For their comments and suggestions, I thank Ed Mansfield, David Spiro, Charles Kupchan, Jack Snyder, and other participants in the Seminar on International Political Economy at Columbia University. I am also grateful to Douglas Baird, Anne-Marie Burley, Dale Copeland, Scott Leuning, Duncan Snidal, Stephen Walt, and other colleagues in the Program on International Politics, Economics, and Security (PIPES) at the University of Chicago.

1. See Funabashi, Yoichi, Managing the Dollar: From the Plaza to the Louvre (Washington, D.C.: Institute for International Economics, 1988Google Scholar); and Kenen, Peter B., Managing Exchange Rates (London: Routledge, 1988Google Scholar). Kenen reproduces key portions of the Plaza Communiqué (22 September 1985) and the Louvre Accord (22 February 1987) on p. 50.

2. See Williamson, John, The Failure of World Monetary Reform, 1971–1974 (New York: New York University Press, 1977Google Scholar); and Dam, Kenneth W., The Rules of the Game: Reform and Evolution in the International Monetary System (Chicago: University of Chicago Press, 1982Google Scholar). For a counterargument focusing on U.S. domestic politics rather than on the breakdown of international commitments, see Gowa, Joanne, Closing the Gold Window: Domestic Politics and the End of Bretton Woods (Ithaca, N.Y.: Cornell University Press, 1983)Google Scholar.

3. See Keal, Paul, Unspoken Rules and Superpower Dominance (London: Macmillan, 1983CrossRefGoogle Scholar). Some diplomatic efforts were made to articulate the rules, but they did little in themselves to clarify expectations. In 1972, as the strategic arms limitation talks (SALT I) were concluded, Nixon and Brezhnev signed the Basic Principles Agreement. It sought to specify some key elements of the superpowers' relationship and thereby facilitate the development of detente. The product was vague and ambiguous. Worse, it seemed to indicate—wrongly—U.S. agreement with the Soviet position on peaceful coexistence and competition in other regions. George, Alexander calls these elements “a pseudoagreement.” For the text of the agreement, see Department of State Bulletin, 26 06 1972, pp. 898–99Google Scholar. For an analysis, see George, Alexander, “The Basic Principles Agreement of 1972,” in George, Alexander L., ed., Managing U.S.-Soviet Rivalry: Problems of Crisis Prevention (Boulder, Colo.: Westview Press, 1983), pp. 107–18Google Scholar.

4. The State Department announced its position in a brief public statement on 4 March 1981. The administration continued to debate its arms control policy, and Reagan continued to criticize Carter's SALT II agreement. In early May 1982, he told a press conference that the agreement “simply legitimizes an arms race” and added that “now the parts [of the agreement] that we're observing … have to do with the monitoring of each other's weaponry, and so both sides are doing that.” In late May 1982, on the eve of the strategic arms reduction talks (START), Reagan finally stated that the United States “would not undercut” the SALT II agreement. He continued to criticize it, however, and left uncertain which portions of the agreement the United States would observe. See the Department of State announcement of 4 March 1981, cited by Talbott, Strobe in Deadly Gambits: The Reagan Administration and the Stalemate in Nuclear Arms Control (New York: Vintage Books, 1985), p. 225Google Scholar; The President's press conference statement of 13 May 1982, quoted in Weekly Compilation of Presidential Documents, vol. 18, no. 19, 17 05 1982, p. 635Google Scholar; and the President's “Remarks at Memorial Day Ceremonies at Arlington National Cemetery,” delivered on 31 May 1982 and included in Public Papers of the President of the United States: Ronald Reagan, 1982, book 1 (Washington, D.C.: Government Printing Office, 1983), p. 709Google Scholar.

5. In 1984, in a confidential report to Congress, President Reagan cited in detail Soviet noncompliance with numerous arms control agreements. Reagan's accompanying message stated that “violations and probable violations have occurred with respect to a number of Soviet legal obligations and political commitments in the arms control field.” SALT II violations were included, and the reference to “political commitments” alludes to them. These criticisms were expanded in another report, issued in 1985. The Soviets rejected these charges and made counterclaims regarding U.S. violations. Relevant documents are cited by Calvo-Goller, Notburga K. and Calvo, Michael A. in The SALT Agreements: Content-Application-Verification (Dordrecht, Netherlands: Martinus Nijhoff, 1987), pp. 318 and 326 ffGoogle Scholar.

6. President Reagan did restate the U.S. commitment not to undercut SALT II in June 1985, some six months before the unratified treaty would have expired. U.S. policy, however, was always contingent on reciprocal Soviet adherence. On that point, Reagan was sharply critical: “The United States has not taken any actions which would undercut existing arms control agreements. The United States has fully kept its part of the bargain; however, the Soviets have not.… Certain Soviet violations are, by their very nature, irreversible. Such is the case with respect to the Soviet Union's flight-testing and steps toward deployment of the SS-X-25 missile, a second new type of 1CBM [intercontinental ballistic missile] prohibited by the unratified SALT II agreement. Since the noncompliance associated with the development of this missile cannot be corrected by the Soviet Union, the United States reserves the right to respond in a proportionate manner at the appropriate time.” See the President's statement of 10 June 1985, quoted in Weekly Compilation of Presidential Documents, vol. 21, no. 24, 17 06 1985, pp. 770–71Google Scholar.

7. It is worth noting that all of these distinctions are ignored in international law. Virtually all international commitments, whether oral or written, whether made by the head of state or a lower-level bureaucracy, are treated as “binding international commitments.” What is missing is not only the political dimension of these agreements, including their status as domestic policy, but also any insight into why states choose more or less formal means for their international agreements.

8. Tacit and oral agreements, by their very nature, do not specify promises in great detail and rarely spell out contingencies or remedies. Consider, for example, the informal cooperation between friendly intelligence agencies such as the U.S. Central Intelligence Agency and Israel's Mossad. Besides exchanging information, both sides engage in unacknowledged spying on each other. But what are the limits? What violates the informal agreement, and what differentiates serious violations from “normal cheating”? To clarify these issues and to encourage regular cooperation, the United States and Israel have signed informal accords, beginning with a secret agreement in 1951. Even so, such agreements are necessarily incomplete, sometimes making it difficult to differentiate cheating from permissible activity. According to Blitzer, “U.S. law enforcement officials … long suspected that Israel was playing fast and loose with the long-standing U.S.-Israeli understanding barring covert operations against each other. Yes, there is always some spying going on, even among very close friends and allies. But that is a far cry from actually planting a ‘mole’ in a friendly country's intelligence community. Thus, there is a huge difference between unobtrusive intelligence-gathering operations, on the one hand, and the actual running of paid spies in each other's country, on the other.” Over time, actors can use sanctions and exhortation to specify these contingent obligations and to signal the limits of their tolerance. But ambiguities will surely remain. See Blitzer, Wolf, Territory of Lies: The Exclusive Story of Jonathan Jay Pollard—The American Who Spied on His Country for Israel and How He Was Betrayed (New York: Harper & Row, 1989), p. 163Google Scholar; and Raviv, Dan and Melman, Yossi, Every Spy a Prince: The Complete History of Israel's Intelligence Community (Boston: Houghton Mifflin, 1990), pp. 77 ffGoogle Scholar.

