The assurances that
accompany
a compellent action- move back a mile and I won't shoot (other- wise I shall) and I won't then try again for a second mile- are
18.
18.
Schelling - The Art of Commitment
?
66
ritualistic commitments, of political and diplomatic commit- ments, of efforts to attach honor and reputation to a commit- ment, is to make the commitment manifestly hard to get out of on short notice. Even the commitments not deliberately in- curred, and the commitments that embarrass one in unforeseen circumstances, cannot be undone cheaply. The cost is the discrediting of other commitments that one would stilllike to be
credited.
If a country does want to get off the hook, to get out of a commitment deliberately incurred or one that grew up unin- tended, the opponent's cooperation can make a difference. The Chinese Communists seemed not to be trying, from 1958on, to make it easy for the United States to decouple itself from Quemoy. They maintained,and occasionally intensified,enough military pressure on the island to make graceful withdrawal difficult, tomake withdrawal looklike retreat under duress. It is hard to escape the judgment that they enjoyed American dis- comfort over Quemoy, their own ability to stir things up at will but to keep crises under their control, and their opportunity to aggravate American differences with Chiang Kai-shek.
Circumventing an Adversary's Commitments
"Salami tactics," we can be sure, were invented by a child; whoever first expounded the adult version had already under- stood the principle when he was small. Tell a child not to go in the water and he'll sit on the bank and submerge his bare feet;
15. The most eloquent rebuff I have come across is the answer the Romans received from the Volciani in Spain, whom they tried to unite with other Spanish cities against Carthage shortly after Rome had declined to defend the allied Spanish town of Saguntum against Hannibal and it had been terribly destroyed. "Men of Rome," said the eldest among them, "it seems hardly decent to ask us to prefer your friendship to that of Carthage, considering the precedent of those who have been rash enough to do so. Was not your betrayal of your friends in Saguntum even more brutal than their destruction by their enemies the Carthaginians? I suggest You look for allies in some spot where what happened to Saguntum has never been heard of. The fall of that town will be a signal and melancholy warning to the peoples of Spain never to count upon Roman friendship nor to trust Rome's word. " The War With Hunnibal, Aubrey de Selincourt, transl. (Baltimore, Penguin Books, 1965), p. 43.
he is not yet "in" the water. Acquiesce, and he'll stand up; no more of him is in the water than before. Think it over, and he'll start wading, not going any deeper; take a moment to decide whether this is different and he'll go a little deeper, arguing that since he goes back and forth it all averages out. Pretty soon we are calling to him not to swim out of sight, wondering whatever happened to all our discipline.
Most commitments are ultimately ambiguous indetail. Some- times they are purposely so, as when President Eisenhower and Secretary Dulles announced that an attack on Quemoy might or might not trigger an American response under the "Formosa Doctrine" according to whether or not it was interpreted as part of an assault, or prelude to an assault, on Formosa itself. Even more commitments are ambiguous because of the plain impos- sibility of defining them in exact detail. There are areas of doubt even in the most carefully drafted statutes and contracts; and even people who most jealously guard their rights and privi- leges have been known to settle out of court, to excuse an honest mistake, or to overlook a minor transgression because of the high cost of litigation. No matter how inviolate our commit- ment to some border, we are unlikely to start a war the first time a few drunken soldiers from the other side wander across the line and "invade" our territory. And there is always the possi- bility that some East German functionary on the Autobahn really did not get the word, or his vehicle really did break down in our lane of traffic. There is some threshold below which the commitment is just not operative, and even that threshold itself is usually unclear.
From this arises the low-level incident or probe, and tactics of erosion. One tests the seriousness of a commitment by probing it in a noncommittal way, pretending the trespass was inadvertent or unauthorized if one meets resistance, both to forestall the reaction and to avoid backing down. One stops a convoy or overflies a border, pretending the incident was accidental or unauthorized; but if there is no challenge, one continues or enlarges the operation, setting a precedent, estab- lishing rights of thoroughfare or squatters' rights, pushing the
l5
ARMS AND INFLUENCE
THE ART OF COMMITMENT 67
? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? 68 ARMS AND INFLUENCE
THE ART OF COMMITMENT 69
candles at night and get out voluntarily, than to start manhan- dling his family and his household goods. Blockade works slowly; it puts the decision up to the other side. To invade Berlin or Cuba is a sudden identifiable action, of an intensity that demands response; but to cut off supplies does little the first day and not much more the second; nobody dies or gets hurt from the initial effects of a blockade. A blockade is compara- tively passive; the eventual damage results as much from the obstinacy of the blockaded territory as from the persistence of the blockading power. And there is no well-defined moment before which the blockading power may quail, for fear of causing the ultimate collapse.
President Truman appreciated the value of this tactic in June 1945. French forces under de Gaulle's leadership had occupied a province in Northern Italy, contrary to Allied plans and
Americanpolicy. Theyannouncedthatanyeffortoftheirallies to dislodge them would be treated as a hostile act. The French intended to annex the area as a "minor frontier adjustment. " It would have been extraordinarily disruptive of Allied unity, of course, to expel the French by force of arms; arguments got nowhere, so President Truman notified de Gaulle that no more supplies would be issued to the French army until it had withdrawn from the Aosta Valley. The French were absolutely dependent on American supplies and the message brought results. This was "nonhostile" pressure, not quite capable of provoking a militant response, therefore safe to use (and effec- tive). A given amount of coercive pressure exercised over an extended period of time, allowed to accumulate its own mo- mentum, is a common and effective technique of bypassing somebody's commitment.
TheDistinctionBetweenDeterrenceand "Compellence"
Blockade illustrates the typical difference between a threat intended to make an adversary do something and a threat intended to keep him from starting something. The distinction is in the timing and in the initiative, in who has to make the first move, in whose initiative is put to the test. To deter an enemy's
commitment back or raising the threshold. The use of "volun- teers" by Soviet countries to intervene in trouble spots was usually an effort to sneak under the fence rather than climb over it, not quite invoking the commitment, but simultaneously making the commitment appear porous and infirm. And if there is no sharp qualitative division between a minor transgression and a major affront, but a continuous gradation of activity, one can begin his intrusion on a scale too small to provoke a reaction, and increase it by imperceptible degrees, never quite presenting a sudden, dramatic challenge that would invoke the committedresponse. Smallviolationsofatruceagreement,for example, become larger and larger, and the day never comes when the camel's back breaks under a single straw.
The Soviets played this game in Cuba for a long time, apparently unaware that the camel's back in that case could stand only a finite weight (or hoping the camel would get stronger and stronger as he got used to the weight). The Korean War may have begun as a low-level incident that was hoped to be beneath the American threshold of response, and the initial American responses (before the introduction of ground troops) may have been misjudged. Salami tactics do not always work. The uncertainty in a commitment often invites a low-level or noncommittal challenge; but uncertainty can work both ways. If the committed country has a reputation for sometimes, unpredictably, reacting where it need not, and not always collaborating to minimize embarrassment, loopholes may be less inviting. If one cannot get areputation for always honoring commitments in detail, because the details are ambiguous, it may help to get a reputation for being occasionally unreason- able. If one cannot buy clearly identifiable and fully reliable trip-wires, an occasional booby trap placed at random may serve somewhat the same purpose in the long run.
