Summer 2018 (32.2) Feature

Threats and Coercive Diplomacy: An Ethical Analysis

Article 2(4) of the UN Charter states that “all Members shall refrain in their international relations from the threat or use of force against the territorial integrity or political independence of any state. . . .” Thus, threatening attack and the corresponding application of armed force are equally prohibited under international law, according to this famous and precedent-setting formulation.1 But are threats as morally and legally problematic as the actual use of armed force? Prima facie it would seem not. Threats, of themselves, cause no direct and tangible harm. And when one state party by way of a threat openly expresses its extreme disapproval of another’s behavior, a space is thereby opened for negotiation and compromise.2 ed., Vol. 1 (New York: Oxford University Press, 2002), p. 124. In this way, threats can prevent the outbreak of war.

Nonetheless, the dark side of threats cannot be denied. Especially where power asymmetries are great, threats can be an effective means of domination. The nations of East Asia still bemoan the “gunboat diplomacy” by which Western colonial powers would send warships with powerful cannons into urban ports, threatening to shell populated areas unless the rulers signed treaties allowing for the establishment of extraterritorial rights.3 More economical than warfare, threats can bring about the same net result—submission—and for this reason they are often placed alongside armed attack under the common heading of “aggression.”

Threats have traditionally occupied a significant place in diplomacy, representing an uncertain ground between consensual agreements on the one hand and armed confrontation on the other. As Thucydides immortalized in the “Melian Dialogue,” and as Hitler exemplified in his negotiations with the Czechs, Poles, and British, threats have long been integral to the practice of diplomacy, statecraft, and generalship. Reaching an apogee in the run-up to the Second World War, the state practice of issuing ultimatums went into decline after the war’s end.4 Beginning with the Yugoslav crisis of 1999, however, ultimatums—particularly as issued by the United States—have made a remarkable comeback. They have been widely used to exert pressure, formulate expectations, and set the threshold for going to war. The Rambouillet talks in 1999 to resolve the status of Kosovo were conducted under the threat of NATO air strikes. In 2001, U.S. President George W. Bush issued an ultimatum to the Taliban to hand over Osama bin Laden. Two years later he threatened Saddam Hussein with invasion. In 2011, President Barack Obama issued an ultimatum to Muammar Qaddafi; and a year later he declared that the use of chemical weapons by Bashar al-Assad in Syria would cross a red line that, he insinuated, would entail a U.S. military response. Most recently, President Donald Trump has exchanged threats with Chairman Kim Jong-un over North Korea’s nuclear weapons program.

The normative evaluation of such coercive diplomacy lies at the crossroads of law and ethics. Within public international law, threats are most often treated as a modality of the jus ad bellum (that part of the law that regulates resort to armed force), usually under the heading of “ultimatums.” Another locus for juridical discussion is the law of treaties, particularly the Vienna Convention of 1969, which provides for the voiding of treaties if they have been signed under duress.5

Despite their prevalence in the international sphere, threats have been little discussed in the literature on just war, apart from the specialized case of nuclear deterrence. This lacuna is regrettable. Most conventional military engagements do not aim simply at achieving victory on the battlefield; they are strategically nested in a broader context where the goal is to influence the political decision-making of the adversary. Such military engagements are typically preceded by verbal threats. And once carried out, these engagements themselves typically assume the character of a threat, since their function is to signal the prospect of renewed attack and to structure the adversary’s incentives so its choices bend to a predetermined outcome. In this respect, threats pertain to the “idiom of military action” (to borrow a phrase from Thomas Schelling) and accordingly merit inclusion within the ethics of war.6 ed. (New Haven: Yale University Press, 2008), p. 126.

In what follows we elucidate how threats represent a distinctive category within the broader field of just war ethics. Although our treatment is not specifically concerned with international law, it overlaps with the legal debate and should be of utility to lawyers as it lays the groundwork for a general normative evaluation of threats.7

Our first section is concerned with the definition and taxonomy of threats. The second section formulates some standard ethical objections that may be raised vis-à-vis the use of threats by state actors, and then goes on to employ just war categories to explain how threats may be morally assessed. Our overall aim is to explain how threats in the international sphere represent a distinctive category that warrants a just war analysis. The attendant moral issues are highly complex, and we do not purport to resolve them all. Our goal is somewhat more modest, namely, to provide a map of this morally fraught domain.


To threaten is to perform a special kind of speech act whereby one person (P) tells another (Q) that she will intentionally bring about some harm x (or allow it to occur) unless Q does (or refrains from doing) the action y. Or, to reformulate this point more simply, we can say (borrowing from Jon Hovi) that “a threat is a contingent assertion signaling an intention to hurt somebody—physically, economically or otherwise—unless that somebody acts in the way prescribed by the threatener.”8 Though we employ language such as “speech acts” and “assertions,” we recognize nonetheless that threats are oftentimes communicated through actions rather than speech, such as when troops are deployed to a border zone so the adversary will understand that an armed attack will be initiated unless previously stated demands are met. Such moves implicitly evoke the communicative nature of a speech act, and hence exemplify the same basic structure.

Donald Trump and Kim Jong-Un have exchanged threats over North Korea's nuclear weapons program.

In the standard fivefold classification of speech acts, threats combine aspects of two categories: namely, commissives (whereby the speaker commits herself in a determinate way, as when a promise is made); and directives (whereby the purpose is to get the hearer to do something, as when soldiers are ordered to advance on the battlefield).9 By contrast, the related act of warning would fall rather into the linguistic category of assertives, whereby a speaker simply commits herself to the truth of what she says, but without committing herself to any further action.10 Martin Gunderson employs the following colorful illustration to drive home this distinction: “If a dancer informs his partner that he will accidentally step on her foot unless she dances in time with the music, he is warning her. If, however, he informs her that he will intentionally step on her foot unless she dances in time with the music, he is threatening her.”11 Schelling sums this up when he writes that “a warning does not constitute a move,”12 whereas “the threat is the commitment to a strategy for a second move.”13

Threats Broadly Construed: Posing Harm

Like warnings, threats are sometimes couched as assertives, in which case they serve to indicate how a person, thing, or event poses harm of some sort. For example, it is common nowadays to speak of “security threats,” a term that can encompass anything such as the accidental launching of an armed missile, the risk of nuclear material falling into the hands of terrorists, the danger that one state will preemptively attack another on erroneous grounds, the probability that global warming will lead to increased rates of violent conflict, and so forth. While intentionality can be identified at different levels in these various cases, the related enunciations are all within the category of assertives.

An “existential threat” also represents a broad usage, as when inveterate ill-will is expressed in the form of an assertive speech act, such as, “This people has no right to statehood, and all necessary steps will be undertaken to impede their claim to independence.” For an existential threat to have the combined character of a commissive and directive—in other words, for one polity intentionally to threaten massive harm against another in order to get the other polity to act in a certain way—the speech act in question would need to be uttered in the form of a conditional do (or refrain from) y or suffer x, as when the Athenian ambassadors insinuated to the leaders of Melos that only by their acceptance of political subservience would destruction (or enslavement) of the island’s inhabitants be forestalled.14

Finally, also to be classed as threats in the broadest sense are the laws enunciated by states for the regulation of conduct within their jurisdictions. It is a necessary aspect of such a law “that it is coercively enforced through the use, when necessary, of physical means to compel compliance.”15 That domestic law is commonly viewed as coercive lends credence to the idea that coercion—hence threats—can justifiably be employed within the sphere of interpersonal and interstate relations. By the same token, however, the example of law highlights how legitimate authority is an indispensable condition for the moral exercise of coercion and, by extension, the issuance of threats. Solely legislators, acting with due process, are entitled to frame laws; likewise, threats of physical harm are justifiably emitted only by individuals possessed of the requisite authority and who act in accordance with a set of (at least tacit) procedures.