9. The State Department's 1981 statement, for example, that the United States “would not undercut” the unratified SALT II treaty if the Soviets reciprocated is an informal commitment. To international lawyers, its status is clear-cut. The State Department has unambiguously committed the United States by using the standard diplomatic language of obligation to a treaty pending ratification. But what about the domestic political status of that promise? The debate within the Reagan administration raged for another year before the President publicly ratified the State Department position. Even then, the Congress and courts need not be bound by these executive branch statements.

10. Recognizing these limitations on oral bargains, domestic courts refuse to recognize such bargains in many cases, thereby creating a powerful incentive for written contracts. There is no such incentive to avoid oral bargains in interstate agreements.

11. According to Downs and Rocke, “A state bargains tacitly with another state when it attempts to manipulate the latter's policy choices through its behavior rather than by relying on formal or informal diplomatic exchange.” Actions, not diplomatic words, are the crucial form of communications, and their aim is joint, voluntary cooperation rather than outright coercion. Downs and Rocke's contribution is to show how imperfect information affects states' strategic choices and may produce inadvertent arms races. Their focus is on uncertain estimates of others' strategies, preferences, and specific actions (either completed or intended), and not on the ambiguous meaning of tacit agreements and other informal bargains. See the following works of Downs, George W. and Rocke, David M.: “Tacit Bargaining and Arms Control,” World Politics 39 (04 1987), p. 297CrossRefGoogle Scholar; and Tacit Bargaining, Arms Races, and Arms Control (Ann Arbor: University of Michigan Press, 1990), p. 3Google Scholar.

12. See Elster's, Jon discussion of “the two problems of social order,” in The Cement of Society: A Study of Social Order (Cambridge: Cambridge University Press, 1989CrossRefGoogle Scholar), chap. 1. Elster's key distinction is between regular behavior patterns and cooperation. He distinguishes five main varieties of cooperation: helping others, voluntarily bearing costs of externalities, physical collaboration in joint ventures, mutual agreements to transfer rights (private orderings), and conventional equilibria (in which no party can improve its outcome by unilaterally deviating). In this article, my discussion of international cooperation focuses only on reciprocal contractual exchanges, which involve future performance and where the possibility of profitable defection might arise.

13. In tacit cooperation, one party in effect takes a chance in the expectation that another will simultaneously take an equivalent chance, leaving both better off. Neither party takes such chances when it maximizes unilaterally and independently. Stable expectations can arise in either case, based upon stable Nash equilibria. It is important not to exaggerate the scale of international cooperation by calling all shared expectations “cooperation.” They may be nothing more than unilateral maximizing.

14. Aust, Anthony, “The Theory and Practice of Informal International Instruments,” International and Comparative Law Quarterly 35 (10 1986), p. 791CrossRefGoogle Scholar.

15. See, for example, Schachter, Oscar, “The Twilight Existence of Nonbinding International Agreements,” American Journal of International Law 71 (04 1977), pp. 296–304CrossRefGoogle Scholar; Glennon, Michael J., “The Senate Role in Treaty Ratification,” American Journal of International Law 77 (04 1983), pp. 257–80CrossRefGoogle Scholar; and Mülnch, Fritz, “Non-Binding Agreements,” in The Encyclopedia of Public International Law, vol. 7 (Amsterdam: North-Holland, 1984), pp. 353–57Google Scholar. The one general (and quite valuable) legal treatment of informal agreements is “The Theory and Practice of Informal International Instruments” by Anthony Aust, a practitioner in the British Foreign Office.

16. The Vienna Convention on the Law of Treaties was opened for signature on 23 May 1969 and entered into force on 27 January 1980, after ratification by thirty-five nations. See UN document A/CONF. 39/27,1969.

17. See, for example, McNair, Lord, The Law of Treaties (Oxford: Clarendon Press, 1961Google Scholar); and Elias, Taslim, The Modern Law of Treaties (Dobbs Ferry, N.Y.: Oceana Publications, 1974)Google Scholar.

18. For a system of contract law to be effective, the parties cannot simply abandon their commitments unilaterally. Or, rather, they cannot abandon these commitments without facing legal penalties. Reflecting this understanding, the key disputes in contract law revolve around what constitutes a binding agreement and what constitutes an appropriate penalty for nonperformance. International legal scholarship largely avoids these fundamental issues, and it says all too little about related issues of renunciation, violation, and monitoring of agreements.

19. Incomplete domestic agreements can be filled in by court decisions. Incomplete international agreements remain incomplete. They are beyond the reach of international court decisions, much less enforcement. For an astute discussion of the weakness of treaties that contemplate further, detailed negotiations, see Baxter, Richard, “International Law in ‘Her Infinite Variety,’” International and Comparative Law Quarterly 29 (10 1980), p. 552CrossRefGoogle Scholar. In these general treaties, Baxter says, the individual provisions “are pacta de contrahendo, which cannot be enforced if the parties do not reach agreement. There is no way in which an agreement can be forced upon them and there is likewise no way in which they can be compelled to negotiate. The assertion that the duty to negotiate or to conclude an agreement implies a duty to negotiate in good faith is an empty one… In the relations of States, a complaint that negotiations have not been carried on in good faith is mere rhetoric.”

20. For an account of the purchase and the litigation, see Petzinger, Thomas Jr, Oil and Honor: The Texaco-Pennzoil Wars (New York: Putnam, 1987)Google Scholar.

21. In the dispute over Getty Oil, compliance issues were complicated by the possible size of any judgment against Texaco. Courts often require bonds to be posted covering any final awards. In this case, a bond of unprecedented size—well beyond the capacity of a bonding agency—was required. The unprecedented scale of Texaco's bond was an important element in the appeals process and the ultimate out-of-court settlement. The dispute also illustrates the high transactions costs that can accompany major litigation—costs that discourage litigation and encourage the establishment of institutions for the private governance of contractual relationships. Besides the direct costs of litigation, the dispute raised uncertainties about Texaco's continuing operations, thus reducing the company's stock price significantly (without adding commensurately to Pennzoil's price). One study indicates that the legal dispute reduced the combined wealth of Texaco and Pennzoil by some $3 billion. Approximately two-thirds of this sum was regained after the final settlement. See Cutler, David M. and Summers, Lawrence H., “The Costs of Conflict Resolution and Financial Distress: Evidence from the Texaco-Pennzoil Litigation,” Rand Journal of Economics 19 (Summer 1988), pp. 157–72CrossRefGoogle Scholar.

22. This backing for promises is qualified in at least two senses. First, it leaves aside the expense and opportunity costs of using the courts (some of which may be recovered in the final judgment). Second, it assumes that the contested promises can somehow be demonstrated to the satisfaction of a third party. For oral promises, this may be a difficult hurdle, as Goldwyn noted.