Landlords rarely evict tenants by strong-arm methods. They havelearnedthatsteadycumulativepressuresworkjustaswell, though more slowly, and avoid provoking a violent response. It is far better to turn off the water and the electricity, and let the tenant suffer the cumulative pressure of unflushed toilets and
? ? ? ? ? ? ? ? 70 ARMS AND INFLUENCE
THE ART OF COMMITMENT 71
might, for want of a better word, call compellence. The dictionary's definition of "deter" corresponds to contemporary usage: to turn aside or discourage through fear; hence, to prevent from action by fear of consequences. A difficulty with our being an unaggressive nation, one whose announced aim has usually been to contain rather than to roll back, is that we have not settled on any conventional terminology for the more active kind of threat. We have come to use "defense" as a euphemism for "military," and have a Defense Department, a defense budget, a defense program, and a defense establish- ment; if we need the other word, though, the English language provides it easily. It is "offense. " We have no such obvious counterpart to "deterrence. " "Coercion" covers the meaning but unfortunately includes "deterrent" as well as "compellent" intentions. "Intimidation" is insufficiently focused on the particular behavior desired. "Compulsion" is all right but its adjective is ''compulsive,''and that has come to carry quite a different meaning. "Compellence" is the best I can do. l7
Deterrence and compellence differ in a number of respects, most of them corresponding to something like the difference between statics and dynamics. Deterrence involves setting the stage- by announcement, by rigging the trip-wire, by incurring the obligation- and waiting. The overt act is up to the oppo- nent. The stage-setting can often be nonintrusive, nonhostile,
17. J. David Singer has used a nice pair of nouns, "persuasion" and "dissuasion," to make the same distinction. It is the adjectives that cause trouble; "persuasive" is bound to suggest the adequacy or credibility of a threat, not the character of its objective. Furthermore, "deterrent" is here to stay, at least in the English language. Singer's breakdown goes beyond these two words and is a useful one; he distinguishes whether the subject is desired to act or abstain, whether or not he is presently acting or abstaining, and whether he is likely (in the absence of threats and offers) to go on acting or abstaining. (If he is behaving, and is likely-
to go on behaving, there can still be reason to "reinforce" his motivation to behave. ) Singer distinguishes also "rewards" and "penalties" as well as threats and offers; while the rewards and "penalties" can be the consequences of threats and offers, they can also be gratuitous, helping to communicate persuasively some new and continuing threat or offer. See his article, "Inter-Nation Influence: A Formal Model," American Political Science Review, 17 (1963), 420-30.
advance it may be enough to burn the escape bridges behind me, or to rig a trip-wire between us that automatically blows us both up when he advances. T o compel an enemy's retreat, though, by some threat of engagement, I have to be committed to move. (This requires setting fire to the grass behind me as I face the enemy, with the wind blowing toward the enemy. ) I can block your car by placing mine in the way; my deterrent threat is passive, the decision to collide is up to you. But if you find me in your way and threaten to collide unless I move, you enjoy no such advantage; the decision to collide is still yours, and I still enjoy deterrence. You have to arrange to have to collide unless I move,andthatisadegreemorecomplicated. Youhavetogetup so much speed that you cannot stop in time and that only I can avert the collision; this may not be easy. If it takes more time to start a car than to stop one, you may be unable to give me the "last clear chance" to avoid collision by vacating the street.
The threat that compels rather than deters often requires that the punishment be administered until the other acts, rather than if he acts. This is because often the only way to become committed to an action is to initiate it. This means, though, that the action initiated has to be tolerable to the initiator, and tolerable over whatever period of time is required for the pressure to work on the other side. For deterrence, the trip-wire can threaten to blow things up out of all proportion to what is being protected, because if the threat works the thing never goes off. But to hold a large bomb and threaten to throw it unless somebody moves cannot work so well; the threat is not believ- able until the bomb is actually thrown and by then the damage is done. 16
There is, then, a difference between deterrence and what we
16. A nice illustration occurs in the movie version of A High Wind in Jamaica. The pirate captain, Chavez, wants his captive to tell where the money is hidden and puts his knife to the man's throat to make him talk. After a moment or two, during which the victim keeps his mouth shut, the mate laughs. "If you cut his throat he can't tell you. He knows it. And he knows you know it. " Chavez puts his knife away and tries something else.
-
? ? ? ? ? ? ? ? ? but not certain
? ? ? ? ? ? ? ? 72 ARMS AND INFLUENCE
nonprovocative. The act that is intrusive, hostile, or provoca- tive is usually the one to be deterred; the deterrent threat only changes the consequences ifthe act in question- the one to be deterred- is then taken. Compellence, in contrast, 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 thecompellentthreat. Todeter,onedigsin,orlaysaminefield, and waits- in the interest of inaction. To compel, one gets up enough momentum (figuratively, but sometimes literally) to maketheotheracttoavoidcollision.
Deterrence tends to be indefinite in its timing. "If you cross the line we shoot in self-defense, or the mines explode. " When? Whenever you cross the line- preferably never, but the timing is up to you. If you cross it, then is when the threat is fulfilled, eitherautomatically,ifwe'veriggeditso,orbyobligationthat immediately becomes due. But we can wait- preferably for- ever; that's our purpose.
Compellence has to be definite: We move, and you must get out of the way. By when? There has to be a deadline, otherwise tomorrow never comes. If the action carries no deadline it is only a posture, or a ceremony with no consequences. If the compellent advance is like Zeno's tortoise that takes infinitely longtoreachtheborderbytraversing,withinfinitepatience,the infinitely small remaining distances that separate him from collision, it creates no inducement to vacate the border. Compellence, to be effective, can't wait forever. Still, it has to wait a little; collision can't be instantaneous. The compellent threat has to be put in motion to be credible, and then the victim must yield. Too little time, and compliance becomes impos- sible; too much time, and compliance becomes unnecessary. Thus compellence involves timing in a way that deterrence typically does not.
In addition to the question of "when," compellence usually involves questions of where, what, and how much. "Do noth- ing" is simple, "Do something" ambiguous. "Stop where you are" is simple; "Go back" leads to "How far? " "Leave me
THE ART OF COMMITMENT 73
alone" is simple; "Cooperate" is inexact and open-ended. A deterrent position- a status quo, in territory or in more figura- tive terms- can often be surveyed and noted; a compellent advance has to be projected as to destination, and the destina- tion can be unclear in intent as well as in momentum and braking power. In a deterrent threat, the objective is often communicated by the very preparations that make the threat credible;thetrip-wireoftendemarcatestheforbiddenterritory. There is usually an inherent connection between what is threat- ened and what it is threatenedabout. Compellent threats tend to communicateonlythegeneraldirectionofcompliance,andare less likely to be self-limiting, less likely to communicate in the very design of the threat just what, or how much, is demanded. ThegarrisoninWestBerlincanhardlybemisunderstoodabout what it is committed to resist; if it ever intruded into East Berlin, though, to induce Soviet or German Democratic Republic forces to give way, there would be no such obvious interpreta- tion of where and how much to give way unless the adventure could be invested with some unmistakable goal or limitation-
a possibility not easily realized.
The Quemoy escapade is again a good example: Chiang's
troops, once on the island, especially if evacuation under fire appeared infeasible, had the static clarity that goes with com- mitmenttoanindefinitestatusquo,whilethecommitmentjust to send troops to defend it (or air and naval support) according to whether a Communist attack there was or was not prelude to anattackonFormosalackedthatpersuasivequality,reminding us that though deterrent threats tend to have the advantages mentioned above they do not always achieve them. (The ambiguous case of Quemoy actually displays the compellent ambiguity, seen in reverse: a "compellent" Communist move against Quemoy was to be accommodated, as long as its extent couldbereliablyprojectedtoaterminusshortofFormosa;ifthe Communiststhought we meant it, it was up to them to design an action that visibly embodied that limitation. ) An American or NATO action to relieve Budapest in 1956- without major engagement but in the hope the Soviets would give way rather
? ? ? ? ? ? ? 74 ARMS AND INFLUENCE
than fight- would have had the dynamic quality of "compellence" in contrast to Berlin: the stopping point would have been a variable, not a constant. Even "Budapest" would have needed a definition, and might have become all of Hun- gary- and after Hungary, what? - if the Soviets initially gave way. The enterprise might have been designed to embody its specificintent,butitwouldhavetakenalotofdesigningbacked up by verbal assurances.
Actually, any coercive threat requires corresponding assur- ances; the object of a threat is to give somebody a choice. To say, "One more step and I shoot," can be a deterrent threat only if accompanied by the implicit assurance, "And if you stop I won't. '' Giving notice of unconditional intent to shoot gives him no choice (unless by behaving as we wish him to behave the opponent puts himself out of range, in which case the effective threat is, "Come closer and my fire will kill you, stay back and it won't''). What was said above about deterrent threats being typically less ambiguous in intent can be restated: the corre- sponding assurances- the ones that, together with the threat- ened response, define the opponent's choice- are clearer than those that can usually be embodied in a compellent action. (Ordinary blackmailers, not just nuclear, find the "assurances" troublesome when their threats are compellent. ) l 8
They are, furthermore, confirmed and demonstrated over time; as long as he stays back, and we don't shoot, we fulfill the assurances and confirm them.
The assurances that accompany a compellent action- move back a mile and I won't shoot (other- wise I shall) and I won't then try again for a second mile- are
18. The critical role of "assurances" in completing the structure of a threat, in making the threatened consequences persuasively conditional on behavior so that the victim is offered a choice, shows up in the offers of amnesty, safe passage, or forgiveness that must often be made credible in inducing the surrender of rebels or the capitulation of strikers or protesters. Even libraries and internal revenue agencies depend on parallel offers of forgiveness when they embark on campaigns to coerce the return of books or payment of back taxes. In personal life I have sometimes relied, like King Lea, on the vague threat that my wrath will be aroused (with who knows what awful consequences) if good behavior is not forthcoming, making a tentative impression on one child, only to have the threat utterly nullified by another's pointing out that "Daddy's mad already. "
THE ART OF COMMITMENT 7s
harder to demonstrate in advance, unless it be through a long past record of abiding by one's own verbal assurances.