George W. Bush delivers an ultimatum to Saddam Hussein to leave Iraq within 48 hours, March 17, 2003. Photo Credit: White House Archives

Threats Narrowly Construed: Intentionality Required

Intentional threats fall into two categories, conditional and unconditional. When George W. Bush declared that Osama bin Laden would “be found dead or alive,”16 he was uttering an unconditional threat. By contrast, when he told Saddam Hussein that he and his sons “must leave Iraq within 48 hours” and that “their refusal to do so will result in military conflict commenced at a time of our choosing,”17 the threat was made conditional upon satisfaction of the stated condition. An unconditional threat has the character of a promise; accordingly, qua speech act it merits classification as a commissive. But since its fulfillment does not require action on the part of the recipient, no directive is thereby implied. Conditional threats, on the other hand, are inseparably commissives and directives; it is in this conjunction that their special character is to be found. The ethics of conditional threats is the special concern of this article; only threats having this character will be taken up in what follows.

As to the intentionality required for a threat, it is necessary that the threatener P deliberately aim to make the target Q aware that consequence x will be imposed in the event of noncompliance. Instead, should Q find out through other means (say, by overhearing a conversation) that P will impose x as penalty for Q’s noncompliance, no threat will have been proffered.18 In addition to deliberateness, Schelling maintains that a second condition is included within the notion of a conditional threat, namely, that the threatener should manifestly prefer not to carry out the penalty.19 Otherwise, Q will easily believe that the threat is unconditional and consequently her compliance will have no decisive effect on the behavior of P.

It has been contended that this second condition is unduly restrictive, since “it requires us to know the threatener’s true intentions before we can determine whether or not a given statement is a threat.”20 Robert Nozick has similarly pointed out that “P can threaten Q with something even though P secretly hopes that Q will resist the threat.”21Q’s noncompliance provides P with a pretext to impose a penalty she had desired to bring about on independent grounds. Indeed, it is easy to construe a situation in which a party issuing a threat wants to implement the threatened sanction (for example, if a leader says, “we will annex this piece of land, which we have always coveted, unless you withdraw your forces”), and nonetheless finds itself legally or morally bound not to do so should the threatened party fulfill the demands placed on it.

It seems, however, that Schelling’s point was to establish what kind of intention must be communicated if an utterance is to function as a threat. Should the threatener merely feign a reluctance to impose the penalty, this would suffice, provided of course the threatener was able to communicate this reluctance credibly. Whether a threat could have proper moral standing under such a circumstance (threatening what one aims to do regardless of the target’s compliance) is a question that we will consider below, especially in the discussion of right intention.

Threats Coerce, but neither Compel nor Persuade

Threats are correlative to coercion. Coercion is the end for the sake of which threats are uttered. The success of a threat is measured by the degree to which it coerces the recipient into a desired line of behavior.22 Coercion is notoriously difficult to define and has accordingly been the subject of much philosophical debate. Most accounts establish a contrast with compulsion—the direct application of physical force to effect desired change in the target. Someone who has been handcuffed, transported in a locked vehicle, and imprisoned has clearly suffered compulsion. That person had no choice in the matter, nor are these actions in any meaningful sense her own.23 But when harm is presented as still in the future—a penalty that can be avoided under the condition that a determinate course of action is adopted—the influence in question is exercised by way of coercion rather than by sheer force. In other words, compulsion removes all free choice (liberum arbitrium) while coercion merely removes full liberty (libertas).

In the case of coercion, it is not only that one’s options are restricted (as can happen without there being a deliberate threat, as when someone is raised in a condition of poverty).24 Rather, someone coerced chooses the lesser of two evils; she intentionally, albeit unwillingly, performs the action in question. Such action is done at the bidding of another who deliberately restricts the range of options that would otherwise be available; in this respect coercion differs from persuasion. Both coercion and persuasion are ways of influencing someone’s behavior; both have the purpose of inducing certain intentional acts. But whereas the latter “attempts to provide a person with new reasons for actions, or to call attention to existing reasons . . . it does not do this in such a way as to exclude all choices but one.”25 In other words, persuasion leaves the recipient’s liberty intact, while coercion restricts it.

To what degree coercion restricts liberty depends on the nature of the coercion that is applied: If the alternative to giving in to a coercive threat is not altogether unthinkable to the party threatened, more liberty is retained than in a case where the threatened outcome is deemed wholly unacceptable. Being coerced by a threat of harm should not be conflated with acting under duress. Within Anglo-Saxon law especially, “duress” designates coercion that is wrongful. Coercion, by contrast, is an “‘ethically neutral’ concept.”26 As David Hoekema writes, “Coercion is not an inherently condemnatory notion: coercion can be justified.”27 That is, abstractly considered, coercion is not inherently bad; concretely, its goodness or badness will depend on the agent’s intention, the circumstances under which it is exercised, and so forth.

Despite this formal moral neutrality, common understanding attaches a negative connotation to coercion and its accompanying threats. To be coerced is akin to being sanctioned. If the sanction is deserved, then the coercion will be justifiable; otherwise it will constitute a wrong.28 In this sense, there is a moral presumption against coercion. If it should be apparent that an individual being threatened has little or no inclination for some wrongful behavior that the threat aims to stop, we would say that this person has been unjustifiably threatened. And someone who views herself as without propensity for committing the wrong in question will invariably feel that her honor has been impugned when she has thus been threatened.29

Distinguishing Deterrent from Compellent Threats

Thomas C. Schelling, Arms and Influence. Photo Credit: Yale University Press.

We have noted how threats are proffered to induce a change of behavior on the part of the one threatened. To threaten is thus to issue a directive. Here, two main possibilities stand out. On the one hand, threats can be used to coerce the recipient into carrying out an action she would otherwise not perform. Such threats are compellent, to borrow a neologism from Schelling.30 On the other hand, threats are sometimes employed to pressure the recipient to refrain from carrying out an act she might otherwise desire to perform. These preventive threats are deterrent. This is the “difference between a threat intended to make an adversary do something and a threat intended to keep him from starting something.”31

Schelling further explains how deterrent and compellent threats follow a contrasting dynamic. Fundamental is the question of “who has to make the first move.”32 When a deterrent threat is issued the initiative is put to the other side; the choice is forced upon the threatened party. As Schelling puts it, “I can block your car by placing mine in the way; my deterrent threat is passive, the decision to collide is up to you.”33 In contrast, with a compellent threat the order of initiative is inverted: here, the threatening party has to be committed to move first. Often (but not always) this requires an infliction of harm in progressively greater degrees until the other side understands how it can prevent a further escalation of pain only by complying with the demands put upon it.34

Schelling notes some other salient differences between compellent and deterrent threats. For instance, “deterrence tends to be indefinite in its timing,” while “compellence has to be definite.” Thus, to be effective, the compellent threat must specify “where, what, and how much” must be done to demonstrate compliance.35 By the same token, “the compellent action has to be one that can be stopped or reversed when the enemy complies, or else there is no inducement.”36 In either case, however, not only the threatened penalty but also the “proffered avoidance or reward” must be communicated so its credibility is apparent. This is especially true in compellence, where “the assurances are a very critical part of the . . . threat.”37

