually Sadat was assassinated by extremists upset at the peace deal with Israel.”[14]  Despite this cost, Israeli defense officials worried that the move was a trick, and, crucially, Israeli President Menachem Begin remained unconvinced of Sadat’s peaceful intent.[15]  What the Sadat/Begin example illustrates is that perceptions of what is costly or a large move may depend not so much on objective factors, but rather subjective ones, including the psychologies of the actors. Diplomats and leaders clearly are making inferences based on the scope of demands, as Trager argues, but whether or not those inferences are the intended ones seems to rely on the notion that it is self-evident what constitutes a large demand versus a small one.

Relatedly, it is intriguing that much of the communication that takes place in the anecdotes, cases, and dataset occurs through face-to-face interaction.  As I have argued elsewhere, face-to-face diplomacy is a particularly useful form of communication for conveying intentions, though for a different reason than Trager identifies.[16] In my case I highlight the psychological and neuroscientific processes that make intention transmittal possible. This leads to the possibility that it may be that there are two discrete processes occurring when intentions are conveyed through private face-to-face diplomacy: the rationalist mechanisms identified in Trager’s book and the more psychological ones identified in mine.  This raises the prospect that these processes might work in tandem, but also that they may diverge.  What happens when a diplomat gains an intuition about his interlocutor’s lack of sincerity related to a specific intention, but the intention is buttressed by the type of information Trager identifies as crucial for revealing an intention?  Here the resolve versus intention distinction may be relevant.  This may be a fruitful area for future research.

Finally, there is a temporal aspect to resolve and intentions that is worth considering.  In most of the examples provided in the book, the intentions that are being conveyed tend to be those that we might consider to be in the present.  Most of the cases in the book involve discussion of intentions that relate to the following days, weeks, or sometimes months.  This represents the synchronic problem of intentions: how can states convey their intentions over a specific issue at the present time.  Trager has undoubtedly made a contribution to our understanding of this problem. But there is also the diachronic problem of intentions, which is to say that the intentions, or the actors, might change in the future.  This might mean changing intentions over a specific issue or the overall intention of the state, what we think of in terms of state ‘type,’ might change as well.  Can the mechanisms of costless communication help ward off the deleterious aspects of this future uncertainty problem?  As with the discussion of endogenous intentions, this is not precisely the question that Trager set out to answer, and thus is not a criticism so much as a question for future research.  The question that does seem relevant, however, is why states should trust the inferences they make about intentions that could easily change and involve a time horizon beyond the immediate?  For example, a large demand might reveal today’s intentions regarding Piedmont, but if the problem states which are negotiating over exists over a longer time horizon, such as the development of a nuclear program, states may not trust that their inferences about today’s resolve will remain the same in the future, leading to a question of whether the mechanisms Trager has identified are particularly useful for immediately-concerned intentions but less useful for ones that are in the not-too-distant future. Intentions, by definition, are about how actors will act in the future.  How far into the future is an important question that remains under-developed.

These questions and criticisms are opportunities for future research and further analysis.  They do not detract from Trager’s important book that will surely cause many to rethink, and perhaps update, priors about the role of costless communication.  It is likely to make a significant contribution to several literatures and provides yet another reason to take diplomacy, even the private type, seriously.



Review by Brian Rathbun, University of Southern California

Robert Trager’s Diplomacy is a landmark in the study of diplomacy specifically, and international communication and signaling more broadly.  For many years, the field has been waiting on a book written in the rationalist tradition that did not tell us again what was so obviously untrue: that diplomacy is cheap talk, lying is the norm, and costly signaling is the only way by which states can communication intentions to one another. More importantly, while relying extensively on formal analysis, the book sets about actually testing its empirical claims on real evidence, and is not an illustrative case study hastily chosen after the fact to give the theory some veneer of real world relevance. This is what formal theory should look like in international relations.

Trager’s book is truly ambitious and comprehensive.  It sets out to establish not just one way by which states might communicate absent costly signaling; it seeks to identify all such mechanisms (at least that the author can think of).  Trager uses extensive case studies but also a completely new data set, based on a coding of British archival sources that aims at looking for every inference made about the intentions of Britain’s rivals in partners over a significant period of time.  Trager’s communication mechanisms are also not, as is often the case with arguments derived from formal theory, trivial or obvious.  They are not mathematical representations of conventional wisdom or platitudes.