Because in the West we deal mainly in deterrence, not com- pellence, and deterrent threats tend to convey their assurances implicitly, we often forget that both sides of the choice, the threatened penalty and the proffered avoidance or reward, need to be credible. The need for assurances- not just verbal but fully credible- emerges clearly as part of "deterrence" in discussions of surprise attack and "preemptive war. " An enemy belief that we are about to attack anyway, not after he does but possibly before, merely raises his incentive to do what we wanted to deter and to do it even more quickly. When we do engage in compellence, as in the Cuban crisis or in punitive at- tacks on North Vietnam that are intended to make the North Vietnamese government act affirmatively, the assurances are a critical part of the definition of the compellent threat.
One may deliberately choose to be unclear and to keep the enemy guessing either to keep his defenses less prepared or to enhance his anxiety. But if one wants not to leave him in doubt about what will satisfy us, we have to find credible ways of com- municating, and communicating both what we want and what we do not want. There is a tendency to emphasize the commu- nication of what we shall do if he misbehaves and to give too little emphasis to communicating what behavior will satisfy us. Again, this is natural when deterrence is our business, because the prohibited misbehavior is often approximately defined in the threatened response; but when we must start something that then has to be stopped, as in compellent actions, it is both harder and more important to know our aims and to communi- cate. It is particularly hard because the mere initiation of an en- ergetic coercive campaign, designed for compellence, disturbs the situation, leads to surprises, and provides opportunity and temptation to reexamine our aims and change them in mid course. Deterrence,if wholly successful, can often afford to con- centrate on the initiating events- what happens next if he mis- behaves. Compellence, to be successful, involves an action that must be brought to successful closure. The payoff comes at the end, as does the disaster if the project fails.
? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? 76 ARMS AND INFLUENCE
THE ART OF COMMITMENT 77
? The compellent action will have a time schedule of its own, and unless it is carefully chosen it may not be reconcilable with the demands that are attached to it. W e cannot usefully threaten to bomb Cuba next Thursday unless the Russians are out by next month, or conduct a six weeks' bombing campaign in North Vietnam and stop it when the Vietcong have been quiescent for six months. There will be limits, probably, to how long the compellent action can be sustained without costing or risking too much, or exhausting itself or the opponent so that he has nothing left to lose. If it cannot induce compliance within that time- and this depends on whether compliance is physi- cally or administratively feasible within that time- it cannot accomplish anything (unless the objective was only an excuse for some act of conquest or punishment). The compellent ac- tion has to be one that can be stopped or reversed when the enemy complies, or else there is no inducement.
surrender statement or acknowledgement of submission, some symbolic knuckling under, will itself achieve the object, verbal compliance may be enough. It is inherent in an intense crisis that the conditions for bringing it to a close have to be of a kind that can be met quickly; that is what we mean by an "intense crisis," one that compresses risk, pain, or cost into a short span of time or that involves actions that cannot be sustained indefi- nitely. If we change our compellent threat from slow pressure to intense, we have to change our demands to make them fit the urgent timing of a crisis.
Notice that to deter continuance of something the opponent is already doing- harassment, overflight, blockade, occupation of some island or territory, electronic disturbance, subversive activity, holding prisoners, or whatever it may be- has some of the character of a compellent threat. This is especially so of the timing, of who has to take the initiative. In the more static case we want him to go on not doing something; in this more dy- namic case we want him to change his behavior. The "when" problem arises in compelling him to stop, and the compellent action may have to be initiated, not held in waiting like the de- terrent threat. The problems of "how much" may not arise if it is some discrete, well defined activity. "At all" may be the obvi- ous answer. For U-2flights or fishing within a twelve-mile limit, that may be the case; for subversive activity or support to in- surgents,"atall"mayitselfbeambiguousbecausetheactivityis complex, ill defined, and hard to observe or attribute.
Blockade, harassment, and "salami tactics" can be interpreted as ways of evading the dangers and difficulties of compel- lence. Blockade in a cold war sets up a tactical "status quo" that is damaging in the long run but momentarily safe for both sides unless the victim tries to run the blockade. President Ken- nedy's overt act of sending the fleet to sea, in "quarantine" of Cuba in October 1962, had some of the quality of deterrent "stage setting"; the Soviet government then had about forty- eight hours to instruct its steamers whether or not to seek collision. Low-levelintrusion,as discussed earlier, can be a way of letting the opponent turn his head and yield a little, or it can be a way of starting a compellent action in low gear, without the
? ? ? If the opponent's compliance necessarily takes time- if it is sustained good behavior, cessation of an activity that he must not resume, evacuation of a place he must not reenter, payment of tribute over an extended period, or someconstructiveactivity that takes time to accomplish- the compellent threat requires some commitment, pledge, or guarantee, or some hostage, or else must be susceptible of being resumed or repeated itself. Particularly in a crisis, a Cuban crisis or a Vietnamese crisis, there is strong incentive to get compliance quickly to limit the risk or damage. Just finding conditions that can be met on the demanding time schedule of a dangerous crisis is not easy. The ultimate demands, the objectives that the compellent threat is really aimed at, may have to be achieved indirectly, by taking pledges or hostages that can be used to coerce compliance after the pressure has been re1ie~ed. IO~f course, if some kind of
? ? ? ? ? 19. Lord Portal's account of the coercive bombing of the villages of recalcitrant Arab tribesmen (after warning to permit evacuation) includes the terms that were
? demanded. Among them were hostages- literal hostages, people -
as well as a fine; otherwise the demand was essentially cessation of the raids or other misbehavior that had brought on the bombing. The hostages were apparently partly to permit subsequent enforcement without repeated bombing, partly to symbolize, together with the fine, the tribe's intent to comply. See Portal, "Air Force Cooperation in
? ? Policing the Empire," pp. 343-58.
? ? ? ? 78 ARMS AND INFLUENCE
THE ART OF COMMITMENT 79
or costly,wecancallita"coercive"or"deterrent"defense. The language is clumsy but the distinction is valid. Resistance that might otherwise seem futile can be worthwhile if, though in- capable of blocking aggression, it can nevertheless threaten to make the cost too high. This is "active" or "dynamic" deter- rence, deterrence in which the threat is communicated by pro- gressive fulfillment. At the other extreme is forcible defense with good prospect of blocking the opponent but little promise of hurting; this would be purely defensive.
Defensive action may even be undertaken with no serious hope of repelling or deterring enemy action but with a view to making a "successful" conquest costly enough to deter repeti- tion by the same opponent or anyone else. This is of course the rationale for reprisals after the fact; they cannot undo the deed but can make the books show a net loss and reduce the incen- tive next time. Defense can sometimes get the same point across, as the Swiss demonstrated in the fifteenth century by the manner in which they lost battles as well as by the way they sometimes won them. "The [Swiss] Confederates were able to reckon their reputation for obstinate and invincible courage as one of the chief causes which gave them political importance. . . . It was no light matter to engage with an enemy who would not retire before any superiority in numbers, who was always ready for a fight, who would neither give nor take quarter. " 2o The Finns demonstrated five hundred years later that the princi- ple still works. The value of local resistance is not measured solely by local success. This idea of what we might call "punitive resistance" could have been part of the rationale for the American commitment of forces in Vietnam. *'
"Compellence" is more like "offense. " Forcible offense is taking something,occupyingaplace,ordisarminganenemyor a territory, by some direct action that the enemy is unable to block. "Compellence" is inducing his withdrawal, or his ac-
20. C. W. C. Oman, The Art of War in the Middle Ages (Ithaca, Cornell University Press, 1953), p. 96.
21. An alternative, but not inconsistent, treatment of some of these distinctions is in Glenn H. Snyder, Deterrence and Defense (Princeton, Princeton University Press, 1961), pp. 5-7, 9-16, 24-40.
conviction that goes with greater momentum but also without the greater risk. Instead of speeding out of control toward our car that blocks his way, risking our inability to see him and get our engines started in time to clear his path, he approaches slowly and nudges fenders, crushing a few lights and cracking some paint. If we yield he can keep it up, if not he can cut his losses. And if he makes it look accidental, or can blame it on an impetuous chauffeur, he may not even lose countenance in the unsuccessful try.