23. Fried and Atiyah represent opposite poles in this debate. Fried argues that the common law of contracts is based on the moral institution of promising, rather than on commercial exchange. To sustain this institution, the recipients of broken promises should be awarded their expectations of profit. Atiyah argues that court decisions have moved away from this strict emphasis, which arose in the nineteenth century, and returned to an older notion of commercial practice, which limits awards to the costs incurred in relying on broken promises. See Fried, Charles, Contract as Promise: A Theory of Contractual Obligation (Cambridge, Mass.: Harvard University Press, 1981Google Scholar); Atiyah, Patrick S., From Principles to Pragmatism (Oxford: Clarendon Press, 1978Google Scholar); and Atiyah, Patrick S., The Rise and Fall of Freedom of Contract (Oxford: Clarendon Press, 1979)Google Scholar.

24. For the classic statement, see Fuller, Lon L. and Perdue, William R., “The Reliance Interest in Contract Damages,” parts 1 and 2, Yale Law Review, vol. 46, 1936, pp. 52–96 and 373–420CrossRefGoogle Scholar.

25. This is usually a monetary award. Occasionally, it is a requirement to perform the specific promises in the contract. See Adams, John and Brownsword, Roger, Understanding Contract Law (London: Fontana, 1987), p. 144Google Scholar; and Fried, Contract as Promise.

26. The courts themselves require some efforts at self-protection. Once a contract has been breached, for instance, the “innocent” party is expected to take reasonable actions to minimize the damages and cannot win awards that cover a failure to do so. For the efficiency implications of this legal doctrine, see Kronman, Anthony and Posner, Richard, The Economics of Contract Law (Boston: Little, Brown, 1979), pp. 160–61Google Scholar.

27. See Mnookin, Robert H. and Kornhauser, Lewis, “Bargaining in the Shadow of the Law: The Case of Divorce,” Yale Law Journal 88 (04 1979), pp. 950–97CrossRefGoogle Scholar. Mnookin and Kornhauser also conclude that the impact of differing legal arrangements on divorce settlements cannot be specified with precision. They attribute that to a more general theoretical gap: a limited understanding of how alternative institutional arrangements can affect bargaining outcomes.

28. These are often ad hoc procedures designed for a specific agreement. Their powers may be quasi-judicial, as in the dispute mechanisms of the General Agreement on Tariffs and Trade (GATT), or merely consultative, as in the procedures of the U.S.-Soviet Standing Consultative Commission, established in SALT I and SALT II. The presence of quasi-judicial bodies attached to specific agreements indicates, once again, the limits of international adjudication. And it points to the ad hoc means devised to manage the risks of international cooperation. See Bilder, Richard B., Managing the Risks of International Agreement (Madison: University of Wisconsin Press, 1981), pp. 56–61Google Scholar.

29. A signatory always has the practical option of withdrawal, whether it is included as a legal option in the treaty or not. For legal analyses, see David, Arie E., The Strategy of Treaty Termination: Lawful Breaches and Retaliations (New Haven, Conn.: Yale University Press, 1975), pp. 203–16Google Scholar; and Briggs, Herbert W., “Unilateral Denunciation of Treaties: The Vienna Convention and the International Court of Justice,” American Journal of International Law 68 (01 1974), pp. 51–68CrossRefGoogle Scholar.

30. There have been proposals, based on efficiency grounds or libertarian principles, that private agents play a much larger role in enforcing domestic laws and contracts and that they be compensated by bounties, paid either by violators or the state. These proposals cannot be applied to international agreements without significant modification, since they ultimately envision authoritative judicial interpretation and enforcement. See Becker, Gary S. and Stigler, George J., “Law Enforcement, Malfeasance and Compensation of Enforcers,” Journal of Legal Studies 3 (01 1974), pp. 1–18CrossRefGoogle Scholar; Becker, Gary S., “Crime and Punishment: An Economic Approach,” Journal of Political Economy 76 (03–03 1968), pp. 169–217CrossRefGoogle Scholar; and Stigler, George J., “The Optimum Enforcement of Laws,” Journal of Political Economy 78 (05–06 1970), pp. 526–36CrossRefGoogle Scholar.

31. As Mnookin and Kornhauser point out in their study of divorce laws, “A legal system might allow varying degrees of private ordering upon dissolution of the marriage. Until recently, divorce law attempted to restrict private ordering severely.” See Mnookin, and Kornhauser, , “Bargaining in the Shadow of the Law,” pp. 952–53Google Scholar.

32. Three standard reasons are given for the legal requirement that contracts be put in writing. First, it should impart caution before an agreement is completed. Second, it should make clear to the parties that they have undertaken specific obligations. Third, if disagreements later arise, it should provide better evidence for courts. See the classic analysis by Fuller, Lon L.: “Consideration and Form,” Columbia Law Review, vol. 41, 1941, pp. 799–824CrossRefGoogle Scholar; and Anatomy of Law (New York: Praeger, 1968), pp. 36–37Google Scholar.

33. There is one restriction worth noting on the legal form of international agreements. The World Court will only consider agreements that have been formally registered with the United Nations. If the World Court were a powerful enforcement body, this restriction would influence the form of major agreements.

34. Criminal organizations such as the Mafia can be understood partly as an institutional response to the problems of providing criminal services when the bargains themselves are illegal. For a fascinating economic study of such institutional arrangements, see Reuter, Peter, Disorganized Crime: Illegal Markets and the Mafia (Cambridge, Mass.: MIT Press, 1983)Google Scholar.

35. On the mechanisms of private governance, see Williamson, Oliver R., The Economic Institutions of Capitalism: Firms, Markets, Relational Contracting (New York: Free Press, 1985)Google Scholar.

36. If a state already has a poor reputation for keeping its promises, then it risks little in staking that reputation on other agreements, and its pledges will fail to convince future partners without special efforts (such as bonds, hostages, or collateral) and careful monitoring, all designed to minimize reliance on “trust.” That does not rule out treaties, but it suggests that they may be disingenuous and cannot be relied upon. Stalin and Hitler, for example, found their pact useful because it produced immediate gains for each: the division of Eastern Europe. The incorporation of the new territories also postponed a confrontation between the two. The pact was useful for these immediate and simultaneous gains, not for any future promises of cooperation it held out.

37. Of course, commitments may be cast aside, no matter how formal, as Hussein, Saddam did when he declared Iraq's border agreement with Iran “null and void” in 1981Google Scholar. The agreement, reached in 1975 in Algiers, stated that “land and river frontiers shall be inviolable, permanent and final.” There is a cost to discarding such an agreement unilaterally, even if that cost seems remote at the time. It virtually rules out the ability to conclude useful agreements on other border disputes. See United Nations, Yearbook of the United Nations, 1981, vol. 35 (New York: United Nations, 1985), pp. 238–39Google Scholar. See also Iran, , Ministry of Foreign Affairs, Legal Department, A Review of the Imposed War (Tehran: Ministry of Foreign Affairs, 1983Google Scholar), including the text of the 1975 treaty, the treaty addendum, and Iran's interpretation.

