This is not an
argument
that "our side" can always win a war
of nerves.
of nerves.
Schelling - The Manipulation of Risk
But this is a far cry from the notion that the two sides just measure up to each other and one bows before the other's superiority and acknowledges that he was only bluffing.
Any situation that scares one side will scare both sides with the danger of a war that neither wants, and both will have to pick their way carefully through the crisis, never quite sure that the other knows how to avoid stumbling over the brink.
Brinkmanship: The Manipulation of Risk If "brinkmanship" means anything, it means manipulating the
shaved risk ofwar. It means exploiting the danger that some- body may inadvertently go over the brink, dragging the other with him. If two climbers are tied together, and one wants to intimidate the other by seeming about to fall over the edge, there has to be some uncertainty or anticipated irrationality or it won't work. If the brink is clearly marked and provides a firm footing, no loose pebbles underfoot and no gusts of wind to catch one off guard, if each climber is in full control of himself and never gets dizzy, neither can pose any risk to the other by approaching the brink. There is no danger in approaching it; and while either can deliberately jump off, he cannot credibly pretend that he is about to. Any attempt to intimi- date or to deter the other climber depends on the threat of slipping or stumbling. With loose ground, gusty winds, and a propensity toward dizziness, there is some danger when a climber approaches the edge; one can credibly threaten to fall off accidentally by standing near the brink.
Without uncertainty, deterrent threats of war would take the form of trip-wires. To incur commitment is to lay a trip-wire, one that is plainly visible, that cannot be stumbled on, and that is manifestly connected up to the machinery of war. And if effective, it works much like a physical barrier. The tripwire will not be crossed as long as it has not been placed in an in- tolerable location, and it will not be placed in an intolerable lo- cation as long as there is no uncertainty about each other's
sians could have rationally denied at the cost of general war.
just might, in spite of all their care and all our care -
the brink and over it in a general war, had some substance. If we were anywhere near the brink of war on that occasion, it was a war that neither side wanted but that both sides might have been unable to forestall.
The idea, expressed by some writers, that such deterrence depends on a "credible first strike capability," and that a coun- try cannot plausibly threaten to engage in a general war over anything but a mortal assault on itself unless it has an appre- ciable capacity to blunt the other side's attack, seems to depend
on the clean-cut notion that war results
- lead up to
? Yet their implicit threat to behave in a way that might
that
? result -
or is expected to
-
only from adeliberate yes- no decision. But if war tends
to result from aprocess, a dynamic process in which both sides get more and more deeply involved, more and more expectant, more and more concerned not to be a slow second in case the war starts, it is not a "credible first strike" that one threatens, but just plain war. The Soviet Union can indeed threaten us with war: they can even threaten us with a war that we eventu- ally start, by threatening to get involved with us in a process that blows up into war. And some of the arguments about "superiority" and "inferiority" seem to imply that one of the two sides, being weaker, must absolutely fear war and concede while the other, being stronger, may confidently expect the other
to yield. There is undoubtedly a good deal to the notion that the country with the less impressive military capability may be less feared, and the other may run the riskier course in a crisis; other things being equal, one anticipates that the strategically
? ? ? 100 ARMS AND INFLUENCE
THE ART OF COMMITMENT
motives and nothing at issue that is worth a war to both sides. Either side can stick its neck out, confident that the other will not chop it off. As long as the process is a series of discrete steps, taken deliberately, without any uncertainty as to the con- sequences, this process of military commitment and maneuver
101
never end in disaster. It could only terminate in disaster if one of the players made a deIiberate move that he knew would cause disaster, and he would not. Second, the possibility of dis- aster will be reflected in the players' tactics. White can effec- tively keep Black's queen on her own side of the board by gettingaknightacrossfirst;orhecankeepbothBlack'sknights on their own side by getting his queen across first. This ability to block or to deter certain moves of the adversary will be an im- portant part of the game; the threat of disaster will be effective, so effective that the disaster never occurs.
In fact, the result is no different from a rule that says no queen can cross a center line if an opponent's knight has al- ready crossed it, and no knight can cross the center line if an opponent's queen has already crossed it. Prohibitive penalties imposed on deliberate actions are equivalent to ordinary rules.
The characteristic that this chess game shares with the trip- wire diplomacy, and that accounts for its peculiar safety, is the absence of uncertainty. There is always some moment, or some final step, in which one side or the other has the last clear chance to turn the course of events away from war (or from disaster in our game of chess) or to turn it away from a political situation that would induce the other to take the final step to- ward war. The skillful chess player will keep the knight across the center line or near enough to cross before his opponent's queen can get across, with due allowance for the cost of having to devote resources to the purpose. Skillful diplomacy, in the absence of uncertainty, consists in arranging things so that it is one's opponent who is embarrassed by having the "last clear
chance" to avert disaster by turning aside or abstaining from what he wanted to do.
But off the chess board the last chance to avert disaster is not always clear. One does not always know what moves of his own would lead to disaster, one cannot always perceive the moves that the other side has already taken or has set afoot, or what interpretation will be put on one's own actions; one does not al- ways understand clearly what situations the other side would not, at some moment, accept in preference to war. When we
? would not lead to war. Imminent war -
- continually threatened, but the threats would work. They would work unless one side were pushed too far; but if the push-
would be ing side knows how far that is, it will not push that far.
The resulting world -
- discriminate in favor of passivity against initiative. It is easier to deter than to compel. Among a group of arthritics moving deli- cately and slowly at a cocktail party, no one can be dislodged from his position near the bar, or ousted from his favorite chair; bodily contact is equally painful to his assailant. By standing in the doorway, one can prevent the entrance or exit of another
the world without uncertainty
would
ailing guest who is unwilling to push his way painfully through. In fact, without uncertainty all the military threats and ma- neuvers would be like diplomacy with rigid rules and can be illustrated with amodified game of chess. A chess game can end in win, lose, or draw. Let's change the game by adding a fourth outcome called "disaster. " If "disaster" occurs, a heavy fine is levied on both players, so that each is worse off than if he had simply lost the game. And the rules specify what causes dis- aster: specifically,if either player has moved his knight across the
center line and the other player has moved his queen across the center line, the game terminates at once and both players are scored with a disaster. If a white knight is already on the black side of the board when the black queen moves across to the white side, the black queen's move terminates the game in dis- aster; if the queen was already across when White moved his knight across the center line, the knight's move terminates the game in disaster for both players. And the same applies for the
white queen and the black knight.
What does this new rule do to the way a game is played? If a
game is played well, and both players play for the best score they can get, we can state two observations. First, a game will
possible war
? ? 102 ARMS AND INFLUENCE
THE ART OF COMMITMENT 103
in Wlxte's favor; he created the pressure, but both are subject to the same risk. White's advantage is that he can back out more quickly, as we have set up the game in this example; even he cannot retreat, though, until Black has made his next move, and for the moment both have the same incentive to come to terms. (White'sabilitytoretreat,andBlack'sinability,mayseemmore of an advantage to White than it actually is; his ability to retreat is an ability to save both players, equally, from disaster. If no bargain is reached, the white knight has to return, because he is the only one who can. If Black can avoid entering any nego- tiation- can absent himself from the room or turn off his hear- ingaid- White's soleremainingobjectivewillbetogethisown knight back before he blows things up. ) If "disaster" is only somewhat worse, not drastically worse, than losing the chess game, the side that is losing may have more incentive to threaten disaster, or more immunity to the other's threat, and perhaps in consequence a stronger bargaining position. Note, in particular, that all of this has nothing to do with whether a knight is more or less potent than a queen in the chess game; queen and knight can be interchanged in the analysis of this paragraph. If the clash of a squad with a division can lead to unintended war, or of a protest marcher with an armed police-
man to an unwanted riot, their potencies are equal in respect of the threats that count.
In this way uncertainty imports tactics of intimidation into the game. One can incur a moderate probability of disaster, sharing it with his adversary, as a deterrent or compellent de- vice, where one could not take, or persuasively threaten to take, a deliberate last clear step into certain disaster. 2
2. To clarify the theoretical point it may be worth observing that the uncertainty and unpredictability need not arise from a genuine random mechanism like the dice. It is unpredictability, not "chance," that makes the difference; it could as well arise in the clumsiness of the players, some uncertainty about the rules of the game or the scoring system, bad visibility or moves made in secret, the need to commit certain moves invisibly in advance, meddling by a third party, or errors made by the referee. Dice are merely a convenient way to introduce unpredictability into an artificial example.
? add uncertainty to this artificial chess game we are not so sure that disaster will be avoided. More important, the risk of disas- ter becomes a manipulative element in the situation. It can be exploited to intimidate.
To see this, make one more change in the rules. Let us not have disaster occur automatically when queen and knight of op- posite color have crossed the center line. Instead, when that occurs, the referee rolls a die. If an ace comes up the game is over and both players are scored with disaster, but if any other number appears the play goes on. If after the next move the queen and knight are still across the center line the dice are rolled again, and so on.
This is a very different game. And not just because disaster may or may not occur when queen and knight get into those positions, instead of occurring with certainty. The difference is that now queen and knight may actually be moved into those positions. One can deliberately move his knight across the line in an attempt to make the queen retreat, if one thinks his ad- versary is less willing to incur a continuing risk of disaster, or thinks his adversary can be persuaded that oneself will not re- treat, and if the momentary risk of disaster is not prohibitive. In fact, getting one's knight across and blocking its return with one's own pieces, so that it clearly takes several moves to re- treat, may persuade the adversary that only he, by withdrawing his queen, can reduce the risk within a tolerable time.