Verbal Threats and Action Threats

Finally, and as already indicated, unlike the case of deterrence, compellent threats must often pass beyond speech; some degree of implementation is ordinarily required. And whereas the pain threatened in deterrence must often be massive (so as to dissuade effectively), in compellence the pain inflicted upon the recalcitrant party must be moderate and calibrated so that further intensification is always possible: each application must be just enough to plausibly signal the possibility of still more. This pain is imposed not so much in view of retribution (although, to be morally justifiable, it must somehow be deserved), but rather to send a message that renewal is in store unless the demands are met. Here, the action that is thus carried out (the pain inflicted) has itself the character of a threat. We would call this an action threat in contrast to the verbal threats discussed above. It is also possible for a compellent threat to be communicated in the form of positive incentives, always with the underlying threat that the positive incentives will disappear if the desired actions are not carried out.38

Threats Convey Commitment

Threats, whether deterrent or compellent, presuppose some show of commitment on the part of the agent. Unless the recipient believes that the threatener will make good on the threats, she will not be motivated to action (if the threat is compellent) or to inaction (if the threat is deterrent). It is for this reason that threats, in addition to being characterized as directives, must also be placed in the category of commissives. In this respect, the threat lies within the same genus as a promise.39 “Both represent a surrender of choice,” in the sense that in each case the agent obligates herself to a particular course of action, and for this reason accepts that other courses of action are consequently closed.40 In other words, to threaten is to convey a commitment to action in the event that the threat should fail. Underlying the threat is consequently a conditional intention to carry out a specified line of action in the event of noncompliance.41

There can be many levels of commitment, ranging from the strongly worded ultimatum or “red line” (both of which communicate a strong commitment and essentially put the threatener’s reputation on the line) to what is often termed “studied ambiguity” (which leaves the threatener’s options more open). Sometimes a commitment may be merely feigned, as in the case of a bluff. Because the intent is to appear committed, bluffs merit inclusion within the category of threats.42 However, since the intent to punish noncompliance is indeed lacking, it is difficult to render a bluff credible, hence their effectiveness at inducing coercion is often in question. Bluffs work best in deterrent settings, when one does not physically have to display one’s commitment or willingness to carry out the sanction that is being threatened.


Having now given an overview of various sorts and levels of threats and coercion, and the sorts of speech acts they represent, let us now analyze their use from an ethical point of view.

Ethically speaking, there are many reasons to assume that threats, whether deterrent or compellent, should have little or no role in the practice of diplomacy. This assumption underlies the prohibition against threats in the UN Charter. Disputes between states (and a fortiori between states and nonstate groups) should ideally be settled by consensual agreement. Respect, openness, and trust create the best climate for reaching such agreements—a climate that can be seriously endangered when one party threatens the other with harm.43 Threats have historically been used by powerful states against their weaker counterparts. Coerced to negotiate “under the barrel of a cannon,” the latter have been subjected to a process akin to extortion. Thereby to exploit power asymmetry within a context of negotiation would seem blatantly inconsistent with the juridical equality of nations, a cornerstone of modern diplomacy. Furthermore, it is reasonable to assume that agreements reached on the basis of threats, if not outright invalid,44 are more likely to be breached at the first opportunity; compliance will be weak and will require renewed threats in order to remain effective. Moreover, countervailing the effectiveness of threats will be the humiliation and consequent resentment experienced by the coerced party, which will lead it to resist implementation of the detested agreement. In short, agreements that depend wholly or in large measure on coercion will arguably be far from sustainable in the long run.

Let us address this initial moral unease and enumerate briefly a set of moral requirements that seem intuitively reasonable if a threat is to be legitimate. Each of these requirements is meant to dispel to varying degrees the moral opprobrium that hangs over threats of armed force:

The target of the threat must knowingly have placed itself in a situation such that it is morally acceptable for it to be threatened, that is, the target has committed a wrong that could merit the response threatened, or has given indications that such a wrong was being planned. In the latter instance, the threatened military action would have a preventive character. The state of affairs to be rectified, avoided, or achieved through the use of threats is of a nature grave or important enough for the threat to be warranted. The party issuing the threat must have the necessary moral and political (and, where relevant, legal) standing to issue the threat, and must have legitimate reason to engage itself in the situation in question. The aim of the party issuing the threat—and the hoped-for outcome of the issuing of the threat—is a stable and peaceful state of affairs for all parties involved. The threat is formulated in such a way so as to make it understandable. The threat is clearly delimited, meaning the party being threatened will not be subject to ever-new threats concerning the same issue, if the demands of the threat are actually acceded to. The penalty for noncompliance would not, if it were imposed, involve an act prohibited by the laws of war or human rights conventions.

A look at the above list quickly shows us why an ethical analysis using just war criteria or categories is appropriate when considering the use of threats. We can indeed link this list of prima facie requirements to just war categories: the first and second deal with the category of just cause; the third is associated with legitimate authority; the fourth echoes the requirement of a rightful intention; the fifth harkens back to the traditional criterion of open declaration; while the sixth and seventh incorporate elements of proportionality, partly ad bellum (“proportionality of ends”) and partly in bello (“proportionality of means”). The seventh also incorporates jus in bello more generally, including the idea that means wrongful in themselves (mala in se) should never be employed.

All of these requirements seek to make threats of use of force legitimate means that contribute to avoiding armed conflict, and if armed conflict nonetheless occurs, to make armed conflict as limited as possible. Building on this, let us attempt to formulate the most important criteria for ethically legitimate threats, based on just war categories.

Just Cause

What might count as a justifiable reason for issuing a threat of armed force? The response will depend on one’s conception of just cause, the linchpin of contemporary just war theory. Rather than rehearse the various accounts here, it can safely be said that whenever there exists a just cause for the use of armed force, there likewise exists just cause to issue the corresponding threat.45 The actual employment of force causes more damage than the threat; hence, it is hard to imagine a case in which using force would be permissible while the threat to use such force would be ruled out. The possible exception would be a situation in which the initial resort to force requires an element of surprise, and thus on pragmatic grounds no prior threat could be given. Imagine, for instance, a case involving military intervention on humanitarian grounds. Atrocities are being carried out by a warlord, and a neighboring state decides that decisive military action will alone end the bloodletting. There is every reason to doubt that threatening intervention will effectively induce the warlord to desist from his armed action against civilians. In other words, a strategy of threats shows little reasonable hope of succeeding. Moreover, there are indications that should he become aware that an attack is imminent, his campaign of violence would only intensify. Supposing that a use of force to stop the atrocities would here be justifiable, it would seem, under the circumstances, that threatening attack beforehand would, on prudential grounds, be excluded. Such a threat would be impermissible even though the corresponding use of force would be allowed and indeed might be considered obligatory.46

In general, not only are threats permitted whenever the corresponding initiation of war would be permitted but, even more strongly, the issuance of a threat may be obligatory. This comports with the just war criterion of last resort (often termed “necessity” by lawyers), which holds that less damaging alternatives should be exhausted before waging war. The conditional threat to impose a consequence provides the addressee with an opportunity to rectify its wrongful behavior, and if successful this would obviate the need for a military solution—obviously a good thing (and always on the supposition that the wrong being responded to is of such nature that its continuance warrants a resort to war).