38. In this sense, secret treaties are similar to informal agreements.

39. In other words, if the future is highly valued, there can be an equilibrium in which the (current discounted) value of a reputation exceeds any short-run gains from taking advantage of it. If the prospective gains from reputation are sufficiently large, then it also pays to invest in reputation. See Kreps, David M., A Course in Microeconomic Theory (Princeton, N.J.: Princeton University Press, 1990), p. 532Google Scholar.

40. Ramseyer, J. Mark, “Legal Rules in Repeated Deals: Banking in the Shadow of Defection in Japan,” Journal of Legal Studies 20 (01 1991), p. 96Google Scholar.

41. Again, the shadow of the future is crucial. If future rewards are sharply discounted, then it pays to exploit prior reputation (to disinvest) to reap short-term rewards.

42. Elsewhere, I have shown that sovereign debtors in the nineteenth century moved to settle their old defaults when they contemplated seeking new loans. Creditors had the greatest bargaining leverage at precisely these moments. See Lipson, Charles, Standing Guard: Protecting Foreign Capital in the Nineteenth and Twentieth Centuries (Berkeley and Los Angeles: University of California Press, 1985), p. 47Google Scholar. See also Marichal, Carlos, A Century of Debt Crises in Latin America: From Independence to the Great Depression, 1820–1930 (Princeton, N.J.: Princeton University Press, 1989), pp. 122–23Google Scholar.

43. The short-term price of reputation may either be foregone opportunities or direct expenditures, such as fixed investments that are most valuable within a specific bilateral relationship. Williamson has explored the use of such fixed investments to make credible commitments in The Economic Institutions of Capitalism.

44. The United States made such an investment in reputation in the late 1970s, after its credibility as leader of the North Atlantic Treaty Organization (NATO) was damaged by the neutron bomb affair. The problem arose after the Carter administration first supported and then opposed NATO's deployment of new antitank weapons, equipped with enhanced radiation warheads or neutron bombs. Key European leaders had already declared their support publicly, at considerable political cost, and now they had to reverse course. After the crisis died down, the Carter administration proposed another approach to nuclear modernization: Pershing II missiles. The administration then held fast (as did the Reagan administration) in support of its new plan. It did so despite a rising tide of public protest abroad and wavering support from European leaders, especially the Germans, who had initially proposed the modernization. According to Garthoff, “The principal effect of the neutron weapon affair was to reduce Western confidence in American leadership in the alliance, and later to lead the United States to seek to undo that effect by another new arms initiative for NATO… The Carter administration itself felt it needed to compensate for its handling of the neutron decision. It sought to do so by responding boldly to a perceived European concern through exercising vigorous leadership Doubts about the military necessity or even desirability of deploying new [long-range tactical nuclear force] systems were overwhelmed by a perceived political necessity within the alliance.” See Garthoff, Raymond L., Detente and Confrontation: American-Soviet Relations from Nixon to Reagan (Washington, D.C.: Brookings Institution, 1985), pp. 853 and 859Google Scholar.

45. Firms can guarantee quality by offering warranties. But what guarantees the warranty? The answer for expensive items may be the threat of litigation. But for less expensive items, it is simply the firm's reputation. “The hostage for performance,” according to Rubin, “must be in the familiar form of a quasirent stream [either of profits or return on capital]. In either case, the price of the product must be above marginal cost, and the difference must be high enough so that cheating by the firm does not pay.” See Rubin, Paul, Managing Business Transactions: Controlling the Cost of Coordinating, Communicating, and Decision Making (New York: Free Press, 1990), p. 147Google Scholar.

46. Telser, L. G., “A Theory of Self-Enforcing Agreements,” Journal of Business 53 (01 1980), pp. 27–28CrossRefGoogle Scholar.

47. Thus, a single agreement can be self-enforcing, even if it is divorced from any reputational concerns. Conversely, even when reputational issues are salient, a treaty may break down if other costs are more important.

48. A poor reputation impedes a state's future agreements because the state cannot use its reputation as a credible and valuable “performance bond.”

49. “Reputation commands a price (or exacts a penalty),” Stigler once observed, “because it economizes on search.” When that search must cover unknown future behavior, such as a partner's likelihood of complying with an agreement, then reputations are particularly valuable. See Stigler, George, “The Economics of Information,” Journal of Political Economy 69 (06 1961), p. 224CrossRefGoogle Scholar.

50. This discount rate refers only to the present value of known future benefits. It assumes perfect information about future payoffs. Greater risk or uncertainty about future benefits can also affect their present value.

51. This logic should apply to all agreements lacking effective third-party enforcement, from modern warfare to premodern commerce. For an application of this approach to medieval economic history, see Veitch, John M., “Repudiations and Confiscations by the Medieval State,” Journal of Economic History 46 (03 1986), pp. 31–36CrossRefGoogle Scholar.

52. Of course, states often do go to war alongside their long-time allies. My point is that if the costs are high (relative to longer-term reputational issues), their decision will be guided largely by their calculus of short-term gains and losses. That determination is largely independent of alliance agreements and formal treaties of mutual support. Knowing that, states facing war are reluctant to count too heavily on prior commitments, however formal or sincere, by alliance partners. By the same token, opponents have considerable incentives to design coalition-splitting strategies by varying the immediate costs and stakes to individual coalition members. This debate over long-term reputation versus short-term costs figured prominently in the British cabinet's debate over commitments to France before World War I.

53. See Baxter, , “International Law in 'Her Infinite Variety,'” p. 550Google Scholar. See also SeidlHohenfeldern, Ignaz, “International Economic Soft Law,” Recueil de cours (Collected Courses of the Hague Academy of International Law), vol. 163, 1979, pp. 169–246Google Scholar.

54. See Baxter, , “International Law in 'Her Infinite Variety,'” p. 551. Baxter refers to alliances and statements of broad political intent (such as the Yalta Agreement) as “political treaties.” He does not define the term further or distinguish it from other kinds of treatiesGoogle Scholar.

55. Ibid.

56. Realists consider cooperation important in only one sphere: military alliances. “In anarchy, states form alliances to protect themselves,” says Walt. “Their conduct is determined by the threats they perceive.” Although such alliances are important, they are simply considered the by-products of a world fundamentally characterized by conflict and the contest for relative gains. As Grieco bluntly puts it, “States are predisposed toward conflict and competition, and they often fail to cooperate even when they have common interests.” See Walt, Stephen M., The Origins of Alliances (Ithaca, N.Y.: Cornell University Press, 1987), p. xGoogle Scholar; and Grieco, Joseph M., Cooperation Among Nations: Europe, America, and Non-Tariff Barriers to Trade (Ithaca, N.Y.: Cornell University Press, 1990), p. 4Google Scholar.

57. Adelman emphasizes the slowness of negotiating formal agreements, especially major agreements with the Soviets. The Limited Test Ban Treaty (1963) took eight years to complete; the Non-Proliferation Treaty (1968) took more than three years; and the SALT I agreement (1972) took more than two years. The SALT II agreement (1979) took seven years and still failed to win Senate ratification. See Adelman, Kenneth, “Arms Control With and Without Agreements,” Foreign Affairs 63 (Winter 1984–1985), pp. 240–63CrossRefGoogle Scholar.

