Essence of Negotiation: Understanding Appeasement and “The Great Munich Stereotype”

Because negotiation is an essential activity in business, government, and international relations, it has attracted research across the social sciences. Economists and some political scientists have described bargaining as a noncooperative game between rational actors (Kreps 1990). Psychologists, sociologists, and other political scientists have studied negotiation as an exchange between boundedly rational actors. Emotional and prone to “cognitive illusions,” such negotiators make mistakes that may generate or prolong conflicts (Thompson 2006). This research comprises the so‐called “behavioral” paradigm. Proponents of one theory often criticize the merits of the other (Lucas 1986; Coursey and Roberts 1992; Walt 1999), though without yet producing much consensus.

Graham Allison (1971) examined the events of the 1962 Cuban Missile Crisis to illustrate the strengths and limitations of three decision‐making paradigms, including the behavioral and rational choice approaches. Because that study predated breakthroughs in noncooperative game theory, he restricted his assessment of rational choice to individual decision theory, which is not well suited to the study of negotiations (Bendor and Hammond 1992). Behavioral research has also advanced considerably in the years since Allison's study. In this article, I adapt Allison's case‐based approach to assess these paradigms in light of an earlier, but even more complex, negotiation, the Paris Peace Conference of 1919. Comparatively recent disclosures of a transcript of the negotiation process (Mantoux 1992) and the diligent work of historians (Schuker 1988; Ferguson 1996) make it possible for scholars to now undertake a new, more thorough analysis of this historically significant negotiation.

Allison contrasted the behavioral and rational perspectives with a third paradigm that he called “bureaucratic politics.” Researchers using this paradigm treat the larger moves made by each government as the product of explicit and implicit bargaining by individuals striving, within their own organizations, to pursue their own interests given their unique perspectives. The decision of the United States to set up a naval quarantine of Cuba rather than launch unilateral airstrikes emerged from “pulling and hauling” by individual actors working within different departments inside the government.

In the years since Allison's study, scholars have proposed alternative behavioral (Pfeffer, Salancik, and Leblebici 1976) and rational choice explanations for bureaucratic politics (Milgrom and Roberts 1988). Jonathan Bendor and Thomas Hammond (1992) argued that the theorist's major task is to determine the proper level of analysis for studying decision making. The bureaucratic bargaining behavior displayed by representatives of the U.S. Defense and State Departments during the Cuban Missile Crisis could be explained by completely rational strategies developed by those actors (Dixit and Skeath 2002). Or, they might reflect the application of highly simplifying heuristics that generate cognitive illusions (Janis 1973).

In considering the case of the Paris Peace Conference, I note up front that the U.S. Senate voted down the treaty drafted by U.S. President Woodrow Wilson on three separate occasions. This vote surprised and greatly dismayed the authors of the treaty, compelling Wilson's successor, Warren G. Harding, to eventually draft and sign a separate peace treaty with Germany. No unitary actor model of the U.S. government, such as the one Allison analyzed under the guise of rational choice, could explain such a sequence of events. Some form of bureaucratic politics perspective will be necessary. But does a rational actor model of bureaucratic politics or a behavioral model of bureaucratic politics provide a more satisfactory explanation?

Actually a distinctive third paradigm has emerged in the decades since Allison's study (1971). What is sometimes referred to as “the new institutional economics” derives from a somewhat intermediate point of view with respect to human rationality. This viewpoint was perhaps best articulated by Oliver Williamson (1985) in his discussion of “contractual man,” an individual who embodies certain elements of the traditional, rational “economic man” as well as the error‐prone actor found in behavioral theories. Williamson's paradigm presumes that people's information‐processing capabilities vary widely but that organizational hierarchies are able to sort these capabilities efficiently.

According to rational choice theory, all decision makers have consistent and complete preferences across all possible future states of affairs that may arise in the world of the decision makers. They process information in a manner consistent with Bayesian probability theory so choosing as if they were maximizing expected utility. David Kreps (1988) deemed them “trade‐off talking rational economic persons” or TOTREPS. Noncooperative game theory further assumes that these TOTREPS have a boundless capacity for perspective taking. They can fully comprehend and accurately analyze the decisions that other actors confront. They choose as if they had awareness of the choices, information, and contingencies facing all relevant actors.

By contrast, the contractual person in the new institutional economics is boundedly rational, self‐interested, and “guileful.” This person behaves in a highly opportunistic manner (Williamson 1999). Because of these human qualities, establishing a hierarchy is the preferred mode of decision making rather than relying on the invisible hand of market forces to efficiently mediate individual decisions. Administrators in a hierarchy efficiently match the capabilities of individuals to the information‐processing demands that will be placed on those individuals, so that complex organizational decisions will not be distorted by the cognitive illusions that mold individual choices. As Williamson (1985: 397) observed, in a hierarchy, “responsibility for processing and displaying the consequences of probabilistic choices is concentrated on the more competent types.”

These three paradigms yield distinctive interpretations of a complex negotiation such as the Paris Peace Conference. Using them to help explain this event provides perspective on the merits and limitations of each paradigm. Understanding the event provides a clearer perspective on the relationship between the three approaches. Though most often seen as competing schools of thought, they should more properly be seen as necessary complements.

A second objective of this article is to better understand this pivotal negotiation. Did it produce “a flexible treaty crafted by relatively well intentioned and rational leaders” (Ikenberry 1999)? Or was “the nightmare of Adolf Hitler at the peak of his power” caused by “the vindictiveness of the British and French peace terms,”“the exclusion of Germany and Russia from the peace conferences,” and “foolish attempts to draw the blood of reparations and war debts” (Kennan 1996)? The three decision‐making paradigms cast light on such claims, which also have contemporary relevance.

Today critics routinely condemn particular diplomatic initiatives by comparing them to the failed appeasement practices of British Prime Minister Neville Chamberlain (Stolberg and Rutenberg 2008). That pattern of critique was so sufficiently pervasive by the middle of the last century that Walter Lippmann (1966) deemed it “the Munich stereotype.” He pointed out how reliance on this stereotype had repeatedly led various world leaders to make serious policy errors in high‐level negotiations. With this article, I seek to develop a better understanding of the origins of the Munich stereotype and the “blind spots” or perceptual errors that its casual application can generate.

This article follows a description of the background of the Paris Peace Conference with a discussion of the different concepts about contracts used in each decision‐making paradigm. Each paradigm generates different postdictions (deductions regarding events that took place in the past) about the bargaining behavior of the conferences participants, the form of the treaty that resulted, and its implementation. I then describe data for testing these postdictions.

Thirty‐three nations participated in the Paris Peace Conference of 1919 to negotiate the end of a war that had consumed much of Europe. The conference arose from the German request for peace based on principles declared by U.S. President Woodrow Wilson in 1917 (see the  Appendix). The Armistice signed in November 1918 forced Germany's emperor, Kaiser Wilhelm II, to abdicate so that a democratic government could negotiate the final treaty. Best known for the Treaty of Versailles between the Allied Powers (primarily the United Kingdom, France, and the Russian Empire) and Germany, the negotiations also produced treaties with each of the other Central Powers (the states allied with Germany during the war, including Austria–Hungary and the Ottoman Empire).¹

Lacking useful precedents, the initial stages of the conference, which began in January, addressed organization. Initially, all stakeholders assumed that a congress with representatives of the defeated nations would follow a conference among the Allied Powers (Tardieu 1921). Without serious debate or much advance warning, that working assumption would later be abandoned. German representatives were summoned to Versailles at the end of April. Isolated, under heavy guard, and cut off from information, they were finally handed a draft on May 7. They were permitted to submit written notes containing questions for clarification of the meaning of provisions but only then learned that they faced not direct negotiation, but an ultimatum (Luckau 1941).

At the head of the conference sat a formal executive overseeing more than fifty committees and subcommittees. But important decisions were ultimately made within a small circle. The Council of Ten, the heads of government and foreign ministers of the five “great powers” (United Kingdom, France, Italy, Japan, and the United States), met seventy‐two times. By May, control had been restricted to the Council of Four consisting of the heads of government from the United States, Great Britain, France, and Italy. They held 145 meetings with an interpreter/transcriber and various experts summoned for particular questions. The council considered issues ranging from recognition of the new communist government in Russia to ascertaining Arab desires for self‐determination (Mantoux 1992).