If, as noted above, a threat will ordinarily be justified when the corresponding employment of force is warranted, then it would seem that the normative assessment of the speech act is derivative upon the permissibility of the action that the speech act serves to signify. But is this always the case? From the point of view of just cause, could an act that would be impermissible to perform nonetheless be permissibly threatened? Can we ever rightly threaten what there can be no moral warrant for doing? The same question can also be asked from the point of view of proportionality: If issuing a severe threat, possibly to do something immoral, does less harm (and more good) than issuing a less severe threat or no threat at all, would the more severe threat then be morally defensible, even if carrying it out would not be? While the two questions of just cause and proportionality are closely related, we treat them here in order, beginning with just cause.

To exemplify how there might be just cause for issuing a threat even when carrying out that threat would be wrong, consider the case of economic harm. Imagine that one country, by its unfair trade practices, has caused grave damage to another. Under today’s understanding of just war and international law, economic harm does not of itself constitute a casus belli.47 But because the harm in question is nonetheless grave, could the victim-state be justified in threatening military attack upon the economic aggressor, precisely as a way to dissuade the latter from persisting in its injustice? Could such a threat have moral warrant? It is our position that it is hard to see how it would. Setting aside a mere bluff (which takes us into the distinct moral domain of truthfulness and lying), and assuming the threat is sincerely meant as a conditional utterance, it must be acknowledged that whosoever threatens such harm must be prepared to follow through in the event of noncompliance. A threat, as we have seen, is a kind of promise. But this fulfillment, under the present scenario, would be wrong. Consequently, to intend such a wrong, even conditionally, would itself be wrong. The gravity of this wrong would certainly be less, in all likelihood significantly so, than actually carrying out the wrongful deed.48 As such, we would judge this sort of wrong (conditionally intending unjustifiable harm) much less severely than the actual performance of the wrongful deed. The large gulf that separates the two wrongs explains why we are inclined to excuse threats that are made in circumstances wherein they enjoy partial but ultimately incomplete justification, as in the example of economic harm. Nonetheless, exculpation must not be confused with justification. Thus, it is difficult to see how threatening to do what would be wrong, even conditionally so, would not be wrong itself.49

Furthermore, upon the issuance of such a threat the threshold for using armed force would be lowered, probably on both sides of the conflict, and preparations would increase for such use, making conflict more likely. In other words, by forming a conditional intention to do an immoral act, albeit reluctantly and precisely to forestall this outcome, one voluntarily assumes (and increases) the risk of having to do it.50 David Gauthier put this well when he noted that “if it is rational to form [a] conditional, deterrent intention, then, should deterrence fail and the condition be realized, it is rational to act on it. The utility cost of acting on the deterrent intention enters, with appropriate probability weighting, into determining whether it is rational to form the intention.”51

A similar problem was on display when Trump, in the summer of 2017, threatened armed action (“fire and fury like the world has never seen”)52 against North Korea if its leader made any further nuclear threats. This could, at first glance, look like a straightforward scenario, where one threat was countered by another. Significant for our purposes, however, was the fact that Trump threatened armed retaliation (apparently by nuclear attack) not in the event that North Korea actually attacked the United States, but merely if the former renewed its verbal threat of such attack. This would, by the reasoning above, be morally indefensible. There can be no just cause for initiating massive military strikes in retaliation against the utterance of a threat. Consequently, to threaten such retaliation would itself be wrong, although obviously not to the same degree as the retaliation itself. In adverting to this example, we leave aside the possibility that Trump did not intend his threat to be taken at face value. Perhaps he meant solely to signal his resolve in the face of threats from North Korea, but not to commit himself to any particular line of action in the event of Kim’s noncompliance.

Ballistic missile submarines are crucial for nuclear deterrence--a type of implied threat. Photo credit: Wikimedia Commons.


The question of whether a threat of armed force may be morally acceptable even if an actual attack would not be can also be raised in relation to what just war theorists term “ad bellum proportionality.” On this criterion, even though a state may very well have just cause to mount an attack (say, because its sovereignty has been violated), it may be obliged to refrain from this resort to force because the resulting damage to life and property and other such goods would outweigh the gravity of the initial wrong. In essence, when the cure would be worse than the illness, resort to force may not be justified.

Intuitively, on a proportionality calculus, the harm caused by a threat is of a lesser magnitude than the direct violence resulting from its physical implementation. The moral standard to be met in issuing a threat is thus, it seems, less stringent than is the corresponding passage to overt military action. On this view, then, there may be cases where threatening military action may be justified even when the action itself may be ruled out on grounds of proportionality. This conclusion does, however, point to a notable difference between deterrence and compellence. As outlined above, a deterrent threat to use armed force implies that if the other party does x, then y will follow. If the other party does not do x, then y will not occur. Thus, if actually performing y is disproportionate to the situation at hand, the mere threat of doing y could still be proportionate if it is effective in hindering the performance of x. If, on the other hand, an action y is being threatened to stop an ongoing action x and it is likely that one will have to initiate some of y to actually stop x (as is the case with compellent threats), the proportionality calculus will differ accordingly. Insofar as compellence combines verbal threat and military action, it will likewise engage proportionality considerations that combine both. Compared to deterrence, a higher justificatory requirement will obtain.

If a difference in proportionality opens up between threats and actual implementation, how wide can this gap become? This was much discussed in the nuclear deterrence literature, where some scholars maintained that from a consequentialist standpoint the threat of attacking a civilian population center could be justified if it would likely deter a nuclear first strike.53 Others, however, maintained that whosoever issues such a threat, however conditionally, nonetheless accepts having to fulfill the threat in the event of noncompliance. As we have already argued, from a deontological standpoint, this acceptance cannot but constitute a wrong,54 albeit of lesser gravity than external commission of the corresponding deed.

True, in the scenario just considered, the threat could be uttered merely as a bluff. Then it would not be condemnable on grounds of wrongful intention, although condemnation on other grounds might be apposite (for example, as a form of lying).55 Bluffs, moreover, are notoriously difficult to pull off successfully: If it becomes clear to the other party that the threat will not be carried out, then its power to coerce disappears.

Combining elements of both—conditional intention on the one hand, and bluffs on the other—is what the literature on threats (in the nuclear context especially) has termed studied ambiguity.56 If ambiguity is cultivated with respect to a deterrent threat, where it is implied that the threat will not be carried out unless some clearly defined step is consciously and clearly taken by the party under threat—and it is possible for this party to avoid taking that step—the threatened penalty would remain merely hypothetical. If we add that the threat may not be carried out at all, and hence no trip wire is involved, merely the option that one might do it, we would be faced with a situation akin to what was much discussed in the literature on nuclear deterrence. (This is admittedly much closer to the current policy of nuclear deterrence than the traditional cold war posture, where the policy of “mutually assured destruction” more closely resembled that of a trip-wire mechanism. However, both sides during the cold war did build up significant conventional forces, which arguably also served the purpose of making a conventional response, even in the face of a nuclear attack, possible.)

Despite this contextual history, adopting a policy of studied ambiguity remains morally worrisome. This is in large part because, in order to maintain sufficient credibility, preparations have to be made for implementing the threat, and thereby one establishes the means and disposition for actually carrying out the wrongful deed. Once put into place, these means could then be utilized at some future point such that the initial threat, even if it was never meant to be carried out, could readily lead to a condition in which the threat might nonetheless be executed, if only by miscalculation of the other’s intention.

