Rationalist Explanations for War Author(s): James D. Fearon
Source: International Organization, Vol. 49, No. 3 (Summer, 1995), pp. 379-414
Published by: The MIT Press
Stable URL: http://www.jstor.org/stable/2706903
Accessed: 26-02-2015 23:42 UTC 战争的理性主义解释 作者:James D. Fearon
资料来源资料来源:《国际组织》,第 49 卷,第 3 期(1995 年夏),第 379-414 页。
出版商:麻省理工学院出版社麻省理工学院出版社
稳定 URL: http://www.jstor.org/stable/2706903
访问时间:26-02-2015 23:42 UTC
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org. JSTOR 是一项非营利性服务,帮助学者、研究人员和学生发现、使用和利用值得信赖的数字档案中的各种内容。我们利用信息技术和工具提高工作效率,促进新形式的学术研究。有关 JSTOR 的更多信息,请联系 support@jstor.org。
Rationalist explanations for war 战争的理性主义解释
James D. Fearon 詹姆斯-D-费伦
The central puzzle about war, and also the main reason we study it, is that wars are costly but nonetheless wars recur. Scholars have attempted to resolve the puzzle with three types of argument. First, one can argue that people (and state leaders in particular) are sometimes or always irrational. They are subject to biases and pathologies that lead them to neglect the costs of war or to misunderstand how their actions will produce it. Second, one can argue that the leaders who order war enjoy its benefits but do not pay the costs, which are suffered by soldiers and citizens. Third, one can argue that even rational leaders who consider the risks and costs of war may end up fighting nonetheless. 战争的核心难题,也是我们研究战争的主要原因,是战争代价高昂,但战争却一再发生。学者们试图通过三种论证来解决这个难题。首先,我们可以认为人(尤其是国家领导人)有时或总是非理性的。他们会受到偏见和病态的影响,从而忽视战争的代价,或误解自己的行为会如何导致战争。其次,我们可以说,下令发动战争的领导人享受了战争带来的好处,但却没有付出代价,而士兵和公民却要为此付出代价。第三,即使是理性的领导人,在考虑到战争的风险和代价后,也可能最终还是会参战。
This article focuses on arguments of the third sort, which I will call rationalist explanations. ^(1){ }^{1} Rationalist explanations abound in the literature on international conflict, assuming a great variety of specific forms. Moreover, for at least two reasons many scholars have given rationalist explanations a certain pride of place. First, historians and political scientists who have studied the origins of particular wars often have concluded that war can be a rational alternative for leaders who are acting in their states’ interest-they find that the expected benefits of war sometimes outweigh the expected costs, however unfortunate 本文重点讨论第三种论点,我称之为理性主义解释。 ^(1){ }^{1} 在有关国际冲突的文献中,理性主义解释比比皆是,其具体形式也多种多样。此外,至少出于两个原因,许多学者都赋予理性主义解释以一定的地位。首先,研究过特定战争起源的历史学家和政治学家通常会得出这样的结论:对于那些为了国家利益而行动的领导人来说,战争可能是一种理性的选择--他们发现,战争的预期收益有时会超过预期成本,无论多么不幸。
This article attempts to provide a clear statement of what a rationalist explanation for war is and to characterize the full set of rationalist explanations that are both theoretically coherent and empirically plausible. It should be obvious that this theoretical exercise must take place prior to testing rationalist explanations against alternatives-we cannot perform such tests unless we know what a rationalist explanation really is. Arguably, the exercise is also foundational for neorealism. Despite its prominence, neorealist theory lacks a clearly stated and fully conceived explanation for war. As I will argue below, it is not enough to say that under anarchy nothing stops states from using force, or that anarchy forces states to rely on self-help, which engenders mutual suspicion and (through spirals or the security dilemma) armed conflict. Neither do diverse references to miscalculation, deterrence failure because of inadequate forces or incredible threats, preventive and preemptive considerations, or free-riding in alliances amount to theoretically coherent rationalist explanations for war. 本文试图清楚地说明什么是理性主义对战争的解释,并描述一整套理论上连贯、经验上合理的理性主义解释。显而易见,这一理论研究必须先于理性主义解释与其他解释的对比测试--除非我们知道理性主义解释到底是什么,否则我们无法进行这样的测试。可以说,这项工作也是新现实主义的基础。尽管新现实主义理论占据重要地位,但它缺乏对战争的清晰阐述和完整构思的解释。正如我将在下文论述的那样,仅仅说在无政府状态下,没有什么能阻止国家使用武力,或者说无政府状态迫使国家依靠自助,从而导致相互猜疑和(通过螺旋式上升或安全困境)武装冲突,是不够的。关于误判、威慑因力量不足或令人难以置信的威胁而失败、预防性和先发制人的考虑或联盟中的搭便车行为等各种提法,也不能构成理论上连贯的理性主义战争解释。
My main argument is that on close inspection none of the principal rationalist arguments advanced in the literature holds up as an explanation because none addresses or adequately resolves the central puzzle, namely, that war is costly and risky, so rational states should have incentives to locate negotiated settlements that all would prefer to the gamble of war. The common flaw of the standard rationalist arguments is that they fail either to address or to explain adequately what prevents leaders from reaching ex ante (prewar) bargains that would avoid the costs and risks of fighting. A coherent rationalist explanation for war must do more than give reasons why armed conflict might appear an attractive option to a rational leader under some circumstances-it must show why states are unable to locate an alternative outcome that both would prefer to a fight. 我的主要论点是,仔细观察文献中提出的主要理性主义论点,没有一个能站得住脚,因为没有一个论点能解决或充分解决核心问题,即战争成本高、风险大,因此理性国家应该有动力通过谈判达成所有人都愿意接受的解决方案,而不是战争赌博。标准理性主义论点的共同缺陷在于,它们既没有解决也没有充分解释是什么阻碍了领导人达成事前(战前)谈判,从而避免战争的成本和风险。对战争进行连贯的理性主义解释,不仅要说明为什么在某些情况下武装冲突可能对理性的领导人有吸引力,还必须说明为什么国家无法找到一个双方都倾向于战争的替代结果。
To summarize what follows, the article will consider five rationalist arguments accepted as tenable in the literature on the causes of war. Discussed at 概括地说,文章将考虑在有关战争原因的文献中被公认为站得住脚的五个理性主义论点。讨论
length below, these arguments are given the following labels: (1) anarchy; (2) expected benefits greater than expected costs; (3) rational preventive war; (4) rational miscalculation due to lack of information; and (5) rational miscalculation or disagreement about relative power. I argue that the first three arguments simply do not address the question of what prevents state leaders from bargaining to a settlement that would avoid the costs of fighting. The fourth and fifth arguments do address the question, holding that rational leaders may miss a superior negotiated settlement when lack of information leads them to miscalculate relative power or resolve. However, as typically stated, neither argument explains what prevents rational leaders from using diplomacy or other forms of communication to avoid such costly miscalculations. 这些论点被贴上以下标签:(1)无政府状态;(2)预期收益大于预期成本;(3)理性的预防性战争;(4)缺乏信息导致的理性误判;(5)理性误判或对相对权力的分歧。我认为,前三个论点根本没有解决是什么阻碍了国家领导人通过讨价还价达成避免战争成本的解决方案这一问题。第四和第五个论点确实解决了这个问题,它们认为,当缺乏信息导致理性的领导者误判相对实力或决心时,他们可能会错过一个更好的谈判解决方案。然而,正如通常所说的那样,这两个论点都没有解释是什么阻碍了理性的领导人利用外交或其他形式的沟通来避免这种代价高昂的误判。
If these standard arguments do not resolve the puzzle on rationalist terms, what does? I propose that there are three defensible answers, which take the form of general mechanisms, or causal logics, that operate in a variety of more specific international contexts. ^(3){ }^{3} In the first mechanism, rational leaders may be unable to locate a mutually preferable negotiated settlement due to private information about relative capabilities or resolve and incentives to misrepresent such information. Leaders know things about their military capabilities and willingness to fight that other states do not know, and in bargaining situations they can have incentives to misrepresent such private information in order to gain a better deal. I show that given these incentives, communication may not allow rational leaders to clarify relative power or resolve without generating a real risk of war. This is not simply a matter of miscalculation due to poor information but rather of specific strategic dynamics that result from the combination of asymmetric information and incentives to dissemble. 如果这些标准论据不能从理性主义角度解决这一难题,那么什么才能解决呢?我提出了三种站得住脚的答案,它们以一般机制或因果逻辑的形式出现,在各种更为具体的国际环境中发挥作用。 ^(3){ }^{3} 在第一种机制中,理性的领导人可能会因为有关相对能力或决心的私人信息以及歪曲这些信息的动机而无法找到双方都喜欢的谈判解决方案。领导人对自己的军事能力和战斗意愿的了解是其他国家所不知道的,在讨价还价的情况下,他们可能有动机歪曲这些私人信息,以获得更好的交易。我的研究表明,在这些动机的驱使下,理性的领导人可能无法通过沟通来澄清相对实力或决心,而又不会引发真正的战争风险。这不仅仅是信息不全导致的误判问题,而是信息不对称和动机失实共同作用下的特定战略动态问题。
Second, rationally led states may be unable to arrange a settlement that both would prefer to war due to commitment problems, situations in which mutually preferable bargains are unattainable because one or more states would have an incentive to renege on the terms. While anarchy (understood as the absence of an authority capable of policing agreements) is routinely cited as a cause of war in the literature, it is difficult to find explanations for exactly why the inability to make commitments should imply that war will sometimes occur. That is, what are the specific, empirically identifiable mechanisms by which the inability to commit makes it impossible for states to strike deals that would avoid the costs of war? I identify three such specific mechanisms, arguing in particular that preventive war between rational states stems from a commitment problem rather than from differential power growth per se. 其次,由于承诺问题,理性主导的国家可能无法达成双方都倾向于战争的解决方案,在这种情况下,双方都倾向于的交易是无法达成的,因为一个或多个国家有动机违背条款。虽然无政府状态(被理解为缺乏能够监督协议的权威机构)在文献中经常被引述为战争的原因,但很难找到确切的解释来说明为什么无法做出承诺意味着战争有时会发生。也就是说,有哪些具体的、经验上可识别的机制,使得无法做出承诺的国家无法达成避免战争代价的协议?我提出了三个这样的具体机制,并特别指出,理性国家之间的预防性战争源于承诺问题,而非实力增长差异本身。
The third sort of rationalist explanation I find less compelling than the first two, although it is logically tenable. States might be unable to locate a peaceful 第三种理性主义解释虽然在逻辑上站得住脚,但我认为不如前两种解释更有说服力。各国可能无法找到一个和平
settlement both prefer due to issue indivisibilities. Perhaps some issues, by their very natures, simply will not admit compromise. Though neither example is wholly convincing, issues that might exhibit indivisibility include abortion in domestic politics and the problem of which prince sits on the throne of, say, Spain, in eighteenth- or nineteenth-century international politics. Issue indivisibility could in principle make war rational for the obvious reason that if the issue allows only a finite number of resolutions, it might be that none falls within the range that both prefer to fighting. However, the issues over which states bargain typically are complex and multidimensional; side-payments or linkages with other issues typically are possible; and in principle states could alternate or randomize among a fixed number of possible solutions to a dispute. War-prone international issues may often be effectively indivisible, but the cause of this indivisibility lies in domestic political and other mechanisms rather than in the nature of the issues themselves. 由于问题的不可分割性,两者都倾向于和解。也许有些问题,就其本质而言,根本无法达成妥协。虽然这两个例子都不完全令人信服,但可能表现出不可分割性的问题包括国内政治中的堕胎问题,以及十八或十九世纪国际政治中的西班牙王位归属问题。问题的不可分割性原则上可以使战争变得合理,原因显而易见,如果问题只允许有限数量的解决方案,那么可能没有一个解决方案是在双方都倾向于战斗的范围内。然而,各国讨价还价的问题通常是复杂的、多层面的;附带利益或与其他问题的联系通常是可能的;原则上,各国可以在解决争端的固定数量的可能方案中交替或随机选择。容易引发战争的国际问题往往实际上是不可分割的,但造成这种不可分割性的原因在于国内政治和其他机制,而不是问题本身的性质。
In the first section of the article I discuss the puzzle posed by the fact that war is costly. Using a simple formalization of the bargaining problem faced by states in conflict, I show that under very broad conditions bargains will exist that genuinely rational states would prefer to a risky and costly fight. The second section argues that rational miscalculations of relative power and resolve must be due to private information and then considers how war may result from the combination of private information and incentives to misrepresent that information in bargaining. In the third section, I discuss commitment problems as the second class of defensible rationalist explanations for war. Throughout, I specify theoretical arguments with simple game-theoretic representations and assess plausibility with historical examples. 在文章的第一部分,我讨论了战争代价高昂这一事实所带来的困惑。通过对冲突中的国家所面临的讨价还价问题进行简单的形式化,我证明了在非常宽泛的条件下,真正理性的国家会选择讨价还价,而不是冒险且代价高昂的战争。第二部分论证了对相对实力和决心的理性误判一定是由私人信息造成的,然后探讨了私人信息和在讨价还价中歪曲信息的动机如何共同导致战争。在第三节中,我将讨论承诺问题,将其作为对战争的第二类站得住脚的理性主义解释。在整个过程中,我用简单的博弈论表述来具体说明理论论点,并用历史实例来评估其合理性。
Before beginning, I should make it clear that I am not presenting either commitment problems or private information and incentives to misrepresent as wholly novel explanations for war that are proposed here for the first time. The literature on the causes of war is massive, and these ideas, mixed with myriad others, can be found in it in various guises. The main theoretical task facing students of war is not to add to the already long list of arguments and conjectures but instead to take apart and reassemble these diverse arguments into a coherent theory fit for guiding empirical research. Toward this end, I am arguing that when one looks carefully at the problem of explaining how war could occur between genuinely rational, unitary states, one finds that there are really only two ways to do it. The diverse rationalist or neorealist explanations commonly found in the literature fail for two reasons. First, many do not even address the relevant question-what prevents states from locating a bargain both sides would prefer to a fight? They do not address the question because it is widely but incorrectly assumed that rational states can face a situation of deadlock, wherein no agreements exist that both sides would prefer to a war. ^(4){ }^{4} 在开始之前,我需要说明的是,我既没有提出承诺问题,也没有提出私人信息和动机失实问题,这些都是我首次提出的关于战争的全新解释。关于战争原因的文献浩如烟海,这些观点以及无数其他观点以各种面目出现在文献中。研究战争的学生所面临的主要理论任务不是在已经很长的论点和猜想清单上再添加新的论点和猜想,而是将这些不同的论点拆分开来,重新组合成一个适合指导实证研究的连贯理论。为此,我认为,当我们仔细研究如何解释战争如何在真正理性的单一国家之间发生的问题时,我们会发现实际上只有两种方法可以做到这一点。文献中常见的各种理性主义或新现实主义解释之所以失败,原因有二。首先,许多解释甚至都没有涉及相关问题--是什么阻碍了国家找到一个双方都倾向于的讨价还价方案,而不是一场争斗?他们之所以没有回答这个问题,是因为人们普遍错误地认为,理性的国家可能会面临僵局,即不存在双方都希望达成的协议,而不是战争。 ^(4){ }^{4}
Second, the rationalist arguments that do address the question-such as (4) and (5) above-do not go far enough in answering it. When fully developed, they prove to be one of the two major mechanisms developed here, namely, either a commitment problem or a problem arising from private information and incentives to misrepresent. These two mechanisms, I will argue, provide the foundations for a rationalist or neorealist theory of war. 其次,理性主义的论证确实涉及到了这一问题--例如上文(4)和(5)--但在回答这一问题时还不够深入。当这些论证得到充分发展时,它们被证明是本文提出的两大机制之一,即要么是承诺问题,要么是由私人信息和虚假陈述动机引起的问题。我认为,这两种机制为理性主义或新现实主义战争理论提供了基础。
The puzzle 谜题
Most historians and political scientists who study war dismiss as naive the view that all wars must be unwanted because they entail destruction and suffering. Instead, most agree that while a few wars may have been unwanted by the leaders who brought them about-World War I is sometimes given as an example-many or perhaps most wars were simply wanted. The leaders involved viewed war as a costly but worthwhile gamble. ^(5){ }^{5} 大多数研究战争的历史学家和政治学家都认为,所有战争都是不受欢迎的,因为它们会带来破坏和痛苦,这种观点太天真了。相反,大多数人都认为,虽然少数战争可能是发动战争的领导人所不希望发生的--第一次世界大战有时就是一个例子--但许多战争,甚至大多数战争都是想要发生的。有关领导人认为战争是一场代价高昂但值得一赌的赌博。 ^(5){ }^{5}
Moreover, many scholars believe that wanted wars are easily explained from a rationalist perspective. Wanted wars are thought to be Pareto-efficient-they occur when no negotiated settlements exist that both sides would prefer to the gamble of military conflict. Conventional wisdom holds that while this situation may be tragic, it is entirely possible between states led by rational leaders who consider the costs and risks of fighting. Unwanted wars, which take place despite the existence of settlements both sides preferred to conflict, are thought to pose more of a puzzle, but one that is resolvable and also fairly rare. 此外,许多学者认为,从理性主义的角度来看,通缉战争很容易解释。被通缉的战争被认为是帕累托效率的战争--当双方都不愿意通过谈判解决问题,而不愿意冒军事冲突的风险时,战争就会发生。传统智慧认为,虽然这种情况可能是悲剧性的,但在由理性领导人领导的国家之间是完全可能发生的,因为他们会考虑战斗的成本和风险。不希望发生的战争,即尽管存在双方都希望的解决办法,但却发生了,这被认为是一个更大的难题,但却是可以解决的,而且也相当罕见。
The conventional distinction between wanted and unwanted wars misunderstands the puzzle posed by war. The reason is that the standard conception does not distinguish between two types of efficiency-ex ante and ex post. As long as both sides suffer some costs for fighting, then war is always inefficient exe x post-both sides would have been better off if they could have achieved the same final resolution without suffering the costs (or by paying lower costs). This is true even if the costs of fighting are small, or if one or both sides viewed the potential benefits as greater than the costs, since there are still costs. Unless states enjoy the activity of fighting for its own sake, as a consumption good, then war is inefficient ex post. 传统上对想要的战争和不想要的战争的区分误解了战争带来的难题。原因在于,标准概念没有区分事前和事后两种效率。只要交战双方都要付出一定的代价,那么战争在事后总是无效率的 exe x --如果双方都能在不付出代价(或付出较低代价)的情况下达成同样的最终解决方案,那么双方都会过得更好。即使战争的成本很小,或者一方或双方认为潜在的利益大于成本,情况也是如此,因为仍然存在成本。除非国家喜欢为战斗而战斗,将其作为一种消费品,否则战争事后就是无效率的。
From a rationalist perspective, the central puzzle about war is precisely this ex post inefficiency. Before fighting, both sides know that war will entail some costs, and even if they expect offsetting benefits they still have an incentive to avoid the costs. The central question, then, is what prevents states in a dispute 从理性主义的角度来看,战争的核心难题正是这种事后的无效率。在开战之前,双方都知道战争会带来一些代价,即使他们预期会有抵消的利益,他们仍然有动机避免代价。因此,核心问题是,是什么阻止了争端中的国家
from reaching an ex ante agreement that avoids the costs they know will be paid ex post if they go to war? Giving a rationalist explanation for war amounts to answering this question. 达成一项事前协议,以避免他们知道一旦开战就会付出事后代价?对战争做出理性主义的解释就等于回答了这个问题。
Three of the most common and widely employed rationalist arguments in the literature do not directly address or answer the question. These are arguments from anarchy, preventive war, and positive expected utility. 文献中最常见和最广泛使用的三个理性主义论证并没有直接涉及或回答这个问题。它们是无政府状态论证、预防性战争论证和正期望效用论证。
Anarchy 无政府状态
Since Kenneth Waltz’s influential Man, the State, and War, the anarchical nature of the international realm is routinely cited as a root cause of or explanation for the recurrence of war. Waltz argued that under anarchy, without a supranational authority to make and enforce law, “war occurs because there is nothing to prevent it. . . . Among states as among men there is no automatic adjustment of interests. In the absence of a supreme authority there is then the constant possibility that conflicts will be settled by force.” ^(6){ }^{6} 自肯尼思-华尔兹(Kenneth Waltz)发表了影响深远的《人、国家与战争》(Man, the State, and War)一书以来,国际领域的无政府主义性质经常被作为战争频发的根本原因或解释。华尔兹认为,在无政府状态下,没有超国家的权威来制定和执行法律,"战争之所以发生,是因为没有任何东西可以阻止它。. . .......国家之间就像人与人之间一样,利益不会自动调整。在没有最高权威的情况下,冲突就始终有可能通过武力来解决"。 ^(6){ }^{6}
The argument focuses our attention on a fundamental difference between domestic and international politics. Within a well-ordered state, organized violence as a strategy is ruled out-or at least made very dangerous-by the potential reprisals of a central government. In international relations, by contrast, no agency exists that can credibly threaten reprisal for the use of force to settle disputes. ^(7){ }^{7} The claim is that without such a credible threat, war will sometimes appear the best option for states that have conflicting interests. 这一论点将我们的注意力集中在国内政治与国际政治的根本区别上。在一个秩序井然的国家中,有组织的暴力作为一种策略是被排除在外的,或者至少会因为中央政府的潜在报复而变得非常危险。与此相反,在国际关系中,没有任何机构可以令人信服地威胁要对使用武力解决争端的行为进行报复。 ^(7){ }^{7} 这种说法认为,如果没有这种可信的威胁,战争有时就会成为利益冲突国的最佳选择。
While I do not doubt that the condition of anarchy accounts for major differences between domestic and international politics, and that anarchy encourages both fear of and opportunities for military conflict, the standard framing of the argument is not enough to explain why wars occur and recur. Under anarchy, nothing stops states from using force if they wish. But if using force is a costly option regardless of the outcome, then why is it ever employed? How exactly does the lack of a central authority prevent states from negotiating agreements both sides would prefer to fighting? As it is typically stated, the argument that anarchy provides a rationalist explanation for war does not address this question and so does not solve the problem posed by war’s ex post inefficiency. 虽然我并不怀疑无政府状态造成了国内政治和国际政治之间的重大差异,而且无政府状态既助长了对军事冲突的恐惧,也助长了军事冲突的机会,但这一论点的标准框架并不足以解释战争为什么会发生和反复发生。在无政府状态下,如果国家愿意,没有什么能阻止它们使用武力。但是,如果无论结果如何,使用武力都是代价高昂的选择,那么为什么还会使用武力呢?缺乏中央权威究竟是如何阻止国家通过谈判达成双方都希望达成的协议,而不是通过战争呢?正如通常所说的那样,无政府状态为战争提供了理性主义的解释这一论点并没有解决这个问题,因此也没有解决战争的事后无效率问题。
Neither, it should be added, do related arguments invoking the security dilemma, the fact that under anarchy one state’s efforts to make itself more secure can have the undesired but unavoidable effect of making another state 应该补充的是,援引安全困境的相关论点也是如此,即在无政府状态下,一个国家为使自身更加安全而做出的努力可能会产生不希望但却不可避免的结果,使另一个国家变得更加安全。
6. The quotation is drawn from Kenneth Waltz, Man, the State, and War: A Theoretical Analysis (New York: Columbia University Press, 1959), p. 188. 6.引文摘自肯尼斯-华尔兹,《人、国家与战争:理论分析》(纽约:哥伦比亚大学出版社,1959 年),第 188 页。
7. For a careful analysis and critique of this standard argument on the difference between the international and domestic arenas, see R. Harrison Wagner, “The Causes of Peace,” in Roy A. Licklider, ed., Stopping the Killing: How Civil Wars End (New York: New York University Press, 1993), pp. 235-68 and especially pp. 251-57. 关于对国际舞台和国内舞台之间差异的标准论点的仔细分析和批评,见 R. Harrison Wagner,"The Causes of Peace",载于 Roy A. Licklider 编辑的《停止杀戮:内战如何结束》(纽约:纽约大学出版社,1993 年),第 235-68 页,特别是第 25-57 页:内战如何结束》(纽约:纽约大学出版社,1993 年),第 235-68 页,尤其是第 251-57 页。
less secure. ^(8){ }^{8} By itself this fact says nothing about the availability or feasibility of peaceful bargains that would avoid the costs of war. More elaborate arguments are required, and those that are typically given do not envision bargaining and do not address the puzzle of costs. Consider, for instance, a spiral scenario in which an insecure state increases its arms, rendering another so insecure that it decides to attack. If the first state anticipated the reaction producing war, then by itself this is a deadlock argument; I argue against these below. If the first state did not anticipate war and did not want it, then the problem would seem to be miscalculation rather than anarchy, and we need to know why signaling and bargaining could not have solved it. As Robert Jervis has argued, anarchy and the security dilemma may well foster arms races and territorial competition. ^(9){ }^{9} But with the exception of occasional references to the preemptive war problem, the standard security dilemma arguments do not explicitly address the question of why the inability to make commitments should necessarily make for war between rational states. ^(10){ }^{10} 更不安全。 ^(8){ }^{8} 这一事实本身并没有说明可以避免战争成本的和平谈判是否存在或可行。我们需要更详尽的论证,而那些通常给出的论证并没有设想讨价还价,也没有解决成本的难题。例如,考虑一个螺旋式上升的情景:一个不安全的国家增加军备,使另一个国家变得如此不安全,以至于决定发动攻击。如果第一个国家预料到了会产生战争的反应,那么这本身就是一个僵局论证;我将在下文反驳这些论证。如果第一个国家没有预料到战争,也不希望战争发生,那么问题似乎就不是无政府状态,而是误判,我们需要知道为什么信号传递和讨价还价不能解决这个问题。正如罗伯特-杰维斯(Robert Jervis)所言,无政府状态和安全困境很可能会助长军备竞赛和领土竞争。 ^(9){ }^{9} 但是,除了偶尔提到先发制人的战争问题之外,标准的安全困境论证并没有明确解决为什么无法做出承诺必然会导致理性国家之间发生战争的问题。 ^(10){ }^{10}
Below I will argue that anarchy is indeed implicated as a cause of specific sorts of military conflict (e.g., preventive and preemptive war and in some cases war over strategic territory). In contrast to the standard arguments, however, showing how anarchy figures in a coherent rationalist explanation entails describing the specific mechanism by which states’ inability to write enforceable contracts makes peaceful bargains both sides would prefer unattainable. 下面,我将论证无政府状态确实是特定军事冲突(如预防性战争和先发制人的战争,以及在某些情况下争夺战略领土的战争)的原因之一。然而,与标准论点不同的是,要说明无政府状态如何体现在一个连贯的理性主义解释中,就必须描述国家无法书写可执行契约的具体机制,从而使双方都希望达成的和平协议无法实现。
Preventive war 预防性战争
It frequently is argued that if a declining power expects it might be attacked by a rising power in the future, then a preventive war in the present may be rational. Typically, however, preventive war arguments do not consider whether the rising and declining powers could construct a bargain, perhaps across time, that would leave both sides better off than a costly and risky preventive war would. ^(11){ }^{11} The incentives for such a deal surely exist. The rising state should not want to be attacked while it is relatively weak, so what stops it from offering concessions in the present and the future that would make the declining state prefer not to attack? Also, if war is inefficient and bargains both sides prefer to a fight will exist, why should the declining power rationally fear being attacked in the future? The standard argument supposes that an 经常有人认为,如果衰落的大国预计未来可能受到崛起大国的攻击,那么在当前发动预防性战争可能是合理的。然而,预防性战争的论点通常不会考虑崛起的大国和衰落的大国是否可以达成一项交易,也许是跨时间的交易,从而使双方都能获得比代价高昂、风险巨大的预防性战争更好的结果。 ^(11){ }^{11} 这种交易的动机肯定是存在的。崛起的国家应该不希望在自己相对弱小的时候受到攻击,那么有什么能阻止它在现在和未来做出让步,让衰落的国家宁愿不进攻呢?另外,如果战争是低效的,而且双方都倾向于讨价还价,那么为什么衰落国在理性上会害怕在未来受到攻击呢?标准论证假设
anticipated shift in the balance of power can by itself be enough to make war rational, but this is not so. 力量对比的预期变化本身就足以使战争变得合理,但事实并非如此。
Positive expected utility 正预期效用
Perhaps the most common informal rationalist explanation found in the literature is that war may occur when two states each estimate that the expected benefits of fighting outweigh the expected costs. As Bruce Bueno de Mesquita argued in an influential formalization of this claim, war can be rational if both sides have positive expected utility for fighting; that is, if the expected utility of war (expected benefits less costs) is greater than the expected utility of remaining at peace. ^(12){ }^{12} 文献中最常见的非正式理性主义解释或许是,当两个国家都认为战争的预期收益大于预期成本时,战争就可能发生。正如布鲁斯-布埃诺-德-梅斯基塔(Bruce Bueno de Mesquita)在对这一主张进行有影响力的形式化时所论证的,如果双方对战争都有正的预期效用,战争就可能是理性的;也就是说,如果战争的预期效用(预期收益减去成本)大于保持和平的预期效用,战争就可能是理性的。 ^(12){ }^{12}
Informal versions of the expected utility argument typically fail to address the question of how or under what conditions it can be possible for two states both to prefer the costly gamble of war to any negotiated settlement. Formal versions have tended to avoid the question by making various restrictive and sometimes nonrationalist assumptions. To support these claims, I need to be more precise about the expected utility argument. 预期效用论证的非正式版本通常无法解决这样一个问题:两个国家如何或在什么条件下都宁愿选择代价高昂的战争而不是任何谈判解决。正式版本往往通过各种限制性的、有时是非理性主义的假设来回避这个问题。为了支持这些主张,我需要更精确地说明预期效用论证。
When will there exist bargains both sides prefer to war? 什么时候才会有双方都喜欢的交易,而不是战争?