The German settlement alone covered a vast range of subjects, from borders to disarmament to the disposition of colonies. Many of the provisions, particularly the decision to restrict German armaments, to occupy the Rhineland, to restrict any Austrian union with Germany, and to grant control of German‐speaking regions in Bohemia and Silesia to other governments, were controversial. (These will be discussed in the second part of the article dealing with implementation of the agreement.) The contract analysis in the first part of this article will focus on Articles 231–247, which concerned reparations. Compared to war crimes, borders, or colonies, financial matters should, in theory, be most amenable to rational calculation. The Council of Ten quickly appointed a multinational “Commission on the Reparation of Damage” (CRD) to construct draft language for a settlement. Negotiations then proceeded within and between the two councils, the CRD, and various subcommittees, before eventually passing to the German delegation in the form of the ultimatum.

Recent research has addressed the form and function of contracts. The contracts devised by purely rational actors with unlimited cognitive resources should be “complete” (Salanié 2002). That term means that the contract explicitly describes all relevant contingencies that could affect the welfare of the parties for the life of the instrument. Rights and obligations are precisely stipulated, permitting a clear allocation of costs and benefits between the parties. Such a contract is only conceivable in a world where cognitive resources are limitless.

According to the contractual and behavioral paradigms, the form and function of contracts are very different than they are according to the rational choice characterization. Because of their bounded rationality, the parties cannot conceive every contingency, even when the transactions address fairly simple problems (Williamson 1985; Robinson and Rousseau 1994). They rely instead on “relational” contracting (Macneil 1978). Such a contract establishes general parameters that govern the ongoing relationship between the parties. Because of unforeseen contingencies that cannot be expressly covered by the terms of the contract, “gaps” in the contract language occur, which are addressed by “fillers,” language that enables the parties to work out solutions to whatever problems may arise. Common fillers include “the use of a standard uncontrolled by either of the parties,”“direct third‐party determination of performance,”“one‐party control of terms,” and “an agreement to agree” whereby they pledge to “fill gaps in their relation” later in time (Macneil 1978: 866–867, 870).

The Paris Peace Conference's formal agreements have long been public record (Allied and Associated Powers 1920), and diplomatic notes, plans, drafts, and reports were eventually also made public with the expiration of the “fifty‐year rule” governing confidentiality of diplomatic papers. A transcription of the Council of Four's deliberations is particularly relevant (Mantoux 1992). From March on, they met daily in Wilson's residence with the interpreter recording each statement so he could translate it for the other participants. Along with the Council of Ten's minutes, these records trace the negotiation process that led to the formulation of various treaty provisions.

The major powers constructed large organizations incorporating delegates, expert advisors, and staff. Germany, for example, sent six plenipotentiaries, twelve commissioners, thirteen secretaries, and fifteen advisors who worked at Versailles from April 29 to June 23 (Luckau 1941). Many participants kept diaries (e.g., Seymour 1928), described events in letters to friends and family members (e.g., Headlam‐Morley 1972), and/or published extended narratives (e.g., Lansing 1921). These sources provide considerable insight into the beliefs, attitudes, intentions, and actions of the negotiators on all sides.

For this analysis, I studied written accounts from each of the major powers, including accounts from plenipotentiaries of each major signatory nation, such as French Prime Minister Georges Clemenceau (1930) and from their staff experts, such as British economist John Maynard Keynes (1919). Where it proved possible, I compared contemporary accounts with statements made in retrospect years later. (A complete table of all these sources is available upon request.) I also considered the work of historians (Ferguson 1996), economists (Schuker 1988), and econometricians who have systematically studied the implementation of the reparations terms (Sargent 1982; Webb 1989; O'Loughlin 2002).

This analysis focuses on reparations, but the later discussion will also address some of the structural similarities between the reparations section and other parts of the Treaty of Versailles. Although reparations were a major issue, other contentious problems included German disarmament, control over the Rhine and the Saar region, the right of Austria to self‐determination, membership in the League of Nations, and the borders of new states to be established, such as Poland and Czechoslovakia. Because these new nations would necessarily contain large numbers of ethnic Germans, protecting their interests and determining their rights to self‐determination presented a difficult problem. Because the many other treaties constructed during the conference were directly patterned on the Versailles model, much of this analysis should pertain to those agreements as well.

The reparations section of the treaty was divided into two parts: general provisions, consisting of thirteen articles and seven annexes, and specific provisions that contain only three articles. The specific provisions were closest in form to the type of contract provisions envisioned by proponents of rational choice theories. But the terms they specified are of considerably less importance than those handled under the general provisions section. Article 246, for example, specified the German government's obligation to return “the original Koran of the Caliph Othman” and “the skull of the Sultan Mkwawa” to the people from which the Germans had taken them.² Of course, these terms reflected far less significant interests compared to such high‐priority issues as the amount, mode, and timing of German financial reparations to the Allied and Associated Nations, which were dealt with in different form under general provisions.

During the Armistice, the United States had proposed — and Germany had agreed to — a framework for “restoration” of damage under Wilson's Fourteen Points (see the  Appendix). U.S. Secretary of State Robert Lansing's clarification of the term “restore” was accepted by all parties: “By it they understand that compensation will be made by Germany for all damage done to the civilian population of the Allies and to their property by the aggression of Germany by land, by sea, and from the air” (Lansing 1921; Burnett 1940: 411). After months of bargaining over these matters, Article 233 of the treaty ultimately indicated that the financial obligation of Germany was a gap to be filled at a later time. “The amount of the above damage for which compensation is to be made by Germany shall be determined by an Inter‐Allied Commission, to be called the Reparation Commission and constituted in the form and with the powers set forth hereunder and in Annexes II to VII inclusive hereto. This Commission shall consider the claims and give to the German Government a just opportunity to be heard. . . . The Commission shall concurrently draw up a schedule of payments” (Allied and Associated Powers 1920). Belgium, France, Great Britain, Italy, Japan, the Serb–Croat–Slovene state, and the United States could each designate one delegate and one assistant to this commission.

So the treaty left gaps with respect to transactions both minor (such as Mkwawa's remains) and major, that is, German financial obligations. The language deferred problem resolution to a multinational organization, a particular type of gap‐filling mechanism. Why did the treaty establish the Reparation Commission and what problem did it solve? Why did the treaty's authors not write out a detailed statement of German obligations, Allied rights, and relevant contingencies? Primary sources as well as secondary accounts by key informants help answer these questions.

Given the complexity of the economic issues involved, the Council of Four turned to experts for ways to settle the reparations issue (Lamont 1921). Wilson relied upon a group of government officials and business leaders including Thomas Lamont of the Morgan Bank. Clemenceau deferred most to armaments minister and industrialist Louis Loucheur (Burnett 1940; Loucheur 1962). British Primer Minister David Lloyd George consulted with Keynes and with banker Lord Walter Cunliffe, among others.

The CRD first considered problem framing. Delegations submitted separate statements outlining their preferred settlement principles (Burnett 1940). The British statement, soon endorsed by every delegation except the United States, declared “Reparation is not a technical word. . . . It is the making good of the losses which a party injured has sustained by wrongful acts and their consequences, so as to replace him in as good a position as that which he occupied before the wrong was done. It is effectuated by material means and affords full compensation for the real effects of the wrong” (Burnett 1940). So each nation but the United States agreed that they sought full reparation from the Central Powers for all the costs incurred as a consequence of the war.

John Foster Dulles, legal counsel to the U.S. delegation, submitted a competing American view that identified a different reference point: “No reparation can be exacted unless: (1) it is clearly due in accordance with accepted principles of international law; or (2) it is stipulated for in the understanding embodied in President Wilson's points regarding restoration of invaded territories and in the qualifications of these clauses embodied in the note of the Secretary of State of the United States of November 5, 1918” (Burnett 1940: 308).