Not only in nuclear settings but also vis-à-vis conventional uses of force it is altogether possible that an attack should be deemed disproportionate based on ad bellum proportionality reasoning, even in the presence of a just cause. An example would be to argue that although the United States or NATO has a just cause to intervene with armed force in Syria, they should refrain, because to do so would cause more harm than good.57 In such a case, would a threat to use force be morally acceptable, if issuing the threat carried a great likelihood of successfully reducing tensions and harm to civilians? The case is a difficult one, because the problem does not here lie in threatening what would per se be immoral to do, but rather in a fault of prudential reasoning.

This is one of the most challenging moral questions that can be raised apropos of threats. We will not attempt to draw a definite conclusion, but merely emphasize how maintaining a stance of studied ambiguity vis-à-vis the use of clearly immoral (or even less clearly immoral, as in the Syria example) means can be hard to maintain without significantly increasing the likelihood that such means may be employed. This drives the strong moral presumption against relying on such threats and, if one seems nonetheless forced to employ them, to obey the moral imperative to remove oneself from this position as quickly as possible through, for instance, balanced disarmament or other tension-reducing measures.

Likelihood of Success

It is worth noting that the preceding discussion is also related to “likelihood of success,” often included as a separate just war category within jus ad bellum (although as a criterion or category it is closely related to proportionality and last resort).58 The question here would be whether the probability that a threat will successfully deter a behavior, thereby obviating the need to use force, will affect whether the threat can be ethically justified. Our discussion of studied ambiguity, as well as the discussion of when a deterrent threat may be justified if it has a high probability of working even though the associated action would not actually be justified, is directly relevant to this question. At the very least, it seems clear that if an ambiguous threat is unlikely to be successful, and its use increases the probability that military means will have to be used, doing so should be avoided. Such a course of action should be avoided especially when means of dubious ethical standing are brought into play. This also bears on the question of nuclear deterrence, but in an ambiguous way: Because the stakes are so high, a nuclear deterrent threat seems more likely to be successful, and this may add some weight to the permissibility of nuclear deterrence. However, precisely because the stakes are so high, threatening such action would be deeply problematic due to both the preparations necessary to make such a threat credible and the catastrophic results of having to act on it in the event of noncompliance. Either way, a greater likelihood of success adds to the permissibility of a deterrent threat, even if we hold that threatening what is immoral would be wrong regardless of the circumstances or likelihood of success.

Legitimate Authority

According to traditional just war teaching, war can be undertaken only by persons or institutions who hold legitimate (lawful and/or rightful) political authority. This criterion is meant to ensure that private wars are not engaged in, and that decision-making about resort to armed force is made in conformity with pre-established procedures. This teaching likewise implies the subordination of military to political authority, on the understanding that decisions associated with war engage the entire political community, and thus should be made only by those entrusted with oversight of the public good.

What can the just war criterion of legitimate authority tell us specifically about the issuing of threats? We suggest five major points:

First, as indicated above, the party issuing a threat must have the requisite moral and legal standing. Put differently, the authority underlying the threat should not be premised solely on physical or military power. The everyday parallel to, say, a school-yard quarrel is clear: the bully who threatens others merely because he is big enough and strong enough to do so is in a very different position from the headmaster who threatens sanctions against that bully unless he desists from his wrongdoing. It goes without saying that reasonable demands, even by legitimate authorities, are often thwarted because of power maneuvers, ideology, historical prejudice, or complex alliances. In spite of—or rather because of—these ever-present dangers, a general ambition to build threats on a “moral high ground” will undoubtedly lend legitimacy to the threats in question. This includes the procedural and moral legitimacy one gains from being recognized as a legitimate party to issue such threats.

Second, the party issuing the threat should be relevantly involved in the matter at hand. Threats that come across as mere posturing or as dubious interference from a distant third party with little knowledge or legitimate interest in the region or state in question will ipso facto stand on shaky ethical ground. Herein lies one of the core challenges faced by imperialist nations. The British in India, for instance, were never able to convincingly persuade the Indian public of the Crown’s rightful place in Indian society. Threats by imperial powers to use force in retaliation against protests by members of the native population often carried little moral strength exactly because the threatener was not taken to be a relevant party to the internal affairs of that society, but rather one to be either ignored or forced out. This clearly limited the strength and moral legitimacy of that party’s speech acts.

Third, the party issuing a threat should, whenever possible, be seen as but one member in a larger alliance. Granted, the existence of more parties is no guarantee of ethical legitimacy. The Axis states of the Second World War formed a highly credible and potent alliance, but did not thereby gain in moral legitimacy. When allies join one state’s threat, this may enhance the perceived legitimacy of the threat (subjective rightness), but it has no essential bearing on its actual legitimacy (objective rightness).59 On the other hand, a state’s inability to muster visible support for its cause, especially from parties whose support would otherwise be expected, will undermine the perceived legitimacy of the threats in question. For example, the ultimatums issued to Saddam Hussein by the United States in 2002 and 2003 failed to gather support from a broad alliance within the United Nations, and even within NATO, significantly decreasing the legitimacy of the subsequent invasion of Iraq. Coercive diplomacy suffers when there is dissension and lack of unity, as also witnessed by the lack of Russian and Iranian support to put pressure on the Assad regime in Syria, and the lack of Chinese support to put pressure on the Kim regime in North Korea. In both cases, the lack of a credible and effective alliance was (and is) arguably due not to the weakness of the (humanitarian or political) cause and the associated threats, but rather to power calculations not directly connected to the plight of the Syrian or North Korean people.

Fourth, it is an ethical as well as practical imperative that threats be credible, and thus also for the issuer of the threats to be credible. Parties that have a track record of bluffing will not be in a position to issue credible threats. Their conditional promise of sanctions will not be taken seriously even if they are so meant. On the other side of the spectrum, threat-makers who are known to move the goal posts—issuing new demands even as their original demands are met—will be unable to issue credible assurances, even if in this particular instance they are sincerely intended. In either case, threats will lose their potency as a tool in negotiations.

Finally, similarly important is the communication of demands in clear and consistent terms, even when a measure of studied ambiguity is involved. What is arguably the most dangerous enactment of coercive diplomacy in modern times, the Cuban Missile Crisis of 1962, became all the more dangerous when conflicting signals came from the Soviet Union at the exact point when a solution was in sight. Had it not been for President Kennedy’s prudent decision to respond only to the less threatening of the two conflicting communiqués received from Moscow, the prospect of a diffused crisis might have given way to destructive failure. Credibility and clarity are not only pragmatic demands; they also say much about the moral standing of the person or persons, that is, the authority issuing the threats. In just war teaching, this is standardly formulated under the heading of “open and transparent declaration.”

Right Intention: Concluding Thoughts

In linking just war categories and the issuance of threats of use of force, we have emphasized the ethical demands we can rightly place on speech acts communicating the possibility of using armed force, this being, in the end, the nature of the threats under consideration. We have also tried to give a taxonomy of what constitutes a threat, and how we can distinguish between different kinds of threats related to the use of force.