58. The slowness and difficulty of ratifying complex agreements and the problems of adapting to meet changing circumstances often lead states to choose less formal mechanisms. The United States and European Community (EC) have made exactly that choice to deal with their conflicts over “competition policy” and antitrust. The two sides “have abandoned the idea of drawing up a special treaty on competition issues,” such as mergers and acquisitions, according to the Financial Times, “because it would be too complicated, and would involve obtaining the approval of both the U.S. congress and EC member states. Instead, they discussed more flexible arrangements providing for a better exchange of information, regular meetings and discussions on current cases, and a means of settling disputes.” See Financial Times, 17 January 1991, p. 6.

59. One reason U.S. diplomats favored the Panama Canal treaties was that Latin American states were so uniformly opposed to continued U.S. ownership of the waterway. The disposition of this issue was crucial to America's role in the hemisphere. A number of Latin American leaders were involved in the negotiations, and some eighteen heads of state attended the final signing ceremony. The whole episode clearly demonstrates how “third parties” can have a direct stake in bilateral treaty arrangements and how these diplomatic relationships can be enhanced or, more likely, embarrassed by frank disclosures during the public ratification process. See Pastor, Robert A. and Castaneda, Jorge G., Limits to Friendship: The United States and Mexico (New York: Knopf, 1988), pp. 159–61Google Scholar.

60. Hogan, J. Michael, The Panama Canal in American Politics: Domestic Advocacy and the Evolution of Policy (Carbondale: Southern Illinois Press, 1986), pp. 190–91 and 200–205Google Scholar.

61. This key treaty won ratification by a single vote, and only then after intense presidential lobbying and some significant amendments. The Carter administration's political dilemma was nicely summarized by the Republican leader in the Senate, Howard Baker: “The Canal has a constituency and the treaty has no constituency.” See Stater, Timothy, “Climax: Senate Ratification, 1977–1978,” in Summ, G. Harvey and Kelly, Tom, eds., The Good Neighbors: America, Panama, and the 1977 Canal Treaties (Athens: Ohio University Center for International Studies, 1988), pp. 90–91Google Scholar; and Moffett, George D., The Limits of Victory: The Ratification of the Panama Canal Treaties (Ithaca, N.Y.: Cornell University Press, 1985)Google Scholar.

62. See “Treaty on Extradition and Mutual Assistance in Criminal Matters Between the United States of America and the Republic of Turkey, with Appendix, Signed June 7,1979, Entered into Force January 1,1981,” in United States Treaties and Other International Agreements, vol. 32, part 3 (Washington, D.C.: Government Printing Office, 1986), pp. 3111 ffGoogle Scholar.

63. Shearer, I. A., Extradition in International Law (Manchester, UK: Manchester University Press, 1971)Google Scholar.

64. See “Treaty on Extradition and Mutual Assistance in Criminal Matters Between the United States of America and the Republic of Turkey.”

65. Just what agreements must be submitted as treaties remains ambiguous. It is a constitutional question, of course, but also a question of the political balance of power between the Congress and the President. At one point, President Carter's chief of staff, Hamilton Jordan, announced that Carter would decide whether the Panama Canal agreements were treaties or not. He “could present [the accords] to the Congress as a treaty, or as an agreement, and at the proper time he'll make that decision.” Interview on “Face the Nation,” CBS News, cited by Johnson, Loch K. in The Making of International Agreements: Congress Confronts the Executive (New York: New York University Press, 1984), p. 141Google Scholar.

66. Stanbrook, Ivor and Stanbrook, Clive, The Law and Practice of Extradition (Chichester, UK: Barry Rose, 1980), p. xxviiGoogle Scholar.

67. The U.S. Constitution, Article II, Section 2, provides that the President “shall have power, by and with the Advice and Consent of the Senate, to make treaties, provided two-thirds of the Senators present concur.” For a detailed study of the constitutional issues, see Henkin, Louis, Foreign Affairs and the Constitution (Mineola, N.Y.: Foundation Press, 1972)Google Scholar.

68. See the War Powers Resolution, 87 Stat. 555,1973; and 50 United States Code 1541–48,1980.

69. See Baxter, , “International Law in 'Her Infinite Variety,'” pp. 554–55Google Scholar. See also U.S. Congress, Senate Committee on the Judiciary, Congressional Oversight of Executive Agreements: Hearings on S. 3475, 92d Congress, 2d sess., 1972Google Scholar.

70. Congress, U.S., Senate Committee on Foreign Relations, Transmittal of Executive Agreements to Congress, Senate Report no. 92–591,92d Congress, 2d sess., 1972, pp. 3–4Google Scholar.

71. The Case Act passed the Senate and House overwhelmingly in 1972 and has been amended several times since then. The State Department finally issued implementation regulations in 1981. The original act was Public Law 92–403. For the amended version, see 1 United States Code 112b, 1988. For implementing regulations, see 22 Code of Federal Regulations, part 181; and 46 Federal Register, 13 July 1981, pp. 35917 ff.

72. A detailed study of the Case Act shows the limits of congressional activism. The Senate, in particular, wanted a more direct role in shaping informal agreements. But during the period from the 1950s through the 1980s, there was never a majority that would require informal agreements to be submitted for ratification. The Case Act succeeded, in part, because it recognized these political limits and required only notification. See Johnson, The Making of International Agreements, chap. 5.

73. For one model of how technical innovations could complicate treaty maintenance, see Downs and Rocke, Tacit Bargaining, Arms Races, and Arms Control, chap. 5.

74. David, The Strategy of Treaty Termination.

75. Article XIX of the GATT covers safeguards. It permits the Contracting Parties to offer emergency protection to industries disrupted by imports. See Bronckers, Marco, Selective Safeguard Measures in Multilateral Trade Relations: Issues of Protectionism in GA TT, European Community, and United States Law (Deventer, Netherlands: Kluwer, 1985Google Scholar); and Kleen, Peter, “The Safeguard Issue in the Uruguay Round: A Comprehensive Approach,” Journal of World Trade 25 (10 1989), pp. 73–92Google Scholar.

76. Typical was the communiquéissued in late 1983 at the conclusion of OPEC's sixty-ninth conference. It restated OPEC's production ceiling of 17.5 million barrels per day and its marker price of $29. See Skeet, Ian, OPEC: Twenty-Five Years of Prices and Politics (Cambridge: Cambridge University Press, 1988), p. 196Google Scholar.

77. Ibid., introduction and chap. 1.

78. See Tanner, James and Sullivan, Allanna, “OPEC Peace May Be Short-Lived as Debate Looms on New Quotas,” Wall Street Journal, European edition, 12 06 1989, pp. 1Google Scholar and 10; and “A Confederacy of Cheats,” Economist, 10 June 1989, p. 88.

79. OPEC's founders had hoped to create an operating international institution, not merely a negotiating forum. In the original organization, there was even an enforcement section. The section was effectively dropped by 1964 and disappeared formally in 1966. This transformation amounted to an “implicit admission … that OPEC and its Secretary General would not act, as had been visualized at its creation, as an operating arm of its members.” See Skeet, , OPEC: Twenty-Five Years of Prices and Politics, p. 237Google Scholar.