This section considers the question of whether and when two rationally led states could both prefer war to any negotiated settlement. 本节探讨的问题是,两个理性主导的国家是否以及何时都会选择战争而非谈判解决。
Consider two states, A and B, who have preferences over a set of issues represented by the interval X=[0,1]X=[0,1]. State A prefers issue resolutions closer to 1 , while B prefers outcomes closer to 0 . Let the states’ utilities for the outcome x in Xx \in X be u_(A)(x)u_{A}(x) and u_(B)(1-x)u_{B}(1-x), and assume for now that u_(A)(*)u_{A}(\cdot) and u_(B)(*)u_{B}(\cdot) are continuous, increasing, and weakly concave (that is, risk-neutral or risk-averse). Without losing any generality, we can set u_(i)(1)=1u_{i}(1)=1 and u_(i)(0)=0u_{i}(0)=0 for both states (i=A,B)(i=A, B). For concreteness we might think of xx as representing the proportion of all territory between A and B that is controlled by A . 考虑两个国家 A 和 B,它们对区间 X=[0,1]X=[0,1] 所代表的一组问题具有偏好。A 国偏好接近 1 的问题决议,而 B 国偏好接近 0 的结果。假设各州对结果 x in Xx \in X 的效用分别为 u_(A)(x)u_{A}(x) 和 u_(B)(1-x)u_{B}(1-x) ,并暂时假设 u_(A)(*)u_{A}(\cdot) 和 u_(B)(*)u_{B}(\cdot) 是连续、递增和弱凹(即风险中性或风险规避)的。在不失一般性的前提下,我们可以为 (i=A,B)(i=A, B) 两种状态设置 u_(i)(1)=1u_{i}(1)=1 和 u_(i)(0)=0u_{i}(0)=0 。具体来说,我们可以把 xx 视为 A 和 B 之间由 A 控制的所有领土的比例。
In order to say whether the set XX contains negotiated settlements that both sides would prefer to conflict, it must be possible to say how the states evaluate the military option versus those outcomes. Almost all analysts of war have 要想知道 XX 集合中是否包含双方都希望通过谈判达成的解决方案,就必须知道各国是如何评估军事选择与这些结果之间的关系的。几乎所有的战争分析家都
FIGURE 1. The bargaining range 图 1.谈判范围
stressed that war is a gamble whose outcome may be determined by random or otherwise unforeseeable events. ^(13){ }^{13} As Bueno de Mesquita argued, this makes expected utility a natural candidate. ^(14){ }^{14} Suppose that if the states fight a war, state A prevails with probability p in[0,1]p \in[0,1], and that the winner gets to choose its favorite outcome in the issue space. It follows that A’s expected utility for war is pu_(A)(1)+(1-p)u_(A)(0)-c_(A)p u_{A}(1)+(1-p) u_{A}(0)-c_{A}, or p-c_(A)p-c_{A}, where c_(A)c_{A} is state A’s utility for the costs of a war. Similarly, state B’s expected utility for war will be 1-p-c_(B)1-p-c_{B}. Since we are considering rationalist theories for war, we assume that c_(A)c_{A} and c_(B)c_{B} are both positive. War is thus represented as a costly lottery. (Note that in this formulation the terms c_(A)c_{A} and c_(B)c_{B} capture not only the states’ values for the costs of war but also the value they place on winning or losing on the issues at stake. That is, c_(A)c_{A} reflects state A’s costs for war relative to any possible benefits. For example, if the two states see little to gain from winning a war against each other, then c_(A)c_{A} and c_(B)c_{B} would be large even if neither side expected to suffer much damage in a war.) 他强调,战争是一场赌博,其结果可能由随机事件或其他不可预见的事件决定。 ^(13){ }^{13} 正如布埃诺德梅斯基塔所言,这使得预期效用成为一个自然的候选方案。 ^(14){ }^{14} 假设如果各国开战,A国获胜的概率为 p in[0,1]p \in[0,1] ,胜者可以在问题空间中选择自己最喜欢的结果。因此,A 国对战争的预期效用为 pu_(A)(1)+(1-p)u_(A)(0)-c_(A)p u_{A}(1)+(1-p) u_{A}(0)-c_{A} ,或 p-c_(A)p-c_{A} ,其中 c_(A)c_{A} 是 A 国对战争成本的效用。同样,B 国对战争的预期效用将是 1-p-c_(B)1-p-c_{B} 。由于我们考虑的是理性主义的战争理论,我们假定 c_(A)c_{A} 和 c_(B)c_{B} 都是正数。因此,战争被表述为代价高昂的抽签。(请注意,在这一表述中, c_(A)c_{A} 和 c_(B)c_{B} 不仅反映了国家对战争成本的估价,也反映了国家对战争输赢的估价。也就是说, c_(A)c_{A} 反映的是A国的战争成本相对于任何可能的利益。例如,如果两个国家都认为赢得战争对彼此没有什么好处,那么即使双方都不希望在战争中遭受太大损失, c_(A)c_{A} 和 c_(B)c_{B} 也会很大。)
We can now answer the question posed above. The following result is easily demonstrated: given the assumptions stated in the last two paragraphs, there always exists a set of negotiated settlements that both sides prefer to fighting. ^(15){ }^{15} Formally, there exists a subset of XX such that for each outcome xx in this set, u_(A)(x) > p-c_(A)u_{A}(x)>p-c_{A} and u_(B)(1-x) > 1-p-c_(B)u_{B}(1-x)>1-p-c_{B}. For example, in the risk-neutral case where u_(A)(x)=xu_{A}(x)=x and u_(B)(1-x)=1-xu_{B}(1-x)=1-x, both states will strictly prefer any peaceful agreement in the interval (p-c_(A),p+c_(B))\left(p-c_{A}, p+c_{B}\right) to fighting. This interval represents the bargaining range, with p-c_(A)p-c_{A} and p+c_(B)p+c_{B} as the reservation levels that delimit it. A risk-neutral case is depicted in Figure 1. 现在我们可以回答上文提出的问题了。下面的结果很容易证明:在上两段所述假设的前提下,总是存在一组双方都喜欢的谈判解决方案,而不是战斗。 ^(15){ }^{15} 形式上,存在一个 XX 子集,使得这个集合中的每个结果 xx 都是 u_(A)(x) > p-c_(A)u_{A}(x)>p-c_{A} 和 u_(B)(1-x) > 1-p-c_(B)u_{B}(1-x)>1-p-c_{B} 。例如,在 u_(A)(x)=xu_{A}(x)=x 和 u_(B)(1-x)=1-xu_{B}(1-x)=1-x 的风险中性情况下,两国都会严格倾向于 (p-c_(A),p+c_(B))\left(p-c_{A}, p+c_{B}\right) 区间内的任何和平协议,而不是战斗。这个区间代表讨价还价的范围, p-c_(A)p-c_{A} 和 p+c_(B)p+c_{B} 是划分这个区间的保留水平。风险中性的情况如图 1 所示。
This simple but important result is worth belaboring with some intuition. Suppose that two people (or states) are bargaining over the division of $100-if they can agree on a split they can keep what they agree to. However, in contrast 这个简单而重要的结果值得用一些直觉来加以阐述。假设有两个人(或两个国家)正在就 100 美元的分割问题讨价还价--如果他们能就分割问题达成一致,他们就能保住他们同意的东西。然而,与此相反
13. See, for classic examples, Thucydides, The Peloponnesian War (New York: Modern Library, 1951), pp. 45 and 48; and Carl von Clausewitz, On War (Princeton, N.J.: Princeton University Press, 1984), p. 85. 经典案例见修昔底德,《伯罗奔尼撒战争》(纽约:现代图书馆,1951 年),第 45 和 48 页;卡尔-冯-克劳塞维茨,《战争论》(新泽西州普林斯顿:普林斯顿大学出版社,1984 年),第 85 页。
14. Bueno de Mesquita, The War Trap. 14.Bueno de Mesquita,《战争陷阱》。
15. A proof is given in the Appendix. 15.证明见附录。
to the usual economic scenarios, in this international relations example the players also have an outside option. ^(16){ }^{16} For a price of $20\$ 20, they can go to war, in which case each player has a 50 -percent chance of winning the whole $100\$ 100. This implies that the expected value of the war option is $30(0.5*100+0.5*0-20)\$ 30(0.5 \cdot 100+0.5 \cdot 0-20) for each side, so that if the players are risk-neutral, then neither should be willing to accept less than $30\$ 30 in the bargaining. But notice that there is still a range of peaceful, bargained outcomes from ($31,$69)(\$ 31, \$ 69) to ($69,$31)(\$ 69, \$ 31) that make both sides strictly better off than the war option. Risk aversion will tend to increase the range yet further; indeed, even if the leaders pay no costs for war, a set of agreements both sides prefer to a fight will still exist provided both are risk-averse over the issues. In effect, the costs and risks of fighting open up a “wedge” of bargained solutions that risk-neutral or risk-averse states will prefer to the gamble of conflict. The existence of this ex ante bargaining range derives from the fact that war is inefficient ex post. 与通常的经济情景不同,在这个国际关系示例中,玩家还有一个外部选择。 ^(16){ }^{16} 只要付出 $20\$ 20 的代价,他们就可以开战,在这种情况下,每一方都有 50%的机会赢得整个 $100\$ 100 。这意味着战争选项的期望值对每一方来说都是 $30(0.5*100+0.5*0-20)\$ 30(0.5 \cdot 100+0.5 \cdot 0-20) ,因此,如果博弈者是风险中性的,那么在讨价还价中,双方都不应该愿意接受低于 $30\$ 30 的价格。但是请注意,从 ($31,$69)(\$ 31, \$ 69) 到 ($69,$31)(\$ 69, \$ 31) ,仍然存在一系列和平的、讨价还价的结果,这些结果使双方的境况都严格优于战争选项。风险规避倾向于进一步扩大这个范围;事实上,即使领导人不为战争付出任何代价,只要双方都在这些问题上规避风险,那么双方都更喜欢的协议集合仍然存在。实际上,战争的成本和风险为讨价还价的解决方案开辟了一个 "楔子",风险中性或风险规避型国家更愿意选择这种解决方案,而不是冒险冲突。这种事前讨价还价范围的存在源于战争事后效率低下这一事实。
Three substantive assumptions are needed for the result, none of which seems particularly strong. First, the states know that there is some true probability pp that one state would win in a military contest. As discussed below, it could be that the states have conflicting estimates of the likelihood of victory, and if both sides are optimistic about their chances this can obscure the bargaining range. But even if the states have private and conflicting estimates of what would happen in a war, if they are rational, they should know that there can be only one true probability that one or the other will prevail (perhaps different from their own estimate). Thus rational states should know that there must in fact exist a set of agreements all prefer to a fight. 这一结果需要三个实质性假设,其中没有一个似乎特别有力。首先,各国都知道,在军事较量中,一国获胜的概率 pp 是真实存在的。正如下文所讨论的,各国对获胜可能性的估计可能存在冲突,如果双方都对自己的胜算持乐观态度,就会模糊讨价还价的范围。但是,即使各国对战争中会发生什么有私人的、相互矛盾的估计,如果它们是理性的,它们就应该知道,一方或另一方获胜的真实概率只有一个(可能与它们自己的估计不同)。因此,理性的国家应该知道,事实上一定存在一套协议,所有协议都倾向于战争。
Second, it is assumed that the states are risk-averse or risk-neutral over the issues. Because risk attitude is defined relative to an underlying metric (such as money in economics), the substantive meaning of this assumption depends on the bargaining context. Loosely, it says that the states prefer a fifty-fifty split or share of whatever is at issue (in whatever metric it comes, if any) to a fifty-fifty chance at all or nothing, where this refers to the value of winning or losing a war. In effect, the assumption means that leaders do not like gambling when the downside risk is losing at war, which seems plausible given the presumption that state leaders normally wish to retain territory and power. A risk-acceptant leader is analogous to a compulsive gambler-willing to accept a sequence of gambles that has the expected outcome of eliminating the state and regime. Even if we admitted such a leader as rational, it seems doubtful that many have held such preferences (Hitler being a possible exception). 其次,假设各国在这些问题上都是规避风险或风险中性的。由于风险态度的定义是相对于一个基本指标(如经济学中的货币)而言的,因此这一假设的实质含义取决于讨价还价的背景。从广义上讲,它指的是国家更倾向于五五分成或分享有争议的东西(无论以何种标准衡量,如果有的话),而不是五五分成的机会,这里的五五分成指的是战争输赢的价值。实际上,这一假设意味着,当战争失败的负面风险时,领导人不喜欢赌博,鉴于国家领导人通常希望保住领土和权力,这一假设似乎是可信的。接受风险的领导人类似于强迫性赌徒--愿意接受一连串的赌博,其预期结果是消灭国家和政权。即使我们承认这样的领导人是理性的,但是否有很多人持有这样的偏好(希特勒可能是个例外)似乎也值得怀疑。
Finally, it was assumed that a continuous range of peaceful settlements (from 0 to 1) exists. In other words, the issues in dispute are perfectly divisible, so that there are always feasible bargains between the states’ reservation levels p-c_(A)p-c_{A} and p+c_(B)p+c_{B}. This third assumption immediately suggests a tenable rationalist explanation for war. Perhaps something about the nature of some international issues, such as which successor will sit on a throne, does not admit finely graded divisions and compromise. If so, then small costs for fighting and bad luck may make for rational war over such issues. 最后,假设存在一个连续的和平解决范围(从 0 到 1)。换句话说,有争议的问题是完全可分的,因此在各国的保留水平 p-c_(A)p-c_{A} 和 p+c_(B)p+c_{B} 之间总是存在可行的讨价还价。第三个假设立即为战争提供了一个合理的解释。也许某些国际问题的本质,比如哪位继承人将坐上王位,并不允许精细分级的分歧和妥协。如果是这样,那么小的战斗成本和坏运气可能会使人们在这些问题上理性地进行战争。
But we would immediately like to know what about the nature of an issue makes it impossible to divide up. On more thought, this seems empirically implausible. In the first place, most issues states negotiate over are quite complex-they have many dimensions of concern and allow many possible settlements. Second, if states can simply pay each other sums of money or goods (which they can, in principle), or make linkages with other issues, then this should have the effect of making any issues in dispute perfectly divisible. Before the age of nationalism, princes often bought, sold, and partitioned land. ^(17){ }^{17} In the nineteenth century the United States purchased the Louisiana territory from France, and Alaska from Russia, and as late as 1898 President McKinley explored the possibility of buying Cuba from Spain in order to avoid a war over it. ^(18){ }^{18} Third, if something about the nature of an issue means that it can be settled in only, say, two ways, then some sort of random allocation or alternation between the two resolutions could in principle serve to create intermediate bargains. Mafia dons, for example, apparently have avoided costly internal wars by using lotteries to allocate construction contracts among families. ^(19){ }^{19} 但是,我们马上就想知道,问题的本质是什么使得它无法分割。再仔细一想,这在经验上似乎不太可信。首先,大多数国家谈判的问题都相当复杂--它们涉及许多方面,有许多可能的解决方案。其次,如果各国可以简单地相互支付金钱或货物(原则上可以),或与其他问题挂钩,那么这就会使任何有争议的问题变得完全可分。在民族主义时代之前,王公们经常买卖和分割土地。 ^(17){ }^{17} 19世纪,美国从法国购买了路易斯安那领土,从俄国购买了阿拉斯加,直到1898年,麦金利总统还在探讨从西班牙购买古巴的可能性,以避免为古巴爆发战争。 ^(18){ }^{18} 第三,如果一个问题的性质意味着它只能通过两种方式来解决,那么某种随机分配或两种解决方案之间的交替原则上可以用来创造中间交易。例如,黑手党的头目们显然通过在家族间分配建筑合同的抽签方式避免了代价高昂的内战。 ^(19){ }^{19} .
In practice, creating intermediate settlements with cash, with linkages to other issues, or with randomization or alternation often seems difficult or impossible for states engaged in a dispute. For example, the immediate issue that gave rise to the Franco-Prussian war was a dispute over which prince would take the Spanish throne. It doubtless occurred to no one to propose that the two candidates alternate year by year, or three years for the Hapsburg and one for the Hohenzollern, or whatever. In this case as in many others, the issue could in principle have been made more continuous and was not for other reasons-here, alternating kings would have violated so many conventions and norms as to have been domestically unworkable. To give a more realistic 在实践中,对于卷入争端的国家来说,用现金、与其他问题的联系、随机化或交替的方式来建立中间解决方案往往显得困难或不可能。例如,引起普法战争的直接问题是关于由哪位王子继承西班牙王位的争端。毫无疑问,没有人提议让两位候选人逐年交替,或者哈布斯堡王朝三年一任,霍亨索伦王朝三年一任等等。在这种情况下,正如在其他许多情况下一样,原则上可以使问题更具连续性,但由于其他原因而没有这样做--在这种情况下,国王轮流制会违反许多惯例和规范,在国内是行不通的。为了更切合实际
example, nineteenth- and twentieth-century leaders cannot divide up and trade territory in international negotiations as easily as could rulers in the seventeenth and eighteenth centuries, due in part to domestic political consequences of the rise of nationalism; contrast, for example, the Congress of Vienna with the negotiations following World War I. 例如,19 世纪和 20 世纪的领导人在国际谈判中不能像 17 世纪和 18 世纪的统治者那样轻易地划分和交易领土,部分原因是民族主义兴起造成的国内政治后果;例如,维也纳会议与第一次世界大战后的谈判形成鲜明对比。
So in principle the indivisibility of the issues that are the subject of international bargaining can provide a coherent rationalist explanation for war. However, the real question in such cases is what prevents leaders from creating intermediate settlements, and the answer is likely to be other mechanisms (often domestic political) rather than the nature of the issues themselves. ^(20){ }^{20} Both the intrinsic complexity and richness of most matters over which states negotiate and the availability of linkages and side-payments suggest that intermediate bargains typically will exist. 因此,原则上,作为国际谈判主题的问题的不可分割性可以为战争提供一个连贯的理性主义解释。然而,在这种情况下,真正的问题在于是什么阻碍了领导人达成中间解决方案,而答案很可能是其他机制(通常是国内政治机制),而不是问题本身的性质。 ^(20){ }^{20} 国家谈判所涉及的大多数问题的内在复杂性和丰富性,以及联系和附带报酬的存在,都表明中间交易通常会存在。
It is thus not sufficient to say that positive expected utility by itself supplies a coherent or compelling rationalist explanation for war. Provided that the issues in dispute are sufficiently divisible, or that side-payments are possible, there should exist a set of negotiated agreements that have greater utility for both sides than the gamble of war does. The reason is that the ex post inefficiency of war opens up an ex ante bargaining range. 因此,仅仅说正期望效用本身为战争提供了一个连贯或令人信服的理性主义解释是不够的。只要争论的问题有足够的可分割性,或者说附带报酬是可能的,那么就应该存在一组谈判达成的协议,这些协议对双方的效用都要大于战争的赌博效用。原因在于,战争的事后无效率开辟了一个事前讨价还价的范围。
So, to explain how war could occur between rationally led states, we need to answer the following question. Given the existence of an ex ante bargaining range, why might states fail either to locate or to agree on an outcome in this range, so avoiding the costs and risks of war? 因此,要解释理性主导的国家之间为何会发生战争,我们需要回答以下问题。既然存在一个事前讨价还价的范围,为什么各国可能无法在这个范围内找到或商定一个结果,从而避免战争的代价和风险?