Of course, as a longstanding critic of the rationalist approach, I have my issues with the book but these are, paradoxically, based on what Trager gets right, in my opinion, not what he gets wrong. Let us take some of the mechanisms by which he hypothesizes that states might signal resolve in their relationship with others.  It seems that in all of them there are implicit social dynamics, perhaps based on universal norms of interaction between any human beings, that go against what I understand to be canonical rationalist assumptions.

Trager argues, for instance, that asking for a lot in a negotiation is a credible signal of resolve because it raises the risk that others might not bargain at all.  It is that very risk that makes the signal credible in the first place.  Why, however, from a rationalist standpoint, should asking for the moon lead the other side to not negotiate or to pull out of negotiations?  One should expect this type of behavior, not take it personally, ignore what the other said, and put forward one’s own outrageous set of demands or some compromise.

We of course know that this is quite rare in international affairs because diplomacy is conducted by humans.  Humans get outraged (or at least are forced to feign outrage to their domestic and international audiences); they take their toys and go home, even when staying and trying to work through disputes is in their interest.  That is an emotional and irrational response, but also one that is quite common.  Trager’s theory seems to work empirically precisely because it is predicated on certain universal truths about human irrationality.  He is laying out how communication can occur given the fact that spite is so common.  The fact that we can expect others to punish us for being greedy, for instance, even if it makes no sense for the punisher, allows us to communicate how much we care about an issue.

The same is true of another of Trager’s mechanisms.  Pushing an issue to a crisis point is credible because it threatens to disrupt one’s overall relations with a target state.  If President Donald Trump, for instance, puts tariffs up on Chinese steel, he can forget about any help from China in dealing with North Korea, which makes his resolve on the trade issue that much more credible.  But why, from the Chinese standpoint, should they make Trump’s life any harder on non-trade issues if doing so would also cost them by disrupting other mutually beneficial areas of cooperation.  This makes little sense in a rationalist framework.  It is only if China responds to Trump irrationally, by cutting off ties and punishing the United States across issues, that Trump’s original action conveys any information at all.

Again, I think that this is precisely how international relations works because they are inter-human in nature.  In our personal lives, we would not expect the butcher to give us the best cut of meat if we intentionally scratched his car in the parking lot.  But from a rationalist standpoint the butcher still profits from the transaction so what has changed?

The last mechanism of signaling I will discuss here is that of approaching a third state.  Trager argues that states can get what they want vis-à-vis some other by improving the general state of their relations with that the adversary of that other.  Again, I agree that this happens empirically all the time.  By why should it be the case from a rationalist standpoint, in which relations are just sets of transactions that have some benefit for both sides?  Trager essentially seems to be saying that states develop positively affective relationships with one another that are threatening to those that are excluded.  International relations is like a particularly clique-ish high school lunch table. Again, I agree, but how exactly does this make any sense without an underlying set of assumptions about the social behavior of states and the humans that lead them?

Is this a criticism of Trager’s book?  Not for me.  I think it is part of why it is so marvelous.  I am merely trying to draw out what I see to be implicit assumptions that make it such an improvement on other rationalist work.  As it is, however, Diplomacy is a model of rigorous empirical analysis and bold theory that should set the standard going forward for formal approaches to international relations.



Review by Anne E. Sartori, Massachusetts Institute of Technology

Diplomacy’s outsized role in interstate relations can be seen by the scarcity of actions that take place without it.  In addition to page-one episodes of crisis or peace-conference diplomacy (such as the to-and-fro between U.S. President John F. Kennedy and Soviet Premier Nikita Khrushchev in the Cuban Missile Crisis, and the Congress of Vienna), diplomacy involves countless, often-tedious exchanges of bureaucrats, whose discussions aim to coordinate or affect policies concerning trade, arming, alliances, migration, global health, territorial claims, nuclear deterrence, and more.  All of this activity has enormous effect.