Unable to reconcile the difference between seeking repayment of all costs the Allies had incurred and seeking only to restore invaded territories, the commission moved on. They attempted to fix a specific sum of money that would represent what Germany owed and a schedule in which this debt would be discharged over time. For weeks they negotiated these issues across subcommittees and working groups (Lamont 1921; Burnett 1940). Figure One summarizes the successive positions the major powers adopted.

The American team was plainly surprised by the demands. According to Wilson's closest advisor Edward “Colonel” House: “I thought the British were as crazy as the French, but they seem only half as crazy which still leaves them a good heavy margin of lunacy. Our people think that the maximum cannot be over twenty‐two billions of dollars (88 milliard[billion] marks) and are inclined to think it should be under that amount” (Seymour 1928: 343).

Confronting unified opposition from Britain and France, the Americans eventually proposed sums far greater than the maximal position their analysts had previously established as the upper limit. But even these numbers were insufficient.

As the British principles declared, their reference point was the financial, physical, and human costs they had incurred. That was the benchmark for all their proposals. The United States had entered the war late, incurring comparatively modest losses, but had also profited enormously from supplying the Allies (Clodfelter 1992). Clemenceau had initially vowed to use the negotiations to close the “most terrible of accounts opened between peoples” (Noble 1935). To Wilson, the U.S. mission was far more positive, it was “ ‘the infinite privilege of fulfilling her [America's] destiny and saving the world' ” (Widenor 1998: 550).

After American arguments that breaking the Armistice agreement was illegal failed to move the other parties to adopt their limited rights as a reference point, the American negotiators shifted course. They moved on to stressing German “capacity to pay” as the proper reference point. Though available information was sketchy, they estimated this capability was well below 100 billion gold marks (Keynes 1919; Burnett 1940). In drafting the treaty, they needed to set up ex ante incentives that would make German compliance self‐enforcing. But the United States pointed out that this would not happen if the treaty included impossibly large reparations demands.

“A definite obligation assumed by Germany under conditions which warrant us in believing that Germany herself has the will and believes she has the capacity to discharge such an obligation, will serve as an immediate basis of credit,” wrote Dulles to Wilson in June 1919. “A far larger amount assumed under equally satisfactory conditions eighteen months from now would not begin to have the same practical value” (Burnett 1940: 109).

The Americans now persuaded the French to lower their claims based on “the play of the percentages” (Lamont 1921; Burnett 1940). France, having sustained a disproportionate percentage of property damage, would ultimately be better served to demand a much smaller sum that reflected only purely civilian property damage caused by the German forces. Inserting this smaller number in the treaty would ultimately reap higher returns for France than an unrealistic claim for the total cost of the war that they would have to share much more equally with Great Britain.

With the French now moderating their claims, Lloyd George voiced opposition to provisions declaring a realistic fixed sum and payment schedule (Loucheur 1962; Trachtenberg 1980). Unsure how to deal with the sudden shift of the French, he vacillated while his experts engaged in the type of “pulling and hauling” that Allison (1971) has described as characteristic of bureaucratic politics. Keynes and Lord Edwin Samuel Montagu, who was at the time Britain's secretary of state for India, argued that Germany's limited capacity created a need for compromise. But British delegates Cunliffe and John Hamilton (Lord Sumner) resolutely demanded reparation in full. On March 10, Lloyd George, Wilson, and Clemenceau asked Montagu for Great Britain, Norman Davis for the United States, and Louis Loucheur for France to privately work out a deal. By designating Montagu, Lloyd George made it more likely that a compromise would be reached. They soon agreed on a fixed sum of 120 billion gold marks, which the Council of Four endorsed five days later. But on March 18, Lloyd George replaced Montagu with Cunliffe and the agreement fell apart (Loucheur 1962; Mantoux 1992; Lentin 2001). By March 25 they were completely deadlocked again.

Lloyd George tried to explain his actions to the other members of the Council of Four. Underscoring the role of domestic political pressure, he told them, “It will be as difficult for me as for M. Clemenceau to dispel the illusions which surround the subject of reparation. Four hundred members of the British Parliament have sworn to extract from Germany the very last penny to which we are entitled; I will have to face them” (Mantoux 1992: 19).

This argument frustrated the other delegates. Wilson wished for “a less slippery customer to deal with than L. G. for he is always temporizing” (Macmillan 2002: 41). Clemenceau concluded that “Lloyd George travels in every direction, so inconsistent is he from day to day” (Lentin 2001: 2).

In their analysis of integrative bargaining solutions, Dean Pruitt and Peter Carnevale (1993) defined a “bridging” solution as one in which a creative alternative is devised that satisfies the most important interests underlying the parties' original demands — neither party is forced to yield to accomplish its larger goals. Two American members of the CRD, Dulles and Davis, finally devised a compromise (Burnett 1940) that creatively bridged the interests the various Allies had been pursuing with what the Americans presumed German interests to be. But unfortunately no German had yet been consulted about that nation's interest. Ignorant of any of the CRD negotiations, their delegation would thus interpret the provisions as anything but a bridge in their direction.

The deal split the search for a fixed sum into pieces. Behavioral decision theory has demonstrated the vital importance of “reference points” in determining how people evaluate possible outcomes. These reference points, which function as the effective zero point for value, also help determine risk tolerance. Decision makers tend to be risk‐averse in evaluating outcomes above a reference point but risk‐seeking in evaluating those lower than the reference point (Kahneman 2003). What became Article 231 in the final treaty recognized the cost incurred by the Allies as a proper reference point for evaluating settlements. “The Allied and Associated Governments affirm and Germany accepts, the responsibility of Germany and her allies for causing all the loss and damage to which the Allied and Associated Governments and their nationals have been subjected as a consequence of the war imposed upon them by the aggression of Germany and her allies” (Allied and Associated Powers 1920). Creating a theoretical responsibility for full restitution, they left intact the illusion that accounts would eventually balance. German capacity to pay, which had been the American reference point for evaluating reparations outcomes, was addressed by the next article, “The Allied and Associated Governments recognize that the resources of Germany are not adequate, after taking into account permanent diminutions of such resources which will result from other provisions of the present Treaty, to make complete reparation for all such loss and damage” (Allied and Associated Powers 1920). Article 234 then established the interallied “Reparation Commission” described previously. The treaty stated that the commission would specify total debt by 1921.

According to William Keylor (1998: 500), “Davis and Dulles thought that they had devised a brilliant solution to the reparation dilemma . . . a ‘psychological sop’ to Allied public opinion as compensation for the loss of the huge German payments that Allied leaders knew could and would never be paid.”

Meeting with the Council of Four, French Finance Minister Louis‐Lucien Klotz stressed the need to cope with the participants' bounded rationality in setting up the commission's task: “The commission must be able to grant extensions. I don't think we can arrive at a satisfactory total estimate as early as 1921. It would require superhuman genius to know at that date what Germany will be able to pay during the next thirty years” (Mantoux 1992: 150).

Lloyd George did not often agree with Klotz, but he did so on this issue: “We will all be in a rather bad state for a number of years, and Germany will need ten years to get back on her feet. We can't really say what Germany's capacity to pay will be!” (Mantoux 1992: 151).

According to game theory, TOTREPS are self‐interested actors with unlimited computational capacity and unlimited understanding of the other side's options, information, and interests. These actors easily conceive all possible moves and their implications for personal well‐being and make similar unbiased calculations for all other relevant actors. The behavioral paradigm, on the other hand, presumes egocentrism, defined as “ignoring cognitions of others” (Tsay and Bazerman 2009). According to behavioral theories, the egocentric individual's attention focuses on his or her personal circumstances, preventing accurate appraisal of others' interests, beliefs, and alternatives. And, finally, the contractual paradigm assumes that a team of individuals working together within a hierarchy of expertise will collectively exercise unbiased judgment necessary to make truly important decisions.