? If the black queen cannot retreat- if her exit is blocked against timely retreat- the white knight's tactic to force her withdrawal is ineffectual and gratuitously risky. But it can pos- sibly serve another end (another risky one), namely, to enforce "negotiation. " By crossing over,once the queen has crossed and cannot readily return, the knight can threaten disaster; White can propose Black's surrender, or a stalemate, or the removal of a bishop or the sacrifice of a pawn. What he gets out of this is wide open; but what began as a chess game has been converted into a bargaining game. Both sides are under similar pressure to settle the game or at least to get the white knight out of mis- chief. The outcome, it should be noticed, will not necessarily be
? ? ? ? ? ? ? 104 ARMS AND INFLUENCE
THE ART OF COMMITMENT 105
The route by which major war might actually be reached would have the same kind of unpredictability. Either side can
engaging in a limited war would usually be such a
gains to the other side. The white knight is as potent as the black queen in creating a shared risk of d i ~ a s t e r . ~
Limited War as a Generator of Risk
Limited war, as a deterrent to continued aggression or as a compellent means of intimidation, often seems to require inter- pretation along these lines, as an action that enhances the risk of a greater war. The danger of major war is almost certainly increased by the occurrence of a limited war; it is almost cer- tainly increased by any enlargement in the scope or violence of a limited war that has already taken place. This being so, the threat to engage in limited war has two parts. One is the threat to inflict costs directly on the other side, in casualties, expendi- tures, loss of territory, loss of face, or anything else. The second is the threat to expose the other party, together with oneself, to a heightened risk of a larger war.
just where the fault,
? take steps
step -
-
that genuinely raise the probability of a blow-up. This
would be the case with intrusions, blockades, occupations of third areas, border incidents, enlargement of some small war, or any incident that involves a challenge and entails a response that may in turn have to be risky. Many of these actions and threats designed to pressure and intimidate would be nothing but noise, if it were reliably known that the situation could not get out of hand. They would neither impose risk nor demon-
strate willingness to incur risk. And if they definitely would lead to major war, they would not be taken. (If war were desired, it would be started directly. ) What makes them significant and
usable is that they create a genuine risk
appreciated -
a danger that can be
-
that the thing will blow up for reasons not fully
under contr01. ~
It has often been said, and correctly, that a general nuclear
war would not liberate Berlin and that local military action in the neighborhood of Berlin could be overcome by Soviet mili- tary forces. But that is not all there is to say. What local mili- tary forces can do, even against very superior forces, is to initi- ate this uncertain process of escalation. One does not have to be able to win a local military engagement to make the threat of it effective. Being able to lose a local war in a dangerous and pro-
Just how the major war would occur -
? ? vocative manner may make the risk
quences, but the possibility of this act
not the sure conse- outweighthe apparent
3. The purest real-life example I can think of in international affairs is "buzzing" an airplane, as in the Berlin air corridor or when a reconnaissance plane intrudes. The only danger is that of an uninterzded collision. The pilot who buzzes obviously wants no collision. (If he did, he could proceed to do it straightforwardly. ) The danger is that he may not avoid accident, through mishandling his aircraft, or misjudging distance, or failure to anticipate the movements of his victim. He has to fly close enough, or recklessly enough, to create an appreciated risk that he may- probably won't, but nevertheless may- fail in his mission and actually collide, to everyone's chagrin including his own.
- -
is not predictable. Whatever it is that makes limited war between great powers a risky thing, the risk is a genuine one that neither side can alto- gether dispel even if it wants to. T o engage in limited war is to start rocking the boat, to set in motion a process that is not al-
4. It may be worth pointing out that, though all attempts to deter or to compel by threat of violence may carry some risk, it is not a necessary character of deterrent threats that they be risky if they are, or try to be, of the full-commitment or trip- wire variety discussed in the preceding chapter. What can make them risky is that they may not work as hoped: they are risky because they may fail. Ideally they would cany no risk. It is part of the logical structure of the threats discussed in this chapter that they entail risk- the risk of being fulfilled--even though they work (or were about to work) as intended. One is risky the way driving a car is always risky: genuine accidents can always occur, no matter how well the car is designed or how carefully it is driven; risk is a fact of life. The other is risky the way certain forms of road- hogging are risky: a genuine risk is incurred, or created, or enhanced, for the purpose of intimidation, a risk that may not be altogether avoided if intimidation is successfully achieved because it may have to operate for a finite period before compliance brings relief. This risk is part of the price of intimidation.
initiative, or misunderstanding may occur -
? ? ? ? ? ? ? ? ? 106 ARMS AND INFLUENCE
THE ART OF COMMITMENT 107
breached. One may then try not to maximize the stability of new limits as one passes certain thresholds, but to pass them in a way that dramatizes and emphasizes that the engagement is a dangerous one and that the other side should be eager to call a halt. Deliberately raising the risk of all-out war is thus a tactic that may fit the context of limited war, particularly for the side most discontent with the progress of the war. Introduction of nuclear weapons undoubtedly needs to be evaluated in these terms.
Discussions of troop requirements and weaponry for NATO have been much concerned with the battlefield consequences of different troop strengths and nuclear doctrines. But the battle- field criterion is only one criterion, and when nuclear weapons are introduced it is secondary. The idea that European arma- ment should be designed for resisting Soviet invasion, and is to be judged solely by its ability to contain an attack, is based on the notion that limited war is a tactical operation. It is not.
What that notion overlooks is that a main consequence of limited war, and potentially a main purpose for engaging in it, is to raise the risk of larger war. Limited war does this whether it is intended to or not.
together in one's control. (In the metaphorical language of our chess game, it is to move a queen or a knight across the center line when the other knight or queen is already across, establish- ing a situation in which factors outside the players' control can determine whether or not the thing blows up. ) The risk has to be recognized, because limited war probably does raise the risk of a larger war whether it is intended to or not. It is a conse- quence of limited war that that risk goes up; since it is a consequence, it can also be a purpose.
If we give this interpretation to limited war, we can give a corresponding interpretation to enlargements, or threats of en- largement, of the war. The threat to introduce new weapons, perhaps nuclear weapons, into a limited war is not, according to this argument, to be judged solely according to the immediate military or political advantage, but also according to the delib- erate risk of still larger war that it poses. And we are led in this way to a new interpretation of the trip-wire. The analogy for limited war forces in Europe, or a blockade about Cuba, or troops for the defense of Quemoy, according to this argument,
is not a trip-wire that certainly detonates all-out war if it is in working order and fails altogether if it is not. We have some- thing more like a minefield, with explosives hidden at random; a mine may or may not blow up if somebody starts to traverse the field. The critical feature of the analogy, it should be empha- sized, is that whether or not one of the mines goes off is at least to some extent outside the control of both parties to the engage- ment.
This argument is pertinent to the question not only of wheth- er, but of how, to cross the boundaries in some limited war. If one can gently erode a boundary, easing across it with-out creating some new challenge or a dramatic bid for enemy reprisal, and if one finds the current bounds intolerable, that may be the way to do it if one wants the tactical advantages of relaxing a rule. But if the tactical advantages are unimpressive, one's purpose in enlarging some limited war may be to con- front the enemy with a heightened risk, to bring into question the possibility of finding new limits once a few have been
This point is fundamental to deterrence of anything other than all-out attack on ourselves. And it is fundamental to the
- would be a real danger and would obsess the strategic commands on both sides. This danger is enhanced in a crisis, particularly one involving military activity. It is en- hancedpartlybecauseofthesheerpreoccupationwithit. Andit is enhanced because alarms and incidents will be more frequent, and those who interpret alarms will be readier to act on them. This is also, to a large extent, the purpose of being prepared to fight a local war in Western Europe. The Soviet anticipation of the risks involved in a large-scale attack must include the danger that general war will result. If they underestimate the scale and duration of resistance and do atrack, a purpose of re- sisting is to confront them, day after day, with an appreciation
strategy of limited war. The danger of sudden large war
of
unpremeditated war -
? ? ? ? 108 ARMS AND INFLUENCE
THE ART OF COMMITMENT 109
weapons signal and dramatize this very danger- a danger that is self-aggravating in that the more the danger is recognized, the more likely are the decisions that cause war to occur. This argu- ment is neither for nor against the use of nuclear weapons, but for recognizing that this consequence of their use equals in im- portance- and could far transcend- their tactical battlefield accomplishments.
It is worth noting that this interpretation suggests that the threat of limited war may be potent even when there is little ex- pectation that one could win it.
It is our sheer inability to predict the consequences of our actions and to keep things under control, and the enemy's sim- ilar inability, that can intimidate the enemy (and, of course, us too). If we were in complete control of the consequen-ces and
that life is risky, and that pursuit of the original objective is not worth the risk. - -
This is distantly but only distantly related to the notion that we deter an attack limited to Europe by the announced threat of all-out war. It is different because the danger of war does not depend solely on whether the United States would coolly resolve to launch general war in response to a limited attack in Europe. The credibility of a massive American re- sponse is often depreciated: even in the event of the threatened loss of Europe the United States would not, it is sometimes said, respond to the fait accompli of a Soviet attack on Europe with
? anything as "suicidal" as general war. But that is a simple- minded notion of what makes general war cr-edible. What can
make it exceedingly credib-
the Chinese in the Far East
war can occur whether we intend it or not.
a war we could make no threat that did not depend on our ultimate willingness to
le to the Russians
and perhaps to is that the triggering of general
knew what would and what would not precip-itate war
General war does not depend on our coolly deciding to retali- ate punitively for the invasion of Western Europe after careful consideration of the material and spiritual arguments pro and con. General war could result because we or the Soviets launched it in the mistaken belief that it was already on, or in the mistaken or correct belief that, if we did not start it in- stantly,theothersidewould,Itdoesnotdependonfortitude: it can result from anticipation of the worse consequences of a war that, because of tardiness, the enemy initiates.