While the field of diplomatic history unfortunately has been under siege recently,[17] the study of diplomacy in the political-science subfield of international relations has been flourishing over the past couple of decades.[18]  Robert F. Trager’s new book, Diplomacy; Communication and the Origins of International Order, is a major addition to this international-relations literature and a social-science masterwork. To speak of it as one contribution to the literature does not do it justice. Focusing on diplomacy as communication, it proposes not one, but four major theoretical innovations and relates these carefully to history using both historical analyses related to the origins of the two World Wars and analyses of two amazing new datasets based on British foreign office documents from the years leading up to World War I.

The success of diplomatic communication is a puzzle because leaders often have incentives to misrepresent the information they possess in order to obtain more favorable settlement terms.[19]  For example, if a leader’s statement that his/her state is resolved to fight a war over a particular issue generally leads to a better outcome, then leaders of irresolute states also will claim resolve, and so the adversary will not be able to learn from such claims. In general, solutions to the puzzle require some reason for speakers to admit truthfully when they do not have the quality that, if communicated credibly, would result in the favorable outcome.  If states generally are willing to acknowledge that they lack the useful quality, states that claim to have it are likely actually to do so.

The early rationalist literature resolved this puzzle by theorizing that states were able to communicate when their diplomacy was a form of what game theorists call “costly signaling.”[20] A costly signal is one that an actor pays a direct cost to send; for example, engaging in a skirmish is costly in lives lost.  However, much of diplomacy involves words, and the direct costs of writing or speaking are negligible.  Moreover, the direct costs of some other forms of diplomacy, such as moving aircraft carriers, or starting weapons programs, are quite small compared to the potential benefits of successful signaling, for example deterring an attack on oneself or an ally. A more-recent body of work deepens the field’s understanding of diplomacy by explaining how communication that involves no direct costs to the sender (often called ‘cheap talk’) can be informative as well.  The costless-signaling literature shows that states do not need the ability to use costly signals (such, as some argue, comes with the existence of a powerful domestic public) for their states’ diplomacy to be effective.  For example, states with any regime type can accrue reputations for using their diplomacy honestly, and states’ diplomacy is more credible when they have such reputations.[21]

Trager’s book adds substantially to this literature on the effectiveness of costless diplomacy. The theoretical contribution of Diplomacy consists of four solutions to the puzzle of how states can communicate in different situations, despite their incentives to use diplomacy deceptively.  The empirical chapters demonstrate that Trager’s mechanisms help to explain when and why diplomacy leads policy-makers to change their minds, focusing in particular on the origins of the two World Wars. The common theme is that diplomacy can be credible and informative under a wide range of circumstances, a much wider range that the literature previously had identified.

In the theoretical chapters, Trager proposes two solutions to the communication problem in the context of bilateral relations, and two in the context of trilateral relations.  In the bilateral context, he argues that states sometimes will refrain from making high demands deceptively (leading such demands to be credible when they do make them) because over-claiming risks reaching no agreement at all. Similarly, they sometimes will refrain from claiming high resolve deceptively because doing so risks provoking an adversary into mobilizing forces, forming new alliances, or taking other action that increases its own security and decreases that of the threatening state.[22]  In the trilateral context, he argues that a state’s commitment to aid a protégé in the event of war can be credible because the commitment may embolden the protégé to start a conflict; states that are very reluctant to fight are also reluctant to commit to the protégé because doing so involves this additional risk of war.  He also argues that a state’s overtures to negotiate or renegotiate an alliance on terms less favorable to itself not only are credible, but indicate that the state is planning for war with a third party, because the reason to make the concessions is to shore up support for the likely conflict.

The chapters that connect the theory to history through empirical work particularly impressed me.  Typically, the purpose of game theoretic models of diplomacy is to explain how, and therefore when, one state’s diplomacy can change the beliefs of leaders of another state, and by doing so can affect the other state’s actions.  Yet, because scholars have so much difficulty ascertaining what was in leaders’ minds, the quantitative empirical investigation of these models’ explanations typically ignore the effect of diplomacy on beliefs, focusing entirely on how a state’s diplomacy affects others’ observable actions, for example the conditions under which a threat to use force is associated with backing down by the adversary.[23]  In other words, while scholars believe that we know how decision-makers’ beliefs should be changing, we largely ignore those ideas when it comes to statistical work because we consider it impossible, or too difficult, to measure the beliefs of historical figures, whom scholars cannot survey and upon whom they cannot experiment. Even experimental studies on this subject rarely attempt to investigate the conditions under which diplomacy changes the beliefs of an adversary; this work, too, focuses on whether it affects the adversary’s actions.[24]