Defense and Deterrence, Offense and Compellence
The observation that deterrent threats are often passive, while compellent threats often have to be active, should not be pressed too far. Sometimes a deterrent threat cannot be made credible in advance, and the threat has to be made lively when the prohibited action is undertaken. This is where defense and deterrence may merge, forcible defense being undertaken in the hope, perhaps with the main purpose, of demonstrating by resistance that the conquest will be costly, even if successful, too costly to be worthwhile. The idea of "graduated deterrence" and much of the argument for a conventional warfare capability in Europe are based on the notion that if passive deterrence initially fails, the more active kind may yet work. If the enemy act to be deterred is a once-for-all action, incapable of with- drawal, rather than progressive over time, any failure of deter- rence is complete and final; there is no second chance. But if the aggressive move takes time, if the adversary did not believe he would meet resistance or did not appreciate how costly it would be, one can still hope to demonstrate that the threat is in force, after he begins. If he expected no opposition, encountering
somemaycausehimtochangehis mind.
There is still a distinction here between forcible defense and
defensive action intended to deter. If the object, and the only hope, is to resist successfully, so that the enemy cannot succeed even if he tries, we can call it pure defense. If the object is to induce him not to proceed, by making his encroachment painful
? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ARMS AND INFLUENCE
quiescence, or his collaboration by an action that threatens to hurt, often one that could not forcibly accomplish its aim but that, nevertheless, can hurt enough to induce compliance. The forcible and the coercive are both present in a campaign that could reach its goal against resistance, and would be worth the cost, but whose cost is nevertheless high enough so that one hopes to induce compliance, or to deter resistance, by making evident the intent to proceed. Forcible action, as mentioned in Chapter 1, is limited to what can be accomplished without enemy collaboration; compellent threats can try to induce more affirmative action, including the exercise of authority by an enemy to bring about the desired results.
War itself, then, can have deterrent or compellent intent, just as it can have defensive or offensive aims. A war in which both sides can hurt each other but neither can forcibly accomplish its purpose could be compellent on one side, deterrent on the other. Once an engagement starts, though, the difference between deterrence and compellence, like the difference between de- fense and offense, may disappear. There can be legal and moral reasons, as well as historical reasons, for recalling the status quo ante; but if territory is in dispute, the strategies for taking it, holding it, or recovering it may not much differ as between the sidethatoriginallypossesseditandthesidethatcovetedit,once the situation has become fluid. (In a local tactical sense, American forces were often on the "defensive" in North Korea and on the "offensive" in South Korea. ) The coercive aspect of warfare may be equally compellent on both sides, the only difference perhaps being that the demands of the defender, the one who originally possessed what is in dispute, may be clearly defined by the original boundaries, whereas the aggressor's demands may have no such obvious definition.
The Cuban crisis is a good illustration of the fluidity that sets in once passive deterrence has failed. The United States made verbal threats against the installation of weapons in Cuba but apparently some part of the threat was unclear or lacked credibility and it was transgressed. The threat lacked the autom- aticity that would make it fully credible, and without some
THE ART OF COMMITMENT 81
80
automaticity it may not be clear to either side just where the threshold is. Nor was it physically easy to begin moderate resistance after the Russians had crossed the line, and to increase the resistance progressively to show that the United States meant it. By the time the President determined to resist, he was no longer in a deterrent position and had to embark on the more complicated business of compellence. The Russian missiles could sit waiting, and so could Cuban defense forces; the next overt act was up to the President. The problem was to prove to the Russians that a potentially dangerous action was forthcoming, without any confidence that verbal threats would be persuasive and without any desire to initiate some irrevers- ible process just to prove, to everybody's grief, that the United States meant what it said.
The problem was to find some action that would communi- cate the threat, an action that would promise damage if the Russians did not comply but minimum damage if they com- plied quickly enough, and an action that involved enough momentum or commitment to put the next move clearly up to the Russians. Any overt act against a well-defended island would be abrupt and dramatic; various alternatives were appar- ently considered, and in the end an action was devised that had many of the virtues of static deterrence. A blockade was thrown around the island, a blockade that by itself could not make the missiles go away. The blockade did, however, threaten aminor military confrontation with major diplomatic stakes- an en- counter between American naval vessels and Soviet merchant ships bound for Cuba. Once in place, the Navy was in a position to wait; it was up to the Russians to decide whether to continue. If Soviet ships had been beyondrecall, the blockade would have been a preparation for inevitable engagement; with modem communications the ships were not beyond recall, and the Russians were given the last clear chance to turn aside. Physi- cally the Navy could have avoided an encounter; diplomati- cally, the declaration of quarantine and the dispatch of the Navy meant that American evasion of the encounter was virtually out of the question. For theRussians, the diplomatic cost of turning
? ? ? ? ? ? ? ? 82 ARMS AND INFLUENCE
THE ART OF COMMITMENT 83
threats of more violent action, had no effect. 23If the North Vietnamese dramatically issue a call to the Vietcong to cease activity and to evacuate South Vietnam, it is a conspicuous act of submission. If the Americans had evacuated Guantanamo when Castro turned off the water, it would have been a con- spicuous act of submission. If an earthquake or change in the weather had caused the water supply to dry up at Guantanamo, and if the Americans had found it wholly uneconomical to supply the base by tanker, they might have quit the place without seeming to submit to Castro's cleverness or seeming afraid to take reprisals against their ungracious host. Similarly, the mere act of bombing North Vietnam changed the status of any steps that the North Vietnamese might take to comply with American wishes. It can increase their desire, if the tactic is successful, to reduce support for the Vietcong; but it also increases the cost of doing so. Secretary Dulles used to say that while we had no vital interest in Quemoy we could not afford to evacuate under duress; intensified Chinese pressure always led to intensified determination to resist it. "
If the object is actually to impose humiliation, to force a showdown and to get an acknowledgement of submission, then the "challenge" that is often embodied in an active compellent threat is something to be exploited. President Kennedy un- doubtedly wanted some conspicuous compliance by the Soviet Union during the Cuban missile crisis, if only to make clear to the Russians themselves that there were risks in testing how much the American government would absorb such ventures. In Vietnam the problem appeared the opposite; what was most
freighters around, or even letting one be examined, proved not to be prohibitive.
Thus an initial deterrent threat failed, acompellent threat was called for, and by good fortune one could be found that had some of the static qualities of a deterrent threat. 22
There is another characteristic of compellent threats, arising in the need for affirmative action, that often distinguishes them from deterrent threats. It is that the very act of compliance- of doing what is demanded- is more conspicuously compliant, more recognizable as submission under duress, than when an act is merely withheld in the face of a deterrent threat. Compli- anceislikelytobelesscasual,lesscapableofbeingrationalized as something that one was going to do anyhow. The Chinese did not need to acknowledge that they shied away from Quemoy or Formosa because of American threats, and the Russians need not have agreed that it was NATO that deterred them from conquering Western Europe, and no one can be sure. Indeed, if a deterrent threat is created before the proscribed act is even contemplated, there need never be an explicit decision not to transgress, just an absence of any temptation to do the thing prohibited. The Chinese still say they will take Quemoy in their own good time; and the Russians go on saying that their intentions against Western Europe were never aggressive.
The Russians cannot, though, claim that they were on the point of removing their missiles from Cuba anyway, and that the President's television broadcast, the naval quarantine and
22. Arnold Horelick agrees with this description. "As an initial response the quarantine was considerably less than a direct application of violence, but considerably more than a mere protest or verbal threat. The U. S. Navy placed itself physically between Cuba and Soviet ships bound for Cuban ports. Technically, it might still have been necessary for the United States to fire the first shot had Khrushchev chosen to defy the quarantine, though other means of preventing Soviet penetration might have been employed. But once the quarantine was effectively established- which was done with great speed- it was Khrushchev who had to make the next key decision: whether or not to risk releasing the trip-wire. '' "The Cuban Missile Crisis," World Politics, 16 (1964), 385. This article and the Adelphi Paper of Albert and Roberta Wohlstetter mentioned in an earlier note are the best strategic evaluations of the Cuban affair that I have discovered.
23. The tendency for affirmative action to appear compliant is vividly illustrated
? ? by the widespread suspicion -
one that could not he effectively dispelled- that the U. S. missiles removed from Turkey in the wake of the Cuban crisis were
? part of a bargain, tacit if not explicit.