And the fear of war that deters the Soviet Union from an at- tack on Europe includes the fear of a general war that they initi- ate. Even if they were confident that they could act first, they would still have to consider the wisdom of an action that might, through forces substantially outside their control, oblige them to start general war.
If nuclear weapons are introduced, the sensed danger of gen- eral war will rise strikingly. Both sides will be conscious of this increased danger. This is partly a matter of sheer expectation; everybody is going to be more tense, and for good reason, once nuclear weapons are introduced. And national leaders will know that they are close to general war if only because nuclear
choose general war.
This is not an argument that "our side" can always win a war
of nerves. (The same analysis applies to "their side" too. ) It is a reminder that between the alternatives of unsuccessful local resistance on the one extreme, and the fruitless, terrifying, and probablyunacceptableandincrediblethreatofgeneralthermo- nuclear war on the other, there is a strategy of risky behavior, of deliberately creating a risk that we share with the enemy, a risk that is credible precisely because its consequences are not entirely within our own and the Soviets' control.
Nuclear Weaponsand the Enhancement of Risk
The introduction of nuclear weapons raises two issues here. One is the actual danger of general war; the other is the role of this danger in our strategy. On the danger itself, one has to guess how likely it is that a sizable nuclear war in Europe can persist, and for how long, without triggering general war. The danger appears great enough to make it unrealistic to expect a tactical nuclear war to "run its course. " Either the nuclear weapons wholly change the bargaining environment, the appreciation of
that we started or a war that the enemy started
? ? ? ? 110 ARMS AND INFLUENCE
THE ART OF COMMITMENT I I I
national leaders as much as anything that is going on in Europe itself. It is the strategic forces whose minute-by-minute behavior on each side will be the main intelligence preoccupation of the other side. 5
Limited and localized nuclear war is not, therefore, a "tacti- cal" war. However few the nuclears used, and however selec- tively they are used, their purpose should not be "tactical" be- cause their consequences will not be tactical. With nuclears, it
as become more than ever a war of risks and threats at the ighest strategic level. It is a war of nuclear bargaining.
There are some inferences for NATO planning. First, nuclear weapons should not be evaluated mainly in terms of what they could do on the battlefield: the decision to introduce them, the way to use them, the targets to use them on, the scale on which to use them, the timing with which to use them, and the com- munications to accompany their use should not be determined (or not mainly determined) by how they affect the tactical course of the local war. Much more important is what they do to the expectation of general war, and what rules or patterns of expectations about local use are created. It is much more a war of dares and challenges, of nerve, of threats and brinkmanship, once the nuclear threshold is passed. This is because the danger of general war, and the awareness of that danger, is lifted an order of magnitude by the psychological and military conse- quences of nuclear explosion.
5. This is why one of the arguments for delegating nuclear authority to theater commanders- as presented in the election campaign of 1964- made little sense. That was the argument that communications between the theater and the American command structure might fail at the moment nuclear weapons were urgently needed. But if the weapons were that urgently needed, especially in the European theater, there would surely be appreciable danger of general war, and to proceed without communicating would guarantee the absence of crucial communication with the Strategic Air Command, the Defense Intelligence Agency, North American Air Defense Command, military forces everywhere, civil defense authorities, and, of course, our diplomatic establishment. It could preclude a choice of what kind of nuclear war to initiate; it could catch the Americans by surprise, and might merely give warning to the Russians.
? risks, and the immediate objectives, and bring about some termination, truce, tranquilization, withdrawal, or pause; or else the local war very likely becomes swamped in a much bigger war. If these are the likely alternatives, we should not take too seriously a nuclear local war plan that goes to great lengths to carry the thing to its bitter end. There is a high probability that the war either will go down by an order of magnitude or go up by an order of magnitude, rather than run the tactical nuclear course that was planned for it.
More important is how we control, utilize, and react to a sud- den increase in the sensed danger of general war. It will be so important to manage this risk properly that the battlefield con- sequences of nuclear weapons may be of minor importance. The hour-by-hour tactical course of the war may not even be worth the attention of the top strategic leadership.
One can question whether we ought to use nuclear weapons deliberately to raise the risk of general war. But unless we are willing to do this, we should not introduce nuclear weapons against an adversary who has nuclear weapons on his side. This raising of risk is so much of the consequence of nuclear weapons that to focus our planning attention on the battlefield may be to ignore what should be getting our main attention (and what would, in the event, get it). Once nuclear weapons are introduced, it is not the same war any longer. The tactical objectives and considerations that governed the original war are no longer controlling. It is now a war of nuclear bargaining and
demonstration.
In a nuclear exchange, even if it nominally involves only the
use of "tactical" weapons against tactically important targets, there will be a conscious negotiating process between two very threatening enemies who are worried that the war will get out of hand. The life expectancy of the local war may be so short that neither side is primarily concerned with what happens on the ground within the next day or two. What each side is doing with its strategic forces would be the main preoccupation. It is the strategic forces in the background that provide the risks and thesenseofdanger;itistheywhosedisposition willpreoccupy
? ? ? ? ? I12 ARMS AND INFLUENCE
Second, as a corollary we should not think that the value or likely success of NATO armed forces depends solely, or even mainly, on whether they can win a local war. Particularly if nu- clears are introduced, the war may never run its course. Even without the introduction of nuclears, a main function of re- sistance forces is to create and prolong a genuine sense of danger, of the potentiality of general war. This is not a danger that we create for the Russians and avoid ourselves; it is a danger we share with them. But it is this deterrent and intimida- tion function that deserves at least as much attention as the tactical military potentialities of the troops.
Third, forces that might seem to be quite "inadequate" by ordinary tactical standards can serve a purpose, particularly if they can threaten to keep the situation in turmoil for some pe- riod of time. The important thing is to preclude a quick, clean Soviet victory that quiets things down in short order.
Fourth, the deployment and equipment of nuclear-armed NATO troops, including the questions of which nationalities have nuclear weapons and which services have them, are affected by the purpose and function and character of nuclear and local war. If what is required is a skillful and well- controlled bargaining use of nuclears in the eventthe decision is taken to go above that threshold, and if the main purpose of nu- clears is not to help the troops on the battlefield, it is much less necessary to decentralize nuclear weapons and decisions to local commanders. The strategy will need tight centralized control; it may not require the kind of close battlefield support that is
often taken to justify distribution of small nuclears to the troops; and nuclears probably could be reserved to some special nuclear forces.
Fifth, if the main consequence of nuclear weapons, and the purpose of introducing them, is to create and signal a height- ened risk of general war, our plans should reflect that purpose.
THE ART OF COMMITMENT 113
veys to the Soviet leadership. Targets should be picked with a view to what the Soviet leadership perceives about the character of the war and about our intent, not for tactical importance. A target near or inside the U. S. S. R. , for example, is important be- cause it is near or inside the U. S. S. R. , not because of its tactical contribution to the European battlefield. A target in a city is important because a city is destroyed, not because it is a local supply or communication center. The difference between one weapon, a dozen, a hundred, or a thousand is not in the number of targets destroyed but in the Soviet (and American) percep- tion of risks, intent, precedent, and implied "proposal" for the conduct or termination of war.
Extra targets destroyed by additional weapons are not a local military "bonus. " They are noise that may drown the message. They are a "proposal" that must be responded to. And they are an added catalyst to general war. This is an argument for a se- lective and threatening use of nuclears rather than large-scale tactical use. (It is an argument for large-scale tactical use only if such use created the level of risk we wish to create. ) Success in the use of nuclears will be measured not by the targets de- stroyed but by how well we manage the level of risk. The So- viets must be persuaded that the war is getting out of hand but is not yet beyond the point of no return.
Sixth, we have to expect the Soviets to pursue their own policy of exploiting the risk of war. W e cannot expect the Sovi- ets to acquiesce in our unilateral nuclear demonstration. We have to be prepared to interpret and to respond to a Soviet nu- clear "counterproposal. " Finding a way to terminate will be as important as choosing how to initiate such an exchange. (We should not take wholly for granted that the initiation would be ours. )
Finally, the emphasis here is that the use of nuclear weapons would create exceptional danger. This is not an argument in favor of their use; it is an argument for recognizing that danger is the central feature of their use.
In other words, nuclears would not only destroy targets but would signal something. Getting the right signal across would be
? ? ? ? ? ? We should plan -
-
a war of nerve, of demonstration, and of bargaining, not just target destruction for local tactical purposes. Destroying a target may be incidental to the message that the detonation con-
in the event of resort to nuclear weapons
for
? ? ? 114 ARMS AND 1NFLUENCE
an important part of the policy. This could imply, for example, deliberate and restrained use earlier than might otherwise seem tactically warranted, in order to leave the Soviets under no illu- sion whether or not the engagement might become nuclear. The only question then would be, how nuclear. It is not necessarily prudent to wait until the last desperate moment in a losing en- gagement to introduce nuclear weapons as a last resort. By the time they are desperately needed to prevent a debacle, it may be too late to use them carefully, discriminatingly, with a view to the message that is communicated, and with the maintenance
of adequate control. Whenever the tactical situation indicates a high likelihood of military necessity for nuclears in the near fu- ture, it may be prudent to introduce them deliberately while there is still opportunity to do so with care, selection, and a properly associated diplomacy. Waiting beyond that point may simply increase the likelihood of a tactical use, possibly an in- discriminate use, certainly a decentralized use, determined by the tactical necessities of the battlefield rather than the strategic necessities of deterrence.