War due to private information and incentives to misrepresent 由于私人信息和虚报动机而引发的战争
Two commonly employed rationalist explanations in the literature directly address the preceding question. Both turn on the claim that war can be and often is the product of rational miscalculation. One explanation holds that a state’s leaders may rationally overestimate their chance of military victory against an adversary, so producing a disagreement about relative power that only war can resolve. The other argues that rationally led states may lack information about an adversary’s willingness to fight over some interest and so may challenge in the mistaken belief that war will not follow. 文献中两种常用的理性主义解释直接涉及了上述问题。这两种解释都基于这样一种主张,即战争可能是而且往往是理性误判的产物。一种解释认为,一个国家的领导人可能会理性地高估自己在军事上战胜对手的机会,从而产生只有战争才能解决的相对实力分歧。另一种解释则认为,理性领导的国家可能缺乏关于对手是否愿意为某些利益而战的信息,因此可能误以为战争不会随之而来,从而提出挑战。
In this section I argue that while these ideas point toward a tenable rationalist explanation for war, neither goes far enough and neither works by itself. Both neglect the fact that states can in principle communicate with each other and so avoid a costly miscalculation of relative power or will. The cause of 在本节中,我将论证,虽然这些观点都指向对战争的合理解释,但两者都不够深入,也都不能单独发挥作用。两者都忽略了一个事实,即各国原则上可以相互沟通,从而避免对相对实力或意愿的错误判断,而这种错误判断代价高昂。战争的原因
war cannot be simply lack of information, but whatever it is that prevents its disclosure. I argue that the fact that states have incentives to misrepresent their positions is crucial here, explaining on rationalist terms why diplomacy may not allow rational states to clarify disagreements about relative power or to avoid the miscalculation of resolve. 战争的原因不能仅仅是缺乏信息,而是任何阻碍信息披露的因素。我认为,国家有动机歪曲自己的立场这一事实在这里至关重要,它从理性主义的角度解释了为什么外交可能无法让理性的国家澄清关于相对实力的分歧或避免对决心的误判。
The mainstream international relations literature recognizes the existence of both private information and incentives to misrepresent, but typically views them as background conditions to be taken for granted rather than as key elements of an explanation of how rationally led states might end up at war. For example, Jack Levy’s impressive review of the literature on the causes of war contains nothing on the role of incentives to misrepresent and discusses private information largely in the context of misperceptions of other states’ intentions (which are linked to psychological biases). This is an accurate reflection of where these factors stand in the mainstream literature. ^(21){ }^{21} 主流国际关系文献承认私人信息和虚假陈述动机的存在,但通常将它们视为理所当然的背景条件,而不是解释理性主导的国家如何最终爆发战争的关键因素。例如,杰克-利维(Jack Levy)对战争原因的文献回顾令人印象深刻,其中没有任何关于虚假陈述动机作用的内容,而且主要是在对其他国家意图的误解(这与心理偏见有关)的背景下讨论私人信息的。这准确地反映了这些因素在主流文献中的地位。 ^(21){ }^{21}
Disagreements about relative power 关于相对权力的分歧
Geoffrey Blainey’s well-known and often-cited argument is that "wars usually begin when two nations disagree on their relative strength. ^('22){ }^{\prime 22} It is easy to see how a disagreement about relative strength—understood as conflicting estimates of the likelihood of military victory-can eliminate any ex ante bargaining range. Recall the example given above, where two states bargain over the division of $100\$ 100, and each has the outside option of going to war. If each expects that it surely would prevail at war, then each side’s expected value for the war option is $80(1*100+0*0-20)\$ 80(1 \cdot 100+0 \cdot 0-20). So given these expectations, neither side will accept less than $80\$ 80 in the bargaining, implying that no negotiated outcome is mutually preferred to war. More generally, suppose that state A expects to win with probability pp, state B expects to win with probability rr, and pp and rr sum to greater than one. Such conflicting expectations will certainly shrink and could eliminate any ex ante bargaining range. 杰弗里-布莱尼(Geoffrey Blainey)提出了一个众所周知且经常被引用的论点:"战争通常始于两国在相对实力上的分歧。 ^('22){ }^{\prime 22} 不难看出,对相对实力的分歧--可以理解为对军事胜利可能性的相互矛盾的估计--是如何消除任何事前讨价还价范围的。回顾上面的例子,两个国家就 $100\$ 100 的分割进行讨价还价,每个国家都有开战的外部选择。如果每一方都预期自己肯定会在战争中获胜,那么每一方对战争选项的预期值就是 $80(1*100+0*0-20)\$ 80(1 \cdot 100+0 \cdot 0-20) 。因此,考虑到这些预期,任何一方都不会在讨价还价中接受小于 $80\$ 80 的结果,这意味着没有任何谈判结果比战争更受双方青睐。更广义地说,假设甲国预期获胜的概率为 pp ,乙国预期获胜的概率为 rr ,且 pp 和 rr 之和大于1。这种相互冲突的预期肯定会缩小并可能消除任何事前讨价还价的范围。
But how could rationally led states have conflicting expectations about the likely outcome of military conflict? In the extreme case, how could both sides rationally expect to win? The literature barely addresses this question in 但是,理性的国家怎么会对军事冲突的可能结果抱有相互冲突的预期呢?在极端情况下,双方又如何能理性地预期获胜呢?文献中几乎没有涉及这个问题
explicit terms. Blainey, whom the literature views as advancing a rationalist explanation for war, in fact explains disagreements about relative power as a consequence of human irrationality. He says that mutual optimism about victory in war is the product of “moods which cannot be grounded in fact” and which “permeate what appear to be rational assessments of the relative military strength of two contending powers.” Mutual optimism is said to result from a “process by which nations evade reality,” which hardly sounds like a rationalist explanation. ^(23){ }^{23} 明确的术语。文献认为布雷尼对战争提出了理性主义的解释,但事实上,他把对相对实力的分歧解释为人类非理性的结果。他说,对战争胜利的相互乐观是 "无法建立在事实基础上的情绪 "的产物,这种情绪 "渗透在对两个竞争大国相对军事实力的看似理性的评估中"。相互乐观主义被认为是 "国家逃避现实的过程 "的结果,这听起来不像是理性主义的解释。 ^(23){ }^{23}
Conflicting expectations about the likely outcome of military conflict may be explained in three ways. First, as Blainey suggests, emotional commitments could irrationally bias leaders’ military estimates. They might, for instance, come to believe nationalist rhetoric holding that their soldiers are more courageous and spirited than those of the adversary. ^(24){ }^{24} Second, the world is a very complex place, and for this reason military analysts in different states could reach different conclusions about the likely impact of different technologies, doctrines, and tactics on the expected course of battle. Third, state leaders might have private information about militarily relevant factors-military capabilities, strategy, and tactics; the population’s willingness to prosecute a long war; or third-state intentions. If a state has superior (and so private) information about any such factor, then its estimate of the probable course of battle may differ from that of an adversary. 对军事冲突可能出现的结果的矛盾预期可以从三个方面来解释。首先,正如布雷尼所言,情感承诺可能会不合理地影响领导人对军事结果的估计。例如,他们可能会相信民族主义言论,认为自己的士兵比对手的士兵更勇敢、更有精神。 ^(24){ }^{24} 其次,世界是一个非常复杂的地方,因此,不同国家的军事分析家可能会就不同技术、理论和战术对预期战斗进程可能产生的影响得出不同的结论。第三,国家领导人可能掌握军事相关因素的私人信息--军事能力、战略和战术;民众进行长期战争的意愿;或第三国的意图。如果一个国家拥有关于任何此类因素的优势信息(因此也是私人信息),那么它对可能的战争进程的估计就可能与对手不同。
Under a strict but standard definition of rationality, only the third explanation qualifies as an account of how rationally led states could have conflicting estimates of the probability of winning in war. As argued by John Harsanyi, if two rational agents have the same information about an uncertain event, then they should have the same beliefs about its likely outcome. ^(25){ }^{25} The claim is that given identical information, truly rational agents should reason to the same conclusions about the probability of one uncertain outcome or another. Conflicting estimates should occur only if the agents have different (and so necessarily private) information. ^(26){ }^{26} 根据理性的严格但标准的定义,只有第三种解释才有资格解释理性领导的国家为何会对战争获胜的概率有相互矛盾的估计。正如约翰-哈桑尼(John Harsanyi)所论证的那样,如果两个理性主体对某一不确定事件拥有相同的信息,那么他们就应该对事件的可能结果拥有相同的信念。 ^(25){ }^{25} 其主张是,在信息完全相同的情况下,真正理性的代理人应该对一种或另一种不确定结果的概率得出相同的推理结论。只有当代理人拥有不同的(因此必然是私人的)信息时,才会出现相互矛盾的估计。 ^(26){ }^{26}
It follows that the second explanation for disagreements about relative power listed above-the complexity of the world-is not a rationalist account. Instead, it is an account that explains conflicting military estimates as a consequence of bounded rationality. In this view, leaders or military analysts with the same information about military technology, strategy, political will, etc. might reason to different conclusions about the likely course of a war because of differential ability to cope with complexity of the problem. This is entirely plausible, but it is a bounded rationality explanation rather than a fully rationalist one. ^(27){ }^{27} 由此可见,上述对相对实力分歧的第二种解释--世界的复杂性--并不是理性主义的解释。相反,它是一种将相互冲突的军事估计解释为有限理性的结果的解释。根据这种观点,在军事技术、战略、政治意愿等方面拥有相同信息的领导人或军事分析家,可能会因为应对问题复杂性的能力不同而对战争的可能进程得出不同的结论。这是完全有道理的,但这只是一种有界理性的解释,而不是完全理性主义的解释。 ^(27){ }^{27}
The rationalist account of how disagreements about the probability of winning might arise also seems empirically plausible. States certainly have private information about factors affecting the likely course of battle-for example, they jealously guard military secrets and often have superior information about what an ally will or will not fight for. Nonetheless, while private information about militarily relevant capabilities provides a first step, it does not provide a coherent rationalist explanation for war. The problem is that even if leaders have such private information, they should understand that their own estimates based on this information are suspect because they do not know the other side’s private information. In principle, both sides could gain by sharing information, which would yield a consensus military estimate (absent bounded rationality). And, as shown above, doing so could not help but reveal bargains that both would prefer to a fight. ^(28){ }^{28} 理性主义关于获胜概率的分歧如何产生的论述在经验上似乎也是可信的。国家当然掌握着影响可能的战斗进程的私人信息--例如,它们严守军事机密,而且往往掌握着关于盟友会为或不会为什么而战的优势信息。然而,虽然关于军事相关能力的私人信息提供了第一步,但它并不能为战争提供一个连贯的理性主义解释。问题在于,即使领导人掌握了这些私人信息,他们也应该明白,自己根据这些信息做出的估计是可疑的,因为他们不知道对方的私人信息。原则上,双方都可以通过共享信息而获益,这样就能达成一致的军事估计(不存在有界理性)。而且,如上文所示,这样做也会揭示出双方都希望达成的协议,而不是一场战斗。 ^(28){ }^{28}
So the question of how rationally led states can disagree about relative power devolves to the question of what prevents states from sharing private information about factors that might affect the course of battle. Before turning to this question, I will consider the second common explanation for how a rational miscalculation may produce war. 因此,关于理性主导的国家如何在相对实力问题上产生分歧的问题,就变成了是什么阻止了国家分享可能影响战争进程的因素的私人信息。在探讨这个问题之前,我将考虑理性误判如何导致战争的第二种常见解释。
War due to the miscalculation of an opponent's willingness to fight 因误判对手的作战意愿而发动战争
Many wars have been given the following so-called rationalist explanation: state A transgressed some interest of state B in the erroneous belief that B would not fight a war over the matter. Though rationally led, state A lacked information about B’s willingness to fight and simply happened to guess wrong, causing a war. Thus, some say that Germany miscalculated Russian and/or British willingness to fight in 1914; Hitler miscalculated Britain and France’s 许多战争都有如下所谓的理性主义解释:甲国错误地认为乙国不会就此事开战,从而侵犯了乙国的某些利益。虽然甲国是理性的,但它缺乏关于乙国是否愿意打仗的信息,只是碰巧猜错了,从而引发了战争。因此,有人说德国在 1914 年错误地估计了俄国和/或英国的战争意愿;希特勒错误地估计了英国和法国的战争意愿。
willingness to resist his drive to the east; Japanese leaders in 1941 miscalculated U.S. willingness to fight a long war over control in the South Pacific; North Korea miscalculated U.S. willingness to defend South Korea; the United States miscalculated China’s willingness to defend North Korea; and so on. In each case, the argument would hold that lack of information led a more-or-less rational actor to guess wrong about the extent of the bargaining range. 1941年,日本领导人错误估计了美国为争夺南太平洋控制权而进行长期战争的意愿;朝鲜错误估计了美国保卫南朝鲜的意愿;美国错误估计了中国保卫朝鲜的意愿;等等。在每种情况下,都会有这样的论点,即由于缺乏信息,一个或多或少理性的行为者对讨价还价的范围猜错了。
Blainey has argued that if states agree on relative power they are very unlikely to go to war against each other. ^(29){ }^{29} It is worth pointing out that in the preceding argument, war can occur despite complete agreement on relative power across states. To show how and for later use, I will introduce a simple model of international bargaining. As in the empirical examples just mentioned, in the model one state unilaterally chooses some revision of the status quo. The second state can then either acquiesce to the revision or can go to war to reverse it. 布莱尼认为,如果各国在相对实力问题上达成一致,它们就不太可能相互开战。 ^(29){ }^{29} 值得指出的是,在前面的论证中,尽管各国在相对实力上完全一致,战争也可能发生。为了说明这一点,我将引入一个简单的国际讨价还价模型。与刚才提到的实证例子一样,在这个模型中,一个国家单方面选择对现状进行某种修正。然后,第二个国家既可以默许这种修改,也可以发动战争来扭转局面。
Formally, suppose there is a status quo resolution of the issues q in Xq \in X and that state A has the opportunity to choose any outcome x in Xx \in X, presenting state BB with a fait accompli. On observing what state A did (which might be nothing, i.e., x=qx=q ), state B can choose whether to go to war or to acquiesce to A’s revision of the status quo. 从形式上看,假设问题的解决现状是 q in Xq \in X ,A 国有机会选择任何结果 x in Xx \in X ,给 B 国带来既成事实 BB 。在观察到 A 国的所作所为(可能什么也没做,即 x=qx=q )后,B 国可以选择是开战还是默认 A 国对现状的修正。
If neither state has any private information, so that all payoffs are common knowledge, state A does best to push the outcome just up to B’s reservation level p+c_(B)p+c_{B}, which makes B just willing to acquiesce rather than go to war. With complete information, then, the states avoid the inefficient outcome of war. ^(30){ }^{30} On the other hand, if state B has private information about either its capabilities (which affect pp ) or its value for the issues at stake relative to the costs of conflict (c_(B))\left(c_{B}\right), then state A may not know whether a particular “demand” xx will yield war or peace. Lacking this information, state A faces a trade-off in deciding whether and how much territory to “grab”: The larger the grab, the greater the risk of war, but the better off A will be if state B acquiesces. 如果两国都没有任何私人信息,因此所有的报酬都是共同的知识,那么甲国就会尽力将结果推高到乙国的保留水平 p+c_(B)p+c_{B} ,这使得乙国只愿意默许而不愿意开战。因此,在完全知情的情况下,各国都会避免战争这种低效率的结果。 ^(30){ }^{30} 另一方面,如果乙国拥有关于其能力(这会影响 pp )或相对于冲突成本 (c_(B))\left(c_{B}\right) 的利害问题价值的私人信息,那么甲国可能不知道特定的 "需求" xx 会带来战争还是和平。由于缺乏这方面的信息,A 国在决定是否 "抢夺 "以及抢夺多少领土时就会面临取舍:攫取的领土越大,战争的风险就越大,但如果乙国默许,甲国就会过得更好。
Suppose, for example, that A and B share a common estimate of pp-they agree about relative power-but that A is unsure about B’s costs for fighting. Under very broad conditions, if A cannot learn B’s private information and if A’s own costs are not too large, then state A’s optimal grab produces a positive chance of war. ^(31){ }^{31} Intuitively, if A is not too fearful of the costs of war relative to 举例来说,假设甲和乙对 pp 有一个共同的估计--他们对相对实力达成一致,但甲不确定乙的战斗成本。在非常宽泛的条件下,如果甲无法得知乙的私人信息,而且甲自己的成本也不是太大,那么甲的最优抢夺就会产生正的战争几率。 ^(31){ }^{31} 直观地说,如果 A 不太担心战争成本,相对于
29. Blainey, The Causes of War. 29.Blainey, The Causes of War.
30. This take-it-or-leave-it model of international bargaining is proposed and analyzed under conditions of both complete and incomplete information in James D. Fearon, “Threats to Use Force: The Role of Costly Signals in International Crises,” Ph.D. diss., University of California, Berkeley, 1992, chap. 1. Similar results for more elaborate bargaining structures are given in my own work in progress. See James D. Fearon, “Game-Theoretic Models of International Bargaining: An Overview,” University of Chicago, 1995. Powell has analyzed an alternative model in which both sides must agree if the status quo is to be revised. See Powell, “Bargaining in the Shadow of Power.” James D. Fearon, "Threats to Use Force. The Role of Costly Signals in International Crises," PhD diss:The Role of Costly Signals in International Crises," Ph.D. diss., University of California, Berkeley, 1992, chap.1.我自己正在进行的研究也给出了更复杂的讨价还价结构的类似结果。见 James D. Fearon, "Game-Theoretic Models of International Bargaining:芝加哥大学,1995 年。鲍威尔分析了另一种模式,在这种模式中,如果要改变现状,双方必须达成一致。见 Powell,"权力阴影下的谈判"。
31. See Claim 2 in the Appendix. 31.见附录中的权利要求 2。
what might be gained in bargaining, it will run some risk of war in hopes of gaining on the ground. So Blainey’s suggestion that a disagreement about relative power is necessary for war is incorrect-all that is necessary is that the states in dispute be unable to locate or agree on some outcome in the bargaining range. Since the bargaining range is determined not just by relative power but also by states’ values for the issues at stake relative to the costs of fighting, uncertainty about the latter can (and apparently does) produce war. 因此,布莱尼认为战争必须要有相对实力的分歧,这是不正确的。因此,布雷尼认为战争必须要有相对实力上的分歧,这种观点是不正确的--所需的条件是有争议的国家无法在讨价还价的范围内找到或商定某种结果。由于讨价还价的范围不仅取决于相对实力,还取决于各国对利害攸关的问题相对于战争成本的价值取向,因此后者的不确定性可能(而且显然确实)导致战争。
Once again, it is entirely plausible that state leaders have private information about their value for various international interests relative to their costs of fighting over them. ^(32){ }^{32} Thus it seems we have a second tenable rationalist explanation for war, again based on the concept of private information. But as in the case of disagreements about relative power, the explanation fails as given because it does not explain why states cannot avoid miscalculating a potential opponent’s willingness to fight. In the model, why cannot state A simply ask state BB whether it would fight rather than acquiesce to a particular demand? To give a concrete example, why did German leaders in 1914 not simply ask their British and Russian counterparts what they would do if Austria were to attack Serbia? If they could have done so and if the answers could have been believed, the Germans might not have miscalculated concerning Russian and, more importantly, British willingness to fight. In consequence they might have avoided the horrendous costs of World War I. 同样,国家领导人拥有关于各种国际利益的价值与争夺这些利益的成本的私人信息,这也是完全可信的。 ^(32){ }^{32} 由此看来,我们对战争有了第二个站得住脚的理性主义解释,同样是基于私人信息的概念。但与相对权力分歧的情况一样,这种解释也是失败的,因为它无法解释为什么国家不能避免误判潜在对手的战争意愿。在这个模型中,为什么国家 A 不能简单地询问国家 BB 它是否会战斗而不是默许某一要求呢?举个具体的例子,1914 年的德国领导人为什么不直接询问他们的英国和俄国同行,如果奥地利进攻塞尔维亚,他们会怎么做?如果他们能这样做,如果答案是可信的,德国人也许就不会对俄国人,更重要的是英国人的战斗意愿产生错误判断。因此,他们可能会避免第一次世界大战的可怕代价。
To recap, I have argued that in a rationalist framework, disagreements about relative power and uncertainty about a potential opponent’s willingness to fight must have the same source: leaders’ private information about factors affecting the likely course of a war or their resolve to fight over specific interests. In order to avoid war’s ex post inefficiency, leaders have incentives to share any such private information, which would have the effect of revealing peaceful settlements that lie within the bargaining range. So, to explain how war could occur between states led by rational leaders who consider the costs of fighting, we need to explain what would prevent them from sharing such private information. 概括地说,我已经论证了在理性主义框架下,关于相对实力的分歧和关于潜在对手作战意愿的不确定性必须有相同的来源:领导人关于影响战争可能进程的因素的私人信息或他们为特定利益而战的决心。为了避免战争的事后无效率,领导者有动力分享任何此类私人信息,因为这将揭示谈判范围内的和平解决方案。因此,要解释在考虑到战争成本的理性领导人所领导的国家之间为何会发生战争,我们需要解释是什么阻止了他们分享此类私人信息。
Incentives to misrepresent in bargaining 在谈判中进行虚假陈述的动机
Prewar bargaining may fail to locate an outcome in the bargaining range because of strategic incentives to withhold or misrepresent private information. While states have an incentive to avoid the costs of war, they also wish to obtain a favorable resolution of the issues. This latter desire can give them an incentive to exaggerate their true willingness or capability to fight, if by doing so they might deter future challenges or persuade the other side to make concessions. States can also have an incentive to conceal their capabilities or resolve, if they are concerned that revelation would make them militarily (and 战前讨价还价可能无法在讨价还价的范围内找到一个结果,这是因为存在着隐瞒或歪曲私人信息的战略动机。虽然国家有避免战争成本的动机,但它们也希望获得有利的问题解决方案。后一种愿望会促使国家夸大其真实的战争意愿或能力,如果这样做可以阻止未来的挑战或说服对方做出让步的话。如果国家担心揭露真相会使其在军事上(和政治上)处于不利地位,那么它们也会有动机隐瞒自己的能力或决心。
hence politically) vulnerable or would reduce the chances for a successful first strike. Similarly, states may conceal their true willingness to fight in order to avoid appearing as the aggressor. 因此在政治上是脆弱的,或者会减少首次打击成功的机会。同样,国家可能会隐瞒其真实的战斗意愿,以避免被视为侵略者。
Combined with the fact of private information, these various incentives to misrepresent can explain why even rational leaders may be unable to avoid the miscalculations of relative will and power that can cause war. This section first considers why this is so theoretically and then discusses two empirical examples. 再加上私人信息的事实,这些导致虚假陈述的各种诱因可以解释为什么即使是理性的领导人也可能无法避免对相对意愿和权力的误判,而这种误判可能导致战争。本节首先从理论上探讨为什么会出现这种情况,然后讨论两个经验案例。