24. Almost everyone in America, surely including the President and the Secretary
of State, would have been relieved in the late 1950s if an earthquake or volcanic action had caused Quemoy to sink slowly beneath the surface of the sea.
ritualistic commitments, of political and diplomatic commit- ments, of efforts to attach honor and reputation to a commit- ment, is to make the commitment manifestly hard to get out of on short notice. Even the commitments not deliberately in- curred, and the commitments that embarrass one in unforeseen circumstances, cannot be undone cheaply. The cost is the discrediting of other commitments that one would stilllike to be
credited.
If a country does want to get off the hook, to get out of a commitment deliberately incurred or one that grew up unin- tended, the opponent's cooperation can make a difference. The Chinese Communists seemed not to be trying, from 1958on, to make it easy for the United States to decouple itself from Quemoy. They maintained,and occasionally intensified,enough military pressure on the island to make graceful withdrawal difficult, tomake withdrawal looklike retreat under duress. It is hard to escape the judgment that they enjoyed American dis- comfort over Quemoy, their own ability to stir things up at will but to keep crises under their control, and their opportunity to aggravate American differences with Chiang Kai-shek.
Circumventing an Adversary's Commitments
"Salami tactics," we can be sure, were invented by a child; whoever first expounded the adult version had already under- stood the principle when he was small. Tell a child not to go in the water and he'll sit on the bank and submerge his bare feet;
15. The most eloquent rebuff I have come across is the answer the Romans received from the Volciani in Spain, whom they tried to unite with other Spanish cities against Carthage shortly after Rome had declined to defend the allied Spanish town of Saguntum against Hannibal and it had been terribly destroyed. "Men of Rome," said the eldest among them, "it seems hardly decent to ask us to prefer your friendship to that of Carthage, considering the precedent of those who have been rash enough to do so. Was not your betrayal of your friends in Saguntum even more brutal than their destruction by their enemies the Carthaginians? I suggest You look for allies in some spot where what happened to Saguntum has never been heard of. The fall of that town will be a signal and melancholy warning to the peoples of Spain never to count upon Roman friendship nor to trust Rome's word. " The War With Hunnibal, Aubrey de Selincourt, transl. (Baltimore, Penguin Books, 1965), p. 43.
he is not yet "in" the water. Acquiesce, and he'll stand up; no more of him is in the water than before. Think it over, and he'll start wading, not going any deeper; take a moment to decide whether this is different and he'll go a little deeper, arguing that since he goes back and forth it all averages out. Pretty soon we are calling to him not to swim out of sight, wondering whatever happened to all our discipline.
Most commitments are ultimately ambiguous indetail. Some- times they are purposely so, as when President Eisenhower and Secretary Dulles announced that an attack on Quemoy might or might not trigger an American response under the "Formosa Doctrine" according to whether or not it was interpreted as part of an assault, or prelude to an assault, on Formosa itself. Even more commitments are ambiguous because of the plain impos- sibility of defining them in exact detail. There are areas of doubt even in the most carefully drafted statutes and contracts; and even people who most jealously guard their rights and privi- leges have been known to settle out of court, to excuse an honest mistake, or to overlook a minor transgression because of the high cost of litigation. No matter how inviolate our commit- ment to some border, we are unlikely to start a war the first time a few drunken soldiers from the other side wander across the line and "invade" our territory. And there is always the possi- bility that some East German functionary on the Autobahn really did not get the word, or his vehicle really did break down in our lane of traffic. There is some threshold below which the commitment is just not operative, and even that threshold itself is usually unclear.
From this arises the low-level incident or probe, and tactics of erosion. One tests the seriousness of a commitment by probing it in a noncommittal way, pretending the trespass was inadvertent or unauthorized if one meets resistance, both to forestall the reaction and to avoid backing down. One stops a convoy or overflies a border, pretending the incident was accidental or unauthorized; but if there is no challenge, one continues or enlarges the operation, setting a precedent, estab- lishing rights of thoroughfare or squatters' rights, pushing the
l5
ARMS AND INFLUENCE
THE ART OF COMMITMENT 67
? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? 68 ARMS AND INFLUENCE
THE ART OF COMMITMENT 69
candles at night and get out voluntarily, than to start manhan- dling his family and his household goods. Blockade works slowly; it puts the decision up to the other side. To invade Berlin or Cuba is a sudden identifiable action, of an intensity that demands response; but to cut off supplies does little the first day and not much more the second; nobody dies or gets hurt from the initial effects of a blockade. A blockade is compara- tively passive; the eventual damage results as much from the obstinacy of the blockaded territory as from the persistence of the blockading power. And there is no well-defined moment before which the blockading power may quail, for fear of causing the ultimate collapse.
President Truman appreciated the value of this tactic in June 1945. French forces under de Gaulle's leadership had occupied a province in Northern Italy, contrary to Allied plans and
Americanpolicy. Theyannouncedthatanyeffortoftheirallies to dislodge them would be treated as a hostile act. The French intended to annex the area as a "minor frontier adjustment. " It would have been extraordinarily disruptive of Allied unity, of course, to expel the French by force of arms; arguments got nowhere, so President Truman notified de Gaulle that no more supplies would be issued to the French army until it had withdrawn from the Aosta Valley. The French were absolutely dependent on American supplies and the message brought results. This was "nonhostile" pressure, not quite capable of provoking a militant response, therefore safe to use (and effec- tive). A given amount of coercive pressure exercised over an extended period of time, allowed to accumulate its own mo- mentum, is a common and effective technique of bypassing somebody's commitment.
TheDistinctionBetweenDeterrenceand "Compellence"
Blockade illustrates the typical difference between a threat intended to make an adversary do something and a threat intended to keep him from starting something. The distinction is in the timing and in the initiative, in who has to make the first move, in whose initiative is put to the test. To deter an enemy's
commitment back or raising the threshold. The use of "volun- teers" by Soviet countries to intervene in trouble spots was usually an effort to sneak under the fence rather than climb over it, not quite invoking the commitment, but simultaneously making the commitment appear porous and infirm. And if there is no sharp qualitative division between a minor transgression and a major affront, but a continuous gradation of activity, one can begin his intrusion on a scale too small to provoke a reaction, and increase it by imperceptible degrees, never quite presenting a sudden, dramatic challenge that would invoke the committedresponse. Smallviolationsofatruceagreement,for example, become larger and larger, and the day never comes when the camel's back breaks under a single straw.
The Soviets played this game in Cuba for a long time, apparently unaware that the camel's back in that case could stand only a finite weight (or hoping the camel would get stronger and stronger as he got used to the weight). The Korean War may have begun as a low-level incident that was hoped to be beneath the American threshold of response, and the initial American responses (before the introduction of ground troops) may have been misjudged. Salami tactics do not always work. The uncertainty in a commitment often invites a low-level or noncommittal challenge; but uncertainty can work both ways. If the committed country has a reputation for sometimes, unpredictably, reacting where it need not, and not always collaborating to minimize embarrassment, loopholes may be less inviting. If one cannot get areputation for always honoring commitments in detail, because the details are ambiguous, it may help to get a reputation for being occasionally unreason- able. If one cannot buy clearly identifiable and fully reliable trip-wires, an occasional booby trap placed at random may serve somewhat the same purpose in the long run.
Landlords rarely evict tenants by strong-arm methods. They havelearnedthatsteadycumulativepressuresworkjustaswell, though more slowly, and avoid provoking a violent response. It is far better to turn off the water and the electricity, and let the tenant suffer the cumulative pressure of unflushed toilets and
? ? ? ? ? ? ? ? 70 ARMS AND INFLUENCE
THE ART OF COMMITMENT 71
might, for want of a better word, call compellence. The dictionary's definition of "deter" corresponds to contemporary usage: to turn aside or discourage through fear; hence, to prevent from action by fear of consequences. A difficulty with our being an unaggressive nation, one whose announced aim has usually been to contain rather than to roll back, is that we have not settled on any conventional terminology for the more active kind of threat. We have come to use "defense" as a euphemism for "military," and have a Defense Department, a defense budget, a defense program, and a defense establish- ment; if we need the other word, though, the English language provides it easily. It is "offense. " We have no such obvious counterpart to "deterrence. " "Coercion" covers the meaning but unfortunately includes "deterrent" as well as "compellent" intentions. "Intimidation" is insufficiently focused on the particular behavior desired. "Compulsion" is all right but its adjective is ''compulsive,''and that has come to carry quite a different meaning. "Compellence" is the best I can do. l7
Deterrence and compellence differ in a number of respects, most of them corresponding to something like the difference between statics and dynamics. Deterrence involves setting the stage- by announcement, by rigging the trip-wire, by incurring the obligation- and waiting. The overt act is up to the oppo- nent. The stage-setting can often be nonintrusive, nonhostile,
17. J. David Singer has used a nice pair of nouns, "persuasion" and "dissuasion," to make the same distinction. It is the adjectives that cause trouble; "persuasive" is bound to suggest the adequacy or credibility of a threat, not the character of its objective. Furthermore, "deterrent" is here to stay, at least in the English language. Singer's breakdown goes beyond these two words and is a useful one; he distinguishes whether the subject is desired to act or abstain, whether or not he is presently acting or abstaining, and whether he is likely (in the absence of threats and offers) to go on acting or abstaining. (If he is behaving, and is likely-
to go on behaving, there can still be reason to "reinforce" his motivation to behave. ) Singer distinguishes also "rewards" and "penalties" as well as threats and offers; while the rewards and "penalties" can be the consequences of threats and offers, they can also be gratuitous, helping to communicate persuasively some new and continuing threat or offer. See his article, "Inter-Nation Influence: A Formal Model," American Political Science Review, 17 (1963), 420-30.