Our emphasis on just cause, clarity, credibility, and legitimacy, and our invocation of the time-honored idea of having the “moral high ground,” can easily mask the fact that the use of threats to avoid the unnecessary use of force will often also be marked by studied ambiguity. Will they, or won’t they? Are they bluffing? Should we take the chance? These are the basic quandaries that both parties to coercive diplomacy must face. This, in turn, leads to a crucial ethical question: Can bluffs and deception ever be acceptable if the ideal speech situation is one of transparency, clarity, and trust?

We believe this question can best be answered by reference to the just war criterion of right intention. Indeed, this also provides a fitting conclusion to our discussion. The ethical issuance of a conditional threat ideally implies a sincere desire not to have to carry out the threat. Threatening harm solely to create a pretext for using force against another party—for instance, by provoking the other party to attack first—falls squarely outside of the underlying right intention that should characterize any legitimate employment of threats. Furthermore, those who issue threats must carefully weigh the consequences of having to fulfill them. Willfully ignoring these consequences or neglecting to reflect carefully upon them evinces a recklessness that is inconsistent with the requirement of right intention. Although deterrent threats of disproportionate harm might be justifiable in extreme cases—if they will very likely succeed in averting the commission of a grave wrong—it is hard to see how there can be a sound moral basis for threatening indiscriminate harm.

When compellence is employed, the challenges to right intention will be especially acute. It can be very tempting to justify the imposition of pain and suffering—whether by economic or other sanctions short of armed attack or by the application of limited military strikes—on the pragmatic grounds that these are needed to make the threat credible. Yet this would be to put the cart before the horse. Such actions must be justified on their own terms before they can rightly be put to use within a strategy of diplomatic coercion. To invert this order would indeed reduce threats to blackmail or extortion, which are means manifestly inconsistent with the requirement of right intention.

Conditional threats of using armed force do constitute a morally problematic and ethically distinct category, for the many reasons we have adduced. Therefore, they should be carefully analyzed using ethical concepts and tools, such as just war criteria. Given the complexity surrounding threats, it is difficult to draw final conclusions about when they may be morally apposite, but we hope herein to have clarified which distinctions must be made and which core concerns must be addressed if we are to make the proper and necessary moral assessments.

A PDF version of this article is available to subscribers. Click here for access.

* Authors' Note: Preparation of this article was supported by a grant from the Norwegian Ministry of Foreign Affairs, as part of a project on “Threats of Force.” We thank the Ministry for this assistance. We are likewise grateful for valuable discussions on this topic with Nobuo Hayashi and Nicolas Stürchler, as well as for the very useful input from the editors of Ethics & International Affairs and their external reviewers.