A drawback of the simple bargaining model given above was that state BB had no opportunity to try to communicate its willingness to fight to state AA. It is easy to imagine that if communication were possible-say, if B could announce what interests in XX it considered vital enough to fight over-this might at least lower the chance of war by miscalculation. To check this, we give state B an initial opportunity to make a foreign policy announcement ff, which can be any statement about its foreign policy or what it considers to be vital or peripheral interests. (Assume as before that A is uncertain about B’s capabilities or costs for fighting.) 上文给出的简单讨价还价模型的一个缺点是,国家 BB 没有机会试图向国家 AA 表达其战斗意愿。不难想象,如果沟通是可能的--比如说,如果 B 可以宣布它认为 XX 中的哪些利益至关重要,足以为之开战--这至少可以降低因误判而发生战争的几率。为了验证这一点,我们给 B 国一个最初的机会来宣布外交政策 ff ,这个宣布可以是关于其外交政策的任何声明,也可以是它认为至关重要的利益或边缘利益。(假定 A 对 B 的作战能力或成本不确定)。
If the announcement itself has no effect on either side’s payoffs, then it can be shown that in any equilibrium in which state A does not choose randomly among demands, A will make the same demand regardless of what state BB says, and the ex ante risk of war will remain the same as in the game without communication by state B. To gain an intuition for these results, suppose that A conditioned its behavior on ff, grabbing more or less depending on what B announced. Then regardless of B’s true willingness to fight, B does best to make the announcement that leads to the smallest grab by A-that is, B has an incentive to misrepresent its actual willingness to resist. But then A learns nothing from the announcement. ^(33){ }^{33} 如果公告本身对任何一方的收益都没有影响,那么可以证明,在任何均衡中,状态 A 都不会在各种需求中随机选择,无论状态 BB 说了什么,A 都会提出相同的需求,并且战争的事前风险将与没有状态 B 公告的博弈中的风险相同。那么,无论 B 的真实战斗意愿如何,B 都会尽最大努力做出能让 A 抢得最少的宣布,也就是说,B 有动机歪曲自己的实际抵抗意愿。但这样一来,A 就什么也学不到了。 ^(33){ }^{33}
This conclusion is slightly altered if the leaders of B can render the announcement ff costly to make. ^(34){ }^{34} In practice, five common methods include building weapons, mobilizing troops, signing alliance treaties, supporting troops in a foreign land, and creating domestic political costs that would be paid if the announcement proves false. Of course, signaling by means of domestic political audience costs lies outside a purely unitary rational-actor framework, since this presumes a state run by an agent on behalf of a principal (the “audience”) rather than a unitary state with a perfectly secure leadership. In the latter case, leaders may be able to make foreign policy announcements 如果 B 国领导人能够使公告 ff 的发布代价高昂,那么这一结论就会略有改变。 ^(34){ }^{34} 在实践中,有五种常见的方法,包括制造武器、动员军队、签署同盟条约、支持驻扎在异国他乡的军队,以及制造国内政治成本,如果声明被证明是假的,就需要支付这些成本。当然,通过国内政治受众成本来发布信号的做法超出了纯粹的单一理性行为者框架,因为这假定了一个由代理人代表委托人("受众")管理的国家,而不是一个领导层完全安全的单一国家。在后一种情况下,领导者可能会宣布外交政策
credible only by engaging an international reputation, taking financially costly mobilization measures, or bearing the costs and risks of limited military engagements. ^(35){ }^{35} ^(35){ }^{35} 只有赢得国际声誉,采取财政上昂贵的动员措施,或承担有限军事行动的成本和风险,才有可信度。 ^(35){ }^{35}
Even when the signal is costly, however, this will not in general completely eliminate all risk of war by miscalculation-indeed, it may even increase it. The reason concerns the nature of the signals that states have incentives to send. To be genuinely informative about a state’s actual willingness or ability to fight, a signal must be costly in such a way that a state with lesser resolve or capability might not wish to send it. Actions that generate a real risk of war-for example, troop mobilizations that engage a leadership’s reputation before international or domestic audiences-can easily satisfy this constraint, since states with high resolve are less fearful of taking them. In other words, a rational state may choose to run a real risk of (inefficient) war in order to signal that it will fight if not given a good deal in bargaining. ^(36){ }^{36} 然而,即使发出的信号代价高昂,一般来说也不会完全消除因误判而引发战争的所有风险--事实上,它甚至可能会增加这种风险。原因在于国家有动机发出的信号的性质。要想真正了解一个国家的实际作战意愿或能力,一个信号的成本必须很高,以至于决心或能力较弱的国家可能不愿意发出这样的信号。产生实际战争风险的行动--例如,在国际或国内受众面前影响领导层声誉的部队调动--很容易满足这一约束条件,因为决心大的国家不太害怕采取这些行动。换句话说,一个理性的国家可能会选择冒着实际的(低效的)战争风险,以表明如果在讨价还价中得不到好的待遇,它就会参战。 ^(36){ }^{36}
The July crisis of World War I provides several examples of how incentives to misrepresent can make miscalculations of resolve hard to dispel. Soon after German leaders secretly endorsed Austrian plans to crush Serbia, they received both direct and indirect verbal indications from St. Petersburg that Russia would fight rather than acquiesce. ^(37){ }^{37} For example, on 21 July, the Russian Foreign Minister told the German ambassador that "Russia would not be able to tolerate Austria-Hungary’s using threatening language to Serbia or taking military measures. ^('38){ }^{\prime 38} Such verbal statements had little effect on German leaders’ beliefs, however, since they knew Russian leaders had a strategic incentive to misrepresent. On 18 July in a cable explaining Berlin’s policy to Ambassador Lichnowsky in London, Secretary of State Jagow wrote that “there is certain to be some blustering in St. Petersburg.” ^(39){ }^{39} Similarly, when on 26 July Lichnowsky began to report that Britain might join with France and Russia in the event of war, German Chancellor Bethmann Hollweg told his personal assistant of the "danger that France and England will commit their support to Russia in order not to alienate it, perhaps without really believing 第一次世界大战中的七月危机提供了几个例子,说明歪曲事实的动机如何使对决心的误判难以消除。在德国领导人秘密批准奥地利镇压塞尔维亚的计划后不久,他们就从圣彼得堡得到了直接和间接的口头暗示,即俄国将战斗而不是默许。 ^(37){ }^{37} 例如,7月21日,俄国外交部长对德国大使说:"俄国不能容忍奥匈帝国对塞尔维亚使用威胁性语言或采取军事措施。 ^('38){ }^{\prime 38} 不过,这种口头声明对德国领导人的信念影响不大,因为他们知道俄罗斯领导人有歪曲事实的战略动机。7月18日,国务卿雅戈夫在一份向驻伦敦大使利希诺夫斯基解释柏林政策的电报中写道:"圣彼得堡肯定会有一些虚张声势的行为"。 ^(39){ }^{39} 同样,7 月 26 日,当利希诺夫斯基开始报道英国可能会在战争爆发时与法国和俄国联手时,德国总理贝特曼-霍尔韦格告诉他的私人助理,"法国和英国为了不疏远俄国而承诺支持俄国的危险,也许他们并不真正相信这一点。
that for us mobilization means war, thinking of it as a bluff which they answer with a counterbluff." ^(40){ }^{40} 对我们来说,动员就意味着战争,认为这是一种虚张声势,而他们则以反虚张声势作为回应"。 ^(40){ }^{40}
At the same time, the Chancellor had an incentive to misrepresent the strength and nature of German support for Austria’s plans. Bethmann correctly anticipated that revealing this information would make Germany appear the aggressor, which might undermine Social Democratic support for his policies in Germany as well as turn British public opinion more solidly against his state. ^(41){ }^{41} This incentive led the Chancellor to avoid making direct or pointed inquiries about England’s attitude in case of war. The incentive also led him to pretend to go along with the British Foreign Secretary’s proposals for a conference to mediate the dispute. ^(42){ }^{42} In consequence, Lord Grey may not have grasped the need for a stronger warning to Germany until fairly late in the crisis (on 29 July), by which time diplomatic and military actions had made backing off more difficult for both Austria and Germany. 同时,首相也有动机歪曲德国支持奥地利计划的力度和性质。贝特曼正确地预料到,透露这些信息会使德国看起来是侵略者,这可能会削弱社会民主党对他在德国的政策的支持,并使英国舆论更加坚定地反对他的国家。 ^(41){ }^{41} 这一动机导致首相避免直接或尖锐地询问英国在战争情况下的态度。这一动机还使他假装同意英国外交大臣关于召开会议调解纠纷的建议。 ^(42){ }^{42} 因此,格雷勋爵可能直到危机后期(7月29日)才意识到有必要向德国发出更强烈的警告,而此时外交和军事行动已经使奥地利和德国更加难以退让。
In July 1914, incentives to misrepresent private information fostered and supported miscalculations of willingness to fight. Miscalculations of relative power can arise from this same source. On the one hand, states at times have an incentive to exaggerate their capabilities in an attempt to do better in bargaining. On the other hand, they can also have the well-known incentive to withhold information about capabilities and strategy. Presumably because of the strongly zero-sum aspect of military engagements, a state that has superior knowledge of an adversary’s war plans may do better in war and thus in prewar bargaining-hence, states rarely publicize war plans. While the theoretical logic has not been worked out, it seems plausible that states’ incentives to conceal information about capabilities and strategy could help explain some disagreements about relative power. 1914 年 7 月,虚报私人信息的动机助长并支持了对战斗意愿的误判。对相对实力的误判也可能源于此。一方面,国家有时有动机夸大自己的能力,试图在讨价还价中取得更好的结果。另一方面,它们也可能有众所周知的动机来隐瞒有关能力和战略的信息。据推测,由于军事交战具有强烈的零和特征,对对手的战争计划了如指掌的国家可能会在战争中表现得更好,从而在战前讨价还价中表现得更好--因此,国家很少公开战争计划。虽然这一理论逻辑尚未被阐明,但国家隐藏能力和战略信息的动机似乎可以帮助解释一些关于相对实力的分歧。
The 1904 war between Japan and Russia serves to illustrate this scenario. On the eve of the war, Russian leaders believed that their military could almost certainly defeat Japan. ^(43){ }^{43} In this conviction they differed little from the view of most European observers. By contrast, at the imperial council of 4 February that decided for war, the Japanese chief of staff estimated a fifty-fifty chance of 1904 年的日俄战争就可以说明这一点。战争前夕,俄国领导人认为他们的军队几乎肯定能打败日本。 ^(43){ }^{43} 在这一信念上,他们与大多数欧洲观察家的观点差别不大。相比之下,在 2 月 4 日决定开战的御前会议上,日本参谋总长估计,日本有五成的可能
prevailing, if their attack began immediately. ^(44){ }^{44} Thus Japanese and Russian leaders disagreed about relative power-their estimates of the likelihood of victory summed to greater than 1. 如果他们立即开始进攻,就会取得胜利。 ^(44){ }^{44} 因此,日本和俄国领导人在相对实力上存在分歧--他们对胜利可能性的估计总和大于 1。
Moreover, historical accounts implicate this disagreement as a major cause of the war: Russia’s refusal to compromise despite repeated offers by the Japanese was motivated in large measure by their belief that Japan would not dare attack them. The Japanese Cabinet finally decided for war after the Tsar and his advisers failed to make any real compromises over Korea or Manchuria in a series of proposals exchanged in 1903. The Tsar and his top advisers were hardly eager to fight, not because they expected to lose but because they saw an Asian war as a costly diversion of resources to the wrong theater. ^(45){ }^{45} Nonetheless, they refused to make concessions from what they viewed as a position of great military strength. They believed that Japan would have to settle for less, given its relative military weakness. ^(46){ }^{46} 此外,历史记载表明,这一分歧是战争的主要原因:俄国之所以在日本的一再提议下拒绝妥协,在很大程度上是因为他们相信日本不敢进攻他们。在 1903 年交换的一系列建议中,沙皇及其顾问未能就朝鲜或满洲问题做出任何真正的妥协,日本内阁最终决定发动战争。沙皇和他的高级顾问们并不热衷于战争,这并不是因为他们预计会失败,而是因为他们认为亚洲战争会把昂贵的资源转移到错误的战场上。 ^(45){ }^{45} 尽管如此,他们拒绝在他们认为拥有强大军事实力的情况下做出让步。他们认为,鉴于日本的军事实力相对较弱,日本只能退而求其次。 ^(46){ }^{46}
The disagreement arose in substantial part from Japanese private information about their military capabilities and how they compared with Russia’s. A far superior intelligence service had provided the Japanese military with a clear picture of Russian strengths and weaknesses in Northeast Asia and enabled them to develop an effective offensive strategy. According to John Albert White, due to this intelligence “the Japanese government apparently faced the war with a far more accurate conception of their task than their enemy had.” ^(47){ }^{47} In addition, compared with the Russians or indeed with any European power, Japanese leaders had much better knowledge of the fighting ability of the relatively untested Japanese army and of the effect of the reforms, training, and capital development of the previous decade. ^(48){ }^{48} 这种分歧在很大程度上源于日本关于其军事能力以及与俄罗斯军事能力比较的私人信息。出色得多的情报部门为日本军方提供了俄国在东北亚的优势和劣势的清晰图景,使他们能够制定有效的进攻战略。根据约翰-阿尔伯特-怀特的说法,由于这些情报,"日本政府在面对战争时,对其任务的认识显然要比敌人准确得多"。 ^(47){ }^{47} 此外,与俄国人相比,甚至与任何欧洲强国相比,日本领导人都更了解相对未经考验的日本军队的战斗力,以及前十年改革、训练和资本发展的效果。 ^(48){ }^{48}
If by communicating this private information the Japanese could have led the Russians to see that their chances of victory were smaller than expected, they might have done so. Almost all historians who have carefully examined the case agree that the Japanese government was not bent on war for its own sake-they 如果日本人能够通过传递这些私人信息,让俄国人看到他们获胜的机会比预期的要小,那么他们可能会这么做。几乎所有仔细研究过这一事件的历史学家都认为,日本政府并不是为了战争而一意孤行--他们
44. J. N. Westwood, Russia Against Japan, 1904-5: A New Look at the Russo-Japanese War (Albany: State University of New York Press, 1986), p. 22. Estimates varied within the Japanese leadership, but with the exception of junior-level officers, few seem to have been highly confident of victory. For example, as the decision for war was taken the Japanese navy requested a two-week delay to allow it to even the odds at sea. See Nish, The Origins of the Russo-Japanese War, pp. 197-200 and 206-7. 44.J. N. Westwood, Russia Against Japan, 1904-5:J. N. Westwood, Russia Against Japan, 1904-5: A New Look at the Russo-Japanese War (Albany: State University of New York Press, 1986), p. 22.日本领导层的估计各不相同,但除了下级军官外,似乎很少有人对胜利抱有很大信心。例如,在做出开战决定时,日本海军要求推迟两周,以便在海上扳回劣势。见 Nish,《日俄战争的起源》,第 197-200 页和第 206-7 页。
45. See, for example, David Walder, The Short Victorious War: The Russo-Japanese Conflict, 1904-5 (London: Hutchinson, 1973), pp. 53-56; and Nish, The Origins of the Russo-Japanese War, p. 253. 45.例如,见 David Walder, The Short Victorious War: The Russo-Japanese Conflict, 1904-5 (London: Hutchinson, 1973), 第 53-56 页;以及 Nish, The Origins of the Russo-Japanese War, 第 253 页。
46. See White, The Diplomacy of the Russo-Japanese War, chaps. 6-8; Nish, The Origins of the Russo-Japanese War, p. 241; and Lebow, Between Peace and War, pp. 244-46. 46.见 White, The Diplomacy of the Russo-Japanese War, chaps.6-8; Nish, The Origins of the Russo-Japanese War, p. 241; and Lebow, Between Peace and War, pp.
47. White, The Diplomacy of the Russo-Japanese War, p. 139. Nish writes that “many Russians certainly took a view of [the Japanese military] which was derisory in comparison with themselves. It may be that this derived from a deliberate policy of secrecy and concealment which the Japanese army applied because of the historic coolness between the two countries.” See Nish, The Origins of the Russo-Japanese War, p. 241. 47.怀特,《日俄战争的外交》,第 139 页。尼什写道:"许多俄国人对[日本军队]的看法肯定是自卑的。这可能是由于两国之间历史上的冷淡关系,日本军队故意采取了保密和隐蔽政策"。见 Nish,《日俄战争的起源》,第 241 页。
48. The British were the major exception, who as recent allies of Japan had better knowledge of its capabilities and level of organization. See Nish, The Origins of the Russo-Japanese War, p. 241. 48.英国是主要的例外,因为他们是日本新近的盟国,对日本的能力和组织水平有更深入的了解。见 Nish,《日俄战争的起源》,第 241 页。
were willing to compromise if the Russians would as well. ^(49){ }^{49} However, it was unthinkable for the Japanese to reveal such information or convince the Russians even if they did. In the first place, the Japanese could not simply make announcements about the quality of their forces, since the Russians would have had no reason to believe them. Second, explaining how they planned to win a war might seriously compromise any such attempt by changing the likelihood that they would win; there is a trade-off between revealing information about resolve or capabilities to influence bargaining and reducing the advantages of a first strike. 如果俄国人也愿意妥协,日本人也愿意妥协。 ^(49){ }^{49} 然而,日本人即使透露这些信息或说服俄国人也是不可想象的。首先,日本人不能简单地宣布其军队的质量,因为俄国人没有理由相信他们。其次,解释他们计划如何赢得战争可能会改变他们获胜的可能性,从而严重损害任何此类尝试;在透露有关决心或能力的信息以影响讨价还价与减少先发制人的优势之间存在权衡。
In sum, the combination of private information about relative power or will and the strategic incentive to misrepresent these afford a tenable rationalist explanation for war. While states always have incentives to locate a peaceful bargain cheaper than war, they also always have incentives to do well in the bargaining. Given the fact of private information about capabilities or resolve, these incentives mean that states cannot always use quiet diplomatic conversations to discover mutually preferable settlements. It may be that the only way to surmount this barrier to communication is to take actions that produce a real risk of inefficient war. 总之,关于相对权力或意愿的私人信息与歪曲这些信息的战略动机相结合,为战争提供了一个站得住脚的理性主义解释。虽然国家总是有动力找到比战争更便宜的和平谈判方式,但它们也总是有动力做好谈判工作。由于存在关于国家能力或决心的私人信息,这些动机意味着国家不可能总是通过平静的外交对话来找到双方都喜欢的解决方案。要克服这种沟通障碍,唯一的办法可能就是采取实际行动,冒着低效战争的风险。
This general mechanism operates in at least two other empirically important ways to produce conflict in specific circumstances. First, private information about the costs of fighting or the value leaders place on international interests can give them an incentive to cultivate a reputation for having lower costs or more far-flung vital interests than they actually do. If cutting a deal in one dispute would lead other states to conclude the leader’s costs for using force are high, then the leader might choose a costly war rather than suffer the depredations that might follow from making concessions. The U.S. interventions in Korea and Vietnam are sometimes explained in these terms, and states surely have worried about such inferences drawn by other states for a long time. ^(50){ }^{50} The same logic operates when a small state or group (for example, Finland or the Chechens) chooses to fight a losing war against a larger one (for example, the Soviet Union or Russia) in order to develop a reputation for being hard to subjugate. In both cases, states employ war itself as a costly signal of privately known and otherwise unverifiable information about willingness to fight. 这种一般机制至少还以另外两种经验上重要的方式在特定情况下产生冲突。首先,关于战争成本或领导人对国际利益的重视程度的私人信息,会促使他们培养自己的声誉,即拥有比实际更低的成本或更广泛的重要利益。如果在一场争端中达成协议会让其他国家认为领导人使用武力的成本很高,那么领导人可能会选择一场代价高昂的战争,而不是因为做出让步而遭受蹂躏。美国在朝鲜和越南的干预行动有时就是从这个角度来解释的,而长期以来,各国肯定都担心其他国家会做出这样的推论。 ^(50){ }^{50} 当一个小国或团体(如芬兰或车臣人)选择与一个大国(如苏联或俄罗斯)打一场失败的战争,以建立难以征服的声誉时,也是同样的逻辑。在这两种情况下,国家都会利用战争本身作为一种代价高昂的信号,以私下已知且无法核实的信息来表明战斗意愿。
Second, since incentives to misrepresent military strength can undermine diplomatic signaling, states may be forced to use war as a credible means to reveal private information about their military capabilities. Thus, a rising state may seek out armed conflict in order to demonstrate that it is more powerful 其次,由于虚报军事实力的动机会破坏外交信号,国家可能被迫利用战争作为披露其军事能力私人信息的可靠手段。因此,一个崛起中的国家可能会寻求通过武装冲突来证明自己更加强大。
than others realize, while a state in apparent decline may fight in hope of revealing that its capabilities remain better than most believe. In both instances, the inefficient outcome of war derives from the fact that states have private information about their capabilities and a strategic incentive to misrepresent it to other states. 而一个明显衰落的国家可能希望通过战争来揭示自己的能力仍然比大多数人认为的要好。在这两种情况下,战争的低效结果都源于国家拥有关于自身能力的私人信息,以及向其他国家歪曲信息的战略动机。
War as a consequence of commitment problems 战争是承诺问题的后果
This section considers a second and quite different rationalist mechanism by which war may occur even though the states in dispute share the same assessment of the bargaining range. Even if private information and incentives to misrepresent it do not tempt states into a risky process of discovery or foster costly investments in reputation, states may be unable to settle on an efficient bargained outcome when for structural reasons they cannot trust each other to uphold the deal. 本节将探讨第二种完全不同的理性主义机制,即即使争端国对谈判范围的评估相同,战争也可能发生。即使私人信息和歪曲信息的诱因不会诱使国家陷入高风险的发现过程,也不会助长代价高昂的声誉投资,但如果由于结构性原因,国家之间无法相互信任对方会维护协议,那么国家就可能无法达成有效的谈判结果。
In this class of explanations, the structural condition of anarchy reemerges as a major factor, although for nonstandard reasons. In the conventional argument, anarchy matters because no hegemonic power exists to threaten states with “jail” if they use force. Without this threat, states become suspicious and worried about other states’ intentions; they engage in self-help by building weapons; and somehow uncertainty-plus-weapons leads them ultimately to attack each other (the security dilemma or spiral model). Below, I show that anarchy does indeed matter but for more specific reasons and in more specific contexts. Anarchy matters when an unfortunate combination of state preferences and opportunities for action imply that one or both sides in a dispute have incentives to renege on peaceful bargains which, if they were enforceable, would be mutually preferred to war. I will consider three such unfortunate situations that can claim some empirical plausibility. 在这一类解释中,无政府状态的结构性条件再次成为主要因素,尽管原因并不标准。按照传统的说法,无政府状态之所以重要,是因为没有霸权国家威胁国家一旦使用武力就会 "坐牢"。如果没有这种威胁,国家就会对其他国家的意图产生怀疑和担忧;它们就会通过制造武器进行自救;不知何故,不确定性加上武器导致它们最终互相攻击(安全困境或螺旋模型)。下面,我将说明无政府状态确实重要,但原因更具体,背景更特殊。当国家偏好和行动机会不幸地结合在一起,意味着争端中的一方或双方有动机背弃和平协议时,无政府状态就很重要。我将考虑三种这样的不幸情况,它们在经验上具有一定的合理性。