To study inferences directly, Trager created two related datasets from the Confidential Print of the British Empire between 1855 and July of 1914. The Confidential Print, “is a record of documents circulated by the British Foreign Office to the cabinet, officials of the Foreign Office, and the King and Queen” (23). The first dataset contains inferences drawn between 1855 and July, 1914 “about the security-related behavior and intentions of other great powers in Europe” (259). The second “comprise(s) all demands, offers, and assurances that occurred between 1900 and the start of the World War in 1914” noted in the Confidential Print, made by European powers other than Great Britain, and classified as “security-related” (261).

The Confidential Print data provide a fascinating window into beliefs and inferences. While historical studies can consider inferences in particular cases, Trager’s data allow him to assess general patterns in how and when diplomacy changes beliefs.  In particular, the data allow him to test whether decision makers change their beliefs in ways that his theories predict they would, which, generally, they seem to. For example, as one of his models predicts, a weaker adversary conveys more information through its threats, since it risks more by making them (201).  The data also allow Trager to answer questions about how decision makers draw inferences that are not directly related to his theories but are important to the field. For example, he provides the percentage of inferences that these political actors made between 1855 and 1914 that came from military factors such as the building of weapons or relative power calculations, and what percentage came from diplomacy (25). That a much greater percentage were drawn from diplomacy suggests that military factors “are by no means the only or even the primary path by which British foreign policy elites reached conclusions” (25).

While the Print data advance our knowledge, there is a limit to what scholars can learn from them and other inference data. Diplomats’ writings, including in the Print, may not tell us exactly what they were thinking for several reasons. First, when government representatives write political memoranda and other intra-government communications, they are not merely communicating information; they also face numerous political considerations, including presenting information in such a way as to convince others of their preferred course of action, presenting themselves or their group (their embassy, desk, etc.) in a favorable light to others, and sometimes presenting competitors or even particular foreign countries in an unfavorable light.  Second, even if we assume that the Print authors were communicating their beliefs sincerely, diplomats are often in a rush and always human, and may not understand at the time of writing exactly why they are reaching particular conclusions. Finally, diplomats select some topics to be covered more than others based on their assigned jobs, their organizational missions.  For example, those who work for the Foreign Office might be more likely to write memos about information received through diplomatic channels, while those who work for the military might be more likely to write them about military matters. These problems are not unique to quantitative data, but in qualitative work, the researcher may be able to substantiate an inference made by a political actor and/or the source of that inference through documents written by other actors involved in the same set of events.

For these reasons, data like these can complement, but not replace, observational and experimental data about statements and actions, as well as the experimental data that researchers could gather about beliefs.  Trager seems to acknowledge this fact implicitly because he combines information about inferences and actions to study diplomacy prior to the World Wars. For example, he argues that the events of 1938 meet the conditions of his scope-of-demands theory, which implies that the initial British concessions would lead German leaders to infer that Britain would accept even greater ones (176-178). Since the Germans then made greater demands, he infers that this implication was borne out in the case.

In Diplomacy, Trager uncovers and demonstrates many of the ways in which words and writings can affect the beliefs, and so presumably the actions, of leaders of other states. He seamlessly melds theory and the qualitative and quantitative study of history, creating a powerful book that is an exemplar of successful social science and a joy to read.


Response by Robert F. Trager, University of California, Los Angeles

Sources of Expectations in Diplomatic Practices

Where these responses critique my book, I find I am often in agreement.  I wish therefore to focus my response on clarifying a couple of areas where my views may diverge somewhat from those of the reviewers.  I am deeply grateful to Todd Hall, Marcus Holmes, Brian Rathbun and Anne Sartori for their insights and careful engagement with the text.