  • 1 This dual prohibition is of course not unqualified, as two exceptions are allowed: self-defense against armed attack (Article 51) and Security Council enforcement action (Article 39).
  • 2 Threats “play the role of a ritualized substitute for the use of force, and as such may help to speed up the peaceful settlement of disputes.” Bruno Simma et al., eds., The Charter of the United Nations: A Commentary, 2nd
  • 3 The United States made use of gunboat diplomacy when Admiral Perry opened Tokyo to trade in 1854. A whole range of European countries did the same in China, forcing the Qing dynasty to sign a series of so-called “unequal treaties.”
  • 4 With of course some notable exceptions, as when the USSR issued ultimatums against Czechoslovakia (1947), Yugoslavia (1949–1951), and Israel and the United States (1973–1974); the U.K. against Israel (1948–1949) and Argentina (1982); the United States against North Korea (1949–1951); and Pakistan against India (1965–1966). For a comprehensive list of settings in which threats (of different kinds, including ultimatums) have been used from 1945 to 2003, see Nikolas Stürchler, The Threat of Force in International Law (New York: Cambridge University Press, 2007), pp. 291–310.
  • 5 For an overview, see H. G. de Jong, “Coercion in the Conclusion of Treaties: A Consideration of Articles 51 and 52 of the Convention on the Law of Treaties,” Netherlands Yearbook of International Law 15 (1984), pp. 209–47.
  • 6 The title of ch. 4 in Thomas Schelling, Arms and Influence, 2nd
  • 7 While far from abundant, the legal literature is considerably more extensive (see the bibliography in Stürchler, Threat of Force in International Law) than what may be found in the parallel discourse on just war ethics. An even more recent discussion may be found in Gro Nystuen, Stuart Casey-Maslen, and Annie G. Bersagel, eds., Nuclear Weapons Under International Law (Cambridge: Cambridge University Press, 2014); chs. 2 and 6 deal explicitly with the issue of threats.
  • 8 Jon Hovi, Games, Threats and Treaties (London: Pinter, 1998), pp. 11–13. The harms may of course be direct, by bringing harm to bear on the person targeted, or indirect, by hurting a third party for whom the target cares. For a technical-philosophical discussion of the logical conditions that are inherent to the idea of a successful (conditional) threat, see Martin Gunderson, “Threats and Coercion,” Canadian Journal of Philosophy 9, no. 2 (1979), pp. 247–59, particularly p. 257.
  • 9 The classification is outlined by John Searle in Expression and Meaning: Studies in the Theory of Speech Acts (Cambridge: Cambridge University Press, 1979). Here, we follow the summary given by Nicholas Fotion, “J. L. Austin, How to Do Things with Words (1962): An Active View of Language,” in Jorge J. E. Gracia, Gregory M. Reichberg, and Bernard Schumacher, eds., The Classics of Western Philosophy: A Reader’s Guide (Malden, Mass.: Blackwell, 2003), p. 578.
  • 10 The remaining two speech-act categories are expressives (which display the emotive state of speakers with respect to what they are talking about), and declarations (whereby a certain state of affairs is brought about simply by issuing the speech act, as when a boss tells her employee that he is fired).
  • 11 Gunderson, “Threats and Coercion,” p. 257.
  • 12 Thomas Schelling, The Strategy of Conflict (Cambridge, Mass.: Harvard University Press, 1980), p. 124, footnote 5. Schelling defines a “strategic move” as “one that influences the other person’s choice, in a manner favorable to oneself, by affecting the other person’s expectations on how one’s self will behave. One constrains the partner’s choice by constraining one’s own behavior” (p. 160). In this sense a threat is something “more than an assertion that is intended to appeal to the other player by power of suggestion” (p. 127).
  • 13 Schelling, Strategy of Conflict, p. 124. In the case of a promise, by contrast, the agent commits herself to a first move. For Schelling’s contrast between threats and promises, see ibid., pp. 177–78. There he explains how “a promise is costly when it succeeds, and a threat is costly when it fails.” Both, however, require the agent to communicate evidence of her commitment, if one or the other is to have the desired impact on the recipient (p. 147).
  • 14 See Thucydides, The Peloponnesian War, 5.84.1–5.111.
  • 15 Neil MacCormick, Institutions of Law (New York: Oxford University Press, 2007), p. 55. As one of our reviewers for this journal pointed out, a state passing laws within its own jurisdiction has a legal and moral authority to do so, whereas an international threatener has no such authority. This makes the former much less problematic than the latter.
  • 16 Wolf Blitzer, “Yet Another Report of bin Laden Escaping Afghanistan; Was American Military Commander Fired Upon?; Richard Reid in Federal Courtroom” (transcript), CNN Wolf Blitzer Reports, December 28, 2001,
  • 17 “Bush: ‘Leave Iraq within 48 Hours,’”, March 18, 2003,
  • 18 In other words, communication of P’s intent to Q is a necessary condition of a threat; see Gunderson, “Threats and Coercion,” p. 257.
  • 19 Schelling, Strategy of Conflict, p. 122: “The distinctive character of a threat is that one asserts that he will do, in a contingency, what he would manifestly prefer not to do if the contingency occurred, the contingency being governed by the second party’s behavior.”
  • 20 Hovi, Games, Threats and Treaties, p. 12.
  • 21 As summed up by Gunderson in “Threats and Coercion,” at pp. 256–57.
  • 22 As David A. Hoekema notes, “coercion” is a success term (Rights and Wrongs [Selinsgrove, Pa.: Susquehanna University Press, 1986]), p. 22.
  • 23 In Rights and Wrongs, Hoekema notes how disablement is “similar to compulsion except that it hinders rather than compels action . . . . Disablement takes away not the opportunity but the necessary means to do a certain range of acts” (p. 51).
  • 24 Grant Lamond thus distinguishes between “coercion” and “coercive situations.” Whereas the former presupposes a threatener who intends to coerce, the latter looks only to the effect, namely that by reason of some policy or condition, people end up restricted in their choices (see Grant Lamond, “Coercion,” in the International Encyclopedia of Ethics [Malden, Mass.: Wiley-Blackwell, 2013], pp. 840–47). Cheyney Ryan similarly distinguishes coercion from both constraint and compulsion (see Cheyney Ryan, “The Normative Concept of Coercion,” Mind 89, no. 356 [1980], pp. 481–98, p. 494). Conflation of these different senses of coercion lies behind Johan Galtung’s famous category of “structural violence,” and accounts for the many objections that have been formulated against it (see, for instance, C. A J. Coady, Morality and Political Violence [Cambridge: Cambridge University Press, 2008], pp. 25–30).
  • 25 Hoekema, Rights and Wrongs, pp. 51–52.
  • 26 Ryan, “The Normative Concept of Coercion,” p. 493.
  • 27 Hoekema, Rights and Wrongs, p. 25. The same author notes that as “a particular way of bringing a person to perform an action by means of a threat,” coercion “sometimes . . . is justified, sometimes not. It is an interference with a person that requires justification; but such justification is not impossible.” For this reason, although “coercion is not a morally evaluative category, it is nevertheless a moral concept. For instances of coercion cannot be identified without relying on moral concepts” (ibid.).
  • 28 This is why we do not usually classify laws as “threatening” or “coercive,” even if they imply sanctions should the law be broken. Now, whether state law is coercive by its very essence is a matter of some dispute. To view such law as coercive is a view supported by positivists in the line of Thomas Hobbes, but denied by natural lawyers in the tradition of Thomas Aquinas. However, even among those who adopt a minimalist line on the coerciveness of state law, it is uncontestable that this law has the character of a threat vis-à-vis the class of persons who are inclined to break this law. Vis-à-vis the upright, the same law merely persuades. In this vein, Thomas Aquinas wrote that “a man is said to be subject to a law as the coerced to the coercer . . . [hence] the virtuous and righteous are not subject to the law, but only the wicked” (Summa Theologiae, I–II, q. 96, art. 5).
  • 29 This reaction is visible, for instance, on the part of the Iranian religious leadership, which has emphatically denied it is pursuing a nuclear weapons capacity and has issued fatwas against this type of weaponry.
  • 30 See Arms and Influence, p. 71, where Schelling explains the rationale behind the choice of this term. He makes clear that “compellence” is one of two modalities within the broader category of “coercion,” with the other modality being “deterrence.” As kinds of coercion, both compellence and deterrence stand apart from what we have here termed “compulsion” (“brute force” in Schelling’s lexicon; see Arms and Influence, pp. 2–6). Thus, despite the fact that “compellence” and “compulsion” have a shared root in the Latin verb “compellere,” within this article we follow Schelling’s usage in defining “compellence” as a species of coercion (not of compulsion).
  • 31 Schelling, Arms and Influence, p. 69.
  • 32 Ibid.
  • 33 Ibid., p. 70.
  • 34 “Compellence, in contrast [to deterrence], usually involves initiating an action (or an irrevocable commitment to action) that can cease, or become harmless, only if the opponent responds. The overt act, the first step, is up to the side that makes the compellent threat. To deter, one digs in, or lays a minefield, and waits—in the interest of inaction. To compel, one gets up enough momentum (figuratively but sometimes literally) to make the other act to avoid collision” (Arms and Influence, p. 72). Thus understood, when (on March 17, 2003) President Bush ordered Saddam Hussein and his sons to leave Iraq within 48 hours or face the consequences, he was issuing a compellent threat. The “momentum” described by Schelling was manifested in the October 2002 “Iraq Resolution” in the U.S. Congress (itself preceded by UN Security Council Resolution 1441), which authorized use of force; this momentum became tangible when the United States started bombing Iraq on March 18, one day before the deadline expired. This bombing represented the president’s “first move.” The economic sanctions that were established against Iran also have the structure of a compellent threat. Iran was told that these sanctions would remain unless determinate steps were taken (allowance of more stringent international inspections, reduction of enrichment levels, etc.). For a recent treatment of compellent threats, see Todd S. Sechser and Matthew Fuhrmann, Nuclear Weapons and Coercive Diplomacy (Cambridge: Cambridge University Press, 2017).