advance it may be enough to burn the escape bridges behind me, or to rig a trip-wire between us that automatically blows us both up when he advances. T o compel an enemy's retreat, though, by some threat of engagement, I have to be committed to move. (This requires setting fire to the grass behind me as I face the enemy, with the wind blowing toward the enemy. ) I can block your car by placing mine in the way; my deterrent threat is passive, the decision to collide is up to you. But if you find me in your way and threaten to collide unless I move, you enjoy no such advantage; the decision to collide is still yours, and I still enjoy deterrence. You have to arrange to have to collide unless I move,andthatisadegreemorecomplicated. Youhavetogetup so much speed that you cannot stop in time and that only I can avert the collision; this may not be easy. If it takes more time to start a car than to stop one, you may be unable to give me the "last clear chance" to avoid collision by vacating the street.
The threat that compels rather than deters often requires that the punishment be administered until the other acts, rather than if he acts. This is because often the only way to become committed to an action is to initiate it. This means, though, that the action initiated has to be tolerable to the initiator, and tolerable over whatever period of time is required for the pressure to work on the other side. For deterrence, the trip-wire can threaten to blow things up out of all proportion to what is being protected, because if the threat works the thing never goes off. But to hold a large bomb and threaten to throw it unless somebody moves cannot work so well; the threat is not believ- able until the bomb is actually thrown and by then the damage is done. 16
There is, then, a difference between deterrence and what we
16. A nice illustration occurs in the movie version of A High Wind in Jamaica. The pirate captain, Chavez, wants his captive to tell where the money is hidden and puts his knife to the man's throat to make him talk. After a moment or two, during which the victim keeps his mouth shut, the mate laughs. "If you cut his throat he can't tell you. He knows it. And he knows you know it. " Chavez puts his knife away and tries something else.
-
? ? ? ? ? ? ? ? ? but not certain
? ? ? ? ? ? ? ? 72 ARMS AND INFLUENCE
nonprovocative. The act that is intrusive, hostile, or provoca- tive is usually the one to be deterred; the deterrent threat only changes the consequences ifthe act in question- the one to be deterred- is then taken. Compellence, in contrast, 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 thecompellentthreat. Todeter,onedigsin,orlaysaminefield, and waits- in the interest of inaction. To compel, one gets up enough momentum (figuratively, but sometimes literally) to maketheotheracttoavoidcollision.
Deterrence tends to be indefinite in its timing. "If you cross the line we shoot in self-defense, or the mines explode. " When? Whenever you cross the line- preferably never, but the timing is up to you. If you cross it, then is when the threat is fulfilled, eitherautomatically,ifwe'veriggeditso,orbyobligationthat immediately becomes due. But we can wait- preferably for- ever; that's our purpose.
Compellence has to be definite: We move, and you must get out of the way. By when? There has to be a deadline, otherwise tomorrow never comes. If the action carries no deadline it is only a posture, or a ceremony with no consequences. If the compellent advance is like Zeno's tortoise that takes infinitely longtoreachtheborderbytraversing,withinfinitepatience,the infinitely small remaining distances that separate him from collision, it creates no inducement to vacate the border. Compellence, to be effective, can't wait forever. Still, it has to wait a little; collision can't be instantaneous. The compellent threat has to be put in motion to be credible, and then the victim must yield. Too little time, and compliance becomes impos- sible; too much time, and compliance becomes unnecessary. Thus compellence involves timing in a way that deterrence typically does not.
In addition to the question of "when," compellence usually involves questions of where, what, and how much. "Do noth- ing" is simple, "Do something" ambiguous. "Stop where you are" is simple; "Go back" leads to "How far? " "Leave me
THE ART OF COMMITMENT 73
alone" is simple; "Cooperate" is inexact and open-ended. A deterrent position- a status quo, in territory or in more figura- tive terms- can often be surveyed and noted; a compellent advance has to be projected as to destination, and the destina- tion can be unclear in intent as well as in momentum and braking power. In a deterrent threat, the objective is often communicated by the very preparations that make the threat credible;thetrip-wireoftendemarcatestheforbiddenterritory. There is usually an inherent connection between what is threat- ened and what it is threatenedabout. Compellent threats tend to communicateonlythegeneraldirectionofcompliance,andare less likely to be self-limiting, less likely to communicate in the very design of the threat just what, or how much, is demanded. ThegarrisoninWestBerlincanhardlybemisunderstoodabout what it is committed to resist; if it ever intruded into East Berlin, though, to induce Soviet or German Democratic Republic forces to give way, there would be no such obvious interpreta- tion of where and how much to give way unless the adventure could be invested with some unmistakable goal or limitation-
a possibility not easily realized.
The Quemoy escapade is again a good example: Chiang's
troops, once on the island, especially if evacuation under fire appeared infeasible, had the static clarity that goes with com- mitmenttoanindefinitestatusquo,whilethecommitmentjust to send troops to defend it (or air and naval support) according to whether a Communist attack there was or was not prelude to anattackonFormosalackedthatpersuasivequality,reminding us that though deterrent threats tend to have the advantages mentioned above they do not always achieve them. (The ambiguous case of Quemoy actually displays the compellent ambiguity, seen in reverse: a "compellent" Communist move against Quemoy was to be accommodated, as long as its extent couldbereliablyprojectedtoaterminusshortofFormosa;ifthe Communiststhought we meant it, it was up to them to design an action that visibly embodied that limitation. ) An American or NATO action to relieve Budapest in 1956- without major engagement but in the hope the Soviets would give way rather
? ? ? ? ? ? ? 74 ARMS AND INFLUENCE
than fight- would have had the dynamic quality of "compellence" in contrast to Berlin: the stopping point would have been a variable, not a constant. Even "Budapest" would have needed a definition, and might have become all of Hun- gary- and after Hungary, what? - if the Soviets initially gave way. The enterprise might have been designed to embody its specificintent,butitwouldhavetakenalotofdesigningbacked up by verbal assurances.
Actually, any coercive threat requires corresponding assur- ances; the object of a threat is to give somebody a choice. To say, "One more step and I shoot," can be a deterrent threat only if accompanied by the implicit assurance, "And if you stop I won't. '' Giving notice of unconditional intent to shoot gives him no choice (unless by behaving as we wish him to behave the opponent puts himself out of range, in which case the effective threat is, "Come closer and my fire will kill you, stay back and it won't''). What was said above about deterrent threats being typically less ambiguous in intent can be restated: the corre- sponding assurances- the ones that, together with the threat- ened response, define the opponent's choice- are clearer than those that can usually be embodied in a compellent action. (Ordinary blackmailers, not just nuclear, find the "assurances" troublesome when their threats are compellent. ) l 8
They are, furthermore, confirmed and demonstrated over time; as long as he stays back, and we don't shoot, we fulfill the assurances and confirm them.
The assurances that accompany a compellent action- move back a mile and I won't shoot (other- wise I shall) and I won't then try again for a second mile- are
18. The critical role of "assurances" in completing the structure of a threat, in making the threatened consequences persuasively conditional on behavior so that the victim is offered a choice, shows up in the offers of amnesty, safe passage, or forgiveness that must often be made credible in inducing the surrender of rebels or the capitulation of strikers or protesters. Even libraries and internal revenue agencies depend on parallel offers of forgiveness when they embark on campaigns to coerce the return of books or payment of back taxes. In personal life I have sometimes relied, like King Lea, on the vague threat that my wrath will be aroused (with who knows what awful consequences) if good behavior is not forthcoming, making a tentative impression on one child, only to have the threat utterly nullified by another's pointing out that "Daddy's mad already. "
THE ART OF COMMITMENT 7s
harder to demonstrate in advance, unless it be through a long past record of abiding by one's own verbal assurances.