Several of the reviewers suggest that the models in the book demonstrate both the promise and the limitations of rationalist models.  I agree, and I wish to attempt to state exactly how I believe this is so.  Brian Rathbun, for instance, argues that irrational emotions are needed to justify some of the behavior predicted by the rationalist models.  I believe he is on to something, but what this may be is not simple to articulate.  The models apply only conventional understandings of rationality, not psychological dynamics that cause choices to diverge from goal following behavior.  He is certainly right that rationalist models and models that do include emotions of the sort Rathbun references, such as outrage, may predict the same behavior.

In fact, I believe he is on to something even more interesting, but first it is instructive to consider just how rationalism fares on its own in the example of signaling dynamics that Rathbun discusses.  He notes that in the rationalist models, a large demand can imply a credible threat, but he argues that emotions like outrage are needed to understand this process.  His view is that only emotional reactions of this sort can explain why a state would take actions against the first state’s interests in response to the large demand.  Yet, if an adversary makes a demand that conveys that it will go to war unless it receives a dramatic concession, what are sensible responses of a state whose leaders feel no outrage but are unwilling to comply?  If the demand is that the state give up its nuclear weapons program, the threatened state would not seek to avoid war through a compromise of giving up some of its conventional capabilities instead.  Since the demand is over nuclear capabilities, this concession only weakens the state without avoiding the conflict.  Concessions often require an underpinning of trust.  The threatened state will instead consider other options, like forming closer ties with a state that can offer protection, or even, if it is convinced that war is coming anyway, striking first.

So, the dynamics of the models do not require outrage or other deviations from rationality, but they do require that actors have certain expectations.  In fact, the models demonstrate that particular sets of expectations of behavior are sufficient to produce communication between rational agents.  Where do these expectations come from?  In some strategic contexts, I argue that the expectations that form part of the equilibria described in the book are the only reasonable equilibria.  But that is not true in all contexts.  I argue, for instance, that in some strategic contexts, if actors believe all concessions are a sign of weakness, then all concession will be so; however, if actors are ready to see concessions as an attempt to achieve mutually beneficial outcomes, they sometimes will be.  A property of these models is that expectations shape behavior that reinforces the expectations – even when other sets of expectations could also be self-reinforcing through producing other behavior.

Sometimes, therefore, we need to go beyond rationalism to understand the sources of expectations and rational responses to strategic context.  It might be that expectations derive from the past: agents have behaved according to these patterns in the past and therefore are expected to do so in the future.  Alternatively, cultural norms that are usually applied in other contexts may imply certain expectations about behavior in a diplomatic context.  Or, and here we return again to Rathbun’s fine points, knowledge of emotional responses may be a basis for expectations.  Locating the sources of expectations in strategic, cultural, or psychological factors—and in individual attributes versus group processes—are fascinating topics for scholars.[25]

Now, to ontology and incommensurability.  Todd Hall worries that the desire to move beyond paradigm wars might lead important distinctions to be lost or at least unappreciated.  I agree, but, as with any Wittgensteinian language-city, I think it depends on how the bridges are built.  If we assign the study of preferences to constructivists and the study of interaction to rationalists, we blind ourselves to significant insights from both sets of scholars.  If, however, we see rationalist accounts as a proposal for one form of social world to be evaluated against others, then all is well.  Both camps would benefit from a greater appreciation of the extent to which rationalist approaches are consistent with a radical contingency of social relations.  Rationalists often forget how much is bundled into specific modeling assumptions, while constructivists appreciate that something has been baked in through the assumptions, but often falsely associate this with the assumptions of rationalism itself rather than the assumptions of particular models.

As an example of a case where differing approaches to diplomacy can have much to say to each other, consider the idea raised by Marcus Holmes of forming intentions through interaction, which remains fascinating to me.[26]  I believe there is a danger that the contrasts we draw between rationalist and constructivist approaches may obscure more than they reveal.  In none of the rationalist models in my book do the actors know, at the start, what actions they will take by the end.  Their later choices will depend on the choices of the other actors.  The actors’ actions depend on the beliefs they have at the start, and on how those beliefs evolve through interaction, which means that the beliefs of one actor influence the later beliefs of other actors and vice versa.  Thus, what they intend changes through a complicated process of social interaction.  In conversation, for instance, it may not be rational to reveal certain information until other information has been revealed by another actor and only once all information is known might it make sense to form what could be called a collective intention.[27]