The Council of Four had failed to anticipate Germany's response to the treaty's provisions. At the formal ceremony for receiving the first draft of the treaty, the Foreign Minister Ulrich von Brockdorff‐Rantzau obviously could not already know its contents. But fearing this could be the only opportunity he would ever have to speak publicly at the conference, he enraged the council members by forcefully rejecting what he believed would be their demand that Germany admit guilt for the war. They did not interrupt him to explain that they had made no such claim in the treaty. They did not explain the complicated negotiations that had culminated in Article 231. They did not explain that the Americans had introduced this term to limit German financial liability. So he went on to argue that only an impartial tribunal, not the Council of Four, could determine responsibility for the war.

Demanding that they stick to terms consistent with the Fourteen Points, von Brockdorff‐Rantzau condemned the Allies for starving noncombatants after the Armistice. According to him, maintaining a blockade on food shipments was a far greater crime than any actions committed during the war. His failure to rise while giving this speech further infuriated the council (Macmillan 2002). In private discussion afterward, Lloyd George remarked that he could not recall having ever been as angry, predicting that “your Brockdorff‐Rantzaus” would ruin Germany's chance for reconstruction. Wilson lamented that the Germans were truly a “stupid people” and that this speech would “turn the whole world against them.” Clemenceau pointed out to his two colleagues that they had finally learned what the French had known all along about Germany (Riddell 1986: 275).

Although the provocative content and tone of the foreign minister's speech had been decided upon only at the last minute, against the advice of his experts, it did reflect concerns that had emerged among the German delegation during their negotiation planning. That planning had been conducted in virtually complete ignorance of the actual discussions ongoing in Paris. Rumor and fear had led Brockdorff‐Rantzau to expect that the Allies would accuse Germany of causing the war. An expert panel, the Heidelberger Veirenigung, led by the imminent social scientist Max Weber (Luckau 1941) had therefore been drafted by the new German chancellor to study the issue. Reviewing events that culminated in the declarations of war, the panel found some clear‐cut German mistakes as well as certain events that the Germans should not be held culpable for, but mostly it found profound ambiguity. The panel advocated formation of the type of impartial investigative body that von Brockdorff‐Rantzau later demanded in his speech.

When the Germans finally studied the draft they prepared a number of written notes that they submitted to the Allies. These expressed alarm at the open‐ended aspects of the treaty, especially the vast powers given to the Reparations Commission, which appeared to enable the maximum possible exploitation of the German economy for Allied ends. But the official, written notes that the German delegation was permitted to pass to the Allies expressed even greater opposition to the other part of the Davis–Dulles compromise mechanism. Working with Max Weber and several other members of the Heidelberger Vereinegung, the delegation identified this article as the dreaded declaration of guilt for having caused the war that they had so feared in advance of the peace conference. Brockdorff‐Rantzau admitted the responsibility to pay reparations but insisted that “the German people should not be held responsible for the faults committed by their government” because “they did not will the war and would never have undertaken a war of aggression” (Luckau 1941: 242).

What had been an American pretext for acknowledging a high theoretical financial obligation that would be translated into a realistic lower payment to be determined later, had been interpreted in a far more sinister way by a delegation that was specifically searching for a “war guilt clause.” To the very end of the conference, the new German prime minister Gustav Bauer vainly attempted to have this excised, writing: “Germany further lays the greatest emphasis on the declaration that she cannot accept Article 231 of the Treaty of Peace which requires Germany to admit herself to be the sole and only author of the war, and does not cover this article by her signature” (Luckau 1941: 479).

Article 231 had been designed as part of a construction that would permit Britain and France to sustain for their public the illusion of debts repaid. It actually reduced the financial burden that Britain and France had sought to impose on Germany. Unfortunately because of the information vacuum the Allies had imposed on the German delegation, it read quite differently to them. They interpreted this language as the admission of guilt that von Brockdorff‐Rantzau (acting on the advice of Weber's group) had already publicly rejected. From the German perspective, because the treaty had taken the unexpected form of an ultimatum, the Allies were compelling the German people to declare their own guilt. It was not merely the Kaiser's responsibility but all of Germany's responsibility for having caused the war.

Based on the tenor of the official notes that the Allies permitted the Germans to write to them to request clarification of the treaty terms, some of the Allied delegates began to doubt whether the Germans would actually sign and began to consider their alternatives. On June 16, in a tardy planning session, the Council of Four summoned Ferdinand Foch, the Allied commander, to explain how they would execute their best alternative to a negotiated agreement (BATNA). “To my mind,” said Clemenceau, “there is no operation to have in mind than a direct march on Berlin . . . a vigorous and unremitting military blow that will force the signing” (Mantoux 1992: 462). But Foch explained that such an invasion was impossible. To avoid a “ruinous occupation which can take us to Berlin too anemic to deliver the decisive blow,” he advised instead only a limited incursion that would nonetheless leave the Germans more willing to compromise, at least in his opinion (Mantoux 1992: 466). In disbelief, the Council sought out opinions from other members of the “Supreme War Council” on June 20. British, American, and French generals all concurred that the farthest Allied troops would reasonably be able to go was Germany's Weser River, well short of Berlin (Mantoux 1992). Abandoning the idea of a congress with Germany in favor of a simple ultimatum was obviously a decision that reflected the Council's overconfidence in the strength of their BATNA.

At this time, a coalition of American experts along with Keynes and Loucheur attempted yet again to replace the two‐piece compromise commission language with a simple fixed sum. They based their argument on both moral and pragmatic grounds. But fatigue and the German reaction against him had greatly affected Wilson's perspective. His determination to teach Germany a lesson had crystallized, so he opposed much further modification (Mantoux 1992).

The reasons for Wilson's dramatic change are not easy to discern. The first of the official written notes of clarification about the treaty signed by Brockdorff‐Rantzau stated that “the basis of the peace of justice mutually agreed upon has been abandoned” (Luckau 1941: 225). Direct German accusations that the Allied behavior was a betrayal of Wilson's own Fourteen Points may have contributed to his much harder position. His precarious physical health may also have been a major factor; he was just then recovering from an illness in April that neurologists today believe may have been a small stroke (Weinstein 1970), a possible precursor to the massive stroke he was to suffer after returning to the United States. Whatever the cause, Wilson rejected late attempts “to soften the treaty,” explaining that these efforts made him “a little tired” and “very sick”“for people to come and say now that they are afraid that the Germans won't sign, and their fear is based upon things that they insisted upon at the time” (Baker 1923: 112).

The record indicates that the treaty was incomplete because the authors were not TOTREPS. They could not conceive or articulate every relevant contingency. They recognized these limitations, discussed what to do about them, and sought alternative forms. Some gaps in the provisions concerned minor matters such as the return of the remains of an obscure African leader to an unspecified party on an unspecified date. But others concerned the most important issues in the treaty, particularly German financial obligations. With flexible gap filling they sought to devise a workable reparations settlement. On these matters, the deliberative record and secondary accounts are consistent with the behavioral and contractual paradigms. The architects knowingly left gaps they tried to fill with flexible mechanisms.

Aspects of the treaty other than reparations were constructed the same way. The treaty created other commissions for particular gaps and the first twenty‐eight articles created a novel governance mechanism. This “covenant” established a General Assembly for a League of Nations run by an Executive Council where the Allied powers enjoyed a majority (Allied and Associated Powers 1920). After considerable debate about the wisdom of doing so, the covenant was included in the treaty itself because, as American banker Thomas Lamont wrote to Wilson, “It was utterly impossible to see far enough into the future to be able in the treaty to devise for each one of these complex situations a solution and a formula that would work. Tribunals, commissions, executive bodies had to be established under the authority of the nations themselves that were involved. How was this authority to be concentrated and invoked? The only possible method was through some medium of some world‐wide agency such as the League of Nations” (Lamont 1919).

But the problems went beyond incompleteness. According to Lamont, they included outright mistakes the league would be required to fix, including, he wrote, “. . . clauses of the treaty which some of us think are not unjust, but simply unworkable” (Lamont 1919).