80. On the importance of surprise for cooperation in security affairs, see Lipson, Charles, “International Cooperation in Economic and Security Affairs,” World Politics 37 (10 1984), pp. 1–23CrossRefGoogle Scholar.

81. Following the same logic, both sides have been cautious about treaties covering chemical and biological weapons, whose production and deployment have been much more difficult to detect. Monitoring a partial ban would be extremely difficult, especially since other kinds of production facilities could be converted to military uses.

82. Note, however, that if the discussions pertained to domestic bargains, a court might interpret these “agreements in principle” as contractually binding, depending on the level of detail and the promissory language. Once again, the absence of effective international courts matters.

83. Public statement issued by U.S. Department of State on 4 March 1981 and cited by Talbott, in Deadly Gambits, p. 225Google Scholar.

84. In his “Remarks at Memorial Day Ceremonies at Arlington National Cemetery,” p. 709, President Reagan made the following statement: “As for existing strategic arms agreements, we will refrain from actions which undercut them so long as the Soviet Union shows equal restraint.”

85. Presidential response to a question on 13 May 1982, quoted in Weekly Compilation of Presidential Documents, vol. 18, no. 19, 17 05 1982, p. 635Google Scholar.

86. President Carter declared that the United States would continue to observe the terms of the interim SALT II agreement after its expiration and was roundly criticized for exceeding his legal powers. His successor, who had campaigned against SALT II, declined to make any “parallel unilateral policy declarations.” The Congress was not so inhibited. In the 1984 Defense Authorization Act, it declared that the United States should not undermine existing international agreements on offensive strategic arms, at least until the SALT II agreement expired in December 1985, provided the Soviet Union did the same. President Reagan made a similar statement but added that the Soviets must also pursue a new agreement on strategic arms (the START talks). In December 1986, the United States exceeded the aggregate SALT limits on strategic weapons when it put a new B-52, equipped with cruise missiles, into service. Even then, one year after the original expiration date of the SALT II agreement, the United States defended its actions as a sanction for Soviet violations. See Calvo-Goller, and Calvo, , The SALT Agreements, p. 330Google Scholar.

87. Bilder, , Managing the Risks of International Agreement, pp. 49–51Google Scholar.

88. Raymond Vernon, writing on foreign investments, has shown the dangers of violating this approach. Even if an agreement provides significant benefits to both sides, it may provide those benefits to one side immediately and to the other much later. Such agreements are vulnerable to noncompliance in midstream, after one side has already received its benefits. This is one element of Vernon's “obsolescing bargain.” It is a variant of Hobbes's critique of covenants, in which one side performs its side of the bargain first. See Vernon, Raymond, Sovereignty at Bay: The Multinational Spread of U.S. Enterprises (New York: Basic Books, 1971Google Scholar). On the general logic of self-sustaining agreements, see Telser, , “A Theory of Self-Enforcing Agreements,” pp. 27–44Google Scholar.

89. This allows negotiators to make reasonable calculations about the various parties’ ex post incentives to defect during the life of the agreement.

90. See “Limitation of Anti-Ballistic Missile Systems, Signed May 26, 1972, with Agreed Interpretations, Common Understandings, and Unilateral Statements,” in United States Treaties and Other International Agreements, vol. 23, part 4 (Washington, D.C.: Government Printing Office, 1973), pp. 3435–61Google Scholar.

91. These issues are the subject of an extensive literature. See, for example, Cimbala, Stephen J., ed., The Technology, Strategy and Politics of SDI (Boulder, Colo.: Westview Press, 1987)Google Scholar.

92. In modern international politics, these hidden agreements are informal because ratification is public and the treaties are registered with the United Nations. In earlier international systems, however, neither condition applied and secret treaties were possible.

93. Garthoff, Raymond L., Reflections on the Cuban Missile Crisis, revised ed. (Washington, D.C.: Brookings Institution, 1989), pp. 86–87Google Scholar.

94. Kennedy, Robert F., Thirteen Days: A Memoir of the Cuban Missile Crisis (New York: Norton, 1971), pp. 108–9Google Scholar.

95. Garthoff, , Reflections on the Cuban Missile Crisis, p. 95 nGoogle Scholar.

96. Ibid.

97. It is sometimes argued that the Turkish and Italian missiles were not part of any deal, since the United States would soon have removed them even if the Soviets had not raised the issue. The counterargument, which I find more convincing, is that the Soviets sought their removal as part of the bargaining on Cuba and the United States did, in effect, agree that it “expected” to remove them. The American decision was timed and disclosed specifically to encourage Soviet acceptance of the larger deal. Garthoff calls this decision “an additional sweetener” that “certainly made it easier for Khrushchev to accept the basic over-the-table settlement.” See Garthoff, , Reflections on the Cuban Missile Crisis, p. 87Google Scholar; see also pp. 88 and 94–95. Given the high stakes in Cuba, Soviet behavior was probably little changed by the Turkish side-payment. We will never know with certainty. From the U.S. viewpoint, however, it was wise to make the concessions part of the bargaining: they were minor, secret, and potentially quite rewarding. When information is imperfect, as it was in this case, clever negotiators can “bargain away” concessions that would have been undertaken anyway. The tactic is to make them appear contingent, with the goal of extracting additional concessions and sealing the final bargain. That is exactly what the United States did with the aging Turkish and Italian missiles, while minimizing the reputational costs of any apparent concessions.

98. In 1906, the British Foreign Minister, Sir Edward Grey, discussed the dilemmas posed by these expectations. The entente agreements, signed by a previous British government, “created in France a belief that we shall support [the French] in war If this expectation is disappointed, the French will never forgive us. There would also I think be a general feeling that we had behaved badly and left France in the lurch On the other hand the prospect of a European war and of our being involved in it is horrible.” See document no. 299, in Gooch, G. P. and Temperley, Harold, eds., British Documents on the Origin of the War, 1898–1914, vol. 3 (London: His Majesty's Stationery Office, 1928), p. 266Google Scholar.

99. Although the State Department did try to persuade Congress that Western Europe was aiding the embargo, its efforts were in vain. Quiet reassurances from the State Department were distrusted by a hard-line, anticommunist Congress, which saw them as self-serving maneuvers to preserve diplomatic ties. See Mastanduno, Michael, “Trade as a Strategic Weapon: American and Alliance Export Control Policy in the Early Postwar Period,” in Ikenberry, G. John, Lake, David A., and Mastanduno, Michael, eds., The State and American Foreign Economic Policy (Ithaca, N.Y.: Cornell University Press, 1988), p. 136Google Scholar.

100. See Mutual Defense Assistance Control Act of 1951 (“Battle Act”), 82d Congress, 1st sess., 65 Stat. 644.

101. Note that Singapore and Hong Kong do not have formal diplomatic relations with either the People's Republic of China or Taiwan. Far from being an impediment, this absence of formal ties contributes to their role as intermediaries. Given the long-standing controversy over diplomatic recognition, the best way to maintain links to both the mainland and the island is through these informal, back channels. The same channels have been used extensively by overseas Chinese communities to arrange for burials or to visit the graves of ancestors interred on the mainland.