These critiques and others suggest research agendas.  Anne Sartori discusses the consequences of focusing on decision-makers’ own accounts of the reasons for their judgments as I do in analyzing the Confidential Print of the British Empire, and Hall expresses similar concerns.  Stated reasons may be rationalizations; what is unsaid – and overlooked in this part of my analysis – may be as important as what is said.  Holmes asks questions about how the judgments of today are altered through time.  How we think about the formation of orders in world politics may hinge on the answers.  For my part, I expect orders to continually reconstitute themselves through diplomatic interaction.  Each interaction reveals and alters complex sets of underlying assumptions about the intentions and beliefs of others, implying both change and path dependence.  These long-term processes appear essential to international politics and remain understudied.[28]  The respondents were not always sure I should be satisfied leaving so many important issues and questions they highlight to future work; but think how interesting that future work will be.



[1] James D. Fearon, “Rationalist Explanations for War.” International Organization 49:3 (1995): 379-414; Jessica L. Weeks, “Autocratic Audience Costs: Regime Type and Signaling Resolve.” International Organization 62:1 (2008): 35-64.

[2] Anne E. Sartori, “The Might of the Pen: A Reputational Theory of Communication in International Disputes.” International Organization 56:1 (2002): 121-149.

[3] Emilie M. Hafner-Burton, Stephan Haggard, David A. Lake, and David G. Victor.  “The Behavioral Revolution and International Relations.” International Organization 71:S1 (2017): S1-S31.

[4] See, for instance, Stephen Van Evera, “The Cult of the Offensive and the Origins of the First World War.” International Security 9:1 (1984): 58-107.

[5] Rudra Sil, and Peter J. Katzenstein.  “Analytic eclecticism in the study of world politics: Reconfiguring problems and mechanisms across research traditions.” Perspectives on Politics 8:2 (2010): 411-431.

[6] Robert Jervis, The Logic of Images in International Relations (Princeton, NJ: Princeton University Press, 1970).

[7] Joshua D. Kertzer, Resolve in International Politics (Princeton: Princeton University Press, 2016).

[8] Kertzer, 3.

[9] Jennifer Mitzen, Power in Concert: The Nineteenth-Century Origins of Global Governance (Chicago: University of Chicago Press, 2013).

[10] Vincent Pouliot, International Security in Practice: The Politics of NATO-Russia Diplomacy (Cambridge: Cambridge University Press, 2010).

[11] Paul Sharp, Diplomatic Theory of International Relations (Cambridge: Cambridge University Press, 2009).

[12] Ole Jacob Sending, Vincent Pouliot, and Iver B. Neumann, “Introduction,” in by Ole Jacob Sending, Vincent Pouliot, and Iver B. Neumann, eds., Diplomacy and the Making of World Politics (Cambridge: Cambridge University Press, 2015), 6.

[13] Nicholas J. Wheeler, Trusting Enemies: Interpersonal Relationships in International Conflict (Oxford: Oxford University Press, 2018), 8.

[14] James Morrow, “The Strategic Setting of Choices: Signaling, Commitments, and Negotiation in International Politics,” in David A. Lake and Robert Powell, eds., Strategic Choice and International Relations, (Princeton: Princeton University Press, 1999), 87.

[15] Wheeler, 29.

[16] Marcus Holmes, Face-to-Face Diplomacy: Social Neuroscience and International Relations (Cambridge: Cambridge University Press, 2018).

[17] William R. Keylor, H-Diplo State of the Field Essay, “The Problems and Prospects of Diplomatic/International History,”, 10 April 2015; Marc Trachtenberg, H-Diplo State of the Field Essay: “The State of International History: Where We’ve Been, Where We Are, and Where We’re Going”

[18] For a recent review of this literature, see Trager’s own piece: Robert F. Trager, “The Diplomacy of War and Peace” Annual Review of Political Science 1 (2016) 9: 205-28.