En route to Paris in December 1918, Wilson had explained that logic to his advisors before the negotiations even started. The “idea of covenants, that is, agreements, pledges, etc. such as could be worked out in general form and agreed to and set in motion, and he particularly emphasized the importance of relying on experience to guide subsequent action” because “alteration would be easier to make in time as passion subsided and matters could be viewed in the light of justice” (Seymour 1928: 280–283). His demand to link the covenant to the treaty reflected Wilson's belief that the treaty must be a relational contract.

Advocates for the treaty (see Table One) were optimistic that the league would correct the mistakes, perfecting the terms over time. Critics of the treaty (see Table Two) believed contracting would fail because the treaty had breached the Armistice with unfeasible demands. Both positions were based on a conviction that the document had been constructed by imperfect, boundedly rational authors.

“The fundamental prediction of non‐cooperative game theory is that the outcome of a strategic interaction will be a Nash equilibrium” (Bendor and Hammond 1992: 157). The outcomes that result from a Nash equilibrium need not be efficient, but the concept provides qualitative constraints on strategies that differentiate it from contractual and behavioral theory. “A Nash equilibrium is an array of strategies, one for each player, such that no player has an incentive (in terms of improving his own payoff) to deviate from his part of the strategy array” (Kreps 1990: 28). How players actually identify an equilibrium strategy set is a bit of a mystery except “in cases where players can gather beforehand for pre‐play negotiation. They sit down together and try, via some procedure, to reason through how they should act. If they come to agreement, and if that agreement is credibly self‐enforcing in the sense that no one who believes that others will conform has the motivation to deviate, then the agreement will be a Nash equilibrium. The set of Nash equilibria contains the set of credibly self‐enforcing agreements that could be made” (Kreps 1990: 32).

The preplay negotiation before the peace conference, the lengthy discussions in Paris, and the exchanging of written notes with the Germans at Versailles provided the opportunity for establishing an equilibrium. The full congress originally envisioned by the parties might have been more effective in accomplishing this objective, but the Allied leadership chose to dispense with it.

The contractual paradigm assumes that decision makers use their scarce cognitive resources efficiently. When transactions are repeated and involve significant uncertainty and idiosyncratic assets, relational contracts economize by eliminating negotiation of possible events that may never come to pass. The contract language establishes efficient governance structures (Williamson 1999). Because parties anticipate that unexpected circumstances can generate ongoing disputes, they build in the flexibility to make adjustments when necessary — efficiency and stability should result.

Behavioral theory predicts inefficiency and instability in relations after the negotiation of a complicated agreement (Tsay and Bazerman 2009). Because the provisions in an agreement are arrived at by people with subjective points of view, disputes will inevitably arise (Robinson and Rousseau 1994). The behavioral paradigm predicts ongoing friction, instability, and renegotiation. Unilateral breaches of contract can be expected to provoke predictable reactions by the injured party (Robinson and Rousseau 1994), which include efforts to negotiate compensation for damages. When the damage is not repaired, the harmed party will experience powerful emotions that compel its members to seek retribution (Trivers 1971; Pillutla and Murnighan 1996).

By 1939 Europe was again at war, following the apparent failure of all the flexible mechanisms in the Treaty of Versailles. How did this happen? Were the critical decisions taken more consistent with the rational, the contractual, or the behavioral paradigms of decision making?

To facilitate analysis, I have divided the interwar period into segments, 1919–1932 and 1932–1939. These periods reflect the rise to power of Adolf Hitler and the Nazi Party in Germany and then his actions once he had taken full control of the German government. Figure Two charts relevant decisions during the second segment, which ended with Great Britain declaring war. These events include Britain's infamous policy of “appeasing” Nazi Germany that even today has been interpreted by many as a warning to American presidents that they should not hesitate to punish aggressors lest they repeat history (Ledeen 2009).

The policy decisions described in the second segment of this section of the article fit a distinct pattern. Unilateral German breach of a provision in the Treaty of Versailles was met with Allied acquiescence — neither France nor Great Britain attempted to enforce provisions of the treaty when they were publicly defied. Figure Two lists treaty articles relevant to each of the German moves. Appeasement was not merely Allied concession making in the face of aggression, but rather it comprised a series of decisions not to enforce the Treaty of Versailles following a unilateral German breach. Each move Hitler initiated was a clear, observable, intentionally verifiable treaty violation.

Obviously adherence to the treaty did not represent a Nash equilibrium outcome. The unilateral German breaches that the Allies failed to respond to improved the German position each time that Hitler moved. Until the Munich agreement of 1938 that renegotiated the borders of Czechoslovakia, these moves were entirely unilateral and unambiguously successful. At Munich, British Prime Minister Neville Chamberlain finally intervened to negotiate a resolution of the dispute. Though given what he ostensibly sought, Hitler was nonetheless furious that Chamberlain had stymied him from unilaterally seizing the Sudeten region of Czechoslovakia (Kershaw 2000). He quickly resumed the pattern of successful, unilateral action when he breached that agreement by taking control of the remaining Czech regions.

Up to this point, Hitler viewed his many accomplishments as fulfillment of goals he had publicly described in the early 1920s. Speaking to the Reichstag in the spring of 1939, he declared, “I have further attempted to tear up page for page that treaty, which contained in its 448 articles the most base violations ever accorded to nations and human beings. I have given back to the Reich the provinces stolen from us in 1919. I have led back into the homeland the millions of deeply unhappy Germans who had been torn away from us. I have recreated the thousand‐year historic unity of the German living space, and I have attempted to do all this without spilling blood and without inflicting on my people or on others the suffering of the war” (Kershaw 2000: 185).

The British and French possessed the military capability to enforce compliance with the treaty in this period. Had they elected to do so, it might have brought an end to Hitler's regime. The highest levels of the German military were vehemently opposed to Hitler's unilateral moves. A coup led by the head of the German General Staff and his immediate predecessor was aborted only by the stunning scope of the Munich victory (Churchill 1948; Kershaw 1998). Confronted by such astonishing success, internal opposition deteriorated and the coup plotters retreated. During this period, the very experienced Allied leaders made a series of critical misjudgments.

Hitler had recognized risks inherent in his strategy to remilitarize the Rhineland in 1936, stating later: “Had the French then marched into the Rhineland, we would have had to withdraw again with our tails between our legs. The military force at our disposal would not have permitted even for limited resistance. . . . It would have been the greatest political defeat for me” (Kershaw 1998: 588).

He gambled successfully that neither the French nor the British would be willing to enforce the treaty even though it had granted them highly favorable terms and that had been devised through preplay negotiation by their own elected representatives. This was a treaty ostensibly so one‐sidedly favorable to the Allies that it had been necessary to impose it by ultimatum. Their subsequent unwillingness to enforce its provisions therefore represents something of a puzzle.

The rational, empathic, and forward‐looking nature of noncooperative game theory is inconsistent with the actions of the various parties in the failed implementation of the terms of the Treaty of Versailles. Similarly, the rational choice paradigm fails to explain these events. The contractual paradigm can explain the form and the intentions of the drafters in creating the League of Nations, but it cannot explain the dramatic failure of these institutional solutions nor the instability that followed. Understanding these failures and resolving the puzzle of French and British nonenforcement requires an analysis of the first decade of the interwar period. Figure Three charts these events with a focus on reparations payments.

The architects of the treaty quickly confronted unanticipated events. The U.S. Constitution requires that the U.S. Senate ratify treaties with a two‐thirds vote. Between November 1919 and March 1920, senators opposed to Wilson urged the Senate to reject the treaty because of their opposition to the League of Nations. An uncompromising Wilson rejected their appeals to amend it, and American participation in the league, the Reparations Commission, and other such ancillary organizations ceased. The Europeans, “who had deferred so much to his (Woodrow Wilson's) opinions in all this business of peace‐making, were told without ceremony that we ought to be better informed about the American Constitution” (Churchill 1948: 11).

The absence of American members greatly altered the structure of the bargaining game in the reconstituted Reparation Commission. Davis and Dulles had reasoned that a combination of time and American leadership would overcome emotional obstacles to specifying a feasible sum for German debt, but without an American commissioner, the bargaining proceeded without this anticipated rational moderation (Trachtenberg 1980; Ferguson 1996).