Because in the West we deal mainly in deterrence, not com- pellence, and deterrent threats tend to convey their assurances implicitly, we often forget that both sides of the choice, the threatened penalty and the proffered avoidance or reward, need to be credible. The need for assurances- not just verbal but fully credible- emerges clearly as part of "deterrence" in discussions of surprise attack and "preemptive war. " An enemy belief that we are about to attack anyway, not after he does but possibly before, merely raises his incentive to do what we wanted to deter and to do it even more quickly. When we do engage in compellence, as in the Cuban crisis or in punitive at- tacks on North Vietnam that are intended to make the North Vietnamese government act affirmatively, the assurances are a critical part of the definition of the compellent threat.
One may deliberately choose to be unclear and to keep the enemy guessing either to keep his defenses less prepared or to enhance his anxiety. But if one wants not to leave him in doubt about what will satisfy us, we have to find credible ways of com- municating, and communicating both what we want and what we do not want. There is a tendency to emphasize the commu- nication of what we shall do if he misbehaves and to give too little emphasis to communicating what behavior will satisfy us. Again, this is natural when deterrence is our business, because the prohibited misbehavior is often approximately defined in the threatened response; but when we must start something that then has to be stopped, as in compellent actions, it is both harder and more important to know our aims and to communi- cate. It is particularly hard because the mere initiation of an en- ergetic coercive campaign, designed for compellence, disturbs the situation, leads to surprises, and provides opportunity and temptation to reexamine our aims and change them in mid course. Deterrence,if wholly successful, can often afford to con- centrate on the initiating events- what happens next if he mis- behaves. Compellence, to be successful, involves an action that must be brought to successful closure. The payoff comes at the end, as does the disaster if the project fails.
? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? 76 ARMS AND INFLUENCE
THE ART OF COMMITMENT 77
? The compellent action will have a time schedule of its own, and unless it is carefully chosen it may not be reconcilable with the demands that are attached to it. W e cannot usefully threaten to bomb Cuba next Thursday unless the Russians are out by next month, or conduct a six weeks' bombing campaign in North Vietnam and stop it when the Vietcong have been quiescent for six months. There will be limits, probably, to how long the compellent action can be sustained without costing or risking too much, or exhausting itself or the opponent so that he has nothing left to lose. If it cannot induce compliance within that time- and this depends on whether compliance is physi- cally or administratively feasible within that time- it cannot accomplish anything (unless the objective was only an excuse for some act of conquest or punishment). The compellent ac- tion has to be one that can be stopped or reversed when the enemy complies, or else there is no inducement.
surrender statement or acknowledgement of submission, some symbolic knuckling under, will itself achieve the object, verbal compliance may be enough. It is inherent in an intense crisis that the conditions for bringing it to a close have to be of a kind that can be met quickly; that is what we mean by an "intense crisis," one that compresses risk, pain, or cost into a short span of time or that involves actions that cannot be sustained indefi- nitely. If we change our compellent threat from slow pressure to intense, we have to change our demands to make them fit the urgent timing of a crisis.
Notice that to deter continuance of something the opponent is already doing- harassment, overflight, blockade, occupation of some island or territory, electronic disturbance, subversive activity, holding prisoners, or whatever it may be- has some of the character of a compellent threat. This is especially so of the timing, of who has to take the initiative. In the more static case we want him to go on not doing something; in this more dy- namic case we want him to change his behavior. The "when" problem arises in compelling him to stop, and the compellent action may have to be initiated, not held in waiting like the de- terrent threat. The problems of "how much" may not arise if it is some discrete, well defined activity. "At all" may be the obvi- ous answer. For U-2flights or fishing within a twelve-mile limit, that may be the case; for subversive activity or support to in- surgents,"atall"mayitselfbeambiguousbecausetheactivityis complex, ill defined, and hard to observe or attribute.
Blockade, harassment, and "salami tactics" can be interpreted as ways of evading the dangers and difficulties of compel- lence. Blockade in a cold war sets up a tactical "status quo" that is damaging in the long run but momentarily safe for both sides unless the victim tries to run the blockade. President Ken- nedy's overt act of sending the fleet to sea, in "quarantine" of Cuba in October 1962, had some of the quality of deterrent "stage setting"; the Soviet government then had about forty- eight hours to instruct its steamers whether or not to seek collision. Low-levelintrusion,as discussed earlier, can be a way of letting the opponent turn his head and yield a little, or it can be a way of starting a compellent action in low gear, without the
? ? ? If the opponent's compliance necessarily takes time- if it is sustained good behavior, cessation of an activity that he must not resume, evacuation of a place he must not reenter, payment of tribute over an extended period, or someconstructiveactivity that takes time to accomplish- the compellent threat requires some commitment, pledge, or guarantee, or some hostage, or else must be susceptible of being resumed or repeated itself. Particularly in a crisis, a Cuban crisis or a Vietnamese crisis, there is strong incentive to get compliance quickly to limit the risk or damage. Just finding conditions that can be met on the demanding time schedule of a dangerous crisis is not easy. The ultimate demands, the objectives that the compellent threat is really aimed at, may have to be achieved indirectly, by taking pledges or hostages that can be used to coerce compliance after the pressure has been re1ie~ed. IO~f course, if some kind of
? ? ? ? ? 19. Lord Portal's account of the coercive bombing of the villages of recalcitrant Arab tribesmen (after warning to permit evacuation) includes the terms that were
? demanded. Among them were hostages- literal hostages, people -
as well as a fine; otherwise the demand was essentially cessation of the raids or other misbehavior that had brought on the bombing. The hostages were apparently partly to permit subsequent enforcement without repeated bombing, partly to symbolize, together with the fine, the tribe's intent to comply. See Portal, "Air Force Cooperation in
? ? Policing the Empire," pp. 343-58.
? ? ? ? 78 ARMS AND INFLUENCE
THE ART OF COMMITMENT 79
or costly,wecancallita"coercive"or"deterrent"defense. The language is clumsy but the distinction is valid. Resistance that might otherwise seem futile can be worthwhile if, though in- capable of blocking aggression, it can nevertheless threaten to make the cost too high. This is "active" or "dynamic" deter- rence, deterrence in which the threat is communicated by pro- gressive fulfillment. At the other extreme is forcible defense with good prospect of blocking the opponent but little promise of hurting; this would be purely defensive.
Defensive action may even be undertaken with no serious hope of repelling or deterring enemy action but with a view to making a "successful" conquest costly enough to deter repeti- tion by the same opponent or anyone else. This is of course the rationale for reprisals after the fact; they cannot undo the deed but can make the books show a net loss and reduce the incen- tive next time. Defense can sometimes get the same point across, as the Swiss demonstrated in the fifteenth century by the manner in which they lost battles as well as by the way they sometimes won them. "The [Swiss] Confederates were able to reckon their reputation for obstinate and invincible courage as one of the chief causes which gave them political importance. . . . It was no light matter to engage with an enemy who would not retire before any superiority in numbers, who was always ready for a fight, who would neither give nor take quarter. " 2o The Finns demonstrated five hundred years later that the princi- ple still works. The value of local resistance is not measured solely by local success. This idea of what we might call "punitive resistance" could have been part of the rationale for the American commitment of forces in Vietnam. *'
"Compellence" is more like "offense. " Forcible offense is taking something,occupyingaplace,ordisarminganenemyor a territory, by some direct action that the enemy is unable to block. "Compellence" is inducing his withdrawal, or his ac-
20. C. W. C. Oman, The Art of War in the Middle Ages (Ithaca, Cornell University Press, 1953), p. 96.
21. An alternative, but not inconsistent, treatment of some of these distinctions is in Glenn H. Snyder, Deterrence and Defense (Princeton, Princeton University Press, 1961), pp. 5-7, 9-16, 24-40.
conviction that goes with greater momentum but also without the greater risk. Instead of speeding out of control toward our car that blocks his way, risking our inability to see him and get our engines started in time to clear his path, he approaches slowly and nudges fenders, crushing a few lights and cracking some paint. If we yield he can keep it up, if not he can cut his losses. And if he makes it look accidental, or can blame it on an impetuous chauffeur, he may not even lose countenance in the unsuccessful try.