In its extreme form the restrained, signaling, intimidating use of nuclears for brinkmanship has sometimes been called the "shotacrossthebow. "Thereisalwaysadanger- Churchilland othershavewarnedagainstit-of makingabolddemonstration on so small a scale that the contrary of boldness is demon- strated. There is no cheap, safe way of using nuclears that scares the wits out of the Russians without scaring us too. Neverthe- less, any use of nuclears is going to change the pattern of expectations about the war. It is going to rip a tradition of
inhibition on their use. It is going to change everyone's expecta- tions about the future use of nuclears. Even those who have arguedthatnuclearsoughttobeconsideredjustamoreefficient kind of artillery will surely catch their breath when the first one goes off in anger. Something is destroyed, even if not enemy targets, if ever-so-few nuclears are used. Whatever a few nu- clears prove, or fail to prove about their user, they will change the environment of expectations. And it is expectations more
THE ART OF COMMlTMENT 115
than anything else that will determine the outcome of a limited East- West military engagement.
It is sometimes argued, quite correctly, that this tradition can be eroded, and the danger of "first use" reduced, by introducing nuclear weapons in some "safe" fashion, gradually getting the world used to nuclear weapons and dissipating the drama of nuclear explosions. Nuclear depth charges at sea, small nuclear warheads in air-to-air combat, or nuclear demolitions on de- fended soil may seem comparativelyfree of the danger of unlim- itedescalation,causenomorecivildisruptionthanTNT,appear responsible, and set new traditions for actual use, includ- ing the tradition that nuclear weapons can be used without sig- nalingall-outwar. Obviouslytoexploitthisideaoneshouldnot wait until nuclear weapons are desperately needed in a serious crisis, but deliberately initiate them in a carefully controlled fashion at a time and place chosen for the purpose. It might not be wise and might not be practical, but if the intent is to remove the curse from nuclear weapons, this may be the way to do it.
Among the several objections there is one that may be over- looked even by the proponents of nuclear "legitimization. " That is the waste involved- the waste of what is potentially the most dramatic military event since Pearl Harbor. President Johnson,remember,referredtoanineteen-yeartraditionofnon- use; the breaking of that tradition (which grows longer with each passing year) will probably be, especially if it is designed to be, a most stunning event. It will signal a watershed in mili- tary history, will instantly contradict war plans and military ex- pectations, wilf generate suspense and apprehension, and will probably startle even those who make the decision. The first post-Nagasaki detonation in combat will probably be evi- dence of a complex and anguished decision, an embarkation on ajourney into a new era of uncertainty. Even those who propose readier use of nuclear weapons must appreciate that this is so, because of the strong inhibitions they encounter during the dispute.
This is not an event to be squandered on an unworthy mili-
? ? ? ? ? ? ? ? ? ? 116 ARMS AND INFLUENCE
THE ART OF COMMITMENT I17
before they were old enough to drive and before automobiles were invented. The earliest instance I have come across, in a race with horse-drawn vehicles, antedates the auto by some time:
The road here led through a gully, and in one part the win- ter flood had broken down part of the road and made a hol-low. Menelaos was driving in the middle of the road, hoping that no one would try to pass too close to his wheel, but Antilochos turned his horses out of the track and followed him a little to one side. This frightened Menelaos, and he shouted at him:
"What reckless driving Antilochos! Hold in your horses. This place is narrow, soon you will have more room to pass. You will foul my car and destroy us both! "
But Antilochos only plied the whip and drove faster than ever, as if he did not hear. They raced about as far as the cast of quoit . . . and then [Menelaos] fell behind: he let the horses go slow himself, for he was afraid that they might all collide in that narrow space and overturn the cars and fall in a struggling heap.
This game of chicken tookplace outside the gates of Troy three thousand years ago. Antilochos won, though Homer says
? tary objective. The first nuclear detonation can convey a mes- sage of utmost seriousness; it may be a unique means of communicationinamomentofunusualgravity. Todegradethe signalinadvance,todepreciatethecurrency,toerodegradually a tradition that might someday be shattered with diplomatic effect, to vulgarize weapons that have acquired a transcendent status, and to demote nuclear weapons to the status of merely efficient artillery, may be to waste an enormous asset of last re- sort. One can probably not, with effect, throw down a gauntlet if he is known to toss his gloves about on every provocation. One may reasonably choose to vulgarize nuclear weapons through a campaign to get people used to them; but to proceed to use them out of expediency, just because they would be tactically advantageous and without regard to whether they
ought to be cheapened, would be shortsighted in the extreme.
Face,Nerve,and Expectations
? Cold war politics have been likened, by Bertrand Russell and others, to the game of "chicken. " This is described as a game in which two teen-age motorists head for each other on a highway - usuallylate at night, with their gangs and girlfriends looking
? -
who does is then called "chicken. "
on
to see which of the two will first swerve aside. The one
The better analogy is with the less frivolous contest of chicken that is played out regularly on streets and highways by people who want their share of the road, or more than their share, or who want to be first through an intersection or at least not kept waiting indefinitely.
"Chicken" is notjust a game played by delinquent teen-agers with their hot-rods in southern California; it is a universal form of adversary engagement. It is played not only in the Berlin air corridor but by Negroes who want to get their children into schools and by whites who want to keep them out; by rivals at a meeting who both raise their voices, each hoping the other will yield the floor to avoid embarrassment; as well as by drivers of both sexes and all ages at all times of day. Children played it
somewhatungenerously
"by trick, not by merit. "
-
-Even the game in its stylized teen-age automobile form is
worth examining. Most noteworthy is that the game virtually disappears if there is no uncertainty, no unpredictability. If the two cars, instead of driving continuously, took turns advancing exactly fifty feet at a time toward each other, a point would be reached when the next move would surely result in collision. Whichever driver has that final turn will not, and need not, drive deliberately into the other. This is no game of nerve. The lady who pushes her child's stroller across an intersection in front of a car that has already come to a dead stop is in no par- ticular danger as long as she sees the driver watching her: even
6 The Iliad, W H D Rouse, trand (Mentor Books, 19501, p 273
? ? ? 118 ARMS AND INFLUENCE
THE ART OF COMMITMENT I19
? ? if the driver prefers not to give her the right of way she has the winning tactic and gets no score on nerve. The more instruc- tive automobile form of the game is the one people play as they crowd each other on thehighway,jockey their way through an intersection, or speed up to signal to a pedestrian that he'd better not cross yet. These are the cases in which, like Antil- ochos' chariot, things may get out of control; no one can trust with certainty that someone will have the "last clear chance" to
Another important characteristic is that, though the two play- ers are cast as adversaries, the game is somewhat collaborative. Even in the stylized version in which they straddle the white line, there is at least an advantage in understanding that, when a player does swerve, he will swerve to the right and not to the left! And the players may try to signal each other to try to co- ordinate on a tie; if each can swerve a little, indicating that he will swerve a little more if the other does too, and if their speeds are not too great to allow some bargaining, they may manage to turn at approximately the same time, neither being proved chicken.
They may also collaborate in declining to play the game. This is a little harder. When two rivals are coaxed by their friends to have it out in a fight, they may manage to shrug it off skillfully,
? avert tragedy and will pull back in time.
These various games of chicken- the genuine ones that in-
volve some real unpredictability- have some characteristics that are worth noting. One is that, unlike those sociable games it takes two to play, with chicken it takes two not to play. If you are publicly invited to play chicken and say you would rather not, you have just played.
at stake but reputations, expectations, and precedents. That is, accommodation or obstinacy, boldness or surrender, merely establishes who is an accommodator, who is obstinate or bold, who tends to surrender or what order of precedence is to be observed. A second, not easily distinguished in practice, occurs when something is consciously put at stake (as in a gambling game or trial by ordeal) such as leadership, deference, popularity, some agreed tangible prize, or the outcome of certain issues in dispute. (The duel between David and Goliath, mentioned in the note on page 144, is an example of putting something at stake. ) The third, which might be called the "real" in contrast to the "conventional," is the case in which yielding or withdrawing yields something that the dispute is about, as in road- hogging or military probes: that is, the gains and losses are part of the immediate structure of the contest, not attached by convention nor resulting entirely from expectations established for future events. The process of putting something at
? Second, what is in dispute is usually not the issue of the mo- ment, but everyone's expectations about how a participant will behave in the future. To yield may be to signal that one can be expected to yield; to yield often or continually indicates acknowl- edgment that that is one's role. To yield repeatedly up to some limit and then to say "enough" may guarantee that the first show of obduracy loses the game for both sides. If you can get a rep- utationforbeingreckless,demanding,orunreliable- and appar- ently hot-rods, taxis, and cars with "driving school" license plates sometimes enjoy this advantage- you may find conces-
sions made to you. (The driver of a wide American car on a narrow European street is at less of a disadvantage than a static calculation would indicate. The smaller cars squeeze over to give him room. ) Between these extremes, one can get a reputa- tion for being firm in demanding an appropriate share of the road but not aggressively challenging about the other's half. Un- fortunately, in less stylized games than the highway version, it is often hard to know just where the central or fair or expected division should lie, or even whether there should be any recog- nition of one contestant's claim. 7
7. Analytically there appear to be at least three different motivational structures in a contest of "chicken.
Brinkmanship: The Manipulation of Risk If "brinkmanship" means anything, it means manipulating the
shaved risk ofwar. It means exploiting the danger that some- body may inadvertently go over the brink, dragging the other with him. If two climbers are tied together, and one wants to intimidate the other by seeming about to fall over the edge, there has to be some uncertainty or anticipated irrationality or it won't work. If the brink is clearly marked and provides a firm footing, no loose pebbles underfoot and no gusts of wind to catch one off guard, if each climber is in full control of himself and never gets dizzy, neither can pose any risk to the other by approaching the brink. There is no danger in approaching it; and while either can deliberately jump off, he cannot credibly pretend that he is about to. Any attempt to intimi- date or to deter the other climber depends on the threat of slipping or stumbling. With loose ground, gusty winds, and a propensity toward dizziness, there is some danger when a climber approaches the edge; one can credibly threaten to fall off accidentally by standing near the brink.