Three milestones distinguish the contentious decade‐long negotiation process: the London Ultimatum of 1921, the Dawes Plan of 1924, and the Young Plan of 1930. The London schedule appeared to fix the total German debt at 135 billion goldmarks. But the true debt burden it imposed has actually been valued as low as 49 billion goldmarks by historians. The negotiators had yet again chosen to mask the obligations in a way that would perpetuate the cognitive illusion that Germany owed more than it actually did.

The German government responded with a genuine attempt to pay reparations on a significant scale (Trachtenberg 1980; Ferguson 1996) in a strategy of treaty revision through “fulfillment.” Actual payments made by the Germans in cash and kind eventually totaled roughly 10.134 billion goldmarks (Schuker 1988: 107). Fulfillment advocates acted on a belief that any sincere attempt to meet the demands would cause such serious economic problems that the Allies would be forced to abandon them. But they failed to fully anticipate the consequences their policies would have on internal German politics. The most prominent fulfillment advocate, Walther Rathenau, was accused of being a Jewish traitor to the German people and assassinated by paramilitaries.

Even with fulfillment, German payments failed to meet the 20 billion gold marks installment demand. Responding in a way that circumvented the treaty provisions, France and Belgium occupied the Ruhr region of Germany in early 1923 to enforce compliance. But the Germans stopped working in boycott, gravely damaging the national economy (Schuker 1988; Ferguson 1996). The value of the mark plummeted and prices exploded. By November 15, one U.S. dollar traded for 4.2 trillion marks. Those without assets in foreign currency saw their wealth wiped out overnight. The United States was compelled to intervene. A committee of experts led by American banker Charles Dawes (who soon after was elected U.S. vice president) proposed the “Dawes Plan” in 1924 to renegotiate the debt provisions, bring an end to the Ruhr occupation, suspend cash reparations payments, and provide a large loan to Germany so they could establish a new currency. The plan itself triggered a boom in private American lending to Germany that set up a curious cycle of payments. Money passed from the United States to Germany in loans, from Germany to France and Great Britain in in‐kind reparations payments, and then from France and Great Britain back to the United States in repayment of war loans (Trachtenberg 1980).

The cycle appeared to function well enough for a time, but how and when Germany would actually complete its obligations remained unspecified. The Dawes Plan was followed by the “Young Plan,” which attempted to answer these questions. Under the leadership of Owen Young, the respected chairman of General Electric Corporation, the Young Committee required Germany to finally resume cash payment of reparations. But the total value of these debts was reduced considerably to figures much closer to the American delegations' estimates back in 1919. Young's plan called for payments to finally end in 1988, obviously long after the deadline set by the original treaty. Unfortunately the plan came into effect in 1929 just as the German and American economies crashed into the Great Depression. The Weimar Republic, the German parliamentary republic that arose after the fall of the German Empire, was nearing an end as voting for the NSDAP, Hitler's National Socialist (Nazi) Party, grew from less than 3 percent in 1924 to 37 percent in 1932. The surge in support for the party came from alienated segments of the bourgeoisie (O'Loughlin 2002; King et al. 2008), a group who had lost all their wealth in the hyperinflation.

If the purpose of the peace conference had been securing repayment for “the most terrible of accounts between peoples,” as Clemenceau had described it (Noble 1935), then the wrong party paid. After reviewing records of all the transfers up to 1931, Stephen Schuker (1988) calculated the total value of German reparation payments under the London, Dawes, and Young regimes at 22.891 billion gold marks. Not only was this total less than the Allies demanded, it was far less than the flow of American funds into Weimar Germany. These resources constituted “one of the greatest proportional transfers of wealth in modern history” but in a direction opposite that called for by the treaty (Schuker 1988: 120). Once Hitler defaulted, these transfers became de facto “reverse reparations” from America to Germany. These resources would enable the military rearmament campaign that Germany soon embarked upon.

Opposition to the Versailles Treaty became integral to the identity and program of the Nazi Party. The twenty‐five‐point “permanent program of the party” announced by Hitler at a meeting on February 24, 1920 illustrates this. The first three principles directly attacked the treaty. For example, principle two demanded “equality of rights for the German people in its dealings with other nations, and the revocation of the peace treaties of Versailles and Saint‐Germain” (Feder 1934).

Hitler's passionate denunciations of the provisions and the process (Baynes 1969) proved to be an effective recruiting tool. “Simple slogans, kindling the fires of anger, resentment, and hatred were what worked” (Kershaw 1998: 137). Addressing more than thirty mass meetings in 1920 and many more in 1921, he succeeded in greatly swelling the party ranks (Kershaw 1998). The speeches returned again and again to the same subject (see Table Three). The German people must have their revenge against “the November criminals” who had undermined the empire from within, betrayed the soldiers on the front line, “stabbed them in the back,” and sold the nation into bondage through the dishonorable “Versailles Diktat.”

The “criminals” that Hitler alluded to in his July 1922 speech included center‐left politicians and the prominent Jewish bankers Max Warburg and Carl Melchior who had served in the German delegation at Versailles. During those negotiations, Melchior had attempted to work around the ultimatum by devising a proposal for Germany to pay 100 billion goldmarks (Luckau 1941).³ By passing this offer in a note to his fellow banker Thomas Lamont, he had hoped the offer might entice the Council of Four to undertake face‐to‐face negotiations to work out the details. Lamont knew the two somewhat professionally before the conference. But when this rejected offer became public in Germany, it was cited as proof that the two bankers and their associates were guilty of the “stab in the back.” Propagandists labeled the treaty “the Warburg Jewish Peace” (Chernow 1993:217). The German interpretation of the reparations deal stigmatized all those connected to the treaty.

The greatest stigma came from interpretation of Article 231 as “the war guilt clause.” This label became so pervasive that long after the treaty unraveled, Lamont was still trying to explain the error to his colleagues at J. P. Morgan. Referring to former U.S. President Herbert Hoover, a fellow member of the American delegation at the peace conference, Lamont (1942) wrote, “If Mr Hoover had had anything to do with drawing up that part of the treaty with Germany, he would have known that there was no war guilt clause. There was a clause so drafted as to hold Germany responsible in a financial sense for all the damage that she had caused. When this was submitted to the Germans they at once came back, I always thought with malice aforethought, and demanded to know if this meant that Germany was solely responsible for causing the war. That put Clemenceau and Lloyd George in the hole. They could not from a political point of view say ‘Oh no, we didn't mean that.’ And although they never intended it that way they just let the matter drop. From that time on the Germans made that terrific hullabaloo about a war guilt clause and that was one of the very good bases for Hitler's start in stirring up his people.”

Lamont apparently believed the German delegation had opportunistically created the war guilt label at Versailles in order to undermine support for the treaty. But their records provide no evidence of such guile (Luckau 1941). If Herbert Hoover, an American delegate at the peace conference, had described the clause as a German declaration of their guilt for initiating the war, then clearly no such malice was necessary.

The unexpected consequences of their decisions greatly affected the Allied policy makers and their successors. It complicated the task of formulating unbiased judgments of Hitler and his objectives. They could not appraise the new German government without reflecting upon the circumstances that many believed had brought the Weimar regime to an end.

Neville Chamberlain's errors have been widely cited. After negotiating with Hitler at Munich, he pronounced, “I believe it is peace for our time” and privately wrote to his sister, that “I had established a certain confidence which was my aim . . . in spite of the hardness and ruthlessness I thought I saw in his (Hitler's) face, I got the impression that here was a man who would be relied upon when he had given his word” (Self 2006: 314).

But Chamberlain was not alone in his delusional thinking: “All those Englishmen who visited the German Fuehrer in those years were embarrassed or compromised. No one was more completely misled than Mr Lloyd George whose rapturous accounts of his conversation make odd reading today” (Churchill 1948: 224–225). Lloyd George's error is notable in several respects. In Paris he had been the ultimate obstacle to specifying a fixed sum for reparations in the treaty. His opposition and vacillations had convinced Davis and Dulles to write the compromise coupling open‐ended financial claims to “the war guilt clause.” In the late 1930s, the former prime minister was still a major force in British politics (Churchill 1948; Lentin 2001). His judgment of Hitler certainly affected many others. It would likely have influenced Chamberlain.