Defense and Deterrence, Offense and Compellence
The observation that deterrent threats are often passive, while compellent threats often have to be active, should not be pressed too far. Sometimes a deterrent threat cannot be made credible in advance, and the threat has to be made lively when the prohibited action is undertaken. This is where defense and deterrence may merge, forcible defense being undertaken in the hope, perhaps with the main purpose, of demonstrating by resistance that the conquest will be costly, even if successful, too costly to be worthwhile. The idea of "graduated deterrence" and much of the argument for a conventional warfare capability in Europe are based on the notion that if passive deterrence initially fails, the more active kind may yet work. If the enemy act to be deterred is a once-for-all action, incapable of with- drawal, rather than progressive over time, any failure of deter- rence is complete and final; there is no second chance. But if the aggressive move takes time, if the adversary did not believe he would meet resistance or did not appreciate how costly it would be, one can still hope to demonstrate that the threat is in force, after he begins. If he expected no opposition, encountering
somemaycausehimtochangehis mind.
There is still a distinction here between forcible defense and
defensive action intended to deter. If the object, and the only hope, is to resist successfully, so that the enemy cannot succeed even if he tries, we can call it pure defense. If the object is to induce him not to proceed, by making his encroachment painful
? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ARMS AND INFLUENCE
quiescence, or his collaboration by an action that threatens to hurt, often one that could not forcibly accomplish its aim but that, nevertheless, can hurt enough to induce compliance. The forcible and the coercive are both present in a campaign that could reach its goal against resistance, and would be worth the cost, but whose cost is nevertheless high enough so that one hopes to induce compliance, or to deter resistance, by making evident the intent to proceed. Forcible action, as mentioned in Chapter 1, is limited to what can be accomplished without enemy collaboration; compellent threats can try to induce more affirmative action, including the exercise of authority by an enemy to bring about the desired results.
War itself, then, can have deterrent or compellent intent, just as it can have defensive or offensive aims. A war in which both sides can hurt each other but neither can forcibly accomplish its purpose could be compellent on one side, deterrent on the other. Once an engagement starts, though, the difference between deterrence and compellence, like the difference between de- fense and offense, may disappear. There can be legal and moral reasons, as well as historical reasons, for recalling the status quo ante; but if territory is in dispute, the strategies for taking it, holding it, or recovering it may not much differ as between the sidethatoriginallypossesseditandthesidethatcovetedit,once the situation has become fluid. (In a local tactical sense, American forces were often on the "defensive" in North Korea and on the "offensive" in South Korea. ) The coercive aspect of warfare may be equally compellent on both sides, the only difference perhaps being that the demands of the defender, the one who originally possessed what is in dispute, may be clearly defined by the original boundaries, whereas the aggressor's demands may have no such obvious definition.
The Cuban crisis is a good illustration of the fluidity that sets in once passive deterrence has failed. The United States made verbal threats against the installation of weapons in Cuba but apparently some part of the threat was unclear or lacked credibility and it was transgressed. The threat lacked the autom- aticity that would make it fully credible, and without some
THE ART OF COMMITMENT 81
80
automaticity it may not be clear to either side just where the threshold is. Nor was it physically easy to begin moderate resistance after the Russians had crossed the line, and to increase the resistance progressively to show that the United States meant it. By the time the President determined to resist, he was no longer in a deterrent position and had to embark on the more complicated business of compellence. The Russian missiles could sit waiting, and so could Cuban defense forces; the next overt act was up to the President. The problem was to prove to the Russians that a potentially dangerous action was forthcoming, without any confidence that verbal threats would be persuasive and without any desire to initiate some irrevers- ible process just to prove, to everybody's grief, that the United States meant what it said.
The problem was to find some action that would communi- cate the threat, an action that would promise damage if the Russians did not comply but minimum damage if they com- plied quickly enough, and an action that involved enough momentum or commitment to put the next move clearly up to the Russians. Any overt act against a well-defended island would be abrupt and dramatic; various alternatives were appar- ently considered, and in the end an action was devised that had many of the virtues of static deterrence. A blockade was thrown around the island, a blockade that by itself could not make the missiles go away. The blockade did, however, threaten aminor military confrontation with major diplomatic stakes- an en- counter between American naval vessels and Soviet merchant ships bound for Cuba. Once in place, the Navy was in a position to wait; it was up to the Russians to decide whether to continue. If Soviet ships had been beyondrecall, the blockade would have been a preparation for inevitable engagement; with modem communications the ships were not beyond recall, and the Russians were given the last clear chance to turn aside. Physi- cally the Navy could have avoided an encounter; diplomati- cally, the declaration of quarantine and the dispatch of the Navy meant that American evasion of the encounter was virtually out of the question. For theRussians, the diplomatic cost of turning
? ? ? ? ? ? ? ? 82 ARMS AND INFLUENCE
THE ART OF COMMITMENT 83
threats of more violent action, had no effect. 23If the North Vietnamese dramatically issue a call to the Vietcong to cease activity and to evacuate South Vietnam, it is a conspicuous act of submission. If the Americans had evacuated Guantanamo when Castro turned off the water, it would have been a con- spicuous act of submission. If an earthquake or change in the weather had caused the water supply to dry up at Guantanamo, and if the Americans had found it wholly uneconomical to supply the base by tanker, they might have quit the place without seeming to submit to Castro's cleverness or seeming afraid to take reprisals against their ungracious host. Similarly, the mere act of bombing North Vietnam changed the status of any steps that the North Vietnamese might take to comply with American wishes. It can increase their desire, if the tactic is successful, to reduce support for the Vietcong; but it also increases the cost of doing so. Secretary Dulles used to say that while we had no vital interest in Quemoy we could not afford to evacuate under duress; intensified Chinese pressure always led to intensified determination to resist it. "
If the object is actually to impose humiliation, to force a showdown and to get an acknowledgement of submission, then the "challenge" that is often embodied in an active compellent threat is something to be exploited. President Kennedy un- doubtedly wanted some conspicuous compliance by the Soviet Union during the Cuban missile crisis, if only to make clear to the Russians themselves that there were risks in testing how much the American government would absorb such ventures. In Vietnam the problem appeared the opposite; what was most
freighters around, or even letting one be examined, proved not to be prohibitive.
Thus an initial deterrent threat failed, acompellent threat was called for, and by good fortune one could be found that had some of the static qualities of a deterrent threat. 22
There is another characteristic of compellent threats, arising in the need for affirmative action, that often distinguishes them from deterrent threats. It is that the very act of compliance- of doing what is demanded- is more conspicuously compliant, more recognizable as submission under duress, than when an act is merely withheld in the face of a deterrent threat. Compli- anceislikelytobelesscasual,lesscapableofbeingrationalized as something that one was going to do anyhow. The Chinese did not need to acknowledge that they shied away from Quemoy or Formosa because of American threats, and the Russians need not have agreed that it was NATO that deterred them from conquering Western Europe, and no one can be sure. Indeed, if a deterrent threat is created before the proscribed act is even contemplated, there need never be an explicit decision not to transgress, just an absence of any temptation to do the thing prohibited. The Chinese still say they will take Quemoy in their own good time; and the Russians go on saying that their intentions against Western Europe were never aggressive.
The Russians cannot, though, claim that they were on the point of removing their missiles from Cuba anyway, and that the President's television broadcast, the naval quarantine and
22. Arnold Horelick agrees with this description. "As an initial response the quarantine was considerably less than a direct application of violence, but considerably more than a mere protest or verbal threat. The U. S. Navy placed itself physically between Cuba and Soviet ships bound for Cuban ports. Technically, it might still have been necessary for the United States to fire the first shot had Khrushchev chosen to defy the quarantine, though other means of preventing Soviet penetration might have been employed. But once the quarantine was effectively established- which was done with great speed- it was Khrushchev who had to make the next key decision: whether or not to risk releasing the trip-wire. '' "The Cuban Missile Crisis," World Politics, 16 (1964), 385. This article and the Adelphi Paper of Albert and Roberta Wohlstetter mentioned in an earlier note are the best strategic evaluations of the Cuban affair that I have discovered.
23. The tendency for affirmative action to appear compliant is vividly illustrated
? ? by the widespread suspicion -
one that could not he effectively dispelled- that the U. S. missiles removed from Turkey in the wake of the Cuban crisis were
? part of a bargain, tacit if not explicit.
24. Almost everyone in America, surely including the President and the Secretary
of State, would have been relieved in the late 1950s if an earthquake or volcanic action had caused Quemoy to sink slowly beneath the surface of the sea.