  • 35 Schelling, Arms and Influence, p. 72.
  • 36 Ibid., p. 76.
  • 37 Ibid. See also Alexander L. George, Forceful Persuasion (Washington, D.C.: United States Institute of Peace, 1991), p. 5 for a useful discussion of Schelling’s term “compellence,” including the distinction between coercive diplomacy and blackmail, which George wants to put into sharper relief, leading him to jettison Schelling’s term.
  • 38 In this way threats have often been used to shift alliances in theaters such as Afghanistan, as for instance when armed groups (and their civilian supporters) were offered incentives to cease cooperation with the Taliban, but with the tacit threat that if compliance was not forthcoming armed attack would follow.
  • 39 Some authors define promises as referring to some benefit that will thereby accrue to the promisee (see Thomas Aquinas, Summa Theologiae, II–II, q. 88, art. 4); on such an understanding, promises and threats are contraries within the genus of commissives. Here, however, we (following Schelling) use the term “promise” more broadly to designate a verbal commitment; hence we speak of promises by reference both to benefits and harms.
  • 40 Schelling, Strategy of Conflict, p. 123. The “conditional self-commitment” (ibid., p. 134), which is characteristic of the threat does allow for various loosening strategies, see ibid., pp. 128–31.
  • 41 This conditionality is most clear in the case of deterrence, because here the first move is left entirely in the hands of the one threatened. The harm will not be meted out unless he acts first. In the case of compellence, often some positive action (a first move) must be undertaken by the one threatening, if the target is to get “the message.” But even here there is conditional intentionality, because the first installments of pain are imposed not for their own sake, but in order to signify the further steps that will follow should compliance not be met.
  • 42 Strictly speaking, as a speech act it is essential to the very notion of a threat that a conditional intention to inflict harm be communicated to the one threatened. For this reason, a mere bluff can count as a threat if this intent is indeed communicated. That said, what differentiates the sincere from a feigned threat is the presence or absence of this conditional intention on the part of the agent who threatens. The sincere threatener has formed this intention, while he who bluffs has not. For elaboration on the special characteristics of conditional versus direct intention, see John M. Finnis, Joseph M. Boyle, Jr., and Germain Grisez, Nuclear Deterrence, Morality and Realism (Oxford: Clarendon Press, 1987), pp. 81–86. As they rightly note, the presence or absence of a conditional intention also differentiates threats from warnings, the former requiring such an intention (at least as communicated) and the latter not. “One who merely predicts ‘If C, I will do A’ (e.g., ‘If the pain gets any worse, I’ll scream’) has a mind and heart different from one who has decided ‘If C, I will do A’ (e.g., ‘If I don’t get my way about this, I’ll scream’).” The latter (if he is not bluffing) “has made a decision; such a decision, fixing one’s will unless one reconsiders and decides otherwise, has the moral significance of an intention” (pp. 83–84).
  • 43 See Roger Fisher, Andrea K. Schneider, Elizabeth Borgwardt, and Brian Ganson, Coping with International Conflict (Upper Saddle River, N.J.: Pearson, 1996), pp. 194–208.
  • 44 Article 51 of the Vienna Convention on Treaties states that “the expression of a State’s consent to be bound by a treaty which has been procured by the coercion of its representative through acts or threats directed against him shall be without any legal effect.” This provision only covers threats that target the individuals who are engaged in interstate negotiations. Threats of broader compass, directed not at determinate individuals, but at one country against another, would not ipso facto have the effect of invalidating an agreement that was reached in this context. Significantly in this connection, Article 52 of the same treaty, which states that “a treaty is void if its conclusion has been procured by the threat or use of force,” adds the rider that this invalidation would obtain only when the threat is issued “in violation of the principles of international law embodied in the Charter of the United Nations.” Threats of force, as such, do not invalidate a resulting agreement; only unlawful threats, as defined by the Charter, would have this effect.
  • 45 For an overview of the various accounts, see Endre Begby, Gregory M. Reichberg, and Henrik Syse, “The Ethics of War. Part II: Contemporary Authors and Issues,” Philosophy Compass 7, no. 5 (2012), pp. 329–32.
  • 46 It is not easy to identify a historical incident that closely fits this scenario, although Tanzania’s 1978 attack on Ugandan forces under Idi Amin may provide an approximation. See Nicholas J. Wheeler, Saving Strangers: Humanitarian Intervention in International Society (Oxford: Oxford University Press, 2000), pp. 111–138.
  • 47 We are grateful to the journal’s editors and referees for highlighting this point.
  • 48 For the sake of clarity, it can be noted that between the conditional intention on the one hand, and the act carried out on the other, there lies the direct intention to perform the latter in the event of noncompliance with the demand that has thus been made. To each of these moments—the conditional intention, the subsequent direct intention, and fulfillment of intention in action—there corresponds a distinct moral assessment. On our analysis, if fulfilling a threat would be wrong, so too would be the antecedent intention, whether direct or conditional. But the three are not wrong in the same way; in other words, the gravity of each, as forms of wrongdoing, varies considerably.
  • 49 In other words, we agree with the assessment of Finnis, Boyle, and Grisez when they deny that the conditional intention can be judged on entirely different grounds than the corresponding direct intention. Discussing threats of nuclear attack, and conceding that “the [nuclear] deterrent strategy’s plan is [to have] the threat succeed so that it need not be carried out,” nonetheless, it would be “a mistake” to conclude “that the deterrent involves no self-determining intention to kill. It necessarily involves the threat to kill, and unless it is a sort of bluff, it expresses a choice, namely the choice to kill if the threat does not succeed” (Nuclear Deterrence, Morality and Realism, p. 83). Thus, if there can be no just cause for killing (say, by nuclear attack on a population center), so too can there be no just cause for the threat thereof.
  • 50 The question whether conditionally intending to do what is wrong, albeit for a good purpose, is itself wrong was at the heart of the 1980s debate on the ethics of nuclear deterrence. On this score, see the contrasting positions of Finnis, Boyle, and Grisez in Nuclear Deterrence, Morality and Realism, who deny that a conditional intention to do wrong can be justified by the greater harm that it prevents, versus Gregory S. Kavka in Moral Paradoxes of Nuclear Deterrence (Cambridge: Cambridge University Press, 1987), who maintained the contrary.
  • 51 David Gauthier, “Deterrence, Maximization, and Rationality,” Ethics 94, no. 3 (1984), pp. 474–95, at p. 486. Gauthier originally employed this argument to justify the rationality of nuclear deterrent threats, but he later pulled back from this position in “Assure and Threaten,” Ethics 104, no. 4 (1994), pp. 690–721.
  • 52 See Jim Sciutto, Barbara Starr, and Zachary Cohen, “Trump Promises North Korea ‘Fire and Fury’ over Nuke Threat,” CNN, August 9, 2017,
  • 53 This viewpoint was most prominently advocated by Kavka, who emphasized the gulf that exists between a direct and conditional intention to cause unjustified harm (for example, targeting a civilian population center with nuclear attack). Whereas the first intention must be excluded, the latter can be allowed on the following rationale: “Deterrent intentions, that is, those conditional intentions whose existence is based on the agent’s desire to thereby deter others from actualizing the antecedent condition of the intention . . . are, by nature, self-stultifying: if a deterrent intention fulfills the agent’s purpose, it ensures that the intended (and possibly) evil act is not performed, by preventing the circumstances of performance from arising. . . . Normally, an agent will form the intention to do something because she either desires doing that thing as an end in itself, or as a means to other ends. . . . But, in the case of deterrent intentions, the ground of the desire to form the intention is entirely distinct from any desire to carry it out” (Moral Paradoxes of Nuclear Deterrence, pp. 20–21).
  • 54 For further elucidation on this debate, see Gregory M. Reichberg, “The Morality of Nuclear Deterrence: A Reassessment,” in Mathias Nebel and Gregory M. Reichberg, eds., Nuclear Deterrence: An Ethical Perspective (Geneva: Caritas in Veritate Foundation, 2015),
  • 55 Bluffs of armed force have traditionally been treated in the just war literature under the heading of “stratagems” (Latin insidiae). In his seminal treatment of this theme (Summa Theologiae, II–II, q. 40, art. 3), Thomas Aquinas distinguished between outright lying and breaches of trust (both he deemed illicit) and dissimulation (which he judged as, in principle, licit).
  • 56 In Schelling’s formulation, “studied ambiguity” goes under the heading of “the threat that leaves something to chance” (Strategy of Conflict, pp. 187–204).
  • 57 We owe this example to one of the journal’s reviewers.
  • 58 We owe some of the formulations in this paragraph to one of the journal’s editors, who helpfully summarized two of our referees’ questions on this exact point.
  • 59 We are grateful to one of the journal’s reviewers for stressing this point.

More in this issue

Summer 2018 (32.2) Review Essay

The Irony of Just War

This review essay examines a series of benchmark books on the ethics of war published over the past year. All three grapple with the hard ...

Summer 2018 (32.2) Review

Gender, UN Peacebuilding, and the Politics of Space: Locating Legitimacy, by Laura J. Shepherd

Through rigorous and rich discourse analysis, Laura J. Shepherd interrogates not only how the UN understands peacebuilding itself but also how it understands gender, women, ...

Summer 2018 (32.2) Essay

Borders of Class: Migration and Citizenship in the Capitalist State

In this essay, Lea Ypi defends the relevance of social class in migration debates. In contrast to both defenders and critics of freedom of movement, ...