His mistakes were also out of character for a man with a vaunted reputation for skillfully turning perspective taking to political advantage. Some of Keynes' (1919: 25) most memorable passages described how “the Wizard,” the “goat‐footed bard” outmaneuvered Woodrow Wilson through “unerring, almost medium‐like sensibility to everyone immediately around him,” his “telepathic instinct” for the “appeal best suited to the vanity, weakness, or self‐interest of his immediate auditor.” But meeting with Hitler, two years before Munich, Lloyd George's skills betrayed him (Lentin 2001). In a glowing editorial that appeared in a major London paper, he praised Hitler as “a born leader, a magnetic dynamic personality with a single‐minded purpose, a resolute will and a fearless heart. . . . Hitler is the George Washington of Germany, the man who made his country independent of all its oppressors” (Lloyd George 1936/1969).

Many historians have explained this gross misjudgment as reflective of bias caused by the events that brought Hitler to power. Churchill (1948), Lewis Namier (1942), and Antony Lentin (2001) have characterized this as a popular delusion in Britain, stemming from having imposed a treaty that produced such widespread economic misery. That the moves in the second segment consisted of Hitler unilaterally reversing what they all already believed to have been serious mistakes, could have made it difficult for them to clearly see larger motives. For Hitler, the treaty was “the blackmailer's lucky find — not the real treaty, but the legends built up around it. He has put them to the fullest and foulest use” (Namier 1942: 57).

In this article, I consider the Treaty of Versailles in light of three decision‐making paradigms in social science that each conceive rationality differently. I argue that only the more limited sense of rationality embodied in the behavioral paradigm explains the implementation problems and systematic unraveling of the treaty during the 1930s.

The gap left in the reparations section stemmed from fixation on recovering the sunk costs of the war. Even when it became obvious that nothing like these costs could ever be meaningfully recovered, the Allies sought to sustain the illusion of reparation in full. Years later Dulles provided a behavioral explanation for the Allied negotiator's “stupidity.”“The participants suffered from ‘blind spots,’ ” he wrote. “Of these the most important is that illustrated by the ‘war guilt’ provision (Article 231). In the light of subsequent developments it may be that this article was the most important single article in the treaty. Thereby, in German eyes, Germany was branded with moral guilt for the World War” (Burnette 1940: xi–xii).

Explaining why he made such a grievous mistake, Dulles pointed to failures of perspective taking. He and his colleagues had simply not anticipated how Germans would interpret Article 231. “It came therefore as a surprise when German observations on the Conditions of Peace showed that this section could plausibly be, and in fact was, considered to be a historical judgment of ‘war guilt.’ It can thus be said that the profound significance of this article of the Treaty came about through accident, rather than design” (Burnett 1940: xi–xii).

TOTREPS would have written a complete contract, structuring incentives so that adherence would constitute a self‐enforcing equilibrium. The negotiators in Paris could not, so they were forced into relational contracting. The contractual man of the new institutional economics would have organized the services of the very best available experts to extend the limitations on rationality, to debias judgments, and craft a highly efficient relational contract. The world leaders gathered in Paris, with the best available resources to recruit and organize, attempted to do just that, but failed. The institutions they created were unstable and maladaptive. By restricting communications with the Germans, they eliminated any opportunity to learn from each other about how the clauses they drafted could and should be interpreted. They could not adjust the drafting to avoid serious errors of interpretation such as the war guilt clause.

The conclusions that can be drawn from case methods of research are limited by the potential selection and filtering of events (Druckman 2003). The close fit between the behavioral paradigm and the negotiating failures in this case reflects more than selection. Nothing like behavioral science existed prior to 1919. Very little in the way of empirical social science and empirical psychological research had been undertaken prior to 1919 (Ruml 1922; Frank 1923). Many of the same individuals who established the institutions needed to support this “new social science” (Lippmann 1922) during the 1920s and 1930s were veterans of the peace negotiations.

Herbert Simon, perhaps the most prolific and best known behavioral scientist (Crowther‐Heyck 2006), actually coined the term “bounded rationality” (1947). He later (Simon 1985) recalled Walter Lippmann's book Public Opinion as having been “the harbinger” of the new science and “still wholly fresh” when he himself began his doctoral studies in 1933. During Woodrow Wilson's administration, Lippmann had been the Assistant Secretary of War charged with the task of organizing “the Inquiry,” a bureau of the best available social science experts whose mission was to analyze the negotiation problem that Wilson would confront at Versailles (Gelfand 1963; Steel 1999).

As a respected writer, Lippmann was also a natural choice to turn the product of the Inquiry's research into the Fourteen Points speech that Wilson used to articulate American war aims. The idealistic speech he constructed became the basis upon which the Germans asked for an armistice. It was the basis for the widely misplaced optimism in Germany that a Wilsonian peace would be negotiated on open, fair, and just rather than punitive terms (Boemeke, Feldman, and Glaser 1998). Lippmann wrote Public Opinion (Steel 1999; Bottom 2009; Newman 2009) as a way of describing why the peace negotiators had failed so completely in their attempt to realize Wilson's vision and what steps were needed to correct their errors before it was too late. In his view the limited information‐ processing capability of the individual, whether private citizen or experienced statesman, biased the negotiators and the general public toward systematic delusions.

His analysis broke significant new ground in epistemology (Dewey 1922) by introducing the concept of a stereotype as the means by which we store knowledge (see Mervis and Rosch 1981; Schneider 2004; Newman 2009). These “pictures in the head” are the building blocks for mental models we all construct to solve problems and reach decisions. But the inherent “blind spots” in these simple models become the source for delusional thinking that, unchecked, can result in serious mistakes in decision making. Although Simon (1947) would advance much further from this starting point, knowledgeable reviewers (Dewey 1922; Wright 1922) saw Lippmann's analysis as a devastating attack on the “false simplicities” of the conventional economic model of rational choice.

The reaction of the Allied leadership to Brockdorff‐Rantzau's speech rejecting war guilt provided Lippmann with his first example of a “stereotype.” Immediately after the ceremony at Versailles, Lloyd George concluded “the old German is still there.” Wilson concluded that “the Germans really are a stupid people, they always do the wrong thing. They don't understand human nature.” Clemenceau pointed out to his colleagues that at last they could now appreciate what the French had long known about the German character (Riddell 1986: 275). To explain how readily Clemenceau generalized when “a junker blustered” but how he dismissed it as inauthentic when “a labor leader (i.e., Johannes Giesberts) declared the guilt of the empire,”Lippmann (1922: 55) explained that the French leader “saw the type . . . and took to heart those reports and, it seems, those only which fitted the type that was in his mind.” In other words, the operation of stereotypes had generated what we now call “confirmation bias.”

But Lippmann did more than describe events, he developed a general theory of judgment and decision (Schneider 2004; Newman 2009). Both officials and the public rely on stereotypes to construct mental models of a situation, developing “blind spots” that pose severe problems for democratic governments.

Lippman advocated interdisciplinary social science as the foundation of what he called “organized intelligence,” which could help leaders correct the existing policy mistakes in the treaty and avoid future blunders of such magnitude. Descriptive studies of bias and human error would provide a means of training experts and developing programs for improving decision making.

Explaining in detail how Lippmann's ideas were actually translated into institutional reality in the United States is well beyond the scope of this article. (The lower half of Figure Three tracks the early stages of this process.) The institutions involved include networks of social scientists dedicated to the behavioral approach, most notably the Social Science Research Council (O'Connor 1999; Crowther‐Heyck 2006; Heaney and Hansen 2006; Bottom 2009); government agencies structured to utilize social science expertise, such as the National Resource Planning Board, the Central Intelligence Agency, and the Council of Economic Advisors (Merriam 1944; Price 2006; Beloff 2007); and a network of intermediary connective “think tanks” and nonprofit foundations that formulates and propagates specific policy ideas, such as the Council on Foreign Relations, the Brookings Institution, and the Rand Corporation (Dockrill 1980; Parmar 1992; Grose 1996; Beloff 2007; Abella 2008). Those attempting to organize diplomatic initiatives today do not confront the severe shortage of experts that Assistant Secretary of War Lippmann described as “a real famine” in 1918. Then his only recourse had been to “train our own experts” (Steel 1999).