102. These figures, which were reported in The New York Times, 14 April 1990, p. 17, are based on data from the Ministry of Economic Affairs in Taiwan and the government in Hong Kong and include trade and investment routed through Hong Kong companies to avoid Taiwanese restrictions.

103. Leifer, Michael and Yahuda, Michael, “Third Party China?” working paper, London School of Economics, 1989, p. 1Google Scholar.

104. Ibid., p. 2.

105. Because informal extradition arrangements are ad hoc, they are easily severed. That is a mixed blessing. It means that extradition issues are directly implicated in the larger issues of bilateral diplomacy. They cannot be treated as distinct, technical issues covered by their own treaty rules. For example, the bloody suppression of popular uprisings in 1989 in the People's Republic of China blocked prisoner exchanges and made trade and investment ties politically riskier.

106. This definition is based on the second meaning of “tacit” in The Oxford English Dictionary, 2d ed., vol. 17 (Oxford: Clarendon Press, 1989), p. 527Google Scholar.

107. See Keal, Unspoken Rules and Superpower Dominance; and Kratochwil, Friedrich, Rules, Norms and Decisions (Cambridge: Cambridge University Press, 1989CrossRefGoogle Scholar), chap. 3.

108. A few international lawyers argued that the Yalta and Potsdam agreements were binding treaty commitments. The U.S. State Department did publish the Yalta Agreement in the Executive Agreements Series (no. 498) and in U.S. Treaties in Force (1963). In 1948, Sir Hersch Lauterpacht said that they “incorporated definite rules of conduct which may be regarded as legally binding on the States in question.” The British and American governments explicitly rejected that view. In 1956, in an aide-memoire to the Japanese government, the State Department declared that “the United States regards the so-called Yalta Agreement as simply a statement of common purposes by the heads of the participating governments and … not as of any legal effect in transferring territories.” See Department of State Bulletin, vol. 35, 1956, p. 484, cited by Schachter, in “The Twilight Existence of Nonbinding International Agreements,” p. 298Google Scholar n. See also L. P. L. Oppenheim, Peace, vol. 1 of Lauterpacht, H., ed., International Law: A Treatise, 7th ed. (London: Longmans, Green, 1948), p. 788Google Scholar, section 487.

109. The one major exception among U.S. politicians was former vice president Henry Wallace, representing the left wing of the Democratic party. Wallace openly stated that the Soviets had legitimate security interests in Eastern Europe and should not be challenged directly there. His views were widely denounced in both parties and won few votes.

110. Schlesinger, Arthur M. Jr, “Origins of the Cold War,” Foreign Affairs 46 (Autumn 1967), pp. 22–52CrossRefGoogle Scholar.

111. See Schelling, Thomas C., “Reciprocal Measures for Arms Stabilization,” Daedalus 89 (Fall 1960), pp. 892–914Google Scholar: Schelling, Thomas C., “What Went Wrong with Arms Control?” Foreign Affairs 64 (Winter 1985—1986), pp. 219–33CrossRefGoogle Scholar; and Schelling, Thomas C. and Halperin, Morton H., Strategy and Arms Control, 2d ed. (Washington, D.C.: Pergamon-Brassey, 1985), pp. 77–90Google Scholar. Schelling's point is strongly endorsed by Adelman in “Arms Control With and Without Agreements.”

112. Axelrod, Robert, The Evolution of Cooperation (New York: Basic Books, 1984Google Scholar).

113. See the following works by Downs and Rocke: “Tacit Bargaining and Arms Control”; and Tacit Bargaining, Arms Races, and Arms Control.

114. In The Cement of Society, Elster makes this distinction between regret and surprise and relates it to two forms of order. Departures from regularized, predictable behavior give rise to surprise. Unreciprocated cooperation produces regret.

115. LaFeber, Walter, The New Empire: An Interpretation of American Expansion, 1860–1898 (Ithaca, N.Y.: Cornell University Press, 1963)Google Scholar.

116. See Garthoff, Detente and Confrontation, chap. 27. Until the Afghanistan invasion, there had been a sharp debate within the Carter administration over the implied terms of detente. The problem with implied terms, after all, is that they may well be ambiguous and differently understood by different actors, across states and within them. The eventual winner was National Security Adviser Zbigniew Brzezinski, who clearly stated his position early in the Carter administration while discussing the Horn of Africa. In March 1978, he wrote to President Carter that “the Soviets must be made to realize that detente, to be enduring, has to be both comprehensive and reciprocal. If the Soviets are allowed to feel that they can use military force in one part of the world—and yet maintain cooperative relations in other areas—then they have no incentive to exercise any restraint.” See Brzezinski, Zbigniew, Power and Principle: Memoirs of the National Security Adviser, 1977–1981 (New York: Farrar, Straus, Giroux, 1983), p. 186Google Scholar.

117. Ulam, Adam B., Dangerous Relations: The Soviet Union in World Politics, 1970–1982 (New York: Oxford University Press, 1983), pp. 260–61Google Scholar.

118. “The Soviet occupation of Afghanistan and the American response led to a sharp break from the whole course of U.S.-Soviet relations over the preceding decade,” according to Garthoff. “It gave the coup de grace to the already seriously eroded and weakened mutual policy of detente established in May 1972 … In many ways January 1980 was a sharper turning point than January 1981, when Ronald Reagan was inaugurated and repudiated detente.” See Garthoff, , Detente and Confrontation, p. 967Google Scholar.

119. See Bowker, Mike and Williams, Phil, Superpower Detente: A Reappraisal (London: Sage, 1988Google Scholar). Garthoff, in his review of this book, points to the contradictory and self-limiting assumptions that supported detente in the United States and Soviet Union: “Detente [according to Bowker and Williams] was not a cause of America's apparent weakness in the 1970s, but a hard-headed strategy devised by Kissinger and Nixon for coping with an adverse situation by managing the emergence of Soviet power. For the Soviet leaders, on the other hand, detente represented an expected opportunity to neutralize superpower relations and exercise Soviet power in a more active way in the Third World. When they did this, however, in Angola, Ethiopia and Afghanistan, the result was to undercut support for detente in the United States and lead to its collapse.” Garthoff, himself a notable student of superpower relations, sees Soviet policies in Africa and Asia as opportunistic rather than strategic. The results, however, were the same for U.S.-Soviet relations. There was a “costly failure of Soviet leaders to appreciate the adverse consequences of their involvements in the Third World on detente with the United States.” See Garthoff, Raymond L., “Review of Bowker and Williams’ Superpower Detente,” International Affairs 65 (Spring 1989), p. 311CrossRefGoogle Scholar.

120. Ullmann-Margalit, Edna, The Emergence of Norms (Oxford: Clarendon Press, 1977Google Scholar).

121. This does not rule out deliberate vagueness on some issues as part of a larger, more detailed settlement. Cooperation is not comprehensive, and some issues have to be finessed if any agreement is to be reached.