Without uncertainty, deterrent threats of war would take the form of trip-wires. To incur commitment is to lay a trip-wire, one that is plainly visible, that cannot be stumbled on, and that is manifestly connected up to the machinery of war. And if effective, it works much like a physical barrier. The tripwire will not be crossed as long as it has not been placed in an in- tolerable location, and it will not be placed in an intolerable lo- cation as long as there is no uncertainty about each other's
sians could have rationally denied at the cost of general war.
just might, in spite of all their care and all our care -
the brink and over it in a general war, had some substance. If we were anywhere near the brink of war on that occasion, it was a war that neither side wanted but that both sides might have been unable to forestall.
The idea, expressed by some writers, that such deterrence depends on a "credible first strike capability," and that a coun- try cannot plausibly threaten to engage in a general war over anything but a mortal assault on itself unless it has an appre- ciable capacity to blunt the other side's attack, seems to depend
on the clean-cut notion that war results
- lead up to
? Yet their implicit threat to behave in a way that might
that
? result -
or is expected to
-
only from adeliberate yes- no decision. But if war tends
to result from aprocess, a dynamic process in which both sides get more and more deeply involved, more and more expectant, more and more concerned not to be a slow second in case the war starts, it is not a "credible first strike" that one threatens, but just plain war. The Soviet Union can indeed threaten us with war: they can even threaten us with a war that we eventu- ally start, by threatening to get involved with us in a process that blows up into war. And some of the arguments about "superiority" and "inferiority" seem to imply that one of the two sides, being weaker, must absolutely fear war and concede while the other, being stronger, may confidently expect the other
to yield. There is undoubtedly a good deal to the notion that the country with the less impressive military capability may be less feared, and the other may run the riskier course in a crisis; other things being equal, one anticipates that the strategically
? ? ? 100 ARMS AND INFLUENCE
THE ART OF COMMITMENT
motives and nothing at issue that is worth a war to both sides. Either side can stick its neck out, confident that the other will not chop it off. As long as the process is a series of discrete steps, taken deliberately, without any uncertainty as to the con- sequences, this process of military commitment and maneuver
101
never end in disaster. It could only terminate in disaster if one of the players made a deIiberate move that he knew would cause disaster, and he would not. Second, the possibility of dis- aster will be reflected in the players' tactics. White can effec- tively keep Black's queen on her own side of the board by gettingaknightacrossfirst;orhecankeepbothBlack'sknights on their own side by getting his queen across first. This ability to block or to deter certain moves of the adversary will be an im- portant part of the game; the threat of disaster will be effective, so effective that the disaster never occurs.
In fact, the result is no different from a rule that says no queen can cross a center line if an opponent's knight has al- ready crossed it, and no knight can cross the center line if an opponent's queen has already crossed it. Prohibitive penalties imposed on deliberate actions are equivalent to ordinary rules.
The characteristic that this chess game shares with the trip- wire diplomacy, and that accounts for its peculiar safety, is the absence of uncertainty. There is always some moment, or some final step, in which one side or the other has the last clear chance to turn the course of events away from war (or from disaster in our game of chess) or to turn it away from a political situation that would induce the other to take the final step to- ward war. The skillful chess player will keep the knight across the center line or near enough to cross before his opponent's queen can get across, with due allowance for the cost of having to devote resources to the purpose. Skillful diplomacy, in the absence of uncertainty, consists in arranging things so that it is one's opponent who is embarrassed by having the "last clear
chance" to avert disaster by turning aside or abstaining from what he wanted to do.
But off the chess board the last chance to avert disaster is not always clear. One does not always know what moves of his own would lead to disaster, one cannot always perceive the moves that the other side has already taken or has set afoot, or what interpretation will be put on one's own actions; one does not al- ways understand clearly what situations the other side would not, at some moment, accept in preference to war. When we
? would not lead to war. Imminent war -
- continually threatened, but the threats would work. They would work unless one side were pushed too far; but if the push-
would be ing side knows how far that is, it will not push that far.
The resulting world -
- discriminate in favor of passivity against initiative. It is easier to deter than to compel. Among a group of arthritics moving deli- cately and slowly at a cocktail party, no one can be dislodged from his position near the bar, or ousted from his favorite chair; bodily contact is equally painful to his assailant. By standing in the doorway, one can prevent the entrance or exit of another
the world without uncertainty
would
ailing guest who is unwilling to push his way painfully through. In fact, without uncertainty all the military threats and ma- neuvers would be like diplomacy with rigid rules and can be illustrated with amodified game of chess. A chess game can end in win, lose, or draw. Let's change the game by adding a fourth outcome called "disaster. " If "disaster" occurs, a heavy fine is levied on both players, so that each is worse off than if he had simply lost the game. And the rules specify what causes dis- aster: specifically,if either player has moved his knight across the
center line and the other player has moved his queen across the center line, the game terminates at once and both players are scored with a disaster. If a white knight is already on the black side of the board when the black queen moves across to the white side, the black queen's move terminates the game in dis- aster; if the queen was already across when White moved his knight across the center line, the knight's move terminates the game in disaster for both players. And the same applies for the
white queen and the black knight.
What does this new rule do to the way a game is played? If a
game is played well, and both players play for the best score they can get, we can state two observations. First, a game will
possible war
? ? 102 ARMS AND INFLUENCE
THE ART OF COMMITMENT 103
in Wlxte's favor; he created the pressure, but both are subject to the same risk. White's advantage is that he can back out more quickly, as we have set up the game in this example; even he cannot retreat, though, until Black has made his next move, and for the moment both have the same incentive to come to terms. (White'sabilitytoretreat,andBlack'sinability,mayseemmore of an advantage to White than it actually is; his ability to retreat is an ability to save both players, equally, from disaster. If no bargain is reached, the white knight has to return, because he is the only one who can. If Black can avoid entering any nego- tiation- can absent himself from the room or turn off his hear- ingaid- White's soleremainingobjectivewillbetogethisown knight back before he blows things up. ) If "disaster" is only somewhat worse, not drastically worse, than losing the chess game, the side that is losing may have more incentive to threaten disaster, or more immunity to the other's threat, and perhaps in consequence a stronger bargaining position. Note, in particular, that all of this has nothing to do with whether a knight is more or less potent than a queen in the chess game; queen and knight can be interchanged in the analysis of this paragraph. If the clash of a squad with a division can lead to unintended war, or of a protest marcher with an armed police-
man to an unwanted riot, their potencies are equal in respect of the threats that count.
In this way uncertainty imports tactics of intimidation into the game. One can incur a moderate probability of disaster, sharing it with his adversary, as a deterrent or compellent de- vice, where one could not take, or persuasively threaten to take, a deliberate last clear step into certain disaster. 2
2. To clarify the theoretical point it may be worth observing that the uncertainty and unpredictability need not arise from a genuine random mechanism like the dice. It is unpredictability, not "chance," that makes the difference; it could as well arise in the clumsiness of the players, some uncertainty about the rules of the game or the scoring system, bad visibility or moves made in secret, the need to commit certain moves invisibly in advance, meddling by a third party, or errors made by the referee. Dice are merely a convenient way to introduce unpredictability into an artificial example.
? add uncertainty to this artificial chess game we are not so sure that disaster will be avoided. More important, the risk of disas- ter becomes a manipulative element in the situation. It can be exploited to intimidate.
To see this, make one more change in the rules. Let us not have disaster occur automatically when queen and knight of op- posite color have crossed the center line. Instead, when that occurs, the referee rolls a die. If an ace comes up the game is over and both players are scored with disaster, but if any other number appears the play goes on. If after the next move the queen and knight are still across the center line the dice are rolled again, and so on.
This is a very different game. And not just because disaster may or may not occur when queen and knight get into those positions, instead of occurring with certainty. The difference is that now queen and knight may actually be moved into those positions. One can deliberately move his knight across the line in an attempt to make the queen retreat, if one thinks his ad- versary is less willing to incur a continuing risk of disaster, or thinks his adversary can be persuaded that oneself will not re- treat, and if the momentary risk of disaster is not prohibitive. In fact, getting one's knight across and blocking its return with one's own pieces, so that it clearly takes several moves to re- treat, may persuade the adversary that only he, by withdrawing his queen, can reduce the risk within a tolerable time.