It should be stressed that in standard security dilemma and spiral model arguments the suspicions and lack of trust engendered by anarchy are understood to originate either from states’ inability to observe each other’s motivations (that is, from private information about greed or desire for conquest) or from the knowledge that motivations can change. ^(51){ }^{51} By contrast, in the arguments given below, states have no private information and motivations never change; thus states understand each other’s motivations perfectly. This is not to argue that private information about the value a leadership places on expansion is unimportant in international politics-it surely is. Indeed, private information about motivation and various incentives to misrepresent it might exacerbate any of the three specific commitment problems discussed below. 需要强调的是,在标准的安全困境和螺旋模型论证中,无政府状态引起的猜疑和缺乏信任被理解为源于国家无法观察到对方的动机(即源于关于贪婪或征服欲望的私人信息),或者源于知道动机会发生变化。 ^(51){ }^{51} 与此相反,在下面的论证中,国家没有私人信息,动机永远不会改变;因此,国家完全了解彼此的动机。这并不是说,关于领导层对扩张的重视程度的私人信息在国际政治中并不重要--它肯定是重要的。事实上,有关动机的私人信息和歪曲动机的各种动机可能会加剧下文讨论的三个具体承诺问题中的任何一个。
However, when they do so this is a matter of an interaction between informational and commitment problems rather than of anarchy per se. Our first task should be to isolate and specify the mechanisms by which anarchy itself might cause war. 然而,当它们这样做时,这是信息问题和承诺问题相互作用的问题,而不是无政府状态本身的问题。我们的首要任务应该是分离并明确无政府状态本身可能引发战争的机制。
Preemptive war and offensive advantages 先发制人的战争和进攻优势
Consider the problem faced by two gunslingers with the following preferences. Each would most prefer to kill the other by stealth, facing no risk of retaliation, but each prefers that both live in peace to a gunfight in which each risks death. There is a bargain here that both sides prefer to “war”-namely, that each leaves the other alone-but without the enforcement capabilities of a third party, such as an effective sheriff, they may not be able to attain it. Given their preferences, neither person can credibly commit not to defect from the bargain by trying to shoot the other in the back. Note that no matter how far the shadow of the future extends, iteration (or repeat play) will not make cooperation possible in strategic situations of this sort. Because being the “sucker” here may mean being permanently eliminated, strategies of conditional cooperation such as tit-for-tat are infeasible. ^(52){ }^{52} Thus, if we can find a plausible analogy in international relations, this example might afford a coherent rationalist explanation for war. 考虑两个枪手面临的问题,他们的偏好如下。每一方都更希望以隐蔽的方式杀死对方,因为这样不会面临报复的风险,但每一方都更希望双方和平共处,而不是冒着死亡的风险进行枪战。在这里,双方都有一个比 "战争 "更可取的交易--即各自放过对方--但如果没有第三方的执法能力,比如一个有效的警长,他们可能无法实现这个交易。考虑到他们的偏好,双方都无法做出可信的承诺,不会试图从背后射杀对方,从而背弃协议。需要注意的是,无论未来的影子能延伸多远,迭代(或重复游戏)都不可能在这种战略情况下促成合作。因为在这里成为 "笨蛋 "可能意味着被永久淘汰,所以有条件的合作策略,如针锋相对,是不可行的。 ^(52){ }^{52} 因此,如果我们能在国际关系中找到一个合理的类比,这个例子也许能为战争提供一个连贯的理性主义解释。
Preemptive war scenarios provide the analogy. If geography or military technology happened to create large first-strike or offensive advantages, then states might face the same problem as the gunslingers. To demonstrate this theoretically, I consider how offensive advantages affect the bargaining range between two states engaged in a dispute. 先发制人的战争场景提供了类比。如果地理位置或军事技术碰巧创造了巨大的先发优势或进攻优势,那么国家可能会面临与枪手同样的问题。为了从理论上证明这一点,我考虑了进攻性优势如何影响陷入争端的两个国家之间的讨价还价范围。
There are at least three ways of interpreting offensive advantages in a formal context. First, an offensive advantage might mean that a state’s odds of winning are better if it attacks rather than defends. Second, an offensive advantage might mean that the costs of fighting are lower for an attacking state than for a defending state. It can be shown that no commitment problem operates in this second case, although lowering the costs of war for attackers does narrow the de facto bargaining range. Third, offensive advantages might mean that military technology and doctrine increase the variance of battlefield outcomes. That is, technology and doctrine might make total victory or total defeat more likely, while rendering stalemate and small territorial changes less likely. In this case, offensive advantages can actually reduce the expected utility of war for both sides, thus increasing the bargaining range and perhaps making war less rather 从形式上解释进攻优势至少有三种方法。首先,进攻优势可能意味着,如果一个国家进攻而不是防守,那么它的胜算就会更大。其次,进攻优势可能意味着进攻国的战斗成本低于防守国。可以证明,在第二种情况下不存在承诺问题,尽管降低进攻方的战争成本确实缩小了事实上的讨价还价范围。第三,进攻优势可能意味着军事技术和理论会增加战场结果的差异。也就是说,技术和理论可能使全胜或全败的可能性更大,而使僵局和小规模领土变化的可能性更小。在这种情况下,进攻性优势实际上会降低双方对战争的预期效用,从而增加讨价还价的余地,也许会使战争变得不那么重要。
than more likely. Intuitively, if states care most of all about security (understood as survival), then offensive advantages make war less safe by increasing the risk of total defeat. ^(53){ }^{53} 而不是更有可能。直观地说,如果国家最关心的是安全(理解为生存),那么进攻优势就会增加彻底失败的风险,从而降低战争的安全性。 ^(53){ }^{53}
A commitment problem of the sort faced by the gunslingers arises only under the first interpretation, in which “offensive advantage” refers to an increase in a state’s military prospects if it attacks rather than defends. To demonstrate this, let p_(f)p_{f} be the probability that state A wins a war if A attacks; p_(s)p_{s} the probability that A wins if A strikes second or defends; and pp the chance that A wins if both states mobilize and attack at the same time. Thus, an offensive advantage exists when p_(f) > p > p_(s)p_{f}>p>p_{s}. 只有在第一种解释下,才会出现枪手们所面临的那种承诺问题,在第一种解释中,"进攻优势 "指的是如果一个国家进攻而不是防御,那么它的军事前景就会增加。为了证明这一点,假设 p_(f)p_{f} 是 A 国在进攻时赢得战争的概率; p_(s)p_{s} 是 A 国在第二次进攻或防御时获胜的概率; pp 是 A 国在两个国家同时动员和进攻时获胜的概率。因此,当 p_(f) > p > p_(s)p_{f}>p>p_{s} 时存在进攻优势。
Since states can always choose to attack if they wish, a peaceful resolution of the issues is feasible only if neither side has an incentive to defect unilaterally by attacking. In the risk-neutral case, we must have x > p_(f)-c_(A)x>p_{f}-c_{A} for state A to prefer not to attack and 1-x > 1-p_(s)-c_(B)1-x>1-p_{s}-c_{B} for state B to prefer the bargained outcome xx. Thus stable outcomes that both sides prefer to conflict exist provided that p_(f)-c_(A) < p_(s)+c_(B)p_{f}-c_{A}<p_{s}+c_{B}, implying a de facto bargaining range represented by the interval (p_(f)-c_(A),p_(s)+c_(B))\left(p_{f}-c_{A}, p_{s}+c_{B}\right). 由于国家总是可以根据自己的意愿选择进攻,因此只有在任何一方都没有动机单方面进攻的情况下,和平解决问题才是可行的。在风险中性的情况下,我们必须有 x > p_(f)-c_(A)x>p_{f}-c_{A} 让甲国倾向于不进攻,而 1-x > 1-p_(s)-c_(B)1-x>1-p_{s}-c_{B} 让乙国倾向于讨价还价的结果 xx 。因此,只要有 p_(f)-c_(A) < p_(s)+c_(B)p_{f}-c_{A}<p_{s}+c_{B} ,双方都倾向于冲突的稳定结果就存在,这意味着事实上的讨价还价范围由区间 (p_(f)-c_(A),p_(s)+c_(B))\left(p_{f}-c_{A}, p_{s}+c_{B}\right) 表示。
Notice that as p_(f)p_{f} increases above pp, and p_(s)p_{s} decreases below it, this interval shrinks and may even disappear. Thus, first-strike advantages narrow the de facto bargaining range, while second-strike (or defensive) advantages increase it. The reason is that when first-strike advantages are large, both states must be given more from the peacetime bargain in order to allay the greater temptation of unilateral attack. 请注意,随着 p_(f)p_{f} 在 pp 以上的增大和 p_(s)p_{s} 在 pp 以下的减小,这个区间会缩小甚至消失。因此,先发优势缩小了事实上的谈判范围,而后发(或防御)优势则扩大了谈判范围。原因在于,当第一次进攻优势较大时,两国必须从和平时期的谈判中获得更多的利益,以减少单方面进攻的更大诱惑。
In the extreme case, if p_(f)-c_(A) > p_(s)+c_(B)p_{f}-c_{A}>p_{s}+c_{B}, or p_(f)-p_(s) > c_(A)+c_(B)p_{f}-p_{s}>c_{A}+c_{B}, no self-enforcing peaceful outcomes exist. This does not mean that no bargains exist that both sides would prefer to war. Since by definition both states cannot enjoy the advantage of going first, agreements that both sides prefer to fighting are always available in principle. The problem is that under anarchy, large enough first-strike incentives (relative to cost-benefit ratios) can make all of these agreements unenforceable and incredible as bargains. 在极端情况下,如果 p_(f)-c_(A) > p_(s)+c_(B)p_{f}-c_{A}>p_{s}+c_{B} 或 p_(f)-p_(s) > c_(A)+c_(B)p_{f}-p_{s}>c_{A}+c_{B} ,就不存在自我强化的和平结果。这并不意味着不存在双方都倾向于战争的协议。因为根据定义,两个国家都不可能享有先发优势,所以原则上总是存在双方都倾向于战争的协议。问题在于,在无政府状态下,足够大的先发制人激励(相对于成本-收益比)会使所有这些协议都无法执行,而且作为讨价还价也是不可思议的。
Does this prisoners’ dilemma logic provide an empirically plausible explanation for war? Though I lack the space to develop the point, I would argue that first-strike and offensive advantages probably are an important factor making war more likely in a few cases, but not because they make mobilization and attack a dominant strategy, as in the extreme case above. In the pure preemptive war scenario leaders reason as follows: “The first-strike advantage is so great that regardless of how we resolve any diplomatic issues between us, one side will always want to attack the other in an effort to gain the (huge) advantage of going first.” But even in July 1914, a case in which European 这种 "囚徒困境 "逻辑是否为战争提供了经验上合理的解释?虽然我没有足够的篇幅来阐述这一点,但我认为,在少数情况下,先发制人和进攻优势可能是使战争更有可能发生的一个重要因素,但并不是因为它们使动员和进攻成为一种主导战略,就像上述极端情况那样。在纯粹的先发制人战争中,领导人的理由如下:"先发制人的优势是如此之大,以至于无论我们之间的外交问题如何解决,一方总是想要攻击另一方,以争取获得先发制人的(巨大)优势"。但是,即使在 1914 年 7 月,欧洲
leaders apparently held extreme views about the advantage of striking first, we do not find leaders thinking in these terms. ^(54){ }^{54} It would be rather surprising if they did, since they had all lived at peace but with the same military technology prior to July 1914. Moreover, in the crisis itself military first-strike advantages did not become a concern until quite late, and right to the end competed with significant political (and so strategic) disadvantages to striking first. ^(55){ }^{55} 虽然各国领导人显然对先发制人的优势持有极端的看法,但我们并没有发现他们有这样的想法。 ^(54){ }^{54} 如果他们真的这么想,那就太令人吃惊了,因为在1914年7月之前,他们都生活在和平环境中,但却拥有相同的军事技术。此外,在这场危机中,先发制人的军事优势直到很晚才成为人们关注的问题,而且直到最后,先发制人在政治上(因此在战略上)还存在着巨大的劣势。 ^(55){ }^{55}
Rather than completely eliminating enforceable bargains and so causing war, it seems more plausible that first-strike and offensive advantages exacerbate other causes of war by narrowing the bargaining range. If for whatever reason the issues in dispute are hard to divide up, then war will be more likely the smaller the set of enforceable agreements both sides prefer to a fight. Alternatively, the problems posed by private information and incentives to misrepresent may be more intractable when the de facto bargaining range is small. ^(56){ }^{56} For example, in 1914 large perceived first-strike advantages meant that relatively few costly signals of intent were sufficient to commit both sides to war (chiefly, for Germany/Austria and Russia). Had leaders thought defense had the advantage, the set of enforceable agreements both would have preferred would have been larger, and this may have made costly signaling less likely to have destroyed the bargaining range. 与其说先发制人和进攻性优势完全消除了可执行的讨价还价从而导致战争,倒不如说先发制人和进攻性优势缩小了讨价还价的范围,从而加剧了战争的其他原因。无论出于何种原因,如果争议问题难以分割,那么双方更倾向于达成的可执行协议的范围越小,战争的可能性就越大。或者说,当事实上的谈判范围较小时,私人信息和虚假陈述的动机所带来的问题可能会更加棘手。 ^(56){ }^{56} 例如,在 1914 年,巨大的先发制人优势意味着相对较少的代价高昂的意图信号就足以让双方(主要是德国/奥地利和俄国)投入战争。如果领导人认为防御具有优势,那么双方都希望达成的可执行协议就会更多,这可能会使代价高昂的信号传递更不可能破坏讨价还价的范围。
I should note that scholars have sometimes portrayed the preemptive war problem differently, assuming that neither state would want to attack unilaterally but that each would want to attack if the other was expected to also. This is a coordination problem known as “stag hunt” that would seem easily resolved by communication. At any rate, it seems farfetched to think that small numbers of states (typically dyads) would have trouble reaching the efficient solution here, if coordination were really the only problem. ^(57){ }^{57} 我要指出的是,学者们有时对先发制人的战争问题有不同的描述,他们假定两国都不想单方面发动战争,但如果预期对方也会发动战争,那么两国都会想发动战争。这是一个被称为 "猎鹿 "的协调问题,似乎很容易通过沟通来解决。无论如何,如果协调真的是唯一的问题,那么认为少数国家(通常是二元对立的国家)很难在这里达成有效的解决方案似乎有些牵强。 ^(57){ }^{57}
Preventive war as a commitment problem 作为承诺问题的预防性战争
Empirically, preventive motivations seem more prevalent and important than preemptive concerns. In his diplomatic history of Europe from 1848 to 1918, A.J.P. Taylor argued that “every war between the Great Powers [in this period] started as a preventive war, not a war of conquest.” ^(58){ }^{58} In this subsection 从经验上看,预防性动机似乎比先发制人的关切更为普遍和重要。泰勒(A.J.P. Taylor)在他的《1848-1918 年欧洲外交史》中认为,"[这一时期]大国之间的每一场战争都是以预防性战争而非征服战争开始的"。 ^(58){ }^{58} 在本小节中
I argue that within a rationalist framework, preventive war is properly understood as arising from a commitment problem occasioned by anarchy and briefly discuss some empirical implications of this view. ^(59){ }^{59} 我认为,在理性主义框架内,预防性战争应被正确理解为源于无政府状态下的承诺问题,并简要讨论了这一观点的一些经验意义。 ^(59){ }^{59}
The theoretical framework used above is readily adapted for an analysis of the preventive war problem. Whatever their details, preventive war arguments are necessarily dynamic-they picture state leaders who think about what may happen in the future. So, we must modify the bargaining model to make it dynamic as well. Suppose state A will have the opportunity to choose the resolution of the issues in each of an infinite number of successive periods. For periods t=1,2,dotst=1,2, \ldots, state A can attempt a fait accompli to revise the status quo, choosing a demand x_(t)x_{t}. On seeing the demand x_(t)x_{t}, state B can either acquiesce or go to war, which state A is assumed to win with probability p_(t)p_{t}. For simplicity, assume the states are risk-neutral; that the winner of a war gets to implement its favorite issue resolution for all subsequent periods; and that the states discount future payoffs by a per-period factor delta in(0,1)\delta \in(0,1). 上文使用的理论框架很容易用于分析预防性战争问题。无论其细节如何,预防性战争的论点必然是动态的--它们描绘了国家领导人对未来可能发生的事情的思考。因此,我们必须修改讨价还价模型,使其也具有动态性。假设 A 国将有机会在无限多个连续时期中的每个时期选择问题的解决方案。在 t=1,2,dotst=1,2, \ldots 期间,A 国可以尝试用既成事实来改变现状,选择一个需求 x_(t)x_{t} 。在看到需求 x_(t)x_{t} 时,B 国可以选择默许或开战,假设 A 国以概率 p_(t)p_{t} 赢得战争。为简单起见,假设各国都是风险中性的;战争的赢家可以在随后的所有时期内执行其最喜欢的问题解决方案;各国按每期系数 delta in(0,1)\delta \in(0,1) 对未来的回报进行贴现。
This model extends the one-period bargaining game considered above to an infinite-horizon case in which military power can vary over time. An important observation about the multiperiod model is that war remains a strictly inefficient outcome. If the states go to war in period tt expected payoffs from period tt on are (p_(t)//(1-delta))-c_(A)\left(p_{t} /(1-\delta)\right)-c_{A} for state AA and ((1-p_(t))//(1-delta))-c_(B)\left(\left(1-p_{t}\right) /(1-\delta)\right)-c_{B} for state B. It is straightforward to show that there will always exist peaceful settlements in XX such that both states would prefer to see one of these settlements implemented in every period from tt forward rather than go to war. 这一模型将上文考虑的单周期讨价还价博弈扩展到了无限视距的情况,在这种情况下,军事实力会随着时间的推移而变化。关于多期模型的一个重要观察结果是,战争仍然是一种严格意义上的无效率结果。如果两国在 tt 期开战,那么从 tt 期开始,两国的预期收益分别为 AA 的 (p_(t)//(1-delta))-c_(A)\left(p_{t} /(1-\delta)\right)-c_{A} 和 ((1-p_(t))//(1-delta))-c_(B)\left(\left(1-p_{t}\right) /(1-\delta)\right)-c_{B} 的 AA 。
The strategic dilemma is that without some third party capable of guaranteeing agreements, state A may not be able to commit itself to future foreign policy behavior that makes B prefer not to attack at some point. Consider the simple case in which A’s chance of winning a war begins at p_(1)p_{1} and then will increase to p_(2) > p_(1)p_{2}>p_{1} in the next period, where it will remain for all subsequent periods. Under anarchy, state A cannot commit itself not to exploit the greater bargaining leverage it will have starting in the second period. In the unique subgame perfect equilibrium, A will demand x_(t)=p_(2)+c_(B)(1-delta)x_{t}=p_{2}+c_{B}(1-\delta) in the second period and in all subsequent periods tt. That is, it will choose a resolution of the issues that makes state BB just willing to acquiesce, given the new distribution of military power. This means that in the first period, when state BB is still relatively strong, B is choosing between going to war and acquiescing to the demand x_(1)x_{1}, which would yield it a total payoff of 1-x_(1)+delta(1-x_(2))//(1-delta)1-x_{1}+\delta\left(1-x_{2}\right) /(1-\delta). The most state 战略困境在于,如果没有能够保证达成协议的第三方,A 国可能无法对未来的外交政策行为做出承诺,使 B 国宁愿在某个时候不发动进攻。考虑这样一个简单的情况:A 国赢得战争的几率从 p_(1)p_{1} 开始,然后在下一个时期增加到 p_(2) > p_(1)p_{2}>p_{1} ,并在随后的所有时期保持不变。在无政府状态下,A 国不能承诺不利用从第二期开始的更大讨价还价筹码。在唯一的次博弈完全均衡中,A 将在第二期要求 x_(t)=p_(2)+c_(B)(1-delta)x_{t}=p_{2}+c_{B}(1-\delta) ,并在随后的各期要求 tt 。也就是说,在新的军事力量分配下,它将选择一种使国家 BB 只愿意默许的问题解决方案。这意味着,在第一个时期,当国家 BB 仍然相对强大时,B 将在参战和默许需求 x_(1)x_{1} 之间做出选择,后者将使其获得总回报 1-x_(1)+delta(1-x_(2))//(1-delta)1-x_{1}+\delta\left(1-x_{2}\right) /(1-\delta) 。最态
A can do for B is to set x_(1)x_{1} to zero, so the largest possible payoff to state B for acquiescing in the first period is 1+delta(1-x_(2))//(1-delta)1+\delta\left(1-x_{2}\right) /(1-\delta). But this can be less than B’s payoff for attacking in the first period, and indeed it will be whenever the following condition holds: deltap_(2)-p_(1) > c_(B)(1-delta)^(2)\delta p_{2}-p_{1}>c_{B}(1-\delta)^{2}. In words (roughly), if B’s expected decline in military power is too large relative to B’s costs for war, then state A’s inability to commit to restrain its foreign policy demands after it gains power makes preventive attack rational for state B. Note also that A’s commitment problem meshes with a parallel problem facing B. If B could commit to fight in the second period rather than accept the rising state’s increased demands, then B’s bargaining power would not fall in the second period, so that preventive war would be unnecessary in the first. A 可以为 B 做的是将 x_(1)x_{1} 设为零,因此,B 国在第一期默许的最大可能报酬是 1+delta(1-x_(2))//(1-delta)1+\delta\left(1-x_{2}\right) /(1-\delta) 。但这可能小于 B 在第一期发动进攻的收益,事实上,只要下面的条件成立,收益就会小于第一期发动进攻的收益: deltap_(2)-p_(1) > c_(B)(1-delta)^(2)\delta p_{2}-p_{1}>c_{B}(1-\delta)^{2} 。换句话说(大致),如果相对于 B 的战争成本而言,B 的军事实力的预期下降幅度过大,那么 A 国在获得权力后无法承诺限制其外交政策要求,这就使得 B 国的预防性进攻变得合理。
Several points about this rationalist analysis of preventive war are worth stressing. First, preventive war occurs here despite (and in fact partially because of) the states’ agreement about relative power. Preventive war is thus another area where Blainey’s argument misleads. Second, contrary to the standard formulation, the declining state attacks not because it fears being attacked in the future but because it fears the peace it will have to accept after the rival has grown stronger. To illustrate, even if Iraq had moved from Kuwait to the conquest of Saudi Arabia, invasion of the United States would not have followed. Instead, the war for Kuwait aimed to prevent the development of an oil hegemon that would have had considerable bargaining leverage due to U.S. reliance on oil. ^(60){ }^{60} 关于这种预防性战争的理性主义分析,有几点值得强调。首先,尽管各国就相对实力达成了一致,但预防性战争还是发生了(事实上部分原因是)。因此,预防性战争是布雷尼论证的另一个误区。其次,与标准的表述相反,衰落国家发动攻击并不是因为它担心将来会受到攻击,而是因为它担心在对手变得更加强大之后不得不接受和平。举例来说,即使伊拉克从科威特转向征服沙特阿拉伯,也不会随之入侵美国。相反,科威特战争的目的是防止石油霸权的发展,由于美国对石油的依赖,石油霸权会有相当大的讨价还价的筹码。 ^(60){ }^{60}
Third, while preventive war arises here from states’ inability to trust each other to keep to a bargain, the lack of trust is not due to states’ uncertainty about present or future motivations, as in typical security-dilemma and spiral-model accounts. In my argument, states understand each other’s motivations perfectly well-there is no private information-and they further understand that each would like to avoid the costs of war-they are not ineluctably greedy. Lack of trust arises here from the situation, a structure of preferences and opportunities, that gives one party an incentive to renege. For example, regardless of expectations about Saddam Hussein’s future motivation or intentions, one could predict with some confidence that decreased competition among sellers of oil would have led to higher prices. My claim is not that uncertainty about intentions is unimportant in such situations-it surely is-but that commitment and informational problems are distinct mechanisms and that a rationalist preventive war argument turns crucially on a commitment problem. 第三,虽然预防性战争的起因是国家之间无法信任对方会遵守协议,但这种信任的缺失并不是因为国家对当前或未来动机的不确定性,正如典型的安全困境和螺旋模型所描述的那样。在我的论证中,各国完全了解彼此的动机--不存在私人信息--而且它们进一步了解彼此都希望避免战争成本--它们并非不可避免地贪婪。在这种情况下,缺乏信任的原因在于偏好和机会的结构使一方有动机出尔反尔。例如,无论人们对萨达姆-侯赛因未来的动机或意图有何预期,都可以有把握地预测,石油销售商之间竞争的减少会导致价格上涨。我的主张并不是说在这种情况下意图的不确定性不重要--它肯定是重要的--而是说承诺问题和信息问题是不同的机制,理性主义的预防性战争论证关键在于承诺问题。