122. The Helsinki Final Act, formally known as the Final Act of the Conference on Security and Cooperation in Europe, was concluded in 1975 and signed by thirty-five states. On the one hand, the states declared their “determination to act in accordance with the provisions contained” in the text. On the other hand, these were not to be the binding commitments of a treaty. The text plainly said that it was not eligible for registration with the United Nations, as a treaty would be. Several democratic states, led by the United States, declared at the time that this document was not a treaty. “There does not appear to be any evidence that the other signatory states disagreed with this understanding,” according to Schachter. The result is a curious contradiction: a nonbinding bargain. It juxtaposes elaborate “commitments” with a claim that they are not to be registered, as a treaty would be. The point, clearly, is to exempt the provisions from the legally binding status of treaty commitments. For an interesting analysis of the Helsinki agreement and its ambiguous status in international law, see Schachter, , “The Twilight Existence of Nonbinding International Agreements,” p. 296Google Scholar. The text of the Helsinki Final Act can be found in International Legal Materials, vol. 14, 1975, pp. 1293 ffGoogle Scholar.

123. This translation of oral agreements into writing is required by the U.S. State Department's regulations implementing the Case Act. See “International Agreement Regulations,” 22 Code of Federal Regulations, part 181; and 46 Federal Register, 13 July 1981, pp. 35917 ff.

124. There is a nice irony here. Goldwyn's disparaging comments about oral agreements are themselves probably apocryphal. He regularly mangled the English language, and quotes like this were often attributed to him, whether he said them or not. The murky origins of this quotation underscore a fundamental problem with oral bargains. How can third parties ever ascertain who really promised what to whom? Goldwyn himself gave one answer to that question: “Two words: im possible.” See Easton, Carol, The Search for Sam Goldwyn (New York: William Morrow, 1976), pp. 150–51Google Scholar; and Marx, Arthur, Goldwyn: A Biography of the Man Behind the Myth (New York: Norton, 1976), pp. 8–10Google Scholar.

125. States on the borders of a guerrilla war are vital allies to the protagonists. They offer a secure launching pad for military operations and a secure site for communications and resupply. If their role becomes too open and prominent, however, the bordering states could be brought directly into the fighting as protagonists themselves. This is clearly a delicate relationship. It is best managed by informal agreements, usually secret ones, such as those reached by the United States and Laos during the Vietnam War. See Johnson, , The Making of International Agreements, p. 68Google Scholar.

126. The importance of secret treaties in European diplomacy was underscored when Woodrow Wilson tried to abolish the practice after World War I. Clemenceau and Lloyd George “said emphatically that they could not agree never to make a private or secret diplomatic agreement of any kind. Such understandings were the foundation of European diplomacy, and everyone knew that to abandon secret negotiations would be to invite chaos. To this [Colonel] House replied … that there was no intention to prohibit confidential talks on delicate matters, but only to require that treaties resulting from such conversations should become ‘part of the public law of the world.’” Quoted by Walworth, Arthur in America's Moment: 1918—American Diplomacy at the End of World War I (New York: Norton, 1977), p. 56Google Scholar.

127. See “Treaty of Non-Aggression Between Germany and the Union of Soviet Socialist Republics, August 23, 1939, Signed by Ribbentrop and Molotov,” document no. 228 in United Kingdom, , Foreign Office, The Last Days of Peace, August 9—September 3, 1939, series D, vol. 7 of Documents on German Foreign Policy, 1918–1945 (London: Her Majesty's Stationery Office, 1956), pp. 245–46Google Scholar. The volume provides official translations of documents from captured archives of the German Foreign Ministry and the Reich Chancellery.

128. See “Telegram, Most Urgent, [from] the Ambassador in the Soviet Union to the [German] Foreign Ministry, August 30,1939, Signed by Schulenburg,” document no. 447 in ibid., pp. 439–40.

129. See “Secret Additional Protocol, August 23, 1939, Signed by Ribbentrop and Molotov,” document no. 229 in ibid., pp. 246–47. For an analysis of the protocol, see Gerhard L. Weinberg, The Foreign Policy of Hitler's Germany: Starting World War II (Chicago: University of Chicago Press, 1980), pp. 602–10.

130. See “Secret Additional Protocol, September 28, 1939, Between the Government of the USSR and the Government of the German Reich,” document no. 159 in United Kingdom, Foreign Office, The War Years, September 4, 1939—March 18, 1940, series D, vol. 8 of Documents on German Foreign Policy, 1918–1945 (London: Her Majesty's Stationery Office, 1954), p. 166Google Scholar. The document, which was signed in Moscow by Ribbentrop and Molotov, stated the following: “The Secret Additional Protocol, signed August 23, 1939, shall be amended in item 1 to the effect that the territory of the Lithuanian states falls to the sphere of influence of the USSR, while, on the other hand, the province of Lublin and parts of the province of Warsaw fall to the sphere of influence of Germany (cf. the map attached to the Boundary and Friendship Treaty signed today).”

131. The relevant maps and microfilms are held in the Federal Republic's Foreign Ministry. The captured documents have all been published, in the original and in translation. The Soviets long claimed that they could not find their copies and that the West German microfilms were forgeries. They held to this formulaic answer through 1989, when Baltic nationalism became a serious political challenge. The nationalists, of course, emphasized the illegitimacy of the protocols and actually published their texts as part of their rising protest of Soviet rule. In August 1989, a senior Soviet official finally acknowledged that the Soviets and Nazis had secretly divided Eastern Europe, but he insisted that the secret agreement had no bearing on the current boundaries of the Soviet Union. See Gumbel, Peter, “Bonn Has Documents Soviets Don't Want Very Much to Find,” Wall Street Journal, European edition, 23–24 06 1989, pp. 1 and 10Google Scholar; Fein, Esther B., “Soviets Confirm Nazi Pacts Dividing Europe,” The New York Times, 19 08 1989, pp. 1Google Scholar and 5; and Fein, Esther B., “Soviet Congress Condemns ‘39 Pact That Led to Annexation of Baltics,” The New York Times, 25 12 1989, pp. 1Google Scholar and 15.

132. Trotsky's release of the secret documents was shrewd and effective. There was a strong, sustained reaction against secret diplomacy, mainly in the Anglo-Saxon countries. Wilson himself was politically embarrassed. Either his wartime allies had not told him of their earlier bargains or they had told him and he had kept the secret, despite his principled attacks on secret diplomacy. See Toscano, Mario, An Introduction to the History of Treaties and International Politics, vol. 1 of The History of Treaties and International Politics (Baltimore, Md.: Johns Hopkins University Press, 1966), pp. 42Google Scholar and 215; and Joll, James, Europe Since 1870, 2d ed. (Harmondsworth, UK: Penguin Books, 1976), p. 233Google Scholar.

133. Wilson's war aims were stated to a joint session of Congress on 8 January 1918. When European leaders later challenged this commitment to open covenants, Wilson announced that he would never compromise the “essentially American terms in the program,” including Point One. See House, Edward M., The Intimate Papers of Colonel House, vol. 4, ed. by Seymour, Charles (London: Ernest Benn, 1928), pp. 182–83Google Scholar.

134. Baxter, , “International Law in ‘Her Infinite Variety,’” p. 549Google Scholar.