? If the black queen cannot retreat- if her exit is blocked against timely retreat- the white knight's tactic to force her withdrawal is ineffectual and gratuitously risky. But it can pos- sibly serve another end (another risky one), namely, to enforce "negotiation. " By crossing over,once the queen has crossed and cannot readily return, the knight can threaten disaster; White can propose Black's surrender, or a stalemate, or the removal of a bishop or the sacrifice of a pawn. What he gets out of this is wide open; but what began as a chess game has been converted into a bargaining game. Both sides are under similar pressure to settle the game or at least to get the white knight out of mis- chief. The outcome, it should be noticed, will not necessarily be
? ? ? ? ? ? ? 104 ARMS AND INFLUENCE
THE ART OF COMMITMENT 105
The route by which major war might actually be reached would have the same kind of unpredictability. Either side can
engaging in a limited war would usually be such a
gains to the other side. The white knight is as potent as the black queen in creating a shared risk of d i ~ a s t e r . ~
Limited War as a Generator of Risk
Limited war, as a deterrent to continued aggression or as a compellent means of intimidation, often seems to require inter- pretation along these lines, as an action that enhances the risk of a greater war. The danger of major war is almost certainly increased by the occurrence of a limited war; it is almost cer- tainly increased by any enlargement in the scope or violence of a limited war that has already taken place. This being so, the threat to engage in limited war has two parts. One is the threat to inflict costs directly on the other side, in casualties, expendi- tures, loss of territory, loss of face, or anything else. The second is the threat to expose the other party, together with oneself, to a heightened risk of a larger war.
just where the fault,
? take steps
step -
-
that genuinely raise the probability of a blow-up. This
would be the case with intrusions, blockades, occupations of third areas, border incidents, enlargement of some small war, or any incident that involves a challenge and entails a response that may in turn have to be risky. Many of these actions and threats designed to pressure and intimidate would be nothing but noise, if it were reliably known that the situation could not get out of hand. They would neither impose risk nor demon-
strate willingness to incur risk. And if they definitely would lead to major war, they would not be taken. (If war were desired, it would be started directly. ) What makes them significant and
usable is that they create a genuine risk
appreciated -
a danger that can be
-
that the thing will blow up for reasons not fully
under contr01. ~
It has often been said, and correctly, that a general nuclear
war would not liberate Berlin and that local military action in the neighborhood of Berlin could be overcome by Soviet mili- tary forces. But that is not all there is to say. What local mili- tary forces can do, even against very superior forces, is to initi- ate this uncertain process of escalation. One does not have to be able to win a local military engagement to make the threat of it effective. Being able to lose a local war in a dangerous and pro-
Just how the major war would occur -
? ? vocative manner may make the risk
quences, but the possibility of this act
not the sure conse- outweighthe apparent
3. The purest real-life example I can think of in international affairs is "buzzing" an airplane, as in the Berlin air corridor or when a reconnaissance plane intrudes. The only danger is that of an uninterzded collision. The pilot who buzzes obviously wants no collision. (If he did, he could proceed to do it straightforwardly. ) The danger is that he may not avoid accident, through mishandling his aircraft, or misjudging distance, or failure to anticipate the movements of his victim. He has to fly close enough, or recklessly enough, to create an appreciated risk that he may- probably won't, but nevertheless may- fail in his mission and actually collide, to everyone's chagrin including his own.
- -
is not predictable. Whatever it is that makes limited war between great powers a risky thing, the risk is a genuine one that neither side can alto- gether dispel even if it wants to. T o engage in limited war is to start rocking the boat, to set in motion a process that is not al-
4. It may be worth pointing out that, though all attempts to deter or to compel by threat of violence may carry some risk, it is not a necessary character of deterrent threats that they be risky if they are, or try to be, of the full-commitment or trip- wire variety discussed in the preceding chapter. What can make them risky is that they may not work as hoped: they are risky because they may fail. Ideally they would cany no risk. It is part of the logical structure of the threats discussed in this chapter that they entail risk- the risk of being fulfilled--even though they work (or were about to work) as intended. One is risky the way driving a car is always risky: genuine accidents can always occur, no matter how well the car is designed or how carefully it is driven; risk is a fact of life. The other is risky the way certain forms of road- hogging are risky: a genuine risk is incurred, or created, or enhanced, for the purpose of intimidation, a risk that may not be altogether avoided if intimidation is successfully achieved because it may have to operate for a finite period before compliance brings relief. This risk is part of the price of intimidation.
initiative, or misunderstanding may occur -
? ? ? ? ? ? ? ? ? 106 ARMS AND INFLUENCE
THE ART OF COMMITMENT 107
breached. One may then try not to maximize the stability of new limits as one passes certain thresholds, but to pass them in a way that dramatizes and emphasizes that the engagement is a dangerous one and that the other side should be eager to call a halt. Deliberately raising the risk of all-out war is thus a tactic that may fit the context of limited war, particularly for the side most discontent with the progress of the war. Introduction of nuclear weapons undoubtedly needs to be evaluated in these terms.
Discussions of troop requirements and weaponry for NATO have been much concerned with the battlefield consequences of different troop strengths and nuclear doctrines. But the battle- field criterion is only one criterion, and when nuclear weapons are introduced it is secondary. The idea that European arma- ment should be designed for resisting Soviet invasion, and is to be judged solely by its ability to contain an attack, is based on the notion that limited war is a tactical operation. It is not.
What that notion overlooks is that a main consequence of limited war, and potentially a main purpose for engaging in it, is to raise the risk of larger war. Limited war does this whether it is intended to or not.
together in one's control. (In the metaphorical language of our chess game, it is to move a queen or a knight across the center line when the other knight or queen is already across, establish- ing a situation in which factors outside the players' control can determine whether or not the thing blows up. ) The risk has to be recognized, because limited war probably does raise the risk of a larger war whether it is intended to or not. It is a conse- quence of limited war that that risk goes up; since it is a consequence, it can also be a purpose.
If we give this interpretation to limited war, we can give a corresponding interpretation to enlargements, or threats of en- largement, of the war. The threat to introduce new weapons, perhaps nuclear weapons, into a limited war is not, according to this argument, to be judged solely according to the immediate military or political advantage, but also according to the delib- erate risk of still larger war that it poses. And we are led in this way to a new interpretation of the trip-wire. The analogy for limited war forces in Europe, or a blockade about Cuba, or troops for the defense of Quemoy, according to this argument,
is not a trip-wire that certainly detonates all-out war if it is in working order and fails altogether if it is not. We have some- thing more like a minefield, with explosives hidden at random; a mine may or may not blow up if somebody starts to traverse the field. The critical feature of the analogy, it should be empha- sized, is that whether or not one of the mines goes off is at least to some extent outside the control of both parties to the engage- ment.
This argument is pertinent to the question not only of wheth- er, but of how, to cross the boundaries in some limited war. If one can gently erode a boundary, easing across it with-out creating some new challenge or a dramatic bid for enemy reprisal, and if one finds the current bounds intolerable, that may be the way to do it if one wants the tactical advantages of relaxing a rule. But if the tactical advantages are unimpressive, one's purpose in enlarging some limited war may be to con- front the enemy with a heightened risk, to bring into question the possibility of finding new limits once a few have been
This point is fundamental to deterrence of anything other than all-out attack on ourselves. And it is fundamental to the
- would be a real danger and would obsess the strategic commands on both sides. This danger is enhanced in a crisis, particularly one involving military activity. It is en- hancedpartlybecauseofthesheerpreoccupationwithit. Andit is enhanced because alarms and incidents will be more frequent, and those who interpret alarms will be readier to act on them. This is also, to a large extent, the purpose of being prepared to fight a local war in Western Europe. The Soviet anticipation of the risks involved in a large-scale attack must include the danger that general war will result. If they underestimate the scale and duration of resistance and do atrack, a purpose of re- sisting is to confront them, day after day, with an appreciation
strategy of limited war. The danger of sudden large war
of
unpremeditated war -
? ? ? ? 108 ARMS AND INFLUENCE
THE ART OF COMMITMENT 109
weapons signal and dramatize this very danger- a danger that is self-aggravating in that the more the danger is recognized, the more likely are the decisions that cause war to occur. This argu- ment is neither for nor against the use of nuclear weapons, but for recognizing that this consequence of their use equals in im- portance- and could far transcend- their tactical battlefield accomplishments.
It is worth noting that this interpretation suggests that the threat of limited war may be potent even when there is little ex- pectation that one could win it.
It is our sheer inability to predict the consequences of our actions and to keep things under control, and the enemy's sim- ilar inability, that can intimidate the enemy (and, of course, us too). If we were in complete control of the consequen-ces and
that life is risky, and that pursuit of the original objective is not worth the risk. - -
This is distantly but only distantly related to the notion that we deter an attack limited to Europe by the announced threat of all-out war. It is different because the danger of war does not depend solely on whether the United States would coolly resolve to launch general war in response to a limited attack in Europe. The credibility of a massive American re- sponse is often depreciated: even in the event of the threatened loss of Europe the United States would not, it is sometimes said, respond to the fait accompli of a Soviet attack on Europe with
? anything as "suicidal" as general war. But that is a simple- minded notion of what makes general war cr-edible. What can
make it exceedingly credib-
the Chinese in the Far East
war can occur whether we intend it or not.
a war we could make no threat that did not depend on our ultimate willingness to
le to the Russians
and perhaps to is that the triggering of general
knew what would and what would not precip-itate war
General war does not depend on our coolly deciding to retali- ate punitively for the invasion of Western Europe after careful consideration of the material and spiritual arguments pro and con. General war could result because we or the Soviets launched it in the mistaken belief that it was already on, or in the mistaken or correct belief that, if we did not start it in- stantly,theothersidewould,Itdoesnotdependonfortitude: it can result from anticipation of the worse consequences of a war that, because of tardiness, the enemy initiates.