Finally, the commitment problem behind preventive war may be undermined if the determinants of military power can reliably be transferred between states. In the model, the rising state can actually have an incentive to transfer away or 最后,如果军事实力的决定因素可以在国家之间可靠地转移,预防性战争背后的承诺问题可能会被削弱。在该模型中,崛起的国家实际上有动机转移或
otherwise limit the sources of its new strength, since by doing so it may avoid being attacked. While such transfers might seem implausible from a realist perspective, the practice of “compensation” in classical balance-of-power politics may be understood in exactly these terms: states that gained territory by war or other means were expected to (and sometimes did) allow compensating gains in order to reduce the incentive for preventive war against them. ^(61){ }^{61} 因为这样做可以避免受到攻击。虽然从现实主义的角度来看,这种转让似乎难以置信,但古典均势政治中的 "补偿 "做法却可以从这些角度来理解:通过战争或其他手段获得领土的国家应该(有时也确实)允许补偿性收益,以减少对其发动预防性战争的动机。 ^(61){ }^{61}
Preventive motivations figured in the origins of World War I and are useful to illustrate these points. One of the reasons that German leaders were willing to run serious risks of global conflict in 1914 was that they feared the consequences of further growth of Russian military power, which appeared to them to be on a dangerous upward trajectory. ^(62){ }^{62} Even if the increase in Russian power had not led Russia to attack Austria and Germany at some point in the future-war still being a costly option-greater Russian power would have allowed St. Petersburg to pursue a more aggressive foreign policy in the Balkans and the Near East, where Austria and Russia had conflicting interests. Austrian and German leaders greatly feared the consequences of such a (pro-Slav) Russian foreign policy for the domestic stability of the AustroHungarian Empire, thus giving them incentives for a preventive attack on Russia. ^(63){ }^{63} 预防性动机在第一次世界大战的起源中占有一席之地,有助于说明这些观点。德国领导人之所以愿意在1914年冒着全球冲突的严重风险,原因之一是他们担心俄国军事力量进一步增长的后果,因为在他们看来,俄国的军事力量正处于危险的上升轨道上。 ^(62){ }^{62} 即使俄国国力的增长没有导致俄国在未来某个时候进攻奥地利和德国--战争仍然是一个代价高昂的选择--俄国国力的增长也会使圣彼得堡在巴尔干半岛和近东地区推行更具侵略性的外交政策,而奥地利和俄国在这些地区的利益是相互冲突的。奥地利和德国领导人非常担心这种(亲斯拉夫的)俄国外交政策会对奥匈帝国的国内稳定造成影响,从而促使他们对俄国发动预防性进攻。 ^(63){ }^{63}
By the argument made above, the states should in principle have had incentives to cut a multiperiod deal both sides would have preferred to preventive war. For example, fearing preventive attack by Austria and Germany, Russian leaders might have wished to have committed themselves not to push so hard in the Balkans as to endanger the Dual Monarchy. But such a deal would be so obviously unenforceable as to not be worth proposing. Leaving aside the serious monitoring difficulties, once Russia had become stronger militarily, Austria would have no choice but to acquiesce to a somewhat more aggressive Russian policy in the Balkans. And so Russia would be drawn to pursue it, regardless of its overall motivation or desire for conquest of Austria-Hungary. 根据上述论点,各国原则上应该有动力达成一个多阶段的协议,双方都会倾向于预防性战争。例如,由于担心奥地利和德国的预防性进攻,俄国领导人可能希望承诺不在巴尔干地区大举进攻,以免危及二元君主制。但这样的协议显然无法执行,不值得提出。撇开监督方面的严重困难不谈,一旦俄国在军事上变得更加强大,奥地利将别无选择,只能默许俄国在巴尔干地区采取更加激进的政策。因此,无论俄国的总体动机或征服奥匈帝国的愿望如何,它都会被吸引去推行这一政策。
While German leaders in July 1914 were willing to accept a very serious risk that Russia might go to war in support of Serbia, they seem to have hoped at the start of the crisis that Russia would accept the Austrian demarche. ^(64){ }^{64} Thus, it is hard to argue that the preventive logic itself produced the war. Rather, as is 虽然德国领导人在1914年7月愿意接受俄国可能会为支持塞尔维亚而参战这一非常严重的风险,但他们似乎在危机一开始就希望俄国会接受奥地利的警告。 ^(64){ }^{64} 因此,很难说是预防性逻辑本身导致了战争。相反,正如
probably true for other cases in which these concerns appear, the preventive logic may have made war more likely in combination with other causes, such as private information, by making Berlin much more willing to risk war. ^(65){ }^{65} How preventive concerns impinge on international bargaining with private information is an important topic for future research. 在出现这些担忧的其他情况下,预防性逻辑可能会使柏林更愿意冒战争的风险,从而使战争更有可能与私人信息等其他原因结合起来。 ^(65){ }^{65} 预防性考量如何影响私人信息的国际讨价还价是未来研究的一个重要课题。
Commitment, strategic territory, and the problem of appeasement 承诺、战略领土和绥靖问题
The objects over which states bargain frequently are themselves sources of military power. Territory is the most important example, since it may provide economic resources that can be used for the military or be strategically located, meaning that its control greatly increases a state’s chances for successful attack or defense. Territory is probably also the main issue over which states fight wars. ^(66){ }^{66} 国家经常讨价还价的对象本身就是军事力量的来源。领土是最重要的例子,因为它可能提供可用于军事的经济资源,或具有战略位置,这意味着对它的控制会大大增加国家成功攻防的机会。领土可能也是国家发动战争的主要问题。 ^(66){ }^{66}
In international bargaining on issues with this property, a commitment problem can operate that makes mutually preferable negotiated solutions unattainable. The problem is similar to that underlying preventive war. Here, both sides might prefer some package of territorial concessions to a fight, but if the territory in question is strategically vital or economically important, its transfer could radically increase one side’s future bargaining leverage (think of the Golan Heights). In principle, one state might prefer war to the status quo but be unable to commit not to exploit the large increase in bargaining leverage it would gain from limited territorial concessions. Thus the other state might prefer war to limited concessions (appeasement), so it might appear that the issues in dispute were indivisible. But the underlying cause of war in this instance is not indivisibility per se but rather the inability of states to make credible commitments under anarchy. ^(67){ }^{67} 在就具有这种性质的问题进行国际讨价还价时,可能会出现承诺问题,使双方都无法通过谈判达成可取的解决方案。这个问题与预防性战争的基本问题类似。在这种情况下,双方可能都更倾向于一揽子领土让步而非战争,但如果有关领土在战略上或经济上非常重要,其转让可能会从根本上增加一方未来讨价还价的筹码(想想戈兰高地)。原则上,一国可能宁愿战争也不愿维持现状,但又无法承诺不利用有限的领土让步所带来的谈判筹码的大幅增加。因此,另一个国家可能宁愿战争也不愿做出有限的让步(绥靖),这样看来,争议问题似乎是不可分割的。但在这种情况下,战争的根本原因并不是不可分割性本身,而是国家在无政府状态下无法做出可信的承诺。 ^(67){ }^{67}
As an example, the 1939 Winter War between Finland and the Soviet Union followed on the refusal of the Finnish government to cede some tiny islands in the Gulf of Finland that Stalin seems to have viewed as necessary for the defense of Leningrad in the event of a European war. One of the main reasons the Finns were so reluctant to grant these concessions was that they believed they could not trust Stalin not to use these advantages to pressure Finland for 举例来说,1939 年芬兰与苏联之间的冬季战争就是因为芬兰政府拒绝让出芬兰湾的一些小岛,而斯大林似乎认为这些小岛是在欧洲战争中保卫列宁格勒所必需的。芬兰人如此不愿意做出这些让步的主要原因之一是,他们认为不能相信斯大林不会利用这些优势向芬兰施压,以换取
more in the future. So it is possible that Stalin’s inability to commit himself not to attempt to carry out in Finland the program he had just applied in the Baltic states may have led or contributed to a costly war both sides clearly wished to avoid. ^(68){ }^{68} 未来还会有更多的战争。因此,斯大林无法承诺不试图在芬兰实施他刚刚在波罗的海国家实施的计划,这有可能导致或促成了一场代价高昂的战争,而这显然是双方都希望避免的。 ^(68){ }^{68}
Conclusion 结论
The article has developed two major claims. First, under broad conditions the fact that fighting is costly and risky implies that there should exist negotiated agreements that rationally led states in dispute would prefer to war. This claim runs directly counter to the conventional view that rational states can and often do face a situation of deadlock, in which war occurs because no mutually preferable bargain exists. 文章提出了两大主张。首先,在广泛的条件下,战斗代价高、风险大,这意味着应该存在谈判达成的协议,理性的争端当事国更倾向于这些协议,而不是战争。这一主张与传统观点直接背道而驰,传统观点认为,理性的国家可能而且经常会面临僵局,在这种情况下,战争的发生是因为不存在双方都更喜欢的谈判协议。
Second, essentially two mechanisms, or causal logics, explain why rationally led states are sometimes unable to locate or agree on such a bargain: (1) the combination of private information about resolve or capability and incentives to misrepresent these, and (2) states’ inability, in specific circumstances, to commit to uphold a deal. Historical examples were intended to suggest that both mechanisms can claim empirical relevance. 其次,基本上有两种机制或因果逻辑可以解释为什么理性主导的国家有时无法找到或达成这样的交易:(1) 关于决心或能力的私人信息与歪曲这些信息的动机的结合;(2) 国家在特定情况下无法承诺维护交易。历史上的例子旨在表明,这两种机制都具有经验意义。
I conclude by anticipating two criticisms. First, I am not saying that explanations for war based on irrationality or “pathological” domestic politics are less empirically relevant. Doubtless they are important, but we cannot say how so or in what measure if we have not clearly specified the causal mechanisms making for war in the “ideal” case of rational unitary states. In fact, a better understanding of what the assumption of rationality really implies for explaining war may actually raise our estimate of the importance of particular irrational and second-image factors. 最后,我预计会有两种批评意见。首先,我并不是说基于非理性或 "病态 "国内政治的战争解释在经验上不那么重要。毫无疑问,它们是重要的,但如果我们没有明确指出在理性单一制国家的 "理想 "情况下战争的因果机制,我们就不能说它们有多重要或重要到什么程度。事实上,更好地理解理性假设对解释战争的真正意义,实际上可能会提高我们对某些非理性和第二形象因素重要性的估计。
For example, once the distinction is made clear, bounded rationality may appear a more important cause of disagreements about relative power than private information about military capabilities. If private information about capabilities was often a major factor influencing the odds of victory, then we would expect rational leaders to update their war estimates during international crises; a tough bargaining stand by an adversary would signal that the adversary was militarily stronger than expected. Diplomatic records should then contain evidence of leaders reasoning as follows: “The fact that the other side is not backing down means that we are probably less likely to win at war than we initially thought.” I do not know of a single clear instance of this sort of updating in any international crisis, even though updating about an opponent’s resolve, or willingness to fight, is very common. 举例来说,一旦区分清楚,有界的理性可能比军事能力的私人信息更能导致对相对实力的分歧。如果关于军事能力的私人信息往往是影响胜算的主要因素,那么我们就会期望理性的领导人在国际危机期间更新他们对战争的估计;对手强硬的讨价还价立场将表明对手在军事上比预期的要强大。因此,外交记录中应包含领导人进行如下推理的证据:"对方不退让的事实意味着,我们在战争中获胜的可能性可能比我们最初想象的要小"。据我所知,在任何国际危机中,都没有一个明确的例子表明这种更新,尽管更新对手的决心或战斗意愿是非常常见的。
Second, one might argue that since both anarchy and private information plus incentives to misrepresent are constant features of international politics, neither can explain why states fail to strike a bargain preferable to war in one instance but not another. This argument is correct. But the task of specifying the causal mechanisms that explain the occurrence of war must precede the identification of factors that lead the mechanisms to produce one outcome rather than another in particular settings. That is, specific models in which commitment or information problems operate allow one to analyze how different variables (such as power shifts and cost-benefit ratios in the preventive war model) make for war in some cases rather than others. 其次,有人可能会说,既然无政府状态和私人信息加上歪曲事实的动机都是国际政治的恒定特征,那么两者都无法解释为什么国家在一种情况下无法达成优于战争的协议,而在另一种情况下却不能。这一论点是正确的。但是,在明确解释战争发生的因果机制之前,必须先找出导致这些机制在特定情况下产生一种结果而非另一种结果的因素。也就是说,在承诺或信息问题发生作用的特定模型中,我们可以分析不同的变量(如预防性战争模型中的权力转移和成本效益比)是如何在某些情况下而不是在其他情况下导致战争的。
This is the sense in which these two general mechanisms provide the foundations for a coherent rationalist or neorealist theory of war. A neorealist explanation for war shows how war could occur given the assumption of rational and unitary (“billiard ball”) states, the assumption made throughout this article. Consider any particular factor argued in the literature to be a cause of war under this assumption-for example, a failure to balance power, offensive advantages, multipolarity, or shifts in relative power. My claim is that showing how any such factor could cause war between rational states requires showing how the factor can occasion an unresolvable commitment or information problem in specific empirical circumstances. Short of this, the central puzzle posed by war, its costs, has not been addressed. 从这个意义上说,这两种一般机制为连贯的理性主义或新现实主义战争理论提供了基础。新现实主义对战争的解释说明了在理性和单一("台球")国家的假设下,战争是如何发生的。考虑文献中被认为是这一假设下战争起因的任何特定因素--例如,力量平衡的失败、进攻优势、多极化或相对力量的变化。我的主张是,要说明任何此类因素如何会导致理性国家之间的战争,就必须说明该因素如何会在特定的经验情况下导致无法解决的承诺或信息问题。如果做不到这一点,战争带来的核心问题--战争的代价--就没有得到解决。
Appendix 附录
Proofs of several claims made in the text are provided here. Throughout, “equilibrium” refers to the solution concept “perfect Bayesian equilibrium.” ^(69){ }^{69} 本文提供了文中若干主张的证明。在全文中,"均衡 "指的是 "完全贝叶斯均衡 "的求解概念。 ^(69){ }^{69}
Claim 1. Under the conditions on u_(A)(*)u_{A}(\cdot) and u_(B)(*)u_{B}(\cdot) given in the text, there will exist a set of issue resolutions such that both states prefer any one of these resolutions to a war. 主张 1.根据文中给出的关于 u_(A)(*)u_{A}(\cdot) 和 u_(B)(*)u_{B}(\cdot) 的条件,将存在一组议题决议,使得两国都倾向于其中任何一个决议而不是战争。
Proof. It is sufficient to show that there exists an interval [a,b]sube[0,1][a, b] \subseteq[0,1] such that for all x in[a,b],u_(A)(x) >= p-c_(A)x \in[a, b], u_{A}(x) \geq p-c_{A} and u_(B)(1-x) >= 1-p-c_(B)u_{B}(1-x) \geq 1-p-c_{B}. Choose an epsilon\epsilon such that 0 < epsilon <0<\epsilon<min{c_(A),c_(B)}\min \left\{c_{A}, c_{B}\right\} and define a=max{0,p-epsilon}a=\max \{0, p-\epsilon\} and b=min{p+epsilon,1}b=\min \{p+\epsilon, 1\}. Consider any x^(')in[a,b]x^{\prime} \in[a, b]. By weak concavity, u_(A)(x^(')) >= x^(')u_{A}\left(x^{\prime}\right) \geq x^{\prime} for all x^(')in[0,1]x^{\prime} \in[0,1]. Further, x^(') > p-c_(A)x^{\prime}>p-c_{A}, since x^(') >= a >=x^{\prime} \geq a \geqp-epsilon > p-c_(A)p-\epsilon>p-c_{A}, by definitions. Thus u_(A)(x^(')) > p-c_(A)u_{A}\left(x^{\prime}\right)>p-c_{A}. A similar argument shows that u_(B)(1-x^(')) > 1-p-c_(B)u_{B}\left(1-x^{\prime}\right)>1-p-c_{B} for all x^(')in[a,b]x^{\prime} \in[a, b]. 证明只需证明存在一个区间 [a,b]sube[0,1][a, b] \subseteq[0,1] ,使得对于所有 x in[a,b],u_(A)(x) >= p-c_(A)x \in[a, b], u_{A}(x) \geq p-c_{A} 和 u_(B)(1-x) >= 1-p-c_(B)u_{B}(1-x) \geq 1-p-c_{B} .选择一个 epsilon\epsilon ,使得 0 < epsilon <0<\epsilon<min{c_(A),c_(B)}\min \left\{c_{A}, c_{B}\right\} ,并定义 a=max{0,p-epsilon}a=\max \{0, p-\epsilon\} 和 b=min{p+epsilon,1}b=\min \{p+\epsilon, 1\} 。考虑任何 x^(')in[a,b]x^{\prime} \in[a, b] 。根据弱凹性, u_(A)(x^(')) >= x^(')u_{A}\left(x^{\prime}\right) \geq x^{\prime} 适用于所有 x^(')in[0,1]x^{\prime} \in[0,1] 。此外, x^(') > p-c_(A)x^{\prime}>p-c_{A} ,因为 x^(') >= a >=x^{\prime} \geq a \geqp-epsilon > p-c_(A)p-\epsilon>p-c_{A} ,根据定义。因此, u_(A)(x^(')) > p-c_(A)u_{A}\left(x^{\prime}\right)>p-c_{A} 。类似的论证表明 u_(B)(1-x^(')) > 1-p-c_(B)u_{B}\left(1-x^{\prime}\right)>1-p-c_{B} 适用于所有 x^(')in[a,b]x^{\prime} \in[a, b] 。
Take-it-or-leave-it international bargaining 要么接受要么放弃的国际谈判
The take-it-or-leave-it game is structured as follows. Nature draws a cost of war for state B,c_(B)\mathrm{B}, c_{B}, from a cumulative distribution H(z)H(z) on the nonnegative real numbers with a 要么接受,要么放弃的博弈结构如下。自然界从非负实数的累积分布 H(z)H(z) 中抽取状态 B,c_(B)\mathrm{B}, c_{B} 的战争成本,其值为 a
69. See Drew Fudenberg and Jean Tirole, Game Theory (Cambridge, Mass.: MIT Press, 1991), chap. 8. 69.见 Drew Fudenberg 和 Jean Tirole, Game Theory (Cambridge, Mass.: MIT Press, 1991),第 8 章。8.
strictly positive density function h(z)h(z) and a nondecreasing hazard rate h(z)//(1-H(z)).^(70)h(z) /(1-H(z)) .{ }^{70} State B observes c_(B)c_{B} but A does not. State A moves first, choosing a demand x in[0,1]x \in[0,1]. B observes the demand and chooses whether to fight or not. As discussed in the text, payoffs are (p-c_(A),1-p-c_(B))\left(p-c_{A}, 1-p-c_{B}\right) if B fights and (x,1-x)(x, 1-x) if B does not fight. 严格的正密度函数 h(z)h(z) 和非递减危险率 h(z)//(1-H(z)).^(70)h(z) /(1-H(z)) .{ }^{70} 国家 B 观察到了 c_(B)c_{B} 但国家 A 没有观察到。A 国首先行动,选择需求 x in[0,1]x \in[0,1] 。乙观察到需求并选择是否战斗。正如文中所讨论的,如果乙出战,回报为 (p-c_(A),1-p-c_(B))\left(p-c_{A}, 1-p-c_{B}\right) ;如果乙不出战,回报为 (x,1-x)(x, 1-x) 。
Claim 2. The take-it-or-leave-it game has a generically unique perfect Bayesian equilibrium in which AA demands x^(**)x^{*} and BB fights if and only if c_(B) < x-pc_{B}<x-p. The demand x^(**)x^{*} is defined as follows: (i) x^(**)=px^{*}=p if h(0) > 1//c_(A)h(0)>1 / c_{A}; (ii) x^(**)=1x^{*}=1 if 主张 2.当且仅当 c_(B) < x-pc_{B}<x-p 时,"要么接受要么放弃 "博弈有一个一般唯一的完全贝叶斯均衡,其中 AA 要求 x^(**)x^{*} 和 BB 打架。需求 x^(**)x^{*} 的定义如下:(i) x^(**)=px^{*}=p 如果 h(0) > 1//c_(A)h(0)>1 / c_{A} ; (ii) x^(**)=1x^{*}=1 如果
Moreover, the ex ante risk of war is always positive for small enough c_(A)c_{A} greater than zero. 此外,在 c_(A)c_{A} 大于零的足够小的范围内,战争的事前风险总是正数。
Proof. That B fights if and only if c_(B) < x-pc_{B}<x-p is immediately implied by subgame perfection. ^(71){ }^{71} Thus in any equilibrium the probability that B fights given demand xx is Pr(1-x < 1-p-c_(B))=Pr(c_(B) < x-p)=H(x-p)\operatorname{Pr}\left(1-x<1-p-c_{B}\right)=\operatorname{Pr}\left(c_{B}<x-p\right)=H(x-p). So A’s expected utility for demanding xx is always u_(A)(x)-=H(x-p)(p-c_(A))+(1-H(x-p))xu_{A}(x) \equiv H(x-p)\left(p-c_{A}\right)+(1-H(x-p)) x. Differentiation shows that u_(A)^(')(x)u_{A}^{\prime}(x) is nonnegative when 证明。亚博弈完美性立即暗示了当且仅当 c_(B) < x-pc_{B}<x-p 时 B 才会战斗。 ^(71){ }^{71} 因此,在任何均衡状态下,给定需求 xx 时,B 打起来的概率都是 Pr(1-x < 1-p-c_(B))=Pr(c_(B) < x-p)=H(x-p)\operatorname{Pr}\left(1-x<1-p-c_{B}\right)=\operatorname{Pr}\left(c_{B}<x-p\right)=H(x-p) 。因此,A 要求 xx 的预期效用总是 u_(A)(x)-=H(x-p)(p-c_(A))+(1-H(x-p))xu_{A}(x) \equiv H(x-p)\left(p-c_{A}\right)+(1-H(x-p)) x 。微分表明,在下列情况下, u_(A)^(')(x)u_{A}^{\prime}(x) 是非负的
Since the left-hand side is nondecreasing in xx and the right-hand side is strictly decreasing for x > px>p, we can conclude that (i)(i) if h(0) > 1//c_(A)h(0)>1 / c_{A} then x^(**)=px^{*}=p maximizes u_(A)(x)u_{A}(x); (ii) if 由于左手边在 xx 中是非递减的,而右手边在 x > px>p 中是严格递减的,因此我们可以得出结论:如果 h(0) > 1//c_(A)h(0)>1 / c_{A} 则 x^(**)=px^{*}=p 使 u_(A)(x)u_{A}(x) 最大;(ii) 如果
will be a unique maximum of u_(A)(x)u_{A}(x). Since the ex ante probability of war is H(x^(**)-p)H\left(x^{*}-p\right), it follows that only in case ( ii ) can this equal zero, and for small enough c_(A)c_{A} case (i) cannot obtain. 将是 u_(A)(x)u_{A}(x) 的唯一最大值。由于事前的战争概率是 H(x^(**)-p)H\left(x^{*}-p\right) ,因此只有在 ( ii ) 的情况下,这个概率才等于零,而对于足够小的 c_(A)c_{A} 情况,则不可能出现情况 (i)。
70. This condition is satisfied for a broad range of distributions. See Fudenberg and Tirole, Game Theory, p. 267. 70.这一条件在很大范围的分布中都能满足。见 Fudenberg 和 Tirole, Game Theory, 第 267 页。
71. The assumption that type c_(B)=x-pc_{B}=x-p chooses not to fight is immaterial. 71. c_(B)=x-pc_{B}=x-p 型选择不战斗的假设并不重要。
Take-it-or-leave-it international bargaining with communication (cheap talk) 要么接受要么放弃的国际沟通讨价还价(廉价谈话)
I now modify the take-it-or-leave-it game by allowing state B to choose (after observing its costs for war) an announcement ff from a large but finite set of possible speeches, FF. After the announcement, the game proceeds exactly as before with identical payoffs. To avoid some measure-theoretic complications, I will consider a finite approximation of the game analyzed above. Nature chooses state B’s type from the set T={c_(0),c_(1),c_(2),dots,c_(i),dots,c_(n)},n≫0T=\left\{c_{0}, c_{1}, c_{2}, \ldots, c_{i}, \ldots, c_{n}\right\}, n \gg 0, according to a discrete prior distribution h(*)h(\cdot), where h(c_(i))=Pr(c_(B)=c_(i))h\left(c_{i}\right)=\operatorname{Pr}\left(c_{B}=c_{i}\right) and h(c_(i)) > 0\mathrm{h}\left(c_{i}\right)>0 for all c_(i)in T\mathrm{c}_{i} \in T. The elements of TT satisfy c_(0)=0c_{0}=0 and (by convention) c_(i) < c_(j)c_{i}<c_{j} for all i < ji<j, where i,j in N-={0,1,2,dots,n}i, j \in N \equiv\{0,1,2, \ldots, n\}. Let H(c_(i))-=sum_(j=0)^(i-1)h(c_(j))H\left(c_{i}\right) \equiv \sum_{j=0}^{i-1} h\left(c_{j}\right) denote the prior probability that B 's cost is strictly less than c_(i)c_{i}, letting H(c_(0))=0H\left(c_{0}\right)=0. 现在,我修改了 "要么接受,要么放弃 "博弈,允许 B 国(在观察其战争成本后)从一个庞大但有限的可能演说集合 FF 中选择一个演说 ff 。在宣布之后,博弈将以完全相同的收益继续进行。为了避免一些计量理论上的复杂问题,我将考虑上述博弈的有限近似。自然根据离散先验分布 h(*)h(\cdot) 从集合 T={c_(0),c_(1),c_(2),dots,c_(i),dots,c_(n)},n≫0T=\left\{c_{0}, c_{1}, c_{2}, \ldots, c_{i}, \ldots, c_{n}\right\}, n \gg 0 中选择状态 B 的类型,其中 h(c_(i))=Pr(c_(B)=c_(i))h\left(c_{i}\right)=\operatorname{Pr}\left(c_{B}=c_{i}\right) 和 h(c_(i)) > 0\mathrm{h}\left(c_{i}\right)>0 适用于所有 c_(i)in T\mathrm{c}_{i} \in T 。对于所有 i < ji<j , TT 的元素满足 c_(0)=0c_{0}=0 和(按照惯例) c_(i) < c_(j)c_{i}<c_{j} ,其中 i,j in N-={0,1,2,dots,n}i, j \in N \equiv\{0,1,2, \ldots, n\} .让 H(c_(i))-=sum_(j=0)^(i-1)h(c_(j))H\left(c_{i}\right) \equiv \sum_{j=0}^{i-1} h\left(c_{j}\right) 表示 B 的成本严格小于 c_(i)c_{i} 的先验概率,让 H(c_(0))=0H\left(c_{0}\right)=0 .