In this article, I have contrasted behavioral, contractual, and rational choice explanations for the negotiation process that created the Treaty of Versailles. Behavioral and contractual theory can both explain the treaty's incomplete form, but only behavioral theory can explain how relational contracting systematically failed, culminating in the unilateral unraveling of the treaty. Relational contracting can only succeed if communication is open enough to permit expectations to align. Cognitive misperceptions, the public's desire for retribution, severe stress, and intense time pressure led the Council of Four to impose an ultimatum without allowing any significant communication with the Germans.

Restricted to swapping notes, the delegations had little chance of reaching genuine understanding. Conflicting interpretations, particularly of the American Article 231, could not be reconciled. The psychological concession that Davis and Dulles gave to Great Britain and France became the reviled “war guilt clause” that Hitler condemned as “a monstrous lie” (Baynes 1969). Dulles himself later admitted his own “blindness and stupidity” (Burnett 1940). By imposing contract provisions that the Allies would later lack the will to enforce, they provided “the blackmailer's lucky find” (Namier 1942) that Hitler so fully exploited. The Allies might have more accurately assessed his intentions had they not so regretted the very provisions he was violating. Lloyd George, whose opposition to conceding on reparations led to the inclusion of Article 231 in the treaty, was among the most biased.

Behavioral theories explain these events so well because they were developed by individuals who participated in negotiating these flawed treaties. The basic elements of the paradigm were created to describe, explain, and correct those very events. The theoretical constructs of stereotypes, blind spots, and mental models sprang from the attempt to understand why relatively well intentioned and relatively rational negotiators could make such serious errors.

Those trained through one of these three paradigms have tended to treat the others as competitors, but they are actually best understood as necessary complements to each other. Behavioral science seeks to reveal “those laws of human nature which can produce social inequality, industrial strife, marital disharmony, juvenile delinquency, mental illness, war, and other widespread miseries” (Alexander 1956: 1–2). Rational choice theory, which emerged in its rigorous axiomatic form after World War II, established mathematics for analyzing strategic situations confronting a TOTREP.⁴ The concepts of complete contract and Nash equilibrium have proven to be invaluable in understanding better how negotiators actually deviate from rationality. The most rapid advances in the behavioral paradigm have been facilitated by psychologists employing the TOTREP concept as a benchmark (Kahneman 2003).

The contractual paradigm predicts efficient organizational solutions based on the delegation of functions to experts in an administrative hierarchy. Recognizing their limitations, the leaders did employ the most “expert probabilists” of that era. Lloyd George solicited advice from Keynes, who soon thereafter published A Treatise on Probability (Keynes 1921). His innovative approach greatly influenced rational choice (Gardenfors and Sahlin 1988) and behavioral decision theories (Camerer and Weber 1992). But breakthroughs in probability and organization theory (Simon 1947) had not yet been made so the experts could not draw upon those insights. The availability of sufficient expertise to extend bounds on rationality for crucial decisions was the very purpose of Lippmann's “organized intelligence.” The contractual paradigm then literally describes an idealized version of the prescriptions that he derived from his behavioral theory. The new institutional economics, as described by Williamson (1985, 1999), describe how organized intelligence would ideally work.

The institutions the early behavioralists established led to significant advances in social science and policy analysis from a very primitive state (Frank 1923). President John F. Kennedy was effectively guided by information and forecasts generated through this system during the missile crisis (Allison 1971). He relied heavily on expert advisers (e.g., Robert MacNamara, MacGeorge Bundy, Dean Rusk, Carl Kaysen, and Walt Rostow) whose approach had been shaped in part by training in the social sciences and active participation in these institutions. Allison's rational actor analysis of the missile crisis revealed this elaborate system operating precisely as it was intended to function. His analysis of organizational behavior and bureaucratic politics during the crisis exposed the real limitations of organized intelligence as it functioned during the middle of the last century.

But have the human failings that led the Paris negotiators astray in 1919 been solved through modern science and organization? Certainly human nature has not fundamentally changed (Kahneman 2003). But complex organizational solutions have been devised to cope with these problems (Williamson 1985; Milgrom and Roberts 1992; Klayman, Heath, and Larrick 1998). Should we now expect the “contractual man” model of new institutional theory to explain more accurately how international diplomacy unfolds? Based on his experiences with the British delegation, Harold Nicolson (1933: 8) warned that the temptation for such overconfident thinking would be inevitable: “Future generations will be equally convinced that they will be immune from the defects which assailed the negotiators of Paris. Yet they in turn will be exposed to similar microbes of infection, to the eternal inadequacy of human intelligence.”

The efficacy of modern organized intelligence cannot be appraised by examining negotiations that predate the development of large‐scale social science, think tanks, and an expansive federal bureaucracy. When primary sources of information about the deliberations that produced the recent decision to invade Iraq and the postwar reconstruction program are disclosed, then scholars may be able to draw firmer conclusions about biases and mistakes in contemporary policy making. The information that has emerged to date (Fallows 2006; Ricks 2006; Chandresekharan 2007) provides quite a bit of evidence that the system for organized intelligence generated policy guidance that failed to correct systematic cognitive biases. Key decisions appear to have been distorted by simplistic stereotypes, delusional thinking, and sizable blind spots. Ironically, a misguided attempt to avoid repeating Neville Chamberlain's mistakes (see Figure Two) may have been responsible for some of the misjudgments made by American policy makers.

Nearly a half century after he first described how stereotypes and blind spots work in Public Opinion, Walter Lippmann had become the consummate Washington insider. He then (Lippmann 1966) described how “the frightful consequences of appeasing Hitler left behind them in the minds of men the great Munich stereotype.” In vain, he attempted to warn President Lyndon Johnson that “Mr Dean Rusk (then secretary of state) is the current user of the Munich stereotype possessed with the notion that in the divided territory of Vietnam there is being reenacted once again the drama of Munich.” Unfortunately, Lippman's warning against the planned troop escalation and stepped‐up bombing program there greatly angered the president but did little to alter his decision making.

Of course, many of those who saw “the drama of Munich” being reenacted in Baghdad appear firmly convinced that it is happening yet again in Teheran. Based on their mental model, the drama will once again require decisive American military response to avoid a tragedy (Fallows 2006; Ledeen 2009). The power and influence of the Munich stereotype suggests that even the most advanced current forms of organized intelligence have yet to solve the negotiating problems posed by bounded rationality. The idealized hierarchical solutions in new institutional theory remain but an aspiration. The real operation of these institutions still represents an imperfect solution to “the microbes of infection” that created such “blindness and stupidity” in Paris in 1919.

(From a speech delivered to the United States Congress on January 8, 1918)

1. Diplomacy shall proceed always frankly and in the public view.

2. Absolute freedom of navigation upon the seas outside territorial waters.

3. An equality of trade conditions among all nations consenting to the peace.

4. Armaments will be reduced to the lowest point consistent with domestic safety.

5. Absolutely impartial adjustment of all colonial claims.

6. Evacuation of all Russian territory and settlement of all questions affecting Russia.

7. Belgium must be evacuated and restored.

8. All French territory should be freed and the invaded portions restored.

9. The frontiers of Italy should be effected along clearly recognized lines of nationality.

10. The peoples of Austria‐Hungary accorded opportunity of autonomous development.

11. The relations of the Balkan States to one another determined by friendly counsel.

12. Turkish portions of Ottoman Empire assured a secure sovereignty. Other nationalities assured an opportunity of autonomous development.

13. An independent Polish state that should include the territories inhabited by indisputably Polish populations assured a free and secure access to the sea.

14. A general association of nations under specific covenants affording guarantees of political independence and territorial integrity.

Source: Gelfand (1963).