And the fear of war that deters the Soviet Union from an at- tack on Europe includes the fear of a general war that they initi- ate. Even if they were confident that they could act first, they would still have to consider the wisdom of an action that might, through forces substantially outside their control, oblige them to start general war.
If nuclear weapons are introduced, the sensed danger of gen- eral war will rise strikingly. Both sides will be conscious of this increased danger. This is partly a matter of sheer expectation; everybody is going to be more tense, and for good reason, once nuclear weapons are introduced. And national leaders will know that they are close to general war if only because nuclear
choose general war.
This is not an argument that "our side" can always win a war
of nerves. (The same analysis applies to "their side" too. ) It is a reminder that between the alternatives of unsuccessful local resistance on the one extreme, and the fruitless, terrifying, and probablyunacceptableandincrediblethreatofgeneralthermo- nuclear war on the other, there is a strategy of risky behavior, of deliberately creating a risk that we share with the enemy, a risk that is credible precisely because its consequences are not entirely within our own and the Soviets' control.
Nuclear Weaponsand the Enhancement of Risk
The introduction of nuclear weapons raises two issues here. One is the actual danger of general war; the other is the role of this danger in our strategy. On the danger itself, one has to guess how likely it is that a sizable nuclear war in Europe can persist, and for how long, without triggering general war. The danger appears great enough to make it unrealistic to expect a tactical nuclear war to "run its course. " Either the nuclear weapons wholly change the bargaining environment, the appreciation of
that we started or a war that the enemy started
? ? ? ? 110 ARMS AND INFLUENCE
THE ART OF COMMITMENT I I I
national leaders as much as anything that is going on in Europe itself. It is the strategic forces whose minute-by-minute behavior on each side will be the main intelligence preoccupation of the other side. 5
Limited and localized nuclear war is not, therefore, a "tacti- cal" war. However few the nuclears used, and however selec- tively they are used, their purpose should not be "tactical" be- cause their consequences will not be tactical. With nuclears, it
as become more than ever a war of risks and threats at the ighest strategic level. It is a war of nuclear bargaining.
There are some inferences for NATO planning. First, nuclear weapons should not be evaluated mainly in terms of what they could do on the battlefield: the decision to introduce them, the way to use them, the targets to use them on, the scale on which to use them, the timing with which to use them, and the com- munications to accompany their use should not be determined (or not mainly determined) by how they affect the tactical course of the local war. Much more important is what they do to the expectation of general war, and what rules or patterns of expectations about local use are created. It is much more a war of dares and challenges, of nerve, of threats and brinkmanship, once the nuclear threshold is passed. This is because the danger of general war, and the awareness of that danger, is lifted an order of magnitude by the psychological and military conse- quences of nuclear explosion.
5. This is why one of the arguments for delegating nuclear authority to theater commanders- as presented in the election campaign of 1964- made little sense. That was the argument that communications between the theater and the American command structure might fail at the moment nuclear weapons were urgently needed. But if the weapons were that urgently needed, especially in the European theater, there would surely be appreciable danger of general war, and to proceed without communicating would guarantee the absence of crucial communication with the Strategic Air Command, the Defense Intelligence Agency, North American Air Defense Command, military forces everywhere, civil defense authorities, and, of course, our diplomatic establishment. It could preclude a choice of what kind of nuclear war to initiate; it could catch the Americans by surprise, and might merely give warning to the Russians.
? risks, and the immediate objectives, and bring about some termination, truce, tranquilization, withdrawal, or pause; or else the local war very likely becomes swamped in a much bigger war. If these are the likely alternatives, we should not take too seriously a nuclear local war plan that goes to great lengths to carry the thing to its bitter end. There is a high probability that the war either will go down by an order of magnitude or go up by an order of magnitude, rather than run the tactical nuclear course that was planned for it.
More important is how we control, utilize, and react to a sud- den increase in the sensed danger of general war. It will be so important to manage this risk properly that the battlefield con- sequences of nuclear weapons may be of minor importance. The hour-by-hour tactical course of the war may not even be worth the attention of the top strategic leadership.
One can question whether we ought to use nuclear weapons deliberately to raise the risk of general war. But unless we are willing to do this, we should not introduce nuclear weapons against an adversary who has nuclear weapons on his side. This raising of risk is so much of the consequence of nuclear weapons that to focus our planning attention on the battlefield may be to ignore what should be getting our main attention (and what would, in the event, get it). Once nuclear weapons are introduced, it is not the same war any longer. The tactical objectives and considerations that governed the original war are no longer controlling. It is now a war of nuclear bargaining and
demonstration.
In a nuclear exchange, even if it nominally involves only the
use of "tactical" weapons against tactically important targets, there will be a conscious negotiating process between two very threatening enemies who are worried that the war will get out of hand. The life expectancy of the local war may be so short that neither side is primarily concerned with what happens on the ground within the next day or two. What each side is doing with its strategic forces would be the main preoccupation. It is the strategic forces in the background that provide the risks and thesenseofdanger;itistheywhosedisposition willpreoccupy
? ? ? ? ? I12 ARMS AND INFLUENCE
Second, as a corollary we should not think that the value or likely success of NATO armed forces depends solely, or even mainly, on whether they can win a local war. Particularly if nu- clears are introduced, the war may never run its course. Even without the introduction of nuclears, a main function of re- sistance forces is to create and prolong a genuine sense of danger, of the potentiality of general war. This is not a danger that we create for the Russians and avoid ourselves; it is a danger we share with them. But it is this deterrent and intimida- tion function that deserves at least as much attention as the tactical military potentialities of the troops.
Third, forces that might seem to be quite "inadequate" by ordinary tactical standards can serve a purpose, particularly if they can threaten to keep the situation in turmoil for some pe- riod of time. The important thing is to preclude a quick, clean Soviet victory that quiets things down in short order.
Fourth, the deployment and equipment of nuclear-armed NATO troops, including the questions of which nationalities have nuclear weapons and which services have them, are affected by the purpose and function and character of nuclear and local war. If what is required is a skillful and well- controlled bargaining use of nuclears in the eventthe decision is taken to go above that threshold, and if the main purpose of nu- clears is not to help the troops on the battlefield, it is much less necessary to decentralize nuclear weapons and decisions to local commanders. The strategy will need tight centralized control; it may not require the kind of close battlefield support that is
often taken to justify distribution of small nuclears to the troops; and nuclears probably could be reserved to some special nuclear forces.
Fifth, if the main consequence of nuclear weapons, and the purpose of introducing them, is to create and signal a height- ened risk of general war, our plans should reflect that purpose.
THE ART OF COMMITMENT 113
veys to the Soviet leadership. Targets should be picked with a view to what the Soviet leadership perceives about the character of the war and about our intent, not for tactical importance. A target near or inside the U. S. S. R. , for example, is important be- cause it is near or inside the U. S. S. R. , not because of its tactical contribution to the European battlefield. A target in a city is important because a city is destroyed, not because it is a local supply or communication center. The difference between one weapon, a dozen, a hundred, or a thousand is not in the number of targets destroyed but in the Soviet (and American) percep- tion of risks, intent, precedent, and implied "proposal" for the conduct or termination of war.
Extra targets destroyed by additional weapons are not a local military "bonus. " They are noise that may drown the message. They are a "proposal" that must be responded to. And they are an added catalyst to general war. This is an argument for a se- lective and threatening use of nuclears rather than large-scale tactical use. (It is an argument for large-scale tactical use only if such use created the level of risk we wish to create. ) Success in the use of nuclears will be measured not by the targets de- stroyed but by how well we manage the level of risk. The So- viets must be persuaded that the war is getting out of hand but is not yet beyond the point of no return.
Sixth, we have to expect the Soviets to pursue their own policy of exploiting the risk of war. W e cannot expect the Sovi- ets to acquiesce in our unilateral nuclear demonstration. We have to be prepared to interpret and to respond to a Soviet nu- clear "counterproposal. " Finding a way to terminate will be as important as choosing how to initiate such an exchange. (We should not take wholly for granted that the initiation would be ours. )
Finally, the emphasis here is that the use of nuclear weapons would create exceptional danger. This is not an argument in favor of their use; it is an argument for recognizing that danger is the central feature of their use.
In other words, nuclears would not only destroy targets but would signal something. Getting the right signal across would be
? ? ? ? ? ? We should plan -
-
a war of nerve, of demonstration, and of bargaining, not just target destruction for local tactical purposes. Destroying a target may be incidental to the message that the detonation con-
in the event of resort to nuclear weapons
for
? ? ? 114 ARMS AND 1NFLUENCE
an important part of the policy. This could imply, for example, deliberate and restrained use earlier than might otherwise seem tactically warranted, in order to leave the Soviets under no illu- sion whether or not the engagement might become nuclear. The only question then would be, how nuclear. It is not necessarily prudent to wait until the last desperate moment in a losing en- gagement to introduce nuclear weapons as a last resort. By the time they are desperately needed to prevent a debacle, it may be too late to use them carefully, discriminatingly, with a view to the message that is communicated, and with the maintenance
of adequate control. Whenever the tactical situation indicates a high likelihood of military necessity for nuclears in the near fu- ture, it may be prudent to introduce them deliberately while there is still opportunity to do so with care, selection, and a properly associated diplomacy. Waiting beyond that point may simply increase the likelihood of a tactical use, possibly an in- discriminate use, certainly a decentralized use, determined by the tactical necessities of the battlefield rather than the strategic necessities of deterrence.