[19] James D. Fearon, “Rationalist Explanations for War,” International Organization 49:3 (1995): 379-414.  Previous waves of international-relations scholarship were skeptical about the prospects for successful diplomacy for various reasons.  Some scholars argued that communication was difficult due to biases in the way that humans process information. Others saw diplomacy as primarily reflecting the will of the powerful.  On the former, see, for example, Robert Jervis, Perception and Misperception in International Relations (Princeton: Princeton University Press, 1976); Richard Ned Lebow, Between Peace and War (Baltimore: The Johns Hopkins University Press, 1981). On the latter, see, for example, Henry Kissinger, Diplomacy (New York: Simon and Schuster, 1994).

[20] For example, see James D. Fearon, “Domestic Political Audiences and the Escalation of International Disputes,” American Political Science Review 88:3 (September 1994): 577-592; Jessica L. Weeks, “Autocratic Audience Costs: Regime Type and Signaling Resolve,” International Organization 62 (Winter 2008): 35-64.

[21] Anne E. Sartori, Deterrence by Diplomacy (Princeton: Princeton University Press, 2005). An early contribution by Trager not included in Diplomacy is “Multidimensional Diplomacy,” International Organization 65:3 (2011): 469-506.

[22] The first idea is related to two other important works on cheap-talk diplomacy: Kristopher Ramsay, “Cheap Talk Diplomacy, Voluntary Negotiations, and Variable Bargaining Power,” International Studies Quarterly 55:4 (December 2011): 1003-1023; Shuhei Kurizaki, “Efficient Secrecy: Public versus Private Threats in Crisis Diplomacy,” American Political Science Review 101:3 (August 2007): 543-558. In the former, states sometimes avoid a hard-line position to convince the adversary to come to the bargaining table; in the latter, they sometimes send the ‘weak’ signal of opting for private diplomacy in order to reach an agreement.

[23] For example, to see whether democracies’ threats are more informative, Kenneth Schultz looks to see whether their threats are more likely to lead adversaries to back down than are threats by leaders of other types of polity. Kenneth A. Schultz, “Do Democratic Institutions Constrain or Inform? Contrasting Two Institutional Perspectives on Democracy and War,” International Organization 53:2 (Spring, 1999): 233-266.

[24] For example, see Dustin H. Tingley and Barbara F. Walter, “Can Cheap Talk Deter? An Experimental Analysis,” The Journal of Conflict Resolution 55:6 (December 2011): 996-1020. Experiments would seem to lend themselves more than observational work to measuring inferences, but in any case they are a complement to rather than a substitute for observational data for a variety of reasons, including that the experimental subjects are not actual decision-makers and do not face actual diplomatic problems.

[25] See, for instance, Todd H. Hall, Emotional diplomacy: Official Emotion on the International Stage (Ithaca: Cornell University Press, 2015); Marcus Holmes, Face-to-Face Diplomacy: Social Neuroscience and International Relations (Cambridge: Cambridge University Press, 2018); Brian Rathbun, Diplomacy’s Value: Creating Security in 1920s Europe and the Contemporary Middle East (Ithaca: Cornell University Press, 2014) and Anne Sartori, 2013. Deterrence by Diplomacy (Princeton: Princeton University Press, 2013).

[26] Vincent Pouliot, International Security in Practice: The Politics of NATO-Russia Diplomacy (Cambridge: Cambridge University Press, 2010); Jennifer Mitzen, Power in Concert: The Nineteenth-Century Origins of Global Governance (Chicago: University of Chicago Press, 2013).  One might also consider Holmes’s point that one of the signaling models requires a prior common understanding of what constitutes a “large” or a “small” demand.  I agree, but this should be seen in light of the fact that intentional communication always requires prior shared understandings to be possible, whether we conceive of these as “logical simples” (Ludwig Wittgenstein, Tractatus logico-philosophicus (London: Routledge, 2013)) or as a shared “form of life” (Ludwig Wittgenstein, Philosophical investigations, trans. G.E.M. Anscombe (New York: Basic Books, 1953)).

[27] Robert J. Aumann, and S. Hart, “Long Cheap Talk,” Econometrica 71:6 (2003):1619-1660.

[28] Paul Pierson, “Big, Slow-Moving, and… Invisible,” Comparative Historical Analysis in the Social Sciences (2003), 177-207.