In the game with talk, state B has a message strategy that gives a probability distribution on all possible messages in FF for each type in the set TT. Let m(f,c_(i))m\left(f, c_{i}\right) be the probability that type c_(i)c_{i} announces speech ff in a given equilibrium. State A’s demand strategy now associates with each ff a probability distribution over [0,1][0,1]. Let x(f)x(f) be the demand made by A on hearing ff whenever A does not mix given the announcement. In a given equilibrium of the game with talk, state A will form posterior beliefs about B’s type for each message ff. Let these be denoted h(c_(i),f)-=Pr(c_(B)=c_(i)∣f)h\left(c_{i}, f\right) \equiv \operatorname{Pr}\left(c_{B}=c_{i} \mid f\right), with H(c_(i),f)-=Sigma_(j=0)^(i-1)H\left(c_{i}, f\right) \equiv \Sigma_{j=0}^{i-1}h(c_(j),f)h\left(c_{j}, f\right). 在会话博弈中,B 国的信息策略给出了集合 TT 中每种类型在 FF 中所有可能信息的概率分布。让 m(f,c_(i))m\left(f, c_{i}\right) 成为 c_(i)c_{i} 类型在给定均衡中宣布讲话 ff 的概率。现在,A 国的需求策略将每个 ff 与 [0,1][0,1] 的概率分布联系起来。让 x(f)x(f) 成为 A 在听到 ff 时提出的需求,只要 A 在公告中没有混合。在有会谈博弈的给定均衡中,A 国将对每条信息 ff 形成关于 B 的类型的后验信念。让我们用 h(c_(i),f)-=Pr(c_(B)=c_(i)∣f)h\left(c_{i}, f\right) \equiv \operatorname{Pr}\left(c_{B}=c_{i} \mid f\right) 和 H(c_(i),f)-=Sigma_(j=0)^(i-1)H\left(c_{i}, f\right) \equiv \Sigma_{j=0}^{i-1}h(c_(j),f)h\left(c_{j}, f\right) 表示这些信念。
For convenience, I will assume that if B is indifferent between rejecting and accepting a demand xx, B accepts for sure. 为了方便起见,我假设如果 B 在拒绝和接受需求 xx 之间无动于衷,那么 B 肯定会接受。
Proposition. If x^(**)x^{*} is the unique equilibrium demand in the game without talk, then in any equilibrium of the game with talk in which state AA uses a pure strategy, (1) state AA demands x^(**)x^{*} regardless of the announcement; and (2) the ex ante risk of war is the same as in the game without talk. 命题。如果 x^(**)x^{*} 是无会谈博弈中唯一的均衡需求,那么在有会谈博弈的任何均衡中,如果国家 AA 使用纯策略,则:(1)国家 AA 无论宣布与否都要求 x^(**)x^{*} ;(2)战争的事前风险与无会谈博弈中的相同。
The proof follows from several lemmas. 证明源于几个定理。
LemmA 1. In any equilibrium A’s payoff is at least p, and in no equilibrium will A respond to any message with a demand that is sure to yield war. 假设 1:在任何均衡中,A 的报酬都至少为 p,而且在任何均衡中,A 都不会对任何信息做出肯定会导致战争的回应。
Proof of Lemma 1. If A sets x=px=p, all types of B accept for sure in any equilibrium (by subgame perfection), so A can assure itself pp in this way. If in some equilibrium, A chooses xx following a message ff such that war certainly follows, then A receives p-c_(A)p-c_{A}, but then A could deviate to x^(')=px^{\prime}=p and do strictly better. 定理 1 的证明。如果 A 设置了 x=px=p ,那么在任何均衡中,所有类型的 B 都肯定会接受(根据子博弈完美性),因此 A 可以通过这种方式向自己保证 pp 。如果在某个均衡中,A 在得到信息 ff 之后选择了 xx ,这样战争肯定会随之而来,那么 A 就会得到 p-c_(A)p-c_{A} ,但此时 A 可以偏离 x^(')=px^{\prime}=p ,并严格地做得更好。
Lemma 2. In any equilibrium, the demand xx is in the support of A’s demand strategy given a message fonly if there exists a c_(i)in Tc_{i} \in T such that x=p+c_(i)x=p+c_{i}. 定理 2:在任何均衡中,如果存在一个 c_(i)in Tc_{i} \in T 使得 x=p+c_(i)x=p+c_{i} 存在,则需求 xx 在给定信息 fonly 的 A 的需求策略的支持中。
Proof of Lemma 2. If not, then A might on hearing ff choose a demand x^(')x^{\prime} such that p+p+c_(i) < x^(') < p+c_(i+1)c_{i}<x^{\prime}<p+c_{i+1} for some ii. But then A could increase its payoff on hearing ff by deviating to x^('')=p+c_(i+1)x^{\prime \prime}=p+c_{i+1}, since doing so has no effect on the risk of war. 如果不是这样,那么 A 在听到 ff 时可能会选择一个需求 x^(')x^{\prime} ,使得 p+p+c_(i) < x^(') < p+c_(i+1)c_{i}<x^{\prime}<p+c_{i+1} 对于某个 ii 。但此时 A 可以通过偏离 x^('')=p+c_(i+1)x^{\prime \prime}=p+c_{i+1} 来增加听到 ff 时的报酬,因为这样做对战争风险没有影响。
Lemma 3. In any equilibrium in which AA does not mix, x(f)=kx(f)=k, a constant, for all messages f inF^(')f \in F^{\prime}, where F^(')F^{\prime} is the set of messages sent with positive probability in the given equilibrium. 定理 3.在 AA 不混合的任何均衡中, x(f)=kx(f)=k 是一个常数,适用于所有信息 f inF^(')f \in F^{\prime} ,其中 F^(')F^{\prime} 是在给定均衡中以正概率发送的信息集合。
Proof of Lemma 3. Let T_(f)-={c_(i):m(f,c_(i)) > 0}T_{f} \equiv\left\{c_{i}: m\left(f, c_{i}\right)>0\right\}. Suppose to the contrary that in some equilibrium there exist two distinct messages ff and f^(')f^{\prime} such that x(f) < x(f^('))x(f)<x\left(f^{\prime}\right). By Lemma 1 , both demands must be accepted with positive probability, implying that there are types c_(i)inT_(f)c_{i} \in T_{f} such that 1-x(f) >= 1-p-c_(i)1-x(f) \geq 1-p-c_{i} and types c_(j)inT_(f^('))c_{j} \in T_{f^{\prime}} such that 1-x(f^(')) >= 1-1-x\left(f^{\prime}\right) \geq 1-p-c_(j)p-c_{j}. But then any such c_(j)inT_(f^('))c_{j} \in T_{f^{\prime}} can do strictly better by deviating to ff, which gives it 1-x(f) > 1-x(f^('))1-x(f)>1-x\left(f^{\prime}\right). But this implies that x(f^('))x\left(f^{\prime}\right) is certainly rejected, contradicting Lemma 1. 定理 3 的证明设 T_(f)-={c_(i):m(f,c_(i)) > 0}T_{f} \equiv\left\{c_{i}: m\left(f, c_{i}\right)>0\right\} 。相反,假设在某个均衡中存在两个不同的信息 ff 和 f^(')f^{\prime} ,使得 x(f) < x(f^('))x(f)<x\left(f^{\prime}\right) .根据 Lemma 1,这两个要求都必须以正概率被接受,这意味着存在 c_(i)inT_(f)c_{i} \in T_{f} 类型,使得 1-x(f) >= 1-p-c_(i)1-x(f) \geq 1-p-c_{i} 和 c_(j)inT_(f^('))c_{j} \in T_{f^{\prime}} 类型,使得 1-x(f^(')) >= 1-1-x\left(f^{\prime}\right) \geq 1-p-c_(j)p-c_{j} 。但是,任何这样的 c_(j)inT_(f^('))c_{j} \in T_{f^{\prime}} 都可以通过偏离 ff 严格地做得更好,从而得到 1-x(f) > 1-x(f^('))1-x(f)>1-x\left(f^{\prime}\right) 。但这意味着 x(f^('))x\left(f^{\prime}\right) 肯定会被拒绝,这与 Lemma 1 相矛盾。
Proof of the proposition. Suppose x^(**)=p+c_(k)x^{*}=p+c_{k} is the unique equilibrium demand in the game without talk. Then c_(k)c_{k} is the only element of TT such that for all j in Nj \in N, 命题证明。假设 x^(**)=p+c_(k)x^{*}=p+c_{k} 是不说话博弈中唯一的均衡需求。那么 c_(k)c_{k} 是 TT 的唯一元素,使得对于所有 j in Nj \in N 、
Suppose to the contrary of the proposition that in the game with talk there is some other demand x^(')=p+c_(l),l!=kx^{\prime}=p+c_{l}, l \neq k, such that A demands x^(')x^{\prime} on hearing any message f inF^(')f \in F^{\prime} (by Lemma 3, any equilibrium without mixing by A must have this form). Then it must be that for each f inF^(')f \in F^{\prime} and for all j in Nj \in N, 假设与这一命题相反,在有交谈的博弈中,存在某种其他需求 x^(')=p+c_(l),l!=kx^{\prime}=p+c_{l}, l \neq k ,使得 A 在听到任何信息 f inF^(')f \in F^{\prime} 时都会要求 x^(')x^{\prime} (根据 Lemma 3,任何没有 A 混合的均衡都必须具有这种形式)。那么,对于每个 f inF^(')f \in F^{\prime} 和所有 j in Nj \in N ,都必须是这样、 H(c_(l),f)(p-c_(A))+(1-H(c_(l),f))(p+c_(l)) >=H\left(c_{l}, f\right)\left(p-c_{A}\right)+\left(1-H\left(c_{l}, f\right)\right)\left(p+c_{l}\right) \geq
By Bayes’s rule, h(c_(j),f)=h(c_(j))m(f,c_(j))//Pr(f)h\left(c_{j}, f\right)=h\left(c_{j}\right) m\left(f, c_{j}\right) / \operatorname{Pr}(f) and H(c_(j),f)=(1//Pr(f))Sigma_(i=0)^(j-1)h(c_(i),f)H\left(c_{j}, f\right)=(1 / \operatorname{Pr}(f)) \Sigma_{i=0}^{j-1} h\left(c_{i}, f\right), where {: Pr(f)=Sigma_(c in T)h(c)m(f,c))\left.\operatorname{Pr}(f)=\Sigma_{c \in T} h(c) m(f, c)\right) is the probability that B chooses ff in the equilibrium. Substitution into equation (2) and multiplication of both sides by Pr(f)\operatorname{Pr}(f) yields, for all j inj \inNN, 根据贝叶斯法则, h(c_(j),f)=h(c_(j))m(f,c_(j))//Pr(f)h\left(c_{j}, f\right)=h\left(c_{j}\right) m\left(f, c_{j}\right) / \operatorname{Pr}(f) 和 H(c_(j),f)=(1//Pr(f))Sigma_(i=0)^(j-1)h(c_(i),f)H\left(c_{j}, f\right)=(1 / \operatorname{Pr}(f)) \Sigma_{i=0}^{j-1} h\left(c_{i}, f\right) ,其中 {: Pr(f)=Sigma_(c in T)h(c)m(f,c))\left.\operatorname{Pr}(f)=\Sigma_{c \in T} h(c) m(f, c)\right) 是 B 在均衡中选择 ff 的概率。对于所有 j inj \inNN ,将其代入公式 (2) 并将两边乘以 Pr(f)\operatorname{Pr}(f) 即可得出、
Since equation (3) holds for each f inF^(')f \in F^{\prime}, we can sum both sides over all f inF^(')f \in F^{\prime} and the inequalities still hold. Thus, for all j in Nj \in N, 由于等式 (3) 对每个 f inF^(')f \in F^{\prime} 都成立,我们可以对所有 f inF^(')f \in F^{\prime} 两边求和,不等式仍然成立。因此,对于所有 j in Nj \in N 、
But this contradicts equation (1), the hypothesis that the game without talk has a unique equilibrium, x^(**)=p+c_(k)x^{*}=p+c_{k} where c_(k)!=c_(l)c_{k} \neq c_{l}. This proves the first part of the proposition. To see the second part, notice that if A demands x^(**)x^{*} after any message sent with positive probability, then all types c_(j) < x^(**)-pc_{j}<x^{*}-p will fight and all types c_(j) >= x^(**)-pc_{j} \geq x^{*}-p will accept the demand, so the ex ante risk of war is just H(x^(**)-p)H\left(x^{*}-p\right), as in the game without talk. 但这与等式(1)相矛盾,即假设不说话博弈有唯一的均衡, x^(**)=p+c_(k)x^{*}=p+c_{k} ,其中 c_(k)!=c_(l)c_{k} \neq c_{l} 。这证明了命题的第一部分。要看第二部分,请注意,如果 A 在以正概率发出任何信息后提出 x^(**)x^{*} 的要求,那么所有类型 c_(j) < x^(**)-pc_{j}<x^{*}-p 都会参战,而所有类型 c_(j) >= x^(**)-pc_{j} \geq x^{*}-p 都会接受该要求,因此事前的战争风险只是 H(x^(**)-p)H\left(x^{*}-p\right) ,就像在没有交谈的博弈中一样。
Remarks. (1) I conjecture that if there is unique equilibrium demand x^(**)x^{*} in the game without talk, then in no equilibrium of the game with talk will A ever mix over demands, so that the proposition should extend to all equilibria of the game with talk. While I have found no counterexamples, I have not been able to demonstrate conclusively that A cannot mix in some equilibrium. (2) Cheap talk can indeed matter in this game in a very limited way-equilibria may exist in which both A’s beliefs and the risk of war differ depending on the message sent. However, as shown in the proposition, in any such equilibrium the variation in A’s beliefs with different messages affects neither A’s behavior nor the ex ante (i.e., premessage) risk of war. 备注(1) 我猜想,如果在无对话博弈中存在唯一的均衡需求 x^(**)x^{*} ,那么在有对话博弈的任何均衡中,A 都不会混合需求,因此该命题应该扩展到有对话博弈的所有均衡。虽然我没有找到反例,但也无法确凿地证明 A 不能在某些均衡中混合。(2) 在这个博弈中,廉价的谈话确实可以以一种非常有限的方式起作用--可能存在这样的均衡,其中 A 的信念和战争风险都因发出的信息不同而不同。然而,如命题所示,在任何这样的均衡中,A 的信念随不同信息的变化既不会影响 A 的行为,也不会影响事前(即信息发出前)的战争风险。
An earlier version of this article was presented at the annual meetings of the American Political Science Association, Washington, D.C., 2-5 September 1993. The article draws in part on chapter 1 of James D. Fearon, “Threats to Use Force: Costly Signals and Bargaining in International Crises,” Ph.D. diss., University of California, Berkeley, 1992. Financial support of the Institute on Global Conflict and Cooperation of the University of California is gratefully acknowledged. For valuable comments I thank Eddie Dekel, Eric Gartzke, Atsushi Ishida, Andrew Kydd, David Laitin, Andrew Moravcsik, James Morrow, Randolph Siverson, Daniel Verdier, Stephen Walt and especially Charles Glaser and Jack Levy. 本文的早期版本于 1993 年 9 月 2-5 日在华盛顿特区举行的美国政治学协会年会上发表。本文部分借鉴了 James D. Fearon 的《威胁使用武力:国际危机中代价高昂的信号与谈判》(Threats to Use Force:国际危机中代价高昂的信号和讨价还价》,加州大学伯克利分校博士论文,1992 年。感谢加州大学全球冲突与合作研究所的资助。感谢 Eddie Dekel、Eric Gartzke、Atsushi Ishida、Andrew Kydd、David Laitin、Andrew Moravcsik、James Morrow、Randolph Siverson、Daniel Verdier、Stephen Walt 以及 Charles Glaser 和 Jack Levy 的宝贵意见。
Of course, arguments of the second sort may and often do presume rational behavior by individual leaders; that is, war may be rational for civilian or military leaders if they will enjoy various benefits of war without suffering costs imposed on the population. While I believe that “second-image” mechanisms of this sort are very important empirically, I do not explore them here. A more accurate label for the subject of the article might be “rational unitary-actor explanations,” but this is cumbersome. 当然,第二种论点可能而且往往确实假定领导人个人的行为是理性的;也就是说,如果文职或军事领导人能够享受战争带来的各种好处,而不付出强加给民众的代价,那么战争对他们来说可能是理性的。虽然我认为这类 "第二形象 "机制在经验上非常重要,但我在此不做探讨。文章主题的一个更准确的标签可能是 "理性的单位行为者解释",但这是累赘。
For the founding work of neorealism, see Kenneth Waltz, Theory of International Politics (Reading, Mass.: Addison-Wesley, 1979). For examples of theorizing along these lines, see Robert Jervis, “Cooperation Under the Security Dilemma,” World Politics 30 (January 1978), pp. 167-214; Stephen Walt, The Origins of Alliances (Ithaca, N.Y.: Cornell University Press, 1987); John J. Mearsheimer, “Back to the Future: Instability in Europe After the Cold War,” International Security 15 (Summer 1990), pp. 5-56; and Charles Glaser, “Realists as Optimists: Cooperation as Self-Help,” International Security 19 (Winter 1994/95), pp. 50-90. 关于新现实主义的奠基之作,见 Kenneth Waltz, Theory of International Politics (Reading, Mass.: Addison-Wesley, 1979)。有关这些理论的例子,见罗伯特-杰维斯(Robert Jervis),《安全困境下的合作》(Cooperation Under the Security Dilemma),《世界政治》第 30 期(1978 年 1 月),第 167-214 页;斯蒂芬-华特(Stephen Walt),《联盟的起源》(The Origins of Alliances)(纽约州伊萨卡:康奈尔大学出版社,1987 年);约翰-米尔斯海默(John J. Mearsheimer),《回到未来》(Back to the Future:International Security 15 (Summer 1990), pp:Cooperation as Self-Help," International Security 19 (Winter 1994/95), pp.
The sense of “mechanism” is similar to that proposed by Elster, although somewhat broader. See Jon Elster, Political Psychology (Cambridge: Cambridge University Press, 1993), pp. 1-7; and Jon Elster, Nuts and Bolts for the Social Sciences (Cambridge: Cambridge University Press, 1989), chap. 1. 机制 "的含义与埃尔斯特提出的含义相似,但更宽泛一些。见 Jon Elster, Political Psychology (Cambridge: Cambridge University Press, 1993), pp.1.
For an influential example of this common assumption see Glenn Snyder and Paul Diesing, Conflict Among Nations (Princeton, N.J.: Princeton University Press, 1977). 关于这种常见假设的一个有影响力的例子,见 Glenn Snyder 和 Paul Diesing,《国家间的冲突》(新泽西州普林斯顿:普林斯顿大学出版社,1977 年)。
See, for examples, Geoffry Blainey, The Causes of War (New York: Free Press, 1973); Michael Howard, The Causes of Wars (Cambridge, Mass.: Harvard University Press, 1983), especially chap. 1; and Arthur Stein, Why Nations Cooperate: Circumstance and Choice in International Relations (Ithaca, N.Y.: Cornell University Press, 1990), pp. 60-64. Even the case of World War I is contested; an important historical school argues that this was a wanted war. See Fritz Fisher, Germany’s Aims in the First World War (New York: Norton, 1967). 例如,见 Geoffry Blainey, The Causes of War (New York: Free Press, 1973);Michael Howard, The Causes of Wars (Cambridge, Mass.: Harvard University Press, 1983),特别是第 1 章;Arthur Stein, Why Nations Cooperate:Why Nations Cooperate:Why Nations Cooperate:Why Nations Cooperate:Why Nations Cooperate:Why Nations Cooperate:Why Nations Cooperate:Why Nations阿瑟-斯坦恩,《国家为何合作》:纽约州伊萨卡:康奈尔大学出版社,1990 年),第 60-64 页。甚至第一次世界大战的情况也存在争议;一个重要的历史学派认为这是一场被通缉的战争。见 Fritz Fisher, Germany's Aims in the First World War (New York: Norton, 1967)。
See John H. Herz, “Idealist Internationalism and the Security Dilemma,” World Politics 2 (January 1950), pp. 157-80; and Jervis, “Cooperation Under the Security Dilemma.” Anarchy is implicated in the security dilemma externality by the following logic: but for anarchy, states could commit to use weapons only for nonthreatening, defensive purposes. 见 John H. Herz, "Idealist Internationalism and the Security Dilemma," World Politics 2 (January 1950), pp.无政府状态与安全困境的外部性有如下联系:如果不是无政府状态,各国可以承诺只将武器用于非威胁性的防御目的。
Jervis, “Cooperation Under the Security Dilemma.” 杰维斯,"安全困境下的合作"。
For an analysis of the security dilemma that takes into account signaling, see Andrew Kydd, “The Security Dilemma, Game Theory, and World War I,” paper presented at the annual meeting of the American Political Science Association, Washington, D.C., 2-5 September 1993. 关于将信号传递考虑在内的安全困境分析,见安德鲁-基德(Andrew Kydd):《安全困境、博弈论与第一次世界大战》(The Security Dilemma, Game Theory, and World War I),1993 年 9 月 2-5 日在华盛顿特区举行的美国政治学协会年会上提交的论文。
The most developed exception I know of is found in Stephen Van Evera, “Causes of War,” Ph.D. diss., University of California, Berkeley, 1984, pp. 61-64. 据我所知,斯蒂芬-范-埃韦拉(Stephen Van Evera)的 "战争原因"(Causes of War)一文是最成熟的例外,该文发表于加利福尼亚大学伯克利分校博士论文,1984 年,第 61-64 页。
See Bruce Bueno de Mesquita, The War Trap (New Haven, Conn.: Yale University Press, 1981), and “The War Trap Revisited: A Revised Expected Utility Model,” American Political Science Review 79 (March 1985), pp. 157-76. For a generalization that introduces the idea of a bargaining range, see James D. Morrow, “A Continuous-Outcome Expected Utility Theory of War,” Journal of Conflict Resolution 29 (September 1985), pp. 473-502. Informal versions of the expected utility argument are everywhere. For example, Waltz’s statement that “A state will use force to attain its goals if, after assessing the prospects for success, it values those goals more than it values the pleasures of peace” appears in different ways in a great many works on war. See Waltz, Man, the State, and War, p. 60. 见 Bruce Bueno de Mesquita, The War Trap (New Haven, Conn.: Yale University Press, 1981) 和 "The War Trap Revisited:A Revised Expected Utility Model," American Political Science Review 79 (March 1985), pp.关于引入讨价还价范围概念的概括,见 James D. Morrow, "A Continuous-Outcome Expected Utility Theory of War," Journal of Conflict Resolution 29 (September 1985), pp.预期效用论的非正式版本随处可见。例如,华尔兹说:"如果一个国家在评估了成功的前景之后,对这些目标的重视超过了对和平乐趣的重视,那么它就会使用武力来实现其目标",这句话以不同的方式出现在许多关于战争的著作中。见华尔兹:《人、国家与战争》,第 60 页。
On the theory of bargaining with outside options, see Martin J. Osborne and Ariel Rubinstein, Bargaining and Markets (New York: Academic Press, 1990), chap. 3; Motty Perry, “An Example of Price Formation in Bilateral Situations,” Econometrica 50 (March 1986), pp. 313-21; and Robert Powell, “Bargaining in the Shadow of Power” (University of California, Berkeley, 1993, mimeographed). See also the analyses in R. Harrison Wagner, “Peace, War, and the Balance of Power,” American Political Science Review 88 (September 1994), pp. 593-607; and Wagner, “The Causes of Peace.” 3; Motty Perry, "An Example of Price Formation in Bilateral Situations," Econometrica 50 (March 1986), pp.另见 R. Harrison Wagner 的分析,"和平、战争与力量平衡",《美国政治科学评论》88(1994 年 9 月),第 593-607 页;以及 Wagner,"和平的原因"。
See, for example, Evan Luard, War in International Society (New Haven, Conn.: Yale University Press, 1992), p. 191. Schroeder notes that “patronage, bribes, and corruption” were “a major element” of eighteenth-century international relations. See Paul Schroeder, The Transformation of European Politics, 1763-1848 (Oxford: Oxford University Press, 1994), p. 579. 例如,参见 Evan Luard, War in International Society (New Haven, Conn.: Yale University Press, 1992),第 191 页。Schroeder 指出,"赞助、贿赂和腐败 "是 18 世纪国际关系的 "主要因素"。见 Paul Schroeder, The Transformation of European Politics, 1763-1848 (Oxford: Oxford University Press, 1994),第 579 页。
On Cuba, see Ernest May, Imperial Democracy (New York: Harper and Row, 1961), pp. 149-50. On the Louisiana purchase, military threats raised in the U.S. Senate apparently made Napoleon more eager to negotiate the sale. See E. Wilson Lyon, Louisiana in French Diplomacy (Norman: University of Oklahoma Press, 1934), pp. 179 and 214ff. 关于古巴,见 Ernest May, Imperial Democracy (New York: Harper and Row, 1961),第 149-50 页。关于路易斯安那州的购买,美国参议院提出的军事威胁显然使拿破仑更急于就出售进行谈判。见 E. Wilson Lyon, Louisiana in French Diplomacy (Norman: University of Oklahoma Press, 1934), pp.