In its extreme form the restrained, signaling, intimidating use of nuclears for brinkmanship has sometimes been called the "shotacrossthebow. "Thereisalwaysadanger- Churchilland othershavewarnedagainstit-of makingabolddemonstration on so small a scale that the contrary of boldness is demon- strated. There is no cheap, safe way of using nuclears that scares the wits out of the Russians without scaring us too. Neverthe- less, any use of nuclears is going to change the pattern of expectations about the war. It is going to rip a tradition of
inhibition on their use. It is going to change everyone's expecta- tions about the future use of nuclears. Even those who have arguedthatnuclearsoughttobeconsideredjustamoreefficient kind of artillery will surely catch their breath when the first one goes off in anger. Something is destroyed, even if not enemy targets, if ever-so-few nuclears are used. Whatever a few nu- clears prove, or fail to prove about their user, they will change the environment of expectations. And it is expectations more
THE ART OF COMMlTMENT 115
than anything else that will determine the outcome of a limited East- West military engagement.
It is sometimes argued, quite correctly, that this tradition can be eroded, and the danger of "first use" reduced, by introducing nuclear weapons in some "safe" fashion, gradually getting the world used to nuclear weapons and dissipating the drama of nuclear explosions. Nuclear depth charges at sea, small nuclear warheads in air-to-air combat, or nuclear demolitions on de- fended soil may seem comparativelyfree of the danger of unlim- itedescalation,causenomorecivildisruptionthanTNT,appear responsible, and set new traditions for actual use, includ- ing the tradition that nuclear weapons can be used without sig- nalingall-outwar. Obviouslytoexploitthisideaoneshouldnot wait until nuclear weapons are desperately needed in a serious crisis, but deliberately initiate them in a carefully controlled fashion at a time and place chosen for the purpose. It might not be wise and might not be practical, but if the intent is to remove the curse from nuclear weapons, this may be the way to do it.
Among the several objections there is one that may be over- looked even by the proponents of nuclear "legitimization. " That is the waste involved- the waste of what is potentially the most dramatic military event since Pearl Harbor. President Johnson,remember,referredtoanineteen-yeartraditionofnon- use; the breaking of that tradition (which grows longer with each passing year) will probably be, especially if it is designed to be, a most stunning event. It will signal a watershed in mili- tary history, will instantly contradict war plans and military ex- pectations, wilf generate suspense and apprehension, and will probably startle even those who make the decision. The first post-Nagasaki detonation in combat will probably be evi- dence of a complex and anguished decision, an embarkation on ajourney into a new era of uncertainty. Even those who propose readier use of nuclear weapons must appreciate that this is so, because of the strong inhibitions they encounter during the dispute.
This is not an event to be squandered on an unworthy mili-
? ? ? ? ? ? ? ? ? ? 116 ARMS AND INFLUENCE
THE ART OF COMMITMENT I17
before they were old enough to drive and before automobiles were invented. The earliest instance I have come across, in a race with horse-drawn vehicles, antedates the auto by some time:
The road here led through a gully, and in one part the win- ter flood had broken down part of the road and made a hol-low. Menelaos was driving in the middle of the road, hoping that no one would try to pass too close to his wheel, but Antilochos turned his horses out of the track and followed him a little to one side. This frightened Menelaos, and he shouted at him:
"What reckless driving Antilochos! Hold in your horses. This place is narrow, soon you will have more room to pass. You will foul my car and destroy us both! "
But Antilochos only plied the whip and drove faster than ever, as if he did not hear. They raced about as far as the cast of quoit . . . and then [Menelaos] fell behind: he let the horses go slow himself, for he was afraid that they might all collide in that narrow space and overturn the cars and fall in a struggling heap.
This game of chicken tookplace outside the gates of Troy three thousand years ago. Antilochos won, though Homer says
? tary objective. The first nuclear detonation can convey a mes- sage of utmost seriousness; it may be a unique means of communicationinamomentofunusualgravity. Todegradethe signalinadvance,todepreciatethecurrency,toerodegradually a tradition that might someday be shattered with diplomatic effect, to vulgarize weapons that have acquired a transcendent status, and to demote nuclear weapons to the status of merely efficient artillery, may be to waste an enormous asset of last re- sort. One can probably not, with effect, throw down a gauntlet if he is known to toss his gloves about on every provocation. One may reasonably choose to vulgarize nuclear weapons through a campaign to get people used to them; but to proceed to use them out of expediency, just because they would be tactically advantageous and without regard to whether they
ought to be cheapened, would be shortsighted in the extreme.
Face,Nerve,and Expectations
? Cold war politics have been likened, by Bertrand Russell and others, to the game of "chicken. " This is described as a game in which two teen-age motorists head for each other on a highway - usuallylate at night, with their gangs and girlfriends looking
? -
who does is then called "chicken. "
on
to see which of the two will first swerve aside. The one
The better analogy is with the less frivolous contest of chicken that is played out regularly on streets and highways by people who want their share of the road, or more than their share, or who want to be first through an intersection or at least not kept waiting indefinitely.
"Chicken" is notjust a game played by delinquent teen-agers with their hot-rods in southern California; it is a universal form of adversary engagement. It is played not only in the Berlin air corridor but by Negroes who want to get their children into schools and by whites who want to keep them out; by rivals at a meeting who both raise their voices, each hoping the other will yield the floor to avoid embarrassment; as well as by drivers of both sexes and all ages at all times of day. Children played it
somewhatungenerously
"by trick, not by merit. "
-
-Even the game in its stylized teen-age automobile form is
worth examining. Most noteworthy is that the game virtually disappears if there is no uncertainty, no unpredictability. If the two cars, instead of driving continuously, took turns advancing exactly fifty feet at a time toward each other, a point would be reached when the next move would surely result in collision. Whichever driver has that final turn will not, and need not, drive deliberately into the other. This is no game of nerve. The lady who pushes her child's stroller across an intersection in front of a car that has already come to a dead stop is in no par- ticular danger as long as she sees the driver watching her: even
6 The Iliad, W H D Rouse, trand (Mentor Books, 19501, p 273
? ? ? 118 ARMS AND INFLUENCE
THE ART OF COMMITMENT I19
? ? if the driver prefers not to give her the right of way she has the winning tactic and gets no score on nerve. The more instruc- tive automobile form of the game is the one people play as they crowd each other on thehighway,jockey their way through an intersection, or speed up to signal to a pedestrian that he'd better not cross yet. These are the cases in which, like Antil- ochos' chariot, things may get out of control; no one can trust with certainty that someone will have the "last clear chance" to
Another important characteristic is that, though the two play- ers are cast as adversaries, the game is somewhat collaborative. Even in the stylized version in which they straddle the white line, there is at least an advantage in understanding that, when a player does swerve, he will swerve to the right and not to the left! And the players may try to signal each other to try to co- ordinate on a tie; if each can swerve a little, indicating that he will swerve a little more if the other does too, and if their speeds are not too great to allow some bargaining, they may manage to turn at approximately the same time, neither being proved chicken.
They may also collaborate in declining to play the game. This is a little harder. When two rivals are coaxed by their friends to have it out in a fight, they may manage to shrug it off skillfully,
? avert tragedy and will pull back in time.
These various games of chicken- the genuine ones that in-
volve some real unpredictability- have some characteristics that are worth noting. One is that, unlike those sociable games it takes two to play, with chicken it takes two not to play. If you are publicly invited to play chicken and say you would rather not, you have just played.
at stake but reputations, expectations, and precedents. That is, accommodation or obstinacy, boldness or surrender, merely establishes who is an accommodator, who is obstinate or bold, who tends to surrender or what order of precedence is to be observed. A second, not easily distinguished in practice, occurs when something is consciously put at stake (as in a gambling game or trial by ordeal) such as leadership, deference, popularity, some agreed tangible prize, or the outcome of certain issues in dispute. (The duel between David and Goliath, mentioned in the note on page 144, is an example of putting something at stake. ) The third, which might be called the "real" in contrast to the "conventional," is the case in which yielding or withdrawing yields something that the dispute is about, as in road- hogging or military probes: that is, the gains and losses are part of the immediate structure of the contest, not attached by convention nor resulting entirely from expectations established for future events. The process of putting something at
? Second, what is in dispute is usually not the issue of the mo- ment, but everyone's expectations about how a participant will behave in the future. To yield may be to signal that one can be expected to yield; to yield often or continually indicates acknowl- edgment that that is one's role. To yield repeatedly up to some limit and then to say "enough" may guarantee that the first show of obduracy loses the game for both sides. If you can get a rep- utationforbeingreckless,demanding,orunreliable- and appar- ently hot-rods, taxis, and cars with "driving school" license plates sometimes enjoy this advantage- you may find conces-
sions made to you. (The driver of a wide American car on a narrow European street is at less of a disadvantage than a static calculation would indicate. The smaller cars squeeze over to give him room. ) Between these extremes, one can get a reputa- tion for being firm in demanding an appropriate share of the road but not aggressively challenging about the other's half. Un- fortunately, in less stylized games than the highway version, it is often hard to know just where the central or fair or expected division should lie, or even whether there should be any recog- nition of one contestant's claim. 7
7. Analytically there appear to be at least three different motivational structures in a contest of "chicken.