Diego Gambetta, The Sicilian Mafia: The Business of Private Protection (Cambridge, Mass.: Harvard University Press, 1993), p. 214. Diego Gambetta, The Sicilian Mafia: The Business of Private Protection (Cambridge, Mass.: Harvard University Press, 1993), p. 214。
In one of the only articles on this problem, Morrow proposes a private information explanation for states’ failures to link issues in many disputes. See James D. Morrow, “Signaling Difficulties with Linkage in Crisis Bargaining,” International Studies Quarterly 36 (June 1992), pp. 153-72. 在关于这一问题的仅有的几篇文章之一中,莫罗对国家在许多争端中未能将问题联系起来提出了一种私人信息解释。见 James D. Morrow, "Signaling Difficulties with Linkage in Crisis Bargaining," International Studies Quarterly 36 (June 1992), pp.
See Jack Levy, “The Causes of War: A Review of Theories and Evidence,” in Philip E. Tetlock et al., eds., Behavior, Society, and Nuclear War, vol. 1 (Oxford: Oxford University Press, 1989), pp. 209-333. Recent work using limited-information game theory to analyze crisis bargaining places the strategic consequences of private information at the center of the analysis. See, for examples, Bruce Bueno de Mesquita and David Lalman, War and Reason (New Haven, Conn.: Yale University Press, 1992); James D. Fearon, “Domestic Political Audiences and the Escalation of International Disputes,” American Political Science Review 88 (September 1994), pp. 577-92; James D. Morrow, “Capabilities, Uncertainty, and Resolve: A Limited Information Model of Crisis Bargaining,” American Journal of Political Science 33 (November 1989), pp. 941-72; Barry Nalebuff, “Brinksmanship and Nuclear Deterrence: The Neutrality of Escalation,” Conflict Management and Peace Science 9 (Spring 1986), pp. 19-30; and Robert Powell, Nuclear Deterrence Theory: The Problem of Credibility (Cambridge: Cambridge University Press, 1990). 见 Jack Levy,"The Causes of War: A Review of Theories and Evidence," in Philip E. Tetlock et al., eds., Behavior, Society, and Nuclear War, vol. 1 (Oxford: Oxford University Press, 1989), pp.近期利用有限信息博弈论分析危机谈判的研究将私人信息的战略后果置于分析的中心。例如,见 Bruce Bueno de Mesquita 和 David Lalman, War and Reason (New Haven, Conn.: Yale University Press, 1992); James D. Fearon, "Domestic Political Audiences and the Escalation of International Disputes," American Political Science Review 88 (September 1994), pp:危机谈判的有限信息模型》,《美国政治学杂志》第 33 期(1989 年 11 月),第 941-72 页;Barry Nalebuff,《边缘技巧与核威慑》,《升级的中立性》,《美国政治学评论》第 88 期(1994 年 9 月),第 577-92 页:The Neutrality of Escalation," Conflict Management and Peace Science 9 (Spring 1986), pp:可信度问题》(剑桥:剑桥大学出版社,1990 年)。
Blainey, The Causes of War, p. 246. 布莱尼,《战争的原因》,第 246 页。
Ibid., p. 54. Blainey also blames patriotic and nationalistic fervor, leaders’ (irrational) tendency to surround themselves with yes-men, and crowd psychology. 同上,第 54 页。布雷尼还将责任归咎于爱国主义和民族主义的狂热、领导人(非理性地)倾向于身边都是唯唯诺诺的人以及从众心理。
See Ralph K. White, Nobody Wanted War: Misperception in Vietnam and Other Wars (New York: Doubleday/Anchor), chap. 7; Blainey, The Causes of War, p. 54; and Richard Ned Lebow, Between Peace and War: The Nature of International Crises (Baltimore, Md.: Johns Hopkins University Press, 1981), p. 247. 见 Ralph K. White, Nobody Wanted War: Misperception in Vietnam and Other Wars (New York: Doubleday/Anchor), chap.7; Blainey, The Causes of War, p. 54; and Richard Ned Lebow, Between Peace and War: The Nature of International Crises (Baltimore, Md.: Johns Hopkins University Press, 1981), p. 247.
John C. Harsanyi, “Games with Incomplete Information Played By ‘Bayesian’ Players, Part III,” Management Science 14 (March 1968), pp. 486-502. John C. Harsanyi,"由'贝叶斯'玩家进行的不完全信息博弈,第三部分",《管理科学》第 14 期(1968 年 3 月),第 486-502 页。
Aumann observed an interesting implication of this doctrine: genuinely rational agents cannot “agree to disagree,” in the sense that it cannot be commonly known that they are rational and that they hold different estimates of the likelihood of some uncertain event. See Robert Aumann, “Agreeing to Disagree,” The Annals of Statistics 4 (November 1976), pp. 1236-39. Emerson Niou, Peter Ordeshook, and Gregory Rose note that this implies that rational states cannot agree to disagree about the probability that one or the other would win in a war in The Balance of Power: Stability in the International System (Cambridge: Cambridge University Press, 1989), p. 59. 奥曼观察到这一学说的一个有趣含义:真正理性的代理人不可能 "同意不同意",也就是说,人们不可能普遍知道他们是理性的,而且他们对某些不确定事件的可能性持有不同的估计。见 Robert Aumann,"Agreeing to Disagree",The Annals of Statistics 4(1976 年 11 月),第 1236-39 页。Emerson Niou、Peter Ordeshook 和 Gregory Rose 在《力量平衡:国际体系中的稳定》(剑桥:剑桥大学出版社,1989 年)中指出,这意味着理性国家无法就战争中一方或另一方获胜的概率达成一致意见。
On bounded rationality, see Herbert A. Simon, “A Behavioral Model of Rational Choice,” Quarterly Journal of Economics 69 (February 1955), pp. 99-118. 关于有界理性,见赫伯特-A-西蒙:《理性选择的行为模式》,《经济学季刊》第 69 期(1955 年 2 月),第 99-118 页。
This analysis runs exactly parallel to work in law and economics on pretrial bargaining in legal disputes. Early studies explained costly litigation as resulting from divergent expectations about the likely trial outcome, while in more recent work such expectations derive from private information about the strength of one’s case. For a review and references, see Robert D. Cooter and Daniel L. Rubinfeld, “Economic Analysis of Legal Disputes and Their Resolution,” Journal of Economic Literature 27 (September 1989), pp. 1067-97. 这一分析与法律和经济学界关于法律纠纷审前讨价还价的研究完全一致。早期的研究将高昂的诉讼费用解释为对可能的审判结果的不同预期所致,而在近期的研究中,这种预期则来自于对自己案件胜负的私人信息。有关评论和参考文献,请参见 Robert D. Cooter 和 Daniel L. Rubinfeld 合著的《法律纠纷及其解决的经济分析》,《经济文献杂志》第 27 期(1989 年 9 月),第 1067-97 页。
For examples and discussion on this point, see Fearon, “Threats to Use Force,” chap. 3. 关于这一点的例子和讨论,见 Fearon,"威胁使用武力",第 3 章。3.
See the Appendix for proofs of these claims. Cheap talk announcements can affect outcomes in some bargaining contexts. For an example from economics, see Joseph Farrell and Robert Gibbons, “Cheap Talk Can Matter in Bargaining,” Journal of Economic Theory 48 (June 1989), pp. 221-37. These authors show how cheap talk might credibly signal a willingness to negotiate seriously that then affects subsequent terms of trade. For an example from international relations, see James D. Morrow, “Modeling the Forms of International Cooperation: Distribution Versus Information,” International Organization 48 (Summer 1994), pp. 387-423. 有关这些说法的证明,请参见附录。在某些讨价还价的情况下,廉价谈话公告会影响谈判结果。经济学中的一个例子,见 Joseph Farrell 和 Robert Gibbons,"Cheap Talk Can Matter in Bargaining," Journal of Economic Theory 48 (June 1989),第 221-37 页。这两位作者展示了 "低声下气 "如何以可信的方式表达出认真谈判的意愿,进而影响随后的贸易条件。国际关系中的一个例子,见 James D. Morrow, "Modeling the Forms of International Cooperation:Distribution Versus Information," International Organization 48 (Summer 1994), pp.
The conclusion is likewise altered if the possibility of repeated interactions in sufficiently similar contexts is great enough that reputation building can be supported. 如果在足够相似的环境中重复互动的可能性足够大,以至于可以支持声誉的建立,那么结论同样也会改变。
On signaling costs in crises and audience costs in particular, see Fearon, “Threats to Use Force,” and “Domestic Political Audiences and the Escalation of International Disputes.” For an excellent analysis of international signaling in general, see Robert Jervis, The Logic of Images in International Relations (Princeton, N.J.: Princeton University Press, 1970). 关于危机中的信号传递成本,尤其是受众成本,见 Fearon, "Threats to Use Force," and "Domestic Political Audiences and the Escalation of International Disputes"。关于国际信号传递的总体分析,见 Robert Jervis,《国际关系中的形象逻辑》(普林斯顿:普林斯顿大学出版社,1970 年)。
For developed models that make this point, see James Fearon, “Deterrence and the Spiral Model: The Role of Costly Signals in Crisis Bargaining,” paper presented at the annual meeting of the American Political Science Association, 30 August-2 September 1990, San Francisco, Calif.; Fearon, “Domestic Political Audiences and the Escalation of International Disputes”; Morrow, “Capabilities, Uncertainty, and Resolve”; Nalebuff, “Brinkmanship and Nuclear Deterrence”; and Powell, Nuclear Deterrence Theory. 关于提出这一点的发达模型,见 James Fearon,"Deterrence and the Spiral Model:The Role of Costly Signals in Crisis Bargaining," paper presented at the annual meeting of the American Political Science Association, 30 August-2 September 1990, San Francisco, Calif.;Fearon, "Domestic Political Audiences and the Escalation of International Disputes";Morrow, "Capabilities, Uncertainty, and Resolve";Nalebuff, "Brinkmanship and Nuclear Deterrence";and Powell, Nuclear Deterrence Theory.
Luigi Albertini, The Origins of the War of 1914, vol. 2 (London: Oxford University Press, 1953), pp. 183-87. Luigi Albertini,《1914 年战争的起源》,第 2 卷(伦敦:牛津大学出版社,1953 年),第 183-87 页。
Ibid., p. 187. 同上,第 187 页。
Ibid., p. 158. For the full text of the cable, see Karl Kautsky, comp., German Documents Relating to the Outbreak of the World War (New York: Oxford University Press, 1924), doc. no. 71, p. 130. 同上,第 158 页。电报全文见卡尔-考茨基编译的《与世界大战爆发有关的德国文件》(纽约:牛津大学出版社,1924 年),第 71 号文件,第 130 页。
Konrad Jarausch, “The Illusion of Limited War: Chancellor Bethmann Hollweg’s Calculated Risk,” Central European History 2 (March 1969), pp. 48-76. The quotation is drawn from p. 65. Konrad Jarausch, "The Illusion of Limited War: Chancellor Bethmann Hollweg's Calculated Risk," Central European History 2 (March 1969), pp.引文摘自第 65 页。
See L. C. F. Turner, Origins of the First World War (New York: Norton, 1970), p. 101; and Jarausch, “The Illusion of Limited War,” p. 63. Trachtenberg writes that “one of Bethmann’s basic goals was for Germany to avoid coming across as the aggressor.” See Marc Trachtenberg, History and Strategy (Princeton, N.J.: Princeton University Press, 1991), p. 90. 见 L. C. F. Turner, Origins of the First World War (New York: Norton, 1970), p. 101;Jarausch, "The Illusion of Limited War," p. 63。特拉滕贝格写道,"贝特曼的基本目标之一是避免德国被视为侵略者"。见 Marc Trachtenberg, History and Strategy (Princeton, N.J.: Princeton University Press, 1991), p. 90。
Albertini concludes that “on the evening of the 27th all the Chancellor sought to do was to throw dust in the eyes of Grey and lead him to believe that Berlin was seriously trying to avert a conflict, that if war broke out it would be Russia’s fault and that England could therefore remain neutral.” See Albertini, The Origins of the War of 1914, vol. 1, pp. 444-45. See also Turner, Origins of the First World War, p. 99. 阿尔贝蒂尼的结论是:"27 日晚,首相所要做的就是向格雷的眼睛里扔灰尘,让他相信柏林正在认真地试图避免冲突,如果战争爆发,那将是俄国的过错,因此英国可以保持中立"。见阿尔贝蒂尼:《1914 年战争的起源》,第 1 卷,第 444-45 页。另见特纳:《第一次世界大战的起源》,第 99 页。
See J. A. White, The Diplomacy of the Russo-Japanese War (Princeton, N.J.: Princeton University Press, 1964), pp. 142-43; and Ian Nish, The Origins of the Russo-Japanese War (London: Longman, 1985), pp. 241-42. 参见 J. A. White,《日俄战争的外交》(普林斯顿:普林斯顿大学出版社,1964 年),第 142-43 页;以及 Ian Nish,《日俄战争的起源》(伦敦:朗文出版社,1985 年),第 241-42 页。
See, for example, William Langer, “The Origins of the Russo-Japanese War,” in Carl Schorske and Elizabeth Schorske, eds., Explorations in Crisis (Cambridge: Harvard University Press, 1969), p. 44. 例如,见 William Langer,"日俄战争的起源",载于 Carl Schorske 和 Elizabeth Schorske 编辑的《危机探索》(剑桥:哈佛大学出版社,1969 年),第 44 页。
For some examples, see Fearon, “Threats to Use Force,” chap. 3. For a formal version of reputational dynamics due to private information, see Barry Nalebuff, “Rational Deterrence in an Imperfect World,” World Politics 43 (April 1991), pp. 313-35. 一些例子见 Fearon,"威胁使用武力",第 3 章。3.关于私人信息导致的声誉动态的正式版本,见 Barry Nalebuff,《不完美世界中的理性威慑》,《世界政治》第 43 期(1991 年 4 月),第 313-35 页。
See, for examples, Robert Jervis, Perception and Misperception in International Politics (Princeton, N.J.: Princeton University Press, 1976), pp. 62-67; Barry Posen, The Sources of Military Doctrine (Ithaca, N.Y.: Cornell University Press, 1984), pp. 16-17; and Charles Glaser, “The Political Consequences of Military Strategy,” World Politics 44 (July 1992), p. 506. 例如,见 Robert Jervis, Perception and Misperception in International Politics (Princeton, N.J.: Princeton University Press, 1976), 第 62-67 页;Barry Posen, The Sources of Military Doctrine (Ithaca, N.Y.: Cornell University Press, 1984), 第 16-17 页;以及 Charles Glaser, "The Political Consequences of Military Strategy," World Politics 44 (July 1992), 第 506 页。
For dynamic game models that demonstrate this, see Robert Powell, “Absolute and Relative Gains in International Relations Theory,” American Political Science Review 85 (December 1991), pp. 1303-20; and James D. Fearon, “Cooperation and Bargaining Under Anarchy,” (University of Chicago, 1994, mimeographed). On tit-for-tat and the impact of the shadow of the future, see Robert Axelrod, The Evolution of Cooperation (New York: Basic Books, 1984); and Kenneth Oye, ed., Cooperation Under Anarchy (Princeton, N.J.: Princeton University Press, 1986). 关于证明这一点的动态博弈模型,见罗伯特-鲍威尔:《国际关系理论中的绝对收益和相对收益》,《美国政治科学评论》第 85 期(1991 年 12 月),第 1303-20 页;以及詹姆斯-费伦:《无政府状态下的合作与谈判》(芝加哥大学,1994 年,油印本)。关于针锋相对和未来阴影的影响,见罗伯特-阿克塞尔罗德:《合作的演变》(纽约:基本书局,1984 年);肯尼斯-奥耶编:《无政府状态下的合作》(新泽西州普林斯顿:普林斯顿大学出版社,1986 年)。
This argument about military variance runs counter to the usual hypothesis that offensive advantages foster war. For a discussion and an empirical assessment, see James D. Fearon, “Offensive Advantages and War since 1648,” paper presented at the annual meeting of the International Studies Association, 21-25 February 1995. On the offense-defense balance and war, see Jervis, “Cooperation Under the Security Dilemma”; and Van Evera, “Causes of War,” chap. 3. 关于军事差异的这一论点与进攻性优势助长战争的通常假设背道而驰。有关讨论和经验评估,见 James D. Fearon,"Offensive Advantages and War since 1648"(《1648 年以来的进攻优势与战争》),在国际研究协会年会上提交的论文,1995 年 2 月 21-25 日。关于攻守平衡与战争,见 Jervis,"安全困境下的合作";Van Evera,"战争原因",第 3 章。3.
For the argument about leaders’ views on first-strike advantages in 1914, see Stephen Van Evera, “The Cult of the Offensive and the Origins of the First World War,” International Security 9 (Summer 1984), pp. 58-107. 关于 1914 年领导人对先发制人优势的看法,见 Stephen Van Evera,"The Cult of the Offensive and the Origins of the First World War",International Security 9 (Summer 1984),pp.
See, for example, Trachtenberg, History and Strategy, p. 90. 例如,见 Trachtenberg,《历史与战略》,第 90 页。
This is suggested by results in Roger Myerson and Mark Satterthwaite, “Efficient Mechanisms for Bilateral Trading,” Journal of Economic Theory 29 (April 1983), pp. 265-81. Roger Myerson 和 Mark Satterthwaite 的 "Efficient Mechanisms for Bilateral Trading"(双边贸易的有效机制),Journal of Economic Theory 29(1983 年 4 月),第 265-81 页。
Schelling suggested that efficient coordination in stag hunt-like preemption problems might be prevented by a rational dynamic of “reciprocal fear of surprise attack.” See Thomas Schelling, The Strategy of Conflict (Cambridge, Mass.: Harvard University Press, 1960), chap. 9. Powell has argued that no such dynamic exists between rational adversaries. See Robert Powell, “Crisis Stability in the Nuclear Age,” American Political Science Review 83 (March 1989), pp. 61-76. 谢林认为,在类似猎鹿的先发制人问题中,"对突然袭击的相互恐惧 "这一理性动态可能会阻碍有效协调。见 Thomas Schelling, The Strategy of Conflict (Cambridge, Mass.: Harvard University Press, 1960), chap.9.鲍威尔认为,理性对手之间不存在这种动态。见 Robert Powell, "Crisis Stability in the Nuclear Age," American Political Science Review 83 (March 1989), pp.
Taylor, The Struggle for Mastery in Europe, 1848-1918 (London: Oxford University Press, 1954), p. 166. Carr held a similar view: "The most serious wars are fought in order to make one’s Taylor, The Struggle for Mastery in Europe, 1848-1918 (London: Oxford University Press, 1954), p. 166。卡尔持类似观点:"最严重的战争是为了使自己的利益得到最大化而进行的。
own country militarily stronger or, more often, to prevent another country from becoming militarily stronger." See E. H. Carr, The Twenty Years’ Crisis, 1919-1939 (New York: Harper and Row, 1964), pp. 111-12. 自己的国家在军事上变得更强大,或者更常见的是,防止另一个国家在军事上变得更强大"。见 E. H. Carr, The Twenty Years' Crisis, 1919-1939 (New York: Harper and Row, 1964), pp.
59. To my knowledge, Van Evera is the only scholar whose treatment of preventive war analyzes at some length how issues of credible commitment intervene. The issue is raised by both Snyder and Levy. See Van Evera, “Causes of War,” pp. 62-64; Jack Snyder, “Perceptions of the Security Dilemma in 1914,” in Robert Jervis, Richard Ned Lebow, and Janice Gross Stein, eds., Psychology and Deterrence (Baltimore, Md.: Johns Hopkins University Press, 1985), p. 160; and Jack Levy, “Declining Power and the Preventive Motivation for War,” World Politics 40 (October 1987), p. 96. 59.据我所知,范-埃韦拉是唯一一位在论述预防性战争时用一定篇幅分析可信承诺问题如何介入的学者。斯奈德和列维都提出了这一问题。参见 Van Evera,"战争原因",第 62-64 页;Jack Snyder,"1914 年对安全困境的看法",载于 Robert Jervis、Richard Ned Lebow 和 Janice Gross Stein 编辑的《心理学与威慑》(巴尔的摩:约翰-霍普金斯大学出版社,1985 年),第 160 页;以及 Jack Levy,"国力衰退与战争的预防动机",《世界政治》第 40 期(1987 年 10 月),第 96 页。
According to Hiro, President Bush’s main concern at the first National Security Council meeting following the invasion of Kuwait was the potential increase in Iraq’s economic leverage and its likely influence on an “already gloomy” U.S. economy. See Dilip Hiro, Desert Shield to Desert Storm: The Second Gulf War (London: Harper-Collins, 1992), p. 108. 根据 Hiro 的说法,在入侵科威特后的第一次国家安全委员会会议上,布什总统的主要担忧是伊拉克经济影响力的潜在增加及其可能对 "已经黯淡的 "美国经济产生的影响。见 Dilip Hiro, Desert Shield to Desert Storm:第二次海湾战争》(伦敦:哈珀-柯林斯出版社,1992 年),第 108 页。
On compensation, see Edward V. Gulick, Europe’s Classical Balance of Power (New York: Norton, 1955), pp. 70-72; and Paul W. Schroeder, The Transformation of European Politics, 1763-1848, pp. 6-7. 关于补偿,见 Edward V. Gulick, Europe's Classical Balance of Power (New York: Norton, 1955), 第 70-72 页;以及 Paul W. Schroeder, The Transformation of European Politics, 1763-1848, 第 6-7 页。
See Trachtenberg, History and Strategy, pp. 56-59; Albertini, The Origins of the War of 1914, vol. 2, pp. 129-30; Turner, Origins of the First World War, chap. 4; James Joll, The Origins of the First World War (London: Longman, 1984), p. 87; and Van Evera, “The Cult of the Offensive and the Origins of the First World War,” pp. 79-85. 见 Trachtenberg, History and Strategy, pp.4; James Joll, The Origins of the First World War (London: Longman, 1984), p. 87; and Van Evera, "The Cult of the Offensive and the Origins of the First World War," pp.
Samuel Williamson, “The Origins of World War I,” Journal of Interdisciplinary History (Spring 1988), pp. 795-818 and pp. 797-805 in particular; and D. C. B. Lieven, Russia and the Origins of the First World War (New York: St. Martins, 1983), pp. 38-49. Samuel Williamson,"第一次世界大战的起源",《跨学科历史杂志》(1988 年春季),第 795-818 页,特别是第 797-805 页;以及 D. C. B. Lieven,《俄罗斯与第一次世界大战的起源》(纽约:圣马丁,1983 年),第 38-49 页。
Jack S. Levy, “Preferences, Constraints, and Choices in July 1914,” International Security 15 (Winter 1990/91), pp. 234-36. Jack S. Levy, "Preferences, Constraints, and Choices in July 1914," International Security 15 (Winter 1990/91), pp.
Levy argues that preventive considerations are rarely themselves sufficient to cause war. See Levy, “Declining Power and the Preventive Motivation for War.” Levy 认为,预防性考虑本身很少足以引发战争。见 Levy,"国力衰退与战争的预防动机"。
See for example Kalevi J. Holsti, Peace and War: Armed Conflicts and International Order 1648-1989 (Cambridge: Cambridge University Press, 1991); and John Vasquez, The War Puzzle (Cambridge: Cambridge University Press, 1993). 例如,见 Kalevi J. Holsti, Peace and War: Armed Conflicts and International Order 1648-1989 (Cambridge: Cambridge University Press, 1991); and John Vasquez, The War Puzzle (Cambridge: Cambridge University Press, 1993)。
The argument is formalized in work in progress by the author, where it is shown that the conditions under which war will occur are restrictive: the states must be unable to continuously adjust the odds of victory by dividing up and trading the land. In other words, the smallest feasible territorial transfer must produce a discontinuously large change in a state’s military chances for war to be possible. See also Wagner, “Peace, War, and the Balance of Power,” p. 598, on this commitment problem. 作者正在进行的研究正式提出了这一论点,研究表明,发生战争的条件具有限制性:国家必须无法通过分割和交易土地来不断调整胜算。换句话说,最小的可行领土转让必须对一国的军事胜算产生不连续的巨大变化,战争才有可能发生。关于这一承诺问题,另见瓦格纳:《和平、战争与均势》,第 598 页。
See Max Jakobson, The Diplomacy of the Winter War: An Account of the Russo-Finnish Conflict, 1939-1940 (Cambridge, Mass.: Harvard University Press, 1961), pp. 135-39; and Van Evera, “Causes of War,” p. 63. Private information and incentives to misrepresent also caused problems in the bargaining here. See Fearon, “Threats to Use Force,” chap. 3. 见 Max Jakobson, The Diplomacy of the Winter War: An Account of the Russo-Finnish Conflict, 1939-1940 (Cambridge, Mass.: Harvard University Press, 1961), pp.私人信息和虚假陈述的动机也给谈判带来了问题。见 Fearon,"威胁使用武力",第 3 